Remote Sensing Monitoring of Summer Heat Waves–Urban Vegetation Interaction in Bucharest Metropolis

Abstract

Through a comprehensive analysis of urban vegetation summer seasonal and interannual patterns in the Bucharest metropolis in Romania, this study explored the response of urban vegetation to heat waves’ (HWs) impact in relation to multi-climatic parameters variability from a spatiotemporal perspective during 2000–2024, with a focus on summer HWs periods (June–August), and particularly on the hottest summer 2024. Statistical correlation, regression, and linear trend analysis were applied to multiple long-term MODIS Terra/Aqua and MERRA-2 Reanalysis satellite and in situ climate data time series. To support the decline in urban vegetation during summer hot periods due to heat stress, this study found strong negative correlations between vegetation biophysical observables and urban thermal environment parameters at both the city center and metropolitan scales. In contrast, during the autumn–winter–spring seasons (September–May), positive correlations have been identified between vegetation biophysical observables and a few climate parameters, indicating their beneficial role for vegetation growth from 2000 to 2024. The recorded decreasing trend in evapotranspiration from 2000 to 2024 during summer HW periods in Bucharest’s metropolis was associated with a reduction in the evaporative cooling capacity of urban vegetation at high air temperatures, diminishing vegetation’s key function in mitigating urban heat stress. The slight decline in land surface albedo in the Bucharest metropolis due to increased urbanization may explain the enhanced air temperatures and the severity of HWs, as evidenced by 41 heat wave events (HWEs) and 222 heat wave days (HWDs) recorded during the summer (June–August) period from 2000 to 2024. During the severe 2024 summer heat wave episodes in the south-eastern part of Romania, a rise of 5.89 °C in the mean annual land surface temperature and a rise of 6.76 °C in the mean annual air temperature in the Bucharest metropolitan region were observed. The findings of this study provide a refined understanding of heat stress’s impact on urban vegetation, essential for developing effective mitigation strategies and prioritizing interventions in vulnerable areas.

1. Introduction

Global climate change, associated with increasingly frequent and intense heat wave events that kill thousands of people each year, is considered a global priority for sustainable development in terms of adaptation and mitigation [1,2]. Heat waves, defined as consecutive days of abnormally high temperatures exceeding local or regional norms, are among the most extreme climate events and negatively impact people’s health, urban ecosystems, and infrastructure. HWs can amplify urban heat island (UHI) effects, exacerbating air and land-surface temperatures and triggering major impacts on urban vegetation ecosystems [3,4].

Urban vegetation (green areas, including parks, gardens, and urban forests) is gaining increasing significance in the context of sustainable urban development. As future projections predict, by 2030, the increase in the frequency, duration, seasonality, and intensity of HWs in the Euro–Mediterranean regions will lead to a decrease in summer precipitation rates [5]. As a consequence, in the context of forecasted climate change, the state of urban vegetation and its heat- and water-stress vulnerability will increasingly affect the urban microclimate, and adaptation strategies [6,7] will be adopted. During summer seasons, urban vegetation faced significant complex environmental heat and water stressors over the past 25 years in the south-eastern part of Romania. Thermal stress, exacerbated by the HWs and UHI effects, elevated day–night time temperatures, and changes in evapotranspiration rates, may alter the cooling efficiency of urban vegetation [8]. However, the interaction between air pollution and high air temperature adds a significant complexity to global warming, particularly the alarming rise during compound climate events, such as heat–ozone extremes driven by the increased temperatures in synergy with non-carbon pollutants [9,10].

Several studies investigating the physical mechanisms driving HWs in the Northern Hemisphere found that large-scale atmospheric dynamics and persistent synoptic anticyclones are responsible for prolonged HWs through subsidence, light winds, and warm-air advection [11,12,13,14,15,16,17]. Also, prolonged droughts can decrease soil moisture and increase surface sensible heat fluxes, amplifying HW [18,19,20]. Other studies have found that strong adiabatic warming within anticyclones and increased solar radiation due to clear skies, which trigger increased surface sensible heating, play a crucial role in producing high near-surface temperatures during HWs [21,22]. According to these studies, the surface sensible-heat flux plays a dominant role in increasing the daytime urban planetary boundary layer (PBL) height during HWs, which become deeper and hotter, providing crucial information for extreme-temperature weather climatology [23,24]. PBL climatology, its characteristics, and its relationship with meteorological parameters play a significant role during HWs, being affected by solar radiation and by deeper PBL during summer seasons [25,26]. Another important parameter in describing the radiative properties of the Earth’s surface, urban land surface albedo (LSA), an essential primary climate variable (ECV), is a crucial tool of climate change at the local urban scale [27]. Moreover, climate parameters at both local and regional scales, and their seasonal variability, affect the interaction between heat waves and vegetation [28,29,30,31]. Heat stress induced in vegetation by high air and land surface temperatures in synergy with water stress can impact the possible physiological or biochemical mechanisms of urban green, being a critical limiting factor of vegetation growth and development [32,33,34,35], from the cellular to whole-vegetation scale, reducing photosynthesis, stomatal conductance, growth, reproduction, and leading to premature leaves deaths depending on the heat stress episode’s duration, frequency, and severity [36,37,38,39].

Also, during heatwaves, air pollution negatively impacts urban vegetation health through the retention of particulate matter (PM2.5 and PM10) by dry and wet atmospheric deposition onto leaf surfaces, increasing light reduction, blocking stomata, and reducing photosynthesis [40,41,42]. Absorption of surface-level gaseous air pollutants (nitrogen dioxide—NO₂, ozone—O₃, and sulfur dioxide—SO₂, carbon monoxide—CO, formaldehyde—HCHO, volatile organic compounds—VOCs) can cause chlorophyll degradation and induce oxidative stress in plants [43,44,45]. CO₂ promotes urban vegetation health through the fertilization effect, while O₃ causes oxidative damage to vegetation [46,47]. Urban aerosols exert dual effects on urban vegetation by altering surface radiation for photosynthesis.

Satellite-based remote sensing monitoring of HW events allows the regional-to-global analysis of urban vegetation ecosystem responses to climate change, and may prove helpful in identifying trends in urban vegetation health, for achieving UN SDGs goals 3, 11, and 13: “Ensure healthy lives and promote well-being for all at all ages”, “Make cities and human settlements inclusive, safe, resilient and sustainable”, and “Take urgent action to combat climate change and its impacts [48].

