CR 41:233-244 (2010)  -  DOI: https://doi.org/10.3354/cr00849

ABSTRACT: Multiple linear regression (MLR) analyses and Mann-Kendall trend analyses were conducted in order to identify the importance of geographical factors (elevation, freshwater area, urbanization, and proximity to coast) and climatic factors (temperature, wind speed, relative humidity, and solar radiation) on the long-term averages and trends of precipitation and pan evaporation. Annual climate data and monthly climate data for January, April, July, and October, from 1973 to 2006, were collected from 54 weather observatories nationwide in South Korea in consideration of seasonal effects. GIS analyses were used to determine geographical characteristics in the 54 study areas, and land cover data were analyzed to determine the degree of urbanization. According to standard MLR analyses, proximity to coast had the most significant effect on annual and monthly precipitation. Accordingly, land areas that were farther inland experienced less precipitation, except in July, when the reverse was true. Except in April and July, proximity to coast had a greater effect on annual and monthly pan evaporation, and land areas that were farther inland experienced less pan evaporation. Average elevation had a greater effect than other geographical factors on annual and monthly precipitation trends. With regard to the annual pan evaporation trend, proximity to coast was the most important factor; however, the monthly pan evaporation trend was affected differently by geographical factors, depending on the season.

KEY WORDS: Precipitation · Pan evaporation · GIS · Multiple linear regression · Trend analysis