Simulations of Watershed Response to Land Use and Climate Change 
in the Saugahatchee Creek Watershed using the WARMF Model 
 
by 
 
Sushban Shrestha 
 
 
 
 
A thesis submitted to the Graduate Faculty of 
Auburn University 
in partial fulfillment of the 
requirements for the Degree of 
Master of Science 
 
Auburn, Alabama 
August 6, 2011 
 
 
 
 
Keywords: watershed, hydrology, water quality,  
land use, climate change, WARMF 
 
 
Copyright 2011 by Sushban Shrestha 
 
 
Approved by 
 
Xing Fang, Chair, Associate Professor of Civil Engineering 
Jose G. Vasconcelos, Assistant Professor of Civil Engineering 
Luke J. Marzen, Associate Professor of Geography
ii 
Abstract 
 
 
During recent years, land surfaces have been markedly transformed with artificial 
land cover limiting natural forests, accompanied by urbanization. The atmospheric 
concentration of greenhouse gases is believed to be increasing, thereby leading to 
anthropogenic climate change. It is very important, from watershed management point of 
view, to know how these alterations would affect water resources. In this study, we have 
tested the ability of a watershed model WARMF to simulate flow and water quality in the 
Saugahatchee Creek Watershed in Alabama and applied the developed model to assess 
the impact due to historical land use change and potential future climate change. Surface 
water temperature, dissolved oxygen, total phosphorus, total nitrogen, and chlorophyll-a 
concentrations were simulated along with flow at the watershed outlet. 
Based on model simulation, historical land use changes from 1991 to 2008 show 
rising pattern in nutrient levels and algal mass depleting water quality in the 
Saugahatchee Creek.  Future climate for the 21
st
 century, derived from HadCM3 A2 and 
B2 emission scenarios, were downscaled to local watershed scale for impact analysis.  
Surface water temperature is projected to increase, mostly in summer and dissolved 
oxygen concentration is projected to decrease. Both HadCM3 scenarios? output predicted 
decrease in flow in the 21
st
 century. Nutrient concentration increased corresponding to 
low flow. These results under different land use and climate change scenarios can be 
useful information for watershed planning and management decision.  
iii 
Acknowledgement 
 
 
Two years have passed since I joined MS program at Auburn University. As I 
approach towards receiving my Master?s Degree, I hereby take this opportunity to share 
my earnest gratitude to one and all for being part of this wonderful journey.  
First and foremost, I owe my deepest gratitude to my advisor, Dr. Xing Fang, for 
his enthusiasm, inspiration, and unique ability to motivate. His mentorship has not only 
guided me during my graduate study, but also, to a greater extent, has sharpened the 
career path I am willing to follow. I would also like to extend my appreciation to my 
committee members Dr. Jose Vasconcelos and Dr. Luke Marzen, for their valuable 
insights and suggestions. I wish to especially thank Joel Herr of Systech for sharing his 
expertise in WARMF; Rajesh Sawant, Mac Martin, and Dr. Luke Marzen of 
AlabamaView for providing land cover data; Eric Reutebuch and Dr. William Deutsch of 
Alabama Water Watch for sharing water quality data. Thanks are due to Prof. Junqi Li 
for his valuable inputs to collaborated research work. 
I do not want to miss this opportunity to thank all the colleagues in my 
department, who have contributed immensely to my professional and personal life at 
Auburn. A sincere thank to this great source of knowledge and friendship. Last but not 
least, I would like to thank my family and friends for their unwavering support 
throughout my education and life. 
iv 
Table of Contents 
 
 
Abstract ............................................................................................................................... ii?
Acknowledgement ............................................................................................................. iii?
List of Tables .......................................................................................................................v?
List of Figures .................................................................................................................... ix?
List of Abbreviations ....................................................................................................... xiii?
Chapter 1. Introduction ........................................................................................................1?
Chapter 2. Development of WARMF Model to Study Hydrology and Water 
Quality in Streams and Reservoirs of the Saugahatchee Creek Watershed .............7?
Chapter 3. Assessing the Impact on Hydrology and Water Quality in the 
Saugahatchee Creek due to Land Use Change and Climate Change using the 
WARMF Model .....................................................................................................50?
Chapter 4. Rainfall Depths of the 95
th
 Percentile Events in the Contiguous U.S. .............92?
Chapter 5. Conclusions and Recommendation ................................................................130?
Appendix A. Downscaling Precipitation Using SDSM 4.2 .............................................136?
Appendix B. Land Use Change Impact: Statistical Summary .........................................141?
Appendix C. Climate Change Impact: Statistical Summary ............................................147?
 
 
v 
List of Tables 
 
 
Table 2.1 2008 ADEM 303(d) List of Impaired Waters in the Saugahatchee Creek 
Watershed ..............................................................................................................34?
Table 2.2 Statistical Summary for Areas of 44 Catchments and Lengths of 40 
Stream Segments of the Saugahatchee Creek Watershed Imported to 
WARMF Model .....................................................................................................34?
Table 2.3 Monthly Average, Standard Deviation, Maximum, and Minimum Values 
of Meteorological Variables for 1997-2009 Period Imported into WARMF 
Model for the Saugahatchee Creek Watershed ......................................................35?
Table 2.4 Land Use Distribution in Percentage for Three Catchments Containing 
Flow/Water Quality Monitoring Stations in the Saugahatchee Creek 
Watershed ..............................................................................................................35?
Table 2.5 Calibrated Parameters of WARMF Model for the Saugahatchee Creek 
Watershed ..............................................................................................................36?
Table 2.6 List of Major Point Source Dischargers in the Saugahatchee Creek 
Watershed ..............................................................................................................36?
Table 2.7 Model Performance for Flow Simulation during Calibration and 
Validation Period ...................................................................................................36?
Table 3.1 Key Parameter Values after WARMF Calibration ............................................76?
Table 3.2 WARMF Performance during Calibration and Validation Periods for 
Daily Flow at USGS 02418230 .............................................................................76?
Table 3.3 Land Use Change in the Saugahatchee Creek Watershed from 1991 to 
2008........................................................................................................................76?
Table 3.4 List of Predictands (Station Climate Parameters) and Corresponding 
Predictors used in SDSM Model to Downscale GCMs Output .............................77?
Table 3.5 Statistical Summary of Maximum Temperature, Minimum Temperature, 
and Precipitation Downscaled Based on HadCM3 A2 and B2..............................77?
vi 
Table 4.1 The 85
th
. 90
th
, 95
th
 Percentile, and 90% Cumulative Rainfall Depths (in.) 
Derived Using Daily Rainfall Data for 206 Weather Stations or Cities in the 
Contiguous U.S. ...................................................................................................113?
Table 4.2 The 95
th
 Percentile Depths (in.) from This Study and Hirschman and 
Kosco (2008) and Their Absolute and Relative Differences (%) in 20 U.S. 
Cities ....................................................................................................................117?
Table 4.3 Storm Depth Mean and Kappa Distribution Paramters for 18 Selected 
Stations in Eastern New Mexico, Oklahoma and Texas ......................................118?
Table 4.4 The 95
th
 Percentile Storm Depths Determined Using Kappa Distribution 
Parameters Derived from Hourly Rainfall Data and Differences between 
95
th
 Percentile Rainfall Depths Estimated from Daily and Hourly Rainfall 
Data ......................................................................................................................119?
Table 4.5 Comparison between NOAA's 1 Year 24-Hour Rainfall and Computed 
95
th
 Percentile Rainfall Depth Using Daily Data .................................................120?
Table 4.6 Statistical Summary of 90
th
 and 95
th
 Percentile Daily Rainfall Depths 
(in.) and the Rainfall Depths that can Capture the Runoff from 90% of 
Average Annual Rainfall at 206 Weather Stations in the Contiguous U.S. ........121?
Table B.1 Monthly Average of Daily Flow and Standard Dev. for Land Use 
Scenarios ..............................................................................................................141?
Table B.2 Relative Change in Monthly Average of Daily Flow from the Baseline ........141?
Table B.3 Monthly Average of Daily Water Temperature and Standard Deviation 
for Land Use Scenarios ........................................................................................142?
Table B.4 Relative Change in Monthly Average of Daily Water Temperature from 
the Baseline ..........................................................................................................142?
Table B.5 Monthly Average of Daily Surface DO and Standard Deviation for Land 
Use Scenarios .......................................................................................................143?
Table B.6 Relative Change in Monthly Average of Daily Surface DO from the 
Baseline ................................................................................................................143?
Table B.7 Monthly Average of Daily TP and Standard Deviation for Land Use 
Scenarios ..............................................................................................................144?
Table B.8 Relative Change in Monthly Average of Daily TP from the Baseline ...........144?
vii 
Table B.9 Monthly Average of Daily TN and Standard Deviation for Land Use 
Scenarios ..............................................................................................................145?
Table B.10 Relative Change in Monthly Average of Daily TN from the Baseline .........145?
Table B.11 Monthly Average of Daily Chlorophyll- a Concentration and Standard 
Deviation for Land Use Scenarios .......................................................................146?
Table B.12 Relative Change in Monthly Average of Daily Chlorophyll-a 
Concentration from the Baseline .........................................................................146?
Table C.1 Monthly Average of Daily Flow (m
3
/s) ..........................................................147?
Table C.2 Standard Deviation of Daily Flow (m
3
/s) ........................................................147?
Table C.3 Relative Change in Monthly Average of Daily Flow from the Baseline 
(m
3
/s) ....................................................................................................................148?
Table C.4 Relative Change in Monthly Average of Daily Flow from the Baseline 
(%)........................................................................................................................148?
Table C.5 Monthly Average of Daily Water Temperature (?C) ......................................149?
Table C.6 Standard Deviation of Daily Water Temperature (?C) ...................................149?
Table C.7 Relative Change in Monthly Average of Daily Water Temperature (?C) ......150?
Table C.8 Relative Change in Monthly Average of Daily Water Temperature (%) .......150?
Table C.9 Monthly Average of Daily Surface DO (mg/l) ...............................................151?
Table C.10 Standard Deviation of Daily Surface DO (mg/l) ...........................................151?
Table C.11 Relative Change in Monthly Average of Daily Surface DO (mg/l) ..............152?
Table C.12 Relative Change in Monthly Average of Daily Surface DO (%) ..................152?
Table C.13  Monthly Average of Daily TP concentration (mg/l) ....................................153?
Table C.14 Standard Deviation of Daily TP concentration (mg/l) ..................................153?
Table C.15 Relative Change in Monthly Average of Daily TP concentration (mg/l) .....154?
Table C.16 Relative Change in Monthly Average of Daily TP concentration (%) .........154?
Table C.17 Monthly Average of Daily TN concentration (mg/l) ....................................155?
viii 
Table C.18 Standard Deviation of Daily TN concentration (mg/l) .................................155?
Table C.19 Relative Change in Monthly Average of Daily TN concentration (mg/l) ....156?
Table C.20 Relative Change in Monthly Average of Daily TN concentration (%) ........156?
Table C.21 Monthly Average of Daily Chlorophyll-a (?g/l) ...........................................157?
Table C.22 Standard Deviation of Daily Chlorophyll-a (?g/l) ........................................157?
Table C.23 Relative Change in Monthly Average of Daily Chlorophyll-a (?g/l) ...........158?
Table C.24 Relative Change in Monthly Average of Daily Chlorophyll-a (%) ..............158?
 
ix 
List of Figures 
 
 
Fig. 2.1 Location map of the Saugahatchee Creek Watershed including surrounding 
counties and locations of three flow and water quality stations ............................37?
Fig. 2.2 Definition sketch for the compartments of a catchment in WARMF model 
(Chen et al. 2001) ...................................................................................................38?
Fig. 2.3 Selection of the Saugahatchee Creek Watershed from the Lower 
Tallapoosa Watershed (HUC 03150110) in BASINS ...........................................38?
Fig. 2.4 Land catchments, stream segments, and reservoirs of the Saugahatchee 
Creek Watershed imported into WARMF including three major point 
source dischargers (Table 2.6) ...............................................................................39?
Fig. 2.5 NLCD 2001 land use of the Saugahatchee Creek Watershed clipped from 
Tallapoosa Basin ....................................................................................................40?
Fig. 2.6 Observed phosphate and nitrate loadings in 2001, 2002 and 90
th
 percentile 
for Auburn Northside WWTP, Opelika Westside WWTP, and West Point 
Stevens Finishing Plant, respectively ....................................................................41?
Fig. 2.7 Monthly summary of observed daily flow from 2000-2009 at USGS 
02418230 station in the Saugahatchee Creek near Loachapoka ............................42?
Fig. 2.8 Flow calibration (2000-05) and validation (2006-09) at USGS 02418230 
station in the Saugahatchee Creek near Loachapoka .............................................43?
Fig. 2.9 Observed and modeled water temperature, DO, TP, and TN concentration 
during calibration and validation period at Station-16 in the Pepperell 
Branch ....................................................................................................................44?
Fig. 2.10 Observed and modeled water temperature, DO, TP, and TN concentration 
during calibration and validation period at Station-8 in the Saugahatchee 
Creek near Loachapoka ..........................................................................................45?
Fig. 2.11 Observed and modeled chlorophyll-a, DO, and TP concentration during 
calibration and validation period at Yates-2 station in the Saugahatchee 
Creek ( Yates Reservoir Embayment) ...................................................................46?
x 
Fig. 2.12 Chlorophyll-a concentration in Yates Reservoir Embayment modeled as a 
reservoir and as a stream segment with observed values (2000-02) ......................47?
Fig. 2.13 Daily standard deviation of dissolved oxygen simulated using hourly time 
step at Yates Reservoir Embayment for the year 2000 and 2002 ..........................48?
Fig. 2.14 Diurnal variation of DO simulated using hourly time step for days with 
observed DO less than 5 mg/l at Yates Reservoir Embayment .............................49?
Fig. 2.15 Vertical profiles of DO for days with observed DO less than 5 mg/l ................49?
Fig. 3.1 Location map of the Saugahatchee Creek Watershed in the Tallapoosa 
Basin including surrounding counties in Alabama ................................................78?
Fig. 3.2 2001 NLCD land use map of the Saugahatchee Creek Watershed ......................79?
Fig. 3.3 Flow calibration (2000-05) and validation (2006-09) at USGS 02418230 
station in the Saugahatchee Creek near Loachapoka .............................................80?
Fig. 3.4 Water quality calibration (2000-01) and validation (2002) at Station-8 near 
Loachapoka ............................................................................................................81?
Fig. 3.5 Land use change scenarios for the Saugahatchee Creek Watershed from 
1991 to 2008 ..........................................................................................................82?
Fig. 3.6 General trend in maximum temperature, minimum temperature, and 
precipitation corresponding to downscaled climate change scenario based 
on HadCM3 A2 ......................................................................................................83?
Fig. 3.7 General trend in maximum temperature, minimum temperature, and 
precipitation corresponding to downscaled climate change scenario based 
on HadCM3 B2 ......................................................................................................84?
Fig. 3.8 Anomaly of average monthly flow, surface water temperature, and DO 
oxygen concentration to the baseline corresponding to land use scenarios of 
2001 and 2008 ........................................................................................................85?
Fig. 3.9 Anomaly of average monthly TP, TN, and chlorophyll-a concentration to 
the baseline corresponding to land use change scenarios of 2001 and 2008 .........86?
Fig. 3.10 A Anomaly of average monthly flow, surface water temperature and DO 
concentration to the baseline corresponding to climate change scenarios 
downscaled with HadCM3 A2 ...............................................................................87?
xi 
Fig. 3.11 Anomaly of average monthly TP, TN and chlorophyll-a concentration to 
the baseline corresponding to climate change scenario downscaled with 
HadCM3 A2 ...........................................................................................................88?
Fig. 3.12 Anomaly of average monthly flow, surface water temperature and DO 
concentration to the baseline corresponding to climate change scenario 
downscaled with HadCM3 B2 ...............................................................................89?
Fig. 3.13 Anomaly of average monthly TP, TN and chlorophyll-a concentration to 
the baseline corresponding to climate change scenario downscaled with 
HadCM3 B2 ...........................................................................................................90?
Fig. 3.14 Flow duration curves for baseline and projected HadCM3 A2 future 
scenarios .................................................................................................................91?
Fig. 3.15 Flow duration curves for baseline and projected HadCM3 B2 future 
scenarios .................................................................................................................91?
Fig. 4.1. Percentile distribution of daily rainfall depths for Montgomery and 
Minneapolis weather stations ...............................................................................122?
Fig. 4.2. 95
th
 percentile rainfall depths estimated using 10, 20, and 30 years of daily 
rainfall data at Montgomery, AL (top) and Minneapolis, MN (bottom) for 
sensitivity analysis ...............................................................................................123?
Fig. 4.3. 95
th
 percentile rainfall map for the contiguous U.S. ..........................................124?
Fig. 4.4. 90
th
 percentile rainfall map for the contiguous U.S. ..........................................125?
Fig. 4.5. 85
th
 percentile rainfall map for the contiguous U.S. ..........................................126?
Fig. 4.6. Distribution of daily rainfall depths and quantiles calculated using Kappa 
distribution with paramters derived from daily data at Abilene Regional 
Airport, Texas ......................................................................................................127?
Fig. 4.7. Regression equations between 95
th
  percentile rainfall depth derived from 
daily rainfall data and 95th percentile rainfall depth derived from hourly 
rainfall data  (left) and NOAA's 1 year 24-hr rainfall (right) for selected 18 
stations (Table 4.4 and Table 4.5) ........................................................................128?
Fig. 4.8 Percentile distribution of cumulative rainfall depths for Montgomery and 
Minneapolis weather stations ...............................................................................129?
Fig. A.1 Correlation between observed precipitation and NCEP predictors for each 
month ...................................................................................................................136?
xii 
Fig. A.2 Correlation matrix and partial correlations between observed precipitation 
and NCEP predictors............................................................................................137?
Fig. A.3 Calibration result for precipitation with selected predictors (p5zh, r500, 
and r850) ..............................................................................................................138?
Fig. A.4 Parameter file generated by SDSM for downscaling precipitation ...................139?
Fig. A.5 Mean monthly precipitation for observed and downscaled results for 
validation..............................................................................................................140?
 
 
xiii 
List of Abbreviations 
 
 
ADEM Alabama Department of Environmental Management 
ASCE  American Society of Civil Engineers 
BASINS Better Assessment Science Integrating Point and Nonpoint Sources 
BMP  Best Management Practice 
CASTNET Clean Air Status and Trends Network 
CCCSN Canadian Climate Change Scenarios Network 
CSTR  Continuously Stirred Tank Reactor 
DEM  Digital Elevation Map 
DO  Dissolved Oxygen 
EISA  Energy Independence Security Act 
GCM  General Circulation Model 
GHG  Greenhouse Gas 
GI  Green Infrastructure 
GIS  Geographic Information System 
GSOD  Global Summary of Day 
HadCM3 Hadley centre Coupled Model, version 3 
HSPF  Hydrological Simulation Program - Fortran 
HUC  Hydrologic Unit Codes 
IPCC  Intergovernmental Panel on Climate Change 
xiv 
LAI  Leaf Area Index 
LID  Low Impact Development 
METF  Maximum Extent Technically Feasible 
NADP  National Atmospheric Deposition Program 
NCAR  National Center for Atmospheric Research 
NCDC  National Climatic Data Center 
NCEP  National Centers for Environmental Prediction 
NHD  National Hydrography Dataset 
NLCD  National Land Cover Dataset 
NOAA  National Oceanic and Atmospheric Administration 
NRCS  National Resources Conservation Service 
NSE  Nash-Sutcliffe?s Efficiency 
PBIAS  Percent Bias 
RSR  RMSE - Standard Deviation Ratio 
SCW  Saugahatchee Creek Watershed 
SDSM  Statistical Downscaling Model 
SRES  Special Report on Emission Scenarios 
SSURGO Soil Survey Geographic Database 
SWAT  Soil and Water Assessment Tool 
TMDL  Total Maximum Daily Load 
TN  Total Nitrogen 
TP  Total Phosphorus 
USDA  United States Department of Agriculture 
xv 
USEPA United States Environmental Protection Agency 
USGS  United States Geological Survey 
WARMF Watershed Analysis Risk Management Framework 
WQV  Water Quality Volume 
WWTP Waste Water Treatment Plant 
1 
Chapter 1.  Introduction 
 
Background 
A watershed, also referred as drainage basin or catchment, is the area of land 
where all of the water that is under it or drains off of it goes into a common waterway, 
such as a stream, reservoir, estuary, or even the ocean (USEPA 2009). A scientist 
geographer, John Wesley Powell defined watershed as ?that area of land, a bounded 
hydrologic system, within which all living thing are inextricably linked by their common 
water course and where, as humans settled, simple logic demanded that they become part 
of a community.? The quantity and quality of water in a water body is directly impacted 
by natural or human activities in its watershed. During the recent years, rapid 
urbanization has lead to massive land use changes. With industrialization and population 
growth, the atmospheric concentrations of greenhouse gases and aerosols are believed to 
be increasing, thereby leading to anthropogenic climate change. Climate change, if it 
occurs, will cause increase in temperature, evaporation, evapotranspiration, and 
precipitation variability and extremes (Kundzewicz et al. 2007). These alterations will 
change the hydrological behavior of watershed ecosystem in physical, chemical, and 
biological terms. 
 To study the watershed response to land use change and future climate scenarios, 
a physically based watershed model WARMF (Watershed Analysis Risk Management  
2 
Framework; Chen et al. 2001) project was set up and applied to the Saugahatchee Creek 
Watershed (SCW) in Alabama. Saugahatchee Creek, which feeds on the parts of rapidly 
growing Auburn-Opelika metropolitan area, has been reported of two portions listed on 
Alabama?s 303(d) list of impaired waters (discussed in Chapter 2) for nutrients and 
organic enrichment/dissolved oxygen according to ADEM (2009). This study focuses on 
simulating driving variables like flow and water temperature , together with nutrients, 
algal and dissolved oxygen concentrations for different land use and climate scenarios. 
 
Scope and Objectives 
Historical land use scenarios and potential future climate scenarios, inputs to the 
WARMF model, are not developed in this study. AlabamaView provided three 
Saugahatchee land use scenarios for the past decade for this study (AlabamaView 2009). 
Future projection of land use has not been addressed in this research study. The output 
from General Circulation Models (GCMs) future climate data, HadCM3 (Hadley Centre 
Coupled Model, version 3; Gordan et al. 2000; Pope et al. 2000) in this case, is used in 
this study and downscaled using SDSM (Statistical Downscaling Model; Wilby and 
Dawson 2007).  The predictor variables derived from HadCM3 scenarios that are 
required for SDSM model are available for download on Canadian Climate Change 
Scenarios Network (CCCSN) website (CCCSN 2010). 
The flow and water quality vary depending upon watershed characteristics, its 
geographical location, and climatic condition. The current study focuses on development 
of the WARMF model for the Saugahatchee Creek Watershed in Alabama.  Although the 
quantitative results from this study may be represented for this particular watershed, the 
3 
relevant knowledge about watershed processes and its response to land use change and 
climate change can be shared and compared with other watershed studies. 
The main objectives of this study are as follows: 
1. To develop the WARMF model project for the SCW, calibrate and validate flow 
and water quality parameters,  
2. To apply the developed model to assess the impact due to historical land use 
change in the SCW, and  
3. To apply the developed model to assess the impact due to future climate scenarios 
in the 21
st
 century 
To accomplish the objective 1, the following tasks were completed: 
1. Delineate land catchments, stream segments, and reservoirs in the SCW, 
2. Prepare necessary input data, such as meteorological data, atmospheric 
deposition, etc. for WARMF simulation, and 
3. Simulate flow and water quality, calibrate and validate the model 
To accomplish the objective 2, the following tasks were completed: 
1. Run the WARMF simulations using three land use scenarios as model inputs 
2. Analyse the changes in modeled flow and water quality parameters using three 
land use change scenarios  
To accomplish the objective 3, the following task were completed: 
1. Downscale to local watershed scale, using statistical techniques, output from 
HadCM3 A2 and HadCM3 B2 future climate scenarios to generate meteorological 
inputs for the WARMF model, 
2. Run the WARMF simulations for three allotted future time period of 2020s, 
2050s, and 2080s to simulate flow and water quality based on downscaled climate 
scenario from (1). 
3. Analyse the anomalies of the modeled flow and water quality parameters using 
climate change scenarios to baseline scenario (1981-2010) 
4 
In addition to these objectives, specific research questions were encountered and dealt 
with: 
1. Is the WARMF model a viable tool to simulate flow and water quality in the 
SCW? 
2. What difference it will bring to chlorophyll-a concentration, if we considered an 
impounded section as a stream instead as a reservoir? 
3. What will be the effect of simulation time step on certain water quality parameters 
such as dissolved oxygen? 
4. How will vertical stratification, if any, in a reservoir will affect dissolved oxygen 
concentration? 
 
Thesis Organization 
This thesis is divided into five chapters. Chapter 1 covers the background, scope, 
objectives, and overall thesis organization. Next three chapters, organized in journal 
paper format, are prepared for ASCE journal publication. 
Chapter 2 documents the processes involved during WARMF project 
development of the Saugahatchee Creek Watershed that includes watershed delineation, 
preparation of model input data, model calibration and validation for flow and water 
quality constituents. 
Next, Chapter 3 describes the application of the WARMF model developed in 
Chapter 2 to assess the impact on hydrology and water quality in the streams and 
reservoirs of the Saugahatchee Creek Watershed due to land use and climate change. In 
addition, development of future climate scenarios using statistical downscaling 
techniques are discussed in this chapter. 
5 
Chapter 4, accompanied by a preface, is a technical note submitted for publication 
in ASCE journal that reports 95
th
 percentile 24-hour rainfall depths computed following 
the U.S. Environmental Protection Agency (USEPA) guidelines at 206 weather stations/ 
cities in the contiguous U.S. This chapter is independent of earlier studies in the thesis 
and is developed with a purpose of developing 95
th
 percentile rainfall isohyetal map for 
the contiguous US. The result obtained herein may provide valuable information for 
engineers and designer to comply with Section 438 of EISA when federal agencies need 
to design, construct, and maintain stormwater management practices for development and 
redevelopment projects in the contiguous U.S 
Lastly, Chapter 5 summarizes the project accomplishment against the set 
objectives. This chapter also outlines some related topics for future exploration.  
 
6 
References 
Alabama Department of Environmental Management (ADEM). (2009). ?2008 Alabama 
303(d) List.? 303(d) Information and Map. <http://adem.alabama.gov/programs/ 
water/wquality/2008AL303dList.pdf> (Aug. 12, 2009). 
AlabamaView (2009). ?Saugahatchee Watershed Landcover.? Saugahatchee Watershed 
Project, <http://www.alabamaview.org/watershed_project.html> (Mar. 20, 2009). 
Canadian Climate Change Scenarios Network (CCCSN). (2010). ? HadCM3 predictors: 
A2(a) and B2(a) experiments.? Statistical Downscaling Input, <http://cccsn.ca/ 
?page=pred-hadcm3> (Sep. 21, 2010)   
Chen, C. W., Herr, J. W., and Weintraub, L. (2001). ?Watershed Analysis Risk 
Management Framework (WARMF): Update One?A decision support system for 
watershed analysis and total maximum daily load calculation, allocation and 
implementation.? Publication No. 1005181. Electric Power Research Institute, 
Palo Alto, California. 
Gordon, C., Cooper, C., Senior, C. A., Banks, H., Gregory, J. M., Johns, T. C., Mitchell, 
J. F. B., and Wood, R. A. (2000). ?The simulation of SST, sea ice extents and 
ocean heat transports in a version of the Hadley Centre coupled model without 
flux adjustments.? Climate Dynamics, 16(2), 147-168. 
 
Kundzewicz, Z. W., Mata, L. J., Arnell, N. W., Doll, P., Kabat, P., Jimenez, B., Miller, 
K. A., Oki, T., Sen, Z., and Shiklomanov, I. A. (2007). "Freshwater resources and 
their management." Climate Change 2007: impacts, adaptation and vulnerability: 
contribution of Working Group II to the fourth assessment report of the 
Intergovernmental Panel on Climate Change, M. L. Parry, O. F. Canziani, J. P. 
Palutikof, P. J. v. d. Linden, and C. E. Hanson, eds., Cambridge Univ Press, 
Cambridge, UK, 173-210. 
Pope, V., Gallani, M. L., Rowntree, P. R., and Stratton, R. A. (2000). ?The impact of new 
physical paramterizations in the Hadley Centre climate model: HadAM3.? 
Climate Dynamics, 16, 123-146. 
U. S. Environmental Protection Agency (USEPA). (2009). ?What is a watershed?? 
Watersheds, <http://water.epa.gov/type/watersheds/> (Nov. 17, 2009). 
Wilby, R. L., and Dawson, C. W. (2007). "SDSM 4.2?A decision support tool for the 
assessment of regional climate change impacts." Lancaster University, UK. 
  
 
7 
Chapter 2.  Development of WARMF Model to Study Hydrology and Water Quality 
in Streams and Reservoirs of the Saugahatchee Creek Watershed 
 
Abstract 
A physically based watershed model WARMF was set up for the Saugahatchee 
Creek Watershed in order to assess the impact on hydrology and water quality in streams 
and reservoirs due to land use and climate change. This paper deals with explaining the 
process involved during WARMF project development that includes watershed 
delineation, preparation of input data required for WARMF model, calibration and 
validation for flow and water quality constituents. Flow calibration (validation) 
performed using 10 years? observed daily flow displayed satisfactory model performance 
resulting Nash-Sutcliffe?s efficiency (NSE) of 0.64 (0.56), ratio of the root mean square 
error to the standard deviation of measured data (RSR) of 0.60 (0.66) and percent bias 
(PBIAS) of -2.78% (-9.53%) . Water quality calibration and validation were performed 
using graphical comparison. Saugahatchee Creek WARMF project was then applied in 
the companion paper to assess the impact of land use change and climate change on 
hydrology and water quality in the watershed. 
 
Introduction 
Flow and water quality conditions in a stream or reservoir depend on not only in-
stream and in-reservoir processes but also inputs of flow and water quality constituents 
8 
from its surrounding and upstream watersheds. Various watershed-scale modeling efforts 
are made to mimic the physical, chemical, and biological processes involved during water 
transport from precipitation through canopy, land surfaces, soil layers, streams, and 
reservoirs. The output of these models under different meteorological, land use, and land 
management scenarios gives hydrological and water quality response of streams and 
reservoirs to these changes over time, and are useful for watershed management and 
planning. 
Complex hydrological models such as AGNPS (Agricultural Nonpoint Source 
Pollution Model; Young et al. 1987); BASINS (Better Assessment Science Integrating 
point and Nonpoint Sources;(USEPA 2004); HSPF (Hydrological Simulation Program ? 
Fortran; Johanson et al. 1980); GWLF (Generalized Watershed Loading Functions; Haith 
and Shoenaker 1987); SWAT (Soil and Water Assessment Tool; Arnold and Soil 1994); 
and WARMF (Watershed Risk Analysis Management Framework; Chen et al. 2001) 
have been frequently applied to study watershed hydrology in the United States (U.S.) 
and all over the world. A physically based, dynamic watershed model WARMF was 
applied in this study to the Saugahatchee Creek Watershed (SCW) for assessing 
hydrology and water quality impact due to land use and climate change.  Although other 
models, discussed above, would yield similar results, WARMF was applied here for its 
integration of stream and one-dimensional (1-D) reservoir water quality models, user 
friendly graphical interface and ability to assess the impact of point and nonpoint sources 
with varying land use and meteorological scenarios. WARMF is also incorporated with 
decision support system designed to support the watershed approach analysis and Total 
Maximum Daily Load (TMDL) calculations. WARMF has undergone several peer 
9 
reviews (Keller 2000; Keller 2001) and has been compared with other renowned models 
(Chen and Herr 2002; Chen et al. 2005; Neilson et al. 2003). WARMF has been applied 
for various practices over the years including TMDL study (Chen et al. 2000; Herr et al. 
2002; Herr et al. 2003; McDonald et al. 2000; Stringfellow et al. 2009); nutrient 
management strategy development (NC DENR 2009; Wang et al. 2004); and watershed 
assessment and planning (RMC Water and Environment 2007). Very few studies have 
demonstrated the capacity of WARMF model for climate change and land use change 
impact study. Rich et al. (2005) enhanced WARMF with a module capable of 
constructing different climate and management scenarios and applied it  to evaluate the 
effects of extended droughts and increased temperature on water budgets in the San Juan 
Basin. 
In this study, WARMF project was developed for the Saugahatchee Creek 
Watershed in Alabama to assess land use and climate change impact. This paper deals 
with explaining the process involved during WARMF project development that includes 
watershed delineation using BASINS, preparation of model input data, model calibration 
and validation for flow and water quality constituents. The Saugahatchee Creek WARMF 
development was applied to the companion paper for assessing the impact on hydrology 
and water quality due to land use and climate change. 
 