While several studies investigated the role of urban vegetation in air pollution removal capacity [49,50,51,52,53], or mitigating heatwaves and urban heat island influenced by high levels of air pollutants in large cities by conducting extensive air quality measurements and using analytical methods [54,55,56], limited research focused on the response of urban vegetation to summer heat stress and climate variability from a spatiotemporal perspective [57,58,59,60]. The year 2024 was the warmest on record at both the global and European scales, with the global mean surface temperature exceeding the pre-industrial level by more than 1.5 °C, and Europe was the fastest-warming continent at twice the global average [61].

In the Bucharest metropolitan area in Romania, urban vegetation has faced increasingly complex environmental stressors over the past 25 years, including the 2024 summer HWs, which were the warmest on record. This study explored the impact of heatwaves on urban/periurban vegetation in Bucharest from a spatiotemporal perspective during the 2000–2024 period, focusing on summer (June–August) months and heatwave periods in relation to climate seasonality. Time series of in situ monitoring data have been used in synergy with biophysical variables generated by the NASA MODIS Terra/Aqua satellites, the NCEP/NCAR and Giovanni MERRA-2 reanalysis datasets, and analytical processing methods, along with statistical cross-correlation and regression analysis. The present study addresses the need to assess the impacts of multiple climatic parameters on biophysical observables of urban vegetation during the 2000–2024 period, for the interannual and summer seasons (June–August), in the Bucharest city center and the metropolitan area. The main objective of this study was to comparatively quantify the effects on vegetation of the main environmental parameters (mean air temperature at 2 m height (TA) and its maximum value (TAmax), land surface temperature (LST), air relative humidity (RH), wind speed intensity (w), surface solar irradiance (SI), total aerosol optical depth at 550 nm (AOD), land surface albedo (LSA)) on vegetation parameters (normalized difference vegetation index (NDVI), leaf area index (LAI), the fraction of absorbed photosynthetically active radiation (FPAR), evapotranspiration, and net primary productivity (NPP)). The number of summer HW events (HWEs) and the number of HW days (HWDs) per summer season have been considered.

2. Materials and Methods

2.1. Study Area

The study site, Bucharest metropolitan region, is situated in the Romanian Plain, on the banks of the Dâmbovița River, a small tributary of the Danube, and the south-eastern part of Romania, and south-eastern Europe. It is bounded by latitudes 44.33° N and 44.66° N and 25.90° E and 26.20° E longitudes. Its center is placed at 44.4355381° N Latitude and 26.100049° E Longitude (Figure 1). The Bucharest metropolitan area, with approximately 2.2 million inhabitants and a 1811 km² area, includes Bucharest city, the capital of Romania, with 1.8 million people and a 240 km² area, of which urban green infrastructure (forests, parks, small green gardens) covers almost 10% of its area [62].

The periurban areas, distributed across five counties, cover approximately 543 km² and have a population of approximately 0.81 million. The most important relief units are Vlăsiei Plain and Mostiştea Plain, with low altitudes ranging between 5 and 150 m. In the Bucharest metropolitan area, significant human-related impacts associated with urbanization and climate change have altered natural vegetation land cover, reducing its extent and altering its floristic structure and composition, thereby favoring aridization processes. Over the last 30 years, the metropolitan vegetation land cover in Bucharest has decreased, from 4839 ha in 1993 to 4506 ha [63], encompassing multiple urban-to-rural transition areas and comprising different vegetation types (shrubs, grass, forest, crops, etc.). The rapid urbanization of Bucharest after 1990, along with urban expansion, reduced vegetation cover and increased impervious surfaces, contributed to higher air and land surface temperatures associated with HWs and UHIs recorded during the summer (June–August) period. According to the Copernicus Urban Atlas in 2018, land use land cover (LULC) distribution (km²) of different surface land cover classes in the Bucharest metropolis was: 53% agricultural area, 33.6% artificial area, 10.7% natural areas, 2.4% water, and 0.2% wetland [64]. Urban green infrastructure in Bucharest’s metropolis, including periurban forests, provides recognized cultural ecosystem services that are important for enhancing social well-being and resilience [65]. With a diverse landscape, including the spatial distribution of different land cover types and flat plains, Bucharest is one of the most polluted cities in Romania and Europe, ranking 9th out of 96 most polluted cities in Europe in 2022 [66].

With a temperate continental climate and Western European Climate Circulation, Mediterranean Cyclones, and the East-European Anticyclone, Bucharest’s climate is characterized by very hot summers, particularly during frequent heat waves recorded since 2000, and cold, humid winters. The annual mean air temperature belongs to the (10.6 to 11.5) °C range, for periurban to urban areas, while the mean annual rainfall belongs to the (613.2 to 578.9) mm range for periurban to urban areas. Frequent HWs occur during the June–August period, and sometimes extend from May to September. Co-occurrence with UHI events results in this area experiencing significantly higher summer temperatures.

2.2. Data Sources

To analyze interannual spatiotemporal urban vegetation greenness and summer heatwaves—urban vegetation interaction in Bucharest metropolitan area during (June–August) periods of 2000–2024, this study used available observational, satellite remote sensing, and reanalysis data provided by various free sources. The climate and other environmental time series datasets were supplied by different monitoring networks and satellite platforms:

