DRASTIC-Fm-URBAN index: an updated and reliable GIS vulnerability mapping for the assessment of fractured rock media in urban areas

Fractured media hydrosystems sources are valuable worldwide for irrigation, water supply, and industrial purposes. Mapping and evaluating intrinsic groundwater vulnerability is a balanced integrative methodology to describe the potential groundwater contamination or pollution. The Geographical Information System (GIS) overlay mapping technique and multi-parametric methodology are largely employed to evaluate vulnerability in groundwater systems. However, the DRASTIC index methodology is often used for regional approaches and is the most recognised for groundwater vulnerability assessment for general purposes. Therefore, several adaptations were proposed, such as the DRASTIC‐Fm methodology focused on fractured rock media. This study proposes an updated multi-parametric DRASTIC-Fm-URBAN methodology’s concept, importance, and applicability, which supports the potential groundwater vulnerability on fractured media assessment for urban areas. This approach is more accurate, particularly on regional to local scales. The analytic hierarchy process (AHP) was applied to define the weight of each assessment parameter. In addition, the index DISCO-URBAN, SI and GODS were used in the historic springs of the Porto and Vila Nova de Gaia riverside urban area. In fact, coupling DRASTIC-Fm-URBAN and DISCO-URBAN vulnerability indexes are reliable approaches to be included as a tool to develop a robust hydrogeological conceptual site model for urban areas.


Introduction
The dependence on groundwater has increased significantly over the past three decades (e.g., Margat and van der Gun 2013;McGrane 2016;Foster et al. 2022;La Vigna 2022). Consequently, the reliance on groundwater has come with the need to protect the water resources from contamination. However, water quality issues are due to hazardous environmental pollutants and various anthropogenic activities in environments, such as mining, industry, urban runoff, sanitary landfill and agriculture (e.g., Barroso et al. 2015;Guimarães et al. 2019;Tahmina et al. 2020;Chaminé et al. 2022). The potential for groundwater contamination is impacted by the area's physical attributes, the contaminant or pollutant chemistry and rated occurrence, as well as the hydrogeological conditions (Shirazi et al. 2012). Groundwater vulnerability describes intrinsic attributes that define the water sensitiveness to be negatively impacted by a potential contaminant load. Human actions and intense agricultural activities deteriorate groundwater quality and endanger vulnerability (e.g., Shirazi et al. 2012;Foster and MacDonald 2014;Afonso et al. 2020;Chaminé et al. 2022;La Vigna 2022). In addition, the effect of the buried intricate system structures (e.g., sewer, storm sewers, pipes, tunnels) and impervious surfaces and/or cover areas should be underlined as anthropogenic disruptions in groundwater media (e.g., Wiles and Sharp 2008;Hibbs and Sharp 2012;Attard et al. 2016).
Assessing intrinsic vulnerability to groundwater contamination and/or pollution is valuable for defining safeguard zones in environments impacted by groundwater contamination, particularly in urban areas (e.g., Panagopoulos et al. 2006;Shirazi et al. 2012;Afonso et al. 2016Afonso et al. , 2020Freitas et al. 2019aFreitas et al. , b, 2021Fannakh and Farsang 2022). The DRASTIC method is a point count system model for evaluating intrinsic groundwater vulnerability methodology based on several parameters. This index was initially designed by Aller et al. (1987a, b) and is mainly based on geology, geomorphology, hydrogeology setting, climate constraints, land use/cover, and other parameters. However, many researchers have modified this method into updated models, such as DRASTIC-Fm (Denny et al. 2007). For example, Denny et al. (2007) propose a modification, combining statistical techniques and specific parameters, and optimises the rating of the DRASTIC model parameters to achieve a proper intrinsic vulnerability mapping focused on the fracture media. Furthermore, compared to the original DRASTIC model as the most revised or updated scheme, that approach improves the accuracy of the vulnerability maps. Thus, the updated or modified DRASTIC model is invaluable for supporting land use/cover planning and groundwater systems management (e.g., Gogu and Dassargues 2000;Gemitziet al. 2006;Panagopoulos et al. 2006;Denny et al. 2007;Shirazi et al. 2012;Sadat-Noori and Ebrahimi 2016;Anantha Rao et al. 2018;Kumar et al. 2019Kumar et al. , 2022Jenifer and Jha 2022;Taghavi et al. 2023).
The vulnerability mapping assessment in a given site, area or region is a powerful tool for evaluating the potential groundwater for contamination. This approach seeks the identification of zones with high contamination risk to assess the possibility of hydrosystems contamination (e.g., Gemitzi et al. 2006;Anantha Rao et al. 2018;Jenifer and Jha 2022). Moreover, it is an excellent tool to support the description, assessment, modelling and communication of groundwater resources (e.g., Robins 1998;Daly and Warren 1998;Witkowski et al. 2007;Kresik and Mikszewski 2013;Chaminé et al. 2015Chaminé et al. , 2016Pisinaras et al. 2016). Thus, vulnerability maps have become important in groundwater protection systems, delineating protection zones in fractured media and groundwater contamination hazard assessment. Furthermore, groundwater vulnerability maps are a valuable primary tool for communicating data related to planning, policy, and operational levels of decision-making. In addition, this facilitates the delineation of safeguard zones for potential contamination and groundwater resource management (Robins et al. 2006). The DRASTIC method's primary purpose is to address mapping groundwater susceptibility to contamination load. Moreover, this method is straightforward and based on geographic information systems (GIS), mostly on a regional scale perspective. Moreover, GIS-based mapping is an effective tool for analysing climatic, geological, geomorphological, hydrogeological and land-use data (e.g., Jha et al. 2007;Neshat et al. 2013;Kumar et al. 2015;Chaminé et al. 2015Chaminé et al. , 2016. An update to the initial DRASTIC methodology called "DRASTIC-Fm" is a consistent approach for regional fractured hard-rock aquifers evaluation (Denny et al. 2007). It comprises an additional parameter considering a higher fracturing degree of rock media connected with geological structures, such as joints, faults, shear zones and tectonized veins (Denny et al. 2007). However, the DRASTIC and DRASTIC-Fm indexes are trustworthy for evaluating the evolution of susceptibility of the groundwater systems to surface pollutants. Therefore, that model presents a visual and quantitative mapping tool oriented to planning and decision-makers challenged by the pressure on groundwater resources management (Erlandson 2014).
The main purpose of this research is to evaluate groundwater vulnerability to contamination by proposing, describing and discussing an updated methodology of the DRAS-TIC-Fm focused on urban areas grounded in fractured rock media called the DRASTIC-Fm-URBAN index. In addition, it is also assessed by an integrative approach using other vulnerability indexes, such as GODS (Foster et al. 2002), DRASTIC-Fm (Denny et al. 2007), SI (Ribeiro et al. 2017) and DISCO-URBAN (Meerkhan et al. 2021) methods.