Study Area 
The watershed of concern is the Saugahatchee Creek Watershed, located mostly 
in Piedmont region of eastern Alabama with an area of approximately 550 km
2 
(Fig. 2.1). 
Beginning with its headwaters in Chamber and Lee counties, Saugahatchee Creek runs 
10 
westward through parts of Macon and Tallapoosa counties until it enters Yates Reservoir 
and converges to Tallapoosa River. Two segments in the SCW are listed on the Alabama 
Department of Environmental Management (ADEM)?s 303 (d) list of impaired waters 
under the federal Clean Water Act (ADEM 2009). Pepperell branch, a tributary to 
Saugahatchee Creek is listed as impaired waters for nutrients and the portion of 
Saugahatchee Creek (Yates Reservoir Embayment) is listed for nutrients and organic 
enrichment/dissolved oxygen (Fig. 2.1). Table 2.1 specifies the causes and sources of 
impairment according to ADEM website (ADEM 2009). 
 
Methods 
Model Description 
WARMF is an integrated watershed model and decision support system (DSS) 
having simulation models, databases, and decision support tools under one GIS-based 
graphical user interface, available as a public domain tool via U.S. Environmental 
Protection Agency (USEPA) website <http://www.epa.gov/athens/wwqtsc/html/ 
warmf.html>. The algorithms embedded in WARMF are adapted from many well 
established codes such as ILWAS (Integrated Lake Watershed Acidification Study (Chen 
et al. 1983; Gherini et al. 1985)), ANSWERS (Areal Nonpoint Source Watershed 
Environmental Response Simulation (Beasley et al. 1980; Beasley and Huggins 1981)), 
SWMM (Storm Water Management Model (Huber et al. 1988)), and WASP (Water 
Quality Analysis Simulation Program (Ambrose et al. 1993)). WARMF is organized into 
five modules ? Data, Engineering, Knowledge, TMDL and Consensus Module. 
11 
Data module provides data to engineering module for model simulations and for 
evaluation of model outputs. Engineering module performs mass balance, heat balance, 
reaction kinetics, chemical equilibrium and other calculations and returns model outputs. 
Knowledge module can be used to store technical documents or other useful information. 
Consensus module helps stakeholders make informed decisions by formulating and 
evaluating management alternatives. TMDL module provides a step-by-step process of 
calculating the total maximum daily load of pollutants entering from upstream of a water 
quality limited section (Chen et al. 2001; Herr et al. 2000). 
WARMF represents a watershed by dividing it into a network of land catchments, 
stream segments, and reservoirs (Fig. 2.2). Major streams and their tributaries in a 
watershed are divided into different stream segments as model compartments in WARMF 
by land catchments. Land catchment is further divided into canopy layer, snowpack and 
soil layers (Fig. 2.2). Each compartment is considered as a seamlessly connected 
continuously stirred tank reactor (CSTR) for flow routing and mass balance calculation. 
WARMF simulates the hydrologic process of canopy interception, snowpack 
accumulation and snowmelt, infiltration through soil layers, evapotranspiration, surface 
runoff and groundwater exfiltration to stream segments. Canopy interception is 
determined as function of leaf area index (LAI), the maximum canopy interception, and 
available precipitation. LAI varies monthly for each land use category and is part of 
system coefficients of WARMF.  When the precipitation is greater than canopy 
interception, the excess becomes throughfall. Throughfall and snowmelt fall onto the 
ground surface.  Impervious surface will result in immediate runoff.  On pervious 
surfaces, the water may infiltrate into the soil layers, remain on the surface (local 
12 
depression storage), and flow as surface runoff. Infiltration is carried out based on 
physical processes in each soil layers depending upon the amount of water available for 
infiltration, the void space available in the layer below and the vertical infiltration rate. 
The lateral flow from the soil layers to adjacent catchment, river or reservoir is based on 
Darcy?s law: 
jhjj
SWZKQ =  (2.1)
Where, Q
j
 is the lateral exfiltration from layer j; K
hj
 is the horizontal hydraulic 
conductivity of layer j; S is slope of the catchment; W is width of the catchment parallel 
to its receiving stream, or perpendicular to the direction of ground water flow; and Z
j
 is 
thickness of layer j. The final water balance is performed from the bottom layer to the top 
layer, one layer at a time. For each soil layer, the overall water balance is as follows, 
jjjjjjj
QELIIVV ??+?+=
+10
 (2.2)
Where, V
j
 is the volume of water in the soil layer j; V
j0
 is the initial volume of water in 
soil layer j; I
j
 is the infiltration to layer j; I
j+1
 is the percolation from layer j to layer j+1; 
L
j
 is the lateral inflow from an upstream segment; Ej is the evapotranspiration from layer 
j; and Q
j
 is lateral exfiltration from layer j. The surface water which does not infiltrate 
into the soil may be ponded on the surface as detention storage or it will runoff as sheet 
flow. The sheet flow is calculated by Manning?s equation: 
3
1
2
1
0
01.0?
=
n
SWZ
Q
s
 
(2.3)
Where, Q
s
 is runoff from the pervious areas (m
3
/s); Z
0
 is the water depth available for 
sheet flow; n is the Manning?s roughness coefficient. Evapotranspiration depends on the 
13 
available amount of water that can be transpired and the potential evapotranspiration 
calculated as a function of latitude according to Hargreaves (1974). 
The water from the upstream stream segment is fully mixed with the water in the stream 
segment from previous time step and the point and nonpoint loads entering the stream 
segment during the time step. For each CSTR (canopy layer, soil layer, stream segment, 
reservoir layer, etc.), the flow continuity equation can be written based on conservation of 
mass: 
??
=?
dt
dV
QQ
outin
 
(2.4)
Where, Q is the flow rate, V is the volume, subscripts in and out represents inflows and 
outflows respectively. The example of inflows to a stream segment will include flows 
from upstream stream segments, reservoirs, surrounding land catchment, soil layers, and 
point sources. 
Heat budget and mass balance calculation are performed to calculate the water 
temperature and concentrations of various water quality constituents in each soil layer, 
stream segment, and reservoir layer (Chen et al. 2001). Dynamic mass balance and heat 
budget equation can be written as follows: 
???
+?= SSCQCQ
dt
VCd
outinin
)(
 
(2.5)
???
+?= SSTQTQ
dt
VTd
outinin
)(
 
(2.6)
Where, V is the volume of each compartment (control volume), C is the constituent 
concentration, T is the temperature, Q is the flow rate, SS is the sources and sinks of the 
constituents, subscripts in and out represents inflows and outflows respectively. 
14 
A reservoir is further divided into about 30 horizontal layers along depth to 
simulate water quality in a stratified reservoir. Each layer is assumed to be horizontally 
mixed. When reservoir elevation rise and fall due to variations of inflow(s) and 
outflow(s), the model correspondingly add or delete layers. The model requires reservoir 
bathymetric data in the form of stage-area relationship for simulation. The reservoir flow 
balance is made according to the Equation (2.4). 
 
Watershed Delineation using BASINS 
The delineated watershed map with land catchments, stream segments, and 
reservoirs, is required for WARMF, to which input data can be given and simulation 
results can be viewed. The functionality of watershed delineation is available in original 
WARMF model (private version developed by Systech Water Resources Inc.) but 
deactivated in USEPA-distributed WARMF model. However, WARMF is compatible 
with data extraction and delineation tools of BASINS. The watershed delineation 
developed from BASINS 3.1 can be imported into WARMF (Systech 2005). 
BASINS is a GIS-based software system developed by U. S. EPA to assist 
regional, states and local agencies in performing watershed analysis and examining 
management alternatives. It integrates a GIS, national watershed data, and state-of-the-art 
environmental assessment and modeling tools into one software package. The 
environmental assessment and modeling tool embedded in BASINS was not used here. 
Instead, watershed delineation tool in BASINS was utilized to delineate watershed based 
on digital elevation map (DEM) and river network. BASINS data download tool extracts 
nationally derived databases like boundaries for states and cataloging units, DEM, 
15 
National Hydrography Dataset (NHD), and National Land Cover Dataset (NLCD). A 
BASINS project can be set up for USGS 8-digit HUC (Hydrologic Unit Codes) 
watershed by selecting the geographic area of interest from among the entire 48 
contiguous United States. 
In this study, BASINS 3.1 was utilized to delineate watershed and prepare 
necessary GIS layers for the WARMF model. We selected the Lower Tallapoosa 
Watershed which is an 8-digit HUC 03150110 watershed that contains Saugahatchee 
Creek Watershed (Fig. 2.3). BASINS watershed delineation tool divided the Lower 
Tallapoosa Watershed into 279 subwatersheds (catchments) when threshold area of 850 
hectares was used. Out of 279 subwatersheds, 44 subwatersheds that drain to the 
Saugahatchee Creek were selected and clipped to generate catchment layer and river 
layer for the study watershed (Fig. 2.3). Systech (2005) provides detailed instructions and 
procedures of creating WARMF application using BASINS delineation tools. 
The stream segments were cut off and replaced with reservoir segment where 171 
ha Lake Saugahatchee and 83 ha Yates Reservoir Embayment existed. Three stream 
segments in place of Lake Saugahatchee and one stream segment in place of Yates 
Reservoir Embayment had to be deleted before these reservoir segments were imported 
into WARMF. The reservoir segments are shown along with 44 land catchments and 40 
remaining stream segments in Fig. 2.4. Table 2.2 provides the statistical summary for 
areas of 44 catchments and 40 stream segments of the SCW imported to WARMF model.  
The catchment area was ranged from 62.0 ha (0.24 mile
2
) to 4035.5 ha (15.6 mile
2
).  The 
length of stream segments was ranged from 419.8 m to 11,694.1 m.  Each stream 
16 
segment is a model compartment or CSTR in WARMF, therefore, large length of stream 
segments may prevent us to mimic variation of water quality constituents along a stream. 
 
WARMF Model Input 
WARMF requires preparation of series of input data before it can be used to 
simulate the watershed hydrology and water quality.  The model inputs can be divided 
into two categories: model coefficients and time series data. Model coefficients are those 
parameters that describe physical, chemical and biological characteristics of the 
watershed. The model coefficients are further classified as catchment, river, reservoir, 
and system coefficients. Model coefficients are adjusted during model calibration. Time 
varying data such as rainfall data is time-series data. Data module manages and stores 
time-series data. The major time-series input data are meteorology, atmospheric 
deposition, and point source discharges data. Observed flow and water quality data are 
required for model calibration and validation. The major meteorological input categories 
to the model are described as follows: 
 
Meteorological Data: 
WARMF requires meteorological variables such as precipitation, minimum and 
maximum temperatures, mean station pressure, dew point temperature, wind speed and 
cloud cover for hydrology simulation. The meteorological data can be downloaded from 
National Climatic Data Center (NCDC) website. The daily precipitation records for a 
station closer to the study watershed were available since 1996 in the station near 
Auburn, AL (Auburn No. 2). The meteorological variables other than precipitation 
17 
required for WARMF input were downloaded and extracted from the Class I weather 
station in Montgomery, AL (Montgomery Dannelly Field) except cloud cover data which 
was calculated using the formula suggested by Systech (2005): 
( )
?
?
?
?
?
?
?
+
=
dew
T
TT
ABSdT
2
maxmin
 
(2.7)
If Precip = 0 then    
       If dT < 4 then 
                 Cloud = 0.6 
              Else If dT < 6 then 
                 Cloud = 0.3 
                                                          Else  
              Cloud = 0 
      Else If Precip > 2 then 
Cloud =1 
     Else If Precip > 1 then 
   Cloud = 0.9 
                                                  Else 
   Cloud = 0.8 
                                                  End If 
 
(2.8)
Where, T
min
 is the minimum temperature in ?C , T
max
 is the maximum temperature in ?C, 
T
dew
 is the dewpoint temperature ?C, Precip is the precipitation in centimeters, and Cloud 
is the cloud cover. Table 2.3 shows the monthly average, standard deviation, maximum 
and minimum values of precipitation (from Auburn No. 2 station) along with minimum 
and maximum temperatures (from Montgomery Dannelly Field station) for 1997-2009 
used for this study. 
 
Land Use: 
Land use describes surface characteristics of the watershed. NLCD 2001 land use 
data was available for Tallapoosa Basin from Alabama Cooperative Extension System 
(ACES) website (ACES 2009). The portion of Saugahatchee Creek Watershed was 
18 
cropped using GIS tools from Tallapoosa Basin (Fig. 2.5). SCW?s land use is dominated 
by forest, cropland, pasture and urban areas.Fig. 2.5 illustrates the distribution of land use 
categories in the watershed according to NLCD 2001. 
GIS shapefile of land use can be imported directly into WARMF after it has been 
unprojected to decimal degrees (Systech 2005). When the land use shapefile is imported, 
WARMF automatically calculates the percentage of land use categories in each 
catchment. Table 2.4 shows examples of the variations in land use distributions in three 
catchments (catchment 11, 22, and 42 in WARMF associated with three monitoring 
stations) within the watershed (Fig. 2.4).  Catchment 11, one of the upstream urban 
catchments, has higher percentage (42%) of residential area whereas Catchment 22 and 
Catchment 42 are mostly forest area comprising about 62% and 88% forestland (Table 
2.4). 
 
Land Application: 
Nitrogen and phosphate fertilizers usage rate for cotton was used to calculate the 
fertilizer application rate for cropland using the following equation (NC DENR 2009): 
PQA ??= 12085.1 (2.9) 
where, A is the average annual nitrogen or phosphate application rate in kg/ha; Q is the 
average annual nitrogen or phosphate application rate in lbs/acre, 1.12085 being 
conversion factor; and P is the percentage of fertilizer used crops area from total 
cropland. United States Department of Agriculture (USDA) provides annual nitrogen and 
phosphate fertilizers usage rate (lbs/acre) for different crops including cotton in Alabama 
(USDA 2011).  Cotton is the major crop, which makes 22% of cropland in Lee County in 
19 
Alabama, 71% are forages and rest 7% are other crops according to 1997 US Census of 
Agriculture (CropMAP 2010). The fertilizer application rate for cotton and other crops 
were considered to be equal and that from forages was assumed to be half of the rate for 
cotton. 
The annual average fertilizer application rate for phosphate was broken down into 
monthly application rate with 4 kg/ha during the growing season (April - October) and 3 
kg/ha for rest of the period. For nitrogen fertilizer, the total annual value was equally 
shared between ammonia and nitrate usage rate. Both ammonia and nitrate fertilizer 
application rate were used as 2.5 kg/ha during growing season and 2 kg/ha for rest of the 
period. The monthly fertilizer application rate for urban areas was calibrated to attain 
good fit between the modeled and observed nutrient concentration during water quality 
calibration. 
 
Soils Data: 
USDA?s National Resources Conservation Service (NRCS) Soil Data Mart 
provides Soil Survey Geographic Database (SSURGO) soils data for each county. Lee 
County soils data were downloaded and three prominent soil coverage based on larger 
percentage coverage were Pacolet sandy loam, Cecil sandy loam and Marvyn loamy 
sand. Based on these soil types, soil parameters such as soil thickness, field capacity, 
saturation moisture, horizontal conductivity and vertical conductivity were estimated. 
The soil parameters were adjusted during model calibration to generate good match 
between modeled and observed data. The soil parameters along with other adjusted 
parameters for calibration are listed in Table 2.5. 
20 
Atmospheric Deposition: 
WARMF requires dry and wet chemical constituents? concentration in air (?g/m
3
) 
and rain water (mg/l) respectively. The Clean Air Status and Trends Network 
(CASTNET) provides dry deposition of particles over 55 site locations mostly in the 
eastern United States. The National Atmospheric Deposition Program (NADP) provides 
wet deposition concentration over 200 sites in the United States, Puerto Rico, Virgin 
Island, Samoa and Islands. CASTNET station GAS153 (Georgia Station, GA) is the best 
source for weekly dry deposition concentration and NADP station AL10 (Black Belt 
Research & Extension Center, AL) is the most appropriate source for weekly wet 
deposition concentration. The dry and wet atmospheric deposition data from these 
stations were extracted, processed and imported into WARMF.  
 
Point Sources Data:  
The major point source dischargers in the SCW were Auburn Northside Waste 
Water Treatment Plant (WWTP), Opelika Westside WWTP and West Point Stevens 
(Table 2.6). The loading from these point discharges for the time period 2000-2002 were 
available from data collected by ADEM and Auburn University (ADEM 2008) and 
presented in the appendices. These point sources were included as input to respective 
streams in the catchment map. 
To extend the point sources data for the time period other than 2000-2002, 90
th
 
percentile of the loadings were calculated and used as the constant rate.  The 90
th
 
percentile loading and the observed loading for three major point source dischargers 
during 2001-2002 period are shown in Fig. 2.6. 
21 
Observed flow: 
The daily flow data can be obtained from USGS water data web site 
<http://waterdata.usgs.gov/nwis/sw>. For the SCW, the only available USGS station with 
continuous daily data is USGS 02418230 in the Saugahatchee Creek at County Road 188 
near Loachapoka. The USGS daily data available from 2000 through 2009 were used for 
model calibration and validation. The monthly summary of daily flow data is provided in 
the box and whisker plot with monthly average, maximum, minimum, median, first and 
third quartiles (Fig. 2.7). 
 
Observed Water Quality: 
ADEM and Auburn University collected water quality data throughout the 
Saugahatchee Creek Watershed in 2000-2002 (ADEM 2008). The stations in the 
Pepperell Branch downstream of West Point Stevens (Station-16), in the Saugahatchee 
Creek near Loachapoka (Station-8) and in the portion of the Saugahatchee Creek entering 
Yates Reservoir (Yates-2 Station) were used in the study for the purpose of model 
calibration and validation for water quality constituents. The surface water temperature, 
dissolved oxygen (DO), total phosphorus (TP) and total nitrogen (TN) concentrations 
were the water quality parameters sampled in Station-16 and Station-8, used in our study 
for model calibration and validation. From Yates-2 station, the surface water temperature, 
DO and chlorophyll-a concentration were derived for our study. 
 
WARMF Simulation 
22 
After completing all of the setup and importing necessary input data outlined in the 
previous sections, WARMF model is ready to run and simulate hydrology and water 
quality in the streams and reservoirs of the watershed. The simulation was performed 
from 1997 through 2009. The model simulation was run three years prior to the 
calibration period (2000 - 2009) to minimize the effect of initial unknown parameters 
used in the model. A daily time step was used for calibration and validation of the model. 
WARMF can also simulate using desired hourly time step. Hourly time step simulation 
can be useful to predict diurnal phenomenon. For example, in an aquatic system with 
high algal concentration, dissolved oxygen largely varies during day and night. 
Simulation using daily time steps may not be able to catch the extreme values which are 
very critical in context of stream and reservoir ecology. But at the same time, it should be 
accounted with change in model coefficients, as simulation using hourly and daily time 
steps do not yield identical results. If we use daily time step, constant moderate amount 
of light is in effect all day which is favorable for algal growth whereas if we simulate 
using hourly time step, during night time there is no light and during some part of day 
there is too much light hindering algal growth. The model coefficients can be adjusted 
separately for simulation using hourly or daily time steps during calibration process. The 
maximum algal growth rate, which is expressed ?per day? was used to be 1 per day. For 
hourly time step, higher growth rate might have to be used to match it with observed 
algal concentration taking into consideration that algal growth requires ideal light 
condition which occur only few hours of the day. We used maximum growth rate of 2.8 
per day for simulation using 1-hour time step. 
 
23 
Results and Discussions 
Calibration Criteria 
The performance of the WARMF model was evaluated using both graphical and 
statistical measures. Moriasi et al. (2007) recommends performance of the model can be 
evaluated based on three quantitative statistics, Nash-Sutcliffe efficiency (NSE), ratio of 
the root mean square error to the standard deviation of measured data (RSR), and percent 
bias (PBIAS) in addition to the graphical techniques. The recommended quantitative 
statistics between simulated (Y
i
sim
) and observed (Y
i
obs
) were computed as follows: 
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?=
?
?
=
=
n
i
meanobs
i
n
i
sim
i
obs
i
YY
YY
NSE
1
2
1
2
)(
)(
1  
(2.10)
?
?
=
=
?
?
=
n
i
meanobs
i
n
i
sim
i
obs
i
YY
YY
RSR
1
2
1
2
)(
)(
 
(2.11)
?
?
?
?
?
?
?
?
?
?
?
?
??
=
?
?
=
=
n
i
obs
i
n
i
sim
i
obs
i
Y
YY
PBIAS
1
1
)(
100)(
 
(2.12)
The simulation of watershed model can be judged satisfactory if NSE > 0.5 and RSR ? 
0.7 and if PBIAS is within ?25% for flow (Moriasi et al. 2007).  
Saugahatchee Creek Watershed was divided into 44 catchments, 40 stream 
segments and 2 reservoirs. The hydrologic and water quality parameters can be assigned 
to these segments individually, referred as catchment coefficients, river coefficients, and 
reservoir coefficients in WARMF manual (Herr et al. 2000). There are another set of 
parameters, known as system coefficients that apply to all catchments in the watershed. 
24 
During the calibration, some of catchment, river, reservoir and system coefficients were 
adjusted to obtain a best fit. Although WARMF includes auto-calibration tool, manual 
calibration was performed prioritizing the critical parameters that are sensitive to flow 
and water quality. Zheng and Keller (2006) and Geza et al. (2009) have listed sensitive 
parameters during WARMF simulation. The adjusted model parameters for our study are 
listed in Table 2.5.  
 
Flow Calibration and Validation 
The observed flow daily data is available for USGS gage station in the 
Saugahatchee Creek near Loachapoka since 2000. The flow was calibrated for the period 
2000-05 and validated for the period 2006-09 (Fig. 2.8). WARMF simulated the flow 
well in the Saugahatchee Creek although there are few underestimations of higher peaks. 
The simulation of flow resulted in NSE, RSR and PBIAS value for calibration 
(validation) as 0.64, 0.60 and -2.78 % (0.56, 0.66 and -9.53 %) respectively (Table 2.7). 
 
Water Quality Calibration and Validation 
Auburn University and ADEM collected water quality data in the SCW in 2000-
2002 (ADEM 2008). Simulated surface water temperature, dissolved oxygen (DO), algal 
and nutrients concentrations from WARMF were compared to the observed data at 
Station 16, Station 8, and Yates-2 Station. The calibration was performed for the year 
2000/01 and validation for the year 2002. As the water quality data were not collected in 
daily basis, water quality calibration and validation was performed based on visual 
comparison of simulated and observed data. Time series plots are used for visual and 
25 
graphical representation. The calibration and validation results are listed according to the 
water quality stations as follows: 
 
At Pepperell Branch downstream of West Point Stevens (Station 16): 
Fig. 2.9 shows the time series plots of the observed and modeled surface water 
temperature, DO, TP and TN concentrations at Pepperell branch downstream of West 
Point Stevens (Station 16). The simulated surface temperature closely followed the 
observed temperature pattern, but seemed to overpredict temperature periodically, 
especially in the summer. The match between the modeled and observed DO shows few 
extreme observed values missing out during simulation. Fig. 2.9 shows satisfactory fit 
between the modeled and observed nutrient (TP and TN) concentrations. 
 
At USGS 02418230 Station in the Saugahatchee Creek near Loachapoka: 
Fig. 2.10 illustrates the time series plot of the observed and modeled surface water 
temperature, DO, TP and TN concentrations at USGS 02418230 Station in the 
Saugahatchee Creek near Loachapoka. The simulated surface water temperature closely 
followed the observed water temperature. For DO simulation, there were few extreme 
observed values, which were not well predicted. But, the overall pattern of observed and 
modeled values was very similar. Fig. 2.10 depicts satisfactory time series plot of 
observed and modeled nutrients concentration.  
 
At Yates Reservoir Embayment: 
26 
Fig. 2.11 shows the time series plot of the observed and modeled surface DO, TP 
and chlorophyll-a concentrations. Chlorophyll-a and TN concentrations were well 
simulated. DO concentration seemed to be over predicted, which could be partly due to 
diurnal variation in dissolved oxygen and partly due to vertical stratification. 
 
Chlorophyll-a and DO Simulation in Yates Reservoir Embayment 
The WARMF model developed for the SCW provides two alternatives to model 
Yates Reservoir Embayment: (1) as a reservoir containing about 30 horizontal layers 
along depth and (2) a stream segment.  Other popular watershed models, e.g., SWAT 
(Arnold and Soil 1994) and HSPF (Johanson et al. 1980), do have a pond model 
component but treat it as well-mixed pond and cannot simulate stratification of any water 
quality parameters.  The modeled chlorophyll-a concentrations when Yates Reservoir 
Embayment is treated as a stream segment got to very low values as compared to treated 
as a reservoir.  Reservoir, as opposed to stream segment, has calm water (slow velocities) 
and higher nutrient enrichment/ organic deposition that support algal growth. When Yates 
Reservoir Embayment was treated as a reservoir, the modeled chlorophyll-a 
concentration, though slightly overpredicted few observed values, showed comparable 
outputs whereas when treated as a stream segment the modeled chlorophyll hugely under-
predicted the observed values (Fig. 2.12).  
Still, model predicted DO values at Yates-2 station are very high compared to the 
observed values. During growing season, the modeled and observed chlorophyll-a 
concentration in Yates Reservoir Embayment showed higher concentration. With 
chlorophyll-values greater than 15 ?g/l, water body is regarded eutrophic (Carlson 1977). 
27 
The studies have shown that, for eutrophic water bodies, the fluctuation of DO 
concentration within a day is larger (Ansa-Asare et al. 2000). DO level are increased 
during day time when algal photosynthesis adds oxygen to the system. When there is no 
light for photosynthesis during night time, algal respiration and decomposition reduces 
oxygen levels from the system. For long-term future analysis, the model was simulated 
on daily time step, from which diurnal fluctuation is not observable. 
To better understand DO diurnal fluctuation at Yates-2 station, the WARMF 
model was set up and run using hourly time step. Fig. 2.13 shows standard deviation of 
surface DO from daily mean for each day for the year 2000 and 2002 simulated at Yates 
Reservoir Embayment using hourly time step. Surface DO standard deviation, 
corresponding to higher chlorophyll-a concentration in growing season, has higher values 
(Fig. 2.13). Simulated surface DO fluctuation during a day for five days for which 
observed DO values were less than 5 mg/l  show the fluctuation of 2.7 mg/l at most.(Fig. 
2.14) However, the modeled values at reservoir surface layer were greater than 5 mg/l. 
The observed values of DO were measured at mid-depth if depth < 10ft and at 5 ft 
if depth ? 10 ft by ADEM and Auburn University (ADEM 2008). WARMF outputs the 
modeled time step values only at the reservoir surface for time-series plots (e.g., Fig. 
2.13). Therefore, the vertical stratification, if any, may provide better understanding for 
DO simulation. Unlike stream segment as a CSTR, reservoir is divided into about 30 
horizontal layers in WARMF to simulate stratification. Usually, the upper layers known 
as epilimnion, close to surface have tendency to be mixed and get aerated but the deeper 
hypolimnion, in case of stratified reservoir, possess diminished oxygen levels. For the 
same five days (observed DO<5mg/l), the vertical profiles of simulated DO were plotted 
28 
with 5 ft line drawn to scale (Fig. 2.15).  Simulated DO at 5 ft are all less than surface 
DO except on September 26, 2002 when fall overturn resulted well mixed profile. 
Simulated DO values got the lowest as 2.6 mg/l near reservoir bottom. 
These results show that temporal variation of DO during a day in algal abundance 
(Fig. 2.13 and Fig. 2.14) and spatial (vertical) variation of DO (Fig. 2.15) in a stratified 
reservoir could be reason for DO mismatch between observed and modeled values at 
Yates-2 station (Fig. 2.11).  
 
Summary 
Saugahatchee Creek WARMF project was set up in order to assess the impact due 
to land use and climate change on hydrology and water quality in the watershed. The 
calibration and validation was performed based on the best available data in the 
watershed. The flow calibration was done using 2000-2005 observed flow, and validated 
using data from 2006-2009 available at the USGS gauge station located in the 
Saugahatchee Creek near Loachapoka. The water quality parameters were calibrated and 
validated at three stations where ADEM and Auburn University collected water quality 
data for 2000-2002. By evaluating model performance criteria and visual inspection of 
time series plot, the calibration and validation results were observed to be satisfactory 
enough for its further application. 
As with all watershed models, there are inherent model limitations associated with 
WARMF. The simplistic ground-water approach in WARMF does not thoroughly 
simulate complex ground-water aquifers. Each stream segment is a model compartment 
29 
or CSTR in WARMF, therefore, large length of stream segments may prevent us to 
mimic variation of water quality constituents along a stream.  
 
 
30 
References 
Alabama Cooperative Extension System (ACES). (2009). ?Land Use and Land Cover 
Data?. Alabama GIS Data Distribution. Alabama Cooperative Extension System. 
< http://www.aces.edu/waterquality/gis_data/>. (Apr. 30, 2009). 
 
Alabama Department of Environmental Management (ADEM). (2008). "Final total 
maximum daily load: Nutrients & OE/DO. Pepperell branch AL03150110-0201-
700 Nutrients; Sougahatchee Creek Embayment (Yates Reservoir) AL03150110-
0204-101 Nutrients & OE/DO.", Alabama Department of Environmental 
Management, Montgomery, AL. 
 
Alabama Department of Environmental Management (ADEM). (2009). ?2008 Alabama 
303(d) List.? 303(d) Information and Map. Alabama Department of 
Environmental Management <http://adem.alabama.gov/programs/water/wquality/ 
2008AL303dList.pdf> (Aug. 12, 2009). 
 
Ambrose, R., Wool, T., and Martin, J. (1993). "WASP 5, The Water Quality Analysis 
Simulation Program Version 5.00." ASCI Corporation, Athens, Georgia. 
 
Ansa-Asare, O. D., Marr, I. L., and Cresser, M. S. (2000). "Evaluation of modelled and 
measured patterns of dissolved oxygen in a freshwater lake as an indicator of the 
presence of biodegradable organic pollution." Water Research, 34(4), 1079-1088. 
 
Arnold, J. G., and Soil, G. (1994). SWAT (Soil and Water Assessment Tool), Grassland, 
Soil and Water Research Laboratory, USDA, Agricultural Research Service. 
 
Beasley, D., Huggins, L., and Monke, E. J. (1980). "ANSWERS: A model for watershed 
planning." 
 
Beasley, D. B., and Huggins, L. F. (1981). "ANSWERS, areal nonpoint source watershed 
environment response simulation: user's manual." 
 
Carlson, R. E. (1977). "A trophic state index for lakes." Limnology Oceanography, 22(2), 
361-369. 
 
Chen, C. W., Gherini, S. A., Hudson, R., and Dean, J. (1983). "The Integrated Lake-
Watershed Acidification Study. Volume 1: Model principles and application 
procedures." Electric Power Research Institute report EA-3221, Palo Alto, 
California, USA. 
 
Chen, C. W., and Herr, J. W. (2002). "Comparison of BASINS and WARMF models: 
Mica Creek watershed." Technial Rep. No. 1005344, Electric Power Research 
Institute, Palo Alto, CA. 
 
31 
Chen, C. W., Herr, J. W., Goldstein, R. A., Ice, G., and Cundy, T. (2005). "Retrospective 
comparison of watershed analysis risk management framework and hydrologic 
simulation program Fortran applications to Mica Creek watershed." Journal of 
Environmental Engineering, 131, 1277. 
 
Chen, C. W., Herr, J. W., and Weintraub, L. (2001). "Watershed Analysis Risk 
Management Framework (WARMF): Update One?A decision support system for 
watershed analysis and total maximum daily load calculation, allocation and 
implementation. Publication No. 1005181." Electric Power Research Institute, 
Palo Alto, California. 
 