• Meteorological driving factors. Air temperature (TA) at 2 m height and its maximum values TAmax, air relative humidity (RH), air pressure (p), wind speed intensity (w) and direction, planetary boundary layer height (PBL), and solar surface irradiance (SI) have been provided by MERRA-2 Version 2 (Modern-Era Retrospective Analysis for Research and Applications) [67], C3S (Copernicus Climate Change Service) [68], and the online database NASA’s Center (GES DISC) Geospatial Interactive Online Visualization and Analysis Infrastructure (GIOVANNI) V4.28 via its portal [69], as well as from NOAA Physical Sciences Laboratory [70].
• Air Pollution load. Time series of total aerosol optical depth (AOD) at 550 nm were supplied by MERRA-2 Version 2 (Modern-Era Retrospective Analysis for Research and Applications) on the GIOVANNI platform.
• MODIS LST data. From the Terrestrial Ecology Subsetting & Visualization Services (TESViS) Global Subsets Tool at the ORNL DAAC, MODIS Terra/Aqua data [71]. Among the available LST products developed through the different retrieval algorithms based on TIR sensors from different satellite missions (MODIS, LandsatTM/ETM+/OLI, AVHRR, SENTINEL, AMSR-E, AATSR, VIRR), this study considered MODIS Terra/Aqua to be a suitable data source for LST monitoring due to its high observation frequency, moderate spatial resolution, and free availability [72,73]. DAAC, MOD11A2 LST_Day_1 km and MOD11A2 LST_Night_1 km collected within 8 days [74], provide time series LST data for the Bucharest metropolitan area. The root mean square error (RMSE) of the MODIS LST data is within 2.0 °C and exhibits high accuracy in the major global cities [75].
• Vegetation MODIS NDVI data. MOD13Q1 MODIS/Terra Vegetation Indices NDVI/EVI 16-Day L3 Global 250 m SIN Grid V06116-day, MODIS NDVI data, from TESViS.
• Vegetation MODIS LAI/FPAR data. MOD15A2H MODIS/Terra Leaf Area Index/FPAR 8-Day with 500 m spatial resolution, mainly for their capacity to detect anthropogenic and climate impacts on urban vegetation land cover changes, from TESViS.
• Vegetation Evapotranspiration MODIS data. MOD16A2ET_500 m, at 8-day data for evapotranspiration monitoring, from TESViS.
• Vegetation Annual Net Primary Production data. MOD17A3HGF v061, MODIS/Terra Net Primary Production Gap-Filled Yearly L4 Global 500 m SIN Grid, from TESViS.
• Land Surface Albedo MODIS LSA data. MCD43A1 MODIS/Terra + Aqua BRDF/Albedo Model Parameters Daily L3 Global—500 m V061, from TESViS.

2.3. Methodology

This study aims to provide insight into the spatiotemporal evolution of the urban vegetation ecosystem in one of the most dynamic urban–rural systems in south-eastern Europe, namely the Bucharest metropolitan area. Our analysis focused on annual and summer (June–August) datasets spanning the entire period of 2000–2024. For this research, a heat wave analysis has been performed using time-series data from the Copernicus GIOVANNI platform. A robust methodology was used to investigate, in a comprehensive manner, the complex interplay between summer intense HWs and heat stress on urban green land cover, in relation to environmental variables, incorporating satellite and in situ time-series datasets and an approach less commonly employed in urban vegetation contexts.

In climate science, HWs are defined as extended periods of excessive heat [76,77], characterized by different measures for each of these features (intensity, maximum amplitude, duration, frequency, and spatial extent) [78,79,80]. This study uses the definition of HWs as persistent periods of at least 3 consecutive days of high air temperature anomalies, during which the daily maximum TA exceeds the 90th percentile in a given region [81]. Also, as a main characteristic of HWs, the heat wave duration (HWD) is defined as n, the total number of heat wave days in the event [82].

Cross-correlation statistical analysis was adopted for the similarity between two time series data of the outdoor daily, monthly or annual averages of meteorological multiparameters (TA, RH, SI, PBL, LST, LSA), and vegetation parameters (NDVI, LAI, FPAR, ET, NPP) in Bucharest capital of Romania (both center defined by a 6.5 km × 6.5 km area, and metropolitan area defined by a 4.5 km × 40.5 km area). NDVI values range from −1.0 to 1.0, representing vegetation for positive values; clouds, water, and snow for negative values; and bare soils and rocks for values close to zero [83]. When vegetation dries out or becomes sick, its color usually turns from green to brown. This change is particularly significant in the near-infrared region of the spectrum, where absorption increases when vegetation is damaged. Therefore, NDVI can accurately detect chlorophyll and is related to plant health [84].

Cross-correlation analysis, nonparametric test coefficients, trend analysis, and linear regression have been used to assess the dependence between dependent and independent variables in the time series of mean environmental parameters. The strength of relationships between variables was described by linear regression values, ranging from 0 to 1, with values closer to 1 signifying stronger connections.

Because climate observables exhibit non-normal distributions during extreme events such as HWs, the Kolmogorov–Smirnov test of normality has been applied to assess the normality of daily or monthly mean time-series datasets [85]. This method is efficient for analyzing the diverse characteristics of vegetation phenology across different spatio-temporal scales of summer seasonal variation, and especially in complex urban environments. The Spearman rank correlation coefficient, r, ranging from −1 to 1, with higher absolute values indicating stronger correlations, is used to measure the strength and direction of both linear and nonlinear relationships between variables. Significant correlations correspond to a p-value between 0.01 and 0.05. These statistical methods were applied to investigate associations between independent variables (explanatory variables), such as NDVI, LAI, FPAR, ET, NPP, and AOD, and the dependent variable (response variable), which represents TA, LST, LSA, HWE, and HWD. The main novelty of this research is the introduction of correlations between urban vegetation and environmental parameters under HWs, which can help understand trends in climate change and temperature variation in the present and future periods. To determine the statistical significance of the correlation, we used the p-value (p < 0.01). ORIGIN 10.0 software version 2021 for Microsoft Windows, and ENVI 5.7 were used for data processing. All time series datasets were preprocessed using the TIMESAT 3.2 software. Several preprocessing levels were tested, including smoothing and phenometrics computation.

3. Results

Climate change intensifies heat stress on urban vegetation ecosystems, creating an urgent need to understand how climate variables influence vegetation health. The present study explored NDVI, LAI, FPAR, and NPP urban vegetation health biophysical variables’ temporal relationship with different climate drivers, such as TA, TAmax, LSA, PBL height, RH, w, SI, and AOD, for a period spanning twenty-five years between 2000 and 2024 in the Bucharest metropolitan area. In a comparative analysis, the heat-stress impact of summer HWs on urban vegetation was examined from June to August.

3.1. Spatiotemporal Patterns of the Urban Vegetation During 2000–2024 in Bucharest

Urban vegetation NDVI daily means exhibit annual patterns in Figure 2 (2000–2024 at 16-day intervals), with clear seasonality: maximum values during summer and minimum during winter.

