Abstract
This paper describes a methodology of early-stage water stress detection for forest vegetation using a newly proposed index, CDNT (Compensated Day Night Temperature difference), which uses only remotely sensed surface temperature and has a close relationship with latent heat flux compared to simple differences of day-night surface temperatures. The comparison between CDNT calculated from the MODIS surface temperature and the ground-measured latent heat flux data at 8 Ameriflux sites in a period of 11 years, from year 2000 to 2010, showed a good relationship. The authors also found that a reduction in precipitation significant enough to cause a drought introduces an anomalously high positive value of CDNT. Anomalous CDNT values, based on the 11-year average daily CDNT value, can be used as indicators of water stress for forest vegetation. Since this method uses only remotely sensed surface temperature data, it is suitable for use in large- or global-scale water stress monitoring.