Chen, C. W., Loeb, C., and Herr, J. W. (2000). "Adaptation of WARMF to Calculate 
TMDL for Chartiers Creek Watershed in Pennsylvania." Final Rep. to USEPA 
Region 3, Philadelphia, PA and The Chartiers Creek Watershed TMDL 
Stakeholder Group, Pittsburgh, PA. 
 
CropMAP (2010). ?Alabama Cropmap Lee County.? A New Crop Information System for 
Alabama. Purdue University. Center for New Corps & Plants Products. 
<http://www.hort.purdue.edu/newcrop/cropmap/alabama/counties/lee.html> 
(Mar. 2, 2011). 
 
Geza, M., Poeter, E. P., and McCray, J. E. (2009). ?Quantifying predictive uncertainty for 
a mountain-watershed model.? Journal of Hydrology, 376,(1-2), 170-181. 
 
Gherini, S., Mok, L., Hudson, R., Davis, G., Chen, C., and Goldstein, R. (1985). "The 
ILWAS model: formulation and application." Water, Air, & Soil Pollution, 26(4), 
425-459. 
 
Haith, D. A., and Shoenaker, L. L. (1987). "Generalized watershed loading functions for 
stream flow nutrients." JAWRA Journal of the American Water Resources 
Association, 23(3), 471-478. 
 
Hargreaves, G. H. (1974). ?Estimation of potential and crop evapotranspiration.? 
Transcations ASAE, Vol. 17, 701-704. 
 
Herr, J., Weintraub, L., and Chen, C. (2000). "User?s Guide to WARMF: Documentation 
of Graphical User Interface." EPRI. Palo Alto, CA. Report EP-P2346/ C, 1054. 
 
Herr, J. W., Chen, C. W., Goldstein, R. A., and and, J. N. B. "A Tool for Sediment 
TMDL Development on Oostanaula Creek." Total Maximum Daily Load (TMDL) 
Environmental Regulations: Proceedings of the March 11-13, 2002 Conference, 
Fort Worth, TX. 
 
Herr, J. W., Chen, C. W., Goldstein, R. A., Herd, R., and Brown, J. M. (2003). "Modeling 
acid mine drainage on a watershed scale for TMDL calculations." JAWRA 
Journal of the American Water Resources Association, 39(2), 289-300. 
32 
 
Huber, W. C., Dickinson, R. E., and Barnwell Jr, T. O. (1988). "Storm water 
management model; version 4." Environmental Protection Agency, United States. 
 
Johanson, R. C., Imhoff, J. D., and Davis, H. H., Jr. (1980). "Users manual for 
hydrological simulation program - Fortran (HSPF)." Environmental Research 
Laboratory, US Environmental Protection Agency, Athens, GA. 
 
Keller, A. (2000). "Peer Review of the Watershed Analysis Risk Management 
Framework (WARMF)?An evaluation of WARMF for TMDL applications by 
independent experts using USEPA guidelines." EPRI, Palo Alto, CA., Report, 
1000252. 
 
Keller, A. (2001). "Peer Review of the Acid Mine Drainage Module of the Watershed 
Analysis Risk Management Framework (WARMF)?An evaluation of 
WARMF/AMD using USEPA guidelines." Electric Power Research Institute, 
Palo Alto, CA, Technical Report. 
 
McDonald, S., Bhimani, S., Weintraub, L., Herr, J., Chan, C., and Dennis, G. (2000). 
"Decision Support Process for Truckee River Watershed Management." 
Watershed Management & Operations Management 2000, ASCE, Fort Collins, 
CO. 
 
Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., and 
Veith, T. L. (2007). "Model Evaluation Guidelines for Systematic Quantification 
of Accuracy in Watershed Simulations." Transactions of the ASABE, 50(3), 885-
900. 
 
NC DENR. (2009). "Falls Lake Watershed Analysis Risk Management Framework 
(WARMF) Development: Final Report." N. C. Department of Environment and 
Natural Resources, Raleigh, NC. 
 
Neilson, B., Horsburgh, J., Stevens, D., Matassa, M., Brogdon, J., and Spackman, A. 
(2003). "Comparison of Complex Watershed Models ?Predictive Capabilities: 
EPRI?s Watershed Analysis Risk Management Framework WARMF vs. 
USEPA?s Better Assessment Science Integrating Point and Nonpoint Sources 
BASINS/WinHSPF." Final Rep. Prepared for Utah Water Research Laboratory, 
Utah State Univ., Logan, Utah. 
 
Rich, P. M., Weintraub, L. H., Chen, L., and Herr, J. "Climate Change Impacts on 
Hydrology and Water Management of the San Juan Basin." American 
Geophysical Union, Fall Meeting 2005, 1054. 
 
RMC Water and Environment. (2007). "Upper Mokelumne River Watershed Assessment 
and Planning Project." Upper Mokelumne River Watershed Authority, Oakland, 
CA. 
33 
 
Stringfellow, W., Herr, J., Litton, G., Brunell, M., Borglin, S., Hanlon, J., Chen, C., 
Graham, J., Burks, R., and Dahlgren, R. (2009). "Investigation of river 
eutrophication as part of a low dissolved oxygen total maximum daily load 
implementation." Water Science & Technology, 59(1), 9-14. 
 
Systech (2005). "Creating a WARMF 6.1 Application Using a BASINS 3.1 Delineation: 
A User's Guide." Systech Water Resources, Inc., Walnut Creek, CA. 
 
USDA (2011). ?Fertilizer use - cotton.? Fertilizer Use and Price. U.S. Department of 
Agriculture. <http://www.ers.usda.gov/Data/FertilizerUse/> (Mar. 2, 2011). 
 
USEPA. (2004). "Better Assessment Science Integrating point and Nonpoint Sources 
(BASINS) Version 3.1: User's Manual." EPA-823-C-04-004, United States 
Environmental Protection Agency. 
 
Wang, P., Shariq, L., Montague, L., Kwaan, R., and Kella, V. (2004). "Developing a 
Nutrient Management Plan for the Napa River Watershed." Donald Bren School 
of Environmental Science and Management, University of California, Santa 
Barbara, CA. 
 
Young, R. A., Onstad, C. A., Bosch, D. D., and Anderson, W. P. (1987). AGNPS, 
agricultural non-point-source pollution model: A watershed analysis tool, U.S. 
Department of Agriculture (USDA), Agricultural Research Service, Report No. 
35. Washington, D.C. 
 
Zheng, Y., and Keller, A. A. (2006). ?Understanding parameter sensitivity and its 
management implications in watershed-scale water quality modeling.? Water 
Resources Research, 42(5), W05402. 
 
 
34 
Table 2.1 2008 ADEM 303(d) List of Impaired Waters in the Saugahatchee Creek 
Watershed 
Water body 
name 
County Size Uses Causes Sources 
Pepperell Branch Lee 6.67 miles 
(10.73 km) 
Fish & 
Wildlife 
Nutrients Industrial 
      
Saugahatchee 
Creek (Yates 
Reservoir 
Embayment) 
Tallapoosa 203.78 acres 
(82.47 ha) 
Public 
Water 
Supply, 
Swimming, 
Fish & 
Wildlife 
Nutrients, 
Organic 
Enrichment 
Industrial, 
Municipal, 
Non-irrigated 
crop 
production, 
Pasture 
grazing 
 
Table 2.2 Statistical Summary for Areas of 44 Catchments and Lengths of 40 Stream 
Segments of the Saugahatchee Creek Watershed Imported to WARMF Model 
  Catchment Area (ha) Stream Segments Length (m) 
Average 1263.9 3616.4 
Standard Deviation 942.8 2820.6 
Maximum 4035.5 11694.1 
Minimum 62.0 419.8 
75th percentile 1866.2 4922.4 
50th percentile 1120.0 2753.1 
25th percentile 268.1 1483.6 
 
35 
Table 2.3 Monthly Average, Standard Deviation, Maximum, and Minimum Values of 
Meteorological Variables for 1997-2009 Period Imported into WARMF Model for the 
Saugahatchee Creek Watershed 
Precipitation
1
 Minimum Temperature
2
 Maximum Temperature
2
 
Ave Stdev Max Min Ave Stdev Max Min Ave Stdev Max Min 
J 3.35 8.09 50.80 0 2.59 6.40 17.20 -11.10 16.30 5.82 28.30 1.10 
F 4.27 10.48 63.50 0 3.97 5.46 17.80 -8.90 18.19 5.17 27.20 5.60 
M 5.06 14.85 136.14 0 7.58 5.24 20.60 -7.80 22.53 4.87 31.70 4.40 
A 3.89 11.80 104.65 0 10.85 4.62 21.70 -0.60 25.82 3.68 34.40 13.00 
M 3.10 8.91 83.57 0 15.66 3.78 23.30 5.60 29.66 2.83 35.00 21.70 
J 4.77 11.44 76.20 0 20.04 2.44 26.70 11.70 32.77 2.76 38.30 23.90 
J 3.70 9.64 103.12 0 21.83 1.51 25.60 15.00 33.82 2.21 40.00 26.70 
A 3.37 9.23 89.92 0 21.66 1.90 26.10 14.40 33.98 2.58 41.10 26.10 
S 2.98 8.87 84.58 0 18.80 3.77 25.00 6.00 31.35 2.76 37.20 21.10 
O 2.85 9.80 71.12 0 12.02 5.74 23.00 -1.00 26.72 4.13 33.30 12.80 
N 3.73 11.52 115.57 0 6.25 5.62 21.10 -6.10 21.62 4.49 31.10 7.80 
D 4.00 9.85 82.55 0 2.70 5.86 18.30 -9.40 16.85 5.26 27.80 2.80 
1
 derived from Auburn No. 2 Station 
2
 derived from Montgomery Dannelly Field Station 
 
Table 2.4 Land Use Distribution in Percentage for Three Catchments Containing 
Flow/Water Quality Monitoring Stations in the Saugahatchee Creek Watershed 
Land Use 
Categories 
Catchment 11 
(Station-16)
Catchment 22 
(Station-8)
Catchment 42 
(Yates-2) 
Deciduous Forest 18.75 32.20 23.05 
Coniferous Forest 14.08 26.53 64.29 
Mixed Forest 2.29 2.86 0.47 
Cropland/Pasture 5.41 15.56 2.26 
Rangeland 3.79 17.90 4.13 
Wetland 0.20 3.07 0.73 
Barren 0.91 0.09 0.11 
Residential 42.31 1.22 1.17 
Comm./Industrial 11.78 0.05 0.00 
Water 0.49 0.54 3.79 
TOTAL 100.00 100.00 100.00 
 
36 
Table 2.5 Calibrated Parameters of WARMF Model for the Saugahatchee Creek 
Watershed 
Parameters Units Literature Range
1
 Calibrated Value  
Precipitation Weighting Factor - 0.5 - 1.5 0.74 
Evaporation Magnitude - 0.6 - 1.4 0.91 
Evaporation Skewness - 0.6 - 1.4 0.9 
Number of Soil Layers - 1 - 5 3 
Thickness of Soil Layers cm > 0 8 ? 79 
Saturation Moisture - 0.2 - 0.6 0.35 ? 0.45 
Field Capacity - 0 - 0.4 0.18 - 0.31 
Initial Moisture - 0 - 0.6 0.25 
Horizontal Conductivity cm/day > 0 3600 ? 5600 
Vertical Conductivity cm/day > 0 1800 ? 2800 
Aeration Factor /day 0.2 - 1 0.5 
Sediment Oxygen Demand g/m
2
/day 0.1 ? 2 0.8 ? 2 
1
 From Herr et al. (2001) 
Table 2.6 List of Major Point Source Dischargers in the Saugahatchee Creek Watershed 
 Facility Name Type Latitude Longitude Discharge 
(MGD) 
1 Auburn 
Northside 
Sewerage System 32.6306 -85.5426 1.6 (0.07 m
3
/s) 
2 Opelika 
Westside 
Sewerage System 32.6607 -85.4505 4.0 (0.18 m
3
/s) 
3 West Point 
Stevens 
Process Water 
(Industrial) 
32.6294 -85.4181 1.6 (0.07 m
3
/s) 
 
Table 2.7 Model Performance for Flow Simulation during Calibration and Validation 
Period 
Statistical Measure Recommended 
Values
1
 
Calibration 
(2000-2005) 
Validation 
(2006-2009) 
NSE 
> 0.5 
0.64 0.56 
RSR 
? 0.7 
0.60 0.66 
PBIAS 
? 25% 
-2.78% -9.53% 
1
 Based on Moriasi et al. (2007)
 
 
Fig. 2.1 Location map of the Saugahatchee Creek Watershed including surrounding counties and locations of three flow and water 
quality stations 
37 
38 
 
Fig. 2.2 Definition sketch for the compartments of a catchment in WARMF model (Chen 
et al. 2001) 
Lower Tallapoosa Watershed
Saugahatchee Creek Watershed
Streams
Lakes
2002040Miles
S
N
EW
BASINS View
 
Fig. 2.3 Selection of the Saugahatchee Creek Watershed from the Lower Tallapoosa 
Watershed (HUC 03150110) in BASINS 
 
C?11
C?22
C?42
 
Fig. 2.4 Land catchments, stream segments, and reservoirs of the Saugahatchee Creek Watershed imported into WARMF including 
three major point source dischargers (Table 2.6) 
39 
 
 
Fig. 2.5 NLCD 2001 land use of the Saugahatchee Creek Watershed clipped from Tallapoosa Basin 
40 
 
0
2
4
6
8
10
12
14
16
18
20
Jan Feb Mar Apr MayJun Jul Aug Sep Oct NovDec
Phos
pha
t
e
?
(k
g
/
d
)
Auburn?Northside
2001
2002
90th?Percentile
0
20
40
60
80
100
120
140
160
Jan FebMar Apr MayJun Jul AugSep Oct NovDec
Ni
tra
t
e
?
(k
g
/
d
)
0
5
10
15
20
25
Jan Feb Mar Apr MayJun Jul Aug Sep Oct NovDec
Phos
pha
t
e
?
(k
g
/
d
)
Opelika?Westside
0
50
100
150
200
250
300
350
400
Jan FebMar Apr MayJun Jul AugSep Oct NovDec
Ni
t
r
a
t
e
?
(k
g
/
d
)
0
1
2
3
4
5
6
7
8
9
10
Jan Feb Mar AprMay Jun Jul Aug Sep Oct NovDec
Pho
s
pha
t
e
?
(k
g
/
d)
West?Point?Stevens
0
20
40
60
80
100
120
140
160
180
Jan FebMarAprMayJun Jul AugSep OctNovDec
Ni
tra
t
e
?
(k
g
/
d
)
 
Fig. 2.6 Observed phosphate and nitrate loadings in 2001, 2002 and 90
th
 percentile for Auburn Northside WWTP, Opelika Westside 
WWTP, and West Point Stevens Finishing Plant, respectively 
41 
42 
0.1
1
10
100
1000
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
St
r
e
a
m
f
l
ow
?
(m
3
/s
)
Maximum
3rd?Quartile
1st?Quartile
Minimum
Average
Median
 
Fig. 2.7 Monthly summary of observed daily flow from 2000-2009 at USGS 02418230 
station in the Saugahatchee Creek near Loachapoka 
 
0
20
40
60
80
100
120
140
Jan?00 Jan?01 Jan?02 Jan?03 Jan?04 Jan?05
S
t
re
am
f
l
o
w
?
(m
3
/s
)
Calibration?2000?05
Modeled Observed
0
20
40
60
80
100
120
140
Jan?06 Jul?06 Jan?07 Jul?07 Jan?08 Jul?08 Jan?09 Jul?09
St
r
e
a
m
f
l
ow
?
(m
3
/s
)
Validation??2006?09
Modeled Observed
 
Fig. 2.8 Flow calibration (2000-05) and validation (2006-09) at USGS 02418230 station in the Saugahatchee Creek near Loachapoka 
43 
44 
?5
0
5
10
15
20
25
30
35
Apr?00 Jul?00 Oct?00 Jan?01 Apr?01 Jul?01 Oct?01 Jan?02 Apr?02 Jul?02 Oct?02
Wa
t
e
r
?
Te
m
p
e
r
at
u
r
e
?
(?
C
)
Modeled
Observed
0
2
4
6
8
10
12
14
Apr?00 Jul?00 Oct?00 Jan?01 Apr?01 Jul?01 Oct?01 Jan?02 Apr?02 Jul?02 Oct?02
Diss
o
l
ve
d
?
Ox
y
g
e
n
?
(m
g
/
l
)
0
1
2
3
4
5
Apr?00 Jul?00 Oct?00 Jan?01 Apr?01 Jul?01 Oct?01 Jan?02 Apr?02 Jul?02 Oct?02
To
t
a
l
?
P
h
os
ph
or
us
?
(m
g
/
l
)
0
5
10
15
20
25
30
35
Apr?00 Jul?00 Oct?00 Jan?01 Apr?01 Jul?01 Oct?01 Jan?02 Apr?02 Jul?02 Oct?02
To
t
a
l
?
Ni
t
r
o
g
en
?
(m
g
/
l
)
 
Fig. 2.9 Observed and modeled water temperature, DO, TP, and TN concentration during 
calibration and validation period at Station-16 in the Pepperell Branch 
45 
?5
0
5
10
15
20
25
30
35
Apr?00 Jul?00 Oct?00 Jan?01 Apr?01 Jul?01 Oct?01 Jan?02 Apr?02 Jul?02 Oct?02
Wa
t
e
r
?
Te
m
p
e
r
a
t
u
r
e
?
(?
C
)
Modeled
Observed
0
2
4
6
8
10
12
14
16
Apr?00 Jul?00 Oct?00 Jan?01 Apr?01 Jul?01 Oct?01 Jan?02 Apr?02 Jul?02 Oct?02
Disso
lve
d
?
Ox
y
g
e
n
?
(m
g
/
l
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Apr?00 Jul?00 Oct?00 Jan?01 Apr?01 Jul?01 Oct?01 Jan?02 Apr?02 Jul?02 Oct?02
To
t
a
l
?
Ph
os
pho
r
u
s
?
(m
g
/
l
)
0
2
4
6
8
10
Apr?00 Jul?00 Oct?00 Jan?01 Apr?01 Jul?01 Oct?01 Jan?02 Apr?02 Jul?02 Oct?02
To
t
a
l
?
Ni
t
r
o
g
en
?
(m
g
/
l
)
Calibration Validation
 
Fig. 2.10 Observed and modeled water temperature, DO, TP, and TN concentration 
during calibration and validation period at Station-8 in the Saugahatchee Creek near 
Loachapoka 
46 
0
2
4
6
8
10
12
14
Jan?00 May?00 Sep?00 Jan?01 May?01 Sep?01 Jan?02 May?02 Sep?02
Disso
l
ve
d
?
Ox
y
g
e
n
?
(m
g
/
l
)
0.00
0.04
0.08
0.12
0.16
0.20
Jan?00 May?00 Sep?00 Jan?01 May?01 Sep?01 Jan?02 May?02 Sep?02
To
t
a
l
?
P
h
os
ph
or
us
?
(m
g
/
l
) Modeled
Observed
0
10
20
30
40
50
60
70
Jan?00 May?00 Sep?00 Jan?01 May?01 Sep?01 Jan?02 May?02 Sep?02
Chl
o
r
o
p
h
y
l
l
?
a
?
(
?
g/
l)
 
Fig. 2.11 Observed and modeled chlorophyll-a, DO, and TP concentration during 
calibration and validation period at Yates-2 station in the Saugahatchee Creek ( Yates 
Reservoir Embayment) 
 
 
47 
0
10
20
30
40
50
60
70
Jan?00 May?00 Sep?00 Jan?01 May?01 Sep?01 Jan?02 May?02 Sep?02
Ch
l
o
r
o
p
h
y
l
l
?
a
?
(
?
g/
l)
Reservoir Stream?Segment Observed
 
Fig. 2.12 Chlorophyll-a concentration in Yates Reservoir Embayment modeled as a 
reservoir and as a stream segment with observed values (2000-02) 
48 
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Jan?00 Feb?00 Mar?00 Apr?00 May?00 Jun?00 Jul?00 Aug?00 Sep?00 Oct?00 Nov?00 Dec?00
DO
??
St
a
n
d
a
r
d
?
De
via
t
io
n
?
(m
g
/
l
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Jan?02 Feb?02 Mar?02 Apr?02 May?02 Jun?02 Jul?02 Aug?02 Sep?02 Oct?02 Nov?02 Dec?02
DO
??
St
a
n
d
a
r
d
?
De
via
t
i
o
n
?
(m
g
/
l
)
 
Fig. 2.13 Daily standard deviation of dissolved oxygen simulated using hourly time step 
at Yates Reservoir Embayment for the year 2000 and 2002 
49 
4
5
6
7
8
9
10
0 2 4 6 8 1012141618202224
Diss
o
l
v
e
d
?
Ox
y
g
e
n
?
(m
g
/
l
)
Time?(hours)
6/19/2000
7/24/2000
8/21/2000
8/27/2002
9/26/2002
 
Fig. 2.14 Diurnal variation of DO simulated using hourly time step for days with 
observed DO less than 5 mg/l at Yates Reservoir Embayment 
0
1
2
3
4
5
6
7
8
9
0
1
2
3
0123456789
De
p
t
h
?
(ft
)
De
p
t
h
?
(m
)
Dissolved?Oxygen?(mg/l)
6/19/2000
7/24/2000
8/21/2000
8/27/2002
9/26/2002
 
Fig. 2.15 Vertical profiles of DO for days with observed DO less than 5 mg/l 
50 
Chapter 3.  Assessing the Impact on Hydrology and Water Quality in the 
Saugahatchee Creek due to Land Use Change and Climate Change using the 
WARMF Model 
 
Abstract 
The hydrology and water quality of a stream or a reservoir gets affected due to 
human activities and land use change in its watershed. Similarly, climate change, if it 
occurs, is likely to have significant impacts on local watershed system. In this study, a 
physically based watershed model WARMF was applied to the Saugahatchee Creek 
Watershed to investigate hydrologic and water quality response to historical land use 
scenarios and statistically downscaled future climate scenarios derived from HadCM3 A2 
and B2 scenarios. Based on monthly average of daily predicted values, nutrients levels 
increased for recent land use scenarios of 2001 and 2008 compared to that of 1991. Based 
on model results, monthly average of daily water temperature is predicted to rise with 
warmer and drier future climate projections. Accordingly, flow will likely decrease, 
nutrient concentration are expected to increase. Algal concentration are predicted to 
increase up to mid 21
st
 century and then decline thereafter. DO concentration are 
predicted to fall, especially in summer. The results of this study can be incorporated into 
watershed management and planning strategies after careful evaluation of uncertainties 
associated with future climate predictions, downscaling, and watershed model output.
51 
Introduction 
During the recent years, rapid urbanization has lead to massive land use changes, 
pervious forest soils have been reduced and industrial and residential areas have been 
increased. Similarly, the atmospheric concentrations of greenhouse gases and aerosols are 
believed to be increasing, thereby leading to climate change. Climate change, if it 
happens, will cause increase in temperature, evaporation, evapotranspiration, 
precipitation variability and extremes (Kundzewicz et al. 2007)). These alterations will 
change the hydrological behavior of watershed ecosystem in physical, chemical, and 
biological terms. The potential effects of land use and climate change are not limited to 
quantity; it can have serious impacts on water quality of streams and reservoirs in the 
watershed. In this study, we utilized the Watershed Analysis Risk Management 
Framework (WARMF) (Chen et al. 2001; Herr et al. 2000) to develop a regional-specific 
modeling framework to systematically assess the impact on hydrology and water quality 
in streams and reservoirs of the Saugahatchee Creek Watershed (SCW) using land use 
and future climate scenarios as model inputs. 
The Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment 
Report summarized that the linear trend over the last 50 years is warming of 0.13 (0.10 to 
0.16) ?C per decade, nearly twice that for the last 100 years, and it is projected to further 
increase by 0.2 ?C per decade for the next two decades (IPCC 2007). The IPCC 
(Kundzewicz et al. 2007) summarized the impact of climate change on freshwater 
systems are mainly due to increases in temperature and precipitation variability. Higher 
water temperatures, increased precipitation intensity, and longer periods of low flows 
52 
exacerbate water pollution, with impacts on ecosystems, human health, water system 
reliability and operating cost (Kundzewicz et al. 2007). 
The outputs from various General Circulation Models (GCMs) like Hadley Centre 
Coupled Model, version 3 (HadCM3) are available to generate future climate scenarios. 
However, the GCMs were not designed to analyze the hydrologic impact at the watershed 
scale and therefore have coarser spatial resolution as compared to what is required for 
watersheds impact studies. GCMs are inherently unable to represent watershed scale 
feature and dynamics for hydrologic impact studies (Samuels et al. 2010; Wigley et al. 
1990). To bridge this gap, the techniques have been developed to downscale GCMs 
output into local meteorological variables required for hydrological modeling, usually 
referred to as downscaling techniques. The downscaling techniques can be statistical or 
dynamic. The statistical downscaling model was used in this study, considering its 
advantages over dynamic method because statistical method is computationally 
undemanding and provides station-scale climate information based on GCM-scale output 
(Wilby and Dawson 2007). 
Hydrological impact of land use change has been investigated in a variety of 
studies using modeling methods (Breuer et al. 2009; Choi et al. 2003; Choi and Deal 
2008; Kim et al. 2002). The water quality in a watershed is affected directly by vegetative 
cover and agricultural and other land management practices (Bhattarai et al. 2008). 
Bhattarai et al. (2008) used BASINS-SWAT model to estimate the effect of land use 
change on nitrogen and phosphorus runoff and sediment deposition in a small watershed 
in the Alabama Wiregrass Region. Similarly, many researchers have pointed out adverse 
53 
effect of increasing urban land use on water quality (Ouyang et al. 2006; Schoonover et 
al. 2005; Sliva and Williams 2001; Tu et al. 2007). 
Many previous studies have assessed the impact of climate change on hydrology 
(Chang et al. 2002; Gleick and Chalecki 1999; Liu et al. 2011; Novotny and Stefan 2007; 
Xu et al. 2009; Zhang et al. 2007). Dibike and Coulibaly (2005) applied statistical 
downscaling techniques to generate future climate scenarios in the Saguenay watershed 
in Canada at local watershed scale and simulated the corresponding flow based on the 
downscaled future climate data as input to hydrological models. Rich et al. (2005) 
applied watershed model (WARMF) to assess impacts of extended droughts and 
increased temperature due to climate change on hydrology of the San Juan Basin in 
Colorado and New Mexico. Simulations showed that drought and increased temperature 
impact water availability and lead to increased frequency of critical shortages. The 
assessment of impact of climate change on water quality in the southeastern United States 
revealed that watersheds are likely to have higher nitrogen levels and lower dissolved 
oxygen problems (Cruise et al. 1999). However, very few have conducted the climate 
change impact study for water quantity as well as quality (Bouraoui et al. 2002; Cruise et 
al. 1999; Neff et al. 2000; Tu 2009). Wang (2010) investigated the individual and 
combined impact of future land use and climate change in the Wolf Bay Watershed using 
SWAT. USEPA is evaluating the impacts of land use and climate change on hydrology 
and water quality in major river basins throughout the United States using watershed 
models, HSPF and SWAT (Butcher et al. 2010). Limited water quality parameters such 
as total nitrogen, total phosphorus concentration, etc. are considered in previous impact 
studies. 
54 
 
To the best of authors? knowledge, no studies have been conducted using a 
physically based complex watershed model and downscaling 100 years? future climate 
data to evaluate the impact on flow and water quality parameters including water 
temperature, dissolved oxygen, total nitrogen, total phosphorus and algal concentration, 
due to land use and climate change, in streams of a local watershed in Alabama. The 
objective of this study was to assess the impact of historical land use change and potential 
climate change downscaled to local watershed scale based on HadCM3 future climate 
projections, on flow and above mentioned five water quality parameters in the streams of 
the Saugahatchee Creek Watershed in Alabama. The WARMF model for the SCW was 
set up, calibrated, and validated for the observed flow and water quality, and then it was 
run for land use change and future climate scenarios, and their potential impacts were 
evaluated. 
 
Study Watershed 
The watershed of concern is the Saugahatchee (otherwise known as 
Sougahatchee) Creek Watershed, located mostly in Piedmont region of eastern Alabama 
with an area of approximately 550 km
2 
(Fig. 3.1). Beginning with its headwaters in 
Chamber and Lee counties, Saugahatchee Creek runs westward through parts of Macon 
and Tallapoosa counties until it enters Yates Reservoir and converges to Tallapoosa 
River. Two segments in the SCW are listed on the State of Alabama?s 303 (d) list of 
impaired waters under the federal Clean Water Act (ADEM 2009). Pepperell branch, a 
tributary to Saugahatchee Creek is listed as impaired waters for nutrients and the portion 
55 
of Saugahatchee Creek entering Yates Reservoir (Yates Reservoir Embayment) is listed 
for nutrients and organic enrichment/dissolved oxygen. 
 
Models, Data, and Methods 
WARMF is an integrated watershed model with simulation models and databases 
under one GIS-based graphical user interface (GUI). The algorithms embedded in 
WARMF are adapted from many well established codes such as ILWAS (Integrated Lake 
Watershed Acidification Study; Chen et al. 1983; Gherini et al. 1985), ANSWERS (Areal 
Nonpoint Source Watershed Environmental Response Simulation; Beasley et al. 1980; 
Beasley and Huggins 1981), SWMM (Storm Water Management Model; Huber et al. 
1988), and WASP (Water Quality Analysis Simulation Program; Ambrose et al. 1993).  
WARMF represents a watershed by dividing it into a network of land catchments, 
river segments, and reservoirs. Land catchment is further divided into a canopy layer, a 
snowpack, and (up to five) soil layers. Each compartment is considered as a seamlessly 
connected continuously stirred tank reactor (CSTR) for flow routing and mass balance 
calculation. WARMF simulates the process of canopy interception, snowpack 
accumulation and snowmelt, infiltration through soil layers, evapotranspiration, surface 
runoff, and groundwater exfiltration to river segments. The water from the upstream river 
segment is mixed with the water in the river segment from previous time step and the 
point and nonpoint loads entering the river segment during the time step. Heat budget and 
mass balance calculation are performed to calculate the temperature and concentrations 
of various water quality constituents in each soil layer, river segment, and reservoir 
(Chen et al. 2001). 
56 
The delineated watershed map with land catchments, stream segments, and 
reservoirs , is required for WARMF, to which input data can be given and simulation 
results can be viewed. BASINS (Better Assessment Science Integrating point and 
Nonpoint Sources; USEPA 2004) provides watershed delineation tool which helps to 
delineate watershed based on DEM (Digital Elevation Map) and river network. To 
delineate watershed, BASINS Data Download tool extracts nationally derived databases 
like States boundaries, cataloging unit boundaries, DEM, National Hydrography Dataset 
(NHD), and National Land Cover Dataset (NLCD). A BASINS project can be set up for 
8-digit HUC (Hydrologic Unit Codes) watershed by selecting the geographic area of 
interest from among the entire 48 contiguous United States. The watershed delineation 
developed from BASINS can be imported into WARMF (Systech 2005). The SCW is a 
subset watershed of the Lower Tallapoosa Watershed in Alabama (HUC 03150110). 
BASINS project for HUC 03150110 was built and only those catchments layers were 
selected which drains into Saugahatchee Creek for the WARMF model. 
To run the simulation, WARMF requires land use data, meteorological data, 
atmospheric deposition, and soil data. Observed hydrology and water quality data are 
required for model calibration and validation. WARMF uses readily available data from 
online sources to predict hydrology and water quality in streams and reservoirs. Land use 
data can be directly imported into WARMF as a GIS shapefile. Land use shapefile was 
obtained for the SCW for 2001 from NLCD website (Fig. 3.2). The meteorological data 
required for WARMF includes precipitation, dew point, minimum and maximum 
temperature, cloud cover, air pressure and wind speed. National Climatic Data Center 
(NCDC) Global Summary of Day (GSOD), online climate dataset was used to download 
57 
the necessary weather data for station in Montgomery, AL (Montgomery Dannelly Field). 
The dry and wet deposition for air quality was obtained from National Atmospheric 
Deposition Program (NADP) and USEPA Clean Air Status and Trends Network 
(CASTNET) website for GAS153 and AL10 station, respectively. There are three major 
point source dischargers, which contribute to the Saugahatchee Creek. The average flow 
discharge from these point source dischargers and constituent loadings were obtained 
from data collected by Alabama Department of Environmental Management (ADEM) 
and Auburn University (ADEM 2008). 
 
