Spatial Homogenization Adjustment and Application of Weather Station Networks in Xinjiang, China

Abstract

In this study, we define the S0 value (buffer zone area centred on a meteorological station) and two inhomogeneity measurement parameters, the station domain area and station network density, for 89 weather stations in the Xinjiang region, and we construct the weight coefficient of the station network according to the station domain area. Applying the weight coefficient, we calculate the mean temperature, maximum temperature, and minimum temperature in January, April, July, October, and annually in the Xinjiang region from 1961 to 2021. The results show that the S0 value of 200,000 km2 is suitable for determining the weight coefficient of the station network in the Xinjiang region. The two measurement parameters can quantitatively reflect the inhomogeneity of the distribution of 89 weather stations in the Xinjiang region. The spatial distribution density of the station network is positively proportional to the station network density and inversely proportional to the station domain area and weight coefficient of the stations. The equal-weighted average is lower than the spatially homogenized revised average, which underestimates the mean temperature in the Xinjiang region, and the spatially homogenized revised average better reflects the real temperature in the Xinjiang region. The annual and monthly mean temperatures, maximum and minimum temperatures calculated by the spatially homogenized revised average, and the equal-weighted average have the same upwards trend, and the mean temperature warming trend calculated by the two methods have differences, but the differences are not significant. The annual, January, April, July, and October minimum temperature warming trends according to the spatial homogenization revised average are greater than the maximum temperature warming trend and the mean temperature warming trend, and the annual minimum temperature warming trend is 3.3 times the annual maximum warming trend and two times the annual mean temperature.