Monitoring the Sustained Environmental Performances of Nature-Based Solutions in Urban Environments: The Case Study of the UPPER Project (Latina, Italy)

Abstract

This follow-up study investigates the long-term environmental sustainability and remediation outcomes of the UPPER (‘Urban Productive Parks for Sustainable Urban Regeneration’-UIA04-252) project in Latina, Italy, focusing on Nature-Based Solutions (NbS) applied to urban green infrastructure. By integrating proximal and satellite-based remote sensing methodologies, the research evaluates persistent improvements in vegetation health, soil moisture dynamics, and overall environmental quality over multiple years. Building upon the initial monitoring framework, this case study incorporates updated data and refined techniques to quantify temporal changes and assess the ecological performance of NbS interventions. In more detail, ground-based data from meteo-climatic, air quality stations and remote satellite data from the Sentinel-2 mission are adopted. Ground-based measurements such as temperature, humidity, radiation, rainfall intensity, PM10 and PM2.5 are carried out to monitor the overall environmental quality. Updated satellite imagery from Sentinel-2 is analyzed using advanced band ratio indices, including the Normalized Difference Vegetation Index (NDVI), the Normalized Difference Water Index (NDWI) and the Normalized Difference Moisture Index (NDMI). Comparative temporal analysis revealed consistent enhancements in vegetation health, with NDVI values significantly exceeding baseline levels (NDVI 2022–2024: +0.096, p = 0.024), demonstrating successful vegetation establishment with larger gains in green areas (+27.0%) than parking retrofits (+11.4%, p = 0.041). However, concurrent NDWI decline (−0.066, p = 0.063) indicates increased vegetation water stress despite irrigation infrastructure. NDMI improvements (+0.098, p = 0.016) suggest physiological adaptation through stomatal regulation. Principal Component Analysis (PCA) of meteo-climatic variables reveals temperature as the dominant environmental driver (PC2 loadings > 0.8), with municipality-wide NDVI-temperature correlations of r = −0.87. These multi-scale findings validate sustained NbS effectiveness in enhancing vegetation density and ecosystem services, yet simultaneously expose critical water-limitation trade-offs in Mediterranean semi-arid contexts, necessitating adaptive irrigation management and continued monitoring for long-term urban climate resilience. The integrated monitoring approach underscores the critical role of continuous, multi-scale assessment in ensuring long-term success and adaptive management of NbS-based interventions.

1. Introduction

As urbanization accelerates worldwide, cities are increasingly burdened by environmental challenges such as rising temperatures, air pollution, flooding, and loss of biodiversity. These pressures are exacerbated by the spread of impervious surfaces and climate change, straining cities’ capacity to adapt to these stressors. As a result, urban areas are often less resilient as ecosystems, leading to diminished environmental quality and a lower standard of living for residents. In this context, Nature-based Solutions (NbS) have emerged as a promising approach to address these complex challenges by harnessing natural processes and integrating them into urban environments to improve resilience and sustainability [1]. NbS have been defined as solutions that leverage natural resources to benefit ecosystems and the human communities that depend on them [2]. In recent years, NbS have gained prominence as a concept in environmental science and nature conservation. Recognized by major international organizations, NbS offer holistic solutions, simultaneously addressing climate change adaptation and mitigation, while preserving ecosystems and enhancing urban sustainability [3]. According to the European Commission (EC), NbS refers to interventions that use natural processes to address issues related to the economy, society, and the environment [4]. Actions to conserve, restore, or regenerate natural and modified ecosystems are emphasized by the International Union for Conservation of Nature (IUCN) as part of NbS [5]. These solutions not only mitigate climate impacts but also enhance human well-being, contribute to biodiversity conservation, and promote sustainable development [6].

In the context of urban environment, NbS interventions represent a promising approach for enhancing urban resilience and sustainability [7]. NbS strategies rely on using or creating natural features and ecosystems within and around urban areas to provide significant ecosystem services, contributing to many aspects of sustainable urban development [8]. A primary advantage of integrating NbS lies in their ability to support climate change mitigation and adaptation efforts [9]. Cities can reduce carbon emissions and improve carbon capture by incorporating green and blue infrastructure, as well as natural elements, into the fabric of their urban spaces and numerous studies report substantial cooling benefits and urban heat island (UHI) effect mitigation potential associated with such interventions, although the magnitude and spatial extent of these effects can be highly context-dependent [10,11,12]. Recent research has also highlighted that the capacity of urban vegetation and NbS to mitigate UHI is not uniform, with some studies questioning or nuancing earlier assumptions by showing that cooling effects vary with climate, spatial configuration, vegetation type, and background urban morphology [13,14].

However, urban green space implementation following an NbS-based approach can offer co-benefits such as improved air quality, noise reduction, and the promotion of mental well-being [15]. The transdisciplinary nature of NbS has been emphasized in several European projects, particularly in contexts facing environmental and socio-cultural challenges. Budoni [16] highlighted how NbS serves as an ‘umbrella concept’ that enables different disciplines (including urban planning, ecology, and social sciences) to collaboratively address complex urban sustainability issues.

In the context of the UPPER (‘Urban Productive Parks for Sustainable Urban Regeneration’; https://uia-initiative.eu/en/uia-cities/latina; accessed on the 25 November 2024) project implemented in the Municipality of Latina (Italy), this integrated approach was operationalized through the development of green and blue infrastructures within a bioregional framework, facilitating participatory co-design processes and fostering community engagement [17]. The UPPER project, funded under the Urban Innovative Actions initiative (UIA04-252), aimed to transform underutilized urban spaces into productive parks that integrate NbS. These parks not only provide green infrastructure but also serve as community hubs that contribute to biodiversity enhancement, climate mitigation, and public well-being. This project introduced a variety of NbS interventions, such as increased urban green spaces, sustainable water management systems, and the use of vegetation to improve air quality and reduce heat stress.

Objectives and Novelty of the Study

Despite the expanding literature on urban NbS, significant knowledge gaps persist. Most existing studies rely on single environmental indicators (e.g., vegetation indices only), focus on short monitoring periods, or lack integration of satellite time-series with ground-based meteorological and air quality data. Consequently, there is limited empirical evidence on the sustained performance of urban NbS across multiple ecological and environmental dimensions over years or decades, particularly in real-world urban regeneration projects. Long-term, multi-scale monitoring frameworks that combine remote sensing with proximal sensing observations are essential to: (i) quantify temporal changes in vegetation and moisture conditions post-intervention; (ii) assess NbS sensitivity to climatic and air quality drivers; and (iii) evaluate the durability of ecological gains beyond the initial implementation phase.

In a previous study, the focus was set on the possibilities of the utilization of a combined approach using both proximal (field-based near-infrared spectroscopy) and remote sensing (satellite data from the Sentinel-2 mission) to monitor the environmental performance of NbS interventions [18]. By employing these sensing techniques, the study comprehensively monitors vegetation health, water content, and urbanization by calculating various vegetation indices (e.g., NDVI, NDWI). Temporal changes among these indices are used to assess the performance of NbS interventions in improving the urban environment. The findings demonstrated the feasibility of a multi-scale monitoring framework that combines proximal and remote sensing, providing valuable insights into the long-term effects of NbS while also supporting sustainable urban development and decision-making based on evidence. In this context, monitoring NbS interventions is fundamental for assessing their multi-dimensional co-benefits as proposed in the comprehensive framework of [19].

