Browsing by Subject "Test for normality"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item Local effects of global warming(2007-12) Lu, Jun; Martin, Clyde F.Currently, global warming is a focus of attention. In order to scientifically evaluate evidence about global warming, and prove the existence of global warming, this research analyzes almost 100 years’ daily temperature records in a local weather station (Period: 1/1/1914-12/31/2006, Station: Lubbock International Airport, TX 79403, CoopID: 415411). Since mean and variance are very important in evaluating that the temperature records are inconsistent with an unchanging climate, this research create effective approaches to test the inequality of the variance and the mean of daily temperature in each year, and find their trend over the time. First, it provides the approach to test the normality of daily temperature because a lot of statistical tests are sensitive to normality assumption. Then, based on the results of invalid normality, it gives the detail procedures to test the equality of variances and means using some statistical tests which are robust to departure from normality. In testing the equality of variances, Brown and Forsythe’s test is used. In testing the equality of variances, Kruskal-Wallis nonparametric test and Z-test are used. In addition, “Loess” Smoother technique is used to find the trend of the variance and the mean of daily temperature over time. The results show that the equality of variances and means are both invalid, and their trends are increasing over time. These results suggest global warming really exists in the local place, and they also provides evidences that support global warming Furthermore, the probability of maximum temperature is found to be a Weibull distribution by applying Extreme Value Theory. Based on fitting data to the distribution, some useful predictions are made, such as estimation of 100-year return level which predicts what the annual maximum temperature will be on average every 100 years. Overall, this research scientifically estimates the local effect of global warming. It gives a good reference for the study of global warming, climate change, weather prediction, or time series data.