The footprint of urban heat island effect in China

6397

Urban heat island (UHI) is one major anthropogenic modification to the Earth system that transcends its physical boundary. Using MODIS data from 2003 to 2012, we showed that the UHI effect decayed exponentially toward rural areas for majority of the 32 Chinese cities. We found an obvious urban/rural temperature “cliff”, and estimated that the footprint of UHI effect (FP, including urban area) was 2.3 and 3.9 times of urban size for the day and night, respectively, with large spatiotemporal heterogeneities. We further revealed that ignoring the FP may underestimate the UHI intensity in most cases and even alter the direction of UHI estimates for few cities. Our results provide new insights to the characteristics of UHI effect and emphasize the necessity of considering city- and time-specific FP when assessing the urbanization effects on local climate.

Urbanization, one major anthropogenic modification to the Earth system, is accelerating at an unprecedented rate in recorded human history worldwide1,2. More than half of world’s population live in urban areas now, and this number is projected to be 67% by 20303. To meet the needs of soaring city dwellers, global urban land is now expanding at twice the population growth rate1 and is expected to nearly triple the area in circa 2000 by 2030 provided with current population density4.

Urbanization can pose many negative impacts on Earth’s environments that transcend far from its physical boundary5. Among these impacts, the urban heat island (UHI), referred as the phenomenon that urban areas tend to have higher temperatures than surrounding areas, has long gained considerable interest among scientists and urban planners6,7,8,9,10,11,12. UHI has many potential impacts on water and air quality, microclimatology, vegetation growth5,9,13, and human health (e.g., increase in morbidity and mortality)5,14,15,16. The UHI effects have been observed in both urban and adjacent suburban areas17,18. Thus, there is a strong impetus to systematically understand the UHI not only in its magnitude (UHI intensity, UHII) but also in its extent (also referred as the footprint of UHI effect in this study, FP).

UHII is loosely defined as the temperature difference between urban and surrounding areas. Due to the poor knowledge of the FP, estimates of UHII can vary dramatically19. For example, the UHII defined as the temperature difference between urban area and suburb10,11,20 might be lower than that between urban and rural areas17,21. Unfortunately, to date nearly all the UHI studies used the areas with static and/or subjective distances away from the urban perimeters as the unaffected references10,11,20,22,23, resulting in large bias in the UHII estimates and difficulty to compare among different studies. Better knowledge of the FP therefore can help not only for better understanding of the UHI phenomenon, but also is essential for an accurate estimate of UHII.

However, a systematic evaluation of the FP over large areas is still lacking. To our knowledge, Zhang et al.18 might be the only study that addressed this issue at a regional level. They found that the FP on average can reach up to 2.4 times of the actual urban land cover (including urban area) in the eastern North America, but they did not explore the spatiotemporal variability. The most recent global study10 also indicated that the UHI effect was mainly limited within the area twice the urban area, but it did not quantify the FP explicitly. Numerous studies documented that the UHII varied substantially across space and time10,11,17,20,21,22,23, suggesting that the FP might also have a great spatiotemporal variability. A comprehensive study on the FP is thus needed across diverse cities to understand the spatial patterns and controlling factors.

In this study, we examined the UHI effect in 32 major cities distributed in different climatic zones of China (Fig. 1) using Moderate Resolution Imaging Spectroradiometer (MODIS) Land Surface Temperature (LST) products (version 5) in conjunction with cloud-free Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) images during the period from 2003 to 2012. Unlike most previous efforts that focused on UHII10,11,21,22, we mainly concentrated on the FP of those cities. China is ideal to investigate the FP at a regional level, since it has been experiencing the rapidest urbanization in the world in recent decades2,3,4 and is characterized by complex zonal variations (from the tropical to subarctic/alpine and from rain forest to desert). Our main objectives were to (1) investigate the trends of UHI effect along urban-rural gradients (Fig. 2), (2) explore the spatiotemporal variability of the FP, and (3) examine the possible UHII bias induced by ignoring the FP by comparing urban-suburban and urban-rural LST differences for those 32 major cities across China. Twelve buffers surrounded urban areas were generated for each city (Fig. 2) and the UHI effect (△T) in urban and nearby buffer zones were defined as their LST differences relative to unaffected rural reference (see Methods). The FP was defined as the continuous extent emanating outward from urban centers to rural areas that have evident UHI effect (i.e., △T was statistically larger than zero).

Figure 1: Locations of the 32 major cities in China.

Figure 1

All of the cities are municipalities or provincial capitals except Shenzhen, which is China’s first special economic zone, and is now considered one of the fastest-growing cities in the world. The red areas on the map were included in this analysis, which excluded the altitude effects and water pixels. BJ: Beijing; CC: Changchun; CS: Changsha; CD: Chengdu; CQ: Chongqing; FZ: Fuzhou; GZ: Guangzhou; GY: Guiyang; HK: Haikou; HZ: Hangzhou; HB: Harbin; HF: Hefei; HT: Hohhot; JN: Jinan; KM: Kunming; LZ: Lanzhou; LS: Lhasa; NC: Nanchang; NJ: Nanjing; NN: Nanning; SH: Shanghai; SY: Shenyang; SZ: Shenzhen; SJZ: Shijiazhuang; TY: Taiyuan; TJ: Tianjin; UQ: Urumqi; WH: Wuhan; XA: Xi’an; XN: Xining; YC: Yinchuan; ZZ: Zhengzhou. Map was generated using ArcGIS 9.3 (www.esri.com/software/arcgis).

Full size image

Figure 2: The delineation of urban area and twelve buffer zones, an example of Beijing.

Figure 2

Landsat TM false color image acquired in Sep 3rd 2005 with a spatial resolution of 30 m × 30 m (A), daytime land surface temperature (LST) (B), and nighttime LST (C) in 2005. The black line stands for the border of urban area, the land within the border was considered as urban area, and the white lines represent the border of buffers (each of them covers half of actual urban area). Pixels that were water body or with elevation more than 50 m higher than the highest point in urban area were excluded from this analysis. Maps were generated using ArcGIS 9.3 (www.esri.com/software/arcgis).

Full size image

Figure 3Full size image
The error bars represent the standard deviation.
Figure 4: Exponential trends of the △T with distance (d) away from urban to rural areas for China’s 32 major cities averaged over 2003–2012.

Figure 4

The function takes form of △T = A × e−S×d + T0, where A indicates the maximum temperature difference, S is the decay rate, and T0 is the asymptotic value that the exponential trend can reach.

Full size image

Figure 5Full size image
The error bars represent the standard deviation.
Figure 6Full size image
The hollow black circle indicates no significant urban heat island effect for the city (NS). Maps were generated using ArcGIS 9.3 (www.esri.com/software/arcgis).
Figure 7Full size image
The boxes represent the 25% to 75% range, the whiskers indicate the minimum and maximum values, and the open pentagrams demonstrates the mean values.
Figure 8: Relationship between the areas of the FP and actual urban land cover across China’s 32 major cities.

Figure 8

The relationship was not significant during daytime in winter (panel B) because 15 of 32 cities have no UHI effect (shown as empty circle), but the correlation was significant at 0.01 level if we excluded those cities with no UHI effect.

Full size image

Figure 9: Relationship between annual mean urban-rural and urban-suburban LST differences averaged over 2003–2012 across China’s 32 major cities.

Figure 9Full size image