This follow-up study is mainly focused on the assessment of integrated environmental performance, with its key goal is to evaluate the long-term impact of these NbS interventions in Latina’s Productive Parks. Specifically, the study seeks to assess whether the ecological and environmental improvements observed during the initial implementation phase have been sustained over time. To achieve this goal, as in the previous study [18], a combined, proximal and remote sensing methods-based approach was adopted.

The study builds on the original monitoring framework, incorporating new data and refined methodologies to provide comprehensive temporal analysis. This includes the integration of ground-based measurements from meteo-climatic and air quality stations with remote satellite data from the Sentinel-2 mission. By using advanced indices such as the Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), and Normalized Difference Moisture Index (NDMI), we aim to capture the nuanced changes in the urban ecosystem and quantify the effectiveness of the NbS interventions over several years.

The findings of this study are intended to provide critical insights into the sustainability and long-term efficacy of NbS in urban environments. By analyzing both the positive trends and the challenges faced in maintaining these interventions, this research contributes to the broader understanding of how NbS can help cities like Latina adapt to the pressing environmental issues of the 21st century.

2. The Case Study Context and Approach

Aerial platforms or satellites allow for the collection of Earth Observation (EO) data, which provides comprehensive and multi-temporal insights into ecosystem health, the dynamics of vegetation, and changes in land cover. For assessing the effects of NbS on a broader spatial scale, information obtained from EO is essential [20]. Proximal sensing, on the other hand, uses portable or ground-based devices to collect high-resolution measurements of specific parameters (such as temperature, humidity, etc.). This combined approach enables localized monitoring and real-time decision-making for NbS management, particularly in the monitoring and evaluation phases [20].

The presented case study is part of the UPPER project, which focuses on experimenting with urban Productive Parks in Latina, Italy, with the goal of co-producing NbS to deal with various social, environmental, and economic challenges. The UPPER project adopted a bioregional approach to urban regeneration, integrating ecological networks, mobility infrastructures, and participatory governance frameworks [17]. This strategy sought to transform underutilized urban spaces into multifunctional ‘productive parks’ that not only enhance ecosystem services but also support social cohesion and economic innovation [16]. The NbS-targeted interventions were strategically implemented to address multiple sustainability objectives, including (i) reduction in impervious surfaces and associated urban heat island effects, (ii) enhancement of water retention and sustainable water management through permeable landscapes and harvesting systems, (iii) promotion of biodiversity using indigenous plant species, and (iv) improvement of urban quality and ecological connectivity. All interventions incorporated water-conserving irrigation systems during plant establishment, minimizing CO₂ emissions from transport and operational management. The overview map of the Municipality of Latina in the European context and the areas targeted for NbS interventions within the UPPER project are shown in Figure 1.

3. Materials and Methods

3.1. Areas of Study and Interventions

The UPPER areas investigated before and after the intervention, as shown in Figure 1c, are: (a) P.Le dei Mercanti–demonstration site (parking area), (b) P.Le dei Mercanti-production park (park/green area), (c) Piazza Ilaria Alpi (demonstration site; park/green area), (d) Via Goya (demonstration site; park/green area), (e) Via Lepanto (demonstration site; parking area), (f) Via Neghelli (demonstration site; parking area) and (g) Via Romagnoli (demonstration site; parking area). The UPPER project NbS interventions by site are summarized in

Table 1. UPPER project NbS interventions by site. A detailed description of the NbS intervention for each site is provided in [18].

Site	Coordinates	Type of Site	Main NbS Intervention	Key Design Elements and Objectives
P.Le dei Mercanti–demonstration site (parking area)	41.4555967° N 12.9131632° E	Parking area	Green retrofit of impervious parking surfaces	Systematic redesign of parking layout to reduce impermeable surfaces; introduction of bioswales and flowerbeds with native trees and shrubs; drip irrigation during establishment; species selected for biodiversity promotion, water conservation and long-term low irrigation demand in Mediterranean/semi-arid climate.
P.Le dei Mercanti–Production Park (park/green area)	41.4542585° N 12.9115145° E	Park/green area	Productive green park with integrated water management	Diverse composition of native trees and shrubs; integrated water management systems to capture and retain rainfall for vegetation; reduction in dependence on external water supplies; emphasis on sustainable resource management and ecological infrastructure.
Piazza Ilaria Alpi (demonstration site; park/green area)	41.463126° N 12.919005° E	Urban park/green area	Urban forest with water and energy-oriented NbS	Establishment of a diverse urban forest of native shrubs and trees; underground irrigation systems for initial growth phase; rainwater harvesting infrastructure; use of eco-friendly materials for urban furniture and pergolas; provisions for photovoltaic installations to support long-term operational sustainability.
Via Goya (demonstration site; park/green area)	41.45188° N 12.883506° E	Urban/food forest	Urban forest connected to existing woodland	Design in continuity with a regenerated wooded area containing centuries-old trees; pedestrian and cycling paths around a central catchment area for irrigation; all plant species selected from indigenous vegetation adapted to local climate; enhancement of recreational use and ecological connectivity.
Via Lepanto (demonstration site; parking area)	41.4718173° N 12.9043426° E	Parking area	Green retrofit with bioswales and native vegetation	Conversion of highly urbanized, impervious parking surfaces into productive ecological zones; introduction of bioswales and flowerbeds with native trees and shrubs; drip irrigation during establishment; species selected for biodiversity, drought tolerance and reduced long-term water demand.
Via Neghelli (demonstration site; parking area)	41.4706980° N 12.9040640° E	Parking area	Vegetation-based retrofit of parking	Redesign of parking to reduce sealed surfaces; planting of native trees and shrubs in bioswales and vegetated beds; drip irrigation systems in early phases; focus on stormwater management, biodiversity support and progressive reduction in irrigation needs.
Via Romagnoli (demonstration site; parking area)	41.4713234° N 12.8937944° E	Parking area near bus station	Ecological retrofitting with climate-resilient trees	Transformation of parking area and adjacent flowerbeds with native tree species; species chosen for ecological function, drought tolerance and aesthetic value; objectives include enhancing ecological connectivity, supporting pollinators, and mitigating the urban heat island effect via increased canopy cover and evapotranspiration; drip irrigation during establishment with long-term reduction in irrigation through deep-rooted, water-efficient species.

. In more detail, the green space sites were designed to integrate ecological infrastructure with renewable energy technologies and sustainable resource management. In contrast to the extensive green space interventions described above, the parking area sites represent intensive retrofitting of highly urbanized, impervious surfaces into productive ecological zones. All parking lot interventions incorporated targeted drip irrigation systems during plant establishment phases, with the long-term objective of reducing irrigation dependency through species selection favoring water-efficient, deep-rooted vegetation adapted to Mediterranean and semi-arid climates.

In Appendix B, pre- and post-intervention photographic documentation for representative sites (Piazza Ilaria Alpi and Via Goya) is provided, illustrating the transition from pre-project conditions (sealed or sparsely vegetated areas) to the current NbS configurations.

