Abstract

Using the reconstructed continuous and homogenized surface air temperature (SAT) series for 16 cities across eastern China (where the greatest industrial developments in China have taken place) back to the nineteenth century, the authors examine linear trends of SAT. The regional-mean SAT over eastern China shows a warming trend of 1.52°C (100 yr)−1 during 1909–2010. It mainly occurred in the past 4 decades and this agrees well with the variability in another SAT series developed from a much denser station network (over 400 sites) across this part of China since 1951. This study collects population data for 245 sites (from these 400+ locations) and split these into five equally sized groups based on population size. Comparison of these five groups across different durations from 30 to 60 yr in length indicates that differences in population only account for between 9% and 24% of the warming since 1951. To show that a larger urbanization impact is very unlikely, the study additionally determines how much can be explained by some large-scale climate indices. Anomalies of large-scale climate indices such as the tropical Indian Ocean SST and the Siberian atmospheric circulation systems account for at least 80% of the total warming trends.

1. Introduction

As long-term and homogenized land surface air temperature (SAT) data are essential for assessing global and regional climate changes in the past century, the establishment of long-term SAT series has attracted extensive attention and achieved great progress (Jones et al. 1999). In recent decades, many datasets of SAT have been constructed and used to study long-term changes (Peterson and Vose 1997; Jones and Moberg 2003; Lawrimore et al. 2011; Jones et al. 2012).

To obtain continuous series of Chinese SAT during the past 100 yr, some authors infilled the series and performed quality assessments of various kinds utilizing instrumental temperature and some documentary data in the early years (e.g., Tang and Lin 1992; Lin et al. 1995; Wang et al. 1998; Yan et al. 2001; Tang and Ren 2005; Tang et al. 2009; Li et al. 2010a). However, these reconstructions indicate a wide divergence in linear trends of nationally averaged SAT [from 0.30° to 1.11°C (100 yr)−1] during the last century (Tang et al. 2009; Li et al. 2010a). This spread likely results from greater uncertainty in the earlier decades and potential flaws in some of the analysis methods used to develop the long-term SAT series. To reduce the uncertainty, a long-term and homogenized instrumental monthly-mean SAT series for 16 locations across central and eastern China back to the nineteenth century was recently reconstructed using quality control, interpolation, and homogeneity methods (Cao et al. 2013).

Detection and attribution of global and regional climate warming over the last century are a central issue in climate change research. Some earlier studies concluded that there was little impact of urban warming on local temperature series (e.g., Jones et al. 1990; Peterson 2003). With the rapid industrial development in China, however, there are arguments that urbanization impacts might be a factor in the wide divergence of the current constructed SAT series. For example, Ren et al. (2008) documented a larger contribution from urban warming to total annual-mean SAT change, based on the 1961–2000 period. Their contribution reaches 38% of the overall warming across north China and almost as much for the country as a whole. Li et al. (2010b), in contrast, found a less than 10% impact of urbanization on the long-term trend of SAT across China.

The aim of this study using the recently reconstructed set of homogenized monthly-mean SAT series back to the nineteenth century is to examine SAT climate changes over eastern China (where the greatest industrial developments in China have taken place) during 1909–2010. Relationships of both local urbanization (indicated by population) and large-scale climate factors such as atmospheric circulation and sea surface temperature (SST) with warming trends over eastern China are also comparatively analyzed.

2. Data

We use the monthly-mean instrumental SAT dataset at 16 stations across eastern China (Table 1) back to the nineteenth century. This dataset was recently developed by the National Meteorological Information Center (NMIC) of the China Meteorological Administration (Cao et al. 2013) (from the website http://cdc.cma.gov.cn), in which some missing records are interpolated using three methods: partial least squares regression, multivariate linear regression, and a gradient inverse distance weighting method. The relative homogeneity test (RHtest) version 3 technique (Wang and Feng 2010) as well as available metadata are applied to perform the homogeneity detection and the adjustment of the series for detected changepoints. The SAT data of over 400 stations across eastern China to the east of 110°E during 1951–2010, reconstructed and archived in NMIC (Ren et al. 2012), are also used, and since 1960 the number of station series in this region exceeds 435. To examine the effects of homogeneity on results, the homogenized SAT data over this denser station network across eastern China (G. L. Tang et al. 2013, unpublished manuscript) are used. Moreover, we also use the Twentieth Century Reanalysis version 2 (20CR) products, at a horizontal resolution of 2° × 2° latitude and longitude, for 1871–2010 (Compo et al. 2011); the global monthly SST from the Hadley Centre Sea Ice and Sea Surface Temperature (HadISST), at a resolution of 1° × 1°, for 1870–2011 (Rayner et al. 2003); and the sea level pressure (SLP) analysis of the Hadley Centre [Second Hadley Centre Sea Level Pressure dataset (HadSLP2)] with a resolution of 5° × 5° for 1850–2004 (Allan and Ansell 2006). The population data, provided by the China Statistics Bureau (2007a,b) and archived in NMIC, are available from more than 300 counties or cities across eastern China during 2002–06. The statistical significance of linear correlation and composite differences is assessed by the Student’s t test.

Table 1.

Station information at 16 stations and their starting dates.

Station information at 16 stations and their starting dates.
Station information at 16 stations and their starting dates.

3. Long-term trend of SAT in eastern China and its links to large-scale atmospheric circulation and SST

a. Warming trend of SAT

Only 4 stations had data before 1900, but the number of stations with data increases rapidly to 12 stations in 1909 and for all 16 stations by 1916 (Table 1), so the time series for the 1909–2010 period are used in the present analysis, giving a length of 102 yr. When we calculate a mean value for SAT averaged over 16 stations, Beijing and Tianjin, and Hong Kong, Macao, and Guangzhou, were first averaged together because they are located close together, and then the large-scale average was calculated by a simple arithmetic average of the 13 series of SAT anomalies rather than absolute SAT. We refer to this average, the regional-mean SAT series over eastern China, as T13.

