From Space to Stream: Combining Remote Sensing and In Situ Techniques for Comprehensive Stream Health Assessment

Abstract

Urban streams undergo significant ecological alterations due to urbanization, including hydrological changes, water contamination, and biodiversity loss. This research employs a combination of satellite and drone imagery alongside traditional chemical and geophysical methods, facilitating a multi-dimensional assessment of Almyros and Gazanos urban stream health in Heraklion (Crete, Greece). The satellite imagery, obtained from the Copernicus program, allows for monitoring land use and impervious surface density around the streams, while drone surveys capture high-resolution images and calculate various water quality indices. In addition, chemical analyses of water samples for pollutants, as well as geophysical measurements using spectral induced polarization (SIP) and electromagnetic scanning (GEM-2), provide insight into the integrity of aquatic and riparian ecosystems. The study reflects on the different types of anthropogenic pressure faced by these two ecosystems. Almyros stream exhibits signs of eutrophication, characterized by elevated levels of chlorophyll and the presence of algal blooms, possibly due to runoff from adjacent agricultural activities. Conversely, the Gazanos stream shows signs of pollution mostly related to urbanization. The findings emphasize that both streams are under increasing anthropogenic pressure, thus highlighting the importance of employing comprehensive methods for effective stream management and policy implementation. This study ultimately advocates for ongoing monitoring initiatives that embrace technological advancements to safeguard urban water ecosystems.

1. Introduction

Urbanization is growing along with the population of the Earth. Most people will be living in cities by 2050 as urban areas expand and rural populations decline; only one in seven people will reside in rural areas [1]. The United Nations’ predictions that 68% of the world’s population is projected to reside in cities by 2050 [2] provide further evidence for this. This “movement” refers to the demand for larger cities, sustainable infrastructure, and an ever-widening resilient metropolitan environment.

The development of the urban environment/ecosystem brings forth significant new ecosystem types like urban blue spaces, which refers to all bodies of water found within the urban environment [1]. While the term “blue space” encompasses various water bodies (oceans, rivers, lakes, and ponds), in this study, we focus on urban streams, i.e., waterways that run through inhabited areas and are influenced by the urban landscape and human activities. Because of the pressures brought on by the surrounding city, urban streams differ from the natural ones [3]. Urban streams are differentiated by several characteristics that are not always the same, but are similar in most cases. These characteristics include alterations in their hydrology, which result in more frequent, higher magnitude flows of shorter duration along with reduced base flows because of the increased impervious surface found, which increases surface runoff and lowers infiltration [2]. Significant changes can also be observed in the chemical composition of water. The water in these cases seems to have high concentrations of nutrients, salts, and other compounds that may pollute. Such pollutants typically include heavy metals, pesticides, petroleum byproducts, and even pharmaceuticals [4]. Also, elevated levels of nitrogen and phosphorus can be detected due to the confluence of fertilizer runoffs, wastewater effluent, and septic systems [3,5]. Urban streams are also often vulnerable to having their morphology altered by civil engineering. They are often channelized and lined with concrete for flood control, thus sacrificing their natural complexity. These changes have direct effects on ecological processes, often reducing the biodiversity of the ecosystem [6,7]. Finally, another effect of urbanization on the streams is the reduction of biological diversity both inside the water and outside. This is credited to the combined effects of the degraded water, altered flow of water, and altered habitat. The effects on biodiversity are often detected in the lack of species that are adapted to cleaner waters and complex habitats, and the dominance of the more resistant species like pollution-tolerant algae, invertebrates, and fish [2,8,9]. In case an urban stream exhibits some or all of these characteristics, it has been impacted by the Urban Stream Syndrome, a term used to characterize the ecological deterioration of streams that flow through urban areas. Some key impacts of this syndrome are loss of sensitive taxa, reduced channel complexity, and increased pollutants found within the channel. Thus, more resistant species like pollution-tolerant algae, invertebrates, and fish manage to dominate the ecosystem [10,11,12]. Although blue spaces have been linked to human well-being [13,14,15] and are becoming a popular subject of research, the monitoring needed to assess their health remains lacking. This gap in knowledge limits our ability to efficiently monitor the water and soil health of the urban stream ecosystem.

Traditional water and soil quality methods of assessment, including on-site sampling and laboratory analysis, are often constrained by spatial and temporal limitations [16]. To address these challenges, remote sensing (RS) technologies, particularly those using Unmanned Aerial Vehicles (UAVs), have emerged as effective tools for large-scale, high-frequency water and soil quality surveillance. UAV-based multispectral and hyperspectral remote sensing enables rapid detection of pollution sources and temporal variations in water quality parameters [17,18]. Furthermore, the use of UAV imagery with machine learning (ML) models has improved the accuracy and efficiency of water quality parameter inversion. For instance, Zheng et al. (2024) [19] demonstrated that feature selection techniques, such as Relief Feature Ranking with Recursive Feature Elimination, enhance the accuracy of ML models for dissolved oxygen, total nitrogen, turbidity, and chemical oxygen demand. Similarly, Chen et al. (2021) [20] employed genetic algorithm–extreme gradient boosting (GA_XGBoost) to predict water quality parameters with high precision, outperforming conventional ML models.

The fusion of multi-source data, such as satellite imagery, meteorological variables, and land use information, has further improved water quality monitoring approaches. Liang et al. (2025) [16] showed that integrating Sentinel-2 multispectral data with land use features and meteorological elements enhances ML model performance, particularly in predicting ammonia nitrogen, total phosphorus, chemical oxygen demand, and dissolved oxygen. Zhang and Yang (2024) [21] proposed a two-stage multidirectional fusion model incorporating probabilistic matrix factorization, achieving robust predictions with reduced dependence on extensive ground samples. Moreover, self-optimizing ML approaches, as demonstrated by Chen et al. (2023) [22], offer solutions for overcoming the scale inconsistencies in multi-source RS data, facilitating more comprehensive urban monitoring. The application of these advanced methodologies is not only limited to monitoring water pollutants but also extends to assessing the overall ecological health of urban streams. Kokolakis et al. (2024) [23] highlighted the significance of integrating remote sensing with GIS-based approaches for evaluating soil and water quality in urban environments, emphasizing the role of continuous monitoring in preserving ecological integrity and human well-being.

The primary goal of this study is to evaluate the health status of urban streams by implementing different monitoring techniques. The approaches selected for this evaluation were RS and in situ methods (chemical [24] and geophysical studies). Evaluating the ecosystem’s health and whether the stream has been impacted by the Urban Stream Syndrome is its secondary objective.

The challenges identified in this research regarding the health assessment of urban streams include the following:

• Complexity of Urban Influences: The urbanization surrounding the streams leads to alterations in hydrology and alterations in chemical composition, making it difficult to discern the impacts of human activities from natural variances. This complexity can complicate assessments of ecological health and environmental quality.
• Monitoring Limitations: Although remote sensing provides broad spatial coverage, it may lack the finer granularity needed to capture local conditions effectively. In situ measurements are required for detailed assessments, but limitations on access to certain areas can restrict data collection efforts.
• Anthropogenic Pressures: Both streams of this study are under considerable anthropogenic pressure from urbanization and agricultural practices. Monitoring and managing these pressures require ongoing efforts and resources, which can be challenging to maintain over time.
• Resource Constraints: Advanced monitoring techniques often require considerable investment in technology and expertise. Limited resources can hinder the establishment of comprehensive monitoring programs, crucial for effective stream management.
• Technological Challenges: RS and in situ methods are fundamentally different, posing a challenge for the combination of monitoring techniques for effective decision-making.

