Modeling the Convective Response to Land-Use Change in the Northern Great Plains

The northern North American Great Plains (NNAGP) have undergone substantial changes in atmospheric boundary layer processes that are thought to arise from changes in land management. Namely, declines in the practice of summer fallow (on the order of millions of hectares) has shifted Bowen ratios from ca. 2 to ca. 1 during the growing season, and concomitant adoption of no-till agriculture has increased surface albedo. Consistent with these large-scale changes in land surface attributes the lifted condensation level and atmospheric boundary layer heights are now lower on average, and the probability of convective likelihood has increased by about 10% in some parts of the NNAGP. The atmospheric boundary layer contains more moisture under these new conditions, increasing convective available energy (CAPE). These changes are also related to observed summer cooling (or lack of warming); cloud cover has increased by 4%, surface net radiation has decreased by 6 W m^-2, and JJA temperatures have decreased by 1 °C across parts of the NNAGP.

Large uncertainties exist in our understanding of these land-atmosphere feedbacks on regional climate. Whereas previous studies have attributed changes in regional climate and precipitation to the decline of the practice of summer fallow, the trends are also coincident with the emergence of no-till agriculture (which reduces surface net radiation by increasing albedo) and other changes to the global climate system. It is further unclear how changes in land management have (or have not) altered climate across different regions within the NNGAP. To further investigate these trends, the Weather Research and Forecasting (WRF) model is used to simulate past and present land cover conditions and feedbacks to convection. Eddy covariance flux measurements are used to inform the land surface parameters initially included in the model that are derived from MODIS. We modify the land use classification of the NOAH model grid points to match those observed for 1980 and then to those to 2015, using unmodified (1990s) runs as controls for each time period. Our domain is centered over northeastern Montana, a region where land use change and precipitation dynamics have seen substantial changes since the 1980s, notably a tripling of May precipitation. Results are discussed in the context of land surface-precipitation feedbacks.