3.2. Analysis of Satellite Data and Indices Derived from Remote Sensing

To evaluate the transition from the pre-intervention to post-intervention phases within the UPPER project areas, a comprehensive study leveraging satellite imagery was conducted. Specific Areas of Interest (AOIs) were delineated to align with the intervention zones, enabling a detailed analysis of vegetation health and plant performance, as reported in Figure 1. This assessment involved mapping key vegetation indices, such as the Normalized Difference Vegetation Index (NDVI), the Normalized Difference Water Index (NDWI), and the Normalized Difference Moisture Index (NDMI).

Historical data from Sentinel-2 L2A satellite imagery were utilized to calculate the average values of NDVI, NDWI, and NDMI within each AOI. This calculation was performed using available satellite images from the last five years (1 January 2019–4 April 2024). The temporal series of these indices enables a direct comparison between the pre-intervention and post-intervention states of the NbS areas, providing valuable insights into their impact on plant health and overall vegetation conditions.

Sentinel-2 is a EO mission that plays a key role in Copernicus Land Monitoring applications, particularly in observing vegetation, soil, water, and coastal areas [21]. Each Sentinel-2 satellite provides imagery with a 10-day revisit period, and the combined constellation enhances this to a 5-day revisit cycle, enabling regular updates on land cover changes. The mission’s multispectral instrument (MSI) captures data across 13 spectral bands at varying spatial resolutions (the size of the area represented by each pixel): (i) 10-m resolution for visible and near-infrared bands, (ii) 20-m resolution for red-edge and shortwave infrared bands and (iii) 60-m resolution for atmospheric bands.

The satellite data processing is carried out using the Copernicus Data Space Ecosystem Browser (https://dataspace.copernicus.eu/browser; accessed on the 25 November 2024), a centralized platform based on the Sentinel Hub EO Browser. This application provides users with the ability to visualize, compare, analyze, and download Earth observation data, supporting a wide range of applications, including environmental monitoring, disaster response, urban planning, and agricultural management. The satellite map used for calculating the vegetation index maps of the post-intervention status is from 4 April 2024 (Figure 2). Sentinel-2 Level-2A products were utilized, which provide Bottom-of-Atmosphere (BOA) reflectance values corrected for atmospheric effects using the Sen2Cor algorithm. This preprocessing ensures comparability across the multi-temporal dataset by mitigating the influence of atmospheric scattering and absorption. Cloud masking was applied using the QA60 band to exclude pixels with cloud coverage >5%, ensuring data quality and consistency.

The Areas of Interest (AOIs), reported in Figure 1c, were obtained from the shapefile corresponding to the UPPER intervention areas, initially accessed through the WebGIS platform Holding Interactive Platform (WHiP), firstly migrated to the WebGIS Territorial Information System of the Province of Latina and then to the novel UPPER-WHiP platform (https://lpdt.cersites.uniroma1.it/index.php/view/map?repository=upper&project=UPPER_WHIP; accessed on the 25 November 2025). The maps of the AOIs considered in this study were subsequently imported and processed within the QGIS environment (version 3.38; QGIS.org, 2024. QGIS Geographic Information System. Open-Source Geospatial Foundation Project. http://qgis.org; accessed on the 25 November 2024). Starting from the selected AOIs the remote sensing indices NDVI, NDWI and NDWI were evaluated.

The NDVI provides useful information about the amount of chlorophyll and photosynthetic activity in plants, as chlorophyll absorbs red light and reflects Near-Infrared (NIR). Healthy vegetation reflects significantly in the NIR spectrum (700 to 1300 nm) while absorbing strongly in the red region of the visible spectrum (Vis). This property makes the red and NIR spectral bands ideal for calculating the NDVI, which serves as a proxy for plant health and biomass. In this study, the NDVI was derived using Sentinel-2 bands B8a (865 nm) for NIR and B4 (665 nm) for the red spectrum:

$$NDVI = {\displaystyle \frac{(NIR - Red)}{(NIR + Red)}} = {\displaystyle \frac{({R }_{865 \mathrm{nm}} - R_{ 665 \mathrm{nm}})}{({R }_{865 \mathrm{nm}}+ R_{ 665 \mathrm{nm}})}}$$

(1)

While the NDWI is usually adopted in evaluating the presence of water and monitoring changes in aquatic environments.

Water bodies typically have a high reflectance within the blue (400–500 nm) spectral range, with reduced reflectance in the green (500–600 nm) and red (600–700 nm) ranges. The NDWI, as introduced by [22], leverages this property to detect and monitor water bodies by assessing the difference between reflectance in the green and NIR bands.

In this analysis, NDWI was evaluated using Sentinel-2 bands B3 (560 nm) for green and B8 (842 nm) for NIR:

$$NDWI = {\displaystyle \frac{(GREEN - NIR) }{(GREEN + NIR)}} = {\displaystyle \frac{({R }_{560 \mathrm{nm}} - R_{ 842 \mathrm{nm}})}{(R_{560 \mathrm{nm}} + R_{842 \mathrm{nm}})}}$$

(2)

Finally, the NDMI measures vegetation moisture content by using the NIR and Short-Wave Infrared (SWIR) bands. The SWIR band reflects variations in both the water content of vegetation and the structure of the spongy mesophyll in plant canopies, while the NIR band primarily reflects leaf structure and dry matter content, unaffected by water content. Water availability in leaf tissues significantly affects spectral reflectance in the SWIR region, with SWIR reflectance being negatively correlated with leaf water content. Therefore, the NDMI is an effective tool for monitoring changes in vegetation water content [23].

In this study, NDMI was calculated using Sentinel-2 bands B8 (842 nm) for NIR and B11 (1610 nm) for SWIR:

$$NDMI = {\displaystyle \frac{(NIR - SWIR) }{(NIR + SWIR)}} = {\displaystyle \frac{(R_{842 \mathrm{nm}} - R_{1610 \mathrm{nm}})}{(R_{842 \mathrm{nm}} +{ R}_{1610 \mathrm{nm}})}}$$

(3)

The NDMI values vary between −1 and 1. Negative NDMI values (near −1) correspond to dry soil, while values near zero (from −0.2 to 0.4) typically indicate water stress. Higher values (around 0.4 to 1) represent dense vegetation with no water stress.

Using historical Sentinel-2 L2A satellite imagery, the average values for the NDVI, NDWI, and NDMI were calculated for the AOIs. These indices were derived from all available Sentinel-2 L2A maps covering the period from 2019 to 2024. The satellite maps selected for the calculation of these indices were chosen with a cloud coverage of <5% to avoid selecting maps covered by clouds, as their presence would have impacted the calculation of average values within the AOIs. Specifically, the value of a single generic index (Index) for a given reference year (Year) was calculated as follows:

$${Index}_{year}={\displaystyle \frac{\sum _i^n{Index }_i}{n}}$$

(4)

where the number of observations (n) represents the number of satellite images available for the calculation related to the year in question.

3.3. Proximal Sensing Monitoring

3.3.1. Meteo-Climatic and Air Quality Data

Assuming that each year presents distinct phenological characteristics, influenced by various factors as both short- and long-term climate fluctuations and weather conditions, meteorological data were considered to ensure an accurate evaluation and comparison of the different indices [24]. The meteo-climatic data were obtained from the website of the integrated agrometeorological service of the Regional Agency for the Development and Innovation of Agriculture in Lazio (ARSIAL Lazio; https://www.siarl-lazio.it; accessed on the 25 November 2024). In more detail, the data were retrieved from B.GO CARSO meteo-climatic station (B.go Carso, LT11CME, Latitude: 41.53777° N, Longitude: 12.85528° E), shown in Figure 1b. Annual data were derived from the monthly data (January 2019–April 2024). In more detail, the observed variables were: (a) Average Humidity (%), that is referred to the monthly mean relative humidity; (b) ETO (mm), that is the Evapotranspiration Reference, measuring the amount of water that evaporates from the soil and is transpired by plants under ideal conditions; (c) Cumulative rainfall depth (mm), that is the monthly total amount of rainfall collected and provides insight into the amount of water supplied to the ecosystem through precipitation over that time; (d) Maximum rainfall intensity (mm), that measures the highest rate at which rain fell; (e) Rainy days, that is the number of days within a month when measurable rainfall (usually more than 0.1 mm) was recorded; (f) Average Daily Radiation (W/m²), that measures the average solar radiation received per square meter of the Earth’s surface during a day and (g) Maximum Minimum and Average Temperature (°C), that are, respectively, the highest, the lowest recorded temperature and the mean temperature of the specified month.

The data regarding pollution levels in the city of Latina (ISTAT code 12059011, ZONE IT1218) were retrieved on 3 September 2024 from the Regional Environmental Protection Agency of Lazio (ARPA Lazio; https://qa.arpalazio.net/datiComunali.php; accessed on the 25 November 2024).

ARPA Lazio carries out its institutional task of monitoring the current air quality situation and understanding the causes that determine it through a complex system. This system includes fixed monitoring stations that form part of the regional air quality network, periodic campaigns carried out using mobile agency equipment, a micro-meteorological network, and daily operational modeling chains at the “Centro Regionale della Qualità dell’Aria” (Regional Air Quality Center; website: https://www.arpalazio.it/web/guest/ambiente/aria/sistema-di-monitoraggio; accessed on the 25 November 2024).

More specifically, the air quality data for PM10 and PM2.5 are derived from the urban traffic monitoring station “LATINA DE CHIRICO—63” (Latitude: 41.451131° N, Longitude: 12.891731° E; Altitude: 21 m above sea level), shown in Figure 1b. The reference station is equipped with the following instrumentation: (i) Envea AC32e for Nox—Nitrogen Oxides, (ii) Philips K50093 for CO—Carbon Monoxide; (iii) Air Toxic for BTEX—Volatile Organic Compounds; (iv) SWAM5a FAI Dual Channel for PM10 and PM2.5—Particulate Matter.

The data spans the years 2019 to 2024, with specific records from 1 January 2019 to 4 April 2024. For the year 2024, the daily averages are based on only 93 observations, covering the period from 1 January 2024 to 4 April 2024. Only PM10 and PM2.5, key indicators for particulate pollution, were used in the analysis because of the inconsistency and missing data of the other measurements.

3.3.2. Exploratory and Correlation Analyses

All the statistical analyses of the meteorological, climatic, air quality data, and remote sensing indices were carried out either in the MATLAB environment (version 23.2; MathWorks, Natick, MA, USA) using the Statistics and Machine Learning Toolbox (version 23.2; MathWorks, Natick, MA, USA) alongside the PLS_Toolbox (version 9.0; Eigenvector Research Inc., Manson, WA, USA) toolboxes, or using Python environment (version 3.13.5) in Visual Studio Code (version 1.107.1; Microsoft, Redmond, WA, USA).

Principal Component Analysis (PCA) was performed to identify dominant pattern of variance in meteorological and air quality data and determine whether the variations observed between 2019 and 2024 can be attributed to factors beyond long-term variability in the Comune di Latina area. PCA is a statistical method commonly used to simplify complex datasets by reducing their dimensionality, allowing for the evaluation of patterns in the analyzed data [25]. By using this method, the initial variables are converted into Principal Components (PCs), which are orthogonal and uncorrelated. The data subjected to PCA were pre-processed using Class Centroid Centering and Scaling. This pre-processing algorithm first centers the data using the class centroid approach and then scales it based on the pooled standard deviation of the classes [26].

Annual averages for NDVI, NDWI, and NDMI were calculated for each AOI. Absolute increments between 2022 (pre-intervention or early-stage post-intervention) and 2024 (post-intervention) were computed as (Value₂₀₂₄ − Value₂₀₂₂), and percentage changes were calculated as [(Value₂₀₂₄ − Value₂₀₂₂)/Value₂₀₂₂] × 100 to enable comparison across indices and sites with different absolute ranges. For each AOI, standard deviations (SD) and 95% confidence intervals (CIs) of the annual mean indices were also computed to quantify inter-annual variability and uncertainty. To assess the statistical significance of changes in NDVI, NDWI, and NDMI between 2022 and 2024, the following statistical tests were performed: Paired t-tests and Effect size (Cohen’s d). In more detail, Two-tailed paired Student’s t-tests were carried out to compare NDVI, NDWI, and NDMI mean values between 2022 and 2024 across all nine AOIs, testing the null hypothesis of no significant difference. Results include t-statistics, degrees of freedom (df = 8), and p-value, with significance threshold α = 0.05. The magnitude of change was quantified using Cohen’s d effect size, calculated as (mean₂₀₂₄ − mean₂₀₂₂)/pooled SD, with interpretation thresholds: d < 0.2 (negligible), 0.2–0.5 (small), 0.5–0.8 (medium), d > 0.8 (large).

The data related to meteorological and climatic variables, as well as the average values of NDVI, NDWI, and NDMI, were analyzed through correlation analysis [27,28], in order to evaluate the impact of meteorological and climatic variables on the indices. Given that meteorological and climatic data are consistent across both the entire municipality of Latina and the individual UPPER areas, the correlation analysis was preliminary carried out on a synoptic scale, focusing on the entire municipality (Comune di Latina). The same meteorological and climatic variables were then analyzed for each individual UPPER area. Still within a synoptic framework, the meteorological and climatic variables, along with the indices, were analyzed to assess the significance and uniqueness of specific years in terms of environmental conditions.

4. Results and Discussion

4.1. Remote Sensing Index Variations

The results related to the processing of satellite data, the vegetation index mappings, and the time series of indices for the entire municipality of Latina and the UPPER areas are presented and discussed in the following.

4.1.1. NDVI Mapping for 2024

The false-color NDVI maps, calculated from the Sentinel-2 satellite imagery of 4 April 2024, for the entire municipality of Latina and the specific UPPER intervention areas are shown in Figure 3. In this analysis, the maps display NDVI values across the landscape in false color, where the intensity of colors correlates to different levels of vegetation health and density. By visually representing this index for the entire urban and peri-urban area of Latina, as well as for the targeted UPPER zones, it is possible to assess and monitor the spatial distribution of vegetation vitality. High NDVI values suggest healthy, dense vegetation, while lower values can indicate sparse or stressed vegetation.

4.1.2. NDWI Mapping for 2024

The false-color NDWI maps, derived from the Sentinel-2 satellite imagery of 4 April 2024, for the entire municipality of Latina and the specific UPPER intervention areas, are shown in Figure 4. Areas with higher NDWI values, displayed in brighter blue, indicate the presence of water bodies or higher moisture levels.

4.1.3. NDMI Mapping for 2024

The false-color NDMI maps, calculated using the Sentinel-2 satellite imagery from April 4th, 2024, for the entire municipality of Latina and the specific UPPER project areas, are displayed in Figure 5. Areas with high NDMI values indicate vegetation with higher water content, typically suggesting healthy and non-stressed vegetation. Conversely, areas with low NDMI values may signal dry soil, urbanized areas or vegetation experiencing water stress.

4.1.4. Temporal Variations in the Indices (2019–2024)

Based on the time series of indices for each Area of Interest (AOI) over the period from 2019 to 2024, annual average values were obtained for each year considered.

Table 2. Variation in NDVI, NDWI, and NDMI over the period 2019–2024, corresponding to each analyzed AOI. The table also includes the number of observations (number of satellite images) used for calculating the average indices.

AOI	Year	NDVI		NDWI		NDMI		Number of Satellite Images
		Mean	Std. Dev.	Mean	Std. Dev.	Mean	Std. Dev.
Comune di Latina	2019	0.444	0.199	−0.480	0.185	0.065	0.126	16
	2020	0.477	0.211	−0.497	0.194	0.095	0.129	30
	2021	0.466	0.208	−0.490	0.191	0.095	0.134	19
	2022	0.460	0.203	−0.488	0.188	0.084	0.124	28
	2023	0.468	0.197	−0.495	0.189	0.070	0.116	20
	2024	0.567	0.217	−0.548	0.198	0.199	0.124	5
P.Le dei Mercanti–Site	2019	0.132	0.101	−0.175	0.093	−0.03	0.051	26
	2020	0.142	0.108	−0.189	0.097	−0.035	0.061	42
	2021	0.13	0.088	−0.184	0.085	−0.06	0.048	32
	2022	0.137	0.099	−0.188	0.092	−0.042	0.055	40
	2023	0.134	0.088	−0.182	0.082	−0.038	0.049	35
	2024	0.189	0.101	−0.233	0.089	0.008	0.055	6
P.Le dei Mercanti–Park	2019	0.471	0.133	−0.48	0.111	0.105	0.068	27
	2020	0.548	0.151	−0.535	0.119	0.159	0.083	43
	2021	0.487	0.138	−0.486	0.108	0.098	0.096	33
	2022	0.499	0.142	−0.501	0.109	0.095	0.094	38
	2023	0.454	0.137	−0.464	0.106	0.086	0.077	33
	2024	0.508	0.156	−0.489	0.116	0.174	0.065	7
Piazza Ilaria Alpi	2019	0.466	0.125	−0.469	0.102	0.06	0.051	26
	2020	0.482	0.132	−0.494	0.109	0.071	0.066	44
	2021	0.434	0.124	−0.455	0.099	0.06	0.065	37
	2022	0.444	0.117	−0.464	0.096	0.058	0.057	38
	2023	0.487	0.134	−0.483	0.108	0.093	0.064	35
	2024	0.54	0.114	−0.528	0.092	0.124	0.055	7
Via Goya	2019	0.598	0.098	−0.581	0.063	0.165	0.075	28
	2020	0.624	0.1	−0.605	0.065	0.173	0.078	46
	2021	0.53	0.113	−0.545	0.073	0.081	0.082	36
	2022	0.566	0.102	−0.562	0.068	0.11	0.071	37
	2023	0.56	0.111	−0.552	0.074	0.129	0.081	35
	2024	0.771	0.091	−0.694	0.062	0.344	0.086	6
Via Lepanto	2019	0.208	0.157	−0.253	0.156	0.026	0.086	25
	2020	0.248	0.192	−0.284	0.18	0.048	0.11	45
	2021	0.19	0.16	−0.228	0.15	0.014	0.085	35
	2022	0.211	0.169	−0.256	0.159	0.03	0.09	36
	2023	0.225	0.167	−0.268	0.157	0.031	0.084	33
	2024	0.272	0.165	−0.301	0.154	0.075	0.081	7
Via Neghelli	2019	0.067	0.037	−0.126	0.05	−0.097	0.04	23
	2020	0.075	0.048	−0.129	0.065	−0.103	0.048	45
	2021	0.065	0.042	−0.115	0.054	−0.083	0.038	35
	2022	0.072	0.04	−0.132	0.053	−0.096	0.04	36
	2023	0.106	0.038	−0.166	0.044	−0.087	0.041	34
	2024	0.182	0.054	−0.226	0.055	−0.033	0.053	7
Via Romagnoli	2019	0.103	0.085	−0.144	0.076	−0.015	0.037	24
	2020	0.112	0.089	−0.153	0.08	−0.01	0.047	46
	2021	0.107	0.087	−0.15	0.078	−0.013	0.046	35
	2022	0.105	0.09	−0.156	0.081	−0.028	0.045	33
	2023	0.108	0.083	−0.156	0.075	−0.021	0.041	34
	2024	0.138	0.056	−0.182	0.058	0.011	0.04	7

shows the average indices for the years from 2019 to 2024, corresponding to each AOI, along with the number of satellite images (i.e., observations) used to calculate the averages.

The temporal variations in the NDVI, NDWI, and NDMI for the considered AOIs are graphically presented in Figure 6. Overall, as shown, in 2024 there were recorded increases in both the NDVI and NDMI, while the NDWI experienced an overall decrease.

4.1.5. Considerations About Indices Variations (2022–2024)

Variation in NDVI (2022–2024). Regarding the NDVI, all areas recorded a positive increase in NDVI between 2022 (pre-intervention) and 2024 (post-intervention), which suggests that the UPPER interventions had a beneficial impact on vegetation across all monitored zones (

Table 3. Variation in NDVI (2022–2024). The 2022–2024 increase is calculated as the difference between the index value in 2024 and that in 2022.

	Comune di Latina	P.Le dei Mercanti–Site	P.Le dei Mercanti–Park	Piazza Ilaria Alpi	Via Goya	Via Lepanto	Via Neghelli	Via Romagnoli
2022	0.460	0.137	0.499	0.444	0.566	0.211	0.072	0.105
2023	0.468	0.134	0.454	0.487	0.560	0.225	0.106	0.108
2024	0.567	0.189	0.508	0.540	0.771	0.272	0.182	0.138
Increment 2022–2024	+0.107	+0.052	+0.009	+0.096	+0.205	+0.061	+0.110	+0.033

). Statistical analysis (paired t-test: t = 2.847, p = 0.024) confirms the significance of this improvement. Across all nine monitored areas, the mean NDVI increase was +0.096 ± 0.062 (95% CI: +0.043 to +0.149), representing a mean relative improvement of +18.8% over the baseline 2022 values. Effect size (Cohen’s d = 1.142) indicates a large magnitude of change.

The areas that experienced the most significant performance improvement were Via Goya (+0.205, +36.2% relative to 2022 baseline) and Via Neghelli (+0.110, +152.8% from a very low baseline of 0.072). Additionally, Via Goya (NDVI = 0.771 in 2024), Piazza Ilaria Alpi (0.540), and the productive park P.Le Dei Mercanti (0.508) registered the highest NDVI values in 2024.

In general, the green areas (P.Le dei Mercanti–park, Piazza Ilaria Alpi, and Via Goya) demonstrated substantially higher NDVI gains (mean Δ = +0.170 ± 0.055, n = 3) compared to urbanized parking areas (mean Δ = +0.056 ± 0.027, n = 4; t-test: t = 2.631, p = 0.041). Via Goya exemplifies this pattern, with the largest increase (+0.205) and the highest post-intervention NDVI value (0.771), indicating very lush vegetation.

