Public Garden Environmental Factors Impact on Land Surface Temperatures of the Adjacent Urban Areas in an Arid Region

Abstract

Urban growth in hot, arid regions intensifies the urban heat island effect, making green spaces vital for climate mitigation. This research investigates the impact of public gardens on the surrounding urban thermal environment and on the mitigation of the urban heat island (UHI) in a hot arid region. This study selects an important public garden in Biskra, the “5 July 1962” Garden, as a case study of significance at the urban scale. To achieve research objectives, onsite measurement using a digital measurement device (5-in-1 Environmental Meter “Extech EN300”) and satellite remote sensing data from LANDSAT8 are employed, capturing summer measurements of key parameters and indices: Land Surface Temperature (LST), Air Temperature (AT), Relative Humidity (RH), Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI) and Normalized Difference Moisture Index (NDMI). The analysis and correlation of these indices with the LST values allow us to evaluate the zoning and distance impacts of the garden studied. Land surface temperature rises gradually from the garden outward, peaking in the North-East with the strongest heat island effect and remaining lower in the cooler, vegetation-rich South-West. The results reveal that air temperature is the primary driver of land surface temperature (72% impact), while relative humidity (17.3%), vegetation index (7.8%), moisture index (2.9%), and water index (1.7%) contribute to cooling, with vegetation and moisture reducing surface temperatures through shading, transpiration, and latent heat exchange.

1. Introduction

Urban growth, especially in hot and arid regions, has promoted significant modifications to the local climate, affected by UHIs depicted by higher temperatures in cities than in their countryside surroundings [1,2,3]. The spreading out of these effects is due to the substitution of natural vegetation with rigid surfaces such as roads and buildings, which absorb and maintain heat [4,5]. Thus, urban environments face increased temperatures, which increase energy consumption, moderate air quality, and negatively affect human health and comfort [6,7]. In reply to these challenges, public green spaces such as parks and gardens have become necessary elements to mitigate urban heat by improving the local microclimate [3,8,9].

Public gardens have a unique potential to improve urban thermal environments. The vegetation in these gardens adjusts microclimatic conditions by influencing daytime and nighttime temperatures. Several studies have used satellite remote sensing to measure the effects of gardens on adjacent urban thermal conditions [10]. Landsat satellite data have been used to examine the thermal impact of public parks in Shanghai, showing a significant reduction in air temperature around large parks during summer, with cooling effects extending several hundred meters beyond park edges and benefiting surrounding neighborhoods [11,12,13].

Vegetation in public gardens regulates temperature among several processes, involving shading, evapotranspiration, and moisture retention [8,14]. These mechanisms offer cooling surrounding areas, which lower LSTs and create a more comfortable environment [15]. However, the degree and efficacy of these cooling effects vary on several factors, including the size of the green space, vegetation density, species composition, and spatial arrangement of plants [16,17]. The study of the radiative influences of public gardens on adjacent urban areas is mainly relevant in hot and arid regions, where extreme temperatures are common and water availability for vegetation is limited [15,18,19].

Vegetation plays a crucial role in reducing heat retention in hot climates, as green spaces have been shown to lower surface and air temperatures in densely urbanized environments [20]. Urban vegetation, particularly parks and gardens, has been shown to play a significant role in reducing the UHI, with tree canopies and grass-covered surfaces providing notable cooling benefits [21]. In addition, the size, configuration, and vegetation type of public gardens significantly influence their cooling capacity [8,22].

The assessment of public gardens’ impact on the adjacent urban thermal environment has gained increasing understanding in recent years, particularly in the context of UHI mitigation [8,23]. As cities around the world experience fast urbanization and environmental challenges such as augmenting temperatures and diminished green spaces, researchers have engaged in the role of urban vegetation in enhancing microclimates [8,19]. Public gardens, as key features of urban green infrastructure, play a valuable role in regulating thermal conditions through several biophysical procedures [15].

Remote sensing has been used to estimate the cooling influence of Beijing parks, showing that the degree of temperature reduction depends on vegetation type and the park’s proximity to built-up areas. Larger parks with denser vegetation were found to provide stronger cooling effects, especially during peak summer months [24]. Urban green spaces have also been recognized for their social and ecological value, as public gardens contribute not only to thermal regulation but also to enhancing urban sustainability and quality of life [25].

