Spatial identification and prediction of thermal hot spots based on mobile measurements in urban areas

Abstract

This study proposes a prediction method for the identification of areas inside and around cities, that—based on their morphometric and built-up conditions—are prone to develop as hot spots in the highly probable case of cities development. To achieve this objective, hot and cold spots of air temperature distribution for the six largest cities from north-eastern Romania were firstly identified using data from mobile measurements, performed during the warm season. From May to September 2022, 64 mobile measurements were made following a standardized observation plan. The measurements were carried out under calm and stable atmospheric conditions, with clear or partly cloudy skies, before sunrise and immediately after sunset in order to ensure representativeness. Global Moran’s Index was used to assess spatial autocorrelation, while Getis-Ord Gi* was used for hot and cold spot identification. The results describe accurately the built-up ratio and landform morphology conditions of the areas that are warmer (hot spots)/colder (cold spots) than their surroundings. In brief, it can be observed that the occurrence of hot spots is mostly controlled by high imperviousness density values (>45%), while cold spots are shaped mainly by local natural conditions that are favorable for the accumulation of cold air below the thermal inversion band. Secondly, based on the relationship of altitude and built-up ratio with hot spots occurrence, we developed a prediction model of the urban areas that are prone to sustain or evolve into hot spots. The analysis results are meant to serve cities stakeholders involved in the mitigation of the urban heat island effects, helping them to identify the regions that are in risk of becoming excessively warm during future summers.

1 Introduction

Urban heat island (UHI), classically defined as the phenomenon whereby urban areas are warmer than their surroundings (Oke et al., 2017), is currently monitored and assessed in various ways. Previously concentrated on large cities and megalopolises, urban climate studies recorded in the last period a surge in interest for medium-sized and even small cities (Lamb et al., 2019; de Costa Trinidad Amorim et al., 2024). This trend aligns itself with experts’ recommendations that highlight the value of understanding urban climate dynamics in rapidly developing small and medium-sized cities (Ting Kwok and Yung Ng, 2021). For these cities, urban climate analysis methodologies are focused on the assessment of spatiotemporal patterns of the UHIs, outlined mainly during the summer or warm season (Lamb et al., 2019; Ting Kwok and Yung Ng, 2021). This is achieved through diverse methods, among which the extensive use of satellite products, with focus on land surface temperature (LST).

The new manufacturing technologies empowered by the Internet of Things (IoT) offer a high potential for urban environmental monitoring (Cecilia and Peng, 2022), especially in the area of mobile measurements (Seidel et al., 2016; Kousis et al., 2022). The mobile meteorological monitoring has been focused primarily on the air temperature due to the interest for the basic UHI assessment (Sundborg, 1951), but is currently extended to environmental aspects, as air quality (Samad and Vogt, 2020), involving a variety of measurement techniques (Seidel et al., 2016). Mobile measurements are considered to overcome and complete nowadays the limitations of remote sensing products or those of sparse fixed points monitoring networks (Shi et al., 2018; Chen et al., 2023), representing a very efficient method to investigate environmental aspects in highly dense urban areas (Tsin et al., 2016; Yokoyama et al., 2018). Due to these aspects, nowadays, innovative mobile measurements methods are proposed in the field of urban climate (Kousis et al., 2021).

The main advantage of mobile measurements is given by the fact that they can provide dense spatio-temporal data that can sample different land use categories inside the cities, achieving accurate local climate characteristics (Herbel et al., 2016), a reason for which they have been extensively used to evaluate the LCZ thermal features (Stewart and Oke, 2012; Shi et al., 2018; Leconte et al., 2020; Chen et al., 2023). Recent reviews also emphasize that mobile measurements are increasingly integrated with remote sensing products, fixed observation networks, and modeling approaches, allowing a more comprehensive characterization of urban thermal environments (Milica et al., 2025). In the same time, mobile measurements are biased by the fact that they are limited to the accessible vehicle tracks (roads, bicycle lanes) and to the selected observation days (Hedquist and Brazel, 2006), while they also may fall apart from some rigorous standardization (Seidel et al., 2016). However, this type of measurements has the advantage to be made mostly at the pedestrian level (Chen et al., 2023), having a higher relevance for human body comfort and people’s daily life (Yokoyama et al., 2018), being considered therefore more relevant than fixed stations when the health risk is investigated (Tsin et al., 2016). The general high density of sampling points obtained through mobile measurements also enables the development of innovative methods for the assessment of UHI intensity (Brandsma and Wolters, 2012; Chen et al., 2023). In this line, it is nowadays considered that mobile measurements generally exacerbate the urban–rural air temperature differences when their results are compared to those from fixed stations (Hedquist and Brazel, 2006; Tsin et al., 2016). Also, due to their high spatial resolution, mobile measurements have been used to validate the heat maps produced using remote sensing products and in-situ measurements (Tsin et al., 2016), while multilinear regression models based on mobile measurements underline that morphological elements perform better than the land use in explaining the normalized cooling rate before and after the sunset (Leconte et al., 2020). This is consistent with the study of Herbel et al. (2016) in Cluj-Napoca (Romania), where local topography and mountain-valley breeze system seem to strongly influence the spatial configuration of UHI. Moreover, mobile measurements remain relatively underused in Romania, despite their potential to provide standardized and comparable assessments of urban thermal variability across multiple cities. Their integration with other approaches may therefore contribute to a more robust identification of urban hot and cold spots and support climate adaptation strategies at the local scale (Crețu et al., 2025).

However, others methods are emerging, as the implementation of mesoscale atmospheric models (Ting Kwok and Yung Ng, 2021), frequently coupled with the development of robust in-situ observation networks, or with the use of mobile measurement campaigns operating at various temporal scales (Muller et al., 2013; Li et al., 2024). Actually, high-resolution urban climate modeling has become an essential tool for investigating the spatial variability of urban thermal environments. Numerical models such as ENVI-met and PALM-4 U simulate air temperature, wind flow, radiation exchange, and human thermal comfort at neighborhood and street scales, allowing a detailed assessment of the physical processes responsible for the development of urban hot and cold spots (Bruse and Fleer, 1998; Maronga et al., 2020; Straub et al., 2025). Consequently, the integration of standardized mobile measurements with high-resolution urban climate modeling provides a comprehensive framework for identifying, explaining, and predicting urban hot and cold spots, supporting both model validation and evidence-based climate adaptation strategies (Ichim et al., 2026).

In Romania, a series of urban climate studies conducted in recent years have analyzed the typologies and effects of the UHI in medium and small size cities, aiming firstly for the identification of the basic features of UHIs (Cheval et al., 2023) or trying to understand LST spatial distribution (Crețu et al., 2025) and bioclimatic characteristics at fine city scale (Ichim and Sfîcă, 2020). The main analytical methods used in urban climate analysis employ satellite LST, from MODIS and Landsat (Cheval et al., 2023; Crețu et al., 2025), meteorological measurements using urban observation networks (Sfîcă et al., 2018; Ichim and Sfîcă, 2020; Sfîcă et al., 2023a), comparisons between the temperature of urban surfaces and temperature values at an altitude of 2 m (Sfîcă et al., 2023a; Mihăilă et al., 2024), and also exploratory standardized mobile measurements during certain specific synoptic patterns (Herbel et al., 2016; Ichim et al., 2018). More recently, satellite-based studies have also integrated Local Climate Zones (LCZ) classification and hot/cold spot analyses to improve the understanding of the spatial distribution of thermal anomalies within cities, highlighting the important role of industrial areas, densely build-up sectors, green spaces and water bodies in shaping urban thermal patterns (Crețu et al., 2025). The current work tackles firstly the assessment of hot and cold thermal spots in six small and medium-sized cities in north-eastern Romania using mobile measurements collected during relevant synoptic conditions for the UHI intensification. Using the fine spatial resolution of the collected data we analyze the distribution of the thermal spots in relation with the impervious density ratio and local morphological conditions. Secondly, based on this fine scale spatial assessment, the final and the main aim of the study is to advance a prediction model of the areas located in the cities surroundings that are prone to develop new hot spots or to intensify existing hot spots, so that the stakeholders could have a clear indication where exactly to concentrate their policies focused on alleviation/mitigation of heat hazard in the cities.

