HR: 1330h
AN: H42A-1057    [PDF]
TI: Soil Water Access Heterogeneity in Land-surface Parameterisation Schemes
AU: Irannejad, P
EM: pix@ansto.gov.au
AF: ANSTO Environment, PMB 1, Menai, NSW 2234 Australia
AU: Phillips, T
EM: phillips14@llnl.gov
AF: PCMDI, Lawrence Livermore National Laboratory, Livermore, CA 94450 United States
AU: Henderson-Sellers, A
EM: ahssec@ansto.gov.au
AF: ANSTO Environment, PMB 1, Menai, NSW 2234 Australia
AU: * McGuffie, K
EM: kendal.mcguffie@uts.edu.au
AF: Department of Applied Physics, University of Technology, Sydney, Broadway, NSW 2007 Australia
AB: Confidence in Atmospheric General Circulation Models' (AGCMs') predictions, the predominant tools with which we predict future climate, depends on successful validation of their simulations. Land-surface climates, particularly important for people as the land is where we live and grow food, suffer from the lack of high quality global observational datasets of variables such as evapotranspiration and soil moisture that are not directly observable at scales appropriate to atmospheric models. We evaluate simulations by 20 AGCMs participating in the second phase of the Atmospheric Model Intercomparison Project (AMIP2) (Phillips {\it et al.}, 2002, Irannejad {\it et al.}, 2003) under the auspices of the Project for Intercomparison of Land-surface Parameterisation Schemes (PILPS, Henderson-Sellers {\it et al.}, 2002), against available observations and reanalyses along with a global off-line land surface simulation by the Variable Infiltration Capacity (VIC) scheme. Our hypothesis is that the largest contribution of soil moisture variability to precipitation variability will arise in models that use land-surface parameterization schemes (LSSs) that derive the largest fraction of their evapotranspiration from rooting zone soil moisture. This is based on the Dickinson {\it et al.} (2003) suggestion regarding differences in the partitioning of evapotranspiration and the Henderson-Sellers {\it et al.} (2003) demonstration that different LSSs cause differences in the prediction of continental surface climates that are not compensated by atmospheric feedbacks in coupled simulations. Underlining the importance of land-surface parameterization for climate prediction, Koster {\it et al.} (2002) demonstrated that different LSSs coupled to different GCMs generate a wide range of simulations of precipitation and its variability resulting from soil moisture variability. Here we use the AMIP2 results to (i) examine the validity of the Dickinson {\it et al.} hypothesis; and (ii) discover a means of ranking LSS parameterizations.
DE: 0315 Biosphere/atmosphere interactions
DE: 0322 Constituent sources and sinks
SC: Hydrology [H]
MN: 2003 Fall Meeting