Research in hot and arid areas, where temperatures can reach extremes, has yielded valuable understandings into the role of public gardens in mitigating urban heat. In Phoenix, Arizona, remote sensing analysis showed that urban green spaces, including public gardens, significantly lowered temperatures in surrounding urban areas, particularly during the day [26]. Conversely, the efficiency of these green spaces was induced by the type of vegetation used, with xerophytic species demonstrating more durability in sustaining cooling effects in arid conditions [27,28]. In Cairo, green spaces were found to play a crucial role in cooling urban areas, highlighting the importance of expanding public gardens to mitigate rising temperatures in arid climates [29].

To quantify the thermal influence of public gardens, researchers typically use satellite-derived indices such as the NDVI, which measures vegetation density, and LST, which assesses surface heat emissions [30]. Studies using these indices have provided valuable insights into the thermal dynamics of green spaces. NDVI and LST have been widely used to investigate the relationship between urban vegetation and thermal environments, revealing that areas with higher NDVI values generally exhibit lower LST [30]. In addition, the NDWI and NDMI are habitually used to assess the water content of vegetation, further influencing the thermal behavior of public gardens [31,32]. NDWI and NDMI values have been found to correlate strongly with cooling effects, especially in areas with high evapotranspiration rates [33]. These indices, used in conjunction with LST data, provide a comprehensive understanding of how public gardens modulate the surrounding thermal environment [34].

This study aims to establish a comprehensive garden model that evaluates the impact of its characteristic vegetation and water elements on the LST of the surrounding urban district. The research aims to formulate a simplified yet effective mathematical expression that quantifies the garden’s influence on the thermal dynamics (AT, RH and LST) of the adjacent urban microclimate. By considering remote sensing key environmental variables such as NDVI, NDMI and NDWI, this study intends to provide a predictive tool that urban planners and environmental designers can use to optimize green spaces for improved thermal comfort of the adjacent districts and climate resilience in hot arid cities.

2. Case Study

2.1. Biskra

The eastern gateway to the Algerian Sahara, Biskra is located 470 km southeast of the country’s capital, and is the main town of Wilaya since 1974. Sited at an altitude of 120 m above sea level, it is one of the lowest cities in the country (Figure 1). The harsh desert environment, combined with the city’s exposure to Saharan winds, presents serious challenges for human habitation and agriculture, requiring the use of traditional architectural and water management techniques and numerous public gardens to mitigate the effects of extreme heat.

2.2. Meteorological Data of the Area Studied

A public garden in Biskra, a city in southeast Algeria’s colonial district, serves as the subject of the case study. Biskra, which is situated at 34°51′ N 5°44′ E (34.850° N 5.733° E), has a hot, dry climate with significant seasonal and night-to-day temperature variations. The city falls under the category of BWh, as per the Köppen climate classification, which describes clear skies, strong sun radiation, and extremely low humidity due to infrequent precipitation. Biskra climate data were obtained from the Weather-Atlas website. From Figure 2, summer average temperatures are over 40 °C in July, which is the hottest month, and between 29 °C at night. For January, the coldest month, winter temperatures ranging from 7 °C to 29 °C are available [https://www.weather-atlas.com/en/algeria/biskra-climate] (accessed on 15 July 2024). These arid conditions, combined with frequent exposure to sunlight and high heat gains, expose city areas to intense radiation, leading to extended thermal stress and elevated outdoor temperatures.

2.3. The Studied Garden: The “5 July 1962” Garden

The “5 July 1962” Garden in Biskra, Algeria, has an area of 5.2 ha with a regular layout and linear shape, with a large central path lined with trees (10 m wide), including tree species such as Duranta Erecta, Ficus Microcarpa, Eucalyptus, and some Palms… etc., which contribute significantly to the garden’s thermal environment. In this garden, a high SVF, which varies from 0.10 to 0.30, has significant control of the thermal regulation. The differences in Sky View Factor (SVF) between the interior of the garden, and the adjacent urban fabric play a critical role in regulating solar exposure and, consequently, surface thermal conditions. Within the garden, lower SVF values result from dense tree canopies and vegetation cover, which obstruct a significant portion of the sky dome and reduce direct solar radiation reaching the ground. This shading effect minimizes heat absorption by the soil and surface materials, thereby maintaining lower LST values. In contrast, the surrounding urban areas exhibit higher SVF values due to sparse vegetation and the predominance of impervious surfaces such as asphalt and concrete (Figure 3).