Overall, while this study advocates for combined monitoring and management approaches, these challenges must be overcome to ensure successful implementation and accurate assessment of urban stream health.

Research Sites

Two (2) urban streams, i.e., Almyros and Gazanos streams, in the Western Heraklion basin of Crete, Greece, are examined in this article (Figure 1).

Almyros is a UNESCO site on Crete in the basin of Heraklion (Figure 1), 8 km from the city of Heraklion on the north-eastern edge of Psiloritis [25]. A 12 m dam near the spring and the beginning of the stream has elevated stream outflow from 6 m to 10 m. This was completed with hopes of uplifting the spring level to (a) increase hydraulic pressure, (b) prevent seawater intrusion, and (c) improve groundwater quality [26]. The stream itself is 1.8 km long, and its width varies between 5 m to 20 m. The flow of the Almyros stream varies throughout the year with an average discharge of 235 × 10⁶ m³. Although the Almyros stream provides enough water to cover the needs of the wider Heraklion region, its salinity makes it impossible to be consumed by humans [27]. Almyros, as the name suggests, when translated into Greek, is indeed salty. The high salinity of the stream is what sets it apart from the rest of the rivers and streams on the island [27]. The brackish nature of the stream is due to the mixture of seawater and freshwater in the aquifer that feeds the spring. This seawater intrusion is possibly due to the overpumping of groundwater, the low elevation of the spring, and the geological characteristics of the area. It should be noted that this phenomenon is more intense in the dry months [26]; conversely, during periods of high rainfall, the flow rate of the stream increases and the salinity decreases, making the water potable in some rare cases [26,27,28,29,30].

Almyros is a karstic wetland, the karst system is a key geological characteristic of the area. Karst systems are shaped by the dissolution of soluble rock, predominantly limestone and dolomite, by acidic water. This leads to unique features such as sinking streams, enclosed depressions, caves, and large springs, all connected by a network of underground conduits [29,30]. Furthermore, the Almyros basin is found on Neogene and pre-Neogene geological units. Its bedrock consists of carbonate rocks, specifically the semi-autochthonous “Platenkalk” unit, dating back to the Triassic–Eocene period [29,31]. Although these characteristics make for a distinguishable ecosystem, they also make the wetland more vulnerable to pollution. Pollutants can easily pass through the porous carbonate rocks and spread rapidly throughout the system. Thus, the pollutants persist for longer within the interconnected underground channels, hindering the wetlands’ recovery from contamination [29,32].

A multitude of plant species that are representative of the Mediterranean climate and the wetland’s distinctiveness can be identified in the Almyros wetland. In the riparian zone of the stream, a mixture of trees, annual plants, and perennial plants can be detected [29,33,34]. A significant portion of these plants are adapted to the distinctive characteristic environmental conditions, including brackish water, high salinity, and the long periods of drought during the summer months (

Table A1. Almyros’ riparian zone plant life [28,29].

Plant Name	Plant Type	Description
Eucalyptus sp.	Perennial tree	Known for rapid growth and adaptability
Olea europea	Perennial tree	Cultivated for olives and oil; a long-lived Mediterranean species
Ceratonia siliqua	Perennial tree	Known as Carob, which is drought tolerant
Tamarix sp.	Perennial shrub/tree	Salt-tolerant; invasive in some regions, used for soil stabilization
Phoenix theophrasti	Perennial palm	Endemic to Crete, drought tolerant
Rhamus alaternus	Perennial shrub	Evergreen shrub
Acacia retinodes	Perennial tree	Known as Swamp Wattle; ornamental and drought tolerant
Nerium oleander	Perennial shrub	Toxic evergreen shrub, used in landscaping
Euphorbia characias	Perennial shrub	Drought-tolerant with unique floral structures; commonly ornamental
Urticia dioica	Perennial herbaceaous	Stinging nettle; used in herbal medicine and as a food source
Nicotina glauca	Perennial shrub	Tree tobacco: invasive in some areas, thrives in arid regions
Drimia maritima	Perennial herbaceous	Known as Sea Squill; medicinal and ornamental uses
Phragmites australis	Perennial grass	Common reed, used in water filtration and habitat restoration
Eryngium maritimum	Perennial herbaceous	Sea holly, ornamental coastal plant with blue spiky flowers
Glebionis coronaria	Annual herbaceous	Garland chrysanthemum, edible leaves used in Mediterranean cuisine
Glycyrrhiza glabra	Perennial herbaceous	Licorice plant: roots used for flavoring and medicinal purposes
Carex pseudocyperus	Perennial sedge	Found in wet habitats; used in erosion control and habitat restoration
Capparis spinosa	Perennial shrub	Caper bush, edible flower buds are used in cooking.
Cota tinctoria	Perennial herbaceous	Dyer’s chamomile; used historically in dyeing fabrics
Matricaria chamomilla	Annual herbaceous	German chamomile.
Salicornia europea	Annual herbaceous	Known as glasswort; thrives in saline conditions, edible shoots
Ricinus communis	Perennial shrub/tree	Castor bean plant: seeds produce castor oil, highly toxic
Cakile maritime	Annual herbaceous	Sea rocket, edible coastal plant adapted to saline soils
Sonchus oleraceus	Annual herbaceous	Sow thistle, a fast-growing weed used for fodder and human consumption
Visnaga daucoides	Biennial herbaceous	Medicinal plant
Arundo donax	Perennial shrub	Giant Reed: invasive in some regions, used for biomass and erosion control
Galactites tomentosus	Annual Herbaceous	Ornamental thistle, native of the Mediterranean

, Appendix A), with the most common plants being Eryngium maritinum, Olea europaea, Visnaga daucoides, Nerium oleander, Phragmites australis, and Phoenix theophrasti (Figure 2).

Almyros is endangered by human activity and urbanization [29] while being a UNESCO-protected site. The surrounding area of Almyros, despite not being highly urbanized as shown in the impervious density (IMD%) map of Figure 3, is affected by human activity because (a) the southern portion of the stream, near the dam, is used for agricultural activities (primarily for olive groves), which are possibly the main cause of the stream’s eutrophication [29] (Figure 4), and (b) the northeastern portion, which is closer to the beach, is mainly occupied by hotels and recreational activities. In addition, there is a functioning desalination plant on the stream that treats brackish water for human consumption [30].

The second site, investigated in this work, is the Gazanos stream (Figure 1). Gazanos also belongs to the Heraklion basin and is about 8 km away from the city center of Heraklion. The total length of the stream is 35 km, while the urban section that is examined in this work is 5.57 km. Gazanos stream has a seasonal flow with an annual discharge of 21.22 hm³ [35,36]. Gazanos, compared to Almyros, fits more with a Mediterranean stream, experiencing seasonal drying and flooding depending on the variable seasonal precipitation. It also showcases spatial variability, emerging as a mosaic of different flow conditions [33]. Gazanos’ riparian zone is home to plants that are expected in a Mediterranean stream, which are a mix of perennial and biennial plants (

Table A2. Gazanos’ riparian zone plant life [33,34].