2 Study area

The six analyzed cities are the most important ones from north-eastern Romania, ranging from 80,000 (Vaslui) to mostly 400,000 (Iași) inhabitants (National Institute of Statistics of Romania, 2023) and summing up, all together, almost 1 million people within their urban agglomerations (Table 1; Supplementary Appendix 1).

The analyzed cities are characterized by a significant local landforms diversity. In brief, the complex regional landforms and significant altitudinal range define the geomorphological specificity of the cities locations (Figure 1). The detailed description presented in Supplementary Appendix 1 shows that most of the investigated cities are located in contact areas between the main physical-geographic units of the region or within the floodplains and on the terraces of major crossing rivers. Generally, they exhibit important differences in terms of topography, urban morphology and urban development patterns. Iași and Bacău are the largest and most densely urbanized cities, whereas Botoșani, Suceava, Piatra-Neamț and Vaslui are medium-sized cities with a stronger influence of local topography on air temperature patterns.

Figure 1

From a regional climate perspective, the investigated cities are located within the temperate climate zone. According to the Köppen-Geiger classification system, 5 from the 6 analyzed cities have a Cfb climate, namely a temperate humid climate with warm summers (Cheval et al., 2023), while the city of Iași has a Cfa climate meaning temperate humid climate with hot summer and long warm season (Ichim and Sfîcă, 2020; Cheval et al., 2023). These general climate characteristics include the frequent occurrence of strong heatwaves during warm seasons (Sfîcă et al., 2017) with appreciable impact on thermal comfort in urban areas (Ichim and Sfîcă, 2020; Mihăilă et al., 2024). Overall, all the cities share a common urban–rural transition structure and background climate conditions, which make them suitable for a comparative assessment of urban thermal hotspots.

3 Data and methods

3.1 Meteorological and ancillary data

In order to overcome the well-known limitation of mobile measurements given by the small number of monitoring observations (Hedquist and Brazel, 2006), we tried to give a sound seasonal image of air temperature distribution in the analyzed cities by performing meteorological measurements during specific synoptic conditions for a full warm season. These observations have been made in the analyzed cities of north-eastern Romania between May and September 2022, being carried out with Meteo Tracker (MT) mobile meteorological mini-station (Supplementary Figure S1). MT weather station¹ is designed for mobile measurements and can be considered a complex measurement tool in a miniaturized form. This tool records the following meteorological parameters during the chosen tracks: air temperature, relative humidity, atmospheric pressure and solar radiation. Geolocation parameters (latitude, longitude and altitude) are also collected during observations. MT data accuracy were tested by Cecilia and Peng (2022), being proved that it is only slightly affected by the moving speed during transect, a common issue in mobile measurements (Seidel et al., 2016). Generally, MT device has a good response in highlighting the air temperature and humidity features for different urban areas (Cecilia and Peng, 2022) and overcome the GPS positioning gaps of previous mobile measurements methods (Yokoyama et al., 2018).

The main sources of uncertainty associated with mobile measurements are related to vehicle movement, temporal variations in air temperature during the transects, and local disturbances caused by traffic conditions. These effects were minimized by performing measurements under stable synoptic conditions and by applying a temporal correction based on closed-loop transects. In addition, the use of identical measurement equipment in all cities contributed to improving data consistency and comparability.

One important advantage in using this tool is given by the possibility to replicate the observations through a standardized observation protocol. All transects were performed along predefined routes crossing the same urban sectors, during similar daytime intervals (before sunrise and after sunset) and comparable synoptic conditions. Therefore, future measurement campaigns can reproduce the same spatial and temporal framework, allowing direct comparisons between different years and facilitating the assessment of long-term changes in urban thermal characteristics.

To represent the spatial distribution of impervious surfaces, we utilized the Imperviousness Density (IMD) dataset provided by the Copernicus Land Monitoring Service (10 × 10 m spatial resolution) for the year 2018. IMD quantifies the percentage of impervious surfaces, highlighting modification of natural soil sealing driven especially by urbanization. This process involves replacing the original (semi-)natural terrain or water surfaces with an artificial, frequently impermeable coating. The IMD layer used in this study was obtained directly from the Copernicus Land Monitoring Service (CLMS, 2018). According to the CLMS technical documentation, it is produced through a semi-automated classification process based on multispectral satellite imagery and vegetation indices (including NDVI) to estimate soil sealing (1–100%). In this study, the IMD dataset was used directly, without additional classification procedures. IMD products generated by the CLMS (2018) tend to under-represent actual IMD values as evaluated by the validation team (GMES, 2020). However, in the literature, the delimitation of urban areas through the IMD ratio uses various IMD thresholds (Lu et al., 2011; Sfîcă et al., 2023a).

The digital elevation model (DEM) used in current study has a spatial resolution of approximately 90 m (Jarvis et al., 2008), covering the current approach needs, taken to account that we used a 100 × 100 m grid system to interpret our observation data from mobile measurements. The slope and aspect models were generated from the DEM using ArcGIS Pro software.

3.2 Methods

3.2.1 Data collection through expeditionary measurements

In order to carry out the mobile meteorological observations, a methodology was developed aiming to collect air temperature data as representative as possible from the perspective of their ability to emphasize general characteristics of the urban climate.

The methodological approach firstly aimed to conduct mobile measurements that cross both the central areas of the cities, as proposed in other studies (Brandsma and Wolters, 2012; Leconte et al., 2015; Leconte et al., 2017; Kim et al., 2022; Samad and Vogt, 2020; Young et al., 2022; de Costa Trinidad Amorim et al., 2024), but also the peri-urban and rural areas that are not influenced directly by urban climate conditions. Therefore, the transects were made following routes that differ from one observation to another, but that necessarily intersect the dense central urban areas of the cities and also their peripheries.

Secondly, the mobile transects were planned to cover as much as possible from the entire geomorphometric diversity of the cities. Also, the elevation range covered during tracks for each city overlaps the natural one, spanning between 123 m in the smallest analyzed city of Vaslui and 313 m on the largest city of Iași (Table 1). Due to their rather limited spatial extension, the analyzed cities offered the possibility to be covered in a limited interval of time by car, generally spanning up to 2.5 h. This represented a time interval during which we can assume, especially taken into account the general high pressure conditions selected for observations, a rather moderate (evening) or small (morning) linear change in air temperature.

Thirdly, given the fact that urban climate is better expressed during high pressure conditions, the mobile observations were carried out in these conditions, preferably associated with atmospheric calm, clear sky and low wind speeds that allow intense radiative exchanges (Leconte et al., 2015; Leconte et al., 2017; de Costa Trinidad Amorim et al., 2024) between the free atmosphere and the urban land surface. In this way, further emphasizing the representativeness of the collected dataset, the monitoring days correspond to clear sky days (taking to account only low cloud cover) that represent some 35–40% of the total number of days between May and September over the analyzed region (Sandu, 2008). Moreover, the observations were made before sunrise, during the peak of nocturnal radiative loss at the land surface and after sunset, on the maximum of the emissivity of the land surface after the diurnal radiative forcing during which the features of urban conditions on air temperature are less pronounced (Xu et al., 2020). It is to mention here that an area acting as hotspot during evening keeps the discomfort level high during the night, while an area acting as a hotspot during morning enables the rise of the temperature to high level during the day. Thus, the obtained hot/cold spots represent key areas for thermal discomfort within the analyzed cities and deserve a detailed description of their urban environmental features.

