Sensitivity of the land surface to sub-grid scale processes: implications for climate simulations

A. J. Pitman*

*Corresponding author for this work

    Research output: Contribution to journalArticlepeer-review

    3 Citations (Scopus)


    Using a state-of-the-art land surface model in a 'stand-alone' mode with prescribed atmospheric forcing, a method for retaining the spatial extent and intensity of precipitation in Atmospheric General Circulation Models (AGCMs) is investigated. It is shown that the surface climatology simulated by this model is strongly dependent upon the fraction of the grid square, μ, receiving precipitation. It is also shown that fundamentally different hydrological regimes (one runoff dominated, the other evaporation dominated) are obtained for the otherwise identical situations. It is argued that the new generation of land surface models which explicitly incorporate vegetation may have to be 're-tuned' if differences between large-scale and small-scale precipitation events are accommodated in AGCMs. If precipitation is always assumed to fall uniformly over the grid square, the precipitation intensity will generally be underestimated. This will lead to an overestimation of canopy interception and interception loss. Furthermore, too little precipitation will reach the soil surface, and therefore surface runoff and the soil moisture will be underestimated. With too much interception loss, and too little soil water (and soil evaporation) there will be a tendency to re-cycle precipitation back to the atmosphere too quickly, leading to the unrealistic simulation of surface-atmosphere interactions. These results, if reproduced within an AGCM, would invalidate previous simulations of the effects of changing the state of the land surface on the atmosphere.

    Original languageEnglish
    Pages (from-to)121-134
    Number of pages14
    Issue number1-2
    Publication statusPublished - Jan 1991


    Dive into the research topics of 'Sensitivity of the land surface to sub-grid scale processes: implications for climate simulations'. Together they form a unique fingerprint.

    Cite this