Analysis of Spatio-Temporal Change in Domestic Meteorological Data

Spatio-temporal data have the spatial and temporal characteristics, making it difficult to interpret accurately using conventional statistical analysis alone. Due to spatial autocorrelation, neighboring regions tend to exhibit similar properties, and the data also change over time, making it challenging to predict future variations based only on the spatial analysis. Therefore, it is essential to apply spatio-temporal statistical analysis methods that consider both spatial dependence and temporal changes, enabling more precise pattern analysis and forecasting. In this study, we analyze the surface average temperature data for July, obtained from the Korea Meteorological Administration's Open Data Portal (https://data.kma.go.kr/cmmn/main.do), covering the period from 1981 to 2022 and collected from 48 observation stations nationwide. July is the hottest month in the Northern Hemisphere, serving as a crucial indicator for assessing the impact of global warming. It is also a key factor in forecasting energy demand and plays a significant role in predicting health indicators such as heatwaves and variations in agricultural productivity. Since the acquired dataset is clearly classified as spatio-temporal data, we employ spatio-temporal statistical methods for analysis. Specifically, we conduct exploratory data analysis by visualizing the spatio-temporal data, the empirical covariance analysis and predict the average surface temperature for July at specific locations and time points using the inverse distance weighting method. Additionally, we fit a spatio-temporal regression model using spatial basis functions and perform diagnostic checks on the model.