Plant Name	Plant Type	Description
Robinia hispida	Perennial shrub	Also called Bristly Locust; nitrogen-fixing and used for erosion control. Adapted to various soils and climates
Glaucium flavum	Perennial herbaceous	Known as Yellow Horned Poppy; drought-tolerant coastal plant with medicinal uses
Arundo donax	Perennial grass	Giant Reed: Invasive in some regions, used for biomass and erosion control.
Visnaga daucoides	Biennial herbaceous	It has a history of being used both as food and poison
Ecballium elaterium	Perennial herbaceous	Known as squirting Cucumber: grows in arid conditions, noted for its medicinal properties but toxic if ingested
Ficus benjamina	Perennial tree	Known as Weeping Figure, a popular indoor plant that thrives in tropical and subtropical climates
Nerium oleander	Perennial shrub	Evergreen shrub, toxic, commonly used in landscaping
Eucalyptus camaldulensis	Perennial tree	River Red Gum: fast-growing, thrives in dry regions, often planted for timber and erosion control
Pandorea jasminoides	Perennial vine	Jasmine-like flowering vine, popular in ornamental horticulture
Ricinus communis	Perennial shrub/tree	Castor Bean Plant: seeds yield castor oil, though highly toxic
Eucalyptus sp.	Perennial tree	Known for rapid growth and adaptability
Olea europea	Perennial tree	Cultivated for olives and oil; a long-lived Mediterranean species

, Appendix A). The plants found on this stream are characterized by their resistance to seasonal droughts. They are also adapted to flooding, exhibiting rapid root extension, a low shoot-to-root ratio, and flexibility to avoid being uprooted by the floods. The most common plants around Gazanos stream are Arundo donax, Visnaga daucoides gaertn, Eucalyptus comoldulensis, Nerium oleander, and Olea euroepea (Figure 5). It is important to note that Arundo donax and Eucalyptus comoldulensis are invasive species [37,38] with the first being highly spread out within the margins of the stream, completely covering any other plant species in some areas. (Figure 6).

Gazanos, compared to Almyros, as shown later in this study, is more heavily influenced by anthropogenic intervention (Figure 3). The southern part of the Gazanos stream is mostly affected by agricultural activities and is surrounded by suburban infrastructure. Urbanization has a greater impact on the Gazanos stream closer to the sea because more bridges cross it, and there are houses and businesses along its edges. Finally, the northeastern section, which faces the sea, is dotted with hotels and is impacted by the same leisure activities as Almyros stream (Figure 3).

2. Materials and Methods

Figure 7 outlines the step-by-step procedure for the stream assessment in this work. The process begins by choosing the right stream sites and then by assessing the land use patterns and impervious surface densities for the chosen urban streams. Simultaneously, an evaluation of the chemical and environmental factors is made to determine the baseline conditions. In parallel, RS is utilized in the calculation of indices for water quality and vegetation health assessment, as well as geophysical investigations regarding streams’ water and soil, and subsurface structure. Statistical methods are used to identify significant patterns and correlations among the water’s physicochemical parameters and various dissolved elements. The results of this work are then discussed and presented in a way that effectively conveys the findings.

2.1. Satellite and Drone Data Acquisition and Processing

As previously stated, the primary objective of this study was to monitor and assess the health of the urban streams Almyros and Gazanos by combining RS and in situ methods. The most popular RS techniques [39,40] were selected for this purpose, supported by chemical and geophysical “in situ” and laboratory methods. Regarding the RS techniques, the study was conducted utilizing both satellite and drone imagery for the Almyros stream and satellite imagery (SI) for the Gazanos stream due to Gazanos’ higher level of urbanization (Figure 3).

The streamlines of SI, used to visualize impervious density, were acquired from the Copernicus program [41], specifically the Copernicus Land Monitoring Service (CLMS) [42], and then were processed in QGIS and ArcGIS Pro. The impervious density data, the streamlines, and the land usage data have been downloaded from CLMS [43,44,45,46]. Concerning the land uses, after visualization, a buffer zone of 200 m was created around each stream to quantify the land usage. The DJI Mavic 3M, which combines an RGB sensor with a multispectral camera, was utilized for drone imagery [47]. On 4 and 5 June 2024, before the airborne imagery acquisition took place, a radiometric correction with a calibration reflectance panel was completed. After that, the drone took more than 7000 pictures and covered 373,078.471 m² in these two days.

The collected satellite data were processed in QGIS and ArcGIS to support the mapping process of both streams and provided information regarding the land uses and the impervious density of land around both streams [43,45], helping to better visualize the information. Finally, the UAV data that were collected during the two-day campaign in the summer of 2024 were processed using Pix4D 4.5.6 Mapper software to obtain the orthomosaic, digital terrain model (DTM), and the digital surface model (DSM). The resulting products were then imported to QGIS and ArcGIS for further processing to (a) create a digital map of the Almyros stream area; (b) determine the sampling stations (Figure 8); (c) complete export of the whole area surrounding the stream; and (d) calculate the vegetation, soil, and water indices used to monitor the overall health of the stream. The final choice of indices was Normalized Difference Chlorophyll Index (NDCI), Normalized Difference Algae Index (NDAI), Maximum Chlorophyll Index (MCI), and Cyanobacteria Index (CI) (

Table A3. Indicators used to monitor stream health from the previous [29] and the current research on Almyros and Gazanos streams.

Water Indices and Formula	Soil Indices and Formula		Geophysical indicators	Chemical Analyses
				Physicochemical Parameters	Nutrients and Pigments
Normalized Difference Chlorophyll Index (RE1 − R)/(RE1 + R)	Soil Adjusted Vegetation Index (1.0 + L) × (N − R)/(N + R + L)		Real conductivity (SIP)	pH	Nitrogen
Normalized Difference Algae Index (NIR − R)/(NIR + R)	Green Normalized Difference Vegetation Index (N − G)/(N + G)		Imaginary conductivity (SIP)	Temperature	Phosphorus
Maximum Chlorophyll Index RE − (N + (N + RE) * (RE − R/N − R))	Normalized Difference Chlorophyll Index (RE1 − R)/(RE1 + R)		Apparent Conductivity (GEM-2)	Conductivity	Ammonium
Cyanobacteria Index RE − (G + (N − G) * (RE − G/N − G)				Dissolved Oxygen	Chlorophyll-a
					Carotenoids

, Appendix A).

It should be noted that the dense vegetation surrounding the Gazanos stream (Figure 6) made it impossible to obtain enough data from UAVs or satellites to compute remote sensing indices.

2.2. Water and Soil Sampling

Both the water (Figure 8, indicated by I–IV) and soil (Figure 8, indicated by 1–4) samples for Almyros were collected in May 2024 for geophysical analysis using spectrally induced polarization (SIP) at four sampling locations (Figure 8). All water samples were obtained from a depth of 10 cm beneath the water surface and loaded into 1.5 L PTE bottles. The bottles were washed with tap water, deionized water, tap water, and then again with deionized water. All samples were analyzed promptly upon returning to the laboratory. Finally, regarding the soil samples, they were also collected near the water sampling sites using a shovel. The soil samples were kept in plastic cases, air-dried, crushed, and sieved to retain fractions below 2 mm to minimize the noise coming from air, water, and pebbles during the analysis for SIP. The same procedure was followed for the Gazanos stream, with two water samples (Figure 8, indicated by I–II) and one soil sample (Figure 8, indicated by 1).

