Urbanization is one of the key factors driving urban temperatures to increase. As cities and urban areas expand and regenerate over time, they experience significant land-use land cover (LULC) change due to urban development and densification. Urbanization and LULC changes influence the landscape pattern of urban areas often resulting in a decrease in aggregation and connectivity of green spaces, along with an increase in built-up areas and their aggregation (Haas et al., 2015; Maheng et al., 2021). As a result, green spaces are becoming more fragmented and clustered into smaller areas. In contrast, built-up areas become more connected and lumped into larger continua such as commercial business districts, industrial areas, and dense residential areas, impacting urban ecosystem services (Depietri et al., 2012; Zhang et al., 2018). The spatial extent and pattern of clustered built-up areas and green spaces in urban areas influence the surface energy balance, which in turn affects urban temperature (Majumdar and Biswas, 2016; Guo et al., 2020; Kumari et al., 2019; Song et al., 2014; Wang et al., 2019). Nowadays, many cities are experiencing periods of elevated temperatures which become more intensive during the summertime leading to public health-related problems (Troeyer et al., 2020; Kasai et al., 2017; Sharma et al., 2019; Steul et al., 2018), and increasing energy usage for cooling buildings (Kolokotroni et al., 2012). The increased urban temperature can also trigges the generation of extreme rainfall events (Pathirana et al., 2014; Tsiringakis et al., 2017).

Numerous observational and modelling studies have been recently conducted to investigate the relationship between LULC change and urban temperature (Maheng et al., 2019; Rizvi et al., 2020; Sharma et al., 2019; Srikanth and Swain, 2022). Most of the studies suggest that the removal of vegetation which reduces transpiration rates of urban surface resulting in a decrease of latent heat leads to an increase of the urban temperature. Other factors associated with increasing urban temperature are the widespread use of high thermal admittance materials, such as concrete and asphalt in urban constructions, changes in urban morphology, and an increase in anthropogenic heat. An increase in the built-up and impervious surface has increased sensible heat (Ridder et al., 2015; Derkzen et al., 2015; Doan and Kusaka, 2016), particularly in areas where the vegetation cover is replaced by intensive use of grey materials, impervious surface, and high thermal admittance materials (Maheng et al., 2019; Ranagalage et al., 2017). Urban morphology can also contribute to an increase in the urban temperature, particularly where the configuration of high-rise buildings hampers wind movement (Kleerekoper et al., 2012; Rajagopalan et al., 2014). Furthermore, the increasing urban population puts physical stress on the urban environment, by increasing anthropogenic heat (AH) fluxes from fossil fuel vehicles, buildings, and human metabolism (Lazarova and Kusaka, 2018).

Many recent studies have investigated the impact of LULC change within a certain time period on the air temperature of an urban area based on mean temperature changes (Darmanto et al., 2019; Kubota et al., 2017; Sharma et al., 2019). These studies, however, do not provide insight into the extent the changes in land-use influence the urban temperature and temperature variations associated with different LULC change classes. For instance, temperature changes in green spaces within urban areas that undergo transformation into built-up areas may be higher compared to those urban areas that experience no land-use transformation, as seen in studies where trees are replaced by grassland or suburban areas transition into urban settings (Maheng et al., 2019). A temperature increase is expected in areas that were previously green spaces, primarily due to a reduction in latent heat and evapotranspiration, as well as alterations in radiative responses stemming from changes in the thermophysical properties of surface materials (Forman, 2013; Hamilton et al., 2014; Oke et al., 2017). This urban climate information is essential for urban planners, designers and local governments to foster sustainable urban development in cities without adversely affecting the comfort and well-being of the inhabitants.