Model Calibration/ Validation 
The SCW was divided into 44 land catchments, 40 stream segments and 2 lake 
layers. The hydrologic and water quality parameters can be assigned to these segments 
individually, referred as catchment coefficients, river coefficients, and reservoir 
coefficients in WARMF manual (Herr et al. 2000). There are another set of parameters 
known as system coefficients that apply to all catchments in the watershed. During the 
calibration, some of catchment, river, reservoir, and system coefficients were adjusted to 
obtain a best fit. Table 3.1 list the adjusted parameters, with their literature range, for 
model calibration.  
The model evaluation techniques to compare simulated and observed data are 
based on graphical and statistical methods. Moriasi et al. (2007) recommended, in 
addition to the graphical techniques, following quantitative statistics to be used for model 
evaluation:  
 
58 
Nash-Sutcliffe efficiency (NSE; Nash and Sutcliffe, 1970): 
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?=
?
?
=
=
n
i
meanobs
i
n
i
sim
i
obs
i
YY
YY
NSE
1
2
1
2
)(
)(
1  
(3.1)
RMSE-observations standard deviation ratio (RSR; Moriasi et al. 2007): 
?
?
=
=
?
?
=
n
i
meanobs
i
n
i
sim
i
obs
i
YY
YY
RSR
1
2
1
2
)(
)(
 
(3.2)
Percent bias (PBIAS; Gupta  et al. 1999): 
?
?
?
?
?
?
?
?
?
?
?
?
??
=
?
?
=
=
n
i
obs
i
n
i
sim
i
obs
i
Y
YY
PBIAS
1
1
)(
100)(
 (3.3)
Where, Y
i
obs
 is the i
th
 observed value, Y
i
sim
 is the i
th
 simulated value, Y
mean
 is the mean of 
the observed values, and n is the total number of observed values. The simulation of 
watershed model can be judged satisfactory if NSE > 0.5 and RSR ? 0.7 and if PBIAS is 
within ?25% for flow (Moriasi et al. 2007). 
The observed flow daily data is available at USGS 02418230 station in the 
Saugahatchee Creek near Loachapoka since 2000. The flow was calibrated for the period 
2000?2005 and validated for the period 2006?2009 (Fig. 3.3). The model simulation was 
run three years prior to the calibration period starting from 1997 to minimize the effect of 
initial unknown parameters used in the model. The simulation of flow resulted in 
satisfactory values of NSE, RSR and PBIAS for calibration (validation) to be 0.64, 0.60, 
and 9.21 % (0.56, 0.66, and 3.79 %), respectively (Table 3.2). 
59 
ADEM and Auburn University collected water quality data in the SCW in 2000?
2002 (ADEM 2008). The observed water quality data such as water temperature, 
dissolved oxygen (DO), total phosphorus (TP), and total nitrogen (TN) were available for 
the same station but not in daily basis. Therefore, water quality calibration was performed 
based on graphical techniques, which include visual comparison of simulated and 
observed data such as time series plots (Fig. 3.4). 
 
Historical Land Use Change Scenarios 
There have been significant changes in land use pattern of Saugahatchee 
watershed over the last two centuries, particularly in recent years (AWW, 2005). The 
upper watershed is undergoing rapid transition from forest to urban/developed land. The 
changes in land use distribution are expected to bring changes in water quality, including 
surface flow, nutrient runoff, and sedimentation levels (Bhattarai et al. 2008). For land 
use change scenarios, we used land use shapefiles downloaded from AlabamaView 
website (AlabamaView 2009) for the SCW for 1991, 2001 and 2008 (Fig. 3.5). From 
1991 to 2008, forest areas decreased from 80% to 72%, while urban land increased from 
4% to 8% and cropland increased from 7% to 9% ( 
Table 3.3).  
 
Climate Change Scenarios and Downscaling 
Hadley Centre Coupled Model, version 3 (HadCM3):  
HadCM3 is a coupled atmospheric-ocean general circulation model (AOGCM) 
developed at the Hadley Centre in the United Kingdom for IPCC Third Assessment 
60 
Report use (Gordan et al. 2000; IPCC 2001; Pope et al. 2000) . The spatial resolution of 
HadCM3, which is 2.5? latitude by 3.75? longitude, represents a global grid of 96 ? 73 
grid cells. This is equivalent to a surface resolution of about 417 km ? 278 km at the 
Equator, reducing to 295 km ? 278 km at 45 degrees of latitude (IPCC 2001). 
 
SRES Emission Scenarios: 
IPCC published a set of future climate scenarios in Special Report on Emission 
Scenarios (SRES) to explore the uncertainties behind potential trends in global 
development and GHG emissions (Nakivenovic et al. 2000). Four SRES storylines and 
scenario families were developed, which assumes a distinctly different direction for 
future developments. Each storyline represents different demographic, social, economic, 
technological, and environmental developments. The major scenario families are 
discussed below (Nakivenovic et al. 2000): 
? A1: The A1 storyline and scenario family describes a future world of very rapid 
economic growth, global population that peaks in mid-century and declines 
thereafter, and the rapid introduction of new and more efficient technologies. 
Major underlying themes are convergence among regions, capacity building and 
increased cultural and social interactions, with a substantial reduction in regional 
differences in per capita income. 
? A2: The A2 storyline and scenario family describes a very heterogeneous world 
with high population growth. The underlying theme is self-reliance and 
preservation of local identities. Economic development is primarily regionally 
61 
oriented and per capita economic growth and technological change are more 
fragmented and slower than in other storylines. 
? B1: The B1 storyline and scenario family describes a convergent world with the 
same low population growth as in the A1 storyline, but with rapid changes in 
economic structures toward a service and information economy, with reductions 
in material intensity, and the introduction of clean and resource-efficient 
technologies. The emphasis is on global solutions to economic, social, and 
environmental sustainability, including improved equity, but without additional 
climate initiatives. 
? B2: The B2 storyline and scenario family describes a world in which the emphasis 
is on local solutions to economic, social, and environmental sustainability. It is a 
world with moderate population growth, intermediate levels of economic 
development, and less rapid and more diverse technological change than in the B1 
and A1 storylines. While the scenario is also oriented toward environmental 
protection and social equity, it focuses on local and regional levels. 
The outputs derived from HadCM3 A2 and HadCM3 B2 were downscaled to local 
watershed scale using statistical techniques and applied to the WARMF model for the 
SCW. 
 
Downscaling: 
Future climate projected using GCMs have coarser spatial resolution than what is 
required for hydrologic impact study. For hydrologic and water quality impact studies, 
local or station scale meteorological variables are required, which can be derived using 
62 
large-scale atmospheric variables available from GCM outputs. Future climate change 
scenarios are downscaled for the SCW at daily time scales, using large-scale GCM 
outputs. SDSM (Statistical DownScaling Model; Wilby and Dawson 2007) was used for 
the purpose of downscaling meteorological variables to the local watershed. 
 In statistical downscaling techniques, the quantitative relationships are established 
between large-scale atmospheric variables (known as predictors) and local or station 
surface variables (known as predictands). The National Centers for Environmental 
Prediction (NCEP) and the National Center for Atmospheric Research (NCAR) worked 
together in a reanalysis project to produce a physically consistent retroactive record of 
more than 50 years of global analyses of atmospheric fields to support the needs of 
research and climate monitoring communities (Kalnay et al. 1996). Reanalysis project 
involved the recovery of data from many observed and measurement systems, quality 
controlled and assimilated with a data assimilation system kept unchanged over the 
reanalysis period. The main objective of reanalysis is to eliminate perceived climate 
jumps associated with changes in data assimilation system and provide consistent records 
of temperature, precipitation, winds and many other variables that describe climatic 
conditions from the past to the present. The daily NCEP/NCAR reanalysis data was 
selected to represent the large-scale predictors in the SDSM model.  NCEP/NCAR data 
has been used in several downscaling studies in different regions over the world (Dibike 
and Coulibaly 2005; Tatli et al. 2004). 
Using SDSM, the appropriate large-scale predictor variables were selected from 
the list of predictors obtained from NCEP/NCAR reanalysis data for the period of 1961-
1990, based on regression techniques, to downscale predictands (such as station 
63 
precipitation, maximum and minimum temperature). Table 3.4 lists the predictor 
variables (from NCEP/NCAR) screened for downscaling and the predictands. SDSM 
constructs a downscaling model with parameters of the model based on multiple 
regression equations, given observed daily weather data (predictand) and the selected 
large-scale NCEP predictors for the same time period. The observed daily weather data at 
Montgomery Dannely Field station from 1961?1990 was downloaded from NOAA?s 
NCDC web site. The data from 1961?1975 was used to develop the regression model and 
the model regression weights produced as a parameter file was then used to validate for 
the period from 1976?1990. The process of selecting predictors, calibration results, 
parameter file generation, and validation for downscaling precipitation in Montgomery 
Dannelly Field station are listed in Appendix A. 
The parameter file was then used to downscale future climate to local watershed 
scale based on predictor derived from HadCM3. The GCM predictor data, derived from 
the HadCM3 A2 and B2 experiments, can be obtained for any global land area through 
data portal maintained by the CCCSN along with the daily observed predictor data, which 
is  derived from NCEP/NCAR reanalyses and interpolated to the same grid as HadCM3 
(CCCSN 2010). Given the latitude and longitude of the SCW, GCM and NCEP/NCAR 
predictors were extracted for the nearest grid. The future climate projection scenarios are 
organized into allotted three time frames and termed as 2020s (2011-2040), 2050s (2041-
2170), and 2080s (2071-2099).  
Fig. 3.6 and Fig. 3.7 show the patterns of downscaled precipitation, maximum and 
minimum temperatures for the study watershed along with standard deviation 
corresponding to HadCM3 A2 and HadCM3 B2 scenarios, respectively. The monthly 
64 
mean of daily maximum and minimum temperatures are projected to increase whereas 
that of precipitation are projected to decrease, especially in summer. Table 3.5 shows the 
statistical summary of maximum temperature, minimum temperature, and precipitation 
for allotted time period of 2020s, 2050s, and 2080s for HadCM3 A2 and B2 scenarios. 
 
Results and Discussions 
For comparing results of land use change scenarios and future climate scenarios, 
flow and water quality parameters were simulated at the watershed outlet. The monthly 
average of daily values over the 30 years? period of baseline and different scenarios were 
calculated.  The anomalies of the monthly average of daily outputs to the baseline 
scenario were computed and plotted. 
 
Impact of Land Use Change on Hydrology and Water Quality 
The hydrological and water quality response of the model corresponding to three 
land use scenarios were analysed using 30 years (1981-2010) simulation results. The 
baseline scenario corresponding to land use of 1991and the other two land use change 
scenarios corresponding to land use of 2001 and 2008 were run to investigate the land use 
change effect. The anomalies of flow and other water quality parameters corresponding 
to land use of 2001 and 2008 to the baseline were further computed. 
Fig. 3.8 shows the relative change from the baseline, of monthly average of daily 
flow, for land use scenarios of 2001 ranged between -0.03 to 0.07 m
3
/s and for 2008 
ranged between -0.09 to 0.16 m
3
/s. The increase or decrease in the flow corresponding to 
land use scenarios of 2001 and 2008 were not more than 1.76% and 3.89%, respectively. 
65 
Similarly, for surface water temperature, there was very little or no change in simulated 
results between three land use scenarios (Fig. 3.8). The increase or decrease in the surface 
water temperature corresponding to land use scenarios of  2001 and 2008 were less than 
0.08% and 0.18%, respectively.  Therefore, land use change from 1991 to 2008 did not 
have significant impact on flow and surface water temperature for the simulation period. 
Surface dissolved oxygen concentration in the Yates Reservoir Embayment, in 
terms of monthly average of daily values, did not experience the expected change due to 
land use change, partly because the simulation was performed on daily time step. The 
water body with predominance of algae shows a larger fluctuations in dissolved oxygen 
than less productive water with low algal concentration (Ansa-Asare et al. 2000). The 
water bodies with algal concentration higher than 15 ?g/l are categorized as eutrophic 
and higher than 40 ?g/l as hypereutrophic waterbodies (Carlson 1977). Higher eutrophic 
level implies algal abundance and hence exhibit higher rate of photosynthesis, 
respiration, and decomposition. For eutrophic and hypereutrophic systems, the dissolved 
oxygen concentration tends to increase during day time due to algal photosynthesis 
dominating over respiration and decomposition; whereas the system will have less 
dissolved oxygen concentration during night as there is no sunlight for photosynthesis 
(O?Connor and Di Toro 1970; Odum 1956). Due to limited data availability, the 
simulation was run on daily time steps in the model, which doesn?t model diurnal 
phenomenon. The diurnal fluctuation in the Saugahatchee Creek (Yates Reservoir 
Embayment) is discussed elsewhere (Chapter 2). Fig. 3.8 shows the relative change in 
monthly average of daily DO concentration to the baseline corresponding to land use 
66 
scenarios of 2001 and 2008 was 0.14 mg/l at most. The relative change in DO 
concentration for 2001 and 2008 were not more than 1.66%. 
In the SCW, forest areas have been transformed into cropland and urban areas 
(Table 3.3). Consequently, fertilizer inputs, livestock and cattle manure, sewage, 
industrial waste, and other sources of nutrients are increased. Therefore, the major impact 
that land use change has on the watershed, is to the nutrient concentration in the streams 
and reservoirs. Land use scenario of 2001 produced TP concentration to be 3.87 to 5.87% 
greater whereas that of 2008 produced TP concentration to be 11.18 to 15.37% greater 
than baseline scenario (Fig. 3.9). The relative change in monthly average of daily TN 
concentration to the baseline corresponding to land use scenario of 2001 ranged between 
-3.71 to 15.04% and that of 2008 ranged between -16.05 to 27.75%, respectively (Fig. 
3.9). 
The algal growth in streams and reservoirs are seasonal. The growing season for 
the SCW has been identified as April through October (ADEM 2008). Monthly 
chlorophyll-a concentration under 1991 land use varied from 1.6 to 33.2 ?g/L during 
growing season. Fig. 3.9 shows the relative change in monthly average of daily algal 
concentration corresponding to land use scenario of 2001 ranged between -0.03 to 1.54 
?g/l and and that of 2008 ranged between -0.09 to 2.39 ?g/l, respectively. The increase in 
algal concentration reached as high as 15.04% and 27.75% for land use scenarios of 2001 
and 2008, respectively.  
 
Impact of Climate Change on Hydrology and Water Quality 
67 
The impact on flow and other water quality parameters due to future climate was 
analyzed under HadCM3 A2 and HadCM3 B2 future climate scenarios. Each GCMs 
output was further categorized into three time frames: 2020s (2011-2040), 2050s (2041-
2070), and 2080s (2071-2099). The simulated results for these time frames were 
compared with baseline scenario (1981-2010). Fig. 3.10 - Fig. 3.13 show the monthly 
average of daily values of simulated flow and water quality parameters under future 
climate scenarios and the relative changes in the future projected values from the baseline 
scenario.  
The monthly average of daily flow is projected to decrease corresponding to 
HadCM3 A2 and B2 scenarios. The flow is projected to decrease by 6.8 m
3
/s and 6.5 
m
3
/s at most for HadCM3 A2 and B2 scenarios, respectively (Fig. 3.10 and Fig. 3.12). 
The surface water temperature is projected to increase, especially in summer having 
increment in monthly average of daily surface water temperature reached as high as 5.5 
?C, and 4.7 ?C for HadCM3 A2 and B2 scenarios, respectively (Fig. 3.10 and Fig. 3.12). 
The monthly average of daily surface DO decreased by 1.9 mg/l, and 1.2 mg/l at most 
corresponding to HadCM3 A2 and B2 scenarios, respectively (Fig. 3.10 and Fig. 3.12). 
The decrease in DO concentration is partly because of increased water temperature as 
dissolved oxygen solubility is inversely proportional to water temperature (APHA 1992). 
 For HadCM3 A2 and B2 scenarios, nutrient concentrations are projected to 
increase in the stream. It corresponds to the flow, as low flow tends to have higher 
concentration of nutrients. The total phosphorus increased as high as 87.4% and 83.2% 
for HadCM3 A2 and B2, respectively (Fig. 3.11 and Fig. 3.13). The total nitrogen 
68 
increased as high as 80.0% and 57.5% for HadCM3 A2 and B2, respectively (Fig. 3.11 
and Fig. 3.13). 
 The algal concentration is significant only during the growing season (April 
through October). For HadCM3 A2 and B2 scenarios, algal concentrations are projected 
to increase most of the time except during late-summer of 2080s when it decreases (Fig. 
3.11 and Fig. 3.13). The algal growth rate increases with the increase in temperature up to 
the optimum level and then decreases with further increase in temperature (Chen et al. 
2001).  
 Fig. 3.14 and Fig. 3.15 are the flow duration curves that compare low-flow and 
high-flow conditions, of daily flows, for future scenarios? time period and baseline 
period. A lower probability of exceedance (< 0.1) represent a high flow that has been 
exceeded by only 10% of the values in that period. On the contrary, a higher probability 
of exceedance represents low flow conditions. Both low flows and high flows for future 
climate scenarios are expected to decrease. However, in the later part of 21
st
 century, it is 
predicted to have increased high flows (Fig. 3.14 and Fig. 3.15).  
 
Conclusions 
In this study, WARMF model was applied to assess the impact on hydrology and 
water quality in streams and reservoirs of the SCW due to historical land use change and 
projected future climate change. Land use change scenarios for 1991, 2001, and 2008 
revealed that forest areas have been reduced and urban and agricultural areas have been 
increased in the watershed. The simulation over the period of 30 years (1981-2010) 
predicted the nutrient (TP and TN) concentrations for land use scenarios of 2001 and 
69 
2008 to be higher than the baseline land use scenario of 1991.  The algal growth was 
predicted to increase during the growing seasons due to increased level of nutrients. 
The downscaled results derived from the output of HadCM3 future climate model 
showed the climate will become warmer and drier in the 21
st
 century. Temperature rise in 
these future climate scenarios are higher especially in summer. The watershed response 
to these changes, based on monthly average of daily values, resulted rise in water 
temperature, which was expected with  rise in air temperatures. Accordingly, flow was 
expected to decrease corresponding to decreasing pattern of precipitation in the 21
st
 
century. Future climate scenarios predicted higher nutrient (TP and TN) concentration. 
Algal concentration are predicted to increase up to mid 21
st
 century and then decline 
thereafter, most probably due to too much heat. DO concentration were expected to fall, 
especially in summer, with the mixed effect of rising temperature and increasing algal 
mass. 
We should not discard possible uncertainties that output of future climate models, 
downscaling, and complex watershed model could bring to the results. However, the 
impact study under different scenarios of land use and climate change, in general, gives 
us better understanding how management alternatives can be launched during watershed 
planning.  
 
70 
References 
Alabama Department of Environmental Management (ADEM). (2008). "Final total 
maximum daily load: Nutrients & OE/DO. Pepperell branch AL03150110-0201-
700 Nutrients; Sougahatchee Creek Embayment (Yates Reservoir) AL03150110-
0204-101 Nutrients & OE/DO.", Alabama Department of Environmental 
Management, Montgomery, AL. 
 
Alabama Department of Environmental Management (ADEM). (2009). ?2008 Alabama 
303(d) List.? 303(d) Information and Map. Alabama Department of 
Environmental Management 
<http://adem.alabama.gov/programs/water/wquality/2008AL303dList.pdf> (Aug. 
12, 2009). 
 
AlabamaView (2009). ?Saugahatchee Watershed Landcover.? Saugahatchee Watershed 
Project. <http://www.alabamaview.org/watershed_project.html> (Mar. 20, 2009). 
 
Alabama Water Watch (AWW). (2005). ?Saugahatchee Creek Watershed: Past, Present, 
& Future.?Saugahatchee Creek Watershed Management Plan. 
  
Ambrose, R., Wool, T., and Martin, J. (1993). "WASP 5, The Water Quality Analysis 
Simulation Program Version 5.00." ASCI Corporation, Athens, Georgia. 
 
American Public Health Association (APHA). (1992). Standard methods for the 
examination of water and wastewater, 18 ed. Washington, DC. 
 
Ansa-Asare, O. D., Marr, I. L., and Cresser, M. S. (2000). "Evaluation of modelled and 
measured patterns of dissolved oxygen in a freshwater lake as an indicator of the 
presence of biodegradable organic pollution." Water Research, 34(4), 1079-1088. 
 
Beasley, D., Huggins, L., and Monke, E. J. (1980). "ANSWERS: A model for watershed 
planning." 
 
Beasley, D. B., and Huggins, L. F. (1981). "ANSWERS, areal nonpoint source watershed 
environment response simulation: user's manual." 
 
Bhattarai, G., Srivastava, P., Marzen, L., Hite, D., and Hatch, U. (2008). "Assessment of 
economic and water quality impacts of land use change using a simple 
bioeconomic model." Environmental Management, 42(1), 122-131. 
 
Bouraoui, F., Galbiati, L., and Bidoglio, G. (2002). "Climate change impacts on nutrient 
loads in the Yorkshire Ouse catchment (UK)." Hydrology and Earth System 
Sciences, 6(2), 197-209. 
 
Breuer, L., Huisman, J. A., Willems, P., Bormann, H., Bronstert, A., Croke, B. F. W., 
Frede, H. G., Gr?ff, T., Hubrechts, L., Jakeman, A. J., Kite, G., Lanini, J., 
71 
Leavesley, G., Lettenmaier, D. P., Lindstrom, G., Seibert, J., Sivapalan, M., and 
Viney, N. R. (2009). "Assessing the impact of land use change on hydrology by 
ensemble modeling (LUCHEM). I: Model intercomparison with current land use." 
Advances in Water Resources, 32(2), 129-146. 
 
Butcher, J. B., Parker, A., Johnson, T., and Weaver, C. P. (2010). "Nationwide Watershed 
Modeling to Evaluate Potential Impacts of Climate and Land Use Change on 
Hydrology and Water Quality." Proceedings of the 2010 Watershed Management 
Conference, K. W. Potter and D. K. Frevert, eds., Madison, WI. 
 
Carlson, R. E. (1977). "A trophic state index for lakes." Limnology Oceanography, 22(2), 
361-369. 
 
Chang, H., Knight, C. G., Staneva, M. P., and Kostov, D. (2002). "Water resource 
impacts of climate change in southwestern Bulgaria." GeoJournal, 57(3), 159-
168. 
 
Chen, C., Gherini, S., Hudson, R., and Dean, J. (1983). "The Integrated Lake-Watershed 
Acidification Study. Volume 1: Model Principles and Application Procedures." 
Electric Power Research Institute report EA-3221, Palo Alto, California, USA. 
 
Chen, C. W., Herr, J. W., and Weintraub, L. (2001). "Watershed Analysis Risk 
Management Framework (WARMF): Update One?A decision support system for 
watershed analysis and total maximum daily load calculation, allocation and 
implementation. Publication No. 1005181." Electric Power Research Institute, 
Palo Alto, California. 
 
Choi, J. Y., Engel, B. A., Muthukrishnan, S., and Harbor, J. (2003). "GIS based long term 
hydrologic impact evaluation for watershed urbanization." Journal of the 
American Water Resources Association, 39(3), 623-635. 
 
Choi, W., and Deal, B. M. (2008). "Assessing hydrological impact of potential land use 
change through hydrological and land use change modeling for the Kishwaukee 
River basin (USA)." Journal of Environmental Management, 88(4), 1119-1130. 
 
Cruise, J. F., Limaye, A. S., and Al-Abed, N. (1999). "Assessment of impacts of climate 
change on water quality in the southeastern United States." Journal of the 
American Water Resources Association, 35(6), 1539-1550. 
 
Dibike, Y. B., and Coulibaly, P. (2005). "Hydrologic impact of climate change in the 
Saguenay watershed: comparison of downscaling methods and hydrologic 
models." Journal of Hydrology, 307(1-4), 145-163. 
 
Gherini, S., Mok, L., Hudson, R., Davis, G., Chen, C., and Goldstein, R. (1985). "The 
ILWAS model: formulation and application." Water, Air, & Soil Pollution, 26(4), 
425-459. 
72 
 
Gleick, P. H., and Chalecki, E. L. (1999). "The impacts of climatic changes for water 
resources of the Colorado and Sacramento San Joaquin River Basins." Journal of 
the American Water Resources Association, 35(6), 1429-1441. 
 
Gordon, C., Cooper, C., Senior, C. A., Banks, H., Gregory, J. M., Johns, T. C., Mitchell, 
J. F. B., and Wood, R. A. (2000). ?The simulation of SST, sea ice extents and 
ocean heat transports in a version of the Hadley Centre coupled model without 
flux adjustments.? Climate Dynamics, 16(2), 147-168. 
 
Gupta, H. V., Sorooshian, S., and Yapo, P. O. (1999). ?Status of automatic calibration for 
hydrologic models: Comparison with multilevel expert calibration.? Journal of 
Hydrologic Engineering, 4(2), 135-143. 
 
Herr, J., Weintraub, L., and Chen, C. (2000). "User?s Guide to WARMF: Documentation 
of Graphical User Interface." EPRI. Palo Alto, CA. Report EP-P2346/ C, 1054. 
 
Huber, W. C., Dickinson, R. E., and Barnwell Jr, T. O. (1988). "Storm water 
management model; version 4." Environmental Protection Agency, United States. 
 
IPCC. (2001). Climate Change 2001: Synthesis Report. Contribution of Working Group I 
to the Third Assessment Report of the Intergovernmental Panel on Climate 
Change, [Houghton, J. T., Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, 
X. Dai, K. Maskell, and C. A. Johnson (eds.)], Cambridge University Press, 
Cambridge, United Kingdom and New York, NY, USA. 
 
IPCC. (2007). Climate Change 2007: The Physical Science Basis. Contribution of 
Working Group I to the Fourth Assessment Report of the Intergovernmental Panel 
on Climate Change, [S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. 
B. Averyt, M. Tignor, and H. L. Miller (eds.)], Cambridge University Press, 
Cambridge, United Kingdom and New York, NY, USA. 
 
Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., 
Saha, S., White, G., and Woollen, J. (1996). "The NCEP/NCAR 40-Year 
Reanalysis Project." Bulletin of the American Meteorological Society, 77(3), 437-
471. 
 
Kim, Y., Engel, B. A., Lim, K. J., Larson, V., and Duncan, B. (2002). "Runoff impacts of 
land-use change in Indian River Lagoon watershed." Journal of Hydrologic 
Engineering, 7(3), 245-251. 
 
Kundzewicz, Z. W., Mata, L. J., Arnell, N. W., Doll, P., Kabat, P., Jimenez, B., Miller, 
K. A., Oki, T., Sen, Z., and Shiklomanov, I. A. (2007). "Freshwater resources and 
their management." Climate Change 2007: impacts, adaptation and vulnerability: 
contribution of Working Group II to the fourth assessment report of the 
Intergovernmental Panel on Climate Change, M. L. Parry, O. F. Canziani, J. P. 
73 
Palutikof, P. J. v. d. Linden, and C. E. Hanson, eds., Cambridge Univ Press, 
Cambridge, UK, 173-210. 
 
Liu, L., Liu, Z., Ren, X., Fischer, T., and Xu, Y. (2011). "Hydrological Impacts of 
Climate Change in the Yellow River Basinfor the 21st Century Using 
Hydrological Model and Statistical Downscaling Model." Quaternary 
International., in press. 
 
Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., and 
Veith, T. L. (2007). "Model Evaluation Guidelines for Systematic Quantification 
of Accuracy in Watershed Simulations." Transactions of the ASABE, 50(3), 885-
900. 
 
Nakicenovic, N., Alcamo, J., Davis, G., de Vries, B., Fenhann, J., Gaffin, S., Gregory, K., 
Grubler, A., Jung, T.Y., Kram, T., La Rovere, E.L., Michaelis, L., Mori, S., 
Morita, T., Pepper, W., Pitcher, H., Price, L., Riahi, K., Roehrl, A., Rogner, H.-
H., Sankovski, A., Schlesinger, M., Shukla, P., Smith, S., Swart, R., van Rooijen, 
S., Victor, N., and Dadi, Z. (2000): IPCC Special Report on Emissions Scenarios. 
Cambridge University Press, Cambridge, United Kingdom and New York, NY, 
USA. 
 
Nash, J. E., and Sutcliffe, J. V. (1970). ?River flow forecasting through conceptual 
models: Part 1. A discussion of principles.? Journal of Hydrology, 10(3), 282-
290. 
 
Neff, R., Chang, H., Knight, C. G., Najjar, R. G., Yarnal, B., and Walker, H. A. (2000). 
"Impact of climate variation and change on Mid-Atlantic Region hydrology and 
water resources." Climate Research, 14(3), 207-218. 
 
Novotny, E. V., and Stefan, H. G. (2007). "Stream flow in Minnesota: Indicator of 
climate change." Journal of Hydrology, 334(3-4), 319-333. 
 
O?Connor, D. J., and Di Toro, D. M. (1970). "Photosynthesis and oxygen balance in 
streams." Journal of the Sanitary Engineering Division, 96(2), 547-571. 
Odum, H. T. (1956). "Primary production in flowing waters." Limnology and 
Oceanography, 1(2), 102-117. 
 
Ouyang, T., Zhu, Z., and Kuang, Y. (2006). "Assessing impact of urbanization on river 
water quality in the Pearl River Delta Economic Zone, China." Environmental 
Monitoring and Assessment, 120(1), 313-325. 
 
Pope, V., Gallani, M. L., Rowntree, P. R., and Stratton, R. A. (2000). ?The impact of new 
physical paramterizations in the Hadley Centre climate model: HadAM3.? 
Climate Dynamics, 16, 123-146. 
 
 
74 
Rich, P. M., Weintraub, L. H., Chen, L., and Herr, J. "Climate Change Impacts on 
Hydrology and Water Management of the San Juan Basin." Amerecan 
Geophysical Union, Fall Meeting 2005, 1054. 
Samuels, R., Rimmer, A., Hartmann, A., Krichak, S., and Alpert, P. (2010). "Climate 
change impacts on Jordan River flow: Downscaling application from a regional 
climate model." Journal of Hydrometeorology, 11(4), 860-879. 
 
Schoonover, J. E., Lockaby, B. G., and Pan, S. (2005). "Changes in chemical and 
physical propertiesof stream water across an urban-rural gradient in western 
Georgia." Urban Ecosystems, 8(1), 107-124. 
 
Sliva, L., and Williams, D. D. (2001). "Buffer zone versus whole catchment approaches 
to studying land use impact on river water quality." Water Research, 35(14), 
3462-3472. 
 
Systech (2005). "Creating a WARMF 6.1 Application Using a BASINS 3.1 Delineation: 
A User's Guide." Systech Water Resources Inc., Walnut Creek, CA. 
 
Tatli, H., N?zhet Dalfes, H., and Sibel Mente (2004). "A statistical downscaling method 
for monthly total precipitation over Turkey." International Journal of 
Climatology, 24(2), 161-180. 
 
Tu, J. (2009). "Combined impact of climate and land use changes on streamflow and 
water quality in eastern Massachusetts, USA." Journal of Hydrology, 379(3-4), 
268-283. 
 
Tu, J., Xia, Z. G., Clarke, K. C., and Frei, A. (2007). "Impact of urban sprawl on water 
quality in eastern Massachusetts, USA." Environmental Management, 40(2), 183-
200. 
 
U.S. Environmental Protection Agency (USEPA). (2004). "Better Assessment Science 
Integrating point and Nonpoint Sources (BASINS) Version 3.1: User's Manual." 
EPA-823-C-04-004, United States Environmental Protection Agency. 
 
Wang, R. (2010). "Modeling hydrologic and water quality responses to changing climate 
and land use/cover in the Wold Bay Watershed, South Alabama," M.S. thesis, 
Auburn University, Auburn, AL. 
 
Wigley, T., Jones, P., Briffa, K., and Smith, G. (1990). "Obtaining sub-grid-scale 
information from coarse-resolution general circulation model output." Journal of 
Geophysical Research, 95(D2), 1943-1953. 
 
Wilby, R. L., and Dawson, C. W. (2007). "SDSM 4.2?A decision support tool for the 
assessment of regional climate change impacts." Lancaster University, UK. 
 
