Study finds that urbanization has considerable influence on the regional climate change, they even blame proximity to air-conditioning as a factor.
Press release from Science China Press (full paper follows)
Urbanization and surface warming in eastern China
This work was led by YANG XiuQun, professor of meteorology in the Institute for Climate and Global Change Research, School of Atmospheric Sciences at Nanjing University. The article entitled “Urbanization and heterogeneous surface warming in eastern China” was published in Chinese Science Bulletin, 2013, No. 12.
Urbanization, as one of the most significant processes in land use/cover change, can not only alter surface vegetation distribution, but also affect surface energy and water balance. Some previous studies indicated that urbanization has little impact on surface warming. However, recent investigations have suggested that urbanization plays an essential role in regional climate change.
China has been experiencing intensive urbanization since the 1980s. Due to close ties in social and economic aspects, single cities have expanded to form distinctive city clusters in eastern China, such as the Beijing-Tianjin-Hebei (BTH), Yangtze River Delta (YRD) and Pearl River Delta (PRD) city clusters. Combining the social needs with scientific issues, Professor YANG XiuQun and his group dedicate to explore the climatic effect of urbanization in eastern China from observation and simulation perspectives. The objective of their work is to estimate the effect of urbanization on surface air temperature (SAT) change, detect the seasonal variation of urban warming in different regions, and analyze the impact of urbanization on maximum and minimum temperatures.
With the homogeneity-adjusted SAT data at 312 stations in eastern China for 1979-2008 and the Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) nighttime light data, the spatial heterogeneities of the SAT trends on different scales are detected and the impact of urbanization in eastern China on surface warming is analyzed.
Results show that the urbanization can induce a remarkable summer warming in YRD city cluster region and a winter warming in BTH city cluster region. The YRD warming in summer primarily results from the significant increasing of maximum temperature, with an estimated urban warming rate at 0.132-0.250°C per decade, accounting for 36%-68% of the total regional warming. The BTH warming in winter is primarily due to the remarkable increasing of minimum temperature, with an estimated urban warming rate at 0.102-0.214°C per decade, accounting for 12%-24% of the total regional warming.
The study finds that urbanization has considerable influence on the regional climate change. Therefore, a more reasonable urban planning should be considered in order to mitigate regional surface warming. In addition, the climatic effect of urbanization features obvious temporal-spatial differences, which may be associated with the variation of regional climatic background and the change of anthropogenic heat release. Detection and assessment of the climatic effect of urbanization is of great significance for further understanding the relationship between urban development and climate change.
This work was supported by the National Basic Research Program of China (2010CB428504).
Wu K, Yang X Q. Urbanization and heterogeneous surface warming in eastern China. Chin Sci Bull, 58(12):1363-1373, doi: 10.1007/s11434-012-5627-8
Science China Press Co., Ltd. (SCP) is a scientific journal publishing company of the Chinese Academy of Sciences (CAS). For 50 years, SCP takes its mission to present to the world the best achievements by Chinese scientists on various fields of natural sciences researches.
Thankfully, they provide full open access to the paper. Here is the abstract:
Urbanization and heterogeneous surface warming in eastern China
Kai Wu, XiuQun Yang
With the homogeneity-adjusted surface air temperature (SAT) data at 312 stations in eastern China for 1979-2008 and the Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) nighttime light data, the spatial heterogeneities of the SAT trends on different scales are detected with a spatial filtering (i.e. moving spatial anomaly) method, and the impact of urbanization in eastern China on surface warming is analyzed. Results show that the urbanization can induce a remarkable summer warming in Yangtze River Delta (YRD) city cluster region and a winter warming in Beijing-Tianjin-Hebei (BTH) city cluster region. The YRD warming in summer primarily results from the significant increasing of maximum temperature, with an estimated urban warming rate at 0.132–0.250°C per decade, accounting for 36%–68% of the total regional warming. The BTH warming in winter is primarily due to the remarkable increasing of minimum temperature, with an estimated urban warming rate at 0.102–0.214°C per decade, accounting for 12%–24% of the total regional warming. The temporal-spatial differences of urban warming effect may be attributed to the variation of regional climatic background and the change of anthropogenic heat release.