Figure 1a shows the time series of T13 during 1909–2010. It is seen from this figure that T13 generally fluctuated between −1° and 0°C before the late 1960s, with smaller values in the 1940s and relatively large values near 1960, and then showed an increasing trend. Figure 1b shows the linear trends of 40-yr running T13 during 1909–2010, in which the number 1 in the abscissa is for the trend of T13 during 1909–48, 2 is for the trend during 1910–49, …, up to 63 for the trend during 1971–2010. Generally speaking, these 40-yr linear trends are weaker in the first half of the twentieth century, showing a decrease from 1.78°C (100 yr)−1 during 1909–48 to −0.53°C (100 yr)−1 during 1919–58 and then varied near zero until 1941–80. The term T13 had a weaker warming trend of 0.7°C (100 yr)−1 during 1909–70. Afterward, the linear trend began increasing and reached peaks over 4°C (100 yr)−1 since the late 1960s, with the maximum trend of 4.89°C (100 yr)−1 during 1969–2008. It is evident that this warming mainly occurs in past 4 decades, for which the linear trend is 2.89°C (100 yr)−1 during 1951–2010, 3.38°C (100 yr)−1 during 1961–2010, and 4.24°C (100 yr)−1 during 1971–2010, which contributed to the warming trend of T13 during the entire study period, with a linear warming trend of 1.52°C (100 yr)−1 during 1909–2010.

Fig. 1.

(a) Time series of T13 (black), TEast (blue), and T245 (red) anomalies (°C) from the 1971–2000 climatology; and (b) time series of trends (°C) of running 40-yr durations of T13 during 1909–2010.

Fig. 1.

(a) Time series of T13 (black), TEast (blue), and T245 (red) anomalies (°C) from the 1971–2000 climatology; and (b) time series of trends (°C) of running 40-yr durations of T13 during 1909–2010.

Since the late 1970s, China has experienced rapid industrial growth and the eastern part of China is the area where this growth has been greatest. This rapid development has probably accelerated the process of local urbanization. Because the 16 stations used in this study are located in large cities across eastern China, one may argue that the linear trend of T13 is likely affected by this urbanization growth. Therefore, we compare the linear trend of T13 with that of a regional annual-mean SAT across eastern China using both nonhomogenized and homogenized SAT data from over 400 stations to the east of 110°E (referred to as TEast) during 1951–2010, in which the time series of TEast is calculated as the arithmetic mean of SAT anomalies (from the climatological mean of 1971–2000) as the distribution of these stations is dense and relative uniform over eastern China in this period.

The variability of the TEast anomaly (from its climatological mean) approaches that of T13 (Fig. 1a), with a correlation coefficient of 0.99 (of the annual averages) between T13 and TEast during 1951–2010. Such a high correlation coefficient indicates that T13 well represents the regional SAT variation over eastern China. Our analysis shows that the linear trend of TEast is 2.44°C (100 yr)−1 for the 1951–2010 period, 2.89°C (100 yr)−1 for the 1961–2010 period, and 3.80°C (100 yr)−1 for the 1971–2010 period. The linear trend of homogenized TEast is 2.53°, 2.94°, and 3.90°C (100 yr)−1 for 1951–2010, 1961–2010, and 1971–2010, respectively. The differences in linear trend between the nonhomogenized and homogenized TEast series are only between 1.7% and 3.7% for these three periods, which indicate a small effect of the homogeneity on the SAT linear trends over eastern China. Compared to the linear trend of SAT in large cities (indicated by T13), the homogenized series of TEast shows 8%–13% lower warming trends during these three periods. In the following section, our first assessment therefore is contributions of local urbanization to the warming trend in eastern China over the past decades.

b. A possible effect of urbanization

The effects of urbanization on SAT are associated with the size of cities, and a change in population over the decades may be used to indicate a change in the size of cities and then an effect of urbanization (Peterson and Vose 1997; Peterson and Owen 2005). Since it is difficult to obtain long-term population data over the past decades in most regions of the world, they defined city or rural stations according to population in one period rather than the change in population and then discussed a possible effect of urbanization on SAT through a comparison of SAT between city/urban and rural stations.

To analyze for a possible impact of urbanization, following previous studies (Peterson and Vose 1997; Peterson and Owen 2005; Hua et al. 2008; Ren et al. 2008), a link between population and the linear trend of SAT during the recent 60 yr is examined. The population data are available from more than 300 counties or cities across eastern China during 2002–06, and an arithmetic mean from these 5 yr is used to indicate a mean value of the population for each county or city. We select 245 stations with population data and station elevations below 500 m (Fig. 2a) (excluding mountainous locations that may be less representative of regional climate change) from the more than 400 stations to the east of 110°E since 1951. The annual-mean SAT at the 245 stations (hereafter T245) shows similar variations (Fig. 1a) to those of T13 and TEast, with a correlation coefficient of 0.99 between T13 and T245 during 1951–2010, which further demonstrates the consistency between these two series.

Fig. 2.

(a) Distributions of stations for groups 1 (purple), 2 (blue), 3 (green), 4 (yellow), and 5 (red); and (b) time series of annual-mean SAT (°C) for groups 1 to 5 during 1951–2010. For details about the construction of the time series see the text.

Fig. 2.

(a) Distributions of stations for groups 1 (purple), 2 (blue), 3 (green), 4 (yellow), and 5 (red); and (b) time series of annual-mean SAT (°C) for groups 1 to 5 during 1951–2010. For details about the construction of the time series see the text.

