Modelling of Daily Reference Evapotranspiration using Fuzzy Logic for Chhattisgarh Plains

Reference evapotranspiration (ET₀) is a key issue in hydrology, water resources management and irrigation scheduling. The standard Food and Agricultural Organization of the United Nations (FAO) – 56 Penman–Monteith (PM) equation requires several meteorological inputs for estimating ET₀, which are generally not available in most of the stations. This study is an attempt to estimate ET₀ using fuzzy logic models with fewer weather parameters for Chhattishgarh plains (Raipur and Bilaspur). Different fuzzy sets have been formed using the parameters, namely maximum temperature (T_max), relative humidity (RH-I) and wind speed (WS). These parameters were classified into nine classes as extremely low, very very low, very low, low, medium, moderate high, high, very high and extremely high. The most common triangular membership function has been adopted. The operations of fuzzy logic models is based on ‘IF-THEN’ rules, where ‘IF’ relates to the fuzzy premises regarding the daily data T_max, RH-I and WS for Raipur and Bilaspur region. The ‘THEN’ relates to the fuzzy response, which means daily ET₀. A total of 127 and 133 rules were formed, respectively, for Raipur and Bilaspur regions, respectively. The fuzzy logic model was trained to estimate ET₀ from daily weather data set as ‘input’ and the PM estimate as ‘output’. Daily data of 14 years (1 January 2000 to 31 December 2014) for Raipur region and 9 years (1 January 2003 to 31 December 2011) for Bilaspur region has been used for analysis. Out of 4,435 patterns for Raipur region, 3,102 have been used for training the model, whereas the remaining 1,333 has been used for testing the model. Similarly, for Bilaspur region, out of 3,041 patterns, 2,034 have been used for training the fuzzy rules, whereas the remaining 1,007 has been used for testing the model. The analysis was carried out in MATLAB-2014 software. The ET₀ estimated by fuzzy logic model was found to be comparable with the ET₀ estimated by FAO PM method. The Performance evaluation of developed models was carried out using correlation coefficient (CC), root mean square error (RMSE), mean absolute deviation (MAD) and coefficient of efficiency (CE). During testing of model, the values of CC = 90.6, RMSE = 0.73, MAD = 0.56 and CE = 82.3 for Raipur region and similarly for Bilaspur region, these values were found as CC = 89.78, RMSE = 0.54, MAD = 0.36 and CE = 80.6. Developed fuzzy logic models require lesser input variable than input required by different empirical methods. Hence, developed fuzzy model may be used to estimate ET₀ with fewer weather parameters.