an improvement to modis tpw products by damping the variation of surface reflectance in channel 2

total perceptible water (tpw) is an important parameter in climatology and weather forecasting and is directly related to any climate process. there are three approaches to estimate this parameter i.e. using radiosonde, using gps and calculating from satellite images where the first two are localized and the last one can give an instant view of tpw in a vast region. the algorithm used for the tpw calculation from modis images is related to the ratio of the reflectance in a water vapor absorbing channel and the reflectance in a non-absorbing channel. due to strong horizontal variation in the surface reflectance in non-absorbing channels, the retrieved tpw varies strongly from one pixel to its neighboring pixels while it is believed that the horizontal gradient of tpw is very weak. to solve this problem, a damping coefficient was added to the non-absorbing channel reflectance. it is found that this coefficient differs for different surface covers. the current work presents a procedure for calculating these coefficients. the results of a comparison between modified tpw and those extracted from gps data showed a r2 of 0.81 whilst this was about 0.67 for non-modified modis tpw.