We evenly divide the 245 stations into five groups according to the population at each station (see Table 2), in which each group has 49 stations. According to this division, all meteorological stations are divided into the largest city stations (with the greatest population), the larger city stations (with greater population), the moderate city stations (with moderate population), the smaller city stations (with less population), and the smallest city stations (with the least population). It is seen in Fig. 2a that stations from the five groups are relatively evenly spread across eastern China. Figure 2b shows the time series of anomalies of annual-mean SAT for the five groups during 1951–2010. Generally speaking, these series are very similar to each other, showing similar interannual and interdecadal features although some greater differences are seen before 1955 and around 1965 and 2000. For 1951–2010, the linear trend of annual-mean temperature series is 2.24°, 2.08°, 1.92°, 2.36°, and 2.18°C (100 yr)−1 for groups 1, 2, 3, 4, and 5 in turn, giving a range of 0.44°C (100 yr)−1 [2.36°C (100 yr)−1 minus 1.92°C (100 yr)−1] due to the population (Table 2). This result also indicates that stations with different population potentially show an additional trend of 0.44°C (100 yr)−1. For group 5 with the largest population, the varying range associated with population accounts for 20.2% of the warming trend of this group [2.18°C (100 yr)−1]. For the 1961–2010 and 1971–2010 periods, the range of the five trends is 0.26° and 0.44°C (100 yr)−1, respectively, accounting for 10.4% and 12.1% of the overall warming trend (based on TEast). Over the recent 30 yr (1981–2010) when the industrial development has been the most rapid in eastern China, the range between the five groups is 1.01°C (100 yr)−1, accounting for 21.3% of the warming trend. Using the homogenized data, we repeat the above analyses and obtain similar results. For 1951–2010, 1961–2010, 1971–2010, and 1981–2010, the ranges of the five groups account for 23.6%, 9.2%, 9.6%, and 19.0% of the warming trend, respectively (Table 2).

Table 2.

Linear trends (°C) of nonhomogenized and homogenized (in bold) SATs for groups 1 to 5 (significant at the 99.9% level), and the range among the five groups and its percentage to the linear trend of group 5.

Linear trends (°C) of nonhomogenized and homogenized (in bold) SATs for groups 1 to 5 (significant at the 99.9% level), and the range among the five groups and its percentage to the linear trend of group 5.
Linear trends (°C) of nonhomogenized and homogenized (in bold) SATs for groups 1 to 5 (significant at the 99.9% level), and the range among the five groups and its percentage to the linear trend of group 5.

It is evident that the warming trend associated with population in this study is much smaller than that of Ren et al. (2008). Their study showed that the linear trend of SAT associated with population accounts for 38% of the local warming trend in north China. Therefore, for eastern China, the warming trend associated with local urbanization likely accounts for 9%–24% of the local warming trend. So, what might be associated with most of the warming trend (about 80%)? In the following section, we will consider relationships of the warming trend over eastern China with large-scale atmospheric circulation and tropical SST.

c. Relationship between SAT and atmospheric circulation

According to the thermodynamic equation, local air temperature changes may be caused by horizontal temperature advection (HTA). Here, we analyze the relationships between SAT, atmospheric circulation, and HTA over eastern China. Figure 3a shows the regressed annual-mean 850-hPa winds against T13 during 1909–2010. It is seen that westerly wind anomalies generally prevail over the tropical Indian Ocean and easterly wind anomalies prevail over the tropical western Pacific. They meet in Southeast Asia and the South China Sea, turning northward into the mainland of eastern China, stretching into northeast China (to north of 40°N). On the climatological map of 850-hPa winds (not shown), northerly (southerly) winds prevail over eastern China during winter (summer). Therefore, the prevalence of the southerly wind anomalies over eastern China (in Fig. 3a) implies the weakening of northerly winds during winter or the strengthening of southerly winds during summer. Figure 3b shows further the regressed annual-mean 1000-hPa geopotential height. In this figure, significant negative values appear over the Asian continent and the tropical Indian Ocean, with the central value of −5 m in the mid- and high latitudes, while positive values cover most of the western and central Pacific. On the climatological-mean map of 1000-hPa geopotential height (figures not shown), a high/low pressure system appears over Siberia during winter/summer, while a high pressure system appears in the subtropical western North Pacific during both winter and summer. Thus, the negative anomalies over Siberia indicate the weakened winter Siberian high and the strengthened summer low system, and the positive anomalies in the subtropical western North Pacific indicate the strengthened subtropical high pressure system over the western North Pacific. It is evident that lower-tropospheric anomalous southerly winds over eastern China (shown in Fig. 3a) are associated with large-scale negative (positive) anomalies of geopotential height in East Asia (the western North Pacific).

Fig. 3.

(a) Regression of the calendar year annual-mean 850-hPa wind (vector; m s−1) and 1000-hPa HTA (color shaded; K s−1) of the 20CR products against the standardized annual-mean T13 during 1909–2010; and (b) as in (a), but for 1000-hPa geopotential height (×10 m), in which shaded areas are significant at the 95% confidence level.

Fig. 3.

(a) Regression of the calendar year annual-mean 850-hPa wind (vector; m s−1) and 1000-hPa HTA (color shaded; K s−1) of the 20CR products against the standardized annual-mean T13 during 1909–2010; and (b) as in (a), but for 1000-hPa geopotential height (×10 m), in which shaded areas are significant at the 95% confidence level.

