Abstract: To acquire knowledge of temperature effect on stable isotope, the relation between various temperature related variables and isotopic composition of modern precipitation was explored, based on the high-resolution meteorological data and the isotope values from WorldClim and China's stable isotope observation sites. We used a linear model to fit the relation between the different temperature variables and the isotopic composition. Although the annual mean temperature does well to explain the annual mean isotope signal, the better correlation between the mean temperature in the hottest quarter and the annual mean isotope of 63 sites in China is found (R²=0.79). The temperature during the coldest quarter is used as an ancillary variable in simple kriging with varying local means (SKlm). In SKlm, the residual isotope values from the regression with mean temperature in the hottest quarter are kriging interpolated, which are then added to the high-resolution spatial distribution of stable isotope (δ¹⁸O) in China's precipitation. So more local isotope effects are accounted for by the spatial interpolation of the residual isotope values. With the good correlation between mean temperature in the hottest quarter and annual mean isotope values, the spatial distribution map can well present the pattern of variability of isotope in China. The low prediction error and a symmetrical distribution of the differences between the true and predicted values demonstrate the successful application of the SKlm approach. In summary, using surface temperature as a factor does improve the prediction of the China's isotope variation in precipitation compared to a combination of latitude and altitude, and also indicates the environmental background of regional climate and local geographic factor.

Key words: China, Kriging interpolation, oxygen isotope, secondary variable, spatial distribution