2.3. Geophysical Analyses of Water and Soil Samples

For the geophysical analyses of water and soil, the portable field/laboratory spectrally induced polarization (PSIP) instrument was used, a tool that allows for in situ and laboratory measurements of SIP, conventional resistivity, self-potential, and induced polarization (IP) in the time domain. Conrad Schlumberger proposed the IP method 100 years ago, and it has been utilized ever since [48] to study the biogeochemical state, flow, and transport properties of soils and water, as well as fluid content and chemistry [29,48,49,50,51,52]. This technique allows the estimation of real conductivity (RC) and imaginary conductivity (IC) by measuring the phase shift and the amount of conductivity of an injected current. The measured RC is the energy loss (conductivity), and the IC corresponds to the energy storage (polarization) [52,53]. In this research, PSIP was employed to measure the SIP response of water and soil samples in the frequency range 0.1 Hz–1000 Hz from Almyros and Gazanos streams, following the same procedure as in [29]. The frequency range 0.1 Hz–1000 Hz is essential in IP measurements as it helps to characterize hydrogeological properties like hydraulic conductivity and delineates features such as contaminant plumes and metallic mineral accumulations. In environmental studies, it tracks changes in redox conditions and supports contamination evaluations.

Furthermore, geophysical electromagnetic induction (apparent electrical conductivity) was conducted for both streams. The GEM-2 from Geophex [54] was used to acquire the electromagnetic data of the traverses along lines (indicated in Figure 8 using orange color) for both Almyros and Gazanos streams. The frequencies of operation for the surveys in both streams were 20,025 Hz, 47,025 Hz, and 90,025 Hz. GEM-2’s ability to use multiple frequencies allows for the measurement of responses at multiple depths, with lower frequencies providing information for deeper depths while higher frequencies provide information for shallower depths, enabling a more complete understanding of the subsurface conditions. These frequencies were chosen because they provide multi-depth imaging with 20,025 Hz penetrating deeper, while 47,025 Hz and 90,025 Hz provide better insights into near-surface features. They also provide an optimal noise-to-signal balance [55,56]. Electromagnetic induction has been used in mineral exploration to discriminate and identify probable deposits [57], and it has also been used in archaeological excavations to map the apparent electrical conductivity to help locate buried structures, features, and even artifacts [58]. Its most important ability is that by mapping subsurface conductivity variations, it can be used to delineate the extent of contamination plumes, identify buried waste, and assess the effectiveness of remediation efforts [59,60].

2.4. Chemical Data

The chemical analyses provided in this study were completed as part of a governmental program regarding management plans for 14 river basins nationwide, including Almyros and Gazanos streams [24]. The streams were monitored from 2012 to 2021 for different parameters affecting the health of the streams, such as pH, electrical conductivity (EC), temperature, inorganic (lead, arsenic, and phosphates, etc.), and organic pollutants (pesticides, polyaromatic hydrocarbons, etc.). The above-mentioned analyses were complemented with additional analyses conducted in 2022 [29].

2.5. Statistical Processing

Statistical processing was conducted using Microsoft Excel along with JASP 0.19.3 software, which is free open-source software backed by the University of Amsterdam, and it offers standard analysis methods in both classical and Bayesian forms [61]. Both Microsoft Excel and JASP 0.19.3 were used to investigate correlations (Pearson’s r correlation coefficient) between the different physicochemical parameters and other compounds in the water.

2.6. Study Requirements

To avoid errors that could arise from the combination of RS and other instruments, certain measures must be taken. Regarding the RS methodology, it is important, before using the drone to collect data, to do radiometric correction with a calibration reflectance panel. Also, all data obtained via RS instruments should be backed by in situ measurements. Concerning the geophysical measurements using PSIP and GEM-2 ski, they have been calibrated according to the manufacturer’s guidelines.

3. Results

3.1. Remote Sensing Imagery

3.1.1. Water Indices

To examine the overall state of the urban streams in this work, over a hundred water indices have been examined. The main criteria for choosing among the indices were to be (a) able to detect pollution, (b) applicable on drones, and (c) applicable on small water bodies.

In 2023, a related study was undertaken on the Almyros stream, which was then classified as eutrophic [29]. This conclusion is also backed by the findings of this study, utilized by the NDCI [62] corresponding to chlorophyll content in the Almyros stream. As shown in Figure 9, the purple color corresponding to eutrophication dominates.

NDAI is used to detect areas with microalgae blooms [56,63]. In the case of Almyros stream, microalgae were detected along the stream, further establishing eutrophism (Figure 10) [57,64].

Next, MCI was used to measure the concentration of phytoplankton in the 10–300 mg/m³ range [58,65], but since it was not possible to further quantify the concentration with chemical analyses, the data were visualized in the 0 to 1 range. This is to showcase the existence of phytoplankton/chlorophyll within the stream’s margins (Figure 11). Higher values (brown) are concentrated along the main stream channel, reflecting areas of significant biological productivity, while lower values (yellow) are observed in surrounding regions, corresponding to areas with less chlorophyll presence (Figure 11).

CI was employed as a supplemental approach to detect cyanobacteria blooms [66]. Similarly, to MCI, this indicator was processed in such a way that the results could be quantified on a scale of 0 to 1, with 1 representing a higher concentration of cyanobacteria and hence chlorophyll-a (chl-a) (Figure 12).

3.1.2. Land and Soil Indices

Following the examination of water indices, it was also vital to evaluate the status of the riparian zone around the stream since it (a) helps maintain water quality by absorbing excess nutrients like phosphorus and nitrogen, (b) acts like a natural filter for pollution, and (c) supports microbial activity, which in turn promotes a healthier stream with activities like denitrification [67,68,69]. To inspect the riparian zone, multiple indices were tested with criteria matching those of the water indices. The indices used were the Soil Adjusted Vegetation Index (SAVI), the Green Normalized Vegetation Index (GNDVI), and the Normalized Difference Chlorophyll Index (NDCI).

To examine the plant’s stress in the riparian zone of Almyros, SAVI was used. SAVI is designed to quantify the plants’ stress found in a designated area while minimizing the influence of the soil background [70]. The outcome of this index indicates that most of the plant life is stressed, as expected, since the data acquisition was conducted during the summer of 2024 (Figure 13); however, plants on agricultural land on the western side of the Almyros stream and some areas near the stream’s margins to the east appear to be exempt (Figure 13).

Then, GNDVI was utilized. Since it is more sensitive to chlorophyll than the Normalized Difference Vegetation Index (NDVI), it seems to be a superior choice for the various types of plants in the Almyros wetland [71]. Again, in the western portion of the stream, where agricultural activity is obvious, plant life appears to be healthier, as do plants located on the stream’s edges in the eastern half of the stream (Figure 14).

Finally, NDCI, which, while not commonly used for land monitoring, follows the same principles as NDVI [62,72], allowing monitoring of chlorophyll even on land (Figure 15) with findings comparable to NDVI. Again, as with the GNDVI, agricultural plants appear to be healthier based on NDCI. What distinguishes the results of this index is that it identifies some spots on the eastern side of the estuary where the stream overflows.

