Impact of precipitation-induced sensible heat on the simulation of land-surface air temperature

Precipitation-induced sensible heat (HPR) which is transferred between the land surface and rainwater can be extremely large during a heavy precipitation event. Thus, the local surface temperature can be sharply altered on an hourly to daily timescale. However, HPR is commonly neglected in current land surface models because of its small magnitude on long timescales. As a consequence, the simulated land-surface air temperature (TLSA) may be biased. In this study, we use satellite and reanalysis data sets to estimate HPR on the global scale, and we use CESM1 (with CAM4 physics and the prescribed sea surface temperature) to investigate the effect of HPR on TLSA simulations over the second half of the 20th century. Our results show that the reanalysis-estimated HPR is largest over Intertropical Convergence Zone (ITCZ) regions, with seasonal mean values of -0.22 W m(-2) (-2.31 to 0.29 W m(-2)) in DJF and -0.28 W m(-2) (-2.89 to 0.24 W m(-2)) in JJA. These values are consistent with the model-simulated HPR patterns. In DJF, the HPR leads to a weaker equator-pole surface heating difference and therefore, weakens northern stationary waves. The consequent changes of the heat and moisture advection result in noticeable TLSA warming effects (for example, +0.53 K over northern Eurasia and +0.46 K over central North America) and cooling effects (for example, -0.62 K over Alaska and -0.42 K over central Asia) at regional scales. These temperature changes help to reduce the TLSA biases in the model. In JJA, the TLSA changes slightly because of the weak stationary wave responses during the northern summer.