Enhanced sensitivity of the urban heat island effect to summer temperatures induced by urban expansion

A suitable thermal environment is important for the economy, society and public health in urban areas. However, the understanding of the relationship between the urban heat island (UHI) effect and background temperature (T-UHI) is very limited. In this study, the UHI effect induced by the urbanization of the megacity Beijing was investigated using the weather research and forecasting model. Urban expansion and heatwaves both considerably enhanced the UHI effect over urban areas in summer. The strengthened UHI effect during the heatwave period can be clearly explained by the positive sensitivity of T-UHI. The urban expansion increased the sensitivity of T-UHI from 0.0207 °C °C−1 in 2000 to 0.0569 °C °C−1 in 2010 in the daytime and from 0.0715 °C °C−1 in 2000 to 0.0995 °C °C−1 in 2010 at nighttime, thus resulting in a much stronger UHI effect mainly by increasing the difference between the latent heat flux and sensible heat flux. This enhanced sensitivity may exacerbate the urban heat stress in the situation of further urban expansion and background climate warming. Our results suggest that the sensitivity of T-UHI is a meaningful indicator to assess the urban thermal environment change and support the designing of heat mitigation strategies in urban planning.


Introduction
Urbanization has been widely occurring around the world, characterized by the dramatic expansion of artificial surfaces (e.g. buildings and roads) and the explosive growth of urban population (Taubenböck et al 2012, Liu et al 2014, UNPD 2014. A considerable impact of urbanization on microclimate is the urban heat island (UHI) effect, which generates higher temperature in urban areas than in suburban or rural areas (Oke 1982, Grimmond 2007. The elevated temperature results in concerns about the urban thermal environmental deterioration, public health issues and economic losses in summer (Frumkin 2002, Tan et al 2010, Laaidi et al 2012, Estrada et al 2017. Therefore, there is a strong motivation to investigate the UHI effect and its dominant influential factors.
Various studies have focused on the changes of the UHI intensity due to the local urbanization (Zhou et al  (Cao et al 2018). These phenomena are linked to the effects of urban structures and heavy usage of building materials (e.g. concretes and asphalts) on the land surface radiation budget and energy balance, which reduce the latent heat flux and increase the sensible heat flux (Oke 1982, Arnfield 2003. The anthropogenic heat emission as an important source of urban heat also shows a positive relationship with the UHI effect (Shahmohamadi et al 2011). In contrast, green lands, water bodies (Zhao et al 2017, Sun et al 2018), and green roofs (Coutts et al 2013, Santamouris 2014 in urban areas show potential as heat sinks to substantially decrease the UHI effect by increasing the surface evaporative cooling. Several studies, on the other hand, investigated the impacts of background climate and weather, including heatwaves  (2018b) demonstrated that there exists synergistic effect between the UHI effect and heatwaves in temperate climate cities. The difference between urban and rural areas in evaporation and anthropogenic heat release contributes to the enhanced UHI effect during heatwaves compared to normal days. Such a phenomenon is found in a wide range of cities (Li et al 2015, Schatz and Kucharik 2015, Ramamurthy and Bou-Zeid 2017, Tewari et al 2019. Considering that the heat risks in urban areas are closely relevant to background temperature and the UHI intensity (Gabriel and Endlicher 2011), it is necessary to investigate the relationship between temperature and the UHI effect (T-UHI). However, few studies have explored the variation of the UHI effect in response to increases in background temperature on fine spatial scales. One such piece of research was conducted by Scott et al (2018), who found that the T-UHI relationship varied in different cities which may be positive or negative. Moreover, there is still lack of research on the influence of urban expansion on the T-UHI relationship, which is a major concern associated with the UHI effect issue.
In this study, we aim to explore the way in which the T-UHI relationship results in the UHI effect difference between heatwaves and normal days and how the urban expansion influences this relationship in the context of Beijing, a highly urbanized and densely populated megacity. The weather research and forecasting (WRF) model was employed to investigate the characteristics of the T-UHI relationship and its responses to urban expansion. The WRF model is a mesoscale numerical weather prediction system and offers full options for representing atmospheric processes (www.mmm.ucar.edu/weather-research-andforecasting-model). Because of its nesting capability and ability to be coupled with the urban canopy models, the WRF has been extensively used to simulate the urban micrometeorology with high resolution. Its capacity to capture the temporal-spatial variations of regional climate caused by urbanization have been shown in many studies across a range of cities and climate conditions (Georgescu et al 2014, Cao et al 2016, Wang et al 2016, Krayenhoff et al 2018. The remainder of this paper is organized as follows: section 2 demonstrates the simulations and statistical methods with details. Next, the results and discussions are presented in section 3. Finally, the conclusions and implications are given in section 4.