For the analyzed period, the daily mean NDVI registered values of (0.2736 ± 0.11848) in the range (−0.00273 to 0.47214) for Bucharest center, defined by a (6.5 km × 6.5 km) area, corresponding to classified as slightly dense and stressed vegetation. At the Bucharest metropolitan scale, MODIS-derived NDVI registered a daily mean value of (0.42705 ± 0.15913) in the range (−0.0386 to 0.6922) for the defined area of (40.5 km × 40.5 km), corresponding to classified as moderately dense vegetation. During all seasons, at the metropolitan scale, LAI recorded a daily mean value of (0.91962 ± 0.68978) in the range of (0.1–2.8473), while FPAR recorded a daily mean value (0.3688 ±0.1761) in the range of (0.0734–0.7392).

The results of this analysis also suggest stress vegetation in both central and metropolitan areas and reveal a range from sparse to rich periurban vegetation influenced by diverse urban microclimates. The time of maximum greenness is similarly distributed from early June to early August, findings which are consistent with previous studies [86,87].

Annual computed mean time-series of the NDVI vegetation index show an overall increasing trend from 2000 to 2024 (Figure 3), explained by climate change at mid-latitudes, and therefore, vegetation appears more active. Similar results were reported in the scientific literature [88,89]. Urban vegetation satellite monitoring in the central part of the city and metropolitan area is essential for efficient planning and management, especially in the south-eastern part of Romania, where HW events are more frequent.

3.2. Variability in Heatwaves and Urban Vegetation Fluctuations over Summer (June-August) Seasons of 2000–2024

Summer 2024 recorded the most extreme and longest heatwaves in the South-eastern part of Europe, especially in July 2024, with persistent and stable drought conditions mostly in the Balkans and Romania [90,91]. From the NOAA NCEP/NCAR reanalysis Europe map in Figure 4, it is clear that an extreme high-air-temperature composite anomaly of more than 3.5 °C at 1000 mb was observed over South-eastern Europe and Bucharest, marked by a star during the summer months of 2024 [90].

Between 2000 and 2024 summer (June–August) seasons, in Bucharest, a total of 41 independent summer heat wave events (HWE) were identified, corresponding to 222 heat wave days (HWD) with considerable variability in both duration and intensity, of which 2024 exhibited the strongest hot period with 5 HWE and 43 HWD. Neither duration nor intensity exhibited significant temporal trends. The most severe HWs associated with higher mean values of air TA max, lower mean air relative humidity RH, and lower values of mean vegetation indices NDVI, LAI, FPAR, and ET have been recorded in Bucharest during mostly summer seasons of 2000, 2012, and 2024 (

Table 1. The most significant heatwaves recorded during the summer (June–August) periods in the Bucharest metropolis, with their main characteristics and corresponding mean values of TAmax, RH, and vegetation indices (NDVI, LAI, and FPAR).

Year (June–August)	HWE	HWD	Mean TA Max (°C)	Mean RH (%)	Mean NDVI	Mean LAI	Mean FPAR
2000	5	22	32.77	38.17	0.43805	1.11454	0.41469
2012	5	29	33.37	44.21	0.51408	1.57357	0.5270
2024	5	43	34.89	44.06	0.4313	0.98797	0.40939

).

Lower values of vegetation parameters recorded in 2000, 2012, and 2024 may indicate that stress induced by higher maximum air temperatures, in synergy with water stress during HWs, can affect the physiological or biochemical mechanisms underlying urban green infrastructure, a critical limiting factor for vegetation growth and development.

The increasing trends of the summer annual mean HWD and TAmax over the 2000–2024 period are clearly illustrated in Figure 5.

3.3. Planetary Boundary Layer Height During Summer (June–August) Seasons of 2000–2024 and Its Impact on Urban Vegetation

Although large-scale persistent synoptic circulation anomalies are considered key factors in generating HWs, urban vegetation–atmosphere interactions and feedbacks seem to play a crucial role in HW intensity. During the summer (June–August) 2000–2024 periods, PBL height in Bucharest displayed a mean of (1170.81 ± 146.20 m in the range of (900.26–1547.63) m, and was positively correlated with daily mean TA, TAmax, and LST. A good correlation between PBL height and air temperature at 2 m height indicates (r = 0.55; p < 0.01), while better positive correlations are found for TAmax (r = 0.75; p < 0.01) and daily mean land surface temperature LST (r = 0.76; p < 0.01). However, the high temperatures during the HWs extend to a higher range, indicating stronger thermal convection and a higher PBL during these periods, results consistent with other studies [25,92]. Also, compared with the same months during 2000–2024, HWs significantly increased PBL height during the June–August period, results similar to those of other studies [93,94].

A strong anticorrelation between PBL height and key biogeophysical variables characterizing vegetation structure and functioning (NDVI, LAI, FPAR, ET, and annual mean NPP) in the Bucharest metropolitan area is shown in

Table 2. Spearman rank correlation coefficients and p-values between the monthly means of vegetation variables and the main climate variables in the Bucharest metropolis during the June–August analysis period (2000–2024).

Vegetation/ Environmental Factors	PBL (m)	TA (°C)	TAmax (°C)	LST (°C)	LSA	RH (%)	SI (W/m2)	w (m/s)	AOD	HWD
NDVI	−0.71 *	−0.76 *	−0.85 *	−0.83 *	−0.41 **	0.71 *	−0.68 *	−0.29 **	−0.72 *	−0.67 *
LAI	−0.74 *	−0.58 *	−0.76 *	−0.82 *	−0.27 ***	0.72 *	−0.63 *	−0.31 **	−0.71 *	−0.68 *
FPAR	−0.79 *	−0.66 *	−0.86 *	−0.86 *	−0.30 ***	0.78 *	−0.68 *	−0.30 **	−0.74 *	−0.68 *
ET (kg m−2 s−1)	−0.95 *	−0.55 *	−0.80 *	−0.28 ***	−0.37 **	0.87 *	−0.61 *	−0.42 *	−0.85 *	−0.60 *
NPP (g C m−2 year− 1)	−0.77 *	−0.61 *	−0.80 *	−0.79 *	−0.10 ***	0.82 *	−0.57 *	−0.30 ***	−0.69 *	−0.64 *

Note: NDVI (normalized difference vegetation index), LAI (leaf area index), FPAR (photosynthetically active radiation), NPP (net primary production), PBL (planetary boundary layer height), TA (air temperature at 2 m height), TAmax (maximum air temperature at 2 m height), LST (land surface temperature), LSA (land surface albedo), RH (air relative humidity), SI (surface solar irradiance), w (wind speed intensity), AOD (total aerosol optical depth at 550 nm), HWD (total number of HW days per summer study period); * and ** indicate p < 0.01 and p < 0.051, respectively, and *** indicates p > 0.05.

and Figure 6. During the summer (June–August) 2000–2024 periods, PBL height in Bucharest displayed a mean of (1170.81 ± 146.20 m in the range of (900.26–1547.63) m, and was inversely correlated with vegetation observables NDVI (r = −0.71; p < 0.01), LAI (r = −0.74; p < 0.01), FPAR (r = −0.79; p < 0.01), and ET (r = −0.95; p < 0.01). It was also reflected in the annual mean NPP.

