Comparison of satellite-derived estimates of gross primary production for Australian old-growth tropical rainforest

The moderate-resolution spectroradiometer (MODIS) sensor has provided estimates of near-real-time gross and annual net primary production (GPP and NPP) since March 2000. We compare 2 years (2000 and 2001) of satellite-based calculations of GPP in the seasonally wet tropical rainforests of north Queensland, Australia, with estimates derived from an ecosystem process model, physiological principles predicting growth using satellite data (3-PGS), that was previously calibrated with ground-based measurements at 21 sites. This comparison provides an independent evaluation of the MOD-17 GPP product because the 3-PGS simulations account for spatial variation in soil properties (available soil water and fertility) and rely on different sources of meteorological data and derived estimates of the fraction of photosynthetically active radiation (fPAR). Results of the comparison show that estimates of averaged annual and monthly GPP from MOD-17 are slightly higher in magnitude than those from 3-PGS, although the two models exhibit comparable seasonal dynamics over the 2-year analysis period. Some of the differences in estimates of productivity and temporal dynamics between 3-PGS and MODIS are a result of limitations in the types of input data used for each model. In addition, the particularly infertile soils of northern Australia account for a reduction in productivity that is not directly incorporated in the MOD-17 algorithm. Further testing and continued improvement of the MOD-17 input datasets and algorithm construction will see improved estimates of satellite-derived productivity in tropical forest regions across the globe.