TITLE:
Estimation of Evapotranspiration by Various Net Radiation Estimation Formulae for Non-Irrigated Grass in Brazil
AUTHORS:
Antonio Ribeiro da Cunha, Edgar Ricardo Schöffel, Clovis Alberto Volpe
KEYWORDS:
Evapotranspiration, Net Radiation, Solar Radiation, Cloud Cover, Empirical Models
JOURNAL NAME:
Journal of Water Resource and Protection,
Vol.6 No.15,
November
24,
2014
ABSTRACT: The
objective of this study was to assess the accuracy of estimating
evapotranspiration (ET) using the FAO-56 Penman-Monteith (FAO-56-PM) model,
with measured and estimated net radiation (Rnmeasured and Rnestimated,
respectively), the latter obtained via five different models. We used meteorological
data collected between August 2005 and June 2008, on a daily basis and on a
seasonal basis (wet vs. dry seasons). The following data were collected:
temperature; relative humidity; global global solar radiation (Rs); wind speed
and soil heat flux. The atmospheric pressure was determined by aneroid
barograph, and sunshine duration was quantified with a Campbell-Stokes
recorder. In addition to the sensor readings (Rnmeasured), five
different models were used in order to obtain the Rnestimated. Four
of those models consider the effects of cloud cover: the original Brunt model;
the FAO-24 model for wet climates; the FAO-24 model for dry climates, and the
FAO-56 model. The fifth was a linear regression model based on Rs. In
estimating the daily ET0 with the FAO-56-PM model, Rnmeasured can be
replaced by Rnestimated, in accordance with the FAO-24 model for dry
climates, with a relative error of 2.9%, or with the FAO-56 model, with an
error of 4.9%, when Rs is measured, regardless of the season. The Rnestimated obtained with the fifth model has a relatively high error. The original Brunt
model and FAO-24 model for wet climates performed more poorly than did the
other models in estimating the Rn and ET0. In overcast conditions, the original
Brunt model, the FAO-24 model for wet climates, the FAO-24 model for dry
climates, the FAO-56 model and the model of linear regression with Rs as the
predictor variable tended to overestimate Rn and ET, those estimates becoming
progressively more accurate as the cloud cover diminished.