CR 76:131-143 (2018)  -  DOI: https://doi.org/10.3354/cr01531

ABSTRACT: Debates over the bias in regional climate records caused by urbanization effects continue, and the diverse estimates of this bias can be partly attributed to different study methods and materials. In this study, we proposed an index (the extent of urban impact, EUI) by quantifying urbanization effects in terms of urban area and station location relative to urbanized areas. The EUI was used to represent the extent of urban impact on the meteorological stations used for observing temperature. Based on the EUI quantification, 569 stations in China were classified into 3 categories: stations with little or no urban impact (R stations), stations with low urban impact (LU stations), and stations with high urban impact (HU stations). Results show that, on average, EUIs for R, LU, and HU stations were 0.3, 2.6, and 15.3%, respectively, implying diverse urbanization influences on the studied stations. Using R stations as a baseline, the average urbanization effects on the warming trend of surface air temperature (SAT) for China during 1961-2012 for the LU and HU stations were estimated to be 0.06 and 0.09°C decade^-1, respectively. Therefore, including LU and HU stations to estimate regional SAT trends would lead to notable biases. The bias of the urbanization effect in the average SAT trend of China based on all stations was estimated to be 0.05°C decade^-1, accounting for 18% of the overall warming over the period 1961-2012. This estimation should be considered the lowest range of the bias. This study highlights the necessity of considering urbanization effect biases when assessing regional climate change.

KEY WORDS: Surface air temperature · Climate change · Urbanization effect · Station classification · China