Model description and experimental design
We used the WRF-ARW version 3.6.1 (Skamarock et al 2008) coupled with a single-layer urban canopy model (Chen and Dudhia 2001, Kusaka et al 2001 to simulate the regional climate features in Beijing. In this study, three nested domains (figure 1(a)) were configured with spatial resolutions of 1, 3 and 9 km to cover the Beijing metropolitan region, the Beijing-Tianjin-Hebei area, and northern China, respectively and to capture the local major topographic features and continental monsoon climate of the temperate zone. Three sets of simulations (see figure 2) were designed to include the hottest months of June-August in two individual years (2000 and 2010), when two of the hottest summers occurred over the past six decades according to observed meteorological records (Sun et al 2014, Cui et al 2017. All experiments' simulation periods started at 00:00 UTC 21 May and ended at 00:00 UTC 01 September to cover the entire summer and the heatwave periods. The initial ten days (21-30 May) were used as the spin-up time to eliminate the bias of the initial conditions, and the following days were analyzed, except for days when intense rainfall events occurred (total rainfall>10 mm) due to their uncertain influence on the UHI effect (Yang et al 2017). The time interval of the model outputs is one hour, and the vertical coordinate level is 50, beginning from the surface to 50 hPa pressure. In our simulation, we defined the ERA-interim dataset (Dee et al 2011) with 0.75°×0.75°spatial resolution and 6 h time interval as the initial boundary conditions of the WRF model as it has a low simulated bias in terms of the temperature pattern in the Beijing area (Yang et al 2016). The main physical parameterization schemes used in this study are summarized in table S1. The sea surface temperature from the ERA-interim reanalysis is also updated every 6 h in the simulation.
Three experiments (named 'Urban_2000', 'Urban_2010', and 'NoUrban', figure 2) were conducted to explore the responses of the UHI effect to urban expansion and heatwaves. Land cover datasets with 30 m resolution for 2000 and 2010 derived from the National Land Cover/Use Dataset of China (NLCD-China, Zhang et al 2014), were aggregated to 1 km spatial resolution (WRF inner scale) based on the dominant land cover approach, and finally were transformed to the US Geological Survey (USGS) 24category system and three-category urban land cover type (figures 1(b) and (c)). The 'Urban_2000' experiment represented the urbanization condition in 2000 and the 'Urban_2010' in 2010. In the 'NoUrban' experiment, all urban areas were replaced by nearby dominant land cover category (cropland in this study). All three sets were performed with both the 2000 and 2010 meteorological forcing to exclude the influence of annual climate variation. Thus, the differences between 'NoUrban' and the other two experiments can be used to present the UHI effect, whose intensity is typically defined as the two-meter air temperature difference (ΔT a ) or the surface temperature difference (ΔT s ).
To evaluate the simulations, the results of the 'Urban_2000' and 'Urban_2010' experiments with the corresponding boundary conditions were compared with the observed data acquired from meteorological stations and satellite remote sensing data (figures S1 and S2 is available online at stacks.iop.org/ERL/14/ 094005/mmedia). The simulated air temperature generated by the WRF model agreed well with the observations. And the surface temperature spatial pattern based on simulations reasonably captured the temperature distribution with high temperatures in the urban areas and the north plain areas and low temperatures in the mountainous and hilly areas. The results confirmed that the WRF model performed well in simulating the climatic characteristics of Beijing, although the simulated land surface temperature was higher than the observed data, perhaps because the MODIS LST data usually underestimated the true value (Hu et al 2014).

Identification and attribution of the UHI effect
The definition of a heatwave provided by Meehl and Tebaldi 2004 was used to identify the heatwave events during a certain period. It relies on two thresholds related to the daily maximum temperature (T max ) of the period: the 97.5st percentile of T max and the 81st percentile of T max . A consecutive air temperature record (from 19600101 to 20101231) of the Beijing WMO station (code: 54511, location: 40.07°N, 116.59°E) was used to identify those two thresholds, which were 34.7°C and 29.6°C, respectively. Four heatwave events in 2000 (total 30 d: 0612-0621, 0630-0702, 0711-0714, 0720-0801) and three heatwave events in 2010 (total 20 d: 0702-0706, 0720-0731, 0814-0816) were identified and referred to as 'HW'; the remaining normal days were referred to as 'ND', hereinafter.
Following the definition above, we computed the average intensity and investigated the spatial distribution of ΔT a and ΔT s during the HW and ND periods. To measure the T-UHI relationship, we calculated the linear regression slope of the UHI effect against the background temperature and defined it as the sensitivity of T-UHI, which reflects the change in the UHI intensity in response to a unit increase in background temperature. The sensitivity of ΔT a and ΔT s against air temperature (termed as T-ΔT a and T-ΔT s , respectively) were estimated using the ordinary least square (OLS) models for both daytime and nighttime over urban area. The Pearson's correlation coefficients were employed to examine the strength of linear regression analyses while the Student's t tests were used to judge the significance at the 0.05 level. To evaluate whether the linear regression model of T-UHI can explain the UHI difference between the HW and ND periods, a 'ratio' index was calculated as: where ΔT refers to the UHI effect (ΔT a or ΔT s ). The subscripts 'HW' and 'ND' represent the different periods, while 'LRM' indicates the predicted values from the linear regression models and 'WRF' represents the average values from the WRF model. where T s is the surface temperature, l es = / 1 4 T 0 s 3 is Stefan-Boltzmann law, reflecting the local background climate, ε is the surface emissivity, σ is the Stefan-Boltzmann constant, SW net is the net shortwave radiation, LW down is the downward longwave radiation, H is the sensible heat flux, LE is the latent heat flux, G is the ground heat flux and Δ refers to the difference between the urban and reference areas.

Results and discussions
3.1. Impact of urban expansion on spatial and temporal variations in UHI effect The intensive urban expansion, shown in figures 1(b) and (c), resulted in an increasing UHI effect in 'Urban_2010' compared with 'Urban_2000' (figure 3). The nighttime ΔT a was noticeably higher than the diurnal value especially over the highly urbanized region, whereas ΔT s displayed the adverse behavior. The spatial distributions of ΔT a and ΔT s were highly consistent with that of the artificial surface (a localscale factor, figure S3), but the influence area differences in figure 3 indicates that ΔT a was more significantly affected by the regional-scale factor (urban size). The downwind areas where the northern suburban areas (e.g. CP and SY) were located, exhibited a clearer urban warming signal than the southern areas (e.g. DX and TZ), as the influence of the transportation of hot urban air from south to north by the summer monsoon wind ( figure S4). Figure 4 shows the mean ΔT a for the 10 districts during the 'HW' and 'ND' periods. Consistent with the results reported above, the HW periods displayed stronger UHI effect than ND periods. The strongest UHI effect occurred in the center of Beijing (the CC area), which had the highest density of population, with the nighttime ΔT a increasing from 1.19°C (ND) to 1.82°C (HW) in 2000 and from 1.56°C (ND) to 2.25°C (HW) in 2010, and the daytime ΔT a increasing from 0.61°C (ND) to 0.68°C (HW) in 2000 and from 0.84°C (ND) to 0.99°C (HW) in 2010. The remaining areas with strong UHI effect followed by CY, HD and FT, which had the three largest population (more than 3 million) among the districts of Beijing. A weak UHI phenomenon was displayed in the suburban areas and less affected by heatwaves in the 'Urban_2000' experiment. However, the gap of the UHI effect between the urban and suburban areas was narrowed in both the 'HW' and 'ND' periods after the intensive urban expansion in 'Urban_2010'. In addition, the nighttime UHI effect was more enhanced than the daytime during heatwaves, raising greater concern of high risk of heat-related mortality during hot summer nights (Laaidi et al 2012).