3.2. Land Use

Understanding the impact of nearby human activities on the wetland is crucial to evaluating the health of the wetland’s ecosystem. The land uses around the Almyros and Gazanos streams were thus visualized by the processing of land use data from CLMS [42] (Figure 16a). To isolate the direct effects on Almyros and Gazanos streams, a 200 m buffer zone was later established around each stream [73]. The data obtained from this process was statistically examined (Figure 16b,c) to gain insights into the land use type that has the greatest influence on the streams. It is evident from Figure 16b, which depicts the land use in Almyros, that agricultural activities account for 24% of the uses that have an impact on the stream, with urban buildings and roads accounting for 26%, and the remaining 8% belongs to other exerted anthropogenic effects. With more than 60% of land uses being agricultural and 35% being other elements, including highways, urban structures, commercial units, and industrial activities, Gazanos (Figure 16c) exhibits a large degree of pressure from anthropogenic causes in comparison to Almyros. Since human activity and urbanization account for more than 90% of the total uses, Gazanos is subject to higher overall pressure. With 41% of the influence belonging to humans within the 200 m buffer zone, Almyros maintains a healthier condition than Gazanos.

3.3. Geophysical and Physicochemical Evaluation

Remote sensing analyses were complemented with ground measurements that took place in two phases; phase 1 was the geophysical measurements regarding water and soil, and phase 2 was the chemical measurements of the stream water.

Geophysical Measurements

The SIP response of Almyros and Gazanos’ water and soil samples is shown in Figure 17 and Figure 18, respectively. The sampling station’s locations are indicated in Figure 8 with blue (water) and red (soil) dots. Specifically, Figure 17a,b present the IC and RC of water samples from the Almyros stream. In Figure 17a (water IC), the values range from 50 to −50 mS/cm, and they seem to be similar to each other while descending from 1000 Hz to 0.25 Hz, where, after that frequency, a sharp descent is noted in the sample coming from Sampling Site I. In Figure 17b (water RC), the values are around 15.5 mS/cm for all sampling stations. In Figure 17c (soil IC), the values generally are around 0 mS/cm; it should be noted that from 1000 Hz to 63 Hz, a transient descent is present for Sampling Site 4, while the other sampling stations reveal similar values close to 0 mS/Hz. In Figure 17d (soil RC), the values range from about 0 mS/cm to 4 mS/cm. The results coming from these samples seem to have different values from each other, with Sampling Site 3 standing at the top with values of a little bit more than 4 mS/cm, Sampling Sites 2 and 1 coming second with conductivity around 2 mS/cm, and Sampling Site 4 coming last, with its conductivity nearing 0 mS/cm.

Figure 18a presents the IC response of two (2) water samples from Gazanos stream, where the values range from 0.125 mS/cm to 0 mS/cm, and there is a transient descend of the IC for Sampling Site I while moving from 1000 Hz to 15 Hz until its values are similar to Sampling Site II which from 1000 Hz to 0.1 Hz remained close to 0 mS/cm. Figure 18b (water RC) shows values ranging from 3 mS/cm to more than 2 mS/cm. Sampling Site I’s values are around 3 mS/cm, and Sampling Site II’s are around 2.25 mS/cm. Figure 18c (soil IC) shows a slight increase after 10 Hz, followed by a sharp fall at 0.40 Hz. Figure 18d (soil RC) presents values ranging around 0.6 mS/cm. The SIP results of this work regarding the Almyros stream are similar to those of Kokinou et al. (2023) [29].

Figure 19 and Figure 20 illustrate the results of GEM-2 electromagnetic scanning (at frequencies 20,025 Hz, 47,025 Hz, and 90,025 Hz) regarding the distribution of the soil apparent EC below the surface of Almyros and Gazanos riparian parts, respectively. The locations of the electromagnetic scanning (GEM-2) are in Figure 8. The results of the 350 m long GEM line (a) of the Almyros stream, which runs parallel to the southern portion of the stream (Figure 8), are shown in Figure 19a. The apparent EC of the GEM line (a) primarily ranges between 0 mS/m and 410 mS/m. The results of the 20 m long GEM line (b) of the Almyros stream, which is oriented vertically to the southern portion of the stream (Figure 8), are displayed in Figure 19b with an apparent EC ranging between 15 mS/m and 210 mS/m. The GEM line (c), which is 40 m long overall and likewise oriented vertically to the southern portion of the stream, displays apparent EC values that range from 40 mS/m to 250 mS/m. Based on Figure 19a–c, it is likely that there are three strata (layers) in the southern portion of the stream up to a depth of roughly 8 m. With a mean apparent EC of 157 mS/m, the shallowest layer appears to stretch down to a depth of roughly 1.5 m, while the underlying layer, which extends down to a depth of roughly 2.3–2.8 m, exhibits a mean apparent EC of 112 mS/m. Lastly, the thickest layer, which reaches a depth of roughly 8 m, has an apparent EC of 79 mS/m.

The apparent EC distribution along the 75 m long GEM line (d), which runs parallel to the stream and is situated close to the Almyros stream’s bifurcation and the artificial channel, is shown in Figure 19d. There are two subsurface strata in this section of the stream. At a depth of around 3 m, the shallowest layer exhibits an apparent EC of 258 mS/m, while the underlying layer exhibits an apparent EC of 232 mS/m, extending up to 8 m; however, these two layers could be considered as one layer corresponding to a mean apparent EC of 245 mS/m.

Almyros GEM lines (e), (f), and (g) (Figure 19e–g) are located near the point where this stream empties into the sea (Figure 8); line (g) is parallel to the stream, while lines (e and f) are vertical to the stream. Two separate layers can be seen in line (e), which has a total length of 100 m. The shallowest layer exhibits an apparent EC of roughly 996 mS/m (down to about 3 m), while the underlying layer displays an apparent EC of around 877 mS/m, extending down to 8 m (Figure 19e). Line (f) is 75 m long overall, and up to 32 m from the stream bank, two distinct layers are visible. The shallowest layer has an apparent EC of roughly 244 mS/m (down to about 3 m), while the underlaying layer has an apparent EC of 227 mS/m, extending down to 8 m (Figure 19f); however, these two layers could be considered as one layer corresponding to a mean apparent EC of 235.5 mS/m. The entry into a more urban setting probably causes a decrease in the mean apparent EC response (135 mS/m) at 32 m (Figure 19f). The total length of the GEM Line (g) is 130 m, and one layer extends from the surface to a depth of 8 m. This layer reveals a mean apparent EC of roughly 1379 mS/m (up to 62 m away from the stream bank) and of roughly 450 mS/m at distances greater than 62 m as it approaches the urban environment (Figure 19g). Lastly, the 380 m long GEM Line (h) runs parallel to the artificial canal (Figure 8). Up to 8 m in depth, three subsurface layers can be identified along this line (Figure 19h). The shallowest layer appears to extend up to a depth of roughly 1.5 m, revealing a mean apparent EC of 278 mS/m, while the underlying layer, which extends to a depth of roughly 2.3 m to 2.8 m, exhibits a mean apparent EC of 245 mS/m. Lastly, the apparent EC of 207 mS/m is revealed by the deepest layer of the GEM line (h), which reaches up to around 8 m.

Figure 20a shows the results of the Gazanos stream’s 25 m GEM line (a), which is vertical to the stream bank (Figure 8). The apparent EC primarily falls between 8 mS/m and 75 mS/m. The GEM line (b) of the Gazanos stream, which is 85 m long and parallel to the stream bank (Figure 8), shows apparent EC values ranging from 8 mS/cm to 70 mS/cm (Figure 20b). In the mid-section of the Gazanos stream, up to a depth of roughly 8 m, three layers are possibly present, as shown in Figure 20a,b. With a mean apparent EC of 62 mS/m, the shallowest layer appears to stretch down to a depth of roughly 1.5 m, while the underlying layer, which extends down to a depth of roughly 2.3–2.8 m, displays a mean apparent EC of 45 mS/m. Lastly, the thickest layer, which reaches a depth of roughly 8 m, has a mean apparent EC of 12 mS/m.

