Determination of Temperature Distribution of Rohri, Sindh using Artificial Neural Network and Regression Analysis

Abstract

As time passes, the world is facing the problem of global warming, which results in a rise in average daily temperature. Proper knowledge of temperature distribution and future prediction may help to cope with the situation in the near future. Climate forecasting has gone through various faces; in the early days’ people used to predict the behavior qualitatively. Now environmental scientists have developed a quantitative method for forest climate behavior with certain uncertainties. Empirical models have been developed based on regression analysis to estimate temperature distribution. Two models, linear and non linear, use dew point temperature and relative humidity as independent variables. In addition to regression analysis, Artificial Neural Network (ANN) has been utilized to predict the average daily temperatures of Rohri Sindh, a city in Pakistan in the Sindh province. Both empirical models and ANN estimates are in good agreement with the known values of average daily temperatures.