75 
Xu, Z. X., Zhao, F. F., and Li, J. Y. (2009). "Response of streamflow to climate change in 
the headwater catchment of the Yellow River basin." Quaternary International, 
208(1-2), 62-75. 
 
Zhang, X., Srinivasan, R., and Hao, F. (2007). "Predicting hydrologic response to climate 
change in the Luohe River basin using the SWAT model." Transactions of the 
ASABE, 50(3), 901-910. 
 
 
 
 
 
76 
Table 3.1 Key Parameter Values after WARMF Calibration 
Parameters Units Literature Range Calibrated Value 
Precipitation Weighting Factor - 0.5 ? 1.5 0.74 
Evaporation Magnitude - 0.6 ? 1.4 0.91 
Evaporation Skewness - 0.6 ? 1.4 0.9 
Number of Soil Layers - 1 ? 5 3 
Thickness of Soil Layers cm > 0 8 ? 79 
Saturation Moisture - 0.2 ? 0.6 0.35 ? 0.45 
Field Capacity - 0 ? 0.4 0.18 ? 0.31 
Initial Moisture - 0 - 0.6 0.25 
Horizontal Conductivity cm/day > 0 3600 ? 5600 
Vertical Conductivity cm/day > 0 1800 ? 2800 
Aeration Factor /day 0.2 ? 1 0.5 
Sediment Oxygen Demand g/m
2
/day 0.1 ? 2 0.8 
 
Table 3.2 WARMF Performance during Calibration and Validation Periods for Daily 
Flow at USGS 02418230 
Statistical Measure Recommended 
Values
1
 
Calibration   
(2000-2005) 
Validation  
(2006-2009) 
NSE 
> 0.5 
0.64 0.56 
RSR ? 0.7 0.60 0.66 
PBIAS ? 25% -2.78% -9.53% 
1
 Recommended values are based on Moriasi et al. (2007) for flow simulation 
 
Table 3.3 Land Use Change in the Saugahatchee Creek Watershed from 1991 to 2008 
Land Use Categories LU 1991 LU 2001 LU 2008 
Water 0.92% 1.56% 1.22% 
Forest 80.08% 71.41% 72.42% 
Urban 3.56% 5.41% 7.92% 
Rangeland 8.49% 14.29% 10.42% 
Cropland 6.94% 7.32% 8.02% 
 
77 
Table 3.4 List of Predictands (Station Climate Parameters) and Corresponding Predictors 
used in SDSM Model to Downscale GCMs Output 
Station Parameters Predictors from NCEP/NCAR and HadCM3
1
 
Precipitation 500 hPa divergence (p5zh),  
Relative humidity at 500 hPa (r500),  
Specific humidity at 500 hPa (s500),  
Relative humidity at 850 hPa (r850), and  
Specific humidity at 850 hPa (s850) 
 
Maximum temperature Mean temperature (temp), and 
500 hPa geopotential height (p500) 
 
Minimum temperature Mean temperature (temp), and  
Near surface specific humidity (shum) 
1
 variable name used in SDSM is given inside parenthesis 
 
Table 3.5 Statistical Summary of Maximum Temperature, Minimum Temperature, and 
Precipitation Downscaled Based on HadCM3 A2 and B2  
  Baseline  HadCM3 A2  HadCM3 B2 
 1961-1990 2020s 2050s 2080s 2020s 2050s 2080s
Maximum Temperature (?C) 
Average 25.0 25.4 26.5 28.2 25.4 26.2 26.9
St. Dev. 7.7 8.0 8.3 8.7 8.1 8.1 8.5
Maximum 40.0 43.7 44.8 47.3 43.8 44.6 45.4
Minimum -5.0 3.3 4.7 5.7 0.0 1.8 4.1
Minimum Temperature (?C) 
Average 12.1 13.5 14.4 15.8 13.5 14.3 15.0
St. Dev. 8.5 8.6 8.8 9.1 8.6 8.7 9.1
Maximum 26.7 28.3 29.3 31.1 26.8 28.5 28.7
Minimum -17.8 -5.8 -5.3 -4.0 -7.0 -6.2 -5.0
Precipitation (mm) 
Average 3.7 3.3 3.1 2.8 3.1 2.9 3.1
St. Dev. 10.8 9.5 9.1 8.4 9.1 8.4 8.9
Maximum 199.1 170.0 152.5 176.6 137.0 168.2 140.1
Minimum 0.0  0.0  0.0  0.0  0.0  0.0  0.0
 
 
Fig. 3.1 Location map of the Saugahatchee Creek Watershed in the Tallapoosa Basin including surrounding counties in Alabama 
78 
 
 
Fig. 3.2 2001 NLCD land use map of the Saugahatchee Creek Watershed 
79 
 
?
0
20
40
60
80
100
120
140
Jan?00 Jan?01 Jan?02 Jan?03 Jan?04 Jan?05
St
r
e
a
m
f
l
o
w
?
(m
3
/s
)
Calibration?2000?05
Modeled Observed
0
20
40
60
80
100
120
140
Jan?06 Jul?06 Jan?07 Jul?07 Jan?08 Jul?08 Jan?09 Jul?09
St
r
e
a
m
f
l
ow
?
(m
3
/s
)
Validation??2006?09
Modeled Observed
 
Fig. 3.3 Flow calibration (2000-05) and validation (2006-09) at USGS 02418230 station in the Saugahatchee Creek near Loachapoka
80 
81 
?
?5
0
5
10
15
20
25
30
35
May?00 Sep?00 Jan?01 May?01 Sep?01 Jan?02 May?02 Sep?02
Wa
t
e
r
?
Te
m
p
e
r
a
t
u
r
e
?
(?
C
)
Modeled
Observed
0
2
4
6
8
10
12
14
16
May?00 Sep?00 Jan?01 May?01 Sep?01 Jan?02 May?02 Sep?02
Di
ss
o
l
v
e
d
?
Ox
y
g
e
n
?
(m
g
/
l
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
May?00 Sep?00 Jan?01 May?01 Sep?01 Jan?02 May?02 Sep?02
To
t
a
l
?
Pho
s
p
h
o
r
us
?
(m
g
/
l
)
0
2
4
6
8
10
May?00 Sep?00 Jan?01 May?01 Sep?01 Jan?02 May?02 Sep?02
To
t
a
l
?
Ni
tro
g
en
?
(m
g
/
l
)
Calibration Validation
 
Fig. 3.4 Water quality calibration (2000-01) and validation (2002) at Station-8 near 
Loachapoka 
82 
 
Fig. 3.5 Land use change scenarios for the Saugahatchee Creek Watershed from 1991 to 
2008 
 
83 
0
5
10
15
20
25
30
35
40
45
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Av
e
r
a
g
e
?
mo
n
t
h
l
y
?
Tm
ax
?
(?
C
)
Observed
2020s
2050s
2080s
0
5
10
15
20
25
30
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Av
e
r
a
g
e
?
m
o
nt
hl
y
?
Tm
i
n
?
(?
C
)
0
1
2
3
4
5
6
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Av
e
r
a
g
e
?
m
o
nt
hl
y
?
p
r
e
c
ip
it
a
t
i
o
n
(
m
m
)
0
1
2
3
4
5
6
7
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Tm
ax
?
s
t
a
nda
r
d
?
de
v
i
a
t
i
o
n
?
(?
C)
0
1
2
3
4
5
6
7
Jan FebMarAprMayJun Jul AugSep OctNovDec
Tm
i
n
?
st
a
n
d
a
r
d
?
de
v
i
a
t
i
o
n
?
(?
C)
0
2
4
6
8
10
12
14
16
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
P
r
e
c
i
p
it
a
t
io
n
?
st
a
n
d
a
r
d
?
de
v
i
a
t
i
o
n
?
(m
m
)
 
Fig. 3.6 General trend in maximum temperature, minimum temperature, and precipitation 
corresponding to downscaled climate change scenario based on HadCM3 A2 
84 
0
5
10
15
20
25
30
35
40
Jan Feb Mar Apr May Jun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
m
o
nt
hl
y
?
Tm
ax
?
(?
C
)
Observed
2020s
2050s
2080s
0
5
10
15
20
25
30
Jan Feb Mar Apr May Jun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
m
o
nt
hl
y
?
Tm
i
n
?
(?
C
)
0
1
2
3
4
5
6
Jan Feb Mar Apr May Jun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
m
o
nt
hl
y
?
p
r
e
c
i
p
it
a
t
io
n
(
m
m
)
0
1
2
3
4
5
6
7
Jan Feb Mar Apr May Jun Jul Aug Sep Oct NovDec
Tm
ax
?
st
a
n
d
a
r
d
?
d
e
via
t
i
o
n
?
(?
C
)
0
1
2
3
4
5
6
7
Jan FebMarAprMayJun Jul AugSep OctNovDec
Tm
i
n
?
st
a
n
d
a
r
d
?
d
e
via
t
i
o
n
?
(?
C
)
0
2
4
6
8
10
12
14
16
Jan Feb Mar Apr May Jun Jul Aug Sep Oct NovDec
P
r
e
c
i
p
it
a
t
io
n
?
st
a
n
d
a
r
d
?
d
e
via
t
i
o
n
?
(
mm)
 
Fig. 3.7 General trend in maximum temperature, minimum temperature, and precipitation 
corresponding to downscaled climate change scenario based on HadCM3 B2 
85 
0
2
4
6
8
10
12
14
16
Jan FebMar Apr MayJun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
mo
nt
h
l
y
?
fl
o
w
?
(m
3
/s
)
Baseline
2001
2008
?0.15
?0.10
?0.05
0.00
0.05
0.10
0.15
0.20
Jan FebMarAprMayJun Jul AugSepOctNovDec
An
o
m
a
l
y
?
of
?
fl
o
w
?
(m
3
/s
)
2001
2008
0
5
10
15
20
25
30
35
Jan FebMar Apr MayJun Jul AugSep Oct NovDec
Av
e
r
a
g
e
?
m
o
nthl
y
?
te
m
p
e
r
a
t
u
r
e
?
(?
C
)
?0.02
?0.01
0.00
0.01
0.02
0.03
Jan FebMarAprMayJun Jul AugSepOctNovDec
An
o
m
a
l
y
?
of
?
te
m
p
e
r
a
t
u
r
e
?
(?
C
)
6
7
8
9
10
11
12
Jan Feb MarAprMayJun Jul AugSep Oct NovDec
Av
e
r
a
g
e
?
mon
t
hl
y
?
DO
?
(m
g
/
l
)
?0.04
?0.02
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
Jan Feb Mar AprMay Jun Jul Aug Sep Oct NovDec
An
o
m
a
l
y
?
of
?
DO
?
(m
g
/
l
)
 
Fig. 3.8 Anomaly of average monthly flow, surface water temperature, and DO oxygen 
concentration to the baseline corresponding to land use scenarios of 2001 and 2008 
86 
0.04
0.06
0.08
0.10
0.12
0.14
Jan FebMarAprMayJun Jul AugSepOctNovDec
Av
e
r
a
g
e
?
m
o
nt
hl
y
?
TP
?
(m
g
/
l
)
Baseline
2001
2008
0.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
Jan FebMarAprMayJun Jul AugSepOctNovDec
An
o
m
a
l
y
?
of
?
TP
?
(m
g
/
l
)
2001
2008
2.0
2.2
2.4
2.6
2.8
3.0
3.2
Jan FebMar Apr MayJun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
mont
hl
y
?
TN
?
(m
g
/
l
)
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
Jan FebMarAprMayJun Jul AugSepOctNovDec
An
o
m
a
l
y
?
of
?
TN
?
(m
g
/
l
)
0
5
10
15
20
25
30
35
40
Jan Feb Mar AprMay Jun Jul AugSep Oct NovDec
Av
e
r
a
g
e
?
mont
hl
y
?
ch
l
o
r
o
p
h
y
l
l
?
a
?
(
?
g/
l
)
?0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Jan FebMarAprMayJun Jul AugSep OctNovDec
An
o
m
a
l
y
?
of
?
ch
l
o
r
o
p
h
y
l
l
?
a
?
(
?
g/
l
)
 
Fig. 3.9 Anomaly of average monthly TP, TN, and chlorophyll-a concentration to the 
baseline corresponding to land use change scenarios of 2001 and 2008 
87 
0
2
4
6
8
10
12
14
16
Jan FebMar Apr MayJun Jul AugSep Oct NovDec
Av
e
r
a
g
e
?
mon
t
hl
y
?
fl
o
w
?
(m
3
/s
)
Baseline
2020s
2050s
2080s
?8
?7
?6
?5
?4
?3
?2
?1
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
An
o
m
a
l
y
?
of
?
fl
o
w
?
(m
3
/s
)
2020s
2050s
2080s
0
5
10
15
20
25
30
35
Jan Feb Mar AprMay Jun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
m
o
nthl
y
?
wa
t
e
r
?
te
m
p
e
r
a
t
u
r
e
?
(?
C
)
?2
?1
0
1
2
3
4
5
6
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
An
o
m
a
l
y
?
of
?
te
m
p
e
r
a
t
u
r
e
?
(?
C
)
4
5
6
7
8
9
10
11
12
Jan Feb Mar AprMay Jun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
mon
t
hl
y
?
DO
?
(m
g
/
l
)
?2.0
?1.5
?1.0
?0.5
0.0
0.5
Jan FebMar Apr MayJun Jul Aug Sep Oct NovDec
An
o
m
a
l
y
?
of
?
DO
?
(m
g
/
l
)
 
Fig. 3.10 A Anomaly of average monthly flow, surface water temperature and DO 
concentration to the baseline corresponding to climate change scenarios downscaled with 
HadCM3 A2 
88 
0.00
0.05
0.10
0.15
0.20
0.25
Jan FebMar Apr MayJun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
mont
hl
y
?
TP
?
(m
g
/
l
)
Baseline
2020s
2050s
2080s
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
Jan FebMarAprMayJun Jul AugSepOctNovDec
An
o
m
a
l
y
?
of
?
TP
?
(m
g
/
l
)
2020s
2050s
2080s
0
1
2
3
4
5
6
7
Jan FebMar Apr MayJun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
mont
hl
y
?
TN
?
(m
g
/
l
)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Jan FebMarAprMayJun Jul AugSep Oct NovDec
An
o
m
a
l
y
?
of
?
TN
?
(m
g
/
l
)
0
5
10
15
20
25
30
35
40
45
Jan Feb MarAprMayJun Jul AugSep Oct NovDec
Av
e
r
a
g
e
?
mo
nt
hl
y
?
ch
l
o
r
o
p
y
l
l
?
a
?
(
?
g/
l
)
?20
?15
?10
?5
0
5
10
15
20
25
30
Jan Feb Mar AprMay Jun Jul Aug Sep Oct NovDec
An
o
m
a
l
y
?
of
?
ch
l
o
r
o
p
h
y
l
l
?
a
?
(
?
g/
l
)
 
Fig. 3.11 Anomaly of average monthly TP, TN and chlorophyll-a concentration to the 
baseline corresponding to climate change scenario downscaled with HadCM3 A2 
89 
0
2
4
6
8
10
12
14
16
Jan FebMar Apr MayJun Jul AugSep Oct NovDec
Av
e
r
a
g
e
?
m
o
nthl
y
?
fl
o
w
?
(m
3
/s
)
Baseline
2020s
2050s
2080s
?7
?6
?5
?4
?3
?2
?1
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
An
o
m
a
l
y
?
of
?
fl
o
w
?
(m
3
/s
)
2020s
2050s
2080s
0
5
10
15
20
25
30
35
Jan Feb Mar AprMay Jun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
m
o
nthl
y
?
wa
t
e
r
?
te
m
p
e
r
a
t
u
r
e
?
(?
C
)
?2
?1
0
1
2
3
4
5
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
An
o
m
a
l
y
?
of
?
te
m
p
e
r
a
t
u
r
e
?
(?
C
)
4
5
6
7
8
9
10
11
12
Jan Feb Mar AprMay Jun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
mon
t
hl
y
?
DO
?
(m
g
/
l
)
?1.4
?1.2
?1.0
?0.8
?0.6
?0.4
?0.2
0.0
0.2
0.4
Jan FebMar Apr MayJun Jul Aug Sep Oct NovDec
An
o
m
a
l
y
?
of
?
DO
?
(m
g
/
l
)
 
Fig. 3.12 Anomaly of average monthly flow, surface water temperature and DO 
concentration to the baseline corresponding to climate change scenario downscaled with 
HadCM3 B2 
90 
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
Jan FebMar Apr MayJun Jul AugSep Oct NovDec
Av
e
r
a
g
e
?
mont
hl
y
?
TP
?
(m
g
/
l
)
Baseline
2020s
2050s
2080s
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
Jan FebMarAprMayJun Jul AugSepOctNovDec
An
o
m
a
l
y
?
of
?
TP
?
(m
g
/
l
)
2020s
2050s
2080s
0
1
2
3
4
5
6
Jan FebMar Apr MayJun Jul AugSep Oct NovDec
Av
e
r
a
g
e
?
m
o
nt
hl
y
?
TN
?
(m
g
/
l
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Jan FebMarAprMayJun Jul AugSepOct NovDec
An
o
m
a
l
y
?
of
?
TN
?
(m
g
/
l
)
0
5
10
15
20
25
30
35
40
45
Jan Feb Mar AprMay Jun Jul Aug Sep Oct NovDec
Av
e
r
a
g
e
?
mo
n
t
h
l
y
?
ch
l
o
r
o
p
h
y
l
l
?
a
?
(
?
g/
l
)
?10
?5
0
5
10
15
20
25
Jan FebMarAprMayJun Jul AugSep OctNovDec
An
o
m
a
l
y
?
of
?
ch
lo
r
o
p
h
y
l
l
?
a
?
(
?
g/
l
)
 
Fig. 3.13 Anomaly of average monthly TP, TN and chlorophyll-a concentration to the 
baseline corresponding to climate change scenario downscaled with HadCM3 B2 
91 
0
100
200
300
400
500
600
700
0.0001 0.001 0.01 0.1 1
Fl
ow
?
(m
3
/s
)
Probability?of?Exceedance
Baseline
2020s
2050s
2080s
0
2
4
6
8
10
12
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
 
Fig. 3.14 Flow duration curves for baseline and projected HadCM3 A2 future scenarios   
 
0
100
200
300
400
500
600
700
0.0001 0.001 0.01 0.1 1
Fl
ow
?
(m
3
/s
)
Probability?of?Exceedance
Baseline
2020s
2050s
2080s
0
2
4
6
8
10
12
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
 
Fig. 3.15 Flow duration curves for baseline and projected HadCM3 B2 future scenarios 
 
92 
Chapter 4.  Rainfall Depths of the 95
th
 Percentile Events in the Contiguous U.S. 
 
Preface 
Stormwater runoff is the major source of water pollution, mostly in urban and 
developed areas. Strict stormwater runoff requirements have been enforced by the U.S. 
Congress for federal projects according to Section 438 of Energy Independence and 
Security Act (EISA) of 2007, which requires federal agencies to develop and redevelop 
facilities with a footprint that exceeds 5000 square feet in a manner that maintains or 
restores the pre-development hydrology to the maximum extent technically feasible. In 
December 2009, the U.S. Environmental Protection Agency (USEPA) developed and 
published a technical guidance to help federal agencies in implementing Section 438 of 
the EISA. Two options listed in the technical guidance are: (1) Retain the 95
th
 percentile 
rainfall event and (2) Site-specific hydrological analysis. The option 1 is one of the 
stormwater management practices that manages rainfall onsite and prevents the off-site 
discharge of the precipitation from all rainfall events less than or equal to the 95
th
 
percentile rainfall event. 
This chapter is independent of earlier studies in the thesis and is developed with a 
purpose of developing 95
th
 percentile rainfall isohyetal map for the contiguous US. The 
95
th
 percentile 24-hour rainfall depths were computed following the U.S. EPA guidelines 
at 206 weather stations in the contiguous U.S. The result obtained herein may provide 
valuable information for engineers and designer to comply with Section 438 of EISA 
93 
when federal agencies need to design, construct, and maintain stormwater management 
practices for development and redevelopment projects in the contiguous U.S 
 
Abstract  
Based on Section 438 of the Energy Independence and Security Act of 2007, U.S. 
Environmental Protection Agency (USEPA) published a technical guidance that 
recommends using the 95
th
 percentile rainfall event for Green Infrastructure/Low Impact 
Development to preserve or restore the hydrology of the site. The 95
th
 storm rainfall 
depths for a few selected cities were reported in the technical guidance. In this technical 
note, the 95
th
 percentile 24-hour rainfall depths were computed following the U.S. EPA 
guidelines at 206 weather stations/cities in the contiguous U.S. The 95
th
 storm rainfall 
depths in 18 selected cities were compared with and smaller than those calculated based 
on hourly rainfall data using the method of L-moments. When compared with NOAA?s 1-
year 24-hour rainfall depths derived from TP-40 rainfall frequency atlas, the 95
th
 
percentile rainfall depths are smaller, and a correlation relation was developed between 
these two rainfall depths. A table with rainfall depths for 206 U.S. cities was generated 
for 85
th
, 90
th
, and 95
th
 percentiles, and the 85
th
 or 90
th
 percentile rainfall depths are 
extensively used in many stormwater management manuals in U.S. to determine water 
quality volume. 
 
Introduction 
When lands are developed and urbanized, the hydrologic characteristics of the 
areas are modified. Urbanization of the land usually results increase in the volume and 
94 
the peak discharge of runoff.  Stormwater runoff in urban and developed areas is one of 
the leading sources of water pollution in the United States (U.S.) and over the world.  In 
December 2007, U.S. Congress enacted the Energy Independence and Security Act 
(EISA), and Section 438 of EISA establishes strict stormwater runoff requirements for 
federal development and redevelopment projects.  In December 2009, the U.S. 
Environmental Protection Agency (USEPA) developed and published a Technical 
Guidance (USEPA 2009) to help federal agencies in implementing Section 438 of the 
EISA.  The intent of Section 438 is to require federal agencies to develop and redevelop 
applicable facilities (exceeding 5000 ft
2
) in a manner that maintains and restores 
stormwater runoff to the maximum extent technically feasible (METF) (USEPA 2009).  
It requires maintaining or restoring the predevelopment site hydrology during the 
development or redevelopment process (USEPA 2009).  U.S. EPA Technical Guidance 
(USEPA 2009) provides two options for site designers to use and to comply with Section 
438: (1) Retain the 95
th
 percentile rainfall event and (2) Site-specific hydrological 
analysis.  The option 1 is one of the stormwater management practices that manages 
rainfall onsite and prevents the off-site discharge of the precipitation from all rainfall 
events less than or equal to the 95
th
 percentile rainfall event to METF (USEPA 2009). 
 U.S. EPA Technical Guidance (USEPA 2009) defines 95
th
 percentile rainfall 
event as ?a precipitation amount which 95 percent of all rainfall events for the period of 
record do not exceed?.  U.S. EPA (2009) further defines (page 10 of the Technical 
Guidance): 
?In more technical terms, the 95
th
 percentile rainfall event is defined as the measured 
precipitation depth accumulated over a 24-hour period for the period of record that ranks 
95 
as the 95
th
 percentile rainfall depth based on the range of all daily event occurrences 
during this period.? 
U.S. EPA Guidance further suggests that implementation of Section 438 can be 
done by applying Green Infrastructure (GI)/Low Impact Development (LID) 
management approaches. GI/LID is a stormwater management strategy designed to 
maintain predevelopment site hydrology by mimicking the natural processes that involve 
infiltration, evapotranspiration, and reuse of stormwater or runoff.  Porous pavements, 
open drainages, rain barrels, rain gardens, and green roofs are few examples. These on-
site design options can be used to manage total volume of rainfall from 95
th
 percentile 
storms. The 95
th
 percentile was recommended because this storm size appears to best 
represent the volume that is fully infiltrated in a natural condition and thus should be 
managed on-site to restore and maintain this pre-development hydrology for duration, 
rate and volume of stormwater flows (USEPA 2009). The 95
th
 percentile rainfall event 
was identified because small, frequently-occurring storms produce large proportion of the 
annual precipitation volume, and the runoff from those storm events significantly alter 
the discharge frequency, rate and temperature of the runoff. 
U.S. EPA Guidance further defines: ?The 24-hour period is typically defined as 
12:00:00 am to 11:59:59 pm. In general, at least a 20-30 year period of rainfall record is 
recommended for such an analysis.?  Storms or rainfall events are defined by a minimum 
interevent time ? a time in which no rainfall occurs.  A hourly record of rainfall data may 
be converted into rainfall events using the specification of a minimum period of no 
rainfall that defines the separation of storm events (USEPA 1986).  Typical minimum 
interevent time is 4 to 5 hours to separate more or less independent storms (Wanielista 
96 
and Yousef 1993). The rainfall event or storm depth increases with the increase of the 
minimum interevent time used to define a storm event.  For design of stormwater runoff 
control structures, e.g., detention ponds, the minimum interevent time to develop rainfall 
statistics are related to the drawdown time, infiltration time, or treatment time for a given 
Best Management Practice (BMP) design (Asquith et al. 2006).  Therefore, the 24-hour 
period recommended by U.S. EPA Guidance to organize rainfall data, i.e. using daily 
rainfall data, is not optimal option, e.g., a rainfall event starting from 10:00 pm and 
ending at 2:00 am could be reported in two days.  The 90
th
 percentile storm depths for 
minimum interevent times of 6, 8, 12, 18, 24, 48, and 72 hours for 774 weather stations in 
eastern New Mexico, Oklahoma and Texas were determined using hourly rainfall data 
and presented by Asquith et al. (2006) in the USGS Professional Paper 1725 (available 
online at http://pubs.water.usgs.gov/pp1725/). The 90
th
 percentile storm depths increase 
with the minimum interevent time, for example, it was determined to be 1.92? and 3.34? 
in Houston, respectively, when the minimum interevent times used to define a storm were 
24- and 72-hours (Asquith et al. 2006).  Therefore, to specify or determine the 95
th
 
percentile rainfall event or depth using daily data recommended by U.S. EPA Guidance 
(USEPA 2009) would be misleading to engineering community and stormwater 
management agencies without considering of the drawdown time for detention pond or 
infiltration time for green infrastructures. 
In this technical note, the 95
th
 percentile rainfall depths at 206 cities or weather 
stations in the contiguous U.S. have been calculated using U.S. EPA recommended 
procedures.  Isohyetal map depicting the patterns of 95
th
 percentile rainfall depths has 
been constructed using calculated rainfall depths from the 206 stations. The tabulation of 
97 
the calculated values for all 206 stations is included. The table and isohyetal map provide 
valuable data for engineers and designer to comply with EISA Section 438 when federal 
agencies need to design, construct, and maintain stormwater management practices for 
development and redevelopment projects in the contiguous U.S.  A sensitivity analysis 
was conducted to examine variations of the 95
th
 percentile rainfall depths on the length 
(e.g., 10, 20, and 30 years) of period of rainfall records used and on specific period of 
rainfall records chosen by analyst (e.g., 10-year data from 1960?s or 1980?s).  The 95
th
 
percentile rainfall depths at 18 selected stations were also calculated using Kappa 
distribution and L moments that were derived from hourly rainfall data with 24-hour 
minimum interevent dry period, and were compared with those determined using U.S. 
EPA Guidance (USEPA 2009).  These two 95
th
 rainfall depths were compared and 
analysed.  The 85
th
 and 90
th
 percentile rainfall depths commonly used to determine the 
water quality volume (WQV) were also computed at the 206 cities and compared with the 
95
th
 percentiles. 
 
Calculating the 95
th
 Percentile Rainfall Depths Using U.S. EPA Guidance 
A long period of precipitation records, e.g., a minimum of 10 years of data, is 
recommended by U.S. EPA Guidance to determine the 95
th
 percentile rainfall depths at a 
location.  In this study, the daily rainfall data at 206 weather stations in the contiguous 
U.S. were obtained from the National Climatic Data Center (NCDC)?s Global Summary 
of Day (GSOD) product (NCDC 2010). The data inventories for the GSOD product start 
from as early as 1931 to recent. However, most of the years before 1973 showed 
discontinuity in rainfall data; with either missing data reported as ?99.99? and/or zero-
98 
rainfall reported as ?0? for the whole year. With few exceptions, the rainfall data were 
available at 206 stations for the period of 1973-2010 (38-year) for our rainfall data 
analysis. 
A computer program was developed to compute n
th
 percentile rainfall depths 
using long-term precipitation data at 206 stations in the contiguous U.S. The precipitation 
records for air temperature less than 32 ?F were considered as snowfall and were 
excluded for the analysis.  Small rainfall depths that are 2.5 mm (0.1 in.) or less were 
excluded for the percentile analysis because they generally do not produce any 
measurable surface runoff due to rainfall loss through interception, depression storage, 
and infiltration (USEPA 2009; Viessman and Lewis 2003). All the remaining daily 
rainfall data were sorted in ascending order and the n
th
 percentile was determined using 
National Institute of Standards and Technology formula (NIST/SEMATECH 2010).  For 
the n
th
 percentile of the data with N records (X
1
, X
2
? X
N
), rank (r) is computed as 
follows:  
()11
100
+?= N
n
r  
(4.1)
The rank r was then splitted into its integer component, i and decimal component, d such 
that r = i + d. After the data were arranged in ascending order, the n
th
 percentile value 
was computed as (NIST/SEMATECH 2010): 
                                                          X
1
               for i = 0 
                              X
n
th =                 X
N 
              for i = N 
                                              X
i
 + d(X
i+1
 ? X
i
)    for 0 < i < N    
 
(4.2)
 
Fig. 4.1 shows daily rainfall frequency spectrum or percentile distributions for 
Minneapolis, MN and Montgomery, AL. The 95
th
 percentile rainfall depth determined for 
99 
Minneapolis is 35.6 mm (1.4 in.) and 53.3 mm (2.1 in.) for Montgomery (Fig. 4.1 and 
Table 4.1).  The 95
th
 percentile rainfall depth determined using U.S. EPA Guidance was 
based on cumulative probability or occurrence frequency of daily rainfall events, i.e., 
representing 95% of daily rainfall events.  Therefore, the 95% of daily rainfall events in 
Minneapolis and Montgomery is less than or equal to 35.6 mm and 53.3 mm, 
respectively. 
U.S. EPA Guidance (USEPA 2009) suggests a minimum of 10 years of data and 
recommends at least a 20-30 year period of rainfall records for the 95
th
 percentile 
analysis.  A sensitivity analysis was conducted to determine the 95
th
 percentile rainfall 
depths using 10 years, 20 years, and 30 years of data.  A moving time window method 
was used for the sensitivity analysis; the 95
th
 percentile rainfall depths were determined 
for 10 years of data from 1961 to 1970, 1962 to 1971, and so on, until the end of rainfall 
record, i.e., 2010 for Minneapolis and Montgomery stations, which have additional 
rainfall data starting from 1961.  Fig. 4.2 shows the 95
th
 percentile rainfall depths 
determined using moving time window for consecutive 10, 20, and 30 years of daily 
rainfall data for Minneapolis and Montgomery.  When 10 years of data were used, the 
95
th
 percentile rainfall depths ranged from 30.5 to 40.6 mm (1.2 to 1.6 in.) with standard 
deviation of 2.5 mm (0.10 in.) for Minneapolis and 43.2 to 53.4 mm (1.7 to 2.1 in.) with 
standard deviation of 3.0 mm (0.12 in.) for Montgomery. The variation of the 95
th
 
percentile rainfall depths determined using 10 years of data is up to 12.7 mm (0.5 in.) for 
these two stations, and calculated depths are sensitive to which period of records was 
used (Fig. 4.2).  When 30 years of data were used, the 95
th
 percentile rainfall depths 
ranged from 33.0 to 38.1 mm (1.3 to 1.5 in.) with standard deviation of 1.3 mm (0.05 in.) 
100 
for Minneapolis and 50.8 to 53.3 mm (2.0 to 2.1 in.) with standard deviation of 0.02 for 
Montgomery (Fig. 4.2). When 20 years of data were used, the ranges of the 95
th
 
percentile rainfall depths were about the same (Fig. 4.2) but the standard deviations were 
slightly larger: 2.0 and 1.0 mm (0.08 and 0.04 in.) for Minneapolis and Montgomery, 
respectively. These results indicate that the 95
th
 percentile rainfall depths determined 
using long-term data, e.g., 20 or more years, have smaller variations (< 0.1?); it can be 
claimed that these values are essential the same regardless the rainfall data used in 1960s 
or 1990s (Fig. 4.2).  Therefore, the 95
th
 percentile rainfall depths determined using 38 
years rainfall data from 1973 to 2010 are accurate and representative to historical rainfall 
conditions. The values of the 95
th
 percentile rainfall depths developed at 206 stations in 
the contiguous U.S. are listed and reported in Table 4.1 for engineering and stormwater 
management communities to use. 
The 95
th
 percentile rainfall depths determined for 206 stations in the contiguous 
U.S. ranged from 17.8 mm to 63.5 mm (0.7 in. to 2.5 in.) and was plotted over a U.S. 
map (Fig. 4.3). Isohyetal map depicting the patterns of 95
th
 percentile rainfall depths was 
constructed using calculated rainfall depths from the 206 stationsfollowed by isohyetal 
maps for 90
th
 percentile (Fig. 4.4) and 85
th
 percentile (Fig. 4.5). Gray scale colors in these 
figures correspond to change or variation of percentile rainfall depths: darker for higher 
values. For central and eastern U.S., the 95
th
 percentile rainfall depths increase from north 
to south, e.g., 27.9 mm (1.1 in.) at Sault St. Marie, MI to 63.5 mm (2.5 in.) at Mobile, AL 
(Fig. 4.3). In the western U.S., there are regions with low rainfall forming decreasing 
contours. The minimum 95
th
 percentile value is 27.9 mm (0.7 in.) obtained at Rock 
Springs, WY; Pendleton, OR; Tonopah, NV; Eagle, CO; and Boise, ID (Table 4.1).  The 
101 
low rainfall in the western U.S. is explained by rain shadow effect due to mountain 
ranges, notably Sierra Nevada and Cascades, which impede the rainfall in the area in the 
lee of the mountain (Haylock and Nicholls 2000). The southeastern U.S. is relatively wet 
with higher 95
th
 percentile rainfall depths, as it receives tropical rainstorms from the Gulf 
of Mexico.  The maximum 95
th
 percentile value is 63.5 mm (2.5 in.) obtained at Mobile, 
AL; Port Arthur, TX, and New Orleans, LA (Table 4.1). 
Hirschman and Kosco (2008) reported the 95
th
 percentile rainfall depths for 
selected stations in U.S. that were also reported by U.S. EPA Guidance (USEPA 2009) 
and are listed in Table 4.2 for a comparison.  The 95
th
 percentile rainfall depths 
determined in this study are generally in agreement with ones reported by Hirschman and 
Kosco (2008) except for Seattle in Washington (Table 4.2).  For Seattle, the relative 
difference was 37.5% or 15.2 mm (0.6 in.).  In a large city, there are several weather 
stations in and surrounding the city; and it is possible that rainfall data from different 
weather stations were used.  In the state of Washington, the 95
th
 percentile rainfall depth 
at Seattle is 25.4 mm (1.0 in.), and at Quillayute, close to the Pacific Ocean, is 50.8 mm 
(2.0 in.) (Table 4.1).  This indicates that there is a large variation of rainfall in that region. 
The upper percentiles (e.g., 85
th
 and 90
th
) of daily (24-hr) rainfall depths have 
been widely applied to numerous urban stormwater management manuals (Haubner et al. 
2001; NYSDEC 2001) and water quality management plans (Riverside County 2004; 
Williardson and Walden 2004).  Therefore, the 85
th
 and 90
th
 percentiles of daily rainfall 
depths at 206 stations were also computed and reported in Table 4.1.  In some states of 
the contiguous U.S., e.g., Georgia, the upper percentiles are about the same for different 
stations, but for some other states they are different from station to station, for example, 
102 
the 90
th
 percentiles in New York ranged from 20.3 to 30.5 mm (0.8 to 1.2 in.) (Table 4.1) 
and are the same reported by NYSDEC (2001).  In Los Angeles County, the 90
th
 
percentiles were determined at 90 gaging stations and ranged from 7.6 mm (0.3 in.) to 
38.1 mm (1.5 in.) (Williardson and Walden 2004).  It is necessary to exercise caution 
when using the upper percentiles in Table 4.1 for regions with large rainfall variations. 
For 206 cities in contiguous U.S., the 85
th
 percentiles of daily rainfall depths ranged from 
10.2 mm (0.4 in.) to 38.1 mm (1.5 in.) (Table 4.1) with average of 22.9 mm (0.9 in.) and 
standard deviation of 5.1 mm (0.2 in.); and the 90
th
 percentiles of daily rainfall depths 
ranged from 12.7 mm (0.5 in.) to 45.7 mm (1.8 in.) (Table 4.1) with average of 27.9 mm 
(1.1 in.) and standard deviation of 7.6 mm (0.3 in.). 
 
