Spin-up time scales of the Off-line Land Surface GEOS Assimilation (OLGA) system. The spin-up time scale of a land surface model is the time that a model takes to arrive at its equilibrium state following initial anomalies in the land surface state with respect to a time series of atmospheric forcing conditions. A robust land surface model can adequately treat realistic anomaly initial conditions to arrive at a realistic equilibrium state. The characteristic spin-up time scales can be used to assess the robustness of land surface models, particularly with respect to the surface moisture and energy conservation. In this study we investigate the spin-up processes in the Off-line Land Surface GEOS Assimilation (OLGA) system. OLGA is a global version of the Mosaic land surface model (LSM, Koster and Suarez 1992) coupled with near surface atmosphere through a turbulence scheme used in the GEOS Data Assimilation System (DAS) at Data Assimilation Office, NASA-GSFC.
We examine the spin-up time scales in a 20-year long perpetual 1992 OLGA integration. The GEOS-DAS 1992 data provided all surface atmospheric forcing. The initial conditions of land surface state were interpolated from the climatology of a general circulation model coupled with the Mosaic LSM. Also, a similar 20-year long perpetual 1992 OLGA integration was performed using an observationally based surface downward longwave and shortwave radiation as a forcing condition. The results show that the spin-up time scale required by the OLGA to arrive at its thermal equilibrium state is much shorter than that required for hydrological equilibrium. The spin-up time scale of available soil moisture varies globally and depends on initial soil moisture anomaly. Almost all the grid points with long spin-up time scales are located poleward of 50N. The factors determining the spin-up time scales will be investigated.