The Sensitivity of Land-Atmosphere Coupling to Modern Agriculture in the Northern Midlatitudes

Abstract

Modern agricultural land cover and management are important as regional climate forcings. Previous work has shown that land cover change can significantly impact key climate variables, including turbulent fluxes, precipitation, and surface temperature. However, fewer studies have investigated how intensive crop management can impact background climate conditions, such as the strength of land-atmosphere coupling and evaporative regime. We conduct sensitivity experiments using a state-of-the-art climate model with modified vegetation characteristics to represent modern crop cover and management, using observed crop-specific leaf area indexes and calendars. We quantify changes in land-atmosphere interactions and climate over intensively cultivated regions situated at transitions between moisture- and energy-limited conditions. Results show that modern intensive agriculture has significant and geographically varying impacts on regional evaporative regimes and background climate conditions. Over the northern Great Plains, modern crop intensity increases the model simulated precipitation and soil moisture, weakening hydrologic coupling by increasing surface water availability and reducing moisture limits on evapotranspiration. In the U.S. Midwest, higher growing season evapotranspiration, coupled with winter and spring rainfall declines, reduces regional soil moisture, while crop albedo changes also reduce net surface radiation. This results overall in reduced dependency of regional surface temperature on latent heat fluxes. In central Asia, a combination of reduced net surface energy and enhanced pre-growing season precipitation amplify the energy-limited evaporative regime. These results highlight the need for improved representations of agriculture in global climate models to better account for regional climate impacts and interactions with other anthropogenic forcings.

FIG. 1.

(a) Fractional crop coverage shown at the native model resolution of 2° latitude × 2.5° longitude. (b) The number of crop cycles represented in each of the modified grid boxes, shown at the 0.5° resolution of the Sacks et al. (2010) dataset. Grid boxes may contain more than one crop, and in some cases, such as the U.S. Midwest, many boxes represent a rotation between crops. The resulting LAIs are aggregated to ModelE resolution using the methods described herein

FIG. 2.

(a) Maximum annual LAI for the NatVeg experiment. (b) The change in maximum LAI between the CropSpec and NatVeg experiments. The blue boxes denote the domains of focus: (i) the Great Plains, (ii) the Midwest, and (iii) central Asia.

FIG. 3.

June–August latent heat flux (Wm⁻²) regressed against soil moisture in the top 50 cm of the soil column. Colors show the correlation between latent heat flux and surface temperature—a measure of the land–atmosphere coupling. Blue colors indicate negative correlations while red colors show positive correlations. Plots are shown for grid boxes lying within the outlined (a) Great Plains, (b) Midwest, and (c) central Asia domains.

FIG. 4.

Monthly average values over the Great Plains domain for (a) LAI, (b) latent heat flux (W m⁻²), (c) total precipitation (mm), (d) soil moisture in the top 50 cm of the soil column (mm), and (e) sensible heat flux (W m⁻²) and (f) the difference in surface temperature between the CropSpec and NatVeg experiments. Pink stars indicate the statistical differences between both experiments for a specific month, following a modified Student’s t test.

FIG. 5.

As in Fig. 4, but for the Midwest domain.

FIG. 6.

As in Fig. 4, but for the central Asian domain.

FIG. 7.

JJA differences between CropSpec and NatVeg for (a) shortwave radiation incident at the surface, (b) upwelling shortwave radiation from the surface, (c) downwelling longwave radiation, (d) upwelling longwave radiation, and (e) net surface radiation. Only grid boxes with significant differences are shown, following a modified Student’s t test unless otherwise specified. Blue boxes highlight the three domains of focus.

FIG. 8.

JJA differences between CropSpec and NatVeg for (a) surface temperature (°C), (b) latent heat flux (W m⁻²), (c) sensible heat flux (W m⁻²), (d) precipitation (% change), and (e) low-level cloud cover (%). Blue boxes highlight the three domains of focus. Only significant differences at the 0.05 level are shaded.

FIG. 9.

(a) JJA soil moisture (mm) in the top 50 cm for the NatVeg experiment. (b) The change in soil moisture (%) between the CropSpec and NatVeg experiments.

FIG. 10.

The JJA Pearson correlation coefficient between interannual latent heat flux and surface temperature for the (a) NatVeg and (b) CropSpec experiments. Blue colors show negative correlations, and red colors indicate positive correlations. (c) Changes in the correlation between the CropSpec and NatVeg experiments. Values have been scaled to show where correlative strength is reduced (brown) and where it is enhanced (teal).