Assessment of High-Resolution Gridded Temperature Data Improved by a Statistical Interpolation Method

Assessment of High-Resolution Gridded Temperature Data Improved by a Statistical Interpolation Method

Abstract High-resolution climate information is an essential factor not only to understand the characteristics of temporal and spatial climate distributions but also to evaluate the impacts of climate change on various sectors, such as water resources, agricultural, and economic fields. The high-resolution grid data generated by a GIS-based regression model might have a few extreme values at some grid points where complex topography or sparse observational networks exist. In this study, the Inverse Distance Weighting (IDW) method was applied to properly smooth the overly predicted values of Improved GIS-based Regression Model (IGISRM), called IDW-IGISRM grid data, at the same resolution for daily maximum temperature (TMAX) and minimum temperature (TMIN) from 2001 to 2010 over South Korea. We carried out a sensitivity test to various radius of influence and decided the most suitable radius of influence of 9 km. IDW-IGISRM was compared with IGISRM to assess the effectiveness of IDW-IGISRM with regard to spatial patterns over 243 AWS observational points and four performance measures at all AWS stations and four selected stations that showed larger errors over the calibration period. The results showed that improving the spatial pattern and the effectiveness of the IDW method were prominent more in extreme temperature occurrence season, i.e. summer for TMAX and winter for TMIN. In addition, all quantitative performance metrics were improved by IDW-IGISRM except for BIAS. Mean absolute error and root mean square error were reduced, more considerable in summer for TMAX and winter for TMIN. Larger improvements were found at the selected four stations when extreme temperatures might occur. These results indicate that IDW-IGISRM can improve the predictive performance of IGISRM for extreme values, consequently

Abstract High-resolution climate information is an essential factor not only to understand the characteristics of temporal and spatial climate distributions but also to evaluate the impacts of climate change on various sectors, such as water resources, agricultural, and economic fields. The high-resolution grid data generated by a GIS-based regression model might have a few extreme values at some grid points where complex topography or sparse observational networks exist. In this study, the Inverse Distance Weighting (IDW) method was applied to properly smooth the overly predicted values of Improved GIS-based Regression Model (IGISRM), called IDW-IGISRM grid data, at the same resolution for daily maximum temperature (TMAX) and minimum temperature (TMIN) from 2001 to 2010 over South Korea. We carried out a sensitivity test to various radius of influence and decided the most suitable radius of influence of 9 km. IDW-IGISRM was compared with IGISRM to assess the effectiveness of IDW-IGISRM with regard to spatial patterns over 243 AWS observational points and four performance measures at all AWS stations and four selected stations that showed larger errors over the calibration period. The results showed that improving the spatial pattern and the effectiveness of the IDW method were prominent more in extreme temperature occurrence season, i.e. summer for TMAX and winter for TMIN. In addition, all quantitative performance metrics were improved by IDW-IGISRM except for BIAS. Mean absolute error and root mean square error were reduced, more considerable in summer for TMAX and winter for TMIN. Larger improvements were found at the selected four stations when extreme temperatures might occur. These results indicate that IDW-IGISRM can improve the predictive performance of IGISRM for extreme values, consequently