3.4. Chemical Analysis

The data used in this work have resulted from the fusion of data provided by a governmental program [24] and recent published works [27,29]. Regarding the streams that are included in this work, a comprehensive list of pollutants has been compiled for both the Almyros (

Table A4. Examined pollutants in Almyros stream [24].

Aluminum dissolved	Chromium dissolved	Mercury dissolved	HCO3
Arsenic dissolved	Copper dissolved	Nickel dissolved	Nitrate
Cadmium dissolved	Dissolved oxygen	Potassium	Nitrite
Calcium	Iron dissolved	Sodium	Total phosphates
Carbonates	Lead dissolved	Sulphate
Chloride	Magnesium	Zinc dissolved
Chromium 6+	Manganese dissolved	Ammonium

, Appendix A) and the Gazanos (

Table A5. Examined pollutants in Gazanos stream [24].

1,1-Dichloroethene	1,1,1-Trichloroethane	1,1,2-Trichloroethane	1,1,2,2-Tetrachloroethene	1,2-Dichlorobenzene	1,2-Dichloroethene	1,2-Dichloroethane	1,3-Dichlorobenzene
1,4-Dichlorobenzene	2-Chlorotoluene	2,2′,3,3′,4,4′-Hexachlorobiphenyl	2,2′,3,4,4′,5-Hexachlorobiphenyl	2,2′,3,4,4′,5,6-Heptachlorobiphenyl	2,2′,3,4,5-Pentachlorobiphenyl	2,4-D	2,4,5-T
3,4-Dichloroaniline	4-Chloroaniline	4-Chlorotoluene	4-Nonylphenol	Aclonifen	Alachlor	Aldrin	Alpha-Endosulfan
Alpha-HCH	Ammonium	Anthracene	Arsenic	Atrazine	Azinphos ethyl	Azinphos methyl	Bentazone
Benzene	Benzo(a)pyrene	Benzo(b)fluoranthene	Benzo(g,h,i)perylene	Benzo(k)fluoranthene	Beta-Endosulfan	Beta-HCH	Bifenox
BOD5	Cadmium	Chlorfenvinphos	Chloridazon	Chlorobenzene	Chlorpyrifos	Chromium	Chromium 6+
Cis-1,2-Dichloroethene	Cl	Cobalt and its compounds	Copper	Coumaphos	Cyanides (as total CN)	Cybutryne	Cyclodiene total
Cypermethrin	DDD, p,p″	DDE, p,p″	DDT total	DDT, o,p″	DDT, p,p″	Delta-HCH	Demeton O+S
Demeton-S-Methyl	Detergents	Di(2-ethylhexyl) phthalate (DEHP)	Dichloromethane	Dichlorprop (2,4-DP)	Dichlorvos	Dicofol	Dieldrin
Dimethoate	Dissolved Oxygen	Disulfoton	Diuron	Endosulfan	Endrin	Ethylbenzene	Fenitrothion
Fenthion	Fluoranthene	Gamma-HCH (Lindane)	Heptachlor	Heptachloroepoxide	Hexachlorobutadiene (HCBD)	Hexachlorocyclohexane (HCH)	InvertebrateEQR
Isodrin	Isoproturon	Lead	Linuron	M-Xylene	Malathion	MCPA	Mecoprop
Mercury	Meta + Para Xylene	Methamidophos	Mevinphos	Molybdenum and its compounds	Monolinuron	Naphthalene	Nickel
Nitrate	Nitrite	O-Xylene	Omethoate	Orthophosphates	Oxydemeton-Methyl	P-Xylene	Para-tert-Octylphenol
Parathion	Parathion-Methyl	PCB’s Total	PCB101 (2,2′,4,5,5′-Pentachlorobiphenyl)	PCB105 (2,3,3′,4,4′-Pentachlorobiphenyl)	PCB114 (2,3,4,4′,5-Pentachlorobiphenyl)	PCB153 (2,2′,4,4′,5,5′-Hexachlorobiphenyl)	PCB156 (2,3,3′,4,4′,5-Hexachlorobiphenyl)
PCB169	PCB170	PCB180	PCB194	PCB28	PCB52	Pentachlorophenol	PFOS
Phenol	Phenols	Propanil	Quinoxyfen	Selenium and its compounds	Simazine	Terbutryn	Tetrachloromethane
Tin and its compounds	Toluene	Total Dissolved Solids	Total Phosphorus	Trans-1,2-Dichloroethene	Triazophos	Tributyltin	Trichlorfon
Trichloromethane	Trifluralin	Zinc

, Appendix A) streams for the years 2013, 2014, 2018, 2019, and 2020. It is noteworthy to mention that the pollutants (Table A4 and Table A5, Appendix A), which are included in the list of priority pollutants intended to be reduced and eventually eliminated from the discharge of wastewater according to Directives (2013/39/EC, 2000/60/EC), were found below the maximum allowed concentrations for surface waters. The chemical analysis data were cleaned and statistically analyzed using the JASP 0.19.3 software to determine correlations between each parameter/compound for the Almyros and Gazanos streams. Heatmaps for each stream were created to better visualize correlations (Figure 21a,b). These heatmaps visualize the Pearson correlation coefficients between various dissolved elements and physicochemical parameters in water samples. Darker shades indicate stronger positive or negative correlations, while lighter shades represent weaker relationships. In Figure 21a, the heatmap shows the correlations between water level, physicochemical parameters, and compounds detected in the Almyros stream. Strongly positive correlation is present between chloride and dissolved copper (0.95), while negative correlations are found between a) dissolved oxygen (DO) and dissolved iron (−0.67), DO and nitrate (−0.85), and pH and water level (−0.72). In Figure 21b, the heatmap shows correlations between physicochemical parameters and compounds in the Gazanos stream. Positive correlations are present between the total dissolved solids (TDS) and DO (0.73), nitrate and nitrite (0.68), and DO and nitrate (0.53). Furthermore, negative correlations can be seen between pH and total dissolved solids (−0.64) and total dissolved solids and temperature (−0.558).

4. Discussion

This study effectively demonstrates the combination of RS technologies with in situ (ground) measurements (Table A3, Appendix A) to carry out a comprehensive environmental evaluation of the Almyros and Gazanos urban streams. The findings reveal critical insights into the environmental health [74,75] of these water bodies, illustrating the significant impacts of both agricultural runoff and urbanization on stream quality.

4.1. Almyros Stream: Eutrophication and Algal Blooms

Regarding the Almyros stream assessment through multiple RS (Table A3, Appendix A), water indices (Figure 9, Figure 10, Figure 11 and Figure 12) designate that the stream is mostly eutrophic or in danger of becoming eutrophic according to NDCI, an indicator designed to detect chlorophyll concentrations and algal blooms in complicated water systems [62,76]. NDCI classified [77,78,79] Almyros stream in the most part as “high eutrophication risk” (yellow color) or eutrophic (purple) (Figure 9). This is further supported by NDAI (Figure 10), which can also identify algal blooms, with an emphasis on microalgae blooms [63]. The outcome of NDAI demonstrates an increased prevalence of microalgae [80] along the stream, which was noticeable even with the naked eye during data acquisition in the field (Figure 3). Algal blooms are a byproduct of the stream’s eutrophication [58,66]. The increased concentration of chlorophyll in the stream is also reinforced by the results of MCI and CI (Figure 11 and Figure 12) [81,82]. These indices were used as complementary tools since further quantification of their results was not possible through chemical analysis. Despite this limitation, the results of MCI and CI provide insights into the health of the stream. MCI allows for the identification of chlorophyll absorption based on algal cells (Figure 11) that further reinforces the results of NDAI (Figure 10) [67,68]. Finally, CI detects cyanobacteria blooms (Figure 12), which are a result of eutrophication [69,70]. Figure 22 depicts the hyperthesis of the four indices used in this study (Table A3, Appendix A). Specifically, the zoomed sections (a) and (b) in Figure 22 exhibit identical high values in common places for all the indices tested, suggesting that the eutrophication discovered using chemical approaches [29] persists to this day.