3.4. Land Surface Albedo Variability and Impact on Vegetation in Bucharest Metropolitan Region

This study analyzed the spatiotemporal variations in urban land surface albedo in the Bucharest metropolitan area in Romania from 2000 to 2024, and assessed the relationships between the spatial and temporal patterns of the albedo and associated influencing factors (vegetation, air pollution, and climate parameters) on a seasonal and interannual basis using satellite products and reanalysis data.

As Figure 7 shows, the monthly mean LSA fluctuated over the 2000–20,024 period and exhibited a clear downward trend after 2020, which may explain the observed climate warming in the Bucharest metropolitan region. The derived MODIS multiannual mean LSA in the Bucharest metropolitan area in the 2000–2024 analyzed period was (0.2547 ± 0.04962), in the range of (0.13467–0.50673) for all seasons.

Also, Spearman correlation coefficients between LSA and NDVI in Bucharest for all seasons during the entire analyzed period 2020–2024 show negative correlations for both the center city/metropolitan area, as follows: r = −0.56 (p < 0.01)/r = −0.36 (p < 0.01).

The 25-year albedo showed a weak decline trend, mainly due to a significant summer (June–August) reduction, reaching 30% at (0.17833 ± 0.0013) in the range (0.17493–0.18135). This research found an inverse correlation between LSA and NDVI, which explains the essential role of urban green infrastructure in reducing land surface temperatures, particularly during hot summers, heat waves, and urban heat islands.

As a consequence, the reduction in LSA enhanced TA, a result that may explain the high summer temperatures of 41 HWE and 222 HWD recorded during the summer (June–August) from 2000 to 2024. A clear anticorrelation was found between LSA and vegetation ET (r = −0.37; p < 0.05) during June–August seasons (Table 2). Also, significant decreases in LSA and ET were found throughout the summer (June–August) over a 25-year study duration (Figure 7).

Lower land surface albedo values recorded after 2020 in the Bucharest metropolitan region are associated with greater absorbed solar radiation and higher air and land surface temperatures, while reduced urban vegetation land cover and increased impermeable surfaces decrease ET. It leads to a higher proportion of energy being converted into sensible heat, driving higher temperatures and enhancing urban heat stress during HWs and the urban heat island effect.

Table 2 shows that the decreasing trends in ET and the strong anticorrelation between air temperature TA and its maximum TAmax, as well as between TA and evapotranspiration ET during summer HWs of the 2000–2024 period, are associated with reduced evaporative cooling effects, likely due to stomatal closure at high air temperatures. This effect will diminish vegetation’s key function in mitigating urban heat stress.

This study also found an anticorrelation between LSA and urban NDVI (r = −0.41; p < 0.05), in good agreement with other studies exploring LSA variability and its impact on vegetation, which found that over vegetation-covered areas, interannual variation in albedo was more sensitive to NDVI [92,93].

Our results reveal a stronger negative correlation between LSA and LST (r = −0.80; p < 0.01) during the summer time in both Bucharest city areas with dominant impervious surfaces, hurting the urban thermal environment. Broadband albedo, which measures urban surface properties, also depends on the atmospheric conditions.

3.5. Air and Land Surface Temperature Variability and Their Impacts on Urban Vegetation During Summer Seasons

Daily and monthly mean surface air temperature (TA) and its maximum value, TAmax, measured at around 2 m above the surface, have significant impacts on both human and natural ecosystems, affecting urban vegetation and biodiversity, agriculture, and energy demand. Due to the synergy of HWs and UHIs during hot summers, this study found in Bucharest’s center, with increased artificial built areas, higher temperatures at 2 m height of 3.6 °C compared to the adjacent, less urbanized periurban zones. The findings of this study support the need to increase urban green spaces, such as street trees, parks, and green roofs, to provide shading and evaporative cooling and mitigate urban heat island effects, especially during heat waves in the Bucharest metropolis.

Also, as one of the most important Earth System variables and a key parameter in urban climate research, land surface temperature (LST) reflects the interaction between solar radiation and the land surface, influencing thermal radiation from the vegetated areas or the ground. In urban areas, LST provides a crucial indicator of HWs and UHIs’ effects, vegetation ecosystem health, energy consumption, and responses to climate change [94,95]. Also, this study analyzed the interannual variations in LST data and their relationships with air temperature, AT. The rank correlation analysis shows that, at the pixel-scale, during summer seasons (June–August) 2000–2024, air temperature at 2 m height, TA, and LST present a strong positive correlation (r = 0.96; p < 0.01). As the physical climate system is highly sensitive to changes in land surface albedo, urban/periurban vegetation systems could significantly feed back into projected climate change scenarios. As Table 2 shows, during the June–August summer analysis period from 2000 to 2024, TA, TAmax, LST, and HWD display strong inverse correlations with the vegetation variables NDVI, LAI, FPAR, and NPP, which means that urban vegetation’s growth is limited during persistent HWs.

Table 2 and Figure 8 clearly show existing anticorrelations between TA, TAmax, LST, and ET, and LSA. Computed for the Bucharest metropolitan area the monthly mean TAmax recorded during the summer (June–August) 2000–2024 periods was (31.06 ± 1.64) °C in the range of (27.36–34.89) °C, while evapotranspiration ET registered very low mean value of (3.02654 × 10⁻⁵ ± 9.02518 × 10⁻⁶) kg m⁻² s⁻¹ in the range of (8.99333 × 10⁻⁶–4.58667 × 10⁻⁵) kg m⁻² s⁻¹.

This finding confirms a decrease in urban vegetation’s cooling effect during summer heat stress associated with HWs and higher temperatures. However, urban vegetation’s evapotranspiration (ET) has significant cooling effects, helping mitigate the summer urban thermal environment and reduce heat exposure risks in large metropolitan areas [96,97]. Despite increased air temperatures during HWs, urban vegetation continued to cool through evapotranspiration, but further studies are needed to identify the best-suited vegetation species to optimize cooling.

