Rainfall trend analysis of different spatial and temporal scales has been a matter of interest due to the global climate change. Global climate change can affect long-term rainfall patterns along with the risk of increasing drought and flooding. Changes in rainfall will lead to changes in the amount of available water and will bring on a serious impact on human life. In Korea, rainfall is concentrated in the summer; therefore, it causes flood damage every year. In contrast, in spring, drought affects agricultural production. Predicting precipitation trends accurately can play an important role in a country's future economic development.
In this study, annual and rainy seasons'(June to September) trends were calculated in order to understand the trend and variability of rainfall in Daegu and Gyeongsangbuk-do. This study established a method of estimating the trend of long-term rainfall fluctuations and analyzed the trend of long-term fluctuations in rainfall in Gyeongsangbuk-do. All steps to this are accomplished by statistical methods. Most statistical data analysis depends on whether the data were sampled from the normal distribution, so it is necessary to check if it follows the normal distribution.
The normality analysis was evaluated by graphical methods and test methods. When normality is determined in this way, various statistical test methods are used to detect the long-term trend of rainfall data. This method is classified into parametric and nonparametric tests. Linear regression analysis was used as the parametric test method, and Mann-Kendall test and Spearman's Rho test were used as the nonparametric test method. In addition, in order to understand the trend of rainfall events during floods or droughts, the effect of regression factors corresponding to various quartiles was identified using quantile regression.
This study quantitatively presented the characteristics of rainfall by region through basic statistical rainfall analysis. It was confirmed that the normality of rainfall data often satisfies the normal distribution in long-term (annual and seasonal rainfall) data, and that the short-term (monthly) data often do not satisfy the normal distribution. Through the rainfall trend analysis, it was found that a few regions satisfied statistical significance. Nevertheless, it was confirmed that the rainfall trend generally decreased in June and generally increased in July, August, and September. In quantile regression analysis, most regions showed an increasing tendency in the heavy rainfall(95%). On the other hand, Andong and Bonghwa showed a decreasing trend in most of the quantiles.