The key takeaway from the above-mentioned indices (Table A3, Appendix A), used to monitor the riparian part of Almyros wetland (Figure 13, Figure 14 and Figure 15), is that regions around the stream, particularly in the western half, are used for agricultural operations because they appear to be less stressed than the “wild” plant life. Figure 13 depicts SAVI, with higher values (red color) in the western part of the stream indicating healthy, moisture-rich, and unstressed plants [62]. In Figure 14, GNDVI appears to follow the same trend as SAVI, with greater values (red color) occurring in the western half of the riparian part, where agricultural activity is more visible. These high values imply that the plants are chlorophyll-rich, healthy, and actively growing [71,83]. Finally, NDCI (Figure 15) appears to follow the same pattern, since it has high values (red color) in the same places as both SAVI and GNDVI. Even though this index was designed to assess algae and water quality by measuring chlorophyll in water, larger values also signal robust plant photosynthetic activity, allowing for the detection of healthy plants, though it is less precise, as shown in Figure 15.

When attempting to monitor the health of a stream, it is also necessary to conduct in situ measurements (Table A3, Appendix A) to supplement remote sensing analyses. The heatmap (Figure 21a) regarding Almyros stream presents strong positive correlations (blue) between (a) chloride and dissolved copper (0.94), indicating a possible shared source. The negative correlations (red) between (a) copper and dissolved iron (−0.67) highlight the possibility of differing geochemical niches [21]; (b) water level and pH (−0.72) indicate that pH is reduced as water levels rise.

Concerning the geophysical analysis (Table A3, Appendix A), the SIP responses are similar to prior research published in 2023 [29], exhibiting identical values in both RC and IC for water samples (Figure 17a,b). The similarities apply even to IC, which shows alterations in the frequencies between 0 and 100 Hz (Figure 17a), indicating medium polarization effects [51,53,84]. These changes are likely caused by nutrition or photosynthetic pigments present in chlorophyll (Table A4, Figure 9, Figure 10, Figure 11 and Figure 12). Peaks in the frequency range 100–10 Hz region are present in IC, particularly at Sampling Sites III and IV, indicating slightly increased polarization effects, which could be caused by interactions between dissolved ions and colloidal particles. Sampling site I shows a quick reduction at lower frequencies, indicating a low concentration of polarizable materials [85,86]. RC appears almost steady, reflecting the ionic concentration that dominates the response, with little influence from frequency-dependent polarization processes [86]. It is noteworthy that sampling sites III and IV have somewhat higher RC, most likely due to higher salinity, as they are closer to the sea. This is consistent with the larger polarization responses in IC at these sites.

The IC response of the soil samples (Figure 17c) of Almyros stream possibly indicates weak soil polarization effects, which may be related to clay content, organic matter, and the presence of metallic minerals [87]. The results from sampling site 4 (green) show a swelling of IC at higher frequencies and a decrease at lower frequencies (Figure 17c). These variations possibly indicate a high concentration of clay minerals or organic debris, both of which have been shown to enhance polarization [88]. Lower frequency fluctuations may be caused by soil heterogeneity (a mixture of conductive and non-conductive elements). Sampling sites 1, 2, and 3 (Figure 17c) have low IC across all frequencies, indicating weak polarization effects and the potential dominance of coarser soil components with little charge storage capacity [86]. Regarding the RC of the soil samples (Figure 17d), Sampling Site 3 (yellow) exhibits the highest RC (nearly 4 mS/cm) across all frequencies. This suggests either a high salinity environment, which is conceivable given Almyros’ salty water, higher ion exchange capacity, or pollution from dissolved inorganic materials [89]. The soils from Sampling Sites 1 and 2 (blue and red) present similar RC (2 mS/cm) while the soil from Sampling Site 4 (green) has the lowest RC (1 mS/cm), indicating lower ionic content, drier soil, or a larger concentration of non-conductive materials [90].

The results of the GEM-2 electromagnetic conductivity (Figure 19a–h) show variation in apparent EC across different distances (Figure 8). The measurements at various frequencies (20,025 Hz, 47,025 Hz, and 90,025 Hz) provide information about subsurface properties that are affected by factors such as soil moisture, salinity, lithology, and possible contamination [91,92]. The apparent EC values at 20,025 Hz are consistently lower than 47,025 Hz and 90,025 Hz, as expected given that lower frequencies penetrate deeper into the ground [93]. The strong peaks seen are most likely the result of localized water saturation and fine-grain sediments. Where apparent EC decreases, it indicates a transition to less conductive or coarser sediments [53,94]. These patterns may also be influenced by factors such as agricultural runoff, salinity variations, or other human activities [25,45,71].

All findings of this study come together when viewed in the context of land use for Almyros wetland (Figure 16a,b), which shows that 43.8% of the 200 m buffer zone around the stream is natural, consisting of 31.7% semi-natural grassland, 8.1% sclerophyllous scrubs, and 1.7% lines of trees and scrub. The remaining 58.4% of land use within the buffer zone is devoted to human activities, the majority of which are agricultural, with olive tree production accounting for 23.1%. The fact that agriculture takes up a significant portion of the human-affected region around the Almyros stream could explain the nutrients found in the water, as well as the stream’s eutrophication.

The findings of this study highlight that the Almyros stream is experiencing pronounced signs of eutrophication, evidenced by elevated chlorophyll concentrations and substantial algal blooms. These conditions have been closely connected to agricultural practices in the surrounding areas, particularly runoff that contributes excess nutrients like nitrogen and phosphorus into the stream. Eutrophication not only degrades water quality but also disrupts aquatic ecosystems by promoting harmful algal blooms that can reduce oxygen levels and produce toxins that are harmful to aquatic life and human health [33].

4.2. Gazanos Stream: Impact of Urbanization

Moving on to the Gazanos stream, as previously stated, gathering satellite and drone data was impossible due to the stream’s dense vegetation cover and the fact that parts of the stream are private property. As a result, the indices utilized in the Almyros stream could not be calculated. As seen in Table A5 (Appendix A), the list of contaminants discovered in Gazanos water is more extensive than Almyros’ (Table A4, Appendix A), indicating the higher impact of anthropogenic processes on the health of the stream. The correlation heatmap (Figure 21b) shows strong positive correlations (blue) between (a) TDS and DO (0.73); this relationship is possibly driven by photosynthetic activity in nutrient-rich waters; (b) nitrate and nitrite (0.68), underlining an ongoing nitrification process in the stream. On the other hand, negative correlations between (a) TDS and pH (−0.64) and temperature (−0.55) suggest that, as the water warms, TDS diminishes, possibly through seasonal hydrological shifts.

