Submitted to: American Meteorological Society Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 1/10/2009
Publication Date: 1/25/2009
Citation: Inamdar, A.K., French, A.N. 2009. ESTIMATION OF POTENTIAL EVAPOTRANSPIRATION FROM MERGED CERES and MODIS OBSERVATIONS. American Meteorological Society Proceedings, January 11-15, 2009, Phoenix, AZ. p. 1-8. Interpretive Summary: Accurate knowledge of evapotranspiration (ET) over agricultural lands is essential for assessing impacts of climate change and for implementing sound water management decisions. However accurate ET estimates are difficult to obtain, either because they are obtained from sparsely distributed ground observations, or because they are derived from satellite imagery that is too coarse (10-20 km) to distinguish between farm land uses. In this study a methodology is proposed to combine observations from multiple satellite sensors that could help derive ET estimates at a much more useful observational scale of 1 km. Visible, near infrared, and long wave radiation data from the NASA MODIS and CERES sensors were combined by calibrating 1-km data with a physically-based CERES energy flux model. Results at the 1 km scale showed energy flux prediction uncertainties less than 20%, suggesting that the method could greatly improve the ability to frequently estimate ET at useful spatial scales. These results will be important for climate modelers and research hydrologists seeking better ways to monitor crop water use.
Technical Abstract: Accurate and timely estimates of potential evapotranspiration (ET) and knowledge of their spatial and temporal distribution are essential for agriculture and water resource management as well as for understanding the impacts of climate variability on terrestrial systems. Because of the paucity and inconsistency of ground-based measurements of potential ET, satellite remote sensing offers the best hope of providing such data. In the present study, we explore a new methodology to estimate potential ET by taking advantage of the new generation of the Earth Observation System (EOS) satellites. While there have been many studies on satellite-based estimation of ET, they are limited to clear-sky conditions and for a fixed overpass time. The approach outlined here utilizes the Clouds and the Earth’s Radiant Energy System (CERES) instrument as a calibration source for the Moderate Resolution Imaging Spectroradiometer (MODIS) imager to produce estimates of net radiation at the surface at the spatial scale of 1 km, which is input into an appropriate surface energy balance formulation. The objective is to take advantage of the extensively developed and well validated CERES top-of-atmosphere (TOA) to surface flux algorithms and extend the same to MODIS by implementing a suitable narrowband-to-broadband conversion scheme for the MODIS imager channels. The diurnal variations of potential ET are further modeled by employing the visible channel data from the geostationary environmental satellite (GOES 10). The potential ET estimates will be compared with measurements at several North American sites with varying land cover types.