Location: Location not imported yet.Title: A modeling framework for deriving daily time series of evapotranspiration maps using a surface energy balance model
|KHAND, KUL - Oklahoma State University
|TAGHVAEIAN, SALEH - Oklahoma State University
|PAUL, GEORGE - Non ARS Employee
Submitted to: Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/26/2019
Publication Date: 3/2/2019
Citation: Khand, K., Taghvaeian, S., Gowda, P.H., Paul, G. 2019. A modeling framework for deriving daily time series of evapotranspiration maps using a surface energy balance model. Remote Sensing. 11(5):508. https://doi.org/10.3390/rs11050508.
Interpretive Summary: Time series of remotely sensed evapotranspiration (ET) maps have extensive applications in agricultural, hydrological and environmental studies as they capture the spatiotemporal variability of vegetation consumptive use from field to continental scales. In this study, an automated modeling framework was developed and tested to construct daily time series of ET maps for Oklahoma for the 2001-2014 period using MODIS imagery and the Surface Energy Balance System model. The ET estimates generated from this modeling framework were validated against observations from three eddy-covariance towers located in central Oklahoma. Although the modeling framework slightly overestimated ET but captured its spatial and temporal variability. Overall, the performance of the proposed ET modeling framework was good with potential for a range of applications.
Technical Abstract: Surface energy balance models have been one of the most widely used approaches to estimate spatially distributed evapotranspiration (ET) at varying landscape scales. However, more research is required to develop and test an operational framework that can address all challenges related to processing and gap filling of non-continuous satellite data to generate time series of ET at regional scale. In this study, an automated modeling framework was developed to construct daily time series of ET maps using MODIS imagery and the Surface Energy Balance System model. The ET estimates generated from this modeling framework were validated against observations of three eddy-covariance towers in Oklahoma, United States during a two-year period at each site. The modeling framework overestimated ET but captured its spatial and temporal variability. The overall performance was good with mean bias errors less than 30 W m-2 and root mean square errors less than 50 W m-2. The model was then applied for a 14-year period (2001–2014) to study ET variations across Oklahoma. The statewide annual ET varied from 841 to 1100 mm yr-1, with an average of 994 mm yr-1. The results were also analyzed to estimate the ratio of estimated ET to reference ET, which is an indicator of water scarcity. The potential applications and challenges of the ET modeling framework are discussed and the future direction for the improvement and development of similar automated approaches are highlighted.