Canadian Forest Service Publications
Estimating afternoon MODIS land surface temperatures (LST) based on morning MODIS overpass, location, and elevation information. 2007. Coops, N.C.; Duro, D.C.; Wulder, M.A.; Han, T. International Journal of Remote Sensing 28(10): 2391-2396.
Available from: Pacific Forestry Centre
Catalog ID: 27345
Remotely sensed predictions of land surface temperature are an important variable that is routinely used to determine surface energy budgets and daily evaporation rates. The MODerate Imaging Spectroradiometer (MODIS) instrument on-board the Terra and Aqua satellites is a critical tool for providing diurnal estimates of LST and allows a global monitoring capacity of the terrestrial biosphere. MODIS, onboard Terra, which has a morning (AM) overpass, was launched in December 1999, and MODIS onboard Aqua, with an afternoon overpass (PM), was launched in May 2002. Generally, LST of vegetation in the afternoon would be expected, under cloudless conditions, to be warmer than the temperature retrieved by the morning satellite overpass, therefore with Aqua PM overpass are likely to be much closer to the maximum daily temperature of the land surface. In addition PM LST retrievals have been shown to give a better indication of the thermal response of rising leaf temperatures. Consequently, LST retrieved by the Aqua satellite are better suited for regional and global change studies. This letter investigates differences between the Aqua MODIS PM and Terra MODIS AM LST estimates over a range of land cover classes, locations and dates, across Canada in 2003. We compare data from both platforms with the aim of developing an adjustment which can be applied to Terra AM LST estimates to approximate a “synthetic” Aqua PM LST product from 2000 – mid 2002 to provide a seamless afternoon LST data from 2000 – 2006. Results indicate that there are, as expected, statistically significant differences between the AM, and PM LST across all cover types and at four key times of the year across Canada ranging from 0.3° - 3.2°C depending on cover type, and between 1.2° and 5° depending on time of year. On average 90% of the variation observed in the PM record can be explained be AM LST with the degree of correspondence dependent on cover type with crops being the lowest and forest classes being the highest. A simple regression model using AM LST, and the three positional variables, predicted PM LST’s statistically indistinguishable from observed TERRA PM LST.
- Date modified: