Downloads & Free Reading Options - Results

The study investigates the applicability of linear regression and ANN models for estimating weekly reference evapotranspiration (ET0) at Tirupati, Nellore, Rajahmundry, Anakapalli and Rajendranagar regions of Andhra Pradesh. The climatic parameters influencing ET 0 were identified through multiple and partial correlation analysis. The sunshine, temperature, wind velocity and relative humidity mostly influenced the study area in the weekly ET0 estimation. Linear regression models in terms of the climatic parameters influencing the regions and, optimal neural network architectures considering these climatic parameters as inputs were developed. The models’ performance was evaluated with respect to ET0 estimated by FAO-56 Penman-Monteith method. The linear regression models showed a satisfactory performance in the weekly ET 0 estimation in the regions selected for the present study. The ANN (4,4,1) models, however, consistently showed a slightly improved performance over linear regression models.

The study investigates the applicability of linear regression and ANN models for estimating weekly reference evapotranspiration (ET0) at Tirupati, Nellore, Rajahmundry, Anakapalli and Rajendranagar regions of Andhra Pradesh. The climatic parameters influencing ET0 were identified through multiple and partial correlation analysis. The sunshine, temperature, wind velocity and relative humidity mostly influenced the study area in the weekly ET0 estimation. Linear regression models in terms of the climatic parameters influencing the regions and, optimal neural network architectures considering these climatic parameters as inputs were developed. The models’ performance was evaluated with respect to ET0 estimated by FAO-56 Penman-Monteith method. The linear regression models showed a satisfactory performance in the weekly ET0 estimation in the regions selected for the present study. The ANN (4,4,1) models, however, consistently showed a slightly improved performance over linear regression models.