3. Materials and Methods

The methodology for evaluating the impact of the public garden on the adjacent urban thermal environment in a hot arid region implies the use of onsite measurement and satellite remote sensing techniques to measure key thermal and vegetation indices. The study centers on a large public garden in the city of Biskra, Algeria, a region known for its extremely arid climate. To record AT and RH values an onsite measurement is conducted using an instrumentation device, (5-in-1 Environmental Meter “Extech EN300”, Teledyne FLIR, Hoogstraten, Belgium) and to collect LST, satellite data are collected using LANDSAT 8 for the summer period (15 July 2024). The most important parameters studied are LST, NDVI, NDWI, and NDMI [35,36]. LST is employed to measure surface heat emissions in the study area, bringing information on temperature differences between vegetated and non-vegetated areas [37]. NDVI is used to assess vegetation density and health, offering a measure of the ability of gardens to mitigate heat stress through photosynthetic activity [38]. NDWI and NDMI are used to assess vegetation water content and humidity, which are critical factors influencing the cooling effects of gardens [39,40,41,42]. Satellite imagery is treated and analyzed using geographic information system (GIS) software version 10.1, where thermal and vegetation indices are extracted and spatially correlated. Buffer zones are determined around the gardens to examine the thermal impact gradient on adjacent urban zones [16]. By correlating LST, AT, RH values with NDVI, NDWI, and NDMI, the study quantifies the extent and intensity of cooling effects exerted by public gardens by the development of a mathematical formula.

Table 1. Spectral Indices (NDVI, LST, NDMI and NDWI) from Landsat 8 Data Sources (Source: https://landsat.gsfc.nasa.gov/satellites/landsat-8/landsat-8-bands/ (accessed on 10 July 2025)).

Index/Data	Landsat 8 Source	Spatial Resolution	Temporal Resolution	Spectral Resolution
LST	OLI (Bands 4—Red, 5—Near IR)	30 m	16 days (revisit cycle of Landsat 8)	Red Band (0.64–0.67 µm) + NIR Band (0.85–0.88 µm)
NDVI	TIRS (Bands 10 and 11—Thermal)	100 m (resampled to 30 m)	16 days	Thermal Band 10 (10.60–11.19 µm) + Thermal Band 11 (11.50–12.51 µm)
NDMI	OLI (Bands 5—NIR, 6—SWIR1)	30 m	16 days	NIR Band (0.85–0.88 µm) + SWIR1 Band (1.57–1.65 µm)
NDWI	OLI (Bands 3—Green, 5—NIR)	30 m	16 days	Green Band (0.53–0.59 µm) + NIR Band (0.85–0.88 µm)

below shows the different resolutions of the obtained LST, NDVI, NDMI and NDWI.

3.1. Measurement Device

To assess the thermal conditions of the Garden studied under extreme summer climatic conditions, onsite measurements were conducted on 15 July 2024, a representative date for peak seasonal heat. The study aims to systematically assess AT and RH within and around the garden using a digital measurement device (5-in-1 Environmental Meter “Extech EN300”), offering insights into microclimatic variations influenced by vegetation density, spatial configuration, and surrounding environmental factors. The EN300 Measures Humidity (10 to 95%), Temperature (0 to 50 °C), Air Velocity, Light and Sound. Rugged 5-in-1 environmental meter with tripod mount and RS-232 PC interface. The accuracy details provided by the manufacturer’s official hardware specifications are shown in

Table 2. Type K Thermocouple Thermometer and Hygrometer (Humidity/Temperature) Accuracy details (Source: https://mci-global.com/UploadedImg/Product/27122021_101909_AM_EN300_UM-en.pdf (accessed on 10 September 2025)).

Units	Range	Resolution	Accuracy
	Type K Thermocouple Thermometer
°F	−148 to 2372 °F	0.1 °F	±(1% rdg + 2 °F)
°C	−100 to 1300 °C	0.1 °C	±(1% rdg + 1 °C)
	Hygrometer (Humidity/Temperature)
%RH	10 to 95%RH	0.1%RH	<70%RH: ±4%RH ≥70%RH: ±(4% rdg + 1.2%RH)
°F	32 to 122 °F	0.1 °F	±2.5 °F
°C	0 to 50 °C	0.1 °C	±1.2 °C

.

3.2. Measurement Protocol

Data collection was carried out along two primary transects (1. North-West–East-South and 2. North-East–South-West) extending from the garden’s central point to its adjacent districts. Multiple pre-determined measurement stations along these axes facilitated a structured assessment of spatial thermal differences, with their exact locations detailed in the accompanying Figure 4.

The onsite measurement is taken at the hottest time of the day (2 p.m.) in the hottest season, allowing the extreme diurnal analysis of temperature and humidity values. High-precision environmental sensors were used to record AT (°C) and RH (%) at each station, with equipment stationed at a standardized height (1.50 m) to minimize inconsistencies caused by ground heat flux. To ensure data accuracy, collection is conducted by standardized protocols, ensuring consistency and reliability across all measurement points.

This study focused on assessing the garden’s perimeter impact on LST, specifically the park’s cool island effect and its variation with distance and direction, rather than modeling the thermal behavior of building blocks. Measurement stations were strategically placed along the garden edge and in adjacent open spaces to capture this cooling signal while minimizing interference from street shading and building materials. Since the Landsat LST data (100 m, resampled to 30 m) cannot resolve individual building blocks, the methodology was designed to evaluate perimeter-scale influences rather than building-scale thermodynamics.

3.3. Choice of Interpolation Method

Using temperature values at discrete distances in different directions, a Radial Basis Function (RBF) interpolation method is appropriate for spatial continuity. Given the smooth transition of temperature values, RBF interpolation using a Gaussian function is selected [43]. The interpolation function can be expressed as follows:

$$\mathrm{LST}\left(\mathrm{x},\mathrm{y}\right)={\displaystyle \sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{w}}_{\mathrm{i}} \cdot \varphi \left(\parallel \mathrm{P}-\mathrm{Pi}\parallel \right)$$

(1)

where

• LST (x, y) is the interpolated temperature at location (x, y);
• wi are the weights determined based on the known LST values;
• P is the location where interpolation is performed;
• Pi are the known locations with measured LST values;
• ϕ(r) is the radial basis function (e.g., Gaussian: ϕ(r) = e (ϵr)2).

Alternatively, in the case of IDW interpolation:

$$\mathrm{LST}\left(\mathrm{x},\mathrm{y}\right)=\frac{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\displaystyle \frac{\mathrm{LSTi}}{{\mathrm{d}}_{\mathrm{i}}^{\mathrm{p}}}}}{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\displaystyle \frac{1}{{\mathrm{d}}_{\mathrm{i}}^{\mathrm{p}}}}}$$

(2)

where

• di is the distance from point i to the location (x, y);
• p is the power parameter controlling the influence of neighboring points.

3.4. Development of a Correlation Formula for LST

A multiple linear regression formula is established that correlates LST with AT, RH, NDVI, NDMI, and NDWI based on the research data. The general regression model is:

$$\mathrm{LST}={\mathrm{a}}_0+{\mathrm{a}}_1\cdot \mathrm{AT}+{\mathrm{a}}_2\cdot \mathrm{RH}+{\mathrm{a}}_3\cdot \mathrm{NDVI}+{\mathrm{a}}_4\cdot \mathrm{NDMI}+{\mathrm{a}}_5\cdot \mathrm{NDWI}+\mathsf{\epsilon }$$

(3)

where

• a₀ is the intercept;
• a₁, a₂, a₃, a₄, and a₅ are regression coefficients;
• ε is the error term.

This methodology presents a robust framework for understanding the role of urban green spaces in adapting the thermal environment, especially in regions with hostile climatic conditions (Figure 5).

4. Results

4.1. The Data Collected

Table 3. The in-situ measurement and LANDSAT 8 datasets used in the Garden model studied.

Data Collected		Central Station	N-W Axe					N-E Axe					S-E Axe					S-W Axe
Source	Indicators	S0	SN1	SN2	SN3	SN4	SN5	SE1	SE2	SE3	SE4	SE5	SS1	SS2	SS3	SS4	SS5	SW1	SW2	SW3	SW4	SW5
In-situ measurement July 15	AT (°C)	36.5	39.2	40.8	43.5	46.3	48.9	40.5	43.4	45.86	46.1	47.5	38.2	39.6	40.3	42.6	44.2	37.2	38.6	40.7	41.5	42.8
	RH (%)	38.56	36.2	31	28.5	26.5	25.5	33.5	30.1	26.21	25.7	25	36	35.6	33.3	31.1	30.2	37	35.5	33.2	32.6	31.4
LANDSAT 8 data July 15	LST (°C)	40.61	42	44.10	48.30	49.10	49.80	44.61	47.20	49.1	49.50	50.50	42	43.30	44.10	45.70	46.10	41.50	42.00	44.60	44.90	45.90
	NDVI	0.25	0.20	0.22	0.09	0.13	0.03	0.23	0.13	0.08	0.06	0.03	0.24	0.18	0.13	0.07	0.05	0.24	0.22	0.13	0.13	0.06
	NDMI	−0.13	−0.14	−0.20	−0.23	−0.25	−0.27	−0.15	−0.20	−0.23	−0.25	−0.28	−0.13	−0.17	−0.21	−0.24	−0.25	−0.13	−015	−0.23	−0.24	−0.25
	NDWI	0.13	0.08	0.10	0.01	−004	−0.05	0.08	0.09	−0.01	−0.04	−0.06	0.12	0.11	0.03	0.02	0.01	0.12	0.09	0.05	0.03	−0.01

shows a comprehensive list of all datasets used in the study research, bringing together both in situ field observations and remotely sensed information. It shows parameters measured manually in field campaigns, such as air temperature (AT), relative humidity (RH), alongside satellite-derived indices like NDVI, NDMI, NDWI, and LST. Therefore, it allows tracing the consistency and reliability of the data inputs. This systematic data not only facilitates reproducibility but also highlights the integration of heterogeneous sources, which supports the accuracy of the model by seizing both microclimatic variations observed onsite and broader-scale environmental conditions derived from remote sensing imagery.

4.2. Results of the Onsite Measurement

The analysis of AT and RH along the N-E axis direction reveals a clear microclimatic reduction. The center station (S0), located in the center of the garden and recording the lowest temperature (36.50 °C) and the maximum humidity value (38.56%), serves as a representation of the cooling effect of thick vegetation and shade. As one proceeds farther out, the temperature gradually increases in both the north (SN1–SN5) and the south (SS1–SS5), reaching 48.90 °C at SN5 and SE5, but the relative humidity decreases to 25%. This points to drier and hotter circumstances as there is a persistent increase in solar exposure and a decrease in vegetation cover. A cooler microclimate within the shaded zones, a reduction in the SVF, and a restriction of direct sun radiation are all supported by the observed tendency (Figure 6).

4.3. Remote Sensing Data Results

4.3.1. LST and NDVI

Landsat 8 data show significant LST variations in Biskra, Algeria, ranging from 40.49 °C in public gardens to 53.26 °C in bare land outside the urban perimeter. Public gardens have the lowest LST due to dense vegetation, which enhances evapotranspiration and reduces solar radiation through tree canopies and low SVF values. In contrast, bare land has the highest LST due to low reflectivity (Albedo), leading to rapid heating and slow cooling. NDVI values range from −0.003 in industrial areas to 0.391 in public gardens and oases, indicating vegetation health. Industrial zones, with negative or near-zero NDVI, lack greenery due to impermeable surfaces replacing natural flora (Figure 7).

4.3.2. NDWI and NDMI

NDWI is essential for identifying water bodies, delineating their borders, and assessing clarity. In Biskra, NDWI values range from −0.53 to 0.011, reflecting moisture variations across land uses. Dry surfaces like urban areas, roads, and industrial zones show the lowest NDWI values (−0.29 to −0.085), displayed in white due to strong SWIR reflection. In contrast, vegetated areas such as public gardens and remnants of old oases have the highest NDWI values (−0.53 to −0.3), shown in red, indicating their role in maintaining humidity and reducing arid conditions. NDMI, which assesses vegetation and soil moisture, varies across Biskra. The lowest NDMI values (−0.23 to −0.071) are found in dry, built-up, and industrial zones due to low moisture content and high SWIR reflectance. Impermeable surfaces in these areas worsen the UHI effect. Even vegetated regions may have low NDMI if affected by water stress from insufficient irrigation or extreme heat (Figure 8).

5. Discussion

5.1. Correlation of LST Values with Indices (NDVI, NDWI and NDMI)

5.1.1. Zoning Analysis

The study of Biskra’s public garden highlights its thermal and environmental benefits in hot, arid regions. Using a zoning approach, key remote sensing indices (LST, AT, RH, NDVI, NDMI, NDWI) were analyzed across zones to assess the gardens’ role in mitigating the UHI effect and improving humidity. The results reveal a gradient in environmental conditions from the garden center to surrounding urban areas. LST in the central garden is significantly lower (40.61 °C) than in adjacent urban areas, indicating a cooling effect due to vegetation, soil moisture, and shading. Spatial temperature distribution was analyzed using spatial analyst tools, interpolation, and the kriging method, generating an LST distribution map (Figure 9).

The interpolation of LST zoning (41–44 °C) reveals that Biskra’s public garden significantly influences surrounding urban areas, particularly in the southeast and southwest. This impact is due to the garden’s extensive size, which helps moderate temperatures in adjacent zones. Lower LSTs inside the garden result from evapotranspiration, dense tree canopies blocking solar radiation, and soil moisture retention, which reduce heat absorption. In contrast, urban areas outside the garden have higher LSTs due to impervious surfaces like asphalt and concrete, contributing to the UHI effect. Moving away from the garden center, temperatures rise by over 8 °C, highlighting its strong local cooling effect. The contrast between LST, NDVI, NDMI, and NDWI inside and outside the garden emphasizes the role of green spaces in mitigating UHIs (Figure 10).

