The impact of agricultural intensification on hydrologic disconnectivity in the Mississippi Delta region of the Mississippi Alluvial Plain, USA

Since the 1970s, the Mississippi Alluvial Plain (MAP - a US EPA Level III Ecoregion) has experienced significant agricultural intensification via irrigation practices. The MAP overlies the Mississippi River Valley Alluvial Aquifer, presently the second most withdrawn aquifer in the USA (~46 million m³ per day). The increased irrigation demand has supported a ~7x increase in irrigated production; whereby, about 70% of all MAP cropland is now irrigated. The irrigation intensification has resulted in similar increases in crop yield. These MAP patterns are exemplified within northwestern portions of the US state of Mississippi, a region referred to as the Mississippi Delta. Within the Mississippi Delta (~18,000 km²), the predominant irrigation application is furrow irrigation, which has been facilitated by precision land leveling. As such, berms are typically placed around the lower elevations of the leveled field to detain runoff. Runoff is then slowly released through one (or more) outlets into a ditch or other drainage system. Consequently, the current topography and hydrography of the Mississippi Delta is vastly different  from historical records and remains dynamic. The alterations in surface hydrology and hydrologic connectivity will also influence how sediments, nutrients, and other agrochemicals are processed ecologically. The size of the Mississippi Delta limits comprehensive monitoring of runoff, transport, and transformation processes across the entire landscape. Therefore, scientists at the U.S. Department of Agriculture, Agricultural Research Service, National Sedimentation Laboratory are developing computer models and supporting databases to evaluate those controlling processes at the basin scale. Current US national databases of surface hydrography in the Mississippi Delta (e.g., the National Hydrography Dataset Plus High Resolution; NHDPlus HR) are based on elevation data (10 m resolution digital elevation model) collected before land leveling. Hence, the NHDPlus HR drainage model is too coarse, and significantly differs from reality. Using machine learning (ML) technology we have characterized the drainage network from high-resolution lidar data (avg. point density > 2 points/m²) collected during the period 2018-2020. The ML-derived drainage model includes ditches (as narrow as 3 m) and identifies field outlets. To assess how surface hydrology and connectivity have changed, and possible implications on water quantity and quality, we are building two HEC-RAS models using the pre-1970 hydrography and the 2018-2020 hydrography. Our test case features the 12-digit hydrologic unit code (HUC) subwatershed: Roundaway Bayou-Quiver River (HUC # 080302070805; surface area is ~162 square kilometers), which occupies central portions of the Mississippi Delta. Because HEC-RAS accounts for the effects of subgrid-scale topography on surface runoff, we can accurately describe the 2018-2020 high-resolution (1 m horizontal) topography and hydrography at larger spatial resolution; in our simulations we therefore used a grid with cells of 10 m horizontal resolution. We will present results on the changes in surface runoff (magnitude, direction, retention, and connectivity) and implications for water quantity and quality in the Mississippi Delta region.