The lower-tropospheric southerly wind anomalies hinder the southward invasion of cold air masses into eastern China during winter or strengthen northward transport of warm air masses during summer, further modulating HTA in eastern China. Figure 3a also shows the regressed 1000-hPa HTA. Positive HTA anomalies generally appear over eastern China, indicating the local strengthened HTA. This result suggests that the warming trend of T13 may be related to the anomalous southerly wind and HTA across eastern China. Then, what large-scale climate anomalies are responsible for these southerly wind and HTA anomalies?

d. Relationships between SAT and large-scale climate signals

In general, most of the observed global-mean land temperature variability during the past 100 yr could result from SST variations, in which the process is mainly that the oceans heat the surface atmosphere and then influence the land SAT through global atmospheric circulation through energy transportation (Hoerling et al. 2008; Compo and Sardeshmukh 2009; Dommenget 2009). Some studies have further addressed influences of tropical Indian Ocean SST on local atmospheric circulations and remote East Asian and western Pacific climates (e.g., Annamalai et al. 2005; Yang et al. 2007; Li et al. 2008; Hu et al. 2011; Gao et al. 2013, manuscript submitted to Climate Dyn.). Here, we calculate the correlation coefficient between the annual-mean T13 and the July–June (previous July to the current June) annual-mean SST during 1909–2010 (Fig. 4). In this figure, there are significant positive correlations over the Indian Ocean, the tropical western Pacific, the high latitudes of the North Pacific, and the mid- and low latitudes of the Atlantic. The large-scale positive correlation exceeds 0.7 (at the 99.9% confidence level) over the tropical Indian Ocean. Referring to the position of the greatest positive correlation in the tropical Indian Ocean, a regional-mean SST over 5°S–10°N/40°–100°E (see the box in Fig. 4) is defined as an Indian Ocean SST index. A correlation analysis shows that the July–June annual-mean Indian Ocean SST index has a correlation coefficient of 0.72 with the calendar year annual-mean T13 during 1909–2010. To assess the robustness of this link, we slightly vary the definition of the annual mean for the Indian Ocean SST index. We experiment using the previous August to present July up to the current January to December. The results still show a high positive correlation from 0.72 to 0.74 between T13 and the tropical Indian Ocean SST, which results from the strong persistence in the Indian Ocean SST index. We note that the Indian Ocean SST index also shows a warming trend (Fig. 5), with a linear trend of 0.83°C (100 yr)−1 during 1909–2010. Therefore, we calculate their correlation after removing their linear trends. The detrended series have correlations from 0.36 to 0.39 during 1909–2010 for the different annual means of the Indian Ocean SST index, significant at the 99.9% level. Moreover, for the 1951–2010 period, there is also a correlation coefficient of 0.68 between T13 and the Indian Ocean SST index. The above analyses indicate a close link between SAT over eastern China and the tropical Indian Ocean SST and despite slight changes to the Indian Ocean “year” used, the result indicates the robustness of this link.

Fig. 4.

Correlation coefficients between the annual-mean temperature average across 16 stations and the July–June (previous July to the current June) annual-mean SST during 1909–2010, in which the box indicates the tropical Indian Ocean. The SST dataset used is from HadISST (Rayner et al. 2003), and the shaded areas are significant at the 95% confidence level.

Fig. 4.

Correlation coefficients between the annual-mean temperature average across 16 stations and the July–June (previous July to the current June) annual-mean SST during 1909–2010, in which the box indicates the tropical Indian Ocean. The SST dataset used is from HadISST (Rayner et al. 2003), and the shaded areas are significant at the 95% confidence level.

Fig. 5.

Time series of T13 (black), the Indian Ocean SST (blue), and Siberian SLP (red) anomalies from the 1971–2000 climatology, in which the dashed lines are for their linear trends. The Siberian high index is calculated from the HadSLP2 data (Allan and Ansell 2006) using the regional mean for 40°–65°N by 80°–120°E suggested by Hasanean et al. (2013). Units of the y axis are °C or hPa.

Fig. 5.

Time series of T13 (black), the Indian Ocean SST (blue), and Siberian SLP (red) anomalies from the 1971–2000 climatology, in which the dashed lines are for their linear trends. The Siberian high index is calculated from the HadSLP2 data (Allan and Ansell 2006) using the regional mean for 40°–65°N by 80°–120°E suggested by Hasanean et al. (2013). Units of the y axis are °C or hPa.

Because of this significant correlation between annual-mean T13 and the July–June annual-mean Indian Ocean SST index, we further examine a possible influence of the Indian Ocean SST on the warming trend of T13 over the last 100 yr. When the variability of the Indian Ocean SST index is removed from the time series of T13 by linear regression, T13 shows a weaker trend of 0.29°C (100 yr)−1, which only accounts for 19% of the total warming trend. This result implies that approximately 80% of the warming trend over eastern China during the last 100 yr may be associated with a larger-scale climate change of the tropical Indian Ocean SST. This link between the tropical Indian Ocean SST and warming trends over eastern China is well explained by changes of the background atmospheric circulation. When the tropical Indian Ocean SST is higher, negative anomalies of the 1000-hPa annual-mean geopotential height appear over the tropical Indian Ocean and most of the Asian continent, while positive anomalies appear over the subtropical North Pacific (Fig. 6a). Corresponding to such height anomalies, an anomalous cyclonic circulation appears over the Indian Peninsula, with westerly wind anomalies over South Asia, while an anomalous anticyclonic circulation appears over the western North Pacific, with easterly wind anomalies over the tropical western North Pacific (Fig. 6b). Meanwhile, southerly wind anomalies generally prevail over eastern China. These anomalous features in the atmospheric circulations are consistent with those in Fig. 3. Previous numerical simulations (Annamalai et al. 2005; Yang et al. 2007) have indicated that a warming anomaly in the tropical Indian Ocean forces a Matsuno–Gill pattern in the troposphere, with a low-level anomalous cyclonic circulation over South Asia and an anomalous anticyclonic circulation over the subtropical western North Pacific. Westerly wind anomalies prevail over the Indian Peninsula, easterly wind anomalies prevail over the subtropical western North Pacific, and southerly wind anomalies are evident across eastern China. These model simulations suggest that the anomalous pattern in height and wind fields shown in Fig. 6 may be forced by a warming tropical Indian Ocean. Accordingly, under the influences of the southerly wind anomalies, positive HTA anomalies at the surface appear over eastern China (Fig. 6b), which increases the local SAT.

