Decreasing Representativeness Scale of Nighttime Surface Temperature of Meteorological Stations in China From 2001 to 2021

Abstract The representativeness scale plays a key role in protection and relocation of meteorological station, and impacts applicability of station observation data. Clarifying the representativeness scale of meteorological stations promotes further research on global warming. The study improved the method used to obtain representativeness scale through spatial heterogeneity. The root mean square error instead of standard deviation was used to derive spatial heterogeneity, which represents the degree of dispersion of all pixels within the region to station point observations. The new method provided a more logical representativeness scale with a temporal series to investigate its variability and trend. The satellite land surface temperature retrievals at 1‐km resolution was used to calculate spatial heterogeneity in the commonly used grid scale (∼1°), and thus, representativeness scale can be calculated with a certain threshold (0.5°C and 1.0°C). This study focuses on nighttime for the similarity between surface temperature and air temperature because both are controlled by surface longwave radiative cooling during the nighttime. The national mean representativeness scale is ∼7 km (0.5°C threshold) and ∼16 km (1°C threshold). Meteorological stations in the western China showed higher spatial heterogeneity than eastern China. From 2001 to 2021, the representativeness scale of meteorological stations in China showed decreasing trends. The inhomogeneous warming caused by urbanization around station increased spatial heterogeneity and decreased representativeness. The conclusions derived here is only suitable for nighttime and the representativeness scale of air temperature was expected much higher during daytime due to turbulent exchange.