Main Article Content

Abstract

The monthly average temperature series of Quetta – Pakistan from 1950 – 2000 is examined. A straight line is fitted to the data and seasonal variation and trend in temperature for each month of the year were obtained. An overall model is constructed as large variations in the monthly slopes were observed. In order to describe the seasonal pattern and trend in temperature, corresponding to the different months, both sine/cosine waves and sine/cosine waves multiplied by the time were included in the model as independent variables. The lag-1 autocorrelation was found in the residual of the model and hence another model was fitted to the pre-whiten series that shows a good fit ( ) and is free from correlated residuals. Both parametric and non-parametric tests applied to each month temperature show significant trend in all months except February and March.

Keywords

Temporal and seasonal variations Pre-whiten series Autocorrelation Man-Kendall test

Article Details

Author Biography

Sohail Chand, College of Statistical and Actuarial Sciences University of the Punjab Quaid-i-Azam Campus Lahore Pakistan

Assistant Professor