Sensitivity Analysis of METRIC–Based Evapotranspiration Algorithm

Document Type : Original Research Paper

Authors

1 1Department of Remote Sensing, Faculty of Geo-information science and engineering, University Technology of Malaysia (UTM), Malaysia a Faculty of natural resources, Yazd University, Yazd, Iran

2 Department of Remote Sensing, Faculty of Geo-information science and engineering, University Technology of Malaysia (UTM), Malaysia

3 Graduate faculty of Environment, University of Tehran, Tehran, Iran

Abstract

METRIC (Mapping Evapotranspiration at High Resolution with Internalized Calibration) is known as an appropriate surface energy balance model for the estimation of the spatial distribution of evapotranspiration (ET) in semi-arid regions. Based on lysimeter measurements, METRIC has shown ET estimates of 10% on a sub-field scale on a daily basis. There is a need to identify how the model is sensitive to the input parameters. Therefore, the most influential parameters in the algorithm can be identified and the
model can be further improved. Sensitivity analysis at three levels of vegetation cover shows that METRIC is
highly sensitive to dT, surface temperature, net radiation, sensible heat flux, surface albedo, soil heat flux, and air temperature. It is also moderately sensitive to friction velocity, aerodynamic resistance to heat transfer, surface emissivity and less sensitive to leaf area index, soil adjusted vegetation index, wind speed(except wind speed at low level of vegetation cover)‘, and roughness length for momentum (except Zom<0.1). A two-factor analysis of the algorithm’s primary inputs showed that the pair albedo-surface temperature is the most and the normalized vegetation index-soil adjusted vegetation index or normalized vegetation index-leaf area index is the least effective pair in this model. In order to improve the accuracy of METRIC, this study suggests upgrading the equations for the above-mentioned effective variables.

Keywords

International Journal of Environmental Research
Volume 7, Issue 2 - Serial Number 2
April 2013
Pages 407-422

Files

Share

How to cite

Statistics