From a climatic perspective, the representativeness of our mobile observations is highlighted by the fact that the composite mean of the monitoring days (20 independent days for morning and 16 days for evening, respectively) indicate, for both intervals, well defined anticyclonic conditions over the analyzed region (Figure 2a,b), and also small positive air temperature anomalies (Figure 2c,d) compared to the multiannual mean of 1981–2010 from ERA−5 reanalysis data (Hersbach et al., 2020).

Figure 2

Last but not least, all mobile measurements were conducted along a closed spatial loop, starting and ending at the same point, so that the temperature gradient per minute could be calculated, overcoming one of the most important gaps of mobile measurements (Kousis et al., 2021), and avoiding the complicated issue of temporal correction using data from fixed weather stations (Liu et al., 2017a; Ichim et al., 2018). Then, assuming the linear temporal evolution of air temperature, the corresponding time gradient was calculated and applied to the entire data series (Ichim et al., 2018).

This way, from May to September 2022 (Table 1), for each of 6 analyzed cities 5 to 6 mobile measurements were made for morning, and 5 to 6 for evening, respectively, summing up a total of 64 mobile tracks (32 tracks for each morning and evening hours, respectively; details and descriptive statistics of these tracks are extensively presented in Supplementary Appendix2).

3.2.2 Assessment of thermal spots

The identification of hot and cold spots ensures the detection of persistent thermal anomalies within urban environments and helps to identify the natural and anthropogenic factors controlling their spatial distribution. This information is essential for understanding the mechanisms underlying UHI development and for supporting urban planning strategies aimed at reducing thermal stress.

The assessment of thermal spots was performed in two successive steps. Firstly, Global Moran’s I (Moran, 1950) was used to evaluate the overall spatial autocorrelation of air temperature values and to verify whether their distribution significantly departed from a random spatial pattern. The formula of this index is as follow:

where:

—is the number of spatial units indexed by and ;

—is the variable of interest;

—in the mean of ;

—are the elements of a matrix of spatial weights with zeros on the diagonal;

—is the sum of all , .

Secondly, hot and cold spots were identified using the Hot Spot Analysis (Getis-Ord Gi*) tool implemented in ArcGIS Pro. This method calculates z-score and p-value for each grid cell by comparing the local concentration. The Getis-Ord Gi* (Getis and Ord, 1992) is calculated in ArcGIS Pro software by using the formula:

where,

—air temperature value at location ;

—spatial weight between location and ;

—total number of observations;

—mean air temperature of all observations;

—standard deviation of all observations.

The Moran’s Index can vary between −1 and 1, where a Moran’s Index close to 1/−1 indicates a strong positive/negative spatial autocorrelation. This means that similar values are found close together in space, suggesting the presence of spatial clusters (Getis and Ord, 1992; Anselin, 1995).

Due to their high spatio-temporal density, mobile measurements are very useful when trying to achieve the spatial distribution of hot/colds spots for different environmental parameters (Samad and Vogt, 2020). In our approach, all mobile measurements were merged into a single spatial database for each daytime period (morning and evening). Subsequently, a 100 m buffer was created round the transects in order to define a continuous analysis area covering the monitored urban sectors. This area was then divided into a regular 100 × 100 m grid, allowing the aggregation and comparison of air temperature values and the subsequent application of spatial statistics. The selected spatial resolution represents a compromise between preserving local thermal variability and ensuring sufficient spatial coverage across all cities. After applying the hot/cold spots method in ArcGIS Pro software, we averaged the obtained z-score for each of two data sets (morning and evening measurements).

Finally, we analyzed the spatial distribution of areas characterized as hot and cold spots through their relationship with morphometric factors (altitude, slope, and aspect), as well as their relation with urban structure characteristics, as they are indicated by IMD.

3.2.3 Prediction of urban hot spots

Based on the identified features of landforms and built-up ratio associated with urban hot spots, outlined using the Getis-Ord Gi* analysis, an extrapolation method was conceived, aiming to predict the areas that may evolve into urban hot spots in response of a very likely increase in IMD under the effect of future urbanization. The extrapolation method uses an additive model built on two main explanatory variables—altitude and IMD—generally accepted as key drivers of the spatial variability (Crétat et al., 2023; Liu and Wang, 2023) of air temperature in urban environments.

For each analyzed city and for the two selected time periods of the day, descriptive statistics of altitude and IMD corresponding to the hotspots areas were extracted. Each city was treated independently, using its own hotspot-derived thresholds for altitude and IMD. Therefore, no direct inter-city referencing or common threshold values were applied. This approach was adopted to account for the differences in urban morphology, topography, and city size among the analyzed cities. After that, these values have been used as reference thresholds, as they capture the typical conditions that accurately characterize the occurrence and persistence of hotspot areas.

To perform the spatial extrapolation, a unidirectional weighting scheme was implemented. Following the methodology adapted from Sfîcă et al. (2013), for each of both analyzed layers (altitude and IMD), values within the specific hotspot thresholds were reclassified and assigned with positive weights (+3), as they indicate an increased potential for hotspot occurrence across the urban surface. The reclassified rasters obtained for the two predictors were aggregated by summation, resulting in an additive model meant to predict the areas prone to hotspots occurrence.

The following formula was used for spatial hotspot prediction:

where:

–hotspot prediction index value for the grid cell ;

, —the weighting for altitude () and IMD () (in this method );

—reclassified values (0 și 1) for each layer, defined as follows:

where:

—altitude range (A) for the identified hotspots areas (H);

—IMD range (I) for the identified hotspots areas (H);

, —altitude and IMD values for grid cell .

In these conditions, cells that simultaneously meet both conditions ( and ) receive the maximum score of +6, indicating a high probability for that gridcell to be included in the hotspot area of the city, which largely corresponds to the UHI core. In the meantime, cells with none of the two conditions met are indicated by 0 values.

Intermediate values (+3) indicate moderate probability (either the IMD or altitude conditions met) of the cell to evolve into a hotspot area, being defined as potential hotspots. Such areas can develop into actual hotspots, if the built-up area increases, provided that altitude conditions are already favorable. Conversely, when altitude conditions are not met but the built-up ratio is high, a further densification of IMD may also lead to hotspots occurrence, owing to the well-known role of built-up density in urban heat island development.

The results were mapped on a uniform 100 × 100 m grid, ensuring spatial comparability among cities and enabling integration with other thematic indicators (e.g., land use, population distribution, or green space structure).

The validation method consisted of a spatial overlapping between the hotspots predicted through the extrapolation model and those identified based on mobile measurements (Supplementary Appendix 3). The accuracy of the model overpasses 78%, expressing a very good reability of the spatial prediction.

4 Results and discussions

4.1 General features of hot and cold spots distribution over the analyzed cities

Our analysis indicate that the spatial distribution and frequency of hot and cold spots are primarily controlled by the intensity and spatial extent of the UHI in each city. Consequently, the most persistent hot spots are generally located within densely built-up urban sectors. However, beyond the influence of intensely built-up urban centers, a subset of cities exhibits altitude elevated hotspots that are not directly correlated with built-up ratio, being rather governed by morphometric characteristics of landforms under the direct influence of synoptic patterns (Figures 3, 4).