In recent years, urban air temperature studies have been supported by a number of different numerical climate models (Darmanto et al., 2019; Ridder et al., 2015; Hürzeler et al., 2022; Michau et al., 2023; Zhang et al., 2017). Climate models offer several advantages for urban climate studies as they provide projections of future urban climate conditions (Oke et al., 2017). UrbClim is one of the climate models suitable for urban temperature studies (García-Díez et al., 2016; Sharma et al., 2019). This model is a boundary climate model designed for urban temperature simulation in the spatial resolution range of hundred meters. It has been developed by using a land surface scheme that is based upon a simplification of urban physics coupled with a 3-D atmospheric boundary layer module (Ridder et al., 2015). The land surface scheme of UrbClim follows energy exchange between soil-vegetation-atmosphere which takes into account urban surface physics. The urban surface in UrbClim is represented as a rough impermeable slab that has values for the albedo, emissivity, thermal conductivity, and volumetric capacity. UrbClim has a finer spatial resolution compared to that of the WRF model, which is the current state-of-the-art Mesoscale model. This model can simulate local climate conditions at a spatial resolution of 1 km. WRF can simulate both urban temperature and urban rainfall (Pathirana et al., 2014), while UrbClim is limited to urban temperature. The ability of UrbClim to simulate urban temperature has been demonstrated in some previous UHI studies, for example, in the studies comprising the four European cities Toulouse, Ghent, Antwerp, and Bilbao (Ridder et al., 2015), Athens (Kourtidis et al., 2015), and Barcelona (García-Díez et al., 2016), as well as Asian cities such as Colombo (Maheng et al., 2019) and New Delhi (Sharma et al., 2019).

An increase in urban temperature due to LULC change is mainly influenced by a decrease in the green spaces and an increase of the built-up areas. The land-use transformation from natural green spaces to built-up areas can change the surface energy balance in an urban area, which decreases latent heat and increases sensible heat (Rios and Ramamurthy, 2022; Wiegels et al., 2021). Important mechanisms of green spaces to control urban temperature are evapotranspiration, and tree shading (Oke et al., 2017). In the evapotranspiration process, green spaces absorb solar radiation which is converted into latent heat which in turn converts water from liquid to water vapor. Latent heat ends up in water vapor, which does not raise the air temperature (Forman, 2013). On the other hand, built-up areas are likely to increase the air temperature. Built-up areas are generally dominated by man-made impervious materials which increase solar radiation absorption. The absorbed energy from solar radiation is converted into sensible heat that is slowly released into the atmosphere causing an increase in air temperature (Oke, 1987). This is the reason why an area becomes warmer after being transformed from a green into a built-up area. The aggregated effect of a gradual and dispersed transformation of green spaces into built-up areas across a city may cause an increase in UHI intensities. Conversely, the LULC change from built-up areas to green spaces can reduce the upward flux of sensible heat and increase latent heat (Roth et al., 2017). Green city policies aiming to increase green and water surface areas in a particular area of a city can influence the air temperature not only at a local scale but also in a larger area (Wiegels et al., 2021).

The present study investigates the impact of urbanization on the spatiotemporal changes of urban air temperature by looking at the temperature anomaly across LULC change classes. LULC change is categorized into four main classes: no built-up changes (BB), no green space changes (GG), built-up to green space (BG), and green space to built-up (GB). The impact of LULC changes on urban temperature is analyzed using land-use data from 1995 to 2014, sourced from the study conducted by Maheng et al. (2021) (Figure 2). Maheng et al. (2021) used the Landsat 5 TM L1TP from 24 August 1995, with a spatial resolution of 30 m for the year 1995, and Landsat 8 OLI TIRS L1TP with a spatial resolution of 30 m from 13 September 2014, for the year 2014. The data shows that there is a significant LULC change in Jakarta between 1995 and 2014 in particular with respect to the natural green spaces which have been converted into impervious built-up areas. The surface area of the urban areas increased by 44% whilst the surface area of the green spaces showed a significant decrease in particularly with respect to the tree-covered areas which showed the highest decrease of 58%. Furthermore, the urban air temperature is simulated by using the boundary climate model UrbClim. Two input data are used comprising LULC data of 1995 and 2014 with the boundary climate data from the European Centre for Medium-Range Weather Forecasting (ECMWF) (ECMWF, 2017). Finally, temperature anomalies were analysed by extracting the temperature difference at each LULC change class.

