Civil Engineering


Date of this Version

Fall 12-2011


A THESIS Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of Requirements For the Degree of Master of Science, Major: Civil Engineering, Under the Supervision of Professor Ayse Irmak. Lincoln, Nebraska: August, 2011

Copyright (c) 2011 Ravi Choragudi


Mapping EvapoTranspiration at high Resolution using Internal Calibration (METRIC) is most widely used to quantify evapotranspiration (ET) spatially and temporally. It is essential to inspect the model’s response to errors in various parameters used in the model. Landsat 5 images from May 30 2009, July 1 2009 and a Landsat 7 image from September 27 2009 are used in this study. Fourteen different fields composed of Corn, Soybeans, Alfalfa are randomly chosen for each crop type.

Two kinds of errors are addressed in this study. One, with the errors that are transferred and potentially compensated by calibration (Global error) and the other is the error that is not passed into the calibration (Local error). For global error, Reflectance at the satellite (ρ), transmissivity (τ), surface temperature (Ts), wind speed (u), Reference Evapotranspiration (ETr) are chosen. In addition, the sensitivity towards selection of hot and cold pixels is also investigated. For local errors, albedo (α), surface temperature (Ts), momentum roughness length (Zom), soil heat flux (G), difference between air and surface temperature (dT) are considered.

In this study, we have found that METRIC is able to compensate most of the global errors passed through the calibration to give consistent results, when the variables considered above has changed to their extremes. ETr should be estimated at a good degree of accuracy to maintain the METRIC’s results to be realistic. Also, selection of hot and cold pixels is the most crucial and sensitive process in METRIC.

In case of local errors: Zom is relatively insensitive to the model. dT is found to be the most sensitive variable for bare soils. However, the other parameters are linearly proportional to their errors.

Adviser: Ayse Irmak