Figure 3

Figure 4

4.1.1 Hot and cold spots spatial distribution during mornings

The measurements aimed to identify hot and cold spots under the UHI effect, but they give also a valuable image of the temperature inversion phenomena (Ichim et al., 2014). During the morning hours, the observed spatial pattern of hot/cold spots is consistent with the manifestation of the UHI in terms of the size and geometry of each analyzed city.

The altitudinal differences and the high degree of land fragmentation jointly contribute to the formation of a dynamic spatial puzzle in the distribution of hot/cold spots. The frequent thermal inversion phenomena that intensify during the early morning hours have the capacity to further increase the vertical temperature differences (Ichim et al., 2014). This phenomenon found extensive conditions for development and manifestation over the analyzed region, since the investigated cities are characterized by significant altitudinal differences, exceeding 100 m in Iași (the largest altitude range; Table 2). As well, during the night the atmospheric stability is higher and commonly associated with the temperature inversions. In these conditions, areas characterized by cold spots are located over the extra-urban areas along the valleys, with corresponding areas of hot spots displaced in altitude. As a result, the mosaic of hot and cold spots is more diverse in areas with greater altitudinal differences (Table 1), such as Iași (Figure 3c) or Suceava (Figure 3a).

Because of its location at the contact between the sub-mountain and mountain areas (Figure 3c), the city of Piatra-Neamț presents very dynamic topo- and micro-climatic differences. Under these conditions, the city is divided into two spatial sectors, with hot and cold spots distributed between sub-mountain and mountain areas. The different fragmentation between the two regions ensures ideal conditions for the occurrence and intensification of thermal inversions through the descent of cold air from altitude into valleys in the north-western part of the city.

An organized distribution of hot and cold spots can be observed in the case of the Bacău city (Figure 3e). The geometry of the hot spots describes accurately the intensely built-up central urban area and its industrial part in the south. The same distributions can be observed in the case of the cities of Botoșani (Figure 3b) and Vaslui (Figure 3f), the last one being more balanced between hot and cold spots. Central urban sectors characterized by intense hot spots are well-defined by moderate temperature buffer zones that separate them from cold spots.

4.1.2 Hot and cold spots spatial distribution during evenings

Mobile measurements conducted in the evening hours reveal a near-identical spatial distribution of thermal hot/cold spots, with hotspots laying over larger areas compared to the morning hours. This distribution can be attributed to the enhanced thermal energy storage capacity of artificial surfaces. During this time, artificial areas located at lower altitudes, which were initially dominated in the morning by cold air from thermal inversions and classified as cold spots or transition zones, change into hotspots areas. The evening configuration of hotspots provides a more accurate representation of the spatial extent and geometry of the UHI as indicated by SUHI for Iași (Sfîcă et al., 2023a) or Bacău (Sfîcă et al., 2023b).

Visually, it can be observed that in the six analyzed cities (Figure 4), regardless of their size, the spatial extension of hotspots varies even under conditions of atmospheric stability, highlighting persistent central warm areas, with a series of hotspots manifesting only during the evening.

4.1.3 Air temperature contrasts between hot and cold spots

Morning observations reveal pronounced thermal contrasts between urban hot- and cold spots (Table 2). The mean air temperature of all morning transects varies between 12.6 °C (Vaslui) and 15.8 °C (Botoșani), with hot spots distributed between 13.7–16.4 °C (min)/ 14.6–18.4 °C (max) and cold spots distributed between 9.6–13.5 °C (min)/ 11.7–15.3 °C (max). Even if these values are not completely comparable, due to the fact that the observation days were partially different from the point of view of air temperature, they capture well the overall magnitude of temperature differences inside the analyzed cities (Table 2).

Generally, the differences between the mean air temperature of hot and cold spots spans between 3.9 °C in Iași, a city that encompasses a larger area with considerable topographical contrast, and 1.9 °C in Bacău, a city with less altitude differences along the transects, while the rest of the cities are recording differences between 2.3 and 3.4 °C. Overall, this difference between the mean temperature of hot and cold spots over the city represents a good estimate of the temperature contrasts in the area of the analyzed cities during summer mornings.

During evening hours (Table 3), the mean difference between hot and cold spots is higher than during morning hours, in complete agreement with in-situ observations over the region that are showing the most intense UHI recorded during the first part of the night (Sfîcă et al., 2018). As well, a considerable contribution to this larger difference is given by the strong temperature inversions observed during the same hours in the evening (Ichim et al., 2018) enforcing the hot spots located at higher altitudes in cities like Iași, Vaslui or Suceava. This way, the mean differences between hot and cold spots range during morning between 2.3 °C in Bacău and 4.3 °C in Iași, on a background of mean air temperature for all transects covering the interval between 19.1 °C in Vaslui/Piatra-Neamț and 22.7 °C in Botoșani. The mean temperature of hot spots spans between 20.0–23.5 °C (min) and 21.4–24.5 °C (max), while cold spots have mean temperature between 14.6–18.8 °C (min) and 18.2–22.2 °C (max).

From this analysis we understand that the differences between the mean air temperature of hot and cold spots represents a very close proxy of UHI intensity, having the great advantage that it can be assessed with less costs and consumed time, through mobile measurements. Overall, altitude is able to increase the thermal contrast over the city with heights, while the UHI increases this contrast at the same altitude.

4.2 Local drivers of hot and cold spots distribution

4.2.1 Geomorphometrical factors

These elements play a significant role in the spatial distribution of thermal parameters, influencing the frequency and intensity of thermal inversion phenomena, as well as the uneven heating of the terrain depending on the magnitude of slope and its aspect, features with a considerable impact on the surface heat radiation balance defined as neighborhood effect (Firozjaei et al., 2024).

• (a) The altitude, a basic widely known geomorphometric element, is known to have a high capacity in explaining complex features of the intra-urban air temperature variation (Lindberg, 2007). The relationship between hot/cold spots over the analyzed cities and the altitude is very complex. The urban relief and air temperature are tightly related (Bokwa et al., 2015; Bokwa et al., 2019), especially within the cities with a contrasted topography (Crétat et al., 2023).

In the morning, the hot spots tend to be localized in higher altitudes (Figure 5; Table 4), as a result of the strong thermal inversions developing during the night and manifesting in the morning hours on the bottom of the valleys with their cold layer and uphill with their warm band. The location of the hot spots reflects a balance between the intensity of thermal inversions and the distribution of the densely built-up areas. Consequently, despite strong air temperature inversions, the city core—although often situated at lower elevations—generally remains warmer, and is therefore identified as hot spot. This is the case especially for those cities that are developed over the low terraces of the rivers (Bacău, Piatra-Neamț). For the evening hours, when the air temperature inversions are not yet very intense in their cold layers, only the cities that do not have their centers at low altitudes (Vaslui and Botoșani) are warmer in altitude, while the other cities maintain their non-disturbed hot spots close to the valleys.

• (b) The terrain slope apparently plays a marginal role for cold and hot spots spatial extension (results not shown). As a general, but inconsistent feature, cold spots tend to develop more often on higher slopes inside the cities, a fact that can be explained by multiple factors. On one hand, higher slopes are less densely built areas due to the construction limitations for large and dense buildings. On the other hand, higher slopes are specific over the analyzed cities for the front of the river terraces and these areas are placed in the proximity of the valleys that are colder due to the extension here of cold air at the bottom of air temperature inversions. It is to note as well that for slope the difference between the means of cold/hot spots areas does not pass regularly the statistical significance thresholds.