This study is divided into two parts. The first part focuses on atmospheric modeling activities, involving the simulation of urban air temperatures for the years 1995 and 2014. This time period was chosen due to the availability of Landsat data with cloud cover of less than 15%. The second part concerns the temperature anomaly analysis which aims to investigate the temperature changes across different LULC change classes.

The LULC changes in Jakarta significantly influenced the spatiotemporal distribution and changes in urban air temperature between 1995 and 2014. This study links temperature changes to LULC change classes, categorized into four main classes: no built-up changes (BB), no green space changes (GG), built-up to green spaces (BG), and green spaces to built-up (GB), each with 25 sub-classes, as listed in Table 1. A notable rise in air temperature was predominantly observed in areas where green spaces were converted into built-up areas, where the temperature anomaly distribution ranged from 0°C to about 0.5°C.

In this study, it was observed that the LULC transition from green spaces to built-up areas has a significant impact on an increase in air temperature. In areas where green spaces transformed into built-up areas such as trees to urban/sub-urban, cropland to urban/sub-urban, and grassland to urban/sub-urban, it was found that the temperature anomaly was higher than that of other classes. Moreover, a slight increase in air temperature was observed in areas where there were not LULC changes. For instance, the temperature anomaly distribution in no built-up changes (BB) class ranged from about 0.1°C to about 0.29°C. The increased temperature in the BB class indicated there was no stationarity in the air temperature in the particular class which could affect an increase in air temperature in its surrounding areas. An increase of air temperature in those LULC transformation classes could be associated with a decrease in latent heat and an increase in sensible heat due to a decrease in green spaces. Decreasing green spaces makes urban areas absorb more solar energy that results in higher temperature compared to that in non-urban areas. The increasing air temperature in the LULC transformed from green spaces to built-up in this study was also observed in other Asian cities, such as in Dhaka city of Bangladesh (Imran et al., 2021) and in Lahore in Pakistan (Arshad et al., 2022).

The spatial distribution of an increased temperature was noted in areas dominated by built-up infrastructure. Notably, in West Jakarta, significant urbanization impacts were observed between 1995 and 2014, with this urban area experiencing a 39% increase in built-up areas and a corresponding rise in mean air temperature of about 0.4°C. This observed temperature increase aligns with findings from previous studies on air temperature in Jakarta. Ramdhoni et al. (2016) discovered that the reduction in green spaces in Jakarta led to an increase in mean air temperatures, rising from 24°C and 30°C in 2001 to 27°C and 30°C in 2014. Moreover, Siswanto et al. (2016) observed an increase in the annual mean air temperature of about 1.6°C per century in the observed period from 1866 to 2012 in Jakarta, which was associated with urbanization. The increasing temperature was higher than the observed global mean temperature rise of 0.85°C in the period from 1880 to 2012 IPCC (2013). The phenomenon of increasing air temperature due to urbanization was also observed in other cities, such as Beijing (Wang et al., 2016), Berlin (Li et al., 2019), Delhi (Sharma et al., 2019), Ho Chi Minh (Doan and Kusaka, 2016), Tokyo (Matsumoto et al., 2017), and American cities of different climate zones (Wang et al., 2019). The observed trend of an increasing temperature is predicted to persist n the future in Jakarta (Darmanto et al., 2019).

