Preliminary study of spin-up processes in land surface models with the first stage data of Project for Intercomparison of Land Surface Parameterization Schemes Phase 1(a)

The spin-up of a land surface model (LSM) is broadly defined as an adjustment process as the model approaches its equilibrium following initial anomalies in soil moisture content or after some abnormal environmental forcings (e.g., drought). The spin-up timescale of LSMs has received little attention in the modeling community. This study uses results from Phase 1(a) of the Project for Intercomparison of Land Surface Parameterization Schemes, and finds that most land surface schemes require many years to come to thermal and hydrologic equilibrium with the forcing meteorology; the time needed depends on the total moisture holding capacity and the initialization of the moisture stores. The linear relationship established for bucket-type models is just a special case of that found for the more sophisticated nonbucket-type models, at least when the models start out with adequate soil moisture. When soil moisture begins at zero or when precipitation is set to zero, there is a nonlinear relationship. Sensitivity studies using the Biosphere-Atmosphere Transfer Scheme confirm that precipitation intensity, solar radiation forcing, vegetation cover, and stomatal resistance also affect the length of spin-up time. The results underline that the accurate calculation of precipitation and solar radiation incident at the Earth's surface is important for a realistic simulation of soil moisture content. Magnitudes of simulated heat fluxes at equilibrium are not related to the thickness of the soil layer below the rooting zone. For most LSMs, initial positive soil moisture anomalies are associated with initial positive evapotranspiration (E) anomalies, while initial negative anomalies of soil moisture are accompanied by the initial negative, but much stronger, E anomalies, as found in past studies performed with general circulation models. In addition, the implications of the spin-up for numerical weather prediction and climate simulation are discussed.