Quantifying the impact of changes in crop area on evapotranspiration regimes in the US corn and soybean belts through phenological modeling and data assimilation

In recent years, fluctuations in food, feed, and fuel prices have led to shifts in the area of cropland dedicated to maize and soybean cultivation in the Northern Great Plains. We report here on a modeling experiment that compares three different simulated scenarios for actual evapotranspiration (ETa) from maize-soybean dominated areas in North Dakota, South Dakota, Nebraska, Iowa, and Minnesota during the 2000-2009 growing seasons. Scenario 1 relies on MODIS-derived crop maps to provide a baseline of subpixel crop proportions; Scenario 2 increases the proportion of maize by to 100 percent; Scenario 3 substitutes grassland for half the maize. We use a simple soil water balance model of ETa linked to an empirically derived crop specific phenology model also capable of producing seasonal trajectories of canopy attributes. This coupled model has been successfully deployed using flux tower records from multiple locations in the central US. Forcing the coupled model using data from NLDAS, we derive seasonal trajectories of daily NDVI and ETa as well as phenological transition points for maize, soybean, and grassland for each scenario. Seasonal differences in ETa among the three scenarios underscore the importance of how land use modulates land surface phenologies and, in turn, water and energy balances.