High Spatial Resolution Of Monthly Air Temperature And Precipitation Gridded Dataset Revision And Comparisons

High spatial resolution climate gridded dataset can be used as a basis for the study of regional climate change, climate type migration and regional hydrological model constrution. China Meteorological Administration has published climate gridded dataset coving China, but because of coarse spatial resolution (0.5°*0.5°), it may not meet the regional climate change and its consequences assessment research demands. LanZhou Univerity has created dataset (LZU0025) using ANUSPLIN software, and its spatial resolution is 0.025 degree, however, the distribution of precipitation dataset error did not agree with the law of "bigger absolute quantity, larger its error", furthermore, the border interpolation is not good because of lacking border station data.This paper added border data and used ANUSPLIN to create LZU0025 temperature dataset and its corresponding error dataset; for precipitataion dataset, according to the software guid book, its metadata needed to go through a root operation to reduce its kurtosis, then the transformation was recovered after interpolation, and then LZU0025 was revised and error dataset was outputted. The results show:(1) the distribution of temperature and precipitation is reasonable, temperature dataset distribution based on ANUSPLIN was in accordance with theory of more site less error", but exited local error, which presetented as error in site was much more than that in surrounding interpolation regions, and this local error was analyzed as a bug when the interpolation was done using ANUSPLIN; (2) The precipitation error in Western area was smaller than that in Eastern and the local error is reduced; (3) Reasonable distribution of error dataset indicates that LZU0025 is accurate. To facilitate the application in meteorology software, this paper also gives a transformation code for the modified LZU0025 (.Asc format) to.Nc format.To compare different climate dataset, given the climate dataset with spatial resolution of 0.5 degree has published by China Meteorological Administration (named CMA), which took the same interpolation method as LZU0025 but used more stations than LZU0025 and other published and widely used datasets such as Climatic Research Unit (CRU), CMA dataset was thus taken as a reference. LZU0025 and other datasets were delt with the same spatial resolution as CMA, and the handled LZU0025 was renamed as LZU. The differences among these datasets were firstly compared from time series, the climate mean state and the trend of climate change visual perspective, then Taylor diagrams were used to compare datasets from standard deviation, mean and trend indexes in national annual and seasonal, regional annual scales. The results show that:(1) the time series changes of LZU and other data indicated a consistent trend, and the trend in each time node almost kept the same. (2) The spatial distribution of standard deviation, average climate state and trends change with time of LZU is most similar to those of CMA, and showed more information of mountains compared with other datasets. (3) From statistical perspective, LZU is more similar to CMA and seasonality regional differences had smaller impact on this similar.Overall, LZU is reliable, and as revised LZU0025 has a higher spatial resolution, it can be used as basic data for the study regional climate change and regional water cycle.