The sensitivity of UHI effect to background temperature
The correlations between ΔT and background temperature showed statistically positive significance (r 0.270, p<0.01) for both 2000 and 2010 conditions (figure 5). It indicates that the UHI effect is controlled by the summer air temperature in Beijing. Furthermore, the sensitivity of T-UHI was enhanced by the urban extension. The daytime sensitivity of T-ΔT a increased from 0.0207°C°C −1 in 2000 to 0.0569°C°C −1 in 2010 over the urban areas, whereas the nighttime sensitivity of T-ΔT a increased from 0.0715°C°C −1 in 2000 to 0.0995°C°C −1 in 2010. Those increases suggest that urban expansion may raise the contribution of the UHI effect on the total warming and impose more serious heat stress on urban environments. Similarly, the sensitivity of T-ΔT s also increased from 2000 to 2010, especially in the daytime, which is consistent with the finding that the relationship between impervious surface and T s varies greatly with the background temperature (He et al 2019). In addition, the differences in the mean daytime ΔT between the HW and ND periods predicted using the LRM were 101%-108% of that simulated using the WRF model ( figure 5). However, it was only 44.3%-63.8% in terms of nighttime. Such a phenomenon may be attributed to the average wind speed difference, which was slightly higher during heatwaves than other normal days ( figure S4). Then the stronger wind speeds may strengthen the magnitude of the surface heat flux over urban areas and thus imposed positive feedbacks to the UHI effect (Li and Bou-Zeid 2013). Another reason could be the lag of accumulated heat storage in hot days, which increases the heat release and UHI effect at nighttime during subsequent days (Zhao et al 2018b). Our results indicate that the strong UHI effect in urbanized areas during heatwaves is mainly associated with the positive sensitivity of T-ΔT a or T-ΔT s . On the pixel scale (figure 6), the sensitivity of T-ΔT a in the daytime was weak (approximately 0.018-0.034°C°C −1 ) in the urban areas in 2000, but it increased to 0.037-0.075°C°C −1 in 2010 over the same region. The sensitivity of T-ΔT a at nighttime was higher than that in daytime, which increased from 0.041-0.119°C°C −1 in 2000 to 0.058-0.148°C°C −1 in 2010. The boundary of the areas with high sensitivity at nighttime generally matched the urban boundary. This may be because the heat released to the air from the urban surface was the major contributor to the increased temperature at nighttime. The CC area, which had nearly completed the urbanization process before 2000, still experienced an enhancement of the sensitivity because of the peripheral urban expansion. Most plain areas located in downwind of urban areas showed a positive but weak sensitivity, which displayed slight warming in the ' Urban_2010' (figure 3). For the surface temperature, the spatial pattern of the sensitivity of T-ΔT s was similar to that of the artificial surface ( figure 6). It indicates that urban expansion has less influence on the sensitivity of T-ΔT s than T-ΔT a . The above results underscore the substantial differences between the sensitivity of T-ΔT a and T-ΔT s in their responses to urban expansion. Many studies use satellite-based surface temperature (T s ) data in the investigation of UHI effect due to the lack of air temperature (T a ) data with high spatial resolution. However, as heat stress is more related to T a than T s in summer, these studies may underestimate the nighttime heat risk caused by urban expansion and heatwaves.

Causes of the T-UHI sensitivity changes induced by urban expansion
The calculated sensitivity of T-ΔT s showed strong agreement with the WRF modeled sensitivity ( figure 7).  We found that the relative contributions of the five biophysical factors to the sensitivity of T-ΔT s showed an apparent difference between 2000 and 2010. In both 'Urban_2000' and 'Urban_2010' experiments, the positive contribution of ΔLE was the dominant reason for the overall positive sensitivity of T-ΔT s although the effect was offset by the negative contribution of ΔH ( figure 7 and S5). This finding indicates that the difference of the evaporative fraction (=LE/(LE+H)) in response to increased background temperature between urban and cropland areas controls the T-UHI variance. In cropland, the land surface is mainly covered by plants and is usually water-sufficient with considerable irrigation, enabling it to have enough ability to increase the evapotranspiration in an environment with high temperature. Therefore, the energy balance over the cropland surface tends to enhance the LE rather than the H to alleviate the high heat stress. In contrast, the urban areas with a high fraction of artificial surface are more likely to decrease the evaporative fraction (Bateni and Entekhabi 2012). Meanwhile, the difference in the increase of H was considerably lower than the decrease of LE with urban expansion. It resulted in increases in the differences in available surface energy (ΔH+ΔLE), contributed the sensitivity of 0.038°C°C −1 in 2000 to 0.088°C°C −1 in 2010. This change made the largest contribution (123.58%) to the increase in T-ΔT s sensitivity.
