Quantifying Errors in FAO Crop Reference Estimates Toward Algorithm Improvement

Abstract:

Crop reference evapotranspiration (ETo) is a hypothetical quantity defined for hypothetical well-irrigated and large grass fields. In practice, such fields do not exist. Estimation of ETo is typically made through the FAO method (based on the Penman-Monteith Equation) using data from weather stations over fields that violate the requirements of ETo. We hypothesize that the FAO ETo estimates are subject to three main error sources: (1) local advection error, resulting from the use of station data over well-irrigated fields surrounded by dry area, (2) surface aridity error, resulting from the use of station data over water-stressed fields, and (3) estimation of net long wave radiation using vapor pressure data. In this study, we quantified and modelled the errors from these error sources, using station data collected over grass fields in various situations: (1) CIMIS station data over well-irrigated grass fields in California, (2) selected MESONET station data over water-stressed grass fields in Oklahoma, and (3) selected Ameriflux data over water-stressed grass fields across the US, and (4) radiation data from the BSRN stations across the U.S. Our results demonstrate that the FAO ETo estimates have large errors, and these errors can be reduced by the use of readily available solar radiation data (obtained using station or satellite data).