Probability of afternoon precipitation in eastern United States and Mexico enhanced by high evaporation

Moisture and heat fluxes from the land surface to the atmosphere form a critical nexus between surface hydrology and atmospheric processes, particularly those relevant to precipitation. Although current theory suggests that soil moisture generally has a positive impact on subsequent precipitation, individual studies have shown support both for1–4 and against5–7 this positive feedback. Broad assessment of the coupling between soil moisture and evapotranspiration, and evapotranspiration and precipitation, has been limited by a lack of large-scale observations. Quantification of the influence of evapotranspiration on precipitation remains particularly uncertain. Here, we develop and apply physically based, objective metrics for quantifying the impacts of surface evaporative and sensible heat fluxes on the frequency and intensity of convective rainfall during summer, using North American reanalysis data. We show that high evaporation enhances the probability of afternoon rainfall east of the Mississippi and in Mexico. Indeed, variations in surface fluxes lead to changes in afternoon rainfall probability of between 10 and 25% in these regions. The intensity of rainfall, by contrast, is largely insensitive to surface fluxes. We suggest that local surface fluxes represent an important trigger for convective rainfall in the eastern United States and Mexico during the summer, leading to a positive evaporation– precipitation feedback. Observational studies of soil moisture–rainfall interactions generally suffer from insufficient data both spatially and temporally, particularly for soil moisture (SM) and surface turbulent fluxes. In models, the strength of the SM-rainfall feedback depends on model parameterizations and grid resolution8–10. The metrics introduced in the present study constitute powerful diagnostic tools for model intercomparison of simulated land–atmosphere coupling, and for validation of fundamental physical processes incorporated in nextgeneration earth systemmodels. Previous work11 identified three necessary conditions for an initial SM anomaly to impact summertime precipitation: (1) the initial SM anomaly must be large; (2) evaporation must be strongly sensitive to SM; and (3) precipitation must be strongly sensitive to evaporation. The first element emphasizes areas with large SM variability, whereas the latter elements emphasize the connection between the land surface and the atmospheric boundary layer (ABL), and the connection between the ABL and overlying free troposphere11. This study emphasizes (3): we demonstrate when