The water SIP measurements for Gazanos (Figure 18a,b) revealed that IC in Sampling Site I is significantly higher at higher frequencies and declines as frequency drops. This may be happening due to the presence of small, suspended particles, reactive mineral surfaces, or metal–organic complexes that improve charge storage mechanisms [95]. At Sampling Site II, the response is essentially flat, indicating minimum polarization, most likely due to the reduced concentration of polarizable elements. The fast decrease in IC at higher frequencies is frequently observed in water contaminated by fine sediments, organic debris, or pollutants [90,96]. The RC at Sampling Site I is higher than at Site II, indicating a higher concentration of dissolved ions, most likely due to industrial or agricultural runoff [53,86]. In the soil sample obtained from the Gazanos stream, IC does not present significant variations (Figure 18c). This could mean that no significant electrochemical processes and capacitive effects take place [88]. It is also worth noting that a severe dip at 0.16 Hz is followed by a quick increase (Figure 18c). This has been observed in soils with high clay content or organic matter, where charge storage processes shift at lower frequencies [97]. The RC reduces slightly as the frequency decreases (Figure 18d). The results from GEM-2 (Figure 20a,b) indicate the presence of three subsurface layers with apparent EC ranging between 8 mS/m and 75 mS/m.

As the chemical analyses indicated, the Gazanos streams are under pressure from human activity and urbanization. This is most evident in the buffer zone (Figure 16a,c) around the stream, where only 3% of the land remains natural, primarily sandy beaches, while 33.4% is occupied by urban structures, and 63.6% is used for agricultural purposes.

Gazanos stream, while not classified as eutrophic, exhibits signs of pollution linked to urbanization—a trend that poses its own set of challenges. The study highlights a critical reduction in natural land cover along the stream’s banks, with over 62% of the surrounding land being utilized for agriculture and urban infrastructure [98]. This rapid urban expansion results in increased impervious surfaces, which exacerbate runoff and pollutant loads. The correlation drawn between urbanization metrics, such as impervious surface density, and the water quality indicators serves as a compelling call for sustainable urban planning practices.

4.3. Combined Monitoring Approach

The results support the efficacy of a combined RS and in situ monitoring strategy. By combining RS technologies—such as drone and satellite imagery—with traditional chemical and geophysical methods (Table A3, Appendix A), we achieve not only a more thorough understanding of environmental conditions but also enable more timely and cost-effective monitoring [1,30]. This combination allows for extensive spatial coverage and detailed local assessments, establishing the detection of pollution sources and trends over time. Moreover, such an integrated approach provides indications about the health of the urban population living in the area, as findings from the COVID-19 era indicate [99].

Furthermore, employing indices like NDCI and SAVI provided key insights into the health of riparian ecosystems. The results indicated that areas of the riparian zone with active agricultural use exhibited healthier vegetation than less managed or natural areas [14], but it should be mentioned that research has also indicated that surrounding imperviousness of an urban environment has a negative relationship with both regulating and provisioning ecosystem services of an urban stream [100]. This insight underscores the complex relationship between land management practices and aquatic health, suggesting that sustainable agricultural practices can coexist with healthy stream ecosystems [39].

Combined environmental monitoring of urban streams (Table A3, Appendix A) can provide valuable insights into the health of the city. Comprehensive data are essential to the development of healthy cities, embedded in a resilient and sustainable ecosystem. A systematic program of assessing urban streams, combining satellite and in situ monitoring, can help support meaningful long-term urban planning. A major limitation in this kind of study is that it is difficult to monitor urban streams that are overwhelmed with vegetation using satellites and UAVs, like the case with the Gazanos stream in this study. However, this can be overcome if the regional and local administration cooperates with the researchers and removes the excess vegetation. Another major limitation is the lack of research that exists about combining such methods to monitor the health of urbanized streams that suffer from the urban stream syndrome [100,101,102]. In contrast, the existing literature that showcases a similar kind of combination of methodologies seems to provide excellent insights into the health status of the streams [22,103].

5. Conclusions

In this work, remote sensing (satellites and UAVs) was combined with in situ measurements (SIP, GEM-2, chemical analyses) to assess the environmental status of two (2) urban streams, Almyros and Gazanos, located in the western part of the Heraklion basin (Crete, Greece). This combination is an enhanced process to monitor stream health by providing spatiotemporal coverage with RS, offering wide area surveillance and high-resolution data, while in situ measurements provide data that captures detailed local conditions, ensuring that remote observations are correct (Table A3, Appendix A). Complementary data types with water and soil indices allow us to estimate and quantify features such as chlorophyll concentration on the surface of the water, enabling us to rapidly assess water quality and potential algal blooms. Laboratory techniques, like SIP, can indicate the existence of contaminants, and chemical analyses provide direct assessments of the water quality and validate remote sensing data. This toolset also offers enhanced validation and calibration to accurately reflect water and soil quality while offering an improved early warning and monitoring system (Table A3, Appendix A). Also, this system is applied to the particularly vulnerable to pollution karstic stream of Heraklion, offering an important case study for understanding anthropogenic impacts on hydrologic systems. Finally, this method is cost-effective and efficient since RS can cover large areas with low cost, making field trips to every part of the stream redundant, with a few strategically placed in situ samplings. We recommend that this monitoring system method could be part of urban health planning and monitoring to drive interventions for resilient and sustainable development.

In conclusion:

• Combined Monitoring Approach: This study successfully demonstrates the effectiveness of combining RS techniques, including satellite and drone imagery, with in situ measurements such as Spectral Induced Polarization (SIP), GEM-2, and chemical analyses. This combination provides a comprehensive, multi-dimensional evaluation of the health of urban streams, critical for timely and accurate assessments (Table A3, Appendix A). We recommend the creation of a dashboard that presents comprehensive data from longitudinal studies.
• Ecosystem Health Insights: The research findings highlight that the Almyros stream is experiencing significant eutrophication, as indicated by high chlorophyll levels and algal blooms primarily due to agricultural runoff. On the other hand, the Gazanos stream, while not classified as eutrophic, is burdened by pollution resulting from urbanization.
• Importance of Land Use Analysis: Analysis of land use surrounding the streams indicated that agricultural practices significantly impact the Almyros stream, accounting for over 40% of the influencing factors. Conversely, the Gazanos stream is heavily influenced by urban development, which comprises 90% of the land use in its vicinity. This finding highlights the critical need to understand local land use impacts for effective stream health assessments.
• Role of Geophysical and Chemical Assessments: The combination of RS data with SIP, GEM-2, and chemical assessments provides valuable insights into subsurface conditions and contaminant distribution. Geophysical and chemical assessments improved the precision of pollution detection and the understanding of its sources, aiding in more effective management interventions.
• Call for Ongoing Monitoring: There is an urgent need for continuous monitoring programs that use integrated approaches to track changes in stream condition over time. Regular data collection using both RS and in situ methods can facilitate the development of early warning systems to detect ecological degradation and enable timely interventions to protect urban aquatic ecosystems.

Finally, this study underlines the importance of taking a holistic perspective when assessing stream health. The interweaving of physical, chemical, and biological data underscores the need for a structured monitoring framework to address the challenges that urbanization and agricultural practices pose to aquatic ecosystems and to ensure their sustainability and resilience for future generations. Practical application measures, such as buffer zone management, regular geophysical monitoring, and pollutant emission limit standards, should be considered in future stream management.