PARENT SESSION
Oral Session #30: Modeling.
Presiding: T. Day
Tuesday, August 6. 8:00 AM to 11:30 AM. Apache Meeting Room, TCC.
Sensitivity of MODIS derived photosynthesis and net primary productivity to relative accuracy of meteorological inputs.
Zhao, Maosheng*,1, Jolly, William2, Kimball, John3, Nemani, Ramakrishna4, Running, Steven5, Kang, Sinkyu6, 1 University of Montana, Missoula, Montana2 University of Montana, Missoula, Montana3 University of Montana, Missoula, Montana4 University of Montana, Missoula, Montana5 University of Montana, Missoula, Montana6 University of Montana, Missoula, Montana
ABSTRACT- The NASA Earth Observing System now produces regular (8-day interval) estimates of terrestrial photosynthesis and annual net primary production at a 1km spatial resolution globally (~150 million cells). MODIS (Moderate Resolution Imaging Spectroradiometer) is the primary EOS sensor for providing data on terrestrial biosphere dynamics and activity, including photosynthesis and net primary production. An entire year (2001) of global PSN data are now available. These data are derived using 1km resolution daily spectral information from MODIS and meteorological inputs obtained from the NASA DAO (Data Assimilation Office) at a relatively coarse 1 degree spatial resolution; meteorological inputs include daily average temperature (Ta), minimum temperature (Tm), average vapor pressure deficit (VPD), and short wave radiation (SWrad). The objective of this investigation is to assess the sensitivity of the MODIS PSN and NPP algorithms to the spatial scale of input meteorological drivers. First, we compare DAO and surface weather station network data in different seasons globally to assess the temporal and spatial quality of the assimilation data, and it is found that DAO data are relatively good. Then we use surface station data to derive MODIS PSN and NPP to assess the relative impact of variability in input meteorology. PSN variability is found more sensitive to SWrad and VPD. These analyses not only evaluate MODIS17 products, but also reveal potential future algorithm improvements.
KEY WORDS: sensitivity, algorithm, inputs