3.6. Air Relative Humidity, Surface Solar Irradiance, and Aerosol Optical Depth Impact on Urban Vegetation During Summer Seasons

This research found that higher values of air relative humidity RH enhance vegetation’s growth and evapotranspiration, and surface heat exchange, thereby strengthening vegetation’s cooling effect. This finding is consistent with other studies [98,99,100,101,102,103]. Knowing the relationship between evapotranspiration and heat stress during HWs is essential for effective water resource management, irrigation planning, and climate change assessments in urban vegetation systems [104,105,106]. As key thermal and water-related factors influencing urban microclimates, air relative humidity and solar radiation have high impacts on urban vegetation growth during summer hot periods.

The results of this study revealed that among several climate factors during frequent HWs recorded in the summers 2000–2024 in the Bucharest metropolitan region, solar surface irradiance SI was an important influential factor limiting the urban vegetation growth. From Figure 9 and Table 2, it is evident that increased solar radiation during heatwaves affects urban vegetation growth, as indicated by lower values of NDVI, LAI, and FPAR.

Figure 9 and Table 2 clearly show decreasing trends in the monthly mean summer (June–August) time series of urban vegetation parameters NDVI, LAI, FPAR, and ET, associated with decreasing air relative humidity and increasing solar radiation during heatwaves, which affect urban vegetation growth.

During the summers 2000–2024 period, the total aerosol load over Bucharest metropolitan area AOD at 550 nm had an increased monthly mean value (0.26413 ± 0.05526) in the range (0.19767–0.39567), being negative correlated with all monthly mean vegetation indices NDVI (r = −0.61; p < 0.01), LAI (r = −0.71; p < 0.01), FPAR (r = −0.74; p < 0.01), ET (r = −0.83; p < 0.01). Positive correlations between AOD and air temperature TA (r = 0.59; p < 0.01), TAmax (r = 0.76; p < 0.01), and HWD (r = 0.73; p < 0.01) confirm that increased air pollution in synergy with air temperatures during summer HWs may have negative impacts on urban vegetation health. Environmental parameters primarily govern the diurnal and seasonal variations in vegetation characteristics, including NDVI, LAI, FPAR, NPP, and ET, in Bucharest’s metropolis. According to Table 2, higher levels of air temperature at 2 m height (AT), ATmax, land surface temperature (LST), planetary boundary layer height (PBL), wind speed intensity (w), surface solar irradiance (SI), and total aerosol optical depth at 550 nm (AOD) decrease urban vegetation’s cooling effect during heatwaves and vegetation’s growth. Long-term changes in vegetation are essential indicators for understanding the response of ecosystems to climate change. NDVI and other vegetation parameters could be used for urban ecosystem monitoring and as an additional criterion for assessing the impact of extreme climatic events.

In a comparison with Table 2,

Table 3. Spearman cross correlation coefficients between monthly mean NDVI variables in Bucharest center city/metropolitan area, and main climate variables during autumn–winter–spring seasons, September–May analysis period 2000–2024.

NDVI Vegetation//Environmental Factors	PBL (m)	TA (°C)	TAmax (°C)	LST (°C)	LSA	RH (%)	SI (W/m2)	w (m/s)	AOD
NDVI Bucharest Center area (6.5 km × 6.5 km)	0.78 *	0.86 *	0.74 *	0.89 *	−0.58 **	−0.65 *	0.85 *	−0.24 **	−0.63 *
NDVI Bucharest metropolitan area (40.5 km × 40.5 km)	0.62 *	0.52 *	0.51 *	0.67 *	−0.35 **	−0.47 *	0.67	−0.29 **	−0.51 *

Note: NDVI (normalized difference vegetation index), LAI (leaf area index), FPAR (photosynthetically active radiation), NPP (net primary production), PBL (planetary boundary layer height), T (air temperature at 2 m height), TAmax (maximum air temperature at 2 m height), LST (land surface temperature), LSA (land surface albedo), RH (air relative humidity), SI (surface solar irradiance), w (wind speed intensity), AOD (total aerosol optical depth at 550 nm); * and ** indicate p < 0.01 and p < 0.05, respectively.

shows that while during summer seasons, vegetation greenness indicators NDVI were inversely correlated with TA, TA max, LST, PBL heights, and SI, during September–May months, corresponding to autumn–winter–spring seasons of 2000–2024, these turned to positive correlations in both the Bucharest center area as well as at the metropolitan scale. Also, opposite correlations were found for air relative humidity RH, which was positively correlated with NDVI during summer seasons and negatively correlated during autumn–winter–spring seasons. In all seasons, land surface albedo LSA, total aerosol optical depth (AOD), and wind speed intensity (w) were negatively correlated with all vegetation parameters, highlighting their damaging effects on urban vegetation.

3.7. Hot Summer 2024 Analysis

Due to rapid urbanization and extensive infrastructure development, Bucharest is highly vulnerable to extreme heat events [107]. By tracking NOAA positive temperature anomalies relative to the 1991–2020 climatology, Figure 3 shows the extreme hot summer of 2024, when temperatures exceeded 3.5 °C. Cumulative positive anomalies highlight the intensity and duration of extreme heat events over time, with high values suggesting the development of severe heat wave conditions. The temporal variations in the monthly mean air temperature at 2 m height (TA), its maximum mean value (TAmax) and its minimum mean value (TAmin) in the Bucharest metropolitan area during the summer (June–August) 2024 period are illustrated in Figure 10 that display very high temperatures (daily mean TAmax = 34.89 °C, daily mean TA = 27.11 °C, and daily mean TAmin = 20.07 °C especially during July observed when some typical heat wave processes took place. This period was important due to 5 significant heat wave events, totaling 43 HW days. Urban vegetation characteristics, NDVI, LAI, and FPAR in the Bucharest metropolitan area registered lower values compared with other summer periods when there have not been HWs (Figure 5).

Also, during the severe 2024 summer heat wave episodes in the south-eastern part of Romania, this study identified increases of 5.89 °C in mean annual LST and 6.76 °C in mean annual TA in the Bucharest metropolitan region.

