Sampling Biases in Datasets of Historical Mean Air Temperature over Land

Abstract:

Global mean surface air temperature have risen by 0.74 °C over the last 100 years. However, the definition of mean surface air temperature is still a subject of debate. The most defensible definition might be the integral of the continuous temperature measurements over a day (T_d0). However, for technological and historical reasons, mean temperatures (T_d1) over land have been taken to be the average of the daily maximum and minimum temperature measurements. All existing principle global temperature analyses over land are primarily based on T_d1. Here, I make a first quantitative assessment of the bias in the use of T_d1 to estimate trends of mean air temperature using hourly air temperature observations at 5600 globally distributed weather stations from the 1970s to 2013. I find that the use of T_d1 has a negligible impact on the global mean warming rate. However, the trend of T_d1 has a substantial bias at regional and local scales, with a root mean square error of over 25% at 5°×5° grids. Therefore, caution should be taken when using mean air temperature datasets based on T_d1 to examine spatial patterns of global warming.