Hydroclimate variability and predictability in the Southeastern United States
Type of DegreePhD Dissertation
Forestry and Wildlife Science
Restriction TypeAuburn University Users
MetadataShow full item record
The hydroclimate is the study of the hydrological cycle in the climate context, it is defined as the intersection of climatology and hydrology, especially energy and moisture exchanges between and within the two systems. There are three hydroclimate projection uncertainties, future emission-scenario uncertainty, model-response uncertainty, and natural variability. The internal variability as unforced variability to climate system fluctuations was unlikely to diminish, even with more accurate projections of greenhouse-gas concentration, or model projection. By analyzing multi-model ensemble and single-model ensemble, the projection of hydroclimate internal variability is quantified on the global and regional scale. This study provides further explanation of its responses to climate warming and predictability. Chapter 1 is a literature review on hydroclimate variability in the climate warming background. Global warming intensifies drought and inundation events and enlarges the climate projection uncertainty. The Southeastern US (SEUS) is the so-called “warming hole”, which is slightly cooling, opposite to the temperature rise in the US for the second-half of the twentieth century (Meehl et al. 2012). It provides a unique test-bed to improve hydroclimate prediction by assessing the role of internal variability. The signal of hydroclimate internal variability became detectable from the background noise in certain model scenarios and with the selection of ensemble members. Strengthening climate model signals and reducing noise will help to improve water cycle prediction in the SEUS, especially soil moisture and streamflow. Chapter 2 is an evaluation of long-term temperature trends and variability in the CMIP6 multimodel ensemble. This study conducts a robust assessment of the Coupled Model Intercomparison Project Phase 6 (CMIP6) to capture the observed temperature trends and variability at global and regional scales. The warming rate in the second half of the twentieth century (0.19°C/decade) is twice as large as in the full analysis period (1901 to 2014; 0.10°C/decade). Multidecadal climate variability results in considerable uncertainties in the regional temperature trend, but the multidecadal variability does not represent a statistically significant trend. Globally, the spatial pattern of trends is most similar among ensemble members of the same model, then among climate models, and the least similar between models and observations. The structural uncertainty and internal variability of climate models provide a range of temperature trends that generally encompass the regional scale observations. Some single-model large ensembles (SLE) also have variability comparable to the multimodel large ensemble (MLE), encompassing the regional scale observations. Chapter 3 is a study of a hybrid physics-AI Model to improve NWM hydrological forecasts. The National Oceanic and Atmospheric Administration (NOAA) has developed a very high-resolution streamflow forecast using the National Water Model (NWM) for 2.7 million stream locations in the United States. However, considerable challenges exist in quantifying uncertainty at ungauged locations and forecast reliability. A data science approach is presented to address the challenge. The long-range daily streamflow forecasts are analyzed from Dec. 2018 to Aug. 2021 for Alabama and Georgia. The forecast is evaluated at 389 observed USGS stream-gauging locations using standard deterministic metrics. Next, the forecast errors are grouped using watersheds’ biophysical characteristics, including drainage area, land use, soil type, and topographic index. The NWM forecasts are more skillful for larger and forested watersheds than smaller and urban watersheds. The NWM forecast considerably overestimates the streamflow in the urban watersheds. The classification and regression tree (CART) analysis confirms the dependency of the forecast errors on the biophysical characteristics. In Chapter 4, the interpretation of the signal-to-noise ratio paradox is revised to constrain the climate projections in the region with a high ratio of predictable component (RPC). The signal-to-noise ratio paradox was assessed in long-term climate projections using Single-model Large Ensemble (SLM) and Multi-model Large Ensemble (MLE) climate data. A null hypothesis was constructed by performing bootstrap resampling of climate model ensembles to test its ability to predict the 20th-century temperature and precipitation trends locally and compare it with the observations. Rejection of the null hypothesis indicates the existence of a paradox. The large multi-model ensemble does not reject the hypothesis in most places globally. The rejection rate in the single-model large ensemble is related to the model's fidelity to simulate internal climate variability rather than its ensemble size. For regions where the null hypothesis is rejected, for example, India, the paradox is caused by a smaller signal strength in the climate model's ensemble. The signal strength was improved by 100% through ensemble selection and based on past performance, which reduced uncertainty in India's 30-year temperature projections by 25%. Chapter 5 quantifies the predictability from the land/ocean/atmosphere initializations, their couplings, and climatology components to sub-seasonal soil moisture forecast skill. Sub-seasonal to seasonal climate prediction is vital for agricultural planning. In this study, we evaluate the contribution of land, atmosphere, and ocean sources to the root zone soil moisture forecast skill using the SubX Community Earth System Model version 2 (CESM2) CAM6 forecast experiments. The anomaly correlation coefficient (ACC) derived from the reforecast is an almost linear combination of the three predictability sources of land, atmosphere, and ocean, their couplings, and climatology components. The soil moisture anomaly correlation shows that land-only initialization contributes 89% of the total forecast skill in the control experiment, and its contributions are greatest in the Great Plains and the central southern US. The majority of the seasonal-averaged predictability is from the soil moisture memory. The CESM2 soil moisture forecast skills are compared with two other SubX climate models: ESRL-FIM and RSMAS-CCSM4. The study identifies soil moisture's initial condition to forecast precipitation feedback as a potential factor for improving the soil moisture forecast skill.