You are here:
Characterizing the Surface Connectivity of Depressional Wetlands: Linking Remote Sensing and Hydrologic Modeling Approaches
Citation:
Christensen, J., G. Evenson, M. Vanderhoof, Q. Wu, H. Golden, AND C. Lane. Characterizing the Surface Connectivity of Depressional Wetlands: Linking Remote Sensing and Hydrologic Modeling Approaches. 2017 AGU Fall Meeting, New Orleans, LA, December 11 - 15, 2017.
Impact/Purpose:
Wetland-to-wetland and wetland-to-stream connections vary both spatially and temporally in pothole watersheds and are important to understanding hydrologic and biogeochemical processes in the landscape. To explore how to best characterize spatial and temporal variability in aquatic connectivity, we compared three approaches, 1) hydrological modeling alone, 2) remotely-sensed data alone, and 3) integrating remotely-sensed data into a hydrological model. These approaches were tested in the Pipestem Creek Watershed, North Dakota across a drought to deluge cycle (1990-2011).
Description:
Surface connectivity of wetlands in the 700,000 km2 Prairie Pothole Region of North America (PPR) can occur through fill-spill and fill-merge mechanisms, with some wetlands eventually spilling into stream/river systems. These wetland-to-wetland and wetland-to-stream connections vary both spatially and temporally in PPR watersheds and are important to understanding hydrologic and biogeochemical processes in the landscape. To explore how to best characterize spatial and temporal variability in aquatic connectivity, we compared three approaches, 1) hydrological modeling alone, 2) remotely-sensed data alone, and 3) integrating remotely-sensed data into a hydrological model. These approaches were tested in the Pipestem Creek Watershed, North Dakota across a drought to deluge cycle (1990-2011). A Soil and Water Assessment Tool (SWAT) model was modified to include the water storage capacity of individual non-floodplain wetlands identified in the National Wetland Inventory (NWI) dataset. The SWAT-NWI model simulated the water balance and storage of each wetland and the temporal variability of their hydrologic connections between wetlands during the 21-year study period. However, SWAT-NWI only accounted for fill-spill, and did not allow for the expansion and merging of wetlands situated within larger depressions. Alternatively, we assessed the occurrence of fill-merge mechanisms using inundation maps derived from Landsat images on 19 cloud-free days during the 21 years. We found fill-merge mechanisms to be prevalent across the Pipestem watershed during times of deluge. The SWAT-NWI model was then modified to use LiDAR-derived depressions that account for the potential maximum depression extent, including the merging of smaller wetlands. The inundation maps were used to evaluate the ability of the SWAT-depression model to simulate fill-merge dynamics in addition to fill-spill dynamics throughout the study watershed. Ultimately, using remote sensing to inform and validate process-based modeling allows us to assess both the spatial and temporal continuum of connections across a watershed, identify approach limitations, and improve efforts to study and map wetland connectivity and hydrologic effects in the PPR and beyond.
