Advancements in mapping potential wetland areas across the conterminous United States
On this page:
Considering wetland restoration at the regional, state, or watershed scale can provide targeted ecosystem services. Highly connected systems, particularly where hydrologic characteristics are foundational to the ecosystem, are more effective at providing ecosystem services compared to smaller, disjointed systems. There are currently limited resources to support landscape-level wetland planning that use consistent methods and data inputs across the conterminous United States. This study develops a map of potential wetland area, which may support better placement of restored or built wetlands. The Random Forest algorithm is used to model potential wetland area in each of 18 two-digit hydrologic units that encompass the conterminous United States. The 17 input variables used in the models include Sentinel-1 SAR imagery, topographic indices and indicators, and soil suitability. The final product of this study is a gridded dataset with 10-m resolution that identifies areas where the landscape, land use, and hydrology would likely support a built or restored wetland. The model predicted that 21.1% of the conterminous United States land can be categorized as potential wetland area, compared to the current estimate of 9.3% existing wetlands in the National Wetland Inventory. Further investigation into the resulting wetland network is being assessed for best placement of wetlands for the highest quality services. Research has shown that targeted placement of wetland systems is critical in ensuring ecosystem services are beneficial to human and environmental health. The resulting maps demonstrate the application of big data and machine learning to identify locations across a diverse landscape likely to support a specific ecosystem type. This product can be used to better place restoration projects to serve ecosystem- and community-wide health where placement of sites is critical to efficacy, including nutrient uptake and flood attenuation.