• (c) The aspect of slopes presents a very important role in cold and hot spots distribution during summer (Figure 6). This was assessed by comparing the terrain exposure frequency over the cold and hot spots with the general exposure frequency along the entire transects. In this way, we can outline that some terrain exposures are overrepresented/underrepresented in hotspots and coldspots distribution in the analyzed cities. The exposure seems to have a clearer association with hot- and coldspots distribution during the evening hours (Figure 6b), due to the highest heat amount stored by the cities during the day, especially after the long summer days with high insolation.

Figure 5

Figure 6

In this case, SSE to SW exposures are the most overrepresented in hotspots distribution, while NE and ENE exposures are the most overrepresented in cold spots distribution. This is clearly balanced by the underrepresentation of N to ENE exposures in hotspots and by the underrepresentation of SSE to SW in cold spots. During the morning hours, the association between terrain exposure and cold/hot spots distribution is weaker, due to the progressive consumption of stored heat during the night.

Nevertheless, especially S and SW exposures are overrepresented/underrepresented in hot/cold spots occurrence. Also, a discreet effect of insolation and heat storage over the night is indicated by the overrepresentation of cold spots over WSW and W slopes, and the underrepresentation over NE and ENE. This indicates a possible effect of higher diffuse radiation before sunrise on NE and ENE slopes that contrast with the sheltering effect from WSW and W. Overall, flat areas of the riverbeds are overrepresented in coldspots distribution due to the higher impact of air temperature inversions in these areas.

4.2.2 Impervious density (IMD)

The ratio of built-up areas represents one of the most important predictor of UHI intensity (Bottyan et al., 2005; Brandsma and Wolters, 2012), being constantly and positively correlated with air temperature (Xu et al., 2020). Moreover, increase in built-up ratio leads to the development of SUHI even in very small cities (Ichim et al., 2024). Both building height and building surface fraction are positively correlated with the night time temperature (Shi et al., 2018). However, Liu et al. (2017b) underlined for Shenzhen city that building height is negatively correlated with UHI intensity due to their increased radiative sheltering effect, as identified also in other studies (Aghazadeh et al., 2026).

The built-up ratio represents by far the most important factor for the occurrence and development of hot and cold spots in the area of the analyzed cities. Generally, the mean built-up ratio of the hot spots is placed between 39 and 51% in the morning and 42–54% in the evening, while cold spots are specific for areas with a built-up ratio between 19 and 31%. Generally, hot spots are associated with a higher built-up ratio in the evening (Figure 7), due to the maximum accumulation of the heat during the day.

Figure 7

For cold spots there is no clear association with built-up ratio, indicating that these areas are under the influence of local climate conditions induced by topography or land use. In this regard, lower temperatures are consistently observed over urban parks (Samad and Vogt, 2020; Corocăescu et al., 2023), since vegetation coverage is known to be negatively correlated with the air temperature (Xu et al., 2020). Vegetation and civil buildings contribute more to higher UHI intensity during the night (Liu et al., 2016; Xu et al., 2020), while anthropogenic heat or industrial buildings impose higher UHI during the noon (Liu et al., 2016). Cities as Iași, with a high built-up ratio in its low altitude area present an increase built-up ratio for the cold spots area, showing that thermal inversions can dismantle the heating capacity of dense urban areas.

4.3 Spatial prediction of hot spots based on altitude and IMD

Due to major challenges triggered by the synergies between urbanization and climate change, leading to increased heat stress in urban areas (Cheval et al., 2023), we further focused our analysis on spatial prediction of potential emerging hotspots. The analysis aggregated the results from mobile measurements, using altitude and IMD as main predictors, into an index capable to estimate where future hotspots may emerge.

During the morning hours (Figure 8), the predicted hotspots and potential hotspots are mainly concentrated in the densely central areas and over the slightly elevated neighborhoods of the cities, while the valley floors and low-lying floodplains show a lower potential to develop into hotspots. This is caused here by the lower temperature associated with thermal inversions. In cities such as Suceava, Iași, and Piatra-Neamț, the model clearly reflects the topographic control over air temperature, with hotspots located on upper terraces and urban hills plateaus where both altitude and IMD values concentrate the air temperature in upper quartile (Q3). These results highlight the amplifying effect of the terrain morphology contrasts and high urban compactness on the configuration of the morning UHI (Esposito et al., 2024).

Figure 8

During the evening hours (Figure 9), the predicted hotspots expand spatially and tend to merge into broader thermal cores, especially into the dense built-up areas of the cities, along the urban corridors that can act as triggers of the UHI intensification (Anupriya, 2016). The transition between the morning and evening hours indicates a progressive increase in the continuity of warm areas, especially in Bacău, Botoșani, and Vaslui, where the central and industrial sectors maintain a high thermal potential even after sunset. The evening thermal pattern corresponds more closely to the nocturnal UHI configuration described in previous studies (Sfîcă et al., 2018, 2023a), where artificial surfaces located at lower altitudes remain significantly warmer due to delayed cooling.

Figure 9

Overall, the comparison between the two temporal time steps highlights the dual role of altitude and urban morphology in modulating hotspots formation. Morning hotspots are primarily topographically driven, whereas evening ones are more closely associated with the IMD, indicating the strong thermal inertia of built-up surfaces. The consistent overlap between validated and predicted hotspots (more than 78%; please see Supplementary Appendix 2) confirms the efficiency of the unidirectional weighting scheme and the suitability of the altitude—IMD composite index for identifying thermally vulnerable urban sectors.

This spatial framework demonstrates the predictive capability of the model in delineating both current and potential hotspots, providing a valuable diagnostic and planning tool for reducing urban heat stress. The areas identified as predicted and potential hotspots should represent priority areas to which the adaptive measures should be targeted (limiting built-up ratio increase, expansion of green areas, restoration of soil permeability, or the use of reflective materials).

The practical relevance of this approach is further supported by recent satellite-based analyses conducted in north-eastern Romanian cities, which identified persistent hotspots in densely built urban cores, industrial areas and major urban corridors, while green spaces and water bodies systematically emerged as cold spots (Crețu et al., 2025). Although land surface temperature and near-surface air temperature describe different components of the urban thermal environment, the recurrence of similar spatial patterns strengthens the robustness of the identified thermally vulnerable sectors and supports their use in urban planning and climate adaptation strategies.

The present results also highlight a new direction in our future urban climate studies, resulting from the complementary role of standardized mobile measurements and high-resolution urban climate modeling. While mobile observations provide high-resolution information on the spatial distribution of urban hot and cold spots under real atmospheric conditions, numerical models such as ENVI-met and PALM-4 U allow a detailed investigation of the physical processes controlling these thermal patterns, including the influence of urban morphology, vegetation, surface materials, shading, radiation exchange and airflow (Bruse and Fleer, 1998; Maronga et al., 2020). In this case, standardized mobile measurements can provide valuable observations for the calibration and validation of these models, improving their capability to reproduce the spatial variability of urban air temperature under real atmospheric conditions (Yang et al., 2013; Crank et al., 2020; Liu et al., 2021).

From an urban climatology perspective, the hotspot maps obtained in the present study may serve not only as reference observations for future urban climate modeling studies but also as a scientific basis for identifying thermally vulnerable sectors where different mitigation strategies can be evaluated through scenario-based simulations (Maronga et al., 2020; Ichim et al., 2026). At the same time, the identified cold spots represent valuable reference areas from which favorable urban configurations and cooling mechanisms can be identified and subsequently adapted or replicated in other urban sectors using high-resolution models (Bruse and Fleer, 1998; Maronga et al., 2020).

