The influence of land surface moisture retention on precipitation statistics

Abstract

The retention of precipitation water in land surface reservoirs damps higher frequencies of evaporation variability and can thereby influence the feedback of evaporation on precipitation. The extent of this influence is examined in a series of general circulation model simulations in which the timescale of surface moisture retention is very carefully controlled. Shorter timescales lead to increased daily precipitation variance and one-day-lagged precipitation autocorrelations but to decreased autocorrelations at longer lags. An explanation for the simulated precipitation statistics is offered in the form of a heuristic model of evaporation feedback that describes precipitation variance and autocorrelation in terms of three parameters: (i) the timescale of precipitation persistence in the absence of feedback; (ii) the surface retention timescale; and (iii) a parameter describing the atmosphere's responsiveness to variations in evaporation. The heuristic model reproduces the statistical trends seen in the GCM diagnostics, and it can be used to explain geographical variations in precipitation statistics generated by a GCM coupled to a full biosphere model.