Parking areas (P.Le dei Mercanti–site, Via Lepanto, Via Neghelli, and Via Romagnoli) showed more modest increases in the NDVI but were still positive. This suggests that despite urbanization, the UPPER interventions contributed to improved vegetation even in these zones. Via Lepanto (+0.061, +28.9%), Via Neghelli (+0.110, +152.8%) and Via Romagnoli (+0.033, +31.4%) all exceeded their 2022 baselines.

This indicates that UPPER interventions, whether in fully green or more urbanized environments, were effective in promoting vegetation growth, as evidenced by the positive trends in NDVI across different types of areas.

NDWI Variation (2022–2024). NDWI analysis (

Table 4. Variation in NDWI (2022–2024). The 2022–2024 increase is calculated as the difference between the index value in 2024 and that in 2022.

	Comune di Latina	P.Le dei Mercanti–Site	P.Le dei Mercanti–Park	Piazza Ilaria Alpi	Via Goya	Via Lepanto	Via Neghelli	Via Romagnoli
2022	−0.488	−0.188	−0.501	−0.464	−0.562	−0.256	−0.132	−0.156
2023	−0.495	−0.182	−0.464	−0.483	−0.552	−0.268	−0.166	−0.156
2024	−0.548	−0.233	−0.489	−0.528	−0.694	−0.301	−0.226	−0.182
Increment 2022–2024	−0.060	−0.044	0.012	−0.064	−0.132	−0.044	−0.095	−0.025

) revealed a consistent decline across all monitored areas between 2022 and 2024 (paired t-test: t = −2.156, p = 0.063), with a mean decrease of −0.066 ± 0.041 (95% CI: −0.109 to −0.023). This pattern indicates increased water stress on vegetation despite concurrent increases in vegetation density (NDVI). This counterintuitive result (higher vegetation cover coupled with lower soil moisture indices) highlights a critical water management challenge in Mediterranean semi-arid urban contexts Moreover, this phenomenon may be linked to rising temperatures or changes in precipitation, which have led to a reduction in water content within the plants.

Green areas exhibited more pronounced declines (mean Δ = −0.087 ± 0.034, n = 3) compared to parking areas (mean Δ = −0.052 ± 0.032, n = 4), suggesting that expanded vegetation canopy from NbS interventions has outpaced available water supply. Via Goya showed the largest absolute decline (−0.132, −23.5% relative to 2022), despite having the highest NDVI gain. This pattern indicates that the denser vegetation established through NbS has increased transpiration demand beyond what precipitation and irrigation systems can sustain. Piazza Ilaria Alpi declined by −0.064 (−13.8%), and P.le dei Mercanti–Park showed a slight improvement (+0.012, +2.4%), the only green area with a positive NDWI trend.

Urbanized parking areas experienced more modest but still significant NDWI reductions: Via Lepanto (−0.044, −17.2%), Via Neghelli (−0.095, −71.9%), and Via Romagnoli (−0.025, −16.0%). These declines suggest that water stress is not confined to green areas but is a system-wide challenge, likely driven by both climatic conditions (elevated temperatures, variable precipitation) and increased vegetation water demand from successful establishment of new plantings.

Overall, the increase in NDVI (an indicator of vegetation) has been accompanied by a decrease in NDWI (a moisture indicator), reflecting a situation where water availability, despite the greater density of vegetation, may be insufficient, causing water stress in plants. In this context, NDWI should be interpreted as an integrative response to both external climatic forcing and internal site management (irrigation scheduling, vegetation density, soil design), rather than as a direct proxy for the performance of NbS interventions in isolation. Moreover, since all NbS sites were designed with irrigation systems specifically for the ‘establishment phase’ (typically 18–24 months post-planting), if irrigation was systematically reduced or eliminated post-2023 per design specifications, NDWI decline represents successful transition to irrigation-independent vegetation rather than drought-induced stress.

NDMI Variation (2022–2024). The analysis of the NDMI variation between 2022 and 2024 shows an overall improvement across all monitored areas, with a general increase in vegetative and moisture conditions (

Table 5. Variation in NDMI (2022–2024). The 2022–2024 increase is calculated as the difference between the index value in 2024 and that in 2022.

	Comune Di Latina	P.Le dei Mercanti–Site	P.Le dei Mercanti–Park	Piazza Ilaria Alpi	Via Goya	Via Lepanto	Via Neghelli	Via Romagnoli
2022	0.084	−0.042	0.095	0.058	0.110	0.030	−0.096	−0.028
2023	0.070	−0.038	0.086	0.093	0.129	0.031	−0.087	−0.021
2024	0.199	0.008	0.174	0.124	0.344	0.075	−0.033	0.011
Increment 2022–2024	+0.115	+0.050	+0.079	+0.066	+0.234	+0.045	+0.063	+0.040

). All monitored areas recorded positive NDMI increases between 2022 and 2024 (paired t-test: t = 3.124, p = 0.016), with a mean improvement of +0.098 ± 0.072 (95% CI: +0.026 to +0.170). However, site-specific patterns reveal substantial heterogeneity between green and urbanized areas, reflecting differential management practices and water infrastructure capacity. The trends vary significantly between green areas (parks) and more urbanized areas (parking lots), reflecting the different management and land use practices.

Green areas demonstrated the most pronounced NDMI gains: Via Goya recorded a marked improvement, increasing from 0.110 (2022) to 0.344 (2024), an increment of +0.234 (+212.7%), the largest absolute gain across all sites and indices. Piazza Ilaria Alpi improved by +0.066 (+113.8%), reaching 0.124 in 2024, though this value still falls in the 0.0–0.4 range, indicating moderate water stress despite improvement. P.le dei Mercanti–Park showed a more modest increase (+0.079, +83.2%), with a final NDMI of 0.174 in 2024; this trajectory suggests ongoing water limitation despite green infrastructure management.

More urbanized areas, such as parking lots, registered less pronounced but still significant NDMI gains. Via Lepanto improved by +0.045 (+150%), reaching 0.075 in 2024 (moderate water stress). Via Neghelli increased by +0.063 (from −0.096 to −0.033), representing a transition from severe water stress (negative NDMI) toward neutral conditions, though still indicating stress. Via Romagnoli improved by +0.040 (+142.9%), reaching 0.011 in 2024, near the neutral threshold. P.le dei Mercanti–Site showed a significant increase (+0.050, +119%), reaching 0.008 in 2024, indicating near-neutral moisture conditions for this previously drought-stressed parking lot.

The divergence between NDWI (declining) and NDMI (increasing) patterns reflects the different physiological processes these indices measure: NDWI primarily responds to soil water availability, while NDMI reflects foliar water status and stomatal conductance. The NDMI improvements suggest that, despite declining soil water availability (NDWI), vegetation is maintaining or improving leaf water content through physiological adaptation (reduced transpiration, stomatal closure) or through irrigation provision during the establishment phase. This adaptive response, while positive for short-term vegetation survival, indicates sustained water-limitation stress that requires long-term management intervention. In conclusion, at this stage, green areas have benefited the most from interventions, with more marked improvements in moisture and vegetation conditions. However, urbanized areas have also shown significant improvements, although in some cases, they remain subject to water stress. The findings suggest that the interventions have had a generally positive impact, with more pronounced effects in green areas, but still relevant improvements in the more urbanized zones.

