Assessing the Impacts of Wetlands on Discharge and Nutrient Loading: Insights from Restoring Past Wetlands with GIS-Based Analysis and Modeling

Abstract

Wetlands play an essential role in overland flow generation by retaining water and providing an environment for the hydrogeochemical functions of a watershed. Few studies have addressed the restoration of past wetlands using wetland-corrected land-use land-cover (LULC) datasets. We evaluated the impacts of wetlands on hydrology and nutrient (nitrogen) loading by comparing Present, Past, and Restored scenarios. A GIS-based data processing procedure was proposed for generating a wetland-corrected proxy LULC raster to mimic the past wetland distribution. This was subsequently incorporated in the Soil and Water Assessment Tool (SWAT) to simulate discharge and nitrogen loading. In addition, two models, with and without using the proxy LULC dataset, were developed to examine the impacts of the proxy LULC on the model performance. The effect of hydroclimatic variations was also assessed by comparing the simulated peak flows and nutrient loads in dry and wet years. The results demonstrated the necessity of using the wetland-corrected proxy LULC data in the water quantity and quality modeling, especially for wet years. Our modeling approach and wetland-oriented LULC correction procedure can be applied to other wetland-dominated watersheds to improve hydrologic and water quality modeling.

Appendix

Raster Data Processing Procedure

A raster data processing procedure was proposed to create wetland-corrected proxy LULC for the Past and Restored scenarios. Figure 12 illustrates the raster data processing procedure used to modify the existing LULC to mimic the past wetland conditions. It was important to ensure that both datasets had the same resolution and used the same projected coordinate system to prevent any errors during the raster processing. First, the NWI dataset was converted to raster data (Fig. 12a) with the same resolution (10 m) as the LULC dataset (Fig. 12b). Then, the converted raster NWI data was reclassified as “0” for all wetland grids and “1” for all non-wetland grids (Fig. 12c). Afterward, the two raster datasets (Fig. 12c and d) were multiplied using the raster calculator (Fig. 12e). For the resulting raster data (Fig. 12e), a value of “195” (since SWAT assigns different values to various land use types, with “195” specifically representing wetlands) was assigned to all cells with a value of “0” (Fig. 12f). Thus, the wetland-corrected proxy LULC raster dataset (Fig. 12g; Table 6), representing the Past scenario in the 1970s, was created. A similar procedure was employed to prepare the proxy LULC for the Restored scenario where the LULC of 2021 was used, assuming that all non-wetland LULC types remain unchanged. In both Past and Restored scenarios, wetland delineation was based on the wetland data in the 1970s.

Fig. 12

Illustration of the raster processing procedure to create wetland-corrected proxy LULC: (a) polygon to raster converted NWI, (b) 2001 LULC without wetland-correction, (c) a small part of NWI raster with “0” for wetland grids and “1” for non-wetland grids, (d) 2001 LULC raster, (e) processed raster from (c) and (d) using the raster calculator, (f) reclassified raster with “195” for wetlands, and (g) wetland-corrected proxy LULC for 1970s (past scenario)

Table 6 SWAT land use types in Fig. 12g (source: SWAT database, 2012)