The zoning analysis confirms the utility of the model as a predictive device for urban planners since it measures both the size and the geographical reach of the cooling and humidity benefits of the garden. For instance, at the central station (S0), LST is 40.61 °C compared to values greater than 49–50 °C at peripheral stations such as SE4 (49.1 °C), SE5 (50.50 °C), and SN5 (49.1 °C), with a temperature increase of approximately 9–10 °C moving from the garden center to exposed urban axes. This cooling gradient is replicated in higher vegetation and soil moisture indices in the garden, where NDVI (0.25) and NDWI (0.13) are considerably greater than at edge stations (e.g., NDVI −0.03 at SE5; NDWI −0.06 at SE4). Similarly, NDMI values fall from −0.13 at S0 to −0.28 at SE5, confirming the drying of surrounding impervious surfaces. Relative humidity also illustrates this contrast, with 38.56% at the garden center dipping to as low as 25% along the NE axis (SE3–SE5).

These results feature the model’s ability to forecast how differences in vegetation density, soil moisture, and water presence immediately affect microclimatic conditions. By joining these associations, the model can be used to simulate alternative greening scenarios, for instance, predicting how the further presence of shaded vegetation along NE and NW axes could attenuate extreme LST values above 50 °C and enhance local humidity. Hence, besides recording true trends, the model provides an option-supportive framework for optimizing the design and planting of green spaces for reducing urban heat islands, increasing thermal comfort, and maximizing climate resilience in tropical deserts such as Biskra.

The scatter plot with regression line shows the LST Distance Values from the Garden Center to Adjacent Districts (Figure 11).

• The scatter plot with regression line shows a clearly increasing trend in LST values as the distance from the garden center increases;
• The LST begins at 40.61 °C at the garden center and peaks above 50 °C in the extreme urban areas;
• This tendency proves that urbanization causes higher temperatures, likely due to diminished vegetation and intensified impervious surfaces;
• The certainty interval around the regression line implies that the predicted LST value closely follows the observed values, indicating a strong fit of the model (Figure 12).

5.1.2. Define the Data

Table 4. LST Values at Different Distances and Directions.

Distance (m)	LST (°C) N-W Direction	LST (°C) N-E Direction	LST (°C) S-E Direction	LST (°C) S-W Direction
0	40.61	40.61	40.61	40.61
60	42.00	44.61	42.00	41.50
120	44.10	47.20	43.30	42.00
180	48.30	49.10	44.10	44.60
240	49.10	49.50	45.70	44.90
300	49.80	50.50	46.10	45.90

presents LST values measured at various distances and directions, providing a spatial distribution of temperature variations that helps identify patterns related to land cover and environmental factors.

Table 5. Maximum Values for NDVI, NDMI, and NDWI.

Indexes	Garden Center	Urban Zone
NDVI	0.25	0.06
NDMI	−0.13	−0.25
NDWI	0.13	−0.01

summarizes the maximum values for three key indices: NDVI, NDMI and NDWI. Simultaneously, these tables contribute to the development of a correlation formula for LST by establishing relationships between thermal variations and surface characteristics.

5.1.3. Regression Model Formulation

In this step, the coefficients of the regression equation are estimated by applying the least squares method:

$$\mathrm{A} = \left({\mathrm{X}}^{\mathrm{T}} \mathrm{X}\right){ \mathrm{X}}^{\mathrm{T}} \mathrm{Y}$$

(4)

where

• X = matrix of independent variables (AT, RH, NDVI, NDMI, NDWI);
• Y = LST values;
• A = vector of regression coefficients.

The final regression model will take the form:

$$\mathrm{LST}={\mathrm{a}}_0+{\mathrm{a}}_1\cdot \mathrm{AT}+{\mathrm{a}}_2\cdot \mathrm{RH}+{\mathrm{a}}_3\cdot \mathrm{NDVI}+{\mathrm{a}}_4\cdot \mathrm{NDMI}+{\mathrm{a}}_5\cdot \mathrm{NDWI}$$

(5)

5.1.4. Implementation in Python

The heatmap presented in Figure 12 (using online Python Version 3.8) visually highlights the spatial variation in LST.

