Determination of urban and rural monsoonal evapotranspiration by neurogenetic models

Abstract

Evaporation measurement is widely used to estimate free water-surface evaporation and is of crucial consideration in water resource development project. Evaporation is influenced by air temperature, relative humidity, wind speed, sunshine, etc. In this chapter, an attempt has been made to study the effect of the above-noted factors on reference evapotranspiration. In the present study, a Clusterized Artificial Neural Network (CANN) model was developed to estimate daily mean evapotranspiration from measured meteorological data of a tropical metro city and a rural area. The CANN model was compared with Time Series Model (TSM), Least Square Estimation Model (LSEM), and Mayer's Method (MM) to validate the estimation. Evapotranspiration estimated by CANN model was found to yield values closest to observe ones and according to the estimation, for extreme values of the input parameters there is a difference between the outputs received for the considered two cities where the main cause for the difference was identified as rainfall. © 2010 Springer Science+Business Media B.V.