Temperature changes in the Heihe River Basin based on high accuracy surface modelling

In the context of global warming, significant differences in temperature changes occur in different regions. Obtaining a more accurate temperature distribution is important for studying regional climate change. A high accuracy surface modeling (HASM) was introduced using the air temperature output from the weather research and forecasting (WRF) model as the driving field, and observation data from meteorological stations as the accuracy control conditions to obtain 30 years of high accuracy temperature fields in the Heihe River Basin. Verified by ground-based observations, the WRF model has a limited ability to simulate temperature conditions and performs worse in low altitude areas and in winter. The fusion results of the HASM decrease the mean absolute error from 2.21 to 0.9 degrees C and decrease the root mean square error from 2.58 to 1.16 degrees C, thus significantly increasing the accuracy compared with the original simulation from the WRF model and resulting in a more reasonable temperature analysis. The Mann-Kendall test shows clear warming in the Heihe River Basin, but there is a mutation year of 1998 in winter and annual average. The space-time distributions of and variations in the temperature from the HASM results are also analysed by dividing the study period into three subperiods. The analysis shows consistent warming except for the winter 2001-2010 period compared with the winter 1991-2000 period, in which the northeast area of the lower reaches of the Heihe River Basin was cool. Temperature changes in the Heihe River Basin also show that the sparsely populated Gobi Desert in the north and the southern high mountains show extreme sensitivity to climate change.