Predictability and Analysis of Weekly Temperature using Autoregressive Time Series Models

Analysis and forecasting of time series have become a major tool for analyzing, forecasting and in other applications in various meteorological and hydrological data like temperature, rainfall, humidity, river flow, stream flow and many more others. In this paper our goal is to forecast the future trend of highest and lowest temperature of Punjab, using non-seasonal and Seasonal autoregressive moving average and Seasonal autoregressive integrated moving average models. For this we collect and analyze the past values and developed a suitable model which describes the inherent structure and characteristics of our series. We calculated the accuracy of our results on the basis of forecast error measures and hypothesis testing. The forecast performances such as mean absolute error, median absolute error, Root mean square error and normalized root mean square error are used for the temperature of Punjab. Forecast results obtained by us in the paper will be useful for various agencies for planning and development of various meteorological and hydrological decisions.