Estimated 95
th
 Percentile Rainfall Depths Using Hourly Data 
To analyze storm characteristics, storms or rainfall events are generally defined 
using a minimum interval of no rainfall, which is known as a minimum interevent time. 
The U.S. EPA (1986) performed statistical analysis of hourly rainfall in the contiguous 
U.S. (dividing into six zones) using a minimum interevent dry period of 3 to 4 hours.  
Driscoll et al. (1989) studied mean rainfall volume and mean duration of runoff-
generating events with a minimum interval between rainfall midpoints of 6 hours.  
Asquith et al. (2006) performed statistical analysis of hourly rainfall data using minimum 
interevent times of 6, 8, 12, 18, 24, 48, and 72 hours and reported L-moments and 
percentiles of storm depth and duration for 774 rainfall-gauging stations in Eastern New 
Mexico, Oklahoma, and Texas. The percentile storm depths depend on the minimum 
interevent times used, for example, the 90
th
 percentile rainfall depth for Abilene, TX is 
103 
reported to be 25.9 mm (1.02 in.) for minimum interevent time of 6 hours and 52.1 mm 
(2.05 in.) for 72 hours (Asquith et al. 2006). Results from Asquith et al. (2006) using the 
minimum interevent time of 24 hours were used to determine the 95
th
 percentile storm 
depths that are compared with the 95
th
 percentile rainfall depths using U.S. EPA 
Guidance developed in the above section. 
The data used by Asquith et al. (2006) were National Weather Service hourly 
rainfall data obtained from Hydrosphere (2003) from 1940-2002 with total of 155 million 
values of hourly rainfall from 774 stations.  For each of the minimum interevent times, 
the time series of hourly rainfall for each station was separated into sequences of storms 
for subsequent statistical analysis to develop L-moments and fit Kappa distribution 
(Asquith et al. 2006). 
L-moments are linear combination of the ranked observations (X
i:n
) used for 
probability distribution fitting. For the n sample of data (storm depths) in ascending 
order, X
1:n
 ? X
2:n
 ? ??. ? X
n:n
, L-moments of a probability distribution is defined by 
Hosking and Wallis (1997) as follows: 
Mean: 
()
1:11
2
1
XE=?  
(4.3)
L-scale: 
()
2:12:22
2
1
XXE ?=?  
(4.4)
 
()
3:13:23:33
2
3
1
XXXE +?=?  
(4.5)
 
()
4:14:24:34:44
33
4
1
XXXXE ?+?=?  
(4.6)
 
 
104 
 
The dimensionless L-moment ratios are 
 
1
2
?
?
=?CVL  
(4.7)
 
2
3
?
?
=? skewL  
(4.8)
 
2
4
?
?
=? kurtosisL  
(4.9)
A quantile function is the inverse of a cumulative distribution function for a 
random variable X, in this case storm depth, and can be defined as: 
 () ( )
FxFX ?= ?  
(4.10)
where X(F) is the variable for non-exceedance probability, F; ? is the arithmetic mean of 
the variable; and x(F) is the dimensionless quantile function (a dimensionless frequency 
curve).  The dimensionless Kappa distribution as quantile function for storm depth was 
preferable in terms of quality of distribution fit (Asquith et al. 2006) and is given as 
(Hosking 1994): 
 
()
?
?
?
?
?
?
?
?
?
?
?
?
?
?
?
? ?
?+=
?
?
?
?
h
F
Fx
h
1
1  
(4.11)
Where, ?, ?, ? and h are parameters of Kappa distribution. These four parameters can be 
computed using the mean, L-CV, L-skew and L-kurtosis (Hosking and Wallis 1997). 
However, Kappa parameter estimation is not manually tractable. Hosking and Wallis 
(1997) report that there are no simple expressions for the Kappa parameters in terms of 
the L-moments. Newton-Raphson iteration can be used for parameter estimation, as 
described by Hosking (1997).  Fig. 4.6 shows an example of non-exceedance probability 
105 
or percentile distribution of daily rainfall depths at Abilene Regional Airport, Texas, and 
quantile values of rainfall depths computed from Kappa distribution using Kappa 
parameters derived from 38-year daily rainfall data (1973 to 2010).  Fig. 4.6 demonstrates 
that Kappa distribution describes percentiles of daily rainfall depths well, and Asquith et 
al. (2006) also showed that Kappa distribution represents well non-exceedance 
probability distribution of storm depths derived from hourly rainfall data using minimum 
interevent time of 24 hours. 
In this study, the 95
th
 percentile of storm depth using hourly data was calculated 
using Kappa distribution parameters derived from site-specific L-moment ratios that were 
calculated by Asquith et al. (2006).  L-moment ratios for 18 selected stations in the 
eastern New Mexico, Oklahoma and Texas were extracted from the USGS Professional 
Paper 1725 (Asquith et al. 2006) and used to calculate Kappa distribution parameters 
(Table 4.3) for each station separately using Newton-Raphson iteration. These 18 
selected stations have 33 to 63 years of hourly rainfall record to compute L-moment 
ratios, and mean storm depth with minimum interevent time of 24 hours ranged from 6.4 
mm (0.25 in.) to 16.8 mm (0.66 in.). The quantiles of storm depths having minimum 
interevent time of 24 hours were calculated using Equation (4.10) and (4.11), e.g., using 
F = 0.95 to compute the 95
th
 percentiles of storm depths for 18 stations, results are listed 
in Table 4.4.  The 95
th
 percentile rainfall depths computed from daily rainfall data using 
U.S. EPA Guidance (USEPA 2009) from these 18 stations were originally given in Table 
4.1 as a subset of the 206 stations and listed again in Table 4.4 for easy comparison.  
Differences of 95
th
 percentile rainfall or storm depths estimated between from daily 
106 
rainfall data and hourly rainfall data (Kappa distribution) are listed in Table 4.4 and 
ranged from -22.9 to 2.5 mm (-0.9 to 0.1 in.). 
The 95
th
 percentile rainfall depths computed using the daily rainfall data are less 
than the 95
th
 storm depths calculated using Kappa distribution with parameters derived 
from hourly rainfall data except El Paso, Texas (dry and desert area).  Underestimation of 
the 95
th
 percentile 24-hr storm depths for the 18 stations was up to 32.0% when daily 
rainfall data was used.  U.S. EPA Guidance (USEPA 2009) states that the 95
th
 percentile 
rainfall event represents a precipitation amount that 95 percent of all rainfall events for 
the period of record do not exceed.  It is relatively easy to compute the rainfall depth for 
the 95
th
 percentile rainfall event based on daily rainfall data (e.g., from NOAA). 
Underestimate of the 95
th
 percentile rainfall depths can be significant in comparison to 
the 95
th
 percentile of storm depths estimated from hourly data organized as storm events 
using minimum interevent time of 24 hours.  Fig. 4.7 (left) shows regression equation 
between the 95
th
 percentile daily rainfall depth and the 95
th
 percentile 24-hr storm depths 
(derived from hourly data) with Pearson?s correlation coefficient (R) of 0.89 for 18 
selected stations studied at the confidence level of 95% (p-value < 0.05). 
 
Discussions 
The U.S. Weather Bureau Technical Paper No. 40 (TP-40) (Hershfield 1963) is a 
rainfall frequency atlas of the United States that includes 24-hour rainfall depths for 
return periods from 1 to 100 years. The 1-year 24-hour rainfall depths for the 18 weather 
stations in Table 4.4 were adopted from TP-40 and listed in the first column of Table 4.5 
for comparison. The 1-year 24-hour rainfall depths for the 18 weather stations are on the 
107 
average 54.0% (ranged from 18.2% to 85.0%) or 25.4 mm (1.0 in.) (ranged from 5.1 to 
43.2 mm or 0.2 to 1.7 in.) larger than the 95
th
 percentile of rainfall depths estimated from 
daily rainfall data. The 95
th
 percentile rainfall depth calculated using U.S. EPA guidance 
correlated well with NOAA?s 1-year 24-hour rainfall (Fig. 4.7) with Pearson?s correlation 
coefficient (R) of 0.93, which implies strong linear correlation between them at the 
confidence level of 95% (p-value < 0.05). 
U.S. EPA suggests that water quality volume (WQV) for BMPs is the storage 
needed to capture and treat ?the runoff from 90% of average annual rainfall? for BMPs in 
urban watersheds (Clar et al. 2004). The Iowa Stormwater Management Manual (ISU 
2009) states ?In numerical terms, it is equivalent to the rainfall depth in inches (the 90% 
cumulative frequency rainfall depth) multiplied by the volumetric runoff coefficient (R
v
) 
for the site, and the site drainage area.?  The 90% cumulative frequency rainfall depth is 
the same as 90
th
 percentile rainfall depth based on cumulative probability or occurrence 
frequency of daily rainfall events, i.e., representing 90% of daily rainfall events.  
However, ?the runoff from 90% of average annual rainfall? suggested by U.S. EPA (Clar 
et al. 2004) could be interpreted as cumulative percent of average annual rainfall 
multiplied by the volumetric runoff coefficient.  The 90 percent of average annual rainfall 
can be determined by sorting long-term 24-hr (daily) rainfall data from the lowest to the 
highest and then computing cumulative rainfall depths and corresponding cumulative 
percents of average annual rainfall over K years of records. 
Cumulative percent of rainfall depth, X
i
 = 
()
()
?
?
=
=
=
N
i
i
m
i
i
KX
KX
Xi
1
1
 
(4.12)
 
108 
Where, m is the number of daily rainfall depths that are less than or equal to the rainfall 
record X
i
 from the total number of N records (X
1
, X
2, 
?, X
i
, ? X
N
) at a weather station. 
The Iowa Stormwater Management Manual (ISU 2009) shows that, using 46-year (1960-
2006) rainfall data at Ames, IA, ?90.6% of the rainfall events (greater than 0.1 inch) had 
a depth of 1.25 inches or less?, and this is the 90
th
 percentile rainfall depth similar to the 
depth reported in Table 4.1 for current study.  Based on cumulative percent of average 
annual rainfall, BMP to capture and treat the runoff of 90% of average annual rainfall is 
2? rainfall at Ames, IA (ISU 2009), and this is 60% larger than the 90
th
 percentile at 
Ames.  Pan et al. (2009) computed design rainfalls for water quality volume at 31 major 
cities in China.  Design rainfall determined for Beijing, China was 35 mm (1.4 in.) using 
90
th
 percentile of daily rainfall events and 70 mm (2.8 in.) using 90% (cumulative 
percent) of average annual rainfall (Pan et al. 2009). 
Using cumulative percent of average annual rainfall as described in Equation 
(4.12), the rainfall depths for BMP to capture and treat the runoff of 90% of average 
annual rainfall were determined at 206 U.S. cities. For example, the rainfall depths from 
90% of average annual rainfall are 45.7 mm (1.8 in.) in Minneapolis and 66.0 mm (2.6 
in.) in Montgomery (Fig. 4.8), which are much larger than the 90
th
 percentile depths 
determined by frequency of the number of events (Table 4.1).  Statistical summary of 
rainfall depths that can capture 90% of average annual rainfall based on cumulative 
percent of rainfall depth (not by number of events) at 206 stations in the contiguous U.S. 
is given in Table 4.6 including statistical summary of 90
th
 and 95
th
 percentile rainfall 
depths for comparison.  The ratio of 90% of average annual rainfall (cumulative percent) 
and 90
th
 percentile of daily rainfall events was calculated for each of the 206 stations, and 
109 
the statistical summary of the ratio is given in Table 4.6 (the last column).  On the 
average, 90% of average annual rainfall determined from cumulative percent of daily 
rainfall depths is about 88% larger than the 90
th
 percentile rainfall depth determined from 
cumulative occurrence frequency of daily rainfall events.  This is similar to the results 
from Pan et al. (2009) and reported by the Iowa Stormwater Management Manual (ISU 
2009).  Unfortunately, U.S. EPA (Clar et al. 2004) did not specifically explain how ?the 
runoff from 90% of average annual rainfall? should be quantified.  Based on the data 
analysis at the 206 stations in the contiguous U.S., depending on interpretation on design 
rainfall, the water quality volume of a BMP can be quite different. 
 
Conclusions 
Rainfall percentile statistics is valuable for estimating design storms, e.g., for 
water quality improvement, low impact development, and green infrastructures. The 95
th
 
percentile rainfall, recommended value for design storm by U.S. EPA, was determined 
and reported at 206 stations/cities in the contiguous U.S. using daily rainfall data from 
1973 to 2010.  The 95
th
 percentiles ranged from 17.8 mm (0.7 in.) to 63.5 mm (2.5 in.).  
The sensitivity analysis has indicated that the 95
th
 percentile rainfall depths determined 
using long-term data, e.g., 20 or more years, have small variations (< 0.1 in.) and are 
independent of data period used.  Comparing with the 95
th
 percentile rainfall depths 
derived from hourly data at 18 selected stations in eastern New Mexico, Oklahoma and 
Texas, the 95
th
 percentile rainfall depths derived from daily rainfall record was typically 
underestimated.  The 95
th
 percentile daily rainfall depths are linearly corrected well with 
NOAA?s 1-yr 24-hour rainfall depth reported in TP-40 but 95
th
 percentiles are smaller.  
110 
Water quality volume determined using 90% cumulative percent of average annual 
rainfall is on average 88% larger than one determined using 90
th
 percentiles of daily 
rainfall. 
 
 
 
 
 
111 
References 
Asquith, W. H., Roussel, M. C., Cleveland, T. G., Fang, X., and Thompson, D. B. (2006). 
"Statistical characteristics of storm interevent time, depth and duration for eastern 
new Mexico, Oklahoma, and Texas." Professional Paper 1725, Rep. No. 5-4194-
01-1, U.S. Geological Survey, Austin, TX. 
Clar, M. L., Barfield, B. J., and O'Connor, T. P. (2004). "Stormwater best management 
practice design guide: Volume 1 general considerations." Rep. No. EPA/600/R-
04/121, U.S. Environmental Protection Agency, Cincinnati, OH. 
Driscoll, E. D., Palhegyi, G. E., Strecker, E. W., and Shelley, P. E. (1989). "Analysis of 
storm event characteristics for selected rainfall gages throughout the United 
States." U.S. Environmental Protection Agency, Washington, D.C. 
Guo, J. C. Y., and Urbonas, B. (1996). "Maximized detention volume determined by 
runoff capture tatio." Journal of Water Resources Planning and Management, 
122(1), 33-39. 
Haubner, S., Reese, A., Brown, T., Claytor, R., and Debo, T. (2001). "Unified stormwater 
sizing criteria." Georgia Stormwater Management Manual Volume 2 Technical 
Handbook, <http://www.georgiastormwater.com/vol2/1-3.pdf> (Nov. 6, 2010) 
Haylock, M., and Nicholls, N. (2000). "Trends in extreme rainfall indices for an updated 
high quality data set for Australia, 1910-1998." International Journal of 
Climatology, 20(13), 1533-1541. 
Hershfield, D. M. (1963). "Rainfall frequency atlas of the United States for durations 
from 30 minutes to 24 hours and return periods from 1 to 100 years." U.S. 
Weather Bureau, U.S. Department of Commerce, Washington, D.C. 
Hirschman, D., and Kosco, J. (2008). "Managing stormwater in your community: A 
guide for building an effective post-construction program." EPA Publication No. 
833-R-08-001, Center for Watershed Protection, Ellicott City, MD. 
Hosking, J. R. M. (1994). "The four-parameter kappa distribution." IBM Journal of 
Research and Development, 38(3), 251-258. 
Hosking, J. R. M. (1997). "Fortran routines for use with the method of L-moments, 
Version 3.02." Research Rep. RC-20525, IBM Research Division, T. C. Watson 
Research Center, Yorktown Heights, NY. 
Hosking, J. R. M., and Wallis, J. R. (1997). Regional frequency analysis: an approach 
based on L-moments, Cambridge Univ Pr., Cambridge, U.K. 
112 
Hydrosphere. (2003). "NCDC hourly precipitation - West." (CD-ROM), Hydrosphere 
Data Products Inc., v. 13.2, Boulder, CO.  
Iowa State University (ISU). (2009). "Stormwater hydrology. " Iowa Stormwater 
Management Manual,  <http://www.intrans.iastate.edu/pubs/stormwater/ 
index.cfm> (Mar. 17, 2011). 
National Climatic Data Center (NCDC). (2010). "Global summary of day. "NNDC 
climate data online, <http://www.ncdc.noaa.gov/oa/climate/onlineprod/ 
drought/xmgr.html> (May 18, 2009). 
NIST/SEMATECH. (2010). "Product and process comparisons." e-Handbook of 
statistical methods,  <http://www.itl.nist.gov/div898/handbook/ > (Aug. 20, 
2010).  
New York State Department of Environmental Conservation (NYSDEC). (2001). " 
Unified stormwater sizing criteria." New York State Stormwater Management 
Design Manual. <http://www.dec.ny.gov/chemical/29072.html> (Nov. 6, 2010). 
Pan G., Che W., Li J., and Li H.Y. (2008). ?Urban runoff pollution control quantity and 
its design rainfall in China.?  China Water & Wastewater, 24(22), 25-29. 
Riverside County. (2004). "Water quality management plan for urban runoff." Storm 
Water Clean Water Protection Program, Riverside County, CA. 
U.S. Environmental Protection Agency (USEPA). (1986). "Methodology for analysis of 
detention basins for control of urban runoff quality." EPA 440/5-87-001, U.S. 
Environmental Protection Agency, Washington, D.C. 
USEPA. (2009). "Technical Guidance on Implementing the Stormwater Runoff 
Requirements for Federal Projects under Section 438 of the Energy Independence 
and Security Act." EPA 841-B-09-001, U.S. Environmental Protection Agency, 
Washington, D.C. 
Viessman, W., and Lewis, G. L. (2003). Introduction to hydrology, Pearson Education, 
Upper Saddle River, NJ. 
Wanielista, M. P., and Yousef, Y. A. (1993). Stormwater management, John Wiley & 
Sons, Inc., New York, NY. 
Williardson, B., and Walden, A. (2004). "Analysis of 85th percentile 24-hour rainfall 
depths within the County of Los Angeles." Water Resources Division/ Hydrology 
Section, Department of Public Works, County of Los Angeles, CA. 
113 
Table 4.1 The 85
th
. 90
th
, 95
th
 Percentile, and 90% Cumulative Rainfall Depths (in.) 
Derived Using Daily Rainfall Data for 206 Weather Stations or Cities in the Contiguous 
U.S. 
State Station/City Latitude Longitude 
Percentile Rainfall Depth (in.) 90% Cumulative 
Rainfall Depth (in.)  85
th
90th 95th 
AL Birmingham 33.564 -86.754 1.2 1.5 2.0 2.4
AL Huntsville 34.644 -86.786 1.2 1.4 2.0 2.6 
AL Mobile 30.688 -88.246 1.4 1.8 2.5 3.6 
AL Montgomery 32.301 -86.394 1.3 1.6 2.1 2.6 
AR Fort Smith 35.334 -94.365 1.1 1.4 1.9 2.5 
AR Little Rock 34.916 -92.146 1.3 1.6 2.1 2.6 
AZ Flagstaff 35.140 -111.672 0.7 0.9 1.3 1.4 
AZ Phoenix 33.443 -111.990 0.7 0.8 1.1 1.3 
AZ Prescott 34.652 -112.421 0.6 0.8 1.0 1.1 
AZ Tucson 32.131 -110.955 0.7 0.9 1.2 1.7 
CA Bakersfield 35.434 -119.056 0.5 0.6 0.8 1.8 
CA Fresno 36.780 -119.719 0.6 0.8 0.9 1.0 
CA Los Angeles 33.938 -118.406 1.1 1.3 2.0 4.7 
CA Sacramento 38.513 -121.493 0.8 0.9 1.2 1.5 
CA San Diego 32.735 -117.169 0.9 1.1 1.5 1.7 
CA San Francisco 37.620 -122.398 0.9 1.1 1.5 2.1 
CA Santa Maria 34.916 -120.465 0.9 1.1 1.5 2.3 
CO Buckley 39.702 -104.752 0.7 0.9 1.2 1.6 
CO Colorado Spring 38.812 -104.711 0.6 0.8 1.0 1.3 
CO Denver 39.833 -104.658 0.6 0.8 1.0 1.3 
CO Eagle 39.643 -106.918 0.5 0.6 0.7 0.9 
CO Grand Junction 39.134 -108.538 0.5 0.6 0.8 4.1 
CO Pueblo 38.290 -104.498 0.6 0.6 0.9 1.0 
CT Bridgeport 41.175 -73.146 1.0 1.2 1.6 1.9 
CT Hartford 41.938 -72.683 1.0 1.3 1.7 2.1 
DC Washington 38.865 -77.034 0.9 1.1 1.5 1.7 
DE Wilmington 39.673 -75.601 1.0 1.2 1.6 1.9 
FL Daytona Beach 29.177 -81.060 1.1 1.4 2.0 2.6 
FL Jacksonville 30.494 -81.693 1.1 1.5 2.0 2.6 
FL Miami 25.824 -80.300 1.2 1.5 2.1 2.9 
FL Tallahassee 30.393 -84.353 1.4 1.7 2.4 3.2 
FL Tampa 27.961 -82.540 1.2 1.4 1.9 2.4 
FL West Palm Beach 26.685 -80.099 1.1 1.4 2.1 3.0 
GA Athens 33.948 -83.327 1.1 1.4 1.8 2.4 
GA Atlanta 33.640 -84.427 1.1 1.4 1.8 2.2 
GA Augusta 33.370 -81.965 1.1 1.3 1.7 2.3 
GA Columbus 32.516 -84.942 1.1 1.4 1.9 2.4 
GA Macon 32.688 -83.654 1.1 1.4 1.8 2.3 
GA Savannah 32.119 -81.202 1.1 1.3 1.9 2.7 
IA Des Moines 41.538 -93.666 1.0 1.2 1.7 2.1 
IA Sioux City 42.391 -96.379 0.9 1.1 1.5 2.1 
IA Waterloo 42.554 -92.401 0.9 1.2 1.5 1.9 
ID Boise 43.565 -116.220 0.4 0.5 0.7 1.6 
ID Coeur D'Alene 47.774 -116.817 0.5 0.6 0.8 0.9 
ID Pocatello 42.920 -112.571 0.4 0.6 0.8 1.3 
IL Chicago 41.986 -87.914 0.9 1.1 1.5 2.0 
IL Greater Peoria 40.668 -89.684 0.9 1.1 1.5 1.9 
IL Greater Rockford 42.196 -89.093 0.9 1.1 1.6 1.9 
IL Moline 41.465 -90.523 1.0 1.2 1.6 2.0 
IL Springfield 39.845 -89.684 0.9 1.1 1.5 2.1 
IN Evansville 38.043 -87.537 1.0 1.3 1.8 2.2 
IN Fort Wayne 41.006 -85.206 0.8 1.0 1.4 1.6 
IN Indianapolis 39.710 -86.272 0.9 1.1 1.5 2.0 
114 
State Station/City Latitude Longitude 
Percentile Rainfall Depth (in.) 90% Cumulative 
Rainfall Depth (in.)  85
th
 90th 95th 
IN South Bend 41.707 -86.333 0.8 1.0 1.5 2.0 
KS Dodge City 37.773 -99.970 0.8 1.0 1.5 1.9 
KS Goodland 39.368 -101.693 0.7 1.0 1.4 2.0 
KS Topeka 39.073 -95.626 1.0 1.3 1.8 2.4 
KS Wichita 37.647 -97.429 1.0 1.2 1.7 2.1 
KY Fort Campbell 36.673 -87.492 1.0 1.3 1.8 2.1 
KY Lexington 38.041 -84.606 0.9 1.1 1.5 2.0 
KY Louisville 38.177 -85.730 0.9 1.1 1.5 2.0 
LA Baton Rouge 29.350 -89.408 1.3 1.6 2.2 2.9 
LA Lake Charles 30.125 -93.228 1.4 1.7 2.4 3.4 
LA New Orleans 29.993 -90.251 1.5 1.8 2.5 3.3 
LA Shreveport 32.447 -93.824 1.3 1.6 2.1 2.8 
MA Boston 42.361 -71.011 0.9 1.2 1.6 1.9 
MA Worchester 42.267 -71.876 1.0 1.2 1.6 2.0 
MD Baltimore 39.172 -76.684 1.0 1.2 1.5 1.8 
ME Caribou 46.867 -68.033 0.7 0.8 1.1 1.4 
ME Portland 43.642 -70.304 1.0 1.3 1.8 2.7 
MI Alpena 45.072 -83.581 0.7 0.8 1.1 1.3 
MI Detroit 42.215 -83.349 0.8 0.9 1.3 1.7 
MI Flint 42.967 -83.749 0.8 0.9 1.4 2.2 
MI Grand Rapids 42.882 -85.523 0.9 1.1 1.6 2.2 
MI Houghton 44.368 -84.691 0.7 0.9 1.3 1.9 
MI Lansing 42.780 -84.579 0.7 0.9 1.2 1.6 
MI Muskegon 43.171 -86.237 0.8 0.9 1.3 1.7 
MI Sault St. Marie 46.467 -84.367 0.7 0.8 1.1 1.3 
MN Duluth 46.844 -92.194 0.9 1.1 1.5 1.8 
MN International Falls 48.566 -93.403 0.8 0.9 1.3 1.7 
MN Minneapolis 44.883 -93.229 0.8 1.0 1.4 1.8 
MN Rochester 43.904 -92.492 0.9 1.2 1.6 2.0 
MN St. Cloud 45.545 -94.052 0.9 1.1 1.5 2.0 
MO Columbia 38.817 -92.218 1.0 1.3 1.9 2.4 
MO Kansas City 39.299 -94.718 1.0 1.3 1.8 2.3 
MO Springfield 37.240 -93.390 1.1 1.4 1.9 2.4 
MO St. Louis 38.753 -90.374 1.0 1.2 1.7 2.3 
MS Jackson 32.320 -90.078 1.2 1.5 2.0 2.6 
MS Meridian 32.333 -88.751 1.3 1.6 2.2 2.7 
MT Billings 45.808 -108.543 0.6 0.8 1.2 1.9 
MT Cut Bank 48.608 -112.376 0.6 0.8 1.0 1.7 
MT Glasgow 48.214 -106.621 0.6 0.8 1.2 1.6 
MT Great Falls 47.473 -111.382 0.6 0.8 1.2 2.1 
MT Helena 46.606 -111.964 0.5 0.7 1.0 1.8 
MT Kalispell 48.304 -114.264 0.5 0.5 0.8 1.1 
MT Lewistown 47.049 -109.467 0.5 0.7 0.9 1.1 
MT Miles City 46.428 -105.886 0.6 0.8 1.0 1.3 
MT Missoula 46.921 -114.093 0.4 0.6 0.8 1.7 
NC Asheville 35.432 -82.538 1.0 1.2 1.7 2.1 
NC Charlotte 35.214 -80.944 1.0 1.3 1.7 2.0 
NC Greensboro 36.098 -79.944 1.0 1.2 1.7 2.3 
NC Hatteras 35.232 -75.623 1.2 1.5 2.1 3.2 
NC Raleigh 35.871 -78.786 1.0 1.2 1.6 2.1 
NC Wilmington 34.268 -77.906 1.2 1.5 2.1 2.9 
ND Bismarck 46.774 -100.748 0.7 0.9 1.4 2.1 
ND Fargo 46.925 -96.811 0.8 1.0 1.4 1.7 
ND Minot 48.259 -101.281 0.8 0.9 1.2 1.6 
NE Grand Island 40.958 -98.313 0.9 1.1 1.6 2.0 
NE Norfolk 41.981 -97.437 1.0 1.2 1.7 2.2 
NE North Platte 41.122 -100.668 0.9 1.0 1.4 1.8 
115 
State Station/City Latitude Longitude 
Percentile Rainfall Depth (in.) 90% Cumulative 
Rainfall Depth (in.)  85
th
 90th 95th 
NE Omaha 41.310 -95.899 0.9 1.0 1.3 1.7 
NE Scottsbluff 41.874 -103.595 0.7 0.8 1.1 1.5 
NH Concord 43.195 -71.501 0.9 1.1 1.5 1.7 
NJ Atlantic City 39.458 -74.577 1.0 1.2 1.5 1.9 
NJ Newark 40.683 -74.169 1.0 1.2 1.6 2.0 
NM Albuquerque 35.042 -106.616 0.5 0.6 0.8 0.9 
NM Tucumcari 35.182 -103.603 0.8 0.9 1.3 1.6 
NV Elko 40.825 -115.792 0.4 0.5 0.7 0.8 
NV Ely 39.295 -114.845 0.5 0.6 0.9 2.1 
NV Las Vegas 36.079 -115.155 0.7 0.8 1.4 2.2 
NV Reno 39.484 -119.771 0.5 0.7 1.0 1.9 
NV Tonopah 38.060 -117.087 0.4 0.5 0.7 0.9 
NV Winnemucca 40.902 -117.807 0.4 0.6 0.9 1.8 
NY Albany 42.748 -73.803 0.8 1.0 1.3 1.5 
NY Binghamton 42.208 -75.981 0.8 0.9 1.2 1.5 
NY Greater Buffalo 42.941 -78.736 0.7 0.8 1.1 1.4 
NY Massena 44.936 -74.846 0.7 0.9 1.1 1.4 
NY New York City 40.779 -73.881 1.0 1.2 1.7 2.1 
NY Rochester 43.117 -77.677 0.7 0.8 1.1 1.4 
NY Syracuse 43.109 -76.103 0.8 0.9 1.2 1.6 
OH Akron 40.918 -81.443 0.8 0.9 1.2 1.5 
OH Cincinnati 39.103 -84.419 0.9 1.0 1.4 1.7 
OH Cleveland 41.405 -81.853 0.8 0.9 1.2 1.5 
OH Columbus 39.991 -82.881 0.8 1.0 1.3 1.6 
OH Dayton 39.906 -84.219 0.9 1.1 1.5 1.9 
OH Mansfield 40.820 -82.518 0.9 1.0 1.3 1.6 
OH Toledo 41.589 -83.801 0.8 0.9 1.2 1.6 
OH Youngstown 41.254 -80.674 0.8 0.9 1.2 1.5 
OK Oklahoma City 35.389 -97.600 1.1 1.3 1.8 2.3 
OK Tulsa 36.198 -95.886 1.1 1.4 1.9 2.5 
OR Astoria 46.158 -123.878 0.9 1.0 1.4 1.8 
OR Burns 43.592 -118.954 0.4 0.5 0.8 2.1 
OR Eugene 44.133 -123.214 0.9 1.0 1.4 1.8 
OR Medford 42.389 -122.871 0.6 0.7 1.0 1.4 
OR North Bend 43.417 -124.250 0.9 1.1 1.5 1.9 
OR Pendleton 45.698 -118.834 0.4 0.5 0.7 1.1 
OR Portland 45.591 -122.600 0.6 0.7 1.0 1.1 
OR Salem 44.908 -122.995 0.7 0.8 1.2 1.4 
PA Allentown 40.651 -75.449 0.9 1.1 1.5 1.9 
PA Bradford 41.803 -78.640 0.8 0.9 1.2 1.4 
PA Erie 42.080 -80.183 0.8 0.9 1.2 1.5 
PA Greater Pittsburgh 40.501 -80.231 0.7 0.9 1.2 1.5 
PA Harrisburg 40.194 -76.763 1.0 1.2 1.5 1.8 
PA Philadelphia 39.868 -75.231 1.0 1.2 1.6 2.0 
PA Wilkes-Barre 41.339 -75.727 0.8 0.9 1.3 1.6 
PA Williamsport 41.243 -76.922 0.9 1.1 1.5 1.8 
RI Providence 41.722 -71.433 1.1 1.3 1.8 2.3 
SC Charleston 32.899 -80.041 1.1 1.4 1.9 2.5 
SC Columbia 33.942 -81.118 1.1 1.3 1.8 2.3 
SC Greenville 34.899 -82.219 1.1 1.4 1.8 2.3 
SD Huron 44.385 -98.229 0.9 1.0 1.4 1.9 
SD Rapid City 44.046 -103.054 0.7 0.9 1.2 2.0 
SD Sioux Falls 43.577 -96.754 0.9 1.2 1.6 2.0 
TN Chattanooga 35.033 -85.200 1.1 1.3 1.7 2.3 
TN Dyersburg 36.019 -89.318 1.1 1.4 1.9 2.4 
TN Knoxville 35.818 -83.986 0.9 1.1 1.5 1.9 
TN Memphis 35.061 -89.985 1.3 1.6 2.1 2.7 
116 
State Station/City Latitude Longitude 
Percentile Rainfall Depth (in.) 90% Cumulative 
Rainfall Depth (in.)  85
th
 90th 95th 
TN Nashville 36.119 -86.689 1.1 1.3 1.8 2.4 
TX Abilene 32.411 -99.682 1.1 1.4 1.8 2.2 
TX Amarillo 35.219 -101.706 0.8 1.0 1.4 1.7 
TX Austin 30.179 -97.681 1.2 1.5 1.9 3.3 
TX Brownsville 25.906 -97.426 1.1 1.5 2.2 3.6 
TX Corpus Christi 27.773 -97.513 1.3 1.6 2.4 3.2 
TX El Paso 31.811 -106.376 0.7 0.8 1.1 1.3 
TX Fort Worth 32.769 -97.441 1.2 1.4 1.9 2.4 
TX Houston 29.607 -95.159 1.2 1.5 2.0 2.9 
TX Lubbock 33.668 -101.821 0.9 1.1 1.6 2.2 
TX Port Arthur 29.951 -94.021 1.4 1.8 2.5 3.4 
TX San Angelo 31.351 -100.494 1.0 1.3 1.8 2.1 
TX San Antonio 29.533 -98.464 1.2 1.6 2.1 3.2 
TX Victoria 28.863 -96.930 1.3 1.6 2.3 3.7 
TX Waco 31.611 -97.229 1.2 1.5 2.2 2.9 
TX Wichita Falls 33.979 -98.493 1.0 1.3 1.7 2.2 
UT Salt Lake City 40.787 -111.968 0.5 0.6 0.8 1.0 
VA Lynchburg 37.338 -79.207 0.9 1.1 1.6 1.9 
VA Norfolk 36.904 -76.192 1.0 1.3 1.8 2.5 
VA Richmond 37.511 -77.323 1.0 1.3 1.6 2.0 
VA Roanoke 37.317 -79.974 0.9 1.2 1.6 2.1 
VT Burlington 44.468 -73.150 0.8 0.9 1.2 1.5 
WA Olympia 46.973 -122.903 0.8 0.9 1.2 1.5 
WA Quillayute 47.934 -124.561 1.2 1.5 2.0 2.5 
WA Seattle 47.461 -122.314 0.6 0.8 1.0 1.4 
WA Spokane 47.621 -117.528 0.5 0.6 0.8 1.4 
WA Yakima 46.564 -120.534 0.4 0.6 0.8 3.1 
WI Green Bay 44.513 -88.120 0.8 1.0 1.3 1.5 
WI La Crosse 43.754 -91.256 0.9 1.1 1.4 1.8 
WI Madison 43.141 -89.345 0.9 1.1 1.5 2.2 
WI Milwaukee 42.947 -87.897 0.9 1.1 1.5 2.3 
WV Beckley 37.795 -81.125 0.8 0.9 1.2 1.5 
WV Elkins 38.885 -79.853 0.8 0.9 1.2 1.4 
WV Huntington 38.382 -82.555 0.8 1.0 1.4 1.7 
WY Casper 42.898 -106.473 0.6 0.7 1.0 1.4 
WY Cheyenne 41.158 -104.807 0.6 0.8 1.0 1.2 
WY Lander 42.817 -108.733 0.6 0.8 1.0 1.3 
WY Rock Springs 41.594 -109.065 0.4 0.5 0.7 0.9 
WY Sheridan 44.774 -106.976 0.6 0.7 1.0 1.8 
 
