Statistical Analysis of Extremes of Temperature, Precipitation and Humidity in Swat Region

Abstract

If a specific area (station) data cannot be obtained due to natural disaster, Meteorological instrument limitations and instability in that area (terrorist attack swat 2009).We need a solid equation to calculate the missing values. Climate change is important factors now a day extreme events of precipitation and temperature are increasing for that purpose we need an equations of extreme.11 year (2006-2016) data of daily basis temperature, precipitation and humidity of three stations Saidu Sharif, Malam Jabba and Kalam is collected quality and controlled of data maintained by removing errors in data. The result is divided in two parts basic statistics of each individual station and dependence structure among different stations. Basic Statics consist of mean, standard deviation, Skewness, kurtosis, histograms and percentiles curves for better understanding of statistical behavior of data of each station. Dependence structure among different station consist Q-Q Plots to verify normality of data. Points closed to reference line means data is normally distributed and give more accurate linear regression model. Correlation and extreme correlation are developed to analyze the strength of relation among different station. Coefficient of determination and extreme coefficient of determination guide us to check the accuracy and validity of linear regression model. Linear regression equations of normal and extreme temperature, precipitation and humidity are developed from linear regression models. These linear regression equations of normal, upper extreme (90%) and lower extreme (10%) used to compute the missing data of one station from data of another station. It does not give us the actual temperature, precipitation and humidity actually happens at that day but closest to the actual value.