• The North-East (N-E) direction shows the maximum temperatures, reaching 50 °C at 300 m, demonstrating the greatest urban heat island effect;
• The South-West (S-W) direction remains cooler, with LST lasting lower at 46 °C, representing more vegetation and other cooling factors in that direction;
• The LST gradient follows a smooth increase from the garden center outward, reinforcing the cooling impact of the garden.

The results of this study confirm the strong cooling and humidity-regulating functions of Biskra’s public garden. At the garden center, the LST reached 40.61 °C, compared to values exceeding 49–50 °C in peripheral districts, revealing a pronounced cooling gradient of nearly 9–10 °C. This amplitude is higher than the 3–8 °C cooling commonly reported in studies of urban parks in temperate and subtropical contexts [44,45]. However, the greater reduction observed in Biskra aligns with evidence from hot-arid cities, where irrigated gardens act as oasis-like systems with stronger localized cooling effects due to dense vegetation and high evapotranspiration rates [46].

The negative correlations between LST and NDVI, NDMI, and NDWI in our study confirm the pivotal role of vegetation and soil moisture in regulating surface temperatures. These results are consistent with prior findings that link vegetation indices to cooling effects in diverse climates [47]. The relatively high NDVI (0.25) and NDWI (0.13) values inside the garden contrast with much lower values in surrounding urban surfaces (e.g., NDVI = −0.03, NDWI = −0.06 at SE5 and SE4), highlighting how vegetation density and soil moisture retention reduce heat absorption. These results are in agreement with Taloor et al. [48], who demonstrated that vegetation and moisture indices serve as reliable proxies for evapotranspiration potential and heat mitigation in semi-arid river basins.

The spatial analysis revealed a clear decay of cooling intensity with distance from the garden center, as indicated by regression modeling of LST against distance. Similar distance-dependent gradients of park cooling have been documented in studies of large urban parks in Asia and Europe, typically extending several hundred meters beyond park boundaries [44]. The persistence of this effect in Biskra, especially toward the southeast and southwest, suggests that garden size and the configuration of adjacent urban fabric significantly shape the magnitude and spatial extent of cooling. These results reinforce global syntheses that identify park size, shape, and vegetation density as primary determinants of cooling intensity [45].

5.1.5. Percentage Impact of Environmental Factors on LST

The bar chart of Figure 13 reveals the relative importance of each factor affecting LST.

• AT, with a coefficient of +0.72 (72% of the total impact), has the highest impact, confirming its direct influence on LST;
• RH has an inverse relationship with a negative connection (−0.45 coefficient, 17.3% impact), meaning higher humidity contributes to cooling effects;
• NDVI (vegetation index) shows a strong negative impact (−0.35 coefficient, 7.8% impact), reinforcing that greener areas significantly reduce LST;
• NDMI, with a coefficient of −0.28 (2.9% impact) and NDWI (−0.20 coefficient, 1.7% impact), also contribute, indicating that soil and water content play essential roles in temperature regulation.

This analysis confirms that vegetation, moisture, and atmospheric factors interact dynamically to influence LST, supporting strategies like urban greening to mitigate heat islands.

To conclude, the modeling framework applied here, combining spatial interpolation methods (kriging and RBF) with vegetation and moisture indices, demonstrated robust predictive capacity, as reflected in the narrow confidence intervals of the regression line. Such integrative approaches are increasingly recommended in urban climatology for their ability to capture both spatial continuity and environmental drivers of LST [43,49]. In comparison with prior studies that often analyze single indices in isolation, the integration of NDVI, NDMI, NDWI, AT, and RH in this study provides a more holistic predictive tool. This aligns with calls for decision-support frameworks that translate remote sensing outputs into practical planning strategies for hot-arid cities [50].

5.2. Limitations of the Research

While this study provides valuable data regarding the cooling and humidity effect of Biskra’s public garden and demonstrates the availability of remote sensing indexes (LST, NDVI, NDMI, NDWI) as explanatory variables, there are some limitations that should be considered. First, the analysis is carried out based on a single temporal moment (summer season Landsat 8 images), which reduces the ability to monitor seasonality and inter-annual variability. Second, the 30 m spatial resolution, although sufficient to capture general patterns, restricts the scope to observe micro-scale heterogeneity for small urban structures such as narrow streets, courtyards, or tiny green patches. Third, climatic variables such as wind speed, shading geometry, and anthropogenic heat emissions were not included, although they may significantly influence local thermal comfort. Finally, the study focuses on a single case study (Biskra’s public garden), which may constrain the generalizability of the findings to other hot arid cities with different urban forms, vegetation typologies, or socio-environmental contexts.