117 
Table 4.2 The 95
th
 Percentile Depths (in.) from This Study and Hirschman and Kosco 
(2008) and Their Absolute and Relative Differences (%) in 20 U.S. Cities 
95
th
 Percentile Rainfall Depth (in.) 
Station/City This Study Hirschman & Kosco (2008) Difference
 1
 
Atlanta 1.8 1.8 0.0? (0.0%) 
Baltimore 1.5 1.6 -0.1? (-6.3%) 
Boston 1.6 1.5 0.1? (6.7%) 
Buffalo 1.1 1.1 0.0? (0.0%) 
Burlington 1.2 1.1 0.1? (9.1%) 
Cincinnati 1.4 1.5 -0.1? (-6.7%) 
Coeur D'Alene 0.8 0.7 0.1? (14.3%) 
Columbus 1.3 1.3 0.0? (0.0%) 
Concord 1.5 1.3 0.2? (15.4%) 
Denver 1.0 1.1 -0.1? (-9.1%) 
Kansas City 1.8 1.7 0.1? (5.9%) 
Knoxville 1.5 1.5 0.0? (0.0%) 
Louisville 1.5 1.5 0.0? (0.0%) 
Minneapolis 1.4 1.4 0.0? (0.0%) 
New York City 1.7 1.7 0.0? (0.0%) 
Phoenix 1.1 1.0 0.1? (10.0%) 
Portland 1.0 1.0 0.0? (0.0%) 
Salt Lake City 0.8 0.8 0.0? (0.0%) 
Seattle 1.0 1.6 -0.6? (-37.5%) 
Washington DC 1.5 1.7 -0.2? (-11.8%) 
1
 Difference was given in inches and in percent that was computed as the difference 
between the values determined in this study and reported by Hirschman and Kosco 
(2008) divided by the value reported by Hirschman and Kosco (2008) 
 
118 
Table 4.3 Storm Depth Mean and Kappa Distribution Paramters for 18 Selected Stations 
in Eastern New Mexico, Oklahoma and Texas 
  
Storm 
Depth Kappa Distribution Parameters
 2
 
Station name Mean (in.)
 1
 ? ? ? h 
Abilene 0.54951 -1.0166 1.4485 -0.0647 2.0081 
Amarillo 0.42832 -1.0425 1.3198 -0.1401 2.1559 
Austin 0.62048 -1.2368 1.4770 -0.1012 2.2560 
Brownsville 0.55926 -1.5950 1.3479 -0.2258 2.8952 
Corpus Christi 0.62332 -1.3302 1.2648 -0.2272 2.6009 
El Paso 0.25273 -0.7447 1.1176 -0.1882 1.9348 
Fort Worth 0.75086 -0.9257 1.4774 0.0040 2.0053 
Houston 0.74602 -0.8873 1.1963 -0.1817 2.0231 
Lubbock 0.44280 -0.9854 1.2629 -0.1610 2.1208 
Port Arthur 0.86073 -0.7344 1.0967 -0.2099 1.8786 
San Angelo 0.46788 -1.2914 1.5072 -0.0936 2.3193 
San Antonio 0.59669 -1.1826 1.2911 -0.1884 2.3525 
Victoria 0.66368 -1.1389 1.2522 -0.2007 2.3330 
Waco 0.63370 -1.0224 1.4888 -0.0444 1.9845 
Wichita Falls 0.59963 -0.9523 1.4329 -0.0572 1.9386 
Oklahoma City 0.62856 -0.8200 1.3398 -0.0789 1.8246 
Tulsa 0.66302 -0.9199 1.4384 -0.0472 1.8874 
Tucumcari 0.43087 -0.5830 1.0862 -0.1384 1.8668 
1
 extracted from Asquith et al. (2006) 
2
 computed station by station using Newton- Raphson method using L-moment ratios 
extracted from Asquith et al. (2006). 
 
119 
Table 4.4 The 95
th
 Percentile Storm Depths Determined Using Kappa Distribution 
Parameters Derived from Hourly Rainfall Data and Differences between 95
th
 Percentile 
Rainfall Depths Estimated from Daily and Hourly Rainfall Data 
95
th
 rainfall depth (in) 
Differences of  
95
th
 rainfall depths (in.)State Station 
From daily 
rainfall data 
From hourly 
rainfall data 
TX Abilene 1.8 2.1 -0.3 
TX Amarillo 1.4 1.7 -0.3 
TX Austin 1.9 2.5 -0.6 
TX Brownsville 2.2 2.4 -0.2 
TX Corpus Christi 2.4 2.6 -0.2 
TX El Paso 1.1 1.0 0.1 
TX Fort Worth 1.9 2.6 -0.7 
TX Houston 2.0 2.9 -0.9 
TX Lubbock 1.6 1.7 -0.1 
TX Port Arthur 2.5 3.3 -0.8 
TX San Angelo 1.8 1.9 -0.1 
TX San Antonio 2.1 2.4 -0.3 
TX Victoria 2.3 2.7 -0.4 
TX Waco 2.2 2.4 -0.2 
TX Wichita Falls 1.7 2.3 -0.6 
OK Oklahoma City 1.8 2.4 -0.6 
OK Tulsa 1.9 2.5 -0.6 
NM Tucumcari  1.3 1.5 -0.2 
 
120 
Table 4.5 Comparison between NOAA's 1 Year 24-Hour Rainfall and Computed 95
th
 
Percentile Rainfall Depth Using Daily Data 
Rainfall Depth (in.) Difference 
State Station/City 1 year 24-hr 
1 
 95
th
 percentile  (in.) 
TX Abilene 2.7 1.8 -0.9 
TX Amarillo 2.2 1.4 -0.8 
TX Austin 3.2 1.9 -1.3 
TX Brownsville 3.6 2.2 -1.4 
TX Corpus Christi 3.4 2.4 -1.0 
TX El Paso 1.3 1.1 -0.2 
TX Fort Worth 3.2 1.9 -1.3 
TX Houston 3.7 2.0 -1.7 
TX Lubbock 2.2 1.6 -0.6 
TX Port Arthur 4.1 2.5 -1.6 
TX San Angelo 2.6 1.8 -0.8 
TX San Antonio 3.1 2.1 -1.0 
TX Victoria 3.4 2.3 -1.1 
TX Waco 3.2 2.2 -1.0 
TX Wichita Falls 2.8 1.7 -1.1 
OK Oklahoma City 2.9 1.8 -1.1 
OK Tulsa 3.2 1.9 -1.3 
NM Tucumcari   1.8 1.3   -0.5 
1
 From U.S. Weather Bureau Technical Paper No. 40 (Hershfield 1963). 
 
121 
Table 4.6 Statistical Summary of 90
th
 and 95
th
 Percentile Daily Rainfall Depths (in.) and 
the Rainfall Depths that can Capture the Runoff from 90% of Average Annual Rainfall at 
206 Weather Stations in the Contiguous U.S. 
Statistical distribution 
parameters 
90
th
 
percentile 
95
th
 
percentile 
90% of average 
annual rainfall
 1
 
Ratio of 90% and 
90
th
 rainfall depth 
Minimum 0.49 0.67 0.79 1.29 
Maximum 1.81 2.50 4.72 7.18 
25% Quartile 0.87 1.19 1.56 1.64 
Median 1.09 1.47 1.94 1.72 
75% Quartile 1.30 1.79 2.28 1.90 
Average  1.08 1.48 1.99 1.88 
Standard deviation 0.31 0.42 0.62 0.59 
1
 To avoid the few very large or extreme events from dominating the annual averages, 
rainfall depths (events) greater than 99.5 percentiles were excluded from the analysis 
(Guo and Urbonas, 1996) 
122 
0
1
2
3
4
5
R
a
i
n
f
a
l
l
 
(
i
n
.
)
0 10 20 30 40 50 60 70 80 90 100
Percentile (%)
95
1.4"
2.1"
Montgomery
Minneapolis
 
Fig. 4.1. Percentile distribution of daily rainfall depths for Montgomery and Minneapolis 
weather stations 
123 
1.0
1.1
1.2
1.3
1.4
1.5
1.6
R
a
i
n
f
a
l
l
 
(
i
n
.
)
1.7
1.8
1.9
2.0
2.1
2.2
2.3
R
a
i
n
f
a
l
l
 
(
i
n
.
)
1960 1965 1970 1975 1980 1985 1990 1995 2000
Beginning Year
10 year data
20 year data
30 year data
Montgomery, AL
Minneapolis, MN
 
Fig. 4.2. 95
th
 percentile rainfall depths estimated using 10, 20, and 30 years of daily 
rainfall data at Montgomery, AL (top) and Minneapolis, MN (bottom) for sensitivity 
analysis 
 
 
 
Fig. 4.3. 95
th
 percentile rainfall map for the contiguous U.S. 
124 
 
 
Fig. 4.4. 90
th
 percentile rainfall map for the contiguous U.S. 
125 
 
 
Fig. 4.5. 85
th
 percentile rainfall map for the contiguous U.S. 
126 
127 
 
0
1
2
3
4
5
6
7
8
9
10
R
a
i
n
f
a
l
l
 
(
i
n
.
)
0 10 20 30 40 50 60 70 80 90 100
Percentile (%)
Sorted from daily rainfall
Kappa distribution
95th
85th
90th
99th
 
Fig. 4.6. Distribution of daily rainfall depths and quantiles calculated using Kappa 
distribution with paramters derived from daily data at Abilene Regional Airport, Texas 
 
128 
0
1
2
3
4
5
9
5
t
h
 
D
a
i
l
y
 
R
a
i
n
f
a
l
l
 
(
i
n
.
)
0
1
2
3
4
5
9
5
t
h
 
D
a
i
l
y
 
R
a
i
n
f
a
l
l
 
(
i
n
.
)
0 1 2 3 4 5
95th 24-hr Rainfall (in.)
0 1 2 3 4 5
1 yr 24-hr Rainfall (in.)
y = 0.4989x + 0.4253
R?  = 0.86
y = 0.5996x + 0.5207
R?  = 0.80
 
Fig. 4.7. Regression equations between 95
th
  percentile rainfall depth derived from daily 
rainfall data and 95th percentile rainfall depth derived from hourly rainfall data  (left) and 
NOAA's 1 year 24-hr rainfall (right) for selected 18 stations (Table 4.4 and Table 4.5) 
129 
0
1
2
3
4
5
R
a
i
n
f
a
l
l
 
(
i
n
.
)
0 10 20 30 40 50 60 70 80 90 100
Percentile (%)
1.8"
2.6"
Montgomery
Minneapolis
 
Fig. 4.8 Percentile distribution of cumulative rainfall depths for Montgomery and 
Minneapolis weather stations 
130 
Chapter 5.  Conclusions and Recommendation 
 
General Summary and Conclusions 
In this study, Watershed Analysis Risk Management Framework (WARMF) 
model was developed and applied to the Saugahatchee Creek Watershed (SCW) to 
simulate flow and water quality, which is based on calibration and validation results, 
supported by quantitative statistics for model evaluation and/or graphical techniques. 
Flow calibration for the period from 2000 ? 2005 and validation for the period from 2006 
? 2009 was performed using observed daily flow data at USGS 02418230 station in the 
Saugahatchee Creek near Loachapoka. The quantitative statistics recommended by 
Moriasi et al (2007), NSE, RSR, and PBIAS were 0.64, 0.60, and -2.78%, respectively 
for calibration and 0.56, 0.66, and -9.53%, respectively for validation. These values 
indicate satisfactory performance of the WARMF model for flow simulation. The 
simulation of water quality parameters such as water temperature, dissolved oxygen 
(DO), total phosphorus (TP), total nitrogen (TN), and chlorophyll-a concentration were 
calibrated to be satisfactorily predicted at various water quality stations in the watershed 
by the WARMF model based on visual inspection. 
 WARMF model simulation was run using daily time step. Certain water quality 
parameter such as DO concentration that vary, especially in an eutrophic system, during a 
day cannot be featured with daily time step. Shorter time step must be used to simulate 
diurnal variations.  
131 
A portion of the Saugahatchee Creek (Yates Reservoir Embayment) that enters 
Yates reservoir, when treated as a stream instead of a reservoir in the WARMF model, 
accounted for significant decrease in chlorophyll-a concentration. Reservoir tends to have 
calm water, which aids in settling of organic matter and nutrient accumulation. Also, 
stratification was predicted in reservoir during summer that will cut off oxygen supply to 
the bottom layer depleting DO concentration. River impounding can possibly be the 
cause, in addition to excess nutrient, of impairment of Saugahatchee Creek (Yates 
Reservoir Embayment) for nutrients and organic enrichment/ dissolved oxygen. 
 Land use scenarios of 1991, 2001, and 2008 for the SCW show the land use 
transformation from forest areas to cropland and urban areas. The major impact of land 
use change was predicted to be rise in nutrient levels. The model predicted as much as 
15.4% and 27.8% increase in concentrations of TP and TN concentration, respectively in 
terms of monthly average of daily values for land use change scenarios. Since two 
portions of the Saugahatchee Creek is already in the list of impaired water for nutrients 
(ADEM 2009), it is essential to exercise effective best management practices (BMPs) for 
nutrient runoff. 
 The outputs derived from Hadley Centre Coupled Model, version 3 (HadCM3) 
future climate model were downscaled to obtain meteorological variables at local 
watershed scale for allotted time frames of 2020s (2011-2040), 2050s (2041-2070), and 
2080s (2071-2099). Two emission scenarios (A2 and B2) for HadCM3, used in this 
study, report that the climate in the SCW for the 21
st
 century will likely be warmer and 
drier. Based on the WARMF model simulation for baseline scenario (1981-2010) and 
future climate scenarios, the general predicted trend was increase in water temperature, 
132 
and nutrient concentration; decrease in flow, and DO concentration; and mixed result for 
chlorophyll-a concentration. The relative decrease in flow, in terms of monthly average 
of daily values for, ranged from 4.1% to 62.2% for HadCM3 A2 and 15.1% to 56.8% for 
HadCM3 B2 scenarios. Water temperature increased as high as 5.5?C and 4.7?C for 
HadCM3 A2  and B2, respectively. DO concentration decreased by 22.1% and 16.5% at 
most for HadCM3 A2 and B2 respectively. The TP concentration climbed as much as 
87.4% and 83.16%; TN concentration climbed as much as 80.0% and 57.5% for 
HadCM3 A2 and B2 respectively.  
 Uncertainties involved with future climate scenarios, downscaling, and watershed 
model outputs should not be discarded while using these results. However, the impact 
study under different scenarios of land use and climate change, in general, is useful for 
decision making in watershed management and sustainable development. 
 
Summary of Chapter 4 
The 95
th
 percentile rainfall, recommended value for design storm by U.S. 
Environmental Protection Agency (USEPA), was determined and reported at 206 
stations/cities in the contiguous U.S. using daily rainfall data from 1973 to 2010.  The 
95
th
 percentiles ranged from 17.8 mm (0.7 in.) to 63.5 mm (2.5 in.).  Comparing with the 
95
th
 percentile rainfall depths derived from hourly data at 18 selected stations in eastern 
New Mexico, Oklahoma and Texas, the 95
th
 percentile rainfall depths derived from daily 
rainfall record was typically underestimated.  The 95
th
 percentile daily rainfall depths 
showed good linear correlation with NOAA?s 1-yr 24-hour rainfall depth but the later 
was greater.  Water quality volume determined using 90% cumulative percent of average 
133 
annual rainfall is on average 88% larger than one determined using 90
th
 percentiles of 
daily rainfall. 
 The rainfall isohyetal map for 85
th
, 90
th
, and 95
th
 percentile rainfall depth, 
accompanied with values in tabular form, was generated for 206 stations in the 
contiguous U.S. Rainfall percentile statistics from this study will be valuable for 
estimating design storms, e.g., for water quality improvement, low impact development, 
and green infrastructures. 
 
Recommendations for Future Exploration 
Simulation at finer temporal scale 
For this study, daily time step was used in the WARMF model to simulate flow 
and other water quality parameters. We learned that some of the features that experience 
diurnal variations are not observable using daily time steps. For example, DO 
concentration tends to fluctuate largely in an eutrophic system. With the availability of 
high temporal resolution data, the simulations of flow and water quality are 
recommended to be simulated using finer (hourly) time step.  
 
Future land use scenarios 
The impact of historical land use scenarios were applied in the WARMF model 
for this research. This research area can be broadened after the completion of future land 
use scenarios for the SCW (Rajesh Sawant, personal communication, Jun. 8, 2011). It 
would be a good suggestion to couple it with the future climate change scenario.  
 
134 
Management scenarios 
Alabama Department of Environmental Management (ADEM 2008)?s Total 
Phosphorus (TP) Total Maximum Daily Load (TMDL) study calls for TP reduction of 
more than 86% from point sources, and 50% from non point sources. Furthermore, it was 
determined that DO concentrations would fall below the standard (5mg/l) even if all the 
point sources are removed during simulation(ADEM 2008). Therefore, reduction in both 
point and non point source pollution is recommended for the SCW. For cost effective 
implementation of best management practices, identification of critical source areas was 
carried out in the SCW (Niraula 2010). Consensus module in WARMF guides the users 
through different management scenarios to compare alternatives. It is recommended to 
further explore the application of WARMF?s consensus module to identify whether water 
bodies meet water quality criteria for particular scenarios, how point and nonpoint source 
affect water quality to evaluate management alternatives. Another unique feature of the 
WARMF, TMDL module is recommended, which is useful for TMDL calculation. 
135 
References 
Alabama Department of Environmental Management (ADEM). (2008). "Final total 
maximum daily load: Nutrients & OE/DO. Pepperell branch AL03150110-0201-
700 Nutrients; Sougahatchee Creek Embayment (Yates Reservoir) AL03150110-
0204-101 Nutrients & OE/DO.", Alabama Department of Environmental 
Management, Montgomery, AL. 
 
ADEM. (2009). ?2008 Alabama 303(d) List.? 303(d) Information and Map. Alabama 
Department of Environmental Management 
<http://adem.alabama.gov/programs/water/wquality/2008AL303dList.pdf> (Aug. 
12, 2009). 
 
Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., and 
Veith, T. L. (2007). "Model Evaluation Guidelines for Systematic Quantification 
of Accuracy in Watershed Simulations." Transactions of the ASABE, 50(3), 885-
900. 
 
Niraula, R. (2010). ?Identifying critical source areas of sediment, nitrogen, and 
phosphorus: A modeling approach.? M.S. thesis, Auburn University, Auburn, AL. 
 
 
 
 
 
 
 
136 
Appendix A.  Downscaling Precipitation Using SDSM 4.2 
 
 
Fig. A.1 Correlation between observed precipitation and NCEP predictors for each month 
137 
 
Fig. A.2 Correlation matrix and partial correlations between observed precipitation and 
NCEP predictors 
138 
 
Fig. A.3 Calibration result for precipitation with selected predictors (p5zh, r500, and 
r850)  
139 
 
Fig. A.4 Parameter file generated by SDSM for downscaling precipitation 
140 
 
Fig. A.5 Mean monthly precipitation for observed and downscaled results for validation 
141 
Appendix B.  Land Use Change Impact: Statistical Summary 
 
Table B.1 Monthly Average of Daily Flow and Standard Dev. for Land Use Scenarios 
    Average flow (m
3
/s)  Standard Deviation (m
3
/s) 
    Baseline 2001 2008 1991 2001 2008 
Jan 10.64 10.63 10.60 15.59 15.47 15.25 
Feb 12.63 12.62 12.57 18.00 17.77 17.44 
Mar 14.26 14.23 14.17 30.74 30.57 30.27 
Apr 10.15 10.20 10.23 18.76 18.72 18.61 
May 6.51 6.58 6.65 5.06 5.06 5.00 
Jun 6.68 6.74 6.81 11.21 11.19 11.11 
Jul 5.75 5.82 5.89 6.70 6.69 6.67 
Aug 4.69 4.74 4.81 3.50 3.60 3.72 
Sep 4.08 4.16 4.24 1.88 2.02 2.26 
Oct 4.36 4.43 4.52 3.43 3.56 3.72 
Nov 6.29 6.35 6.42 8.64 8.53 8.41 
Dec   9.27 9.27 9.26  13.89 13.69 13.39 
 
Table B.2 Relative Change in Monthly Average of Daily Flow from the Baseline 
    Relative Change (m
3
/s)  Relative Change (%) 
    2001 2008 2001 2008 
Jan 0.00 -0.03 -0.04 -0.31 
Feb -0.01 -0.06 -0.11 -0.45 
Mar -0.03 -0.09 -0.23 -0.64 
Apr 0.04 0.08 0.41 0.76 
May 0.06 0.13 0.96 2.05 
Jun 0.06 0.13 0.95 2.01 
Jul 0.07 0.15 1.17 2.54 
Aug 0.06 0.12 1.23 2.64 
Sep 0.07 0.16 1.76 3.89 
Oct 0.07 0.16 1.66 3.57 
Nov 0.06 0.14 0.99 2.21 
Dec  0.01 -0.01  0.06 -0.07 
142 
Table B.3 Monthly Average of Daily Water Temperature and Standard Deviation for 
Land Use Scenarios 
    Average temperature (?C)  Standard Deviation (?C) 
    Baseline 2001 2008 1991 2001 2008 
Jan 8.76 8.76 8.77 3.91 3.91 3.92 
Feb 11.11 11.11 11.12 3.53 3.53 3.53 
Mar 15.06 15.07 15.07 3.55 3.55 3.55 
Apr 19.37 19.38 19.38 3.01 3.01 3.01 
May 24.42 24.42 24.42 2.41 2.41 2.41 
Jun 28.48 28.48 28.47 1.98 1.98 1.99 
Jul 30.02 30.01 30.00 1.42 1.43 1.44 
Aug 29.93 29.93 29.92 1.62 1.62 1.63 
Sep 26.73 26.72 26.72 2.46 2.47 2.47 
Oct 20.42 20.42 20.42 3.24 3.24 3.24 
Nov 14.70 14.71 14.73 3.23 3.23 3.23 
Dec   10.12 10.12 10.13  3.92 3.92 3.91 
 
Table B.4 Relative Change in Monthly Average of Daily Water Temperature from the 
Baseline 
    Relative Change (?C)  Relative Change (%) 
    2001 2008 2001 2008 
Jan 0.01 0.02 0.08 0.18 
Feb 0.01 0.02 0.07 0.14 
Mar 0.01 0.01 0.06 0.08 
Apr 0.01 0.01 0.03 0.05 
May 0.00 0.00 0.01 0.01 
Jun 0.00 0.00 0.00 -0.01 
Jul -0.01 -0.02 -0.02 -0.05 
Aug -0.01 -0.01 -0.02 -0.04 
Sep -0.01 -0.01 -0.02 -0.04 
Oct 0.00 0.00 0.00 0.01 
Nov 0.01 0.02 0.06 0.15 
Dec  0.01 0.02  0.06 0.18 
 
 
 
143 
Table B.5 Monthly Average of Daily Surface DO and Standard Deviation for Land Use 
Scenarios 
    Average DO (mg/l)  Standard Deviation (mg/l) 
    Baseline 2001 2008 1991 2001 2008 
Jan 10.39 10.39 10.37 0.92 0.92 0.92 
Feb 10.05 10.04 10.02 0.87 0.86 0.85 
Mar 9.23 9.22 9.21 0.78 0.78 0.78 
Apr 8.45 8.47 8.49 0.70 0.71 0.72 
May 8.07 8.12 8.18 1.12 1.15 1.19 
Jun 8.16 8.25 8.29 1.21 1.21 1.22 
Jul 7.46 7.47 7.49 0.96 0.97 0.97 
Aug 7.54 7.56 7.60 1.19 1.20 1.20 
Sep 8.53 8.54 8.57 0.99 0.98 0.98 
Oct 8.06 8.06 8.07 0.52 0.51 0.51 
Nov 8.76 8.75 8.74 0.71 0.70 0.70 
Dec   9.92 9.91 9.90  0.97 0.97 0.95 
 