Increased urban temperature in Jakarta can also be associated with increased concentrations of pollutants and greenhouse gases, which are not investigated in this study. The high rate of urbanization and LULC change in Jakarta trigger anthropogenic activities that may increase air pollutants and greenhouse gas concentrations in the urban atmosphere. This increase will contribute to an increase in the urban temperature. The relation between LULC change and air pollution have been discussed in some previous studies. For example, increased urban temperature in the city of Bengaluru in India was associated with an increase in the concentration of PM 2.5 , SO 2 , NO 2 , and NH 3 (Suthar et al., 2022). Moreover, the relation between increased urban temperature and SO 2 , CO , and NO 2 , was also identified in Seoul, South Korea (Ngarambe et al., 2021). Increasing urban temperature and the concentration of air pollutants will be greater risk of loss of health and life in the city due to an unfavorable change in the local budget as well as the inappropriate ventilation of urban areas (Shi et al., 2023). Furthermore, the local LULC change observed in urban areas is one of the main contributor to the global LULC change that has given an impact on the global warming. The impact of global LULC change on the global increasing temperature is mainly influenced by the land-use transition from secondary land to cropland and from cropland to urban land, which contributed over 70% to LULC change’s temperature effect at a rate of 0.0025°C/century (Yu and Leng, 2022).

The observed air temperature anomaly between 1995 and 2014 stresses the essential role of green spaces in reducing high air temperature in Jakarta and hence the impact of urbanization on the urban temperature characteristics. The important role of green spaces is generally associated with their ecosystem functions such as evapotranspiration, tree shading, and air movement modification (Oke et al., 2017). The influence on the composition and configuration of green spaces is acknowledged but less understood (Bao et al., 2016). In this study it is demonstrated that green spaces have the ability to reduce high temperatures based on a comparison of the mean temperature anomaly in the BG class with that of other classes. The observed lower values indicate that an increase of the temperature is mitigated by the presence of green spaces.

The present study investigates the impact of urbanization on the spatiotemporal distribution and the air temperature change in Jakarta between 1995 and 2014. It is found that the LULC change has altered the spatial distribution and increased, the air temperature. Moreover, the mean temperature anomaly in areas transformed from green spaces to built-up areas, GB, is higher than in other LULC change transformations. In the GB class, the mean values of the temperature anomaly was about 0.24°C that was the highest mean values compared to that of GG (0.20°C), BB (0.19°C), and BG (0.17°C).

The LULC change resulted in a significant impact on the variations and spatial distribution of mean air temperature in Jakarta between 1995 and 2014. The expansion of built-up areas contributed to the rise of 0.5°C in air temperature, attributed to a reduction in evapotranspiration resulting from decreased green spaces. Significant increases in air temperature were observed in areas that are transformed from green spaces into built-up areas. These transformations were mainly observed in newly built-up areas of Jakarta where the highest temperature anomaly was about 0.5°C. However, a decreased temperature was observed in areas which were transformed from built-up areas into green spaces where the temperature anomaly could lower than −0.1°C. This indicates that an increasing air temperature can be mitigated by greening solutions in which built-up areas are transformed into green spaces. The ability of green spaces to provide such a climate regulation is influenced by several factors, including the proportion and the distribution of green spaces, as shown in this study. Hence, urban planning and urban land-use management have to play an important role to safeguard that there is an appropriate quantity, composition, and distribution of green spaces in urban (re)development programs to prevent a further increase of the urban air temperature.

The method proposed in this study enables to analyse the detailed impact of urbanization on urban temperature. It offers comprehensive information on land use and land cover (LULC) changes and their implications for temperature variations, a crucial aspect in urban climate studies. The method also highlights that internal factors within the urban environment, such as LULC change, can play a significant role in urban climates alongside the impact of global climate change. Furthermore, the method can serve as an alternative to many existing approaches that offer general information about temperature changes within a specific timeframe, but do not provide detailed insights into temperature changes resulting from LULC transformation. Hence, this study has a novelty that explicitly assess the contribution of different LULC change classes on urban temperature changes which is remain under examined. At the same time, the proposed method also can be used to identify the potential main contributor of the increasing urban temperature, that is important for urban community and related stakeholders in preparing the mitigation and adaptation planning for the climate change in urban areas. Moreover, this research contributes to the existing body of knowledge on the impact of LULC change on increased urban temperature and suggest that the impact of LULC change on urban temperature should consider temperature changes at different LULC change classes.