Synthesis and Interpretation. In synthesis, the 2022–2024 temporal analysis reveals a complex pattern of ecological change (Figure 7): NDVI improvements across all sites (mean +18.8%, p = 0.024) demonstrate that NbS interventions successfully established vegetation, with significantly larger gains in green areas (mean +27.0%) than in parking retrofits (mean +11.4%). However, this vegetation establishment has not been accompanied by proportional improvements in water availability, as evidenced by widespread NDWI declines (mean −0.066, p = 0.063), indicating that vegetation water demand has exceeded water supply despite the presence of irrigation and rainwater harvesting infrastructure at several sites. The positive NDMI trends (mean +0.098, p = 0.016) suggest partial physiological compensation through stomatal regulation and adaptive water-use efficiency, particularly in green areas where infrastructure maintenance (via Goya: +0.234 NDMI increment) has been prioritized. This adaptive response, however, masks ongoing water limitation that, if not addressed through enhanced water supply or irrigation management, may limit long-term vegetation resilience and sustainability in the Mediterranean climate context. These quantitative findings underscore the need for adaptive management strategies that balance vegetation establishment success with water resource constraints.

4.2. Meteo-Climatic Variables Influence on UPPER Sites

4.2.1. Exploratory Analysis of Meteo-Climatic and Air Quality Measurements in Comune Di Latina

The PCA of meteo-climatic variables for the Comune di Latina reveals clear temporal trends and highlights key factors influencing environmental variability between 2019 and 2024 (Figure 8). In this case, PCA was based on the analysis of PC2 capturing 15% of the total variance. By focusing on PC2, the analysis avoids the bias introduced by the limited data in 2024 and provides a more robust representation of the underlying environmental patterns. The PCA scores plot for PC2, reported in Figure 8a, indicates that 2019 and 2021 are distinctly separated from the other years; this suggests unique environmental or climatic patterns during these periods. Conversely, 2022 and 2024 exhibit similar scores, indicating comparable environmental conditions. The confidence intervals show consistent variation within each year, although some years, like 2019, display slightly greater variability.

The loadings plot, shown in Figure 8b, provides insight into the contributions of individual variables to PC2. In more detail, it can be observed that temperature-related factors, such as Average Temperature, Maximum Temperature, and ETO (Evapotranspiration), contribute strongly and positively to PC2, underscoring their primary role in driving environmental variability. This indicates that periods with higher temperatures and greater water demand are characterized by higher PC2 scores (i.e., the most affected year is 2019). In contrast, variables like PM10 Annual Average and Average Daily Radiation exhibit strong negative contributions, suggesting that higher values of air pollution and radiation correspond to lower PC2 scores (i.e., 2020, 2021 and 2023). While precipitation-related variables, including Rainy Days, Cumulative Rainfall Depth, and Maximum Rainfall Intensity, alongside PM2.5 Annual Average and Humidity, show moderate contributions.

Overall, the analysis highlights that temperature-related variables are the most significant drivers of environmental variability, closely followed by air quality (PM10) and radiation. Precipitation and humidity play smaller but notable roles. The temporal trends emphasize notable shifts in environmental conditions over the six-year period, with years like 2019 and 2021 standing out as distinct, potentially reflecting unique climatic events or anomalies.

4.2.2. Correlation Analysis of Climatic Variables and Vegetation Indices

Figure 9 shows the correlation analysis heatmaps for the Comune di Latina and the individual monitoring sites. The analysis reveals notable differences in the influence of climatic variables on vegetation indices across these areas, highlighting site-specific dynamics.

Comune di Latina. The overall analysis for Comune di Latina indicates strong correlations between NDVI and NDWI (r = −0.98), suggesting a close inverse relationship between vegetation greenness and water availability. NDVI also showed strong negative correlations with meteorological variables such as ETO (r = −0.73), Maximum Temperature (r = −0.87), and Average Temperature (r = −0.85), emphasizing the role of thermal conditions in driving vegetation variability. Additionally, NDWI and NDMI exhibited a significant negative correlation (r = −0.93), further linking water and moisture indices. Conversely, rainfall-related variables and humidity showed weaker correlations with NDWI, suggesting that these factors have less influence on water availability trends compared to temperature and radiation.

Differences Among the Sites. For Piazza Ilaria Alpi, the NDVI and NDWI remained strongly correlated (r = −0.98), with NDWI and NDMI also showing a strong relationship (r = −0.86). However, correlations with rainfall-related variables (e.g., cumulative rainfall depth, r = −0.29) and humidity (r = −0.48) were weaker than in the overall Comune di Latina, suggesting reduced precipitation influence on vegetation indices at this site. For P.Le dei Mercanti–Park, the temperature variables were the dominant drivers, with NDVI showing strong correlations with Maximum, Average, and Minimum Temperatures (e.g., r = −0.84). Rainfall-related variables and humidity had much weaker correlations (e.g., Max rainfall intensity, r = −0.14; humidity, r = −0.39), indicating limited precipitation influence in this vegetated park area. While for P.Le dei Mercanti–Site, the NDVI and NDWI exhibited moderately strong correlations (r = −0.92), though slightly lower than in the park. Temperature variables continued to drive NDVI trends (e.g., NDVI and Max Temp, r = −0.88), though the influence was less pronounced than in the demonstration park. Even for Via Goya, strong correlations between NDVI and NDWI (r = −0.97) and NDWI and NDMI (r = −0.89) were observed, like Piazza Ilaria Alpi. However, rainfall-related variables (e.g., Max rainfall intensity, r = −0.24) showed weak correlations, suggesting that precipitation has minimal influence on vegetation dynamics in this area. Concerning Via Lepanto, the correlations between NDVI and NDWI (r = −0.91) were slightly weaker than in other sites, and NDWI and NDMI showed only moderate associations (r = −0.45). Temperature and radiation variables exhibited consistently strong correlations across indices (e.g., NDVI and Max Temp, r = −0.78), while rainfall-related variables (e.g., Max rainfall intensity, r = −0.14) had very weak relationships. The site of Via Neghelli showed the weakest overall correlations. For instance, NDWI had negligible correlations with rainfall-related variables (e.g., cumulative rainfall depth: r = 0.00; rainy days: r = −0.01). NDVI and temperature variables (e.g., Max Temp, r = −0.46) also displayed weaker correlations compared to other areas, indicating minimal climatic influence on vegetation indices at this site. Finally, for Via Romagnoli, the NDWI and NDMI had negligible correlations (r = −0.04), while NDVI showed moderate correlations with meteorological variables, such as Max Temp (r = −0.61). However, these correlations were weaker than those observed in the overall Comune di Latina data.

