Civil Engineering


First Advisor

Ayse Kilic

Date of this Version



Blankenau, Philip. Bias and Other Error in Gridded Weather Data Sets and Their Impacts on Estimating Reference Evapotranspiration. MS Thesis. University of Nebraska-Lincoln, 2017.


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 Kilic. Lincoln, Nebraska: May, 2017

Copyright (c) 2017 Philip A. Blankenau


Gridded weather data sets are increasingly used in a variety of hydrologic and agricultural applications due to their complete spatial and temporal coverage. One application of gridded data sets is the estimation of evapotranspiration (ET). Several operational remote sensing (RS) approaches for estimating ET, such as the SEBAL, METRIC and EEFlux models, require estimates of reference ET (ETref), where ETref is expected ET from a hypothetical reference crop of clipped grass or alfalfa. Gridded weather data provide for the computation of ETref in all areas of a remote sensing image, and therefore potentially remove the need for dense weather station data.

Given the increasing use of gridded weather data to estimate ETref, this study assessed the quality of gridded weather data estimates of ETref. To accomplish this evaluation, several gridded weather data sets – GLDAS-1, NLDAS-2, CFSv2 operational analysis, GRIDMET, RTMA and NDFD – were compared to weather station data that were considered to represent ‘ground truth’ across the continental United States. The stations were selected to represent reference conditions when possible. The four primary weather variables – near-surface air temperature, vapor pressure, wind speed and shortwave solar radiation - required to compute ETref, plus calculated ETref itself were compared.

The application of the same analysis to multiple gridded data sets made comparisons among the gridded data sets possible. Generally, the gridded weather data sets overestimated ETref. This was mainly due to overestimation of air temperature, shortwave radiation and wind speed, and underestimation of vapor pressure. RTMA had the most accurate weather data and the most accurate estimates of ETref due to its assimilation of vast amounts of surface weather data and its continual refinement. Surprisingly, the global data sets, GLDAS and CFSv2, generally performed better than their North American counterparts – NLDAS and GRIDMET. All gridded weather data sets may be useful for estimating ETref and employment in remote sensing ET models provided some procedures for correcting biases are developed.

Advisor: Ayse Kilic