Fig. 6.

(a) Regression of the calendar year annual-mean 1000-hPa geopotential height (×10 m) of the 20CR products against the standardized July–June (previous July to the current June) annual-mean Indian Ocean SST index during 1909–2010, in which shaded areas are significant at the 95% confidence level; and (b) as in (a), but for 850-hPa wind (vector; m s−1) and 1000-hPa HTA (color shaded; K s−1).

Fig. 6.

(a) Regression of the calendar year annual-mean 1000-hPa geopotential height (×10 m) of the 20CR products against the standardized July–June (previous July to the current June) annual-mean Indian Ocean SST index during 1909–2010, in which shaded areas are significant at the 95% confidence level; and (b) as in (a), but for 850-hPa wind (vector; m s−1) and 1000-hPa HTA (color shaded; K s−1).

We also note from Fig. 3b that significant anomalies appear over Siberia, which indicates a link between SAT over eastern China and the Siberian pressure systems. Previous studies have shown that both the winter high pressure (anticyclonic) circulation and the summer low pressure (cyclonic) circulation in Siberia exerts influences on climate anomalies in eastern China by modulating the southward invasions of East Asian cold air masses (e.g., Ding and Krishnamurti 1987; Wu and Chan 1997; Zhao and Zhou 2009; Qian et al. 2011). Here, we further examine the combined relationship of both the Indian Ocean SST and the Siberian winter high pressure system/summer low system, in which the Siberian pressure system index is defined as the regional-mean SLP over 40°–65°N, 80°–120°E (Hasanean et al. 2013). The Siberian annual-mean SLP shows a decreasing trend, particularly since the early 1970s (Fig. 5), with a linear trend of −0.45 hPa (100 yr)−1 during 1909–2004, which indicates a weakening of Siberian winter high pressure system and a strengthening of Siberian summer low pressure system on the climatological-mean maps. Figure 7 shows the regression of the annual-mean 850-hPa wind against the Siberian SLP index during 1909–2004. Corresponding to the negative SLP anomalies over Siberia is a local cyclonic circulation anomaly. Southerly wind anomalies to the east of the anomalous cyclonic circulation center mainly appear over northeastern China, with relatively weaker southerly wind anomalies over eastern China to the south of 30°N. This hinders the southward invasion of cold air masses into eastern China, instead favoring local positive HTA anomalies, especially over northeastern China. These features are similar to those associated with the SAT warming (shown in Fig. 3a), which shows that the Siberian atmospheric circulation systems are mainly related to the warming trends in northeastern China. The correlation analysis further reveals a correlation of −0.61 between the annual-mean Siberian SLP index and T13 during 1909–2004. After removing their linear trends, the correlation coefficient is −0.59 during 1909–2004, significant at the 99.9% level, which supports the robustness between the Siberian SLP and T13.

Fig. 7.

Regression of the calendar year annual-mean 850-hPa wind (vector; m s−1) and 1000-hPa HTA (color shaded; K s−1) of the 20CR products against the standardized annual Siberian SLP index with an opposite sign during 1909–2004.

Fig. 7.

Regression of the calendar year annual-mean 850-hPa wind (vector; m s−1) and 1000-hPa HTA (color shaded; K s−1) of the 20CR products against the standardized annual Siberian SLP index with an opposite sign during 1909–2004.

Moreover, there is a correlation of −0.45 between T13 and the Siberian SLP index when the influence of the tropical Indian Ocean SST index is removed. Here, the influence of the Indian Ocean SST on the Siberian SLP index is removed using linear regression between the Siberian SLP index and the tropical Indian Ocean SST index. Using the T13 time series after subtracting the influence of the tropical Indian Ocean SST, we further remove the influence of the Siberian SLP index. In this case, T13 (without the variations of both the Indian Ocean SST and the Siberian SLP) shows a warming trend of 0.20°C (100 yr)−1, and it accounts for just 15% of the total warming trend [1.36°C (100 yr)−1] during 1909–2004.

The above analyses imply that at least 80% of the warming trend in eastern China may be explained by two large-scale climate indices, lending further support to the previous conclusion that 9%–24% of the warming trend in eastern China may be related to local urbanization.

4. Summary and discussion

Using the reconstructed continuous and homogenized SAT series at 16 large city stations across eastern China, we examine linear trends of regional-mean SAT in eastern China since 1909. The SAT average over this region shows a warming trend of 1.52°C (100 yr)−1 during 1909–2010, and this warming trend mainly occurred in the past 4 decades. Moreover, the variation of SAT series at the 16 stations generally represents that from a much denser station network (TEast) across this part of China since 1951. However, compared to the linear trend of SAT in these large cities, the nonhomogenized and homogenized series of TEast with a denser station network shows 8%–16% lower warming trends during the 1951–2010 period.

A comparison between warming trends of SAT and population over eastern China reveals a small influence of population growth on local warming trends during 1951–2010. The linear trends associated with population show an additional potential component of 0.44°C (100 yr)−1 [0.52°C (100 yr)−1] for the nonhomogenized (homogenized) SAT data. This accounts for 20.2% (23.6%) of the total warming trend for the largest cities. For the more rapid period of urbanization (1981–2010), this additional warming is 1.01°C (100 yr)−1 [0.92°C (100 yr)−1], accounting for 21.3% (19.0%) of the warming trend. The present result suggests that the warming trend associated with urbanization likely accounts for 9%–24% of the local warming trend in the past 60 yr, which is similar to the result of Wang et al. (2013). Their result addressed that the urban warming in Beijing accounts for 10% (20%) of the total warming trend in local-mean (minimum) temperature since the late 1970s. All these results reveal a much lower influence of urbanization than the estimate of Ren et al. (2008) but slightly higher than the estimate for all of China by Li et al. (2010b).

