EMC/NCEP has completed a set of more than 4000 retrospective integrations of the new Climate Forecast System (CFS), which became operational in August 2004. The CFS is a fully coupled atmosphere-ocean-land modeling system. The ocean model is the GFDL MOM3, with 1/3rd degree resolution near the equator, and the atmospheric model is the NCEP GFS at T62L64 resolution. The land model is composed of a 2-layer soil model and canopy treatment as described by Pan and Mahrt (1987), with coupling to the planetary boundary layer scheme described by Hong and Pan (1996). The 4000+ retrospective runs were produced from 15-member ensembles corresponding to initial conditions from all 12 months within the 23-year period 1981-2003. All integrations were for 9-months. The initial ocean states were obtained from the NCEP/GODAS, and Reanalysis-2 data were used to provide initial conditions for the atmosphere and land states. The enormous retrospective data set is ideal for diagnostic and empirical studies. Here we pursue issues regarding the feedback of the land surface in models, compared to the feedback as determined from a global “observed” soil moisture data set (Fan and van den Dool 2004). In Huang and van den Dool (1993; see their Fig.2) we demonstrated that models (in use at that time) produced a much greater feedback from soil moisture onto local surface temperature and local precipitation over the U.S. than the magnitude of the feedback in the observed data. Here, one decade later, we have repeated these calculations using the CFS retrospective dataset, not just for the US, but globally, in order to investigate both local and non-local feedbacks. Quantifying non-local feedbacks (especially on precipitation) presents an even greater challenge. In order to investigate non-local impacts, we employed an approximate Green’s function to construct idealized soil moisture anomalies (van den Dool et al 2003, see example below), each centered at all points over land in the CFS global domain. This method was applied to the CFS model data and to the observed soil moisture data, in order to locate ‘sensitive’ land areas, both in reality and in the “model world.” A selection of these sensitive areas is explored with additional model integrations, as described in a companion paper (Saha and van den Dool 2005).

References: Fan, Y., and H. van den Dool (2004), Climate Prediction Center global monthly soil moisture data set at 0.5° resolution for 1948 to present, J. Geophys. Res., 109, D10102, doi:10.1029/2003JD004345.

Hong, S.-Y. and H.-L. Pan, 1996: Nonlocal boundary layer vertical diffusion in a medium-range forecast model. Mon. Wea. Rev., 124, 2322-2339.

Huang, J. and H. M. van den Dool, 1993: Monthly precipitation-temperature relation and temperature prediction over the U.S. J. Climate, 6, 1111-1132.

Pan, H-L. and L. Mahrt, 1987: Interaction between soil hydrology and boundary layer developments. Boundary Layer Meteor., 38, 185-202.

Van den Dool, H. M., Jin Huang and Yun Fan, 2003: Performance and Analysis of the Constructed Analogue Method Applied to US Soil Moisture over 1981-2001. J. Geophys. Res., 108(D16), 8617, doi:10.1029/2002JD003114,2003.