Polynomial Regression Model for Rainfall

This study investigated the rainfall pattern in Caraga region based on graph and generated a polynomial functions as mathematical model. Collection of Rainfall data of PAGASA-Caraga from year 2009-2012 used as the main source of analysis. Simulation was done using Microsoft- Excel polynomial regression program of order 2 and 3. Generated expressions and square correlation coefficients were noted. The accuracy of two simulated models is tested for rainfall predictions. Data showed that Caraga rainfall patterns behaved closely to a polynomial function with R-2-value of similar to 0.81 for order 2 and 0.89 for order 3. It implies that the fitted functions were resembled 81-89% to that of the rainfall data from PAGASA. Results also revealed that the accuracy of the two models in rainfall prediction was 65%. Based on the two mathematical predictors, October 2012 will have a rainfall of about 143.4 mm (Model 1) and about 180.6 mm (Model 2). In November 2012, it will have a rainfall value of about 275.7 mm (Model 1) and about 204.41 mm (Model 2). By December 2012, a rainfall of about 275.7 mm (Model 1) and about 214.17 mm (Model 2) will possibly be observed. Results suggested the possibility of using the polynomial function of degree 2 and 3 to predict rainfall. This mathematical rainfall predictor, if optimized, may be useful in farm and water resource management.