Compared to local urbanization, larger-scale climate variability in the tropical Indian Ocean SST and the Siberian atmospheric circulation systems is more closely associated with this warming trend. For example, after subtracting the variability of the Indian Ocean SST from the SAT series over eastern China, the SAT shows a weaker trend of 0.29°C (100 yr)−1 that only accounts for 19% of the total warming trend. After further subtraction of the variability of the Siberian high/low atmospheric circulation system, T13 has an even weaker warming trend of 0.20°C (100 yr)−1, only accounting for 15% of the total warming trend. These results suggest that at least 80% of warming trends over eastern China during the last 100 yr may be associated with two larger-scale climate indices rather than local urbanization.

These links between SAT over eastern China, the tropical Indian Ocean SST, and the Siberian SLP are well supported by changes in the background atmospheric circulation. Corresponding to a warming Indian Ocean, there is an anomalous cyclonic circulation over South Asia and an anomalous anticyclonic circulation over the subtropical western North Pacific, which may lead to westerly and easterly anomalies over South Asia and the tropical western North Pacific, with southerly wind anomalies over eastern China. When there is an anomalous cyclonic circulation over Siberia, southerly wind anomalies to the east of the anomalous cyclonic center are mainly evident over northeastern China. All these southerly wind anomalies hinder the southward invasion of cold air masses from higher latitudes during winter or strengthen the northward transport of warm air masses from lower latitudes during summer, cause warm advection at the surface across eastern China, and then possibly contribute to the increasing trends of local SAT. The above physical processes imply a potential effect of these large-scale climate signals on the warming trend over eastern China. On the other hand, the link between SAT and the climate signals might also suggest an effect of the former on the latter (Li et al. 2010). To further understand the reasons for the warming over eastern China, therefore, simulations will be needed in the future work.

Because the reconstructed long-term SAT series in eastern China is not adequate to represent variability across all of China, it will be necessary to construct more long-term series of temperature in western China during the last century by combining limited observational data in western China with those from neighboring countries before 1950. It needs to be remembered that industrial and urban development has been slower in western China than in eastern China. For all of China, an influence of urbanization on warming trends is probably smaller than for eastern China and the urbanization impact in eastern China is an upper limit for the case of China as a whole. The western region will be addressed in future work.

Natural factors (e.g., effective solar radiation and volcanic aerosols) and human influences (e.g., greenhouse gases, aerosols, and large-scale land use changes) are responsible for SAT trends. Previous studies have shown that greenhouse gases are accelerating global warming during the recent 100 yr, while aerosols generally reduce this warming trend to some extent (Solomon et al. 2007). For the warming trend in China, a similar conclusion was also obtained from observations and simulations (e.g., Dai et al. 2001; Ding et al. 2006; Zhou and Yu 2006; Chen et al. 2009; Chen and Wu 2011; Zhang et al. 2012; Man et al. 2012; Guo et al. 2013). However, different types of aerosols can also cause different responses of SAT (Chang et al. 2009). Therefore, it is useful to analyze model simulations such as the phase 5 of the Climate Model Intercomparison Project (CMIP5) to gain further insights into the attribution of SAT trends, especially influences of aerosols. This will also be addressed in future work.

Acknowledgments

This work is supported by the Strategic Priority Research Program-Climate Change “Carbon Budget and Relevant Issues of CAS SPRP” (Grant XDA05090100), the National Natural Science Foundation of China (41221064), and the Third Tibetan Plateau Atmospheric Experiment (GYHY201406001). PDJ has been supported by the U.S. Department of Energy (Grant DE-SC0005689).