Future research can overcome these challenges by integrating multi-temporal satellite data to examine seasonal changes, employing higher-resolution remote sensing (e.g., Sentinel-2 or UAV imagery) for finer spatial detail, and incorporating additional climatic and morphological parameters into the model. Extending the analysis to multiple urban contexts will further validate and refine the predictive framework, ultimately enhancing its applicability as a decision-support tool for urban planners and environmental designers in hot arid cities.

6. Conclusions

This study provides clear evidence of the significant cooling and humidity-regulating role of Biskra’s public garden within a hot-arid urban context. By blending in situ measurements (RH, AT) and satellite-derived indices (NDVI, NDWI, NDMI, LST), the study developed a strong model that was capable of effectively measuring the amplitude and area of the garden-induced microclimatic benefits. Results demonstrated an ample temperature gradient of about 9–10 °C between the garden center (40.61 °C) and adjacent urban areas (>49 °C), and the reduction in relative humidity from 38.56% to 25%. These variations were directly related to vegetation cover, evapotranspiration, and low SVF, highlighting the capacity of high vegetal cover to act as an oasis-like system that reduces the UHI effect. Furthermore, the integration of heterogeneous datasets enhanced model accuracy, establishing reproducibility while establishing a holistic view of both regional and local conditions.

The spatial analysis also confirmed that the garden’s cooling effect extends far beyond its boundaries, with lower LST values in the surrounding built-up areas, especially in the southeast and southwest directions. The negative correlations of LST with vegetation/moisture indices (NDVI, NDWI, NDMI) point to the contribution of soil moisture and canopy density in reducing surface heat loads. These results are consistent with international evidence demonstrating that park size, vegetation cover, and design are major determinants of cooling magnitude. LST raises with distance from the public garden, strengthening the cooling effect of urban green spaces, with the North-East direction being the hottest, likely due to more urban expansion. Vegetation indices (NDVI, NDMI, NDWI) play a critical role in decreasing LST, supporting the importance of urban green planning. Additionally, air temperature (AT) and humidity (RH) significantly impact LST, confirming climate-based urban heat island studies.

From the Specific Factors Studied, important conclusions are:

• Air Temperature (AT): With a coefficient of +0.72 (72% of the total impact), the regression model shows that AT has the greatest impact on LST. This suggests that there is a high correlation between atmospheric and surface temperatures, with a 1 °C increase in AT resulting in a 0.72 °C increase in LST;
• A negative connection (−0.45 coefficient, 17.3% impact) was established between relative humidity (RH) and LST, indicating that higher RH reduces LST. This is because surface heating has decreased by evaporative cooling effects and increased latent heat flow;
• The Normalized Difference Vegetation Index (NDVI) has a considerable negative effect on LST (−0.35 coefficient, 7.8% impact), indicating that better shading and transpiration cooling result in significantly lower temperatures in places with denser vegetation;
• Normalized Difference Moisture Index (NDMI): With a coefficient of −0.28 (2.9% impact), NDMI plays a secondary role in cooling. Higher soil moisture leads to increased latent heat exchange, preventing excessive surface heating;
• Normalized Difference Water Index (NDWI): The least significant factor (−0.20 coefficient, 1.7% impact), NDWI still contributes to LST regulation, mainly through water body cooling effects, though its influence is weaker compared to AT and NDVI;
• Predicted LST values closely match observed values, with a mean absolute error (MAE) of 0.43 °C, showing minimal deviation;
• Comparing LST at 0 m (garden center, 40.61 °C) and 300 m (district center, 50.50 °C), a 9.89 °C increase is observed, with AT contributing ~7.12 °C of this rise, NDVI accounting for ~0.77 °C reduction, and RH offsetting ~1.71 °C of warming;
• The heatmap’s highest LST values align with the lowest RH, NDVI, and NDMI values, confirming the dominant effect of vegetation and moisture loss in urban heat islands.

Research demonstrates that public gardens are essential in mitigating urban heat, particularly through their effect on the thermal microclimate of adjacent areas. Although significant progress has been made in quantifying these effects, including through onsite measurement and satellite remote sensing, additional research is needed to address gaps in temporal and spatial dynamics, particularly in hot and arid regions where green spaces are necessary for urban sustainability. Further studies should focus on optimizing garden design, vegetation selection, and management practices to maximize their thermal benefits, particularly in facing increasing urbanization and climate change. Furthermore, although remote sensing techniques have advanced the field, there is still a need for long term onsite measurements, ground validation, and numerical simulations to complement satellite data. This would allow for a better understanding of how public gardens affect surface and air temperatures in complex urban environments.