5 Conclusion

Firstly, our study tests with very good results a very effective IoT tool for urban climate exploration through mobile measurements. This indicates that a finely planned mobile observation campaign, overpassing the common limitations of the mobile measurements, is able to deliver enough and sufficiently accurate data in order to support a comprehensive assessment of UHI, as made by a classical identification of cold and hot spots. Beyond this, the complex interaction of morphological factors and IMD creates a diversity of hot and cold spots, with a specific distribution within each city. Our analysis highlights that high IMD ratios represent the main factor in the emergence of hot spots in all six studied cities, regardless of their size. The areas with intense hot spots overlap the central urban areas, characterized by high built-up ratios, emphasizing the role of this factor in the development of UHI.

Landforms morphological factors can amplify/diminish the hots/cold spots distribution. Altitude plays the most important role in this regard, with higher areas being generally cooler, especially in the morning. Despite this general rule, thermal inversions intensify altitude-related temperature differences, favoring the occurrence of cold spots also over the floodplain. Morning thermal inversions amplify the air temperature differences between hot and cold spots, especially in cities with large altitude differences. This effect is more pronounced in low-lying areas, where cold air is trapped by the inversion, resulting in significantly lower temperatures. The aspect of slopes has also a significant influence in the evening, with slopes facing south-southeast and southwest being warmer due to higher insolation. Cold spots are more likely to occur on higher slopes (due to less development of buildings and in the proximity of the valleys) and on NE-ENE slopes (due to lower insolation and sheltering effect). Hot spots, however, show a clearer association with SSE-SW slopes, particularly in the evening, due to their higher heat storage capacity.

Based on the above mentioned characteristics of cold/hot spots a simple prediction model of hotpots was proposed, developed using altitude and built-up ratios. The area identified as potential hot spots represent a valuable indicator for the regions inside and outside the cities where an increase in built-up ratio can lead directly to a hot spot occurrence. This could represent a very useful tool for stakeholder in cautionary development of the urban space, avoiding the extension of the city in area prone to increase heat stress risk. The fine spatial resolution of the obtained results offers the opportunity of detailed comparison with results of urban climate modeling.

It should be emphasized that the identified hot and cold spots represent relative thermal features within each city and should not be interpreted as absolute indicators of thermal risk. An identified hotspot does not automatically require intervention from local authorities. Instead, hotspot identification should be considered a first step toward prioritizing areas for further investigation, particularly when they overlap with densely populated sectors, vulnerable social groups, or areas exposed to prolonged thermal stress. Conversely, cold spots represent thermally favorable sectors that contribute to urban cooling during summer and should be preserved and integrated into future climate adaptation strategies. Nevertheless, these interventions should be made with caution in the current time in order to avoid also a possible intensification of cold spots during winter, when cold stress still poses major problems.

References

• 1

  AghazadehF.OndrejickaV.RahimiA.ChevalS.AkbariM. A.FirozjaeiM. K.et al. (2026). When dense built-up areas cool: multi-scale neighborhood pattern changes of a semi-arid urban cooling paradox in Tehran (2000–2020). Sustain. Cities Soc.147:107540. doi: 10.1016/j.scs.2026.107540

  • CrossRef
  • Google Scholar

• 2

  AnselinL. (1995). Local indicators of spatial association—LISA. Geogr. Anal.27, 93–115.

  • Google Scholar

• 3

  AnupriyaS. (2016). Exploratory study on the relation between urban landscapes and urban corridors for outdoor thermal comfort. Procedia Technol.24, 1801–1807. doi: 10.1016/j.protcy.2016.05.224

  • CrossRef
  • Google Scholar

• 4

  BokwaA.GeletičJ.LehnertM.Žuvela-AloiseM.HollósiB.GálT.et al. (2019). Heat load assessment in central European cities using an urban climate model and observational monitoring data. Energ. Buildings201, 53–69. doi: 10.1016/j.enbuild.2019.07.023

  • CrossRef
  • Google Scholar

• 5

  BokwaA.HajtoB. A.WalawenderM. J.JakubP.SzymanowskiM. (2015). Influence of diversified relief on the urban heat island in the city of Kraków, Poland. Theor. Appl. Climatol.122, 365–382. doi: 10.1007/s00704-015-1577-9

  • CrossRef
  • Google Scholar

• 6

  BottyanZ.KircsiA.SzegediS.UngerJ. (2005). The relationships between built-up areas and the spatial development of the mean maximum urban heat island in Debrecen, Hungary. Int. J. Climatol.25, 405–418. doi: 10.1002/joc.1138

  • CrossRef
  • Google Scholar

• 7

  BrandsmaT.WoltersD. (2012). Measurement and statistical modeling of the urban heat island of the city of Utrecht (the Netherlands). J. Appl. Meteorol. Climatol.51, 1046–1060. doi: 10.1175/JAMC-D-11-0206.1

  • CrossRef
  • Google Scholar

• 8

  BruseM.FleerH. (1998). Simulating surface–plant–air interactions inside urban environments with a three dimensional numerical model. Environ. Model. Softw.13:373.

  • Google Scholar

• 9

  CeciliaA.PengZ. (2022) “Validating a Cycle-Based IoT Sensor for Mapping Intra-Urban Air Temperature and Humidity,” Conference Paper, PLEA SANTIAGO 2022 Will Cities Survive? November, 2022.

  • Google Scholar

• 10

  ChenY.LiJ.HuY.LiuL. (2023). Spatial and temporal characteristics of nighttime UHII based on local climate zone scheme using mobile measurement-a case study of Changsha. Build. Environ.228:109869. doi: 10.1016/j.buildenv.2022.109869

  • CrossRef
  • Google Scholar

• 11

  ChevalS.DumitrescuA.AmihăeseiV. A.IrașocA.ParaschivM. G.GhentD. (2023). A country scale assessment of the heat hazard-risk in urban areas. Build. Environ.229:109892. doi: 10.1016/j.buildenv.2022.109892

  • CrossRef
  • Google Scholar

• 12

  CLMS (2018) Copernicus Land Monitoring Service. Impervious Density. Available online at: https://land.copernicus.eu/pan-european/high-resolution-layers/imperviousness (Accessed on 14 October 2023.

  • Google Scholar

• 13

  CorocăescuA.IchimP.CrețuC. Ș.DiorA.ȘerbanL. (2023) “Assessment of climate characteristics of an Urban Park using satellite imagery and in-situ measurements. Study case of Cancicov Park from Bacău City (Romania),” Air And Water – Components of the Environment - conf Proceedings, Cluj-Napoca, Romania 33–46

  • Google Scholar

• 14

  CrankP. J.MiddelA.WagnerM.HootsD.SmithM.BrazelA. (2020). Validation of seasonal mean radiant temperature simulations in hot arid urban climates. Sci. Total Environ.749:141392. doi: 10.1016/j.scitotenv.2020.141392,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 15

  CrétatJ.RichardY.PohlB.EmeryJ.DudekJ.RoyD.et al. (2023). Impact of topography and land cover on air temperature space-time variability in an urban environment with contrasted topography (Dijon, France, 2014-2021). Theor. Appl. Climatol.155, 1941–1958. doi: 10.1007/s00704-023-04742-4

  • CrossRef
  • Google Scholar

• 16

  CrețuS. C.SfîcăL.IchimP.AmihăeseiV. A.BreabănI. G.RoșuL.et al. (2025). Warm season land surface temperature and its relationship with local climate zones in post-socialist cities. Theoretical and Applied Climatology, 156, 191. doi: 10.1007/s00704-025-05409-y

  • CrossRef
  • Google Scholar

• 17

  de Costa Trinidad AmorimM. C.DubreuilV.TeixeiraD. C. F.AmorimA. T.BrabantC. (2024). Exceptional heat island intensities also occur in medium-sized cities. Urban Clim.53:101821. doi: 10.1016/j.uclim.2024.101821