Key differences among the analyzed sites can be seen in Temperature Sensitivity, Rainfall and Humidity Influence, NDWI-NDMI Relationship and Overall Climatic Sensitivity. In more detail, sites such as P.Le dei Mercanti–Park and Via Lepanto exhibited stronger correlations between NDVI and temperature variables, indicating higher sensitivity to thermal conditions. While rainfall-related variables displayed minimal correlations with vegetation indices at sites like Via Neghelli, Piazza Ilaria Alpi, and Via Lepanto, that can suggest limited precipitation influence. On the other hand, the strength of the NDWI-NDMI relationship varied across sites, with the strongest correlations in Piazza Ilaria Alpi (r = −0.86) and Via Goya (r = −0.89), while Via Lepanto had the weakest correlation (r = −0.45). Finally, it is observed that Via Neghelli emerged as the site least influenced by climatic variables, while P.Le dei Mercanti–Park and Comune di Latina showed stronger overall correlations, indicating a greater dependency of vegetation indices on environmental conditions. These findings emphasize the variability in climatic influence across urban and peri-urban areas. The weaker correlations in certain sites, such as Via Neghelli, may reflect the success of NBS interventions, potentially decoupling vegetation indices from broader climatic trends. Conversely, areas with stronger correlations (e.g., P.Le dei Mercanti–Park) highlight greater environmental sensitivity, which could inform future climate adaptation and management strategies.

4.3. Attribution and Methodological Limitations

Although the 2022–2024 analysis reveals consistent improvements in NDVI and NDMI and site-specific patterns that align with the design and management of UPPER NbS interventions, the current monitoring framework does not allow strict causal attribution of these changes to NbS alone. The study is based on observational time series without formally established control sites, and the detected trends therefore reflect the combined influence of local NbS implementation and broader regional drivers such as climate variability, background vegetation dynamics, and seasonal fluctuations. The inclusion of the municipality-wide “Comune di Latina” as a synoptic reference series provides contextual information at the urban scale but does not substitute for true control plots. Therefore, the results should be interpreted as evidence of association rather than proof of causality, and future work should incorporate explicit control areas and experimental designs to more robustly isolate the effect of NbS from regional-scale environmental dynamics.

A further limitation concerns the unequal number of Sentinel-2 images available per year. While annual index values are based on all cloud-free scenes (cloud cover ≤5%), the 2024 means are derived from only 5–7 images per site, compared with more than 40 in some pre-intervention years. As a result, 2024 averages may be more sensitive to the timing of individual acquisitions (e.g., specific phenological stages or short-term climatic anomalies) and should be interpreted with caution. Nevertheless, the consistency of the direction of change across all AOIs and indices, together with the statistically significant paired t-test results, suggests that the main trends reported are robust, although they would benefit from confirmation with a longer post-intervention time series.

A third limitation is the absence of site-specific quantitative hydrological data. Since irrigation volumes, rainwater harvesting performance, and soil moisture profiles were not measured, this precluded attribution of NDWI/NDMI changes to (i) irrigation scheduling, (ii) harvested rainwater utilization, or (iii) climatic precipitation deficits. For these reasons, future monitoring protocols should implement: (i) automated irrigation flow meters to track applied volumes and timing; (ii) rainwater harvesting performance metrics; and (iii) continuous multi-depth soil moisture profiles (0–30 cm surface, 30–100 cm root zone) to enable complete water budget analysis distinguishing management-driven from climate-driven effects on spectral moisture indices.

5. Conclusions and Future Perspectives

This follow-up research underscores the positive outcomes of UPPER interventions in Latina Municipality, demonstrating their effectiveness in promoting vegetation growth across various urban and green areas. The observed increase in NDVI between 2022 and 2024 indicates enhanced vegetation density, particularly in green areas, while urbanized spaces have also shown significant improvement. Within the Nature-Based Solutions (NbS) theoretical framework, these NDVI trends can be interpreted as evidence of successful ecosystem establishment and regenerative capacity, which represent core NbS principles related to ecosystem restoration and multifunctionality. However, the concurrent decrease in NDWI highlights water stress as a key challenge, suggesting that greater vegetation density has increased water demand, which is not fully met by available resources. Rather than indicating a failure of NbS interventions, this pattern reflects an intrinsic trade-off that is widely acknowledged within the NbS framework, especially in Mediterranean semi-arid contexts, where vegetation expansion may amplify competition for limited water resources. This emphasizes the need for integrated and adaptive water management strategies to balance vegetation growth with long-term hydrological sustainability. At the same time, the observed increase in NDMI provides additional insight into vegetation functioning, indicating improved leaf-level moisture status and adaptive physiological responses. When interpreted jointly, the divergent trajectories of NDVI, NDWI and NDMI reveal that NbS operate as dynamic ecological systems, where increased biomass (NDVI) may coexist with reduced soil water availability (NDWI), while vegetation maintains functionality through adaptive water-use efficiency and stomatal regulation (NDMI). This integrative interpretation directly aligns with the NbS conceptualization of resilient, self-regulating systems rather than static, equilibrium-based solutions.

The exploratory analysis of meteo-climatic and air quality data revealed that temperature-related variables (e.g., maximum and average temperatures) are the most significant drivers of environmental variability, followed by air quality (PM10) and radiation. Precipitation and humidity, though secondary factors, still contribute to environmental trends. Temporal analysis between 2019 and 2024 further revealed significant environmental shifts, with years like 2019 and 2021 showing unique climatic patterns, likely due to specific anomalies. These results reinforce the NbS framework assumption that climatic forcing (particularly thermal stress) plays a dominant role in shaping the long-term performance and limits of urban nature-based interventions, thereby highlighting the importance of adaptive management under changing climate conditions.

The correlation analysis highlighted site-specific dynamics in climatic sensitivity. For instance, P.Le dei Mercanti–Park and Via Lepanto exhibited strong correlations between NDVI and temperature variables, indicating higher sensitivity to thermal conditions, while rainfall-related variables showed minimal influence across most sites. Conversely, weaker correlations in areas like Via Neghelli may reflect the success of NBS interventions to partially decouple vegetation performance from broader climatic variability through site-specific design and management choices. This spatial heterogeneity confirms that NbS effectiveness is strongly context-dependent, as emphasized in NbS theory, and underscores the necessity of localized, site-adapted strategies rather than uniform implementation approaches.

In summary, UPPER interventions have fostered vegetation growth while also revealing issues related to water stress, underscoring the need for ongoing adaptation and monitoring in urban areas. The combined analysis of vegetation, soil moisture and plant water status demonstrates that multi-indicator monitoring is essential for capturing the complexity of NbS performance and avoiding oversimplified success metrics. However, while the observed trends are consistent with expected NbS dynamics, the absence of dedicated control sites and the influence of regional-scale climatic variability limit the possibility of unambiguous causal attribution. In this context, future work should include dedicated control plots (e.g., unsealed urban areas without NbS) and before–after–control–intervention (BACI) experimental designs to better isolate NbS effects from background environmental drivers.

This study highlights the fundamental importance of monitoring and evaluating NbS in urban settings. The effectiveness of a multi-scale approach for tracking environmental changes is demonstrated by combining EO data with proximal sensing techniques. Both methods revealed consistent trends across the areas studied; however, it remains of fundamental importance to carry out local assessments of vegetation health to address species-specific requirements and environmental differences. This integrated monitoring strategy enhances understanding of NbS performance, supports evidence-based decision-making, and provides a transferable methodological reference for other Mediterranean and semi-arid urban contexts. Additionally, it offers a holistic perspective on environmental impacts, which is essential for crafting effective NbS strategies. As technology evolves, these monitoring techniques will improve NbS management, promoting more sustainable and resilient urban environments.