Evaluating daily surface maximum temperature interpolation error by adding climate stations near border areas over China

Based on the observed daily surface maximum temperature data from 610 stations over mainland China and 162 sites from neighbouring countries during 1981-2010, the impacts of adding climate stations near border areas on daily surface maximum temperature interpolation errors over China were studied using the thin-plate spline method. The results showed that large improvements of interpolation accuracy in daily surface maximum temperature were achieved in border areas by adding climate information from neighbouring countries. Mean absolute error (MAE) was reduced by an average of 0.6 degrees Cyear(-1) and the frequency distribution of daily bias narrowed and became more peaked at 21 boundary stations that were withheld from the model-building process. Although pronounced variation was not found in most part of mainland China at a 10kmx10km grid resolution, the interpolated temperature transition between different thresholds became more stable and aligned with elevation gradients in western and northern border areas of China after incorporating data from the 162 foreign stations into the spatial interpolations. Therefore, caution should be taken when generating gridded data sets only based on observed sites in study areas to investigate regional climate change, particularly if border areas with complicated terrain are insufficiently covered by data networks.