  • CrossRef
  • Google Scholar

• 18

  EspositoA.PappaccogliG.DonateoA.SalizzoniP.MaffeisG.SemeraroT.et al. (2024). Urban morphology and surface urban heat island relationship during heat waves: a study of Milan and Lecce (Italy). Remote Sens.16:4496. doi: 10.3390/rs16234496

  • CrossRef
  • Google Scholar

• 19

  FirozjaeiM. K.MijaniN.KiavarzM.DuanS. B.AtkinsonP. M.AlavipanahS. K. (2024). A novel surface energy balance-based approach to land surface temperature downscaling. Remote Sens. Environ.305:114087. doi: 10.1016/j.rse.2024.114087

  • CrossRef
  • Google Scholar

• 20

  GetisA.OrdJ. K. (1992). The analysis of spatial association by use of distance statistics. Geogr. Anal.24, 189–206. doi: 10.1111/j.1538-4632.1992.tb00261.x

  • CrossRef
  • Google Scholar

• 21

  GMES (2020) Initial operations / Copernicus land monitoring services – validation of products, HRL IMPERVIOUSNESS DEGREE 2018 VALIDATION REPORT. Available online at: https://land.copernicus.eu/en/products/high-resolution-layer-imperviousness (Accessed on 20 January 2024).

  • Google Scholar

• 22

  HedquistB. C.BrazelA. J. (2006). Urban, residential, and rural climate comparisons from mobile transects and fixed stations: Phoenix, Arizona. J. Ariz. Nev. Acad. Sci.38, 77–87. doi: 10.2181/1533-6085(2006)38[77:URARCC]2.0.CO;2

  • CrossRef
  • Google Scholar

• 23

  HerbelI.CroitoruA. E.RusI.HarpaG. V.CiuperteaA. F. (2016). Detection of atmospheric urban heat island through direct measurements in Cluj-Napoca city, Romania. Hung. Geogr. Bull.65, 117–128. doi: 10.15201/hungeobull.65.2.3

  • CrossRef
  • Google Scholar

• 24

  HersbachH.BellB.BerrisfordP.HiraharaS.HorányiA.Muñoz SabaterJ.et al. (2020). The ERA5 global reanalysis. Q. J. R. Meteorol. Soc.146, 1999–2049. doi: 10.1002/qj.3803

  • CrossRef
  • Google Scholar

• 25

  IchimP.ApostolL.SfîcăL.Kadhim-AbidA.-L.IstrateV. (2014). Frequency of thermal inversions between Siret and Prut rivers in 2013. Pres Envir and Sustain Develop8, 267–285. doi: 10.2478/pesd-2014-0040

  • CrossRef
  • Google Scholar

• 26

  IchimP.MironR.CorocăescuA. C.CrețuC.SfîcăL. (2024). Land use as key element for urban heat island inception in small cities. Case study: Bârlad City, Romania. Present Environ. Sustain. Dev.18:79–94. doi: 10.47743/pesd2024181006

  • CrossRef
  • Google Scholar

• 27

  IchimP.SfîcăL. (2020). The influence of urban climate on bioclimatic conditions in the city of Iași, Romania. Sustainability12:9652. doi: 10.3390/su12229652

  • CrossRef
  • Google Scholar

• 28

  IchimP.SfîcăL.Kadhim-AbidA. L.UrsuA.JitaruV. (2018) “Characteristics of nocturnal Urban Island of Iași during a summer heat wave (1-6 of august 2017)”, Air and Water Components of the Environment – Conference Proceedings, 253–260

  • Google Scholar

• 29

  IchimP.ȚurcanuR.SfîcăL. (2026). Optimizing biometeorological comfort in the central pedestrian area of Iași city through various urban design scenarios implemented with ENVI-met. Geographia Technica21, 132–149. doi: 10.21163/GT_2026.213.09,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 30

  JarvisA.ReuterH. I.NelsonA.GuevaraE. (2008). Hole - Filled Seamless SRTM data V4. Cali, Colombia: International Centre for Tropical Agriculture (CIAT).

  • Google Scholar

• 31

  KimH.KimS. W.JoY.KimE. J. (2022). Findings from a field study of urban microclimate in Korea using mobile meteorological measurements. Open House Int.47, 473–493. doi: 10.1108/OHI-12-2021-0280

  • CrossRef
  • Google Scholar

• 32

  KousisI.ManniM.PiselloA. L. (2022). Environmental mobile monitoring of urban microclimates: a review. Ren. Sust. En. Rev.169:112847. doi: 10.1016/j.rser.2022.112847

  • CrossRef
  • Google Scholar

• 33

  KousisI.PigliautileI.PiselloA. L. (2021). Intra-urban microclimate investigation in urban heat island through a novel mobile monitoring system. Sci. Rep.11:9732. doi: 10.1038/s41598-021-88344-y,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34

  LambW. F.CreutzigF.CallaghanM. W.MinxJ. C. (2019). Learning about urban climate solutions from case studies. Nat. Clim. Chang.9, 279–287. doi: 10.1038/s41558-019-0440-x

  • CrossRef
  • Google Scholar

• 35

  LeconteF.BouyerJ.ClaverieR. (2020). Nocturnal cooling in local climate zone: statistical approach using mobile measurements. Urban Clim.33:100629. doi: 10.1016/j.uclim.2020.100629

  • CrossRef
  • Google Scholar

• 36

  LeconteF.BouyerJ.ClaverieR.PétrissansM. (2015). Using local climate zone scheme for UHI assessment: evaluation of the method using mobile measurements. Build. Environ.83, 39–49. doi: 10.1016/j.buildenv.2014.05.005

  • CrossRef
  • Google Scholar

• 37

  LeconteF.BouyerJ.ClaverieR.PétrissansM. (2017). Analysis of nocturnal air temperature in districts using mobile measurements and a cooling indicator. Theor. Appl. Climatol.130, 365–376. doi: 10.1007/s00704-016-1886-7

  • CrossRef
  • Google Scholar

• 38

  LiY.YangT.ZhaoG.MaC.YanY.XuY.et al. (2024). A systematic review of studies involving canopy layer urban heat island: monitoring and associated factors. Ecol. Indic.158:111424. doi: 10.1016/j.ecolind.2023.111424

  • CrossRef
  • Google Scholar

• 39

  LindbergF. (2007). Modeling the urban climate using a local governmental geo-database. Meteorol. Appl.14, 263–273. doi: 10.1002/met.29

  • CrossRef
  • Google Scholar

• 40

  LiuZ.ChengW.JimC. Y.MorakinyoT. E.ShiY.NgE. (2021). Heat mitigation benefits of urban green and blue infrastructures: a systematic review of modeling techniques, validation and scenario simulation in ENVI-met V4. Build. Environ.200:107939. doi: 10.1016/j.buildenv.2021.107939

  • CrossRef
  • Google Scholar

• 41

  LiuL.LinY.LiuJ.WangL.WangD.ShuiT.et al. (2017b). Analysis of local-scale urban heat island characteristics using an integrated method of mobile measurement and GIS-based spatial interpolation. Build. Environ.117, 191–207. doi: 10.1016/j.buildenv.2017.03.013

  • CrossRef
  • Google Scholar

• 42

  LiuL.LinY.WangD.LiuJ. (2017a). An improved temporal correction method for mobile measurement of outdoor thermal climates. Theor. Appl. Climatol.129, 201–212. doi: 10.1007/s00704-016-1769-y

  • CrossRef
  • Google Scholar

• 43

  LiuL.LiuJ.LinY. (2016). Spatial-temporal analysis of the urban Heat Island of a subtropical city by using mobile measurements. Proc Engin169, 55–63. doi: 10.1016/j.proeng.2016.10.007