Understanding the relationship between HWs’ interaction with urban vegetation in the Bucharest area, in a comparative analysis for both the city center and the metropolitan area, and urban heat during HWs is crucial for mitigating the thermal environment and improving environmental quality.

4. Discussion

Successive record-breaking summer temperatures, both globally and in south-eastern Europe, including Romania, raise the urgent question of how to better protect vulnerable people’s health and urban vegetation ecosystems. Also, compound drought–air pollution–heat wave events are associated with several risks, including vegetation fire risk during summer extreme hot temperatures [85]. Several recent studies analyzed the impacts of HWs on vegetation in large metropolitan areas.

As essential metrics for assessing urban vegetation greenness and health from satellite sensors, the normalized difference vegetation index (NDVI) indicating photosynthetic activity, is a strong indicator for essential climate variables (ECVs) like as leaf area index (LAI), and fraction of absorbed photosynthetic radiation (FPAR) provided by satellite images, which show their condition based on how much light of different wavelengths plants absorb and reflect without in situ measurement data. In urban ecosystems with high spatial heterogeneity, vegetation responses to climate and air pollution variability at the city scale exhibit diverse spatiotemporal patterns. As a global index used to analyze the characteristics of vegetation land cover and its dynamic spatiotemporal patterns, NDVI is a key parameter for assessing vegetation stability and its complex interaction with climate and anthropogenic stressors. LAI and FPAR indicators were used to characterize the effects of climate parameters and urbanization on vegetation health and growth, and to assess their changes. Urban vegetation phenology in LAI is the primary driver of seasonal variation in FPAR, which, in turn, drives photosynthesis. While FPAR absorbed by urban vegetation is a key biophysical variable for monitoring vegetation growth, productivity, and the energy balance of urban ecosystems, evapotranspiration (ET) is a key process in the water cycle. Moreover, monitoring variations in these indices over time allows scientists to track seasonal development, detect early signs of stress, and plan interventions before damage becomes critical. During summer hot periods, a sudden decline might indicate drought stress, disease or pest outbreaks, nutrient deficiencies, or harvesting.

Urban vegetation in the Bucharest metropolitan area is sensitive to climate change, being significantly affected by variations in climate parameters and exhibiting considerable differences in distribution patterns during summer heat waves and during the autumn–winter–spring seasons. However, during the 2000–2024 period, as reported in other studies in Europe [102,103,104,105,106], vegetation greenness in Bucharest’s metropolitan area, as measured by NDVI, showed a slight increasing trend after 2000, with 2000 identified as a turning point. This was explained by climate change in European mid-latitudes, and therefore, vegetation is observed to be more active.

There is an important interplay between deeper and hotter planetary boundary layer PBL, especially during the daytime HWs, and its synergies with the enhanced urban heat island (UHI) effect, and urban vegetation, which can alter how heat is mixed and how it interacts with large-scale synoptic weather patterns, affecting turbulence and air quality [107,108,109]. It was demonstrated that heatwaves are driven by persistent large-scale circulation anomalies, and that land–atmosphere interactions intensify and propagate them. In urban areas, various vegetation types play a vital role in land–atmosphere interactions by modulating energy and water exchange through different pathways. Similar to other studies, the present research found a strong anticorrelation of urban greenness during the summer seasons with PBL heights between 2000 and 2024 in Bucharest [110,111,112].

While HWs are significantly associated with high air and land surface temperatures, it is also crucial to examine changes in the other meteorological factors, especially air relative humidity, wind speed intensity, and solar surface irradiance, which are critical controls of urban thermal comfort and heat stress on vegetation [113,114,115,116]. A serious impact on local and regional climate due to the alteration of the effective surface albedo, a change in land surface albedo can have a positive (cooling) or negative (warming) effect on climate change [117,118].

Land surface albedo (LSA), air temperature at 2 m (TA), and land surface temperature (LST) are thermophysical parameters that define the urban thermal environment, including urban heat islands (UHIs) and heat waves (HWs) [119,120]. These parameters are correlated such that materials with low albedo (low reflectivity of solar radiation) exhibit higher heat absorption and, consequently, higher TA and LST values, intensifying the effects of UHI and HWs. While Bucharest center predominantly consists of impervious surfaces such as roofs, parking lots, streets, and sidewalks, which have low albedo, absorbing significant amounts of heat, in contrast, periurban metropolitan counties with increased vegetation land cover mitigate heat through natural shading and evapotranspiration. However, the ongoing conversion of green spaces with concrete and asphalt exacerbates the effects of urban heat stress on the vegetation ecosystem [121,122]. However, long-term vegetation dynamics that significantly impact the climate are reflected by the dramatic variations in urban land surface albedo. The slight decline in land surface albedo in the Bucharest metropolis due to increased urbanization may explain the enhanced air temperatures and the severity of HWs, as evidenced by 41 heat wave events and 222 heat wave days recorded during the summer (June–August) period from 2000 to 2024. Bucharest, densely packed with heat-retaining concrete and limited green space, is highly vulnerable to summer heatwaves and the urban heat island (UHI) effect. Land surface albedo and heat waves in cities are closely linked because low-albedo urban surfaces (like dark, impermeable materials) absorb more solar radiation, significantly increasing urban heat island (UHI) effects and intensifying heat waves. Conversely, increasing the albedo of urban surfaces—making them more reflective—is a key strategy to reduce daytime heat absorption, lower surface temperatures, and mitigate the negative impacts of heat waves on urban environments and human thermal comfort [123,124].

Quantifying the impact of TA, TAmax, and LST during HWs on urban vegetation in the highly heterogeneous metropolis like Bucharest is challenging. However, this research found that the normalized difference vegetation index (NDVI) is the most significant indicator of urban vegetation affecting LST with a significant cooling effect, particularly in summer. The total effects of all vegetation indicators are anticorrelated with LST, with the greatest impact in spring and the least in winter. Also, seasonal metropolitan morphology has a high impact on the urban thermal environment during HWs [125,126].

The interplay between urban heat environments and air pollution is deeply interconnected; their synergistic effects, associated with stagnant synoptic conditions and lower PBL heights, reduced wind speeds, and reduced turbulent flux, reduce the dispersion of air pollutants [127].

