Valuing Ecosystem Services from Forested Landscapes: How Urbanization Influences Drinking Water Treatment Cost

Abstract

For two decades high total organic carbon (TOC) levels in Converse Reservoir, a water source for Mobile, Alabama, have concerned water treatment officials due to the potential for disinfection byproduct (DBP) formation. TOC reacts with chlorine during drinking water treatment to form DBPs, some of which are carcinogenic and regulated under the Safe Drinking Water Act. Previous studies have shown that raw water TOC concentration >2.7 mg L-1 in Converse Reservoir can cause elevated DBPs during warm weather (May to October). Additional chemical treatment, such as use of powdered activated carbon (PAC) at the water treatment plant is necessary at this plant when raw water TOC concentration exceeds 2.7 mg L-1. TOC in drinking water reservoirs originates from either watershed sources or internal algal growth. This study evaluated, through paired watershed and reservoir modeling with actual atmospheric data from 1991 to 2005, how urbanization may alter chlorophyll a, total nitrogen (TN), total phosphorus (TP), and TOC concentrations in Converse Reservoir. The Converse Watershed on the urban fringe of Mobile is projected to undergo considerable urbanization by 2020. A base scenario using 1992 land cover was paired with 2020 projections of land use. The Loading Simulation Program C++ (LSPC) watershed model was used to evaluate changes in nutrient concentrations (mg L-1) and loads (kg) to Converse Reservoir. Combined urban and suburban area was simulated within the watershed from an initial 3% in 1992 to 22% in 2020. From 1992 to 2020, forest to urban land conversion increased TN and TP loads to Converse Reservoir by 109 and 62%, respectively. TOC load increased by 26% compared to base land use. Forest to urban land conversion increased monthly stream flows in 94% of months simulated (1991 to 2005) by a mean increase of 14%. Simulated urbanization generally increased streamflow, but decreased monthly streamflow by 2.9% during drought months. Simulated future overall median TN and TP concentrations (0.82 and 0.017 mg L-1, respectively) were 59 and 66% higher than base concentrations (0.52 and 0.010 mg L-1, respectively); but future median TOC concentration (3.3 mg L-1) was 16% lower than base concentrations. Increased total urban flow caused overall TOC loads (kg) to increase by 26% during the simulation period despite lower TOC concentrations. Monthly analysis indicated significantly elevated TOC concentrations in June, July and August (p<0.05) following simulated urbanization. Simulated annual TOC export ranged from 12.7 kg ha-1 y-1 in a severe drought year to 52.8 kg ha-1 y-1 in the year with the highest precipitation. Post-urbanization source water TOC concentrations in the receiving water body will likely increase more than predicted by the watershed model since larger TP loads following urbanization will support increased reservoir algae growth, further increasing internal generation of TOC. To evaluate reservoir nutrient concentrations in response to urbanization, LSPC watershed model streamflow and selected water quality constituents were input into the Environmental Fluid Dynamics Code (EFDC) reservoir model. EFDC calibration and validation performance ratings for chlorophyll a, TN, TP and TOC ranged from ‘satisfactory’ to ‘very good’. Between 1992 and 2020, simulated forest to urban land conversion increased median overall TOC concentration in the reservoir by 1.1 mg L-1 (41%). From 1992 to 2020, monthly median TOC concentrations between May and October increased 33 and 49% as a result of urbanization. Simulated chlorophyll a, indicating algae growth, accounted for most of the variance in simulated TOC concentration at the reservoir intake between May and November. Base scenario daily TOC concentrations between May and October exceeded 2.7 mg L-1 on 47% of days simulated. Daily TOC concentrations between May and October using the 2020 land use continuously exceeded 2.7 mg L-1. Consequently, based upon simulated urbanization, increased urban land use will result in elevated reservoir TOC concentrations from both autochthonous and allochthonous sources and the need for additional water treatment between May and October. The cost for additional chemical treatment to offset DBP formation was based on simulated values for raw water TOC at the source water intake. Assuming a PAC cost of $1.72 kg-1, the daily mean increase in treatment cost following forest to urban land conversion was between $4,700 and $5,000 d-1. This corresponds to a value of $91 to $95 km2 d-1 converted from forest to urban land use. The economic value of forested watersheds for source water protection related to drinking water quality has long been recognized but rarely quantified within an existing cost structure. This research determined that the ecosystem services for reservoir water TOC provided by forest land in the Converse Watershed were $91 to $95 km2 d-1 or $12,080 to $25,190 km2 y-1. Since the influence of forest to urban land use change on TOC concentrations varies, this value is watershed specific. The ecosystem services provided by forested land related to source water TOC in the Converse Watershed were within previously reported estimates for all water provision ecosystem services from forested catchments. Simulated reservoir TOC concentrations indicated that without additional chemical treatment at the drinking water plant, expected urbanization will likely increase carcinogenic DBP formation in the drinking water supply distribution system. The 3% urban land use of the 1992 base simulation maintained TOC concentrations near the TOC threshold such that additional treatment was likely unnecessary. Simulation of future urban land increased May to October reservoir TOC concentrations such that additional treatment would be necessary to mitigate DBP formation and safeguard human health.