Here is the full paper:
Some highlights: (I bolded the part about blaming air conditioning systems for increase in Tmax and the end sentence)
It is evident that although warming signal becomes more obvious with the expansion of filtering window, the spatial patterns roughly keep unchanged. In summer (Figure 3(a)), the remarkable warming primarily confines to the YRD and eastern Inner-Mongolian regions. In winter (Figure 3(b)), the significant warming mainly locates in the BTH region.
The increasing of seasonal mean temperatures in YRD and BTH regions may result from the continuing urbanization process, and that in eastern Inner-Mongolia may be due to the desertification occurred over recent decades . Temporal-spatial variations of urbanization effect in the YRD and BTH regions examined in this study are consistent with
those in Du et al.  and Ren et al. .
In comparison with the MSA for seasonal mean SAT trends in Figure 3, Figures 4 and 5 illustrate the spatial distributions of the MSA for maximum and minimum temperature
trends, respectively. It can be found that the urban warming in the YRD region in summer is primarily determined by the significant increasing of maximum temperature (Figure 4(a)), while that in the BTH region in winter is primarily due to the remarkable increasing of minimum temperature (Figure 5(b)).
The temporal-spatial differences of urbanization effect may be related to the anthropogenic heat release . Several studies have indicated that anthropogenic heat in the YRD region is mainly caused by the air conditioning refrigeration, which is more intensive during daytime than night and eventually can increase the daily maximum temperature in summer [56,57]. Conversely, anthropogenic heat in the BTH region is primarily caused by winter heating that is more intensive during night, which favors the increasing of daily minimum temperature in winter [58–60]. Additionally, winter maximum temperature also has increased in some other regions of northern and southeastern China (Figure 4(b)). The reason for such occurrences requires further investigation.
Conclusions and discussion
With homogeneity-adjusted SAT data at 312 stations in eastern China for 1979–2008, this study investigates spatial heterogeneities of the SAT trends and their association with
the urbanization. Main conclusions are as follows.
(1) The SAT is characterized by a significant large-scale
increasing trend from 1979 to 2008 in eastern China. The
annual mean warming rate is estimated to be 0.5°C per
decade. This large-scale warming features an obvious seasonal
variation. The warming is more remarkable in spring
and autumn than winter and summer. Meanwhile, the spatial
pattern of the warming is more heterogeneous in summer
(2) The difference between urban and non-urban stations
indicates that the warming rate of urbanization effect on
annual mean temperature is 0.057°C per decade, accounting
for 11.4% of total averaged warming in whole eastern China.
Overall, the urbanization effect has obvious seasonal variation,
which appears to be more significant in winter and
spring, and relatively weak in autumn and summer.
(3) A spatial filtering (i.e. the MSA) method is proposed
to detect the spatial heterogeneity of the SAT trends. The
MSA distribution exhibits a significant summer urban warming
in the YRD region and a significant winter urban
warming in the BTH region. The YRD warming in summer
primarily results from the increasing of maximum temperature
while the BTH warming in winter is due to the increasing
of minimum temperature.
(4) The urbanization in the BTH region has significant
impact on the increasing of winter minimum temperature
with an estimated urban warming rate at 0.102–0.214°C per
decade, while the urbanization in the YRD region primarily
affects the summer maximum temperature with an urban
warming rate at 0.132–0.250°C per decade. Winter maximum
temperature in the PRD region is also increased by
urbanization at a rate of 0.076–0.125°C per decade.
(5) The urban warming accounts for 12%–24%, 36%–
68% and 20%–32% of the total regional warming, respectively,
in the BTH, YRD and PRD regions. Most strikingly,
the contribution rate of urban warming in the YRD region is
the largest and explains around half of the total regional
As most meteorological stations have been migrated for several times, the homogeneity-adjusted dataset, which could mitigate the impact of subjective factors in observed records,
is taken to analyze the SAT change in eastern China in this study. Despite there are a few controversies about the method of homogeneity adjustment, the adjusted data reflects temperature series more objectively. Furthermore, Peterson  argued that there was no significant impact of urbanization in observed records once the temperature series were adjusted for homogeneities. However, this study confirmed that an urban warming effect is still remarkable in the homogeneity-adjusted data.
This is consistent with what we have learned in the surfacestations project and in the draft paper Watts et al 2012, especially about the homogenized data set.