Like other scientific publications [128], this study found that heat waves and high air temperatures can severely stress urban vegetation in Bucharest, thereby reducing the cooling effect it provides to the metropolis. This effect is explained by vegetation stomatal closure, which conserves water and reduces evapotranspiration, thereby diminishing its cooling effect. However, this stress cycle can damage vegetation health, limit its growth, and make cities more vulnerable to extreme heat. It was demonstrated that local and regional background climate conditions, vegetation types, and spatial configuration influence the cooling effectiveness of urban vegetation [129]. In densely built-up urban areas with limited vegetation cover, there is a significant lack of summer thermal regulation during the day, especially in the early morning and late evening, thus exacerbating thermal stress. Other studies considered that thermal impacts of heat waves vary by urban green space type, with broad-leaved forests showing the strongest cooling [130]. Also, urban impervious-surface expansion reduces ET, offsetting the increase from anthropogenic heat. The replacement of natural landscapes like forests and grasslands by built impervious surfaces is leading to changes in meteorological factors, such as air temperature, humidity, precipitation, and wind speed near the ground in urban areas, and further affecting the energy and water balance. An additional heat load during heat waves must account for the heat generated by densely populated areas and frequent human activities. However, the mechanisms linking urban vegetation to climate drivers are very complex at temperate latitudes; therefore, a targeted investigation is required to detect possible trends in evapotranspiration and in urban surface morphology and roughness, to support understanding of the dynamics of urban vegetation during heat waves. Several studies found that heat stress negatively affects vegetative photosynthesis, thereby affecting tree growth and, eventually, survivability [122,123,124].

The mechanisms underlying the interactions among urban vegetation health and growth, heat stress, and HWs are of immense importance for assessing vegetation cooling effects and developing mitigation measures. Another aspect to consider is the existing differences among plant and tree species in their tolerance and adaptation to climate change [82,84,125]. Understanding the interplay between summer heatwaves, urban vegetation, air pollution, and climate variables is essential for developing evidence-based public health strategies and mitigating environmental health risks.

Like other studies in the field [126,127,128], this study has revealed the mechanism of a bidirectional influence between environmental and climate factors and urban vegetation greenness, emphasizing the green space protection needs during heat wave periods in reducing air pollutant emissions, promoting environmental technology innovation, and safeguarding ecological measures

Study Limitations and Future Directions

Despite providing valuable insights into the impact of summer heatwave variability on Bucharest metropolitan vegetation, this study has several limitations that underscore the need for a more complex approach. Our results might not fully align with those spatial and contextual variations. Also, MODIS Terra/Aqua datasets have coarse spatial resolution; higher-resolution imagery, such as Worldview-3 (WV-3) satellite imagery [129] or ESA Copernicus Sentinel imagery, will be preferred in urban greenness studies.

Recent studies highlight that urban vegetation (forest trees, parks, yards, streets, etc.) structure and vertical morphology, particularly mean tree height, canopy aggregation, and spatial continuity, can have a greater impact on LST than total green land cover, with effects modulated by urban form in seasonal LST patterns [130]. Conventional metrics used in this study for urban greenness —NDVI, LAI, and FPAR—derived from MODIS satellite observations do not capture vertical vegetation structure or spatial configuration, offering limited insight into intraurban spatiotemporal HWs—vegetation interactions and their cooling effects during the day.

However, the use of remotely sensed LST and LSA datasets to quantify the heat waves–urban vegetation interaction may introduce uncertainties, because the derived MODIS satellite LST and LSA reflect emitted radiation from the urban surface of an object seen from the point of view of the thermal sensor, which is not a good proxy for the air temperature TA perceived by people, compared to in situ TA monitoring, which is important to be considered for the realistic mitigation strategies of alleviating urban heat stress on vegetation and people during synergies of HWs with UHIs through increasing vegetation land cover [131].

In general, urban vegetation plays a significant role in the interaction between heatwaves, urban heat islands, and vegetation, lowering daytime temperatures, but the magnitude of their cooling effect is notably amplified when using only satellite datasets [132,133]. In the context of increasing frequency of globally deadly heatwaves, declining urban vegetation land cover, and rising imperviousness, which intensify decision-makers’ concerns, there is an urgent need, especially in European Southern cities, to improve the adaptive capacity of biosystems.

One significant hypothesis that can be made is that the conditions for vegetation growth and health during hot summers and heat waves have deteriorated (i.e., high air and land surface temperatures, low precipitation, and heat stress), caused by climate change at mid-latitudes, and, therefore, vegetation can be observed from satellites to be more affected. Similar patterns have been observed in the other parts of the World [134,135,136], and further studies are needed to confirm and characterize this change across the different biogeographical regions and urban landscapes.

These insights are essential for developing human and alth urban vegetation health adaptation strategies during summer heat stress and informing sustainable metropolitan environmental management practices in the face of ongoing climatic changes [137,138,139,140].

5. Conclusions

This study aims to explore the response of urban vegetation to the impacts of heat waves in relation to multi-climatic parameter variability, from a spatiotemporal perspective during 2000–2024, with a focus on summer HWs (June–August), particularly the hottest summer of 2024 in the Bucharest metropolitan area, Romania. In the frame of worsening heat wave severity under future warming, this research revealed the interannual and inter-seasonal relationships between HWs and urban vegetation parameters NDVI, LAI, FPAR, ET, and their interplay with environmental variables and background climate conditions, with implications for the management of urban resources.

The results of this study highlight the importance of considering land surface albedo, land surface temperature, aerosol optical depth, surface solar irradiance, and other climate variables when assessing mitigation strategies to reduce summer extreme heat in urban areas. In the frame of sustainable urban development, optimizing the structure of green space and the built environment in cities helps mitigate the adverse impacts of extreme summer heat on public health and the environment.

Urban heat stress–vegetation-induced biophysical feedbacks have been considered for moderating increases in land-surface air temperature and reducing warming during summer heat waves and under urban heat island effects. Recognized for the cooling effect associated with changes in surface convection and reduced transmissivity of shortwave radiation, evapotranspiration from urban vegetation has shown high spatiotemporal variability in response to the urban thermal environment in Bucharest. The impacts of heat waves on urban vegetation evapotranspiration during hot summers limit the cooling benefits of urban vegetation. In such conditions, the preservation of urban ecosystems becomes a priority for urban decision-makers, who need scientific information to develop mitigation and adaptation strategies for sustainable planning and design of green-blue infrastructure, thereby guaranteeing the conservation of biodiversity in cities.