Investigating the issues of scale and land surface representation in the Community Climate System Model (CCSM) with the Community Land Model (CLM), Oleson et al. (2004) recently found that climate modelling experiments using satellite derived land surface datasets that capture sub-grid heterogeneity and land biogeophysics, have significant impacts on the land surface climate simulated in the model. In addition to the changes in land surface climate, the satellite data also diminished the magnitude and spatial extent of climate changes that resulted in continental United States land cover change experiments, compared to experiments conducted with biome derived datasets at the model grid increment. The reduction in magnitude of the climate changes was attributed to the ecological similarity between land cover types involved in the land use change, and the reduced intensity of agriculture identified in the satellite-derived datasets. The study suggested that the realism of the datasets used to represent land cover change, and the parameterizations selected to represent land cover biogeophysics, needed to be carefully considered to understand and quantify the impacts of land cover change on climate.

In the context of this sensitivity to land surface representation and data, there may be significant implications for the differences recently identified by Tian et al (2004) and Oleson et al. (2003) between the current land surface parameters of the CLM, and Moderate Resolution Imaging Spectroradiometer (MODIS) satellite derived data for Leaf Area Index (LAI), Fraction of Photosynthetically Active Radiation (FPAR), and land surface albedo. To asses the impacts on climate modeled in the CLM and the CCSM from these differences in land surface properties, this research has developed new land surface parameters using land surface data derived from fine scale (0.05 degree) NOAA Pathfinder Advanced Very High Resolution Radiometer (AVHRR) and MODIS satellite imagery. The new land surface data globally describes the current land surface at a scale that captures land surface heterogeneity and land cover change. The new methods and land surface data also ensure that the new CLM parameters reproduce the land surface conditions in the model as closely as possible to the land surface conditions represented in the AVHRR and MODIS satellite data for any given month.

The mapping of CLM Plant Functional Type (PFT) distributions was performed at the 0.05 degree resolution, following the methods used in CLM2 by Bonan et al (2002), with the physiology and climate rules developed by Nemani and Running (1996). The percentage Tree Cover, Leaf Type, and Leaf Longevity were derived from AVHRR Continuous Fields Tree Cover Project data from DeFries et al (2000). The 0 – 80% cover values of the original data were scaled to 0 – 100% PFT values so that at the densest tree cover values the tree PFTs were given 100% value. This addressed the over prescription of understorey PFTs, in particular grass, from the raw tree cover values alone. Understorey and herbaceous PFT distributions were derived from IGBP Global Land Cover Characterization land cover mapping, with the distribution of C3/C4 grass adapted from the fractional C3/C4 mapping methods of Still et. al. (2003). Climate data was compiled from 1970 – 2000 monthly surface air temperature and precipitation surfaces generated by Willmont and Matsuura (2001) from the Global Historical Climate Network.

The monthly LAI parameters of the CLM were derived from quality assured and averaged MODIS V4 LAI data, to give mean 2001 – 2003 monthly values at the 0.05 degree resolution. The grid average monthly LAI values were allocated to the various PFTs in each grid cell by the PFT percentages and the PFT LAI max values. Additional rules were used to enforce known leaf phenology and Stem Area Index (SAI) from SAI min and LAI values. The leaf phenology of Summer Green Deciduous Trees was enforced using Growing Degree Days (GDD) from a Tbase Temperature following the LPJ dynamic vegetation model of Sitch et al (2003). Evergreen leaf phenology was enforced, and SAI calculated from the methods of Zeng et al (2002).

The visible and near infrared soil and leaf optical parameters used in the Two Stream Radiation Model of the CLM were derived from MODIS V4 monthly Black Sky (Direct Beam) and White Sky (Diffuse Beam) surface albedo mapping in each spectrum. The MODIS albedo data was quality assured with snow and cloud pixel replaced with closest monthly values to produce mean 2001 – 2003 monthly snow free surface albedo. The PFT optical properties and fine scale soil reflectance values were reverse calculated over a range of values to find the reflectance values that reproduced the monthly snow free MODIS albedo values in the Two Stream Radiation Model, given the PFT percentages with the monthly PFT LAI and SAI values.

The new land surface parameters are currently being run in sensitivity experiments with the CCSM 3.0 to assess the climate response of the model to the differences in land surface representations of the existing parameters and the new fine scale AVHRR and MODIS derived parameters. Land cover change studies also are to be performed with “Natural” land surface conditions reconstructed using historical and potential vegetation mapping in conjunction with the current conditions, with the remnant natural vegetation used to develop surrogate parameters for the “Natural” land cover.