REFERENCES

REFERENCES
Allan
,
R. J.
, and
T. J.
Ansell
,
2006
:
A new globally complete monthly historical mean sea level pressure data set (HadSLP2): 1850–2004
.
J. Climate
,
19
,
5816
5842
, doi:.
Annamalai
,
H.
,
P.
Liu
, and
S.-P.
Xie
,
2005
:
Southwest Indian Ocean SST variability: Its local effect and remote influence on Asian monsoons
.
J. Climate
,
18
,
4150
4167
, doi:.
Cao
,
L. J.
,
P.
Zhao
,
Z. W.
Yan
,
P.
Jones
,
Y. N.
Zhu
,
Y.
Yu
, and
G. L.
Tang
,
2013
: Instrumental temperature series in eastern and central China back to the nineteenth century. J. Geophys. Res. Atmos.,118, 8197–8207, doi:.
Chang
,
W. Y.
,
H.
Liao
, and
H. J.
Wang
,
2009
:
Climate responses to direct radiative forcing of anthropogenic aerosols, tropospheric ozone, and long-lived greenhouse gases in eastern China over 1951-2000
.
Adv. Atmos. Sci.
,
26
,
748
762
, doi:.
Chen
,
L. X.
,
B.
Zhang
,
W. Q.
Zhu
,
X.-J.
Zhou
,
Y.-F.
Luo
,
Z.-J.
Zhou
, and
J.-H.
He
,
2009
:
Variation of atmospheric aerosol optical depth and its relationship with climate change in China east of 100°E over the last 50 years
.
Theor. Appl. Climatol.
,
96
,
191
199
, doi:.
Chen
,
X. M.
, and
J.
Wu
,
2011
:
Numerical simulation of the direct effects on climate in East Asia induced by carbonaceous aerosol
.
Procedia Environ. Sci.
,
10
,
178
184
, doi:.
China Statistics Bureau
,
2007a
: China County Statistical Yearbook (2001–2006). China Statistics Press, 3395 pp.
China Statistics Bureau
,
2007b
: China City Statistical Yearbook (2002–2006). China Statistics Press, 2741 pp.
Compo
,
G. P.
, and
P. D.
Sardeshmukh
,
2009
:
Oceanic influences on recent continental warming
.
Climate Dyn.
,
32
,
333
342
, doi:.
Compo
,
G. P.
, and Coauthors
,
2011
:
The Twentieth Century Reanalysis Project
.
Quart. J. Roy. Meteor. Soc.
,
137
,
1
28
, doi:.
Dai
,
A. G.
,
G. A.
Meehl
,
W. M.
Washington
, and
T. M. L.
Wigley
,
2001
:
Climate changes in the 21st century over the Asia-Pacific region simulated by the NCAR CSM and PCM
.
Adv. Atmos. Sci.
,
18
,
639
658
.
Ding
,
Y. H.
, and
T. N.
Krishnamurti
,
1987
:
Heat budget of the Siberian high and the winter monsoon
.
Mon. Wea. Rev.
,
115
,
2428
2449
, doi:.
Ding
,
Y. H.
, and Coauthors
,
2006
:
National assessment report of climate change (I): Climate change in China and its future trend (in Chinese)
.
Adv. Climate Change Res.
,
2
,
3
8
.
Dommenget
,
D.
,
2009
:
The ocean’s role in continental climate variability and change
.
J. Climate
,
22
,
4939
4952
, doi:.
Gao
,
L. H.
,
Z. W.
Yan
, and
X. W.
Quan
,
2014
:
Observed and SST-forced multidecadal variability in global land surface air temperature
.
Climate Dyn.
, doi:, in press.
Guo
,
L.
,
E. J.
Highwood
,
L. C.
Shaffrey
, and
A. G.
Turner
,
2013
:
The effect of regional changes in anthropogenic aerosols on rainfall of the East Asian summer monsoon
.
Atmos. Chem. Phys.
,
13
,
1521
1534
, doi:.
Hasanean
,
H. M.
,
M.
Almazroui
,
P. D.
Jones
, and
A. A.
Alamoudi
,
2013
:
Siberian high variability and its teleconnections with tropical circulations and surface air temperature over Saudi Arabia
.
Climate Dyn.
, 41, 2003–2018, doi:.
Hoerling
,
M.
,
A.
Kumar
,
J.
Eischeid
, and
B.
Jha
,
2008
:
What is causing the variability in global mean land temperature?
Geophys. Res. Lett.
,
35
, L23712, doi:.
Hu
,
K. M.
,
G.
Huang
, and
R. H.
Huang
,
2011
:
The impact of tropical Indian Ocean variability on summer surface air temperature in China
.
J. Climate
,
24
,
5365
5377
, doi:.
Hua
,
L. J.
,
Z. G.
Ma
, and
W. D.
Guo
,
2008
:
The impact of urbanization on air temperature across China
.
Theor. Appl. Climatol.
,
93
,
179
194
, doi:.
Jones
,
P. D.
, and
A.
Moberg
,
2003
:
Hemispheric and large-scale surface air temperature variations: An extensive revision and an update to 2001
.
J. Climate
,
16
,
206
223
, doi:.
Jones
,
P. D.
,
P. Ya.
Groisman
,
M.
Coughlan
,
N.
Plummer
,
W. C.
Wang
, and
T. R.
Karl
,
1990
:
Assessment of urbanization effects in time series of surface air temperature over land
.
Nature
,
347
,
169
172
, doi:.
Jones
,
P. D.
,
M.
New
,
D. E.
Parker
,
S.
Martin
, and
L. G.
Rigor
,
1999
:
Surface air temperature and its changes over the past 150 years
.
Rev. Geophys.
,
37
,
173
199
, doi:.
Jones
,
P. D.
,
D. H.
Lister
,
T. J.
Osborn
,
C.
Harpham
,
M.
Salmon
, and
C. P.
Morice
,
2012
:
Hemispheric and large-scale land surface air temperature variations: An extensive revision and an update to 2010
.
J. Geophys. Res.
,
117
, D05127, doi:.
Lawrimore
,
J. H.
,
M. J.
Menne
,
B. E.
Gleason
,
C. N.
Williams
,
D. B.
Wuertz
,
R. S.
Vose
, and
J.
Rennie
,
2011
:
An overview of the Global Historical Climatology Network monthly mean temperature data set, version 3
.
J. Geophys. Res.
,
116
, D19121, doi:.
Li
,
H. M.
,
A. G.
Dai
,
T. J.
Zhou
, and
J.
Lu
,
2010
:
Responses of East Asian summer monsoon to historical SST and atmospheric forcing during 1950–2000
