Data from: Projected land use changes will cause water quality degradation at drinking water intakes across a regional watershed
Data files
Feb 27, 2025 version files 13.66 GB
-
dryad_data.zip
13.66 GB
-
README.md
3.30 KB
Abstract
Forest conversion to development threatens the ability of watersheds to provide stable and clean water supplies. Water managers are targeting forest conservation as a source water protection strategy to maintain healthy watershed function in developing areas, especially upstream of drinking water treatment facilities. Understanding the role of current forest cover in safeguarding these facilities is therefore crucial. We used the Soil and Water Assessment Tool (SWAT) to examine the relationship between upstream forest cover and downstream water resources under both current (2020) conditions and multiple projected land use scenarios for 2070 in the Middle Chattahoochee watershed, USA. We investigated the extent to which existing forest cover benefits water quality at 15 drinking water intake facilities within the watershed. Our analysis of four land use change scenarios revealed that forest conversion and increased development resulted in higher average annual concentrations of total suspended sediment (TSS) and total nitrogen (TN) at 13 out of 15 intake facilities, with potential increases of up to 318% for sediment and 220% for nitrogen. Conversely, concentrations decreased relative to the baseline when upstream agricultural land was converted to forest cover or new, low-intensity development, suggesting that certain types of development may improve water quality compared to maintaining agricultural land. Our results also indicated that extreme nitrogen and sediment concentration events – defined as days exceeding the highest 10th percentile of baseline concentrations – could become 3.6 to 6.6 times more frequent in the future, respectively. Notably, forest conversion to new development upstream of intakes with smaller subwatersheds could reduce water quality for utilities serving smaller towns and rural areas, which may have limited resources for managing this challenge. Our findings highlight vulnerable intake locations and underscore the benefit of forest conservation for source water protection under future land use change.
https://doi.org/10.5061/dryad.3r2280gsm
Description of the data and file structure
This dataset contains the data required to replicate figures and tables from Gay et al. (2025), which assessed the effect of land use change on water quality at drinking water intakes in the Middle Chattahoochee watershed. Data is comprised of spatial and tabular data. Spatial data includes the boundary from SWAT watershed delineation, a shapefile of the Chattahoochee River, and land cover from the various scenarios. Tabular data is comprised of outputs from the SWAT model scenarios. All scenarios, sources, and methodologies are detailed in Gay et al. (2025).
Files and variables
File: dryad_data.zip
Description: To access data, download the dryad_data.zip folder. Data is structed in two main folders: spatial_data and swat_outputs. File structure is as follows:
Spatial Data:
- flow_lines: this shapefile line data represents the Chattahoochee River derived from the National Hydrography Dataset (NHD).
- watershed_boundary: this shapefile polygon data represents the boundary for the Middle Chattahoochee produced through the SWAT watershed delineation step.
- land_cover:
- all_forested: a raster where all terrestrial land in the 2011 National Land Cover Dataset (NLCD) raster for the Middle Chattahoochee was transformed to mixed forest cover (NLCD class 43; SWAT class FRST).
- all_developed: a raster where all terrestrial land in the 2011 NLCD raster for the Middle Chattahoochee was transformed to medium-intensity development (NLCD class 23; SWAT class URHD).
- 2020_basemap_NLCD: a raster representing the basemap for the baseline scenario.
- 2070_hadSSP3_NLCD: a raster representing land use for the LG-1 scenario (low growth; HadGEM2-ES and SSP3)
- 2070_hadSSP5_NLCD: a raster representing land use for the HG-1 scenario (high growth; HadGEM2-ES and SSP5).
- 2070_mriSSP3_NLCD: a raster representing land use for the LG-2 scenario (low growth; MRI-CGCM3 and SSP3).
- 2070_mriSSP5_NLCD: a raster representing land use for the HG-2 scenario (high growth; MRI-CGCM3 and SSP5).
SWAT Outputs:
- all_developed: SWAT model outputs from the all_developed scenario, including outputs for .rch, .cio, .std, .hru., .std., .sub.
- all_forested: SWAT model outputs from the all_forested scenario, including outputs for .rch, .cio, .std, .hru., .std., .sub.
- baseline2011: SWAT model outputs from the scenario used to calibrate the model (2011 NLCD), including outputs for .rch, .cio, .std, .hru., .std., .sub.
- basemap2020: SWAT model outputs from baseline scenario, including outputs for .rch, .cio, .std, .hru., .std., .sub.
- HADSSP3: SWAT model outputs from the LG-1 scenario, including outputs for .rch, .cio, .std, .hru., .std., .sub.
- HADSSP5: SWAT model outputs from the HG-1 scenario, including outputs for .rch, .cio, .std, .hru., .std., .sub.
- MRISSP3: SWAT model outputs from the LG-2 scenario, including outputs for .rch, .cio, .std, .hru., .std., .sub.
- MRISSP5: SWAT model outputs from the HG-2 scenario, including outputs for .rch, .cio, .std, .hru., .std., .sub.