  • CrossRef
  • Google Scholar

• 44

  LiuY.WangZ. (2023). Research progress and hotspot analysis of urban heat island effects based on cite space analysis. Land12:1154. doi: 10.3390/land12061154

  • CrossRef
  • Google Scholar

• 45

  LuD.MoranE.HetrickS. (2011). Detection of impervious surface change with multitemporal Landsat images in an urban–rural frontier. ISPRS J of Photogram and Rem Sens66, 298–306. doi: 10.1016/j.isprsjprs.2010.10.010,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 46

  MarongaB.BanzhafS.BurmeisterC.EschT.ForkelR.FröhlichD.et al. (2020). Overview of the PALM model system 6.0. Geosci. Model Dev.13, 1335–1372. doi: 10.5194/gmd-13-1335-2020

  • CrossRef
  • Google Scholar

• 47

  MihăilăD.BistriceanP. I.SfîcăL.HorodnicV. D.PrisăcariuA.AmihăeseiV. A. (2024). Summer discrepancies between 2 m air temperature and Landsat LST in Suceava City, northeastern Romania. Remote Sens.16:2967. doi: 10.3390/rs16162967

  • CrossRef
  • Google Scholar

• 48

  MilicaV.JelenaD.StevanS.OndřejD. (2025). Applications of local climate zone classification in European cities: a review of in situ and mobile monitoring methods in urban climate studies. Open Geosci.17:20250878. doi: 10.1515/geo-2025-0878

  • CrossRef
  • Google Scholar

• 49

  MoranP. A. P. (1950). Notes on continuous stochastic phenomena. Biometrika37, 17–23.

  • Google Scholar

• 50

  MullerC. L.ChapmanL.GrimmondC. S. B.YoungD. T.CaiX. (2013). Sensors and the city: a review of urban meteorological networks. Int. J. Climatol.33, 1585–1600. doi: 10.1002/joc.3678

  • CrossRef
  • Google Scholar

• 51

  National Institute of Statistics of Romania, (2023). National Institute of statistics of Romania, 2023. Available online at: https://www.recensamantromania.ro/ (Accessed on: 13.02.2024).

  • Google Scholar

• 52

  OkeT. R.MillsG.ChristenA.VoogtJ. A. (2017). “Urban Heat Island,” in Urban Climates, (Cambridge: Cambridge University Press), 197–237.

  • Google Scholar

• 53

  SamadA.VogtU. (2020). Investigation of urban air quality by performing mobile measurements using a bicycle (MOBAIR). Urban Clim.33:100650. doi: 10.1016/j.uclim.2020.100650

  • CrossRef
  • Google Scholar

• 54

  SanduI. (2008). Climate of Romania. Bucharest, Romania: Romanian Academy Publishing House, 365.

  • Google Scholar

• 55

  SeidelJ.KetzlerG.BechtelB.ThiesB.PhilippA.BöhnerJ.et al. (2016). Mobile measurement techniques for local and micro-scale studies in urban and topo-climatology. Erde147 p 15–39. doi: 10.12854/erde-147-2

  • CrossRef
  • Google Scholar

• 56

  SfîcăL.AndreiA.BărcăcianuF.CojocaruŞ.EneaA.HapciucO. E.et al. (2013). Analysis of snow-drifting vulnerability. Application to Botoşani county. Lucrările Seminarului Geografic" Dimitrie Cantemir"36, 43–54.

  • Google Scholar

• 57

  SfîcăL.CorocăescuA.-C.CrețuC.-S.AmihăeseiV.-A.IchimP. (2023b). Spatiotemporal features of the surface urban heat island of Bacau City (Romania) during the warm season and local trends of LST imposed by land use changes during the last 20 years. Remote Sens.15:3385. doi: 10.3390/rs15133385

  • CrossRef
  • Google Scholar

• 58

  SfîcăL.CrețuC. Ș.IchimP.HrițacR.BreabănI. G. (2023a). Surface urban heat island of Iași city (Romania) and its differences from in situ screen-level air temperature measurements. Sustain. Cities Soc.94:104568. doi: 10.1016/j.scs.2023.104568

  • CrossRef
  • Google Scholar

• 59

  SfîcăL.CroitoruA.-E.IordacheI.CiuperteaF.-A. (2017). Synoptic conditions generating heat waves and warm spells in Romania. Atmos.8:50. doi: 10.3390/atmos8030050

  • CrossRef
  • Google Scholar

• 60

  SfîcăL.IchimP.ApostolL.UrsuA. (2018). The extent and intensity of the urban heat island in Iași city, Romania. Theor. Appl. Climatol.134, 777–791. doi: 10.1007/s00704-017-2305-4

  • CrossRef
  • Google Scholar

• 61

  ShiY.Ka-Lun LauK.RenC.Ng (2018). Evaluating the local climate zone classification in high-density heterogeneous urban environment using mobile measurement. Urban Clim.25, 167–186. doi: 10.1016/j.uclim.2018.07.001

  • CrossRef
  • Google Scholar

• 62

  StewartI. D.OkeT. R. (2012). Local climate zones for urban temperature studies. Bull. Am. Meteorol. Soc.93, 1879–1900. doi: 10.1175/BAMS-D-11-00019.1

  • CrossRef
  • Google Scholar

• 63

  StraubA.BeckC.du PreezD. J.KnoteC.HolstC. C.PhilippA. (2025). Characterization of potential temperature hotspots during urban heat island episodes examined by large eddy simulation and land use regression for the city of Augsburg. Meteorol. Z.34, 125–148. doi: 10.1127/metz/1252

  • CrossRef
  • Google Scholar

• 64

  SundborgÅ. (1951). Climatlogical studies in Uppsala with special regard to the temperature conditions in the urban area. Geographica22:111.

  • Google Scholar

• 65

  Ting KwokY.Yung NgE. Y. (2021). Trends, topics, and lessons learnt from real case studies using mesoscale atmospheric models for urban climate applications in 2000–2019. Urban Clim.36:100785. doi: 10.1016/j.uclim.2021.100785

  • CrossRef
  • Google Scholar

• 66

  TsinP. K.KnudbyA.KrayenhoffE. S.HoH. C.BrauerM.HendersonS. B. (2016). Microscale mobile monitoring of urban air temperature. Urban Clim.18, 58–72. doi: 10.1016/j.uclim.2016.10.001

  • CrossRef
  • Google Scholar

• 67

  XuH.ChenH.ZhouX.WuY.LiuY. (2020). Research on the relationship between urban morphology and air temperature based on mobile measurement: a case study in Wuhan, China. Urban Clim.34:100671. doi: 10.1016/j.uclim.2020.100671

  • CrossRef
  • Google Scholar

• 68

  YangX.ZhaoL.BruseM.MengQ. (2013). Evaluation of a microclimate model for predicting the thermal behavior of different ground surfaces. Build. Environ.60, 93–104. doi: 10.1016/j.buildenv.2012.11.008

  • CrossRef
  • Google Scholar

• 69

  YokoyamaH.OokaR.KikumotoH. (2018). Study of mobile measurements for detailed temperature distribution in a high-density urban area in Tokyo. Urban Clim.24, 517–528. doi: 10.1016/j.uclim.2017.06.006

  • CrossRef
  • Google Scholar

• 70

  YoungE.KastnerP.DoganT.ChokhachianA.MokhtarS.ReinhartC. (2022). Modeling outdoor thermal comfort along cycling routes at varying levels of physical accuracy to predict bike ridership in Cambridge. Build. Environ.208:108577. doi: 10.1016/j.buildenv.2021.108577

  • CrossRef
  • Google Scholar