Estimating evapotranspiration (ET) using satellite sensors in the North American Monsoon region in western Mexico

The North American Monsoon (NAM) system brings abundant summer rains to the western slopes of the Sierra Madre Occidental between latitudes 15º N and 35º N in Mexico. This rainfall produces large increases in foliage so that a significant proportion of the precipitation is recycled back to the atmosphere through soil and leaf evaporation and, more importantly, plant transpiration. Two flux towers were installed in 2004 for the NAM Experiment and another two more recently, so that there are energy balance and ET measurements over four important cover types: Tropical Deciduous Forest, Subtropical Shrub, Oak Grassland and Mesquite Woodland. In this work, these datasets are used to validate estimates for the distribution of ET over the region obtained using satellite remote sensors. The estimation method is based on energy balance constraints. The maximum or potential ET is limited by the available energy for transforming liquid water into vapor and can be equated to a large fraction of the incoming solar energy at ground level. For clear sky conditions, this can be accurately determined using sun-earth geometry and estimates of atmospheric turbidity. To determine incoming solar energy under cloudy conditions, we include a cloud index (between 0 and 1) that can be obtained from GOES Channel 11 data, available every 15 to 30 minutes during daylight hours at 1 km. The maximum ET calculated in this way corresponds to an area (pixel) with complete vegetation cover, so that we must multiply this by an estimate for the actual vegetation to obtain the actual ET. The MODIS 16 day composite EVI product was used here to obtain a cover factor between 0 and 1 since the change in vegetation condition is relatively slow. This method has been used successfully in previous studies, mainly in cultivated lands in Europe and Africa. We will show maps of the ET distribution obtained in NW Mexico and its comparison to flux tower data.