Table B.6 Relative Change in Monthly Average of Daily Surface DO from the Baseline 
    Relative Change (mg/l)  Relative Change (%) 
    2001 2008 2001 2008 
Jan -0.01 -0.02 -0.07 -0.22 
Feb -0.01 -0.03 -0.10 -0.28 
Mar 0.00 -0.01 -0.04 -0.15 
Apr 0.02 0.03 0.25 0.41 
May 0.05 0.11 0.64 1.35 
Jun 0.09 0.14 1.09 1.66 
Jul 0.01 0.03 0.15 0.44 
Aug 0.01 0.06 0.19 0.74 
Sep 0.01 0.04 0.10 0.50 
Oct 0.00 0.01 0.03 0.18 
Nov -0.01 -0.02 -0.09 -0.25 
Dec   -0.01 -0.02  -0.06 -0.22 
 
 
 
144 
Table B.7 Monthly Average of Daily TP and Standard Deviation for Land Use Scenarios 
    Average TP (mg/l)  Standard Deviation (mg/l) 
    Baseline 2001 2008 1991 2001 2008 
Jan 0.070 0.074 0.080 0.036 0.039 0.043 
Feb 0.066 0.070 0.075 0.031 0.033 0.035 
Mar 0.067 0.070 0.076 0.026 0.028 0.031 
Apr 0.066 0.070 0.076 0.021 0.022 0.025 
May 0.074 0.078 0.085 0.018 0.020 0.023 
Jun 0.083 0.087 0.093 0.020 0.022 0.025 
Jul 0.087 0.091 0.098 0.026 0.027 0.031 
Aug 0.093 0.097 0.105 0.030 0.032 0.036 
Sep 0.100 0.104 0.112 0.029 0.030 0.034 
Oct 0.102 0.106 0.114 0.033 0.035 0.038 
Nov 0.095 0.100 0.108 0.041 0.043 0.048 
Dec   0.080 0.084 0.090  0.041 0.043 0.048 
 
Table B.8 Relative Change in Monthly Average of Daily TP from the Baseline 
    Relative Change (mg/l)  Relative Change (%) 
    2001 2008 2001 2008 
Jan 0.004 0.010 5.87 14.41 
Feb 0.003 0.009 5.16 12.98 
Mar 0.004 0.009 5.25 13.87 
Apr 0.004 0.010 5.58 15.37 
May 0.004 0.010 5.17 14.15 
Jun 0.004 0.011 4.64 12.73 
Jul 0.004 0.011 4.80 13.07 
Aug 0.004 0.012 4.56 12.62 
Sep 0.004 0.012 4.23 12.00 
Oct 0.004 0.011 3.87 11.18 
Nov 0.005 0.012 5.11 13.00 
Dec   0.004 0.010  5.06 12.96 
 
 
145 
Table B.9 Monthly Average of Daily TN and Standard Deviation for Land Use Scenarios 
    Average TN (mg/l)  Standard Deviation (mg/l) 
    Baseline 2001 2008 1991 2001 2008 
Jan 2.28 2.33 2.40 0.74 0.75 0.78 
Feb 2.13 2.18 2.26 0.72 0.74 0.77 
Mar 2.06 2.11 2.19 0.65 0.66 0.69 
Apr 2.24 2.29 2.38 0.66 0.67 0.72 
May 2.43 2.50 2.59 0.65 0.66 0.69 
Jun 2.57 2.63 2.72 0.70 0.71 0.76 
Jul 2.63 2.69 2.77 0.74 0.75 0.79 
Aug 2.81 2.88 2.97 0.76 0.78 0.82 
Sep 2.92 2.98 3.05 0.70 0.71 0.75 
Oct 2.94 2.99 3.06 0.70 0.71 0.75 
Nov 2.84 2.89 2.95 0.77 0.79 0.84 
Dec   2.51 2.56 2.62  0.83 0.83 0.85 
 
Table B.10 Relative Change in Monthly Average of Daily TN from the Baseline 
    Relative Change (mg/l)  Relative Change (%) 
    2001 2008 2001 2008 
Jan 0.05 0.13 2.33 5.59 
Feb 0.05 0.13 2.30 5.94 
Mar 0.05 0.13 2.52 6.47 
Apr 0.06 0.15 2.58 6.53 
May 0.06 0.15 2.49 6.33 
Jun 0.06 0.15 2.38 5.97 
Jul 0.06 0.14 2.16 5.28 
Aug 0.07 0.16 2.33 5.53 
Sep 0.06 0.14 2.02 4.66 
Oct 0.05 0.12 1.73 3.98 
Nov 0.05 0.11 1.82 4.00 
Dec   0.05 0.11  1.92 4.36 
 
 
 
 
146 
Table B.11 Monthly Average of Daily Chlorophyll- a Concentration and Standard 
Deviation for Land Use Scenarios 
    Average Chl-a (?g/l)  Standard Deviation (?g/l) 
    Baseline 2001 2008 1991 2001 2008 
Jan 0.26 0.25 0.24 1.82 1.70 1.52 
Feb 0.56 0.52 0.47 3.22 2.97 2.57 
Mar 0.05 0.05 0.05 0.03 0.03 0.03 
Apr 1.60 1.84 2.04 4.72 5.22 5.62 
May 9.45 10.15 10.92 11.90 12.29 12.87 
Jun 24.56 26.10 26.95 20.19 20.07 20.22 
Jul 20.43 20.70 21.03 16.43 16.62 16.60 
Aug 22.31 22.64 23.22 16.70 16.88 16.75 
Sep 33.21 33.50 34.07 16.49 16.45 16.36 
Oct 10.56 10.69 10.94 11.45 11.50 11.71 
Nov 1.04 1.01 1.00 3.66 3.61 3.63 
Dec   0.05 0.05 0.06  0.10 0.11 0.14 
 
Table B.12 Relative Change in Monthly Average of Daily Chlorophyll-a Concentration 
from the Baseline 
    Relative Change (?g/l)   Relative Change (%) 
    2001 2008 2001 2008 
Jan -0.01 -0.03 -3.71 -10.43 
Feb -0.03 -0.09 -5.79 -16.05 
Mar 0.00 0.00 -0.78 0.18 
Apr 0.24 0.44 15.04 27.75 
May 0.70 1.47 7.46 15.58 
Jun 1.54 2.39 6.27 9.74 
Jul 0.26 0.60 1.29 2.92 
Aug 0.33 0.92 1.47 4.11 
Sep 0.28 0.85 0.85 2.57 
Oct 0.13 0.38 1.26 3.59 
Nov -0.02 -0.03 -2.16 -3.28 
Dec   0.00 0.00   2.46 8.71 
 
147 
Appendix C.  Climate Change Impact: Statistical Summary 
 
Table C.1 Monthly Average of Daily Flow (m
3
/s) 
      HadCM3 A2  HadCM3 B2 
 Baseline 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 10.79 7.33 7.03 4.66 7.38 5.83 5.93
Feb 12.63 9.08 7.80 5.86 8.59 6.16 6.91
Mar 14.16 10.93 9.73 6.87 8.90 7.76 9.05
Apr 10.18 9.77 7.24 5.97 7.08 6.84 8.27
May 6.56 5.76 5.10 4.55 5.12 4.78 5.57
Jun 6.92 4.52 4.08 3.41 4.01 3.67 4.23
Jul 5.62 3.66 3.29 2.84 3.46 3.17 3.49
Aug 4.72 3.05 2.83 2.62 3.01 2.75 2.91
Sep 4.14 2.76 2.60 2.21 3.16 2.52 2.62
Oct 4.41 2.67 2.55 2.39 2.76 2.31 2.43
Nov 6.38 3.44 3.30 2.86 3.14 2.73 2.68
Dec 9.42   6.11 6.10 3.56  4.99 4.07 4.38
 
Table C.2 Standard Deviation of Daily Flow (m
3
/s) 
      HadCM3 A2   HadCM3 B2 
 Baseline 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 16.26 7.00 9.15 4.36 8.14 6.31 5.96
Feb 17.40 11.31 9.46 5.65 12.30 4.01 8.09
Mar 30.71 16.95 13.68 16.24 14.49 10.66 15.98
Apr 18.80 14.52 5.63 4.10 6.78 7.83 10.05
May 5.07 1.82 1.72 2.07 1.73 3.70 3.97
Jun 12.30 2.09 1.41 0.92 1.34 1.08 1.25
Jul 4.50 1.06 0.91 0.76 1.10 0.98 1.31
Aug 3.54 0.74 0.75 0.91 0.93 0.86 0.85
Sep 1.97 0.92 1.20 0.76 2.39 1.09 0.76
Oct 3.51 0.97 1.24 1.37 1.32 0.79 0.87
Nov 8.68 2.13 3.17 2.02 1.92 1.74 1.23
Dec 14.03   5.10 8.50 2.47  4.89 3.35 3.58
148 
Table C.3 Relative Change in Monthly Average of Daily Flow from the Baseline (m
3
/s) 
    HadCM3 A2   HadCM3 B2 
  2020s 2050s 2080s   2020s 2050s 2080s 
Jan -3.46 -3.76 -6.13 -3.41 -4.96 -4.87
Feb -3.55 -4.83 -6.77 -4.04 -6.48 -5.73
Mar -3.22 -4.42 -7.28 -5.26 -6.40 -5.11
Apr -0.41 -2.95 -4.22 -3.10 -3.35 -1.91
May -0.79 -1.46 -2.00 -1.44 -1.77 -0.99
Jun -2.40 -2.85 -3.51 -2.91 -3.25 -2.69
Jul -1.96 -2.32 -2.78 -2.16 -2.45 -2.13
Aug -1.67 -1.89 -2.10 -1.72 -1.98 -1.81
Sep -1.38 -1.54 -1.93 -0.98 -1.62 -1.52
Oct -1.74 -1.86 -2.02 -1.65 -2.10 -1.98
Nov -2.94 -3.09 -3.52 -3.24 -3.65 -3.70
Dec   -3.31 -3.33 -5.86  -4.43 -5.35 -5.04
 
Table C.4 Relative Change in Monthly Average of Daily Flow from the Baseline (%) 
    HadCM3 A2   HadCM3 B2 
 2020s 2050s 2080s 2020s 2050s 2080s 
Jan -32.09 -34.87 -56.78 -31.62 -45.94 -45.08
Feb -28.10 -38.24 -53.62 -32.00 -51.25 -45.33
Mar -22.77 -31.26 -51.44 -37.15 -45.20 -36.07
Apr -4.06 -28.94 -41.41 -30.43 -32.88 -18.78
May -12.11 -22.27 -30.57 -21.93 -27.06 -15.10
Jun -34.73 -41.13 -50.78 -42.04 -47.00 -38.92
Jul -34.94 -41.37 -49.41 -38.39 -43.61 -37.89
Aug -35.34 -40.00 -44.43 -36.33 -41.85 -38.33
Sep -33.37 -37.13 -46.53 -23.63 -39.09 -36.66
Oct -39.55 -42.20 -45.82 -37.49 -47.62 -44.89
Nov -46.12 -48.35 -55.22 -50.79 -57.15 -58.05
Dec   -35.13 -35.30 -62.23  -47.02 -56.76 -53.48
 
 
 
 
149 
Table C.5 Monthly Average of Daily Water Temperature (?C) 
      HadCM3 A2  HadCM3 B2 
 Baseline 2020s 2050s 2080s  2020s 2050s 2080s 
Jan 8.77 8.30 9.52 10.16 9.10 9.31 9.01
Feb 11.16 10.20 10.87 12.28 10.04 10.87 10.89
Mar 15.17 13.92 14.52 15.67 13.97 14.80 14.92
Apr 19.49 20.00 19.92 21.93 19.61 20.18 20.46
May 24.51 25.46 25.90 27.72 25.10 26.29 26.75
Jun 28.51 28.85 29.96 31.91 29.12 29.24 30.22
Jul 30.04 30.53 31.53 32.94 30.75 31.45 31.98
Aug 29.91 30.80 31.87 33.03 30.77 31.55 32.55
Sep 26.62 28.15 30.23 31.41 28.40 29.24 30.53
Oct 20.33 22.22 23.53 25.82 22.12 23.67 25.03
Nov 14.63 16.15 17.05 18.82 15.81 16.54 17.99
Dec 10.07   10.48 11.48 12.34  10.40 11.34 11.75
 
Table C.6 Standard Deviation of Daily Water Temperature (?C) 
      HadCM3 A2   HadCM3 B2 
 Baseline 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 3.91 2.99 3.33 3.20 3.52 3.49 3.05
Feb 3.53 3.31 3.07 3.50 3.51 3.53 3.72
Mar 3.54 3.52 3.56 3.72 3.30 3.69 3.68
Apr 3.01 3.41 3.60 3.55 3.23 3.15 3.63
May 2.42 2.63 2.75 2.83 2.71 2.50 2.94
Jun 1.98 2.13 2.23 2.00 1.94 2.09 1.90
Jul 1.42 1.30 1.31 1.18 1.23 1.08 1.20
Aug 1.62 1.37 1.29 1.18 1.35 1.28 1.27
Sep 2.48 2.55 2.26 2.46 2.53 2.38 2.15
Oct 3.23 3.45 3.33 3.22 3.25 2.99 3.46
Nov 3.23 3.48 3.51 3.93 3.43 3.78 3.66
Dec 3.90   3.32 3.38 3.62  3.35 3.50 3.25
 
 
 
 
150 
Table C.7 Relative Change in Monthly Average of Daily Water Temperature (?C) 
    HadCM3 A2   HadCM3 B2 
  2020s 2050s 2080s   2020s 2050s 2080s 
Jan -0.47 0.76 1.39 0.33 0.54 0.24
Feb -0.96 -0.29 1.12 -1.12 -0.30 -0.27
Mar -1.26 -0.65 0.49 -1.20 -0.38 -0.26
Apr 0.51 0.43 2.45 0.13 0.69 0.98
May 0.95 1.39 3.21 0.59 1.78 2.24
Jun 0.33 1.45 3.40 0.61 0.72 1.71
Jul 0.49 1.49 2.90 0.71 1.41 1.95
Aug 0.89 1.97 3.12 0.86 1.64 2.64
Sep 1.53 3.61 4.79 1.78 2.61 3.91
Oct 1.89 3.20 5.49 1.79 3.35 4.70
Nov 1.52 2.42 4.18 1.18 1.91 3.36
Dec   0.40 1.41 2.26  0.32 1.26 1.68
 
Table C.8 Relative Change in Monthly Average of Daily Water Temperature (%) 
    HadCM3 A2   HadCM3 B2 
  2020s 2050s 2080s 2020s 2050s 2080s 
Jan -5.38 8.63 15.84 3.74 6.16 2.77
Feb -8.62 -2.57 9.99 -10.01 -2.66 -2.41
Mar -8.29 -4.31 3.26 -7.92 -2.47 -1.69
Apr 2.63 2.23 12.56 0.64 3.55 5.01
May 3.89 5.67 13.10 2.41 7.25 9.12
Jun 1.17 5.07 11.93 2.14 2.54 6.00
Jul 1.64 4.96 9.64 2.37 4.68 6.48
Aug 2.97 6.57 10.43 2.87 5.48 8.84
Sep 5.74 13.54 18.00 6.69 9.82 14.70
Oct 9.29 15.73 27.00 8.80 16.45 23.12
Nov 10.39 16.54 28.60 8.08 13.02 22.96
Dec   3.99 13.97 22.47  3.21 12.51 16.63
 
 
 
 
151 
Table C.9 Monthly Average of Daily Surface DO (mg/l) 
      HadCM3 A2  HadCM3 B2 
 Baseline 2020s 2050s 2080s  2020s 2050s 2080s 
Jan 10.38 10.45 10.16 9.75 10.31 10.11 10.17
Feb 10.02 10.22 10.01 9.50 10.31 10.01 9.97
Mar 9.20 9.49 9.28 8.92 9.41 9.16 9.19
Apr 8.43 8.28 8.22 8.04 8.40 8.30 8.23
May 8.02 7.73 7.87 7.77 7.96 7.97 7.52
Jun 8.22 8.36 7.84 6.81 8.51 8.54 7.68
Jul 7.40 7.75 7.01 5.98 7.59 7.15 6.62
Aug 7.45 7.54 6.77 5.85 7.61 7.06 6.22
Sep 8.46 8.48 7.47 6.59 8.18 7.99 7.22
Oct 8.04 8.09 8.15 8.01 8.09 8.18 8.09
Nov 8.77 8.36 8.14 7.90 8.33 8.20 7.94
Dec 9.93   9.74 9.41 9.06  9.67 9.35 9.28
 
Table C.10 Standard Deviation of Daily Surface DO (mg/l) 
      HadCM3 A2   HadCM3 B2 
 Baseline 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 0.91 0.77 0.81 0.72 0.95 0.80 0.69
Feb 0.86 0.83 0.90 0.83 1.16 1.12 0.98
Mar 0.77 0.77 0.72 0.84 0.81 0.78 0.76
Apr 0.70 0.65 0.64 0.91 0.77 0.92 0.79
May 1.10 0.97 1.12 1.30 1.25 1.31 1.00
Jun 1.22 1.13 1.31 1.23 1.21 1.19 1.18
Jul 0.97 0.95 0.94 0.67 0.92 0.81 0.79
Aug 1.19 1.01 0.86 0.52 1.06 0.96 0.71
Sep 1.03 0.96 1.15 1.35 1.03 1.24 1.31
Oct 0.52 0.50 0.54 0.80 0.49 0.54 0.69
Nov 0.71 0.64 0.57 0.54 0.59 0.61 0.53
Dec 0.96   0.77 0.66 0.75  0.85 0.72 0.73
 
 
 
 
152 
Table C.11 Relative Change in Monthly Average of Daily Surface DO (mg/l) 
    HadCM3 A2   HadCM3 B2 
  2020s 2050s 2080s   2020s 2050s 2080s 
Jan 0.07 -0.22 -0.62 -0.07 -0.27 -0.21
Feb 0.20 -0.01 -0.52 0.29 -0.01 -0.05
Mar 0.28 0.08 -0.28 0.21 -0.04 -0.01
Apr -0.15 -0.21 -0.39 -0.03 -0.13 -0.20
May -0.29 -0.15 -0.25 -0.06 -0.05 -0.51
Jun 0.14 -0.38 -1.41 0.29 0.32 -0.54
Jul 0.35 -0.39 -1.42 0.18 -0.26 -0.79
Aug 0.09 -0.68 -1.60 0.17 -0.39 -1.23
Sep 0.02 -1.00 -1.87 -0.29 -0.47 -1.24
Oct 0.05 0.11 -0.03 0.05 0.14 0.05
Nov -0.41 -0.63 -0.88 -0.44 -0.57 -0.83
Dec   -0.18 -0.52 -0.87  -0.26 -0.58 -0.64
 
Table C.12 Relative Change in Monthly Average of Daily Surface DO (%) 
    HadCM3 A2   HadCM3 B2 
  2020s 2050s 2080s 2020s 2050s 2080s 
Jan 0.66 -2.10 -6.02 -0.63 -2.59 -2.04
Feb 1.95 -0.10 -5.16 2.86 -0.15 -0.53
Mar 3.07 0.85 -3.05 2.25 -0.43 -0.16
Apr -1.75 -2.46 -4.62 -0.35 -1.57 -2.35
May -3.60 -1.89 -3.16 -0.79 -0.61 -6.33
Jun 1.74 -4.57 -17.19 3.57 3.85 -6.61
Jul 4.70 -5.27 -19.19 2.49 -3.51 -10.65
Aug 1.26 -9.11 -21.51 2.23 -5.17 -16.49
Sep 0.20 -11.76 -22.12 -3.37 -5.59 -14.71
Oct 0.63 1.34 -0.32 0.58 1.73 0.66
Nov -4.67 -7.21 -9.98 -5.03 -6.53 -9.43
Dec   -1.84 -5.22 -8.72  -2.60 -5.81 -6.49
 
 
 
 
153 
Table C.13  Monthly Average of Daily TP concentration (mg/l) 
      HadCM3 A2  HadCM3 B2 
 Baseline 2020s 2050s 2080s  2020s 2050s 2080s 
Jan 0.080 0.093 0.106 0.141 0.116 0.123 0.106
Feb 0.079 0.090 0.098 0.134 0.112 0.113 0.103
Mar 0.080 0.091 0.093 0.124 0.102 0.114 0.095
Apr 0.075 0.083 0.089 0.112 0.098 0.108 0.094
May 0.080 0.084 0.093 0.116 0.099 0.113 0.097
Jun 0.089 0.100 0.109 0.135 0.114 0.129 0.109
Jul 0.092 0.113 0.126 0.158 0.130 0.144 0.125
Aug 0.097 0.129 0.141 0.173 0.142 0.162 0.143
Sep 0.104 0.141 0.156 0.194 0.150 0.175 0.156
Oct 0.106 0.149 0.163 0.196 0.159 0.188 0.166
Nov 0.102 0.140 0.158 0.190 0.161 0.186 0.167
Dec 0.089   0.102 0.127 0.165  0.134 0.152 0.130
 
Table C.14 Standard Deviation of Daily TP concentration (mg/l) 
      HadCM3 A2   HadCM3 B2 
 Baseline 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 0.039 0.030 0.037 0.047 0.037 0.038 0.036
Feb 0.035 0.031 0.037 0.045 0.035 0.033 0.029
Mar 0.037 0.035 0.038 0.042 0.031 0.035 0.029
Apr 0.028 0.026 0.029 0.030 0.024 0.034 0.033
May 0.020 0.019 0.023 0.022 0.019 0.031 0.028
Jun 0.022 0.017 0.025 0.028 0.020 0.028 0.023
Jul 0.026 0.020 0.026 0.028 0.023 0.030 0.024
Aug 0.030 0.020 0.025 0.032 0.024 0.033 0.026
Sep 0.029 0.021 0.025 0.034 0.025 0.038 0.025
Oct 0.034 0.028 0.035 0.041 0.033 0.039 0.028
Nov 0.041 0.035 0.040 0.049 0.039 0.045 0.036
Dec 0.043   0.035 0.051 0.051  0.040 0.044 0.039
 
 
 
 
154 
Table C.15 Relative Change in Monthly Average of Daily TP concentration (mg/l) 
    HadCM3 A2   HadCM3 B2 
  2020s 2050s 2080s   2020s 2050s 2080s 
Jan 0.013 0.026 0.061 0.036 0.043 0.026
Feb 0.012 0.020 0.055 0.034 0.034 0.024
Mar 0.011 0.013 0.044 0.022 0.034 0.015
Apr 0.009 0.014 0.037 0.023 0.033 0.020
May 0.004 0.013 0.036 0.019 0.033 0.017
Jun 0.010 0.020 0.046 0.025 0.040 0.020
Jul 0.021 0.033 0.065 0.038 0.052 0.032
Aug 0.032 0.044 0.076 0.045 0.065 0.046
Sep 0.037 0.052 0.090 0.046 0.071 0.052
Oct 0.043 0.057 0.090 0.053 0.082 0.060
Nov 0.039 0.056 0.089 0.059 0.085 0.066
Dec   0.014 0.038 0.076  0.045 0.064 0.041
 
Table C.16 Relative Change in Monthly Average of Daily TP concentration (%) 
    HadCM3 A2   HadCM3 B2 
 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 16.30 32.19 76.20 45.09 53.18 32.78
Feb 14.68 24.77 70.16 42.55 43.61 30.66
Mar 13.40 16.05 55.20 27.73 42.92 18.43
Apr 11.85 18.98 50.13 30.94 44.88 26.62
May 5.25 16.62 44.46 23.56 41.62 21.34
Jun 11.66 22.11 51.97 27.82 44.82 22.33
Jul 22.73 35.91 70.89 40.95 56.08 34.88
Aug 33.14 45.43 78.57 46.41 66.97 47.32
Sep 35.50 50.10 87.07 44.41 68.45 50.00
Oct 40.33 53.42 84.66 49.96 77.55 57.07
Nov 37.97 55.23 87.37 58.29 83.16 64.72
Dec   15.21 43.03 85.50  50.76 71.49 46.24
 
 
 
 
155 
Table C.17 Monthly Average of Daily TN concentration (mg/l) 
      HadCM3 A2  HadCM3 B2 
 Baseline 2020s 2050s 2080s  2020s 2050s 2080s 
Jan 2.57 3.03 3.29 4.26 3.26 3.57 3.49
Feb 2.43 2.90 3.18 4.07 3.16 3.47 3.48
Mar 2.36 2.71 2.95 3.97 3.06 3.36 3.13
Apr 2.56 2.83 3.17 3.96 3.23 3.46 3.27
May 2.78 3.19 3.50 4.30 3.48 3.86 3.61
Jun 2.90 3.53 3.84 4.82 3.82 4.22 3.94
Jul 2.96 3.81 4.18 5.22 4.07 4.47 4.27
Aug 3.15 4.09 4.48 5.38 4.30 4.75 4.63
Sep 3.24 4.28 4.71 5.83 4.33 4.98 4.81
Oct 3.26 4.37 4.74 5.66 4.51 5.13 4.94
Nov 3.13 4.04 4.52 5.40 4.43 4.92 4.84
Dec 2.80   3.28 3.77 4.80  3.80 4.22 4.01
 
Table C.18 Standard Deviation of Daily TN concentration (mg/l) 
      HadCM3 A2   HadCM3 B2 
 Baseline 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 0.80 1.03 1.09 1.28 1.04 0.94 0.92
Feb 0.79 0.91 0.99 1.25 0.92 0.85 0.89
Mar 0.71 0.85 0.90 1.15 0.88 0.90 0.93
Apr 0.73 0.79 0.88 1.07 0.87 0.84 0.90
May 0.71 0.84 0.91 1.10 0.82 0.84 0.87
Jun 0.76 0.87 0.97 1.16 0.85 0.86 0.89
Jul 0.79 0.90 0.94 1.15 0.84 0.83 0.93
Aug 0.81 0.89 1.02 1.23 0.86 0.83 0.94
Sep 0.75 0.94 1.02 1.26 0.92 0.94 0.92
Oct 0.74 0.98 1.08 1.45 0.88 0.91 0.93
Nov 0.83 1.02 1.15 1.50 0.91 1.12 1.04
Dec 0.87   1.06 1.33 1.31  0.96 1.08 1.06
 
 
 
 
156 
Table C.19 Relative Change in Monthly Average of Daily TN concentration (mg/l) 
    HadCM3 A2   HadCM3 B2 
 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 0.46 0.72 1.70 0.69 1.00 0.92
Feb 0.47 0.74 1.64 0.73 1.04 1.05
Mar 0.35 0.59 1.60 0.69 1.00 0.77
Apr 0.27 0.61 1.39 0.66 0.90 0.71
May 0.41 0.73 1.52 0.70 1.09 0.84
Jun 0.63 0.94 1.92 0.92 1.32 1.05
Jul 0.85 1.22 2.25 1.11 1.50 1.31
Aug 0.95 1.34 2.23 1.15 1.60 1.48
Sep 1.04 1.48 2.59 1.10 1.75 1.57
Oct 1.11 1.48 2.40 1.26 1.87 1.68
Nov 0.91 1.38 2.27 1.30 1.79 1.71
Dec   0.48 0.97 2.01  1.00 1.42 1.21
 
Table C.20 Relative Change in Monthly Average of Daily TN concentration (%) 
    HadCM3 A2   HadCM3 B2 
 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 18.03 28.04 65.97 26.77 38.86 35.69
Feb 19.42 30.63 67.43 29.86 42.81 43.08
Mar 14.92 25.00 67.96 29.38 42.22 32.42
Apr 10.53 23.83 54.43 25.85 35.10 27.75
May 14.85 26.13 54.84 25.36 39.09 30.09
Jun 21.67 32.47 66.10 31.73 45.54 36.07
Jul 28.67 41.05 76.05 37.40 50.67 44.21
Aug 30.14 42.54 70.95 36.70 50.95 47.17
Sep 32.17 45.61 80.01 33.92 53.99 48.50
Oct 34.11 45.42 73.59 38.53 57.49 51.53
Nov 28.92 44.18 72.46 41.53 57.00 54.49
Dec   17.31 34.64 71.75  35.66 50.90 43.25
 
 
 
 
157 
Table C.21 Monthly Average of Daily Chlorophyll-a (?g/l) 
      HadCM3 A2  HadCM3 B2 
 Baseline 2020s 2050s 2080s  2020s 2050s 2080s 
Jan 0.25 0.72 1.15 0.63 1.76 0.67 0.39
Feb 0.52 1.25 1.70 0.80 2.81 1.89 1.16
Mar 0.05 0.49 0.07 0.72 0.63 0.50 0.26
Apr 1.52 0.23 0.47 3.73 2.07 2.40 1.90
May 9.15 8.13 11.96 17.97 10.94 15.55 9.01
Jun 25.76 30.66 28.03 19.65 36.88 40.35 25.44
Jul 19.72 31.21 21.89 9.24 29.53 25.97 16.10
Aug 20.98 30.09 20.55 8.06 32.09 25.92 12.89
Sep 32.11 41.92 29.48 17.25 37.34 37.33 23.96
Oct 10.05 22.86 28.42 33.48 22.07 31.33 32.09
Nov 0.95 2.95 3.90 8.08 2.62 4.62 5.76
Dec 0.05   0.08 0.07 0.37  0.43 0.09 0.10
 
Table C.22 Standard Deviation of Daily Chlorophyll-a (?g/l) 
      HadCM3 A2   HadCM3 B2 
 Baseline 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 1.73 3.89 3.95 2.75 5.32 2.32 1.47
Feb 3.08 3.84 6.02 3.60 8.20 7.77 4.57
Mar 0.03 2.60 0.18 5.22 3.38 2.47 1.13
Apr 4.85 1.04 1.78 8.71 6.84 8.40 7.77
May 12.02 13.10 14.81 19.19 14.68 17.78 13.09
Jun 20.31 16.50 15.90 15.62 16.57 15.41 15.16
Jul 16.88 12.58 11.93 9.52 12.11 10.94 11.11
Aug 17.27 11.99 10.28 6.73 12.86 12.08 9.14
Sep 17.49 12.51 12.96 16.46 14.54 16.73 15.57
Oct 11.28 14.83 13.95 14.29 14.06 13.79 13.18
Nov 3.59 4.81 4.94 8.50 4.70 6.22 7.01
Dec 0.07   0.27 0.10 1.15  2.48 0.16 0.17
 
 
 
158 
Table C.23 Relative Change in Monthly Average of Daily Chlorophyll-a (?g/l) 
    HadCM3 A2   HadCM3 B2 
 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 0.48 0.90 0.38 1.51 0.43 0.14
Feb 0.72 1.18 0.28 2.29 1.37 0.64
Mar 0.45 0.03 0.68 0.58 0.46 0.21
Apr -1.29 -1.05 2.21 0.55 0.88 0.38
May -1.02 2.80 8.82 1.79 6.40 -0.15
Jun 4.90 2.28 -6.11 11.13 14.60 -0.31
Jul 11.49 2.17 -10.48 9.80 6.25 -3.62
Aug 9.11 -0.44 -12.93 11.10 4.94 -8.10
Sep 9.81 -2.63 -14.85 5.24 5.23 -8.14
Oct 12.80 18.37 23.43 12.02 21.28 22.04
Nov 1.99 2.95 7.12 1.67 3.66 4.80
Dec   0.03 0.02 0.32  0.38 0.04 0.05
 
Table C.24 Relative Change in Monthly Average of Daily Chlorophyll-a (%) 
    HadCM3 A2   HadCM3 B2 
 2020s 2050s 2080s 2020s 2050s 2080s 
Jan 194.22 367.35 156.13 613.95 173.13 57.78
Feb 139.37 226.62 54.54 440.17 263.90 122.23
Mar 963.61 58.83 1455.49 1250.46 985.46 450.85
Apr -84.55 -68.82 145.18 36.14 57.82 24.75
May -11.17 30.62 96.34 19.54 69.91 -1.63
Jun 19.03 8.85 -23.71 43.21 56.68 -1.22
Jul 58.23 11.00 -53.14 49.71 31.66 -18.38
Aug 43.39 -2.09 -61.59 52.90 23.52 -38.60
Sep 30.56 -8.20 -46.26 16.31 16.28 -25.37
Oct 127.40 182.76 233.13 119.55 211.76 219.32
Nov 209.10 309.06 747.33 174.73 384.29 504.08
Dec  62.59 36.13 644.11  765.99 74.22 91.00