The ΔSW net contributed 0.092°C°C −1 in 2000 and only 0.051°C°C −1 in 2010 to the sensitivity of T-ΔT s (figure 7), but the ΔSW down showed an insignificant correlation with T a (r0.116, p0.17, figure S6). Because the ΔSW net is highly sensitive to ΔSW down due to the relatively low albedo of urban canopy, the result of ΔSW net led to less confidence in the change in the sensitivity of T-ΔT s . The sensitivity change contributed by the ΔLW down was small as it was neither closely related to urban expansion nor highly sensitive to small temperature change (figure S7). The last contributor was the change in ΔG which was negative in the daytime (heat storage in the surface) and positive at nighttime (heat release from the surface). ΔG change caused a slight reduction in the sensitivity of T-ΔT s from 2000 to 2010 because of the increased heat absorption and storage during daytime (figure S5). It is worth to note that our method of calculating the contributions of the biophysical factors is still unable to reveal the differences between daytime and nighttime. However, the dominant contributors to the UHI effect are different between daytime and nighttime. Figure S5 shows that ΔLE controlled the daytime UHI effect whereas ΔG (change in stored heat) was the main contributor to the nighttime UHI effect. As a result, there was a larger increase in the daytime sensitivity of T-ΔT s than the nighttime from 2000 to 2010 ( figure 5(b)).

Conclusions and implications
We employed the WRF model to investigate the UHI effect in response to urban expansion and heatwaves during the summer in Beijing. As expected, the UHI effect described by ΔT a or ΔT s increased dramatically in ten districts due to urban expansion and heatwaves. Moreover, we found a positive relationship between the UHI effect and background temperature, which resulted in an enhanced UHI effect in heatwaves. We examined this relationship as the sensitivity of T-ΔT a and T-ΔT s , finding the spatial pattern of the sensitivity was similar to that of the artificial surface over urban and suburban areas. The sensitivity of T-ΔT a increased from 0.0207°C°C −1 in 2000 to 0.0569°C°C −1 in 2010 in the daytime and from 0.0715°C°C −1 in 2000 to 0.0995°C°C −1 in 2010 at nighttime in Beijing as a result of urban expansion. Such changes in sensitivity can clearly explain the changes in UHI effect associated with urban expansion and heatwaves. Our investigation of the biophysical factors revealed that the changes in the sensitivity of T-ΔT s were mainly due to the increase in the available energy difference (ΔH+ΔLE). With increasing temperature, the surface energy balance impeded the cooling contribution of LE but enhanced the heat storage during daytime and the heat release during nighttime in urban canopy.
The sensitivity of T-UHI is an indicator to provide the valuable information about the change of the UHI effect under different background temperature for urban planning. A positive sensitivity of T-UHI indicates a higher UHI effect in a warming environment over an urbanized region, while an enhanced sensitivity resulted from urban expansion means a larger contribution of the UHI effect to the total warming. Those extra warmings aggravate the heat stress and harm the urban ecosystem and economies (Ma et al 2015, Estrada et al 2017, Battles and Kolbe 2019. In this situation, the city governments, such as Beijing, should pay more attention to the additive UHI effect and continued background temperature warming in urban planning. The attribution analysis of biophysical factors offers insight into how mitigation strategies of the UHI effect can impact on the sensitivity of T-UHI. We found the largest contribution to the enhanced T-UHI sensitivity was from the deficit of evaporation over urban areas. Thus, the measures aiming to increase the evaporation, including cover improvement of green lands and water bodies (Ellison et al 2017) and employment of green roofs (Santamouris 2014), can depress the sensitivity. Another widely used strategy to mitigate the UHI effect is the usage of cool roofs with high albedo, which can greatly reduce the SW net (Vahmani et al 2016, Macintyre andHeaviside 2019), thus possibly producing a negative contribution to the sensitivity. A novel adaptation strategy is to use the lightweight urban materials, with the purposes to decrease the heat release from urban canopy in nighttime (Krayenhoff et al 2018). This may reduce the sensitivity of T-UHI at nighttime.
In conclusion, our findings show that the positive sensitivity of T-UHI which resulted in the obviously increased UHI effect during heatwaves was enhanced by urban expansion, exacerbating heat risks over the urban areas of Beijing. However, the sensitivity of T-UHI could be various in different climate zones. Meanwhile, the regional climate context may shift dramatically with future global warming (Fitzpatrick and Dunn 2019), which may change the local sensitivity of T-UHI for the city. To better understand the impacts of urbanization and global warming on the urban thermal environment, additional work should aim to examine the variations of the sensitivity of T-UHI across a wide range of cities with geographic and climate contexts.