Estimation of land surface Evapotranspiration in Nepal using Landsat based METRIC model

TL;DR

This study demonstrates the use of Landsat 8 imagery and the METRIC model to accurately estimate land surface evapotranspiration across Nepal's diverse terrain, aiding agricultural planning.

Contribution

First application of high-resolution remote sensing-based ET estimation in Nepal's complex topography using the METRIC model.

Findings

Remote sensing ET estimates closely match ground measurements.

ET inversely related to elevation over Nepal.

Model achieves low RMSE and MBE in validation.

Abstract

Nepal has a topographically diverse terrain with variations from plain flatland to high hills and mountains within a small region. Due to the lack of abundant ground-based measurement sites, it is difficult to estimate evapotranspiration (ET) covering most of the areas of Nepal from lower to higher elevations. In this study, we proposed to use a remote sensing-based METRIC (Mapping Evapotranspiration at high Resolution with Internalized Calibration) model for estimating ET in Nepal. Landsat 8 imagery, which provides a fine spatial resolution (30 m), was used for the estimation. The results obtained from the METRIC model were compared with ground-based measurements from an independent Eddy-Covariance (EC) station. Besides ET, the estimated surface temperatures (Ts) from the remote sensing model were also compared with ground-based measurements for model validation. The results obtained from the remote sensing model were close to the ground-based measurements which establish the accuracy of the model. Root mean square error (RMSE) for hourly and daily ET was obtained as 0.06 mm/hr and 1.24 mm/day while mean bias error (MBE) for hourly and daily ET was observed as 0.03 mm/hr and 0.29 mm/day, respectively. Further, we analyzed the variations of ET with elevation for six different months and found that ET was inversely related to elevation, in general, over the regions of Nepal. This was established as an effect of Ts and vegetation distribution over different elevations. To the best of our knowledge, our study is a first that has investigated ET estimation over a topographically diverse region of Nepal with fine spatial resolution afforded by Landsat 8. Such large-scale ET estimation in Nepal has several applications. This can be leveraged for agricultural planning and forecasting in a country where most of the population still relies on agriculture for their daily livelihood.