Summary Results from:

Thermal-based techniques for land cover change detection using a new dynamic MODIS multispectral emissivity product (MOD21)
As they relate to the validation of mod21

Authors: Hulley, G., S. Veraverbeke, S. Hook

Source: Rem. Sens. Environ, 140, p755-765

Link to: Access Publication


Land Surface Temperature and Emissivity (LST&E) data determine the amount of net longwave radiation emitted from the Earth's surface and are therefore critical variables for studying a variety of surface-atmosphere processes over land such as evapotranspiration, land cover change, and surface composition. Because emissivity is an intrinsic property of the surface, multispectral thermal infrared emissivity data have the potential for enhancing our ability to monitor landscape changes in environmentally sensitive zones beyond what is currently possible from standard practices used today. The most common of these practices is the use of visible to short-wave infrared data, in particular the Normalized Difference Vegetation Index (NDVI). Two algorithms are currently used to generate the LST&E products from MODIS data, but studies have identified several issues with both these algorithms that limit their usefulness for land cover change detection. These issues have been recently addressed by applying the ASTER Temperature Emissivity Separation (TES) algorithm to MODIS thermal infrared data to generate LST and a dynamically varying multispectral emissivity product for bands 29, 31, and 32 at 1-km resolution. The new product (MOD21) will be released with MODIS Collection 6 during fall 2013. This study demonstrates the utility of the dynamic MOD21 multispectral emissivity product to detect land cover changes over a broad range of different Earth surface domains including land degradation in dryland regions, snow melt characteristics on glaciers and ice sheets, extreme ecosystem disturbances, and agricultural activities. The MOD21 spectral emissivity provided increased sensitivity to land cover change in a more consistent manner than is currently possible with other emissivity products and, depending on the application, standard visible to near infrared (VNIR) data. The results suggest that synergistic use of thermal and VNIR data will help us to better identify and understand changes in the Earth surface system, and reduce uncertainties in estimating their magnitudes and trends