Seasonal and Global Representation of Land Surface Properties from MODIS and other EOS instruments and their Implications for Application in Climate ModelsProposal to NASA NRA-03-OES-02 Earth Science System Research Using Data and Products from Terra, Aqua, and Acrim SatellitesPrincipal Investigator: Robert E. Dickinson Tel: Fax: Email: robted@eas.gatech.edu Co-Investigators: Liming Zhou1, Yuhong
Tian1,  Abstract The proposed
work is to explore and better quantify issues raised about the application of
MODIS land data to climate models in our current IDS investigation. We have
been examining the seasonal and spatial variations of albedo,
FPAR, and LAI and their connections to other MODIS fields. In doing so, we
have established that the largest variability and greatest uncertainties
occur in the winter-spring seasons in high latitudes of the Northern
Hemisphere and in semi arid regions or regions with extensive cloud cover. Of
particular interest are the effects of snow on the retrieval of albedo and LAI information. The spectral signal of a snow
background obscures the presence of vegetation with the current LAI-FPAR
retrievals but may be adequately accounted for in the current albedo retrievals. However, the compositing method of the
albedo retrievals only uses either the snow-free or
the snow-covered scenes, depending on which occur more often over the
compositing period. Climate models currently are unable to accurately specify
land albedos for boreal forests or high latitude
shrubs underlain by snow from their descriptions of snow cover and land use
because they do not know enough about the structure and openness of such
vegetation. In addition, they require a poorly known characterization of stems
and dead leaves called the stem-area index (SAI). Sparse vegetation underlain
by mineral soil presents a similar difficulty in climate models since the albedos of such soils is inadequately known. With
extensive cloud cover, the primary issue is to ensure that cloud
contamination effects have been removed in the data used to establish climate
model parameters. We have found that the seasonal cycle LAI-FPAR data from
MODIS for mid to high latitudes over a small region can be adequately
described by 4 to 6 parameters – that is the minimum winter values, the
maximum summer values, and 2 to 4 values describing the periods of greenup and leaf browning and drop. Any information as to
time variation in the winter is obscured by the biases from snow and in
summer from signal saturation. In addition, the season transition periods may
occur over only a few weeks and so their details may be obscured by
compositing over 8 days or more as currently done for LAI and albedo. The objectives of the proposed investigation will
be to further explore the issues mentioned above in the application of MODIS
terra and aqua data to climate models. In particular, can we establish better
estimates of LAI in winter by use of the retrieved albedo
fields? Can we likewise characterize |
Atmospheric Dynamics and
Climate
Georgia Tech