.
Climate Dyn.
,
34
,
501
514
, doi:.
Li
,
Q. X.
,
W. J.
Dong
,
W.
Li
,
X. R.
Gao
,
P.
Jones
,
J.
Kennedy
, and
D.
Parker
,
2010a
:
Assessment of the uncertainties in temperature change in China during the last century
.
Chin. Sci. Bull.
,
55
, 1974–1982, doi:.
Li
,
Q. X.
,
W.
Li
,
P.
Si
,
X. R.
Gao
,
W. J.
Dong
,
P.
Jones
,
J. Y.
Huang
, and
L. J.
Cao
,
2010b
:
Assessment of surface air warming in northeast China, with emphasis on the impacts of urbanization
.
Theor. Appl. Climatol.
,
99
,
469
478
, doi:.
Li
,
S. L.
,
J.
Lu
,
G.
Huang
, and
K. M.
Hu
,
2008
:
Tropical Indian Ocean basin warming and East Asian summer monsoon: A multiple AGCM study
.
J. Climate
,
21
,
6080
6088
, doi:.
Lin
,
X. C.
,
S. Q.
Yu
, and
G. L.
Tang
,
1995
:
Series of average air temperature over China for the last 100-year period (in Chinese)
.
Sci. Atmos. Sin.
,
19
,
525
534
.
Man
,
W. M.
,
T. J.
Zhou
,
L. X.
Zhang
,
B.
Wu
, and
J.
Zhang
,
2012
:
Contributions of natural and anthropogenic forcings in twentieth-century temperature change: Results from LASG/IAP climate system model FGOALS_gl (in Chinese)
.
Chin. J. Geophys.
,
55
,
10
22
, doi:.
Peterson
,
T. C.
,
2003
:
Assessment of urban versus rural in situ surface temperature in the contiguous United States: No difference found
.
J. Climate
,
16
,
2941
2959
, doi:.
Peterson
,
T. C.
, and
R. S.
Vose
,
1997
:
An overview of the global historical climatology network temperature database
.
Bull. Amer. Meteor. Soc.
,
78
,
2837
2849
, doi:.
Peterson
,
T. C.
, and
T. W.
Owen
,
2005
:
Urban heat island assessment: Metadata are important
.
J. Climate
,
18
,
2637
2646
, doi:.
Qian
,
C.
,
Z.
Yan
,
Z.
Wu
,
C.
Fu
, and
K.
Tu
,
2011
:
Trends in temperature extremes in association with weather-intraseasonal fluctuations in eastern China
.
Adv. Atmos. Sci.
,
28
,
297
309
, doi:.
Rayner
,
N. A.
,
D. E.
Parker
,
E. B.
Horton
,
C. K.
Folland
,
L. V.
Alexander
,
D. P.
Rowell
,
E. C.
Kent
, and
A.
Kaplan
,
2003
:
Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century
.
J. Geophys. Res.
,
108
, 4407, doi:.
Ren
,
G.
,
Y.
Zhou
,
Z.
Chu
,
J.
Zhou
,
A.
Zhang
,
J.
Guo
, and
X.
Liu
,
2008
:
Urbanization effects on observed surface air temperature trends in north China
.
J. Climate
,
21
,
1333
1348
, doi:.
Ren
,
Z. H.
,
Y.
Yu
,
F. L.
Zou
, and
Y.
Xu
,
2012
:
Quality detection of surface historical basic meteorological data (in Chinese)
.
J. Appl. Meteor. Sci.
,
23
,
739
747
.
Solomon
,
S.
,
D.
Qin
,
M.
Manning
,
Z.
Chen
,
M.
Marquis
,
K.
Averyt
,
M.
Tignor
, and
H. L.
Miller
Jr.
, Eds.,
2007
: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.
Tang
,
G. L.
, and
X. C.
Lin
,
1992
:
Average air temperature series and its variations in China (in Chinese)
.
Meteor. Monogr.
,
18
,
3
6
.
Tang
,
G. L.
, and
G. Y.
Ren
,
2005
:
Reanalysis of surface air temperature change of the last 100 years over China (in Chinese)
.
Climate Environ. Res.
,
10
,
791
798
.
Tang
,
G. L.
,
Y. H.
Ding
,
S. W.
Wang
,
G. Y.
Ren
,
H. B.
Liu
, and
L.
Zhang
,
2009
:
Comparative analysis of the time series of surface air temperature over China for the last 100 years (in Chinese)
.
Adv. Climate Change Res.
,
5
,
71
78
.
Wang
,
J.
,
Z.
Yan
,
Z.
Li
,
W.
Liu
, and
Y.
Wang
,
2013
:
Impact of urbanization on changes in temperature extremes in Beijing during 1978–2008
.
Chin. Sci. Bull.
,
58
,
4679
4686
, doi:.
Wang
,
S. W.
,
J. L.
Ye
,
D. Y.
Gong
,
J. H.
Zhu
, and
T. D.
Yao
,
1998
:
Construction of mean annual temperature series for the last one hundred years in China (in Chinese)
. Quart. J. Appl. Meteor.,
9
,
392
401
.
Wang
,
X. L.
, and
Y.
Feng
,
2010
: RHtestsV3 user manual. Climate Research Division, Science and Technology Branch, Environment Canada Rep., 26 pp.
Wu
,
M. C.
, and
J. C. L.
Chan
,
1997
:
Upper-level features associated with winter monsoon surges over south China
.
Mon. Wea. Rev.
,
125
,
317
340
, doi:.
Yan
,
Z. W.
,
C.
Yang
, and
P. D.
Jones
,
2001
:
Influence of inhomogeneity on the estimation of mean and extreme temperature trends in Beijing and Shanghai
.
Adv. Atmos. Sci.
,
18
,
309
322
, doi:.
Yang
,
J. L.
,
Q. Y.
Liu
,
S. P.
Xie
,
Z. Y.
Liu
, and
L. X.
Wu
,
2007
:
Impact of the Indian Ocean SST basin mode on the Asian summer monsoon
.
Geophys. Res. Lett.
,
34
, L16604, doi:.
Zhang
,
H.
, and Coauthors
,
2012
:
Simulation of direct radiative forcing of aerosols and their effects on East Asian climate using an interactive AGCM-aerosol coupled system
.
Climate Dyn.
,
38
,
1675
1693
, doi:.
Zhao
,
P.
, and
Z. J.
Zhou
,
2009
:
An East Asian subtropical summer monsoon index and its relationship to summer rainfall in China
.
Acta Meteor. Sin.
,
23
,
18
28
.
Zhou
,
T. J.
, and
R. C.
Yu
,
2006
:
Twentieth century surface air temperature over China and the global simulated by coupled climate models
.
J. Climate
,
19
,
5843
5858
, doi:.