Study area: framework
The studied urban region comprises the riverside area of Porto and Vila Nova de Gaia (NW Portugal), Fig. 1. Bordered by the river Douro, Porto is an ancient European city (Costa-Lobo 1991; Lochery 2020) with a historic centre that UNESCO has classified as a World Heritage Site. Both cities are settled on the hill slopes of the Douro riverside. The two urban regions assembled 535.624 inhabitants, corresponding to 5.2% of the total population of Portugal (INE 2022). The population density of Porto city is 5597 inhabitants/ km 2 , and Vila Nova de Gaia municipality is 1805 inhabitants/km 2 . The urban water supply in Porto and Vila Nova de Gaia cities was delivered throughout the centuries by fountains fed by numerous springs, water mines, and wells (e.g., Rebello da Costa 1788; Bourbon e Noronha 1885; Monteiro e Azevedo and Santos 1861; Carteado Mena 1908;Fontes 1908;Bahia Junior 1909;Chaminé et al. 2010Chaminé et al. , 2014Freitas et al. 2014Freitas et al. , 2019aFreitas et al. , 2021Afonso et al. 2016Afonso et al. , 2019. The urban area of Porto and the Vila Nova de Gaia riverside area are geomorphologically part of the Iberian Massif, where there is a flattened surface (littoral platform) that develops from altitudes of 125 m a.s.l and rises to 150 m a.s.l south of Porto (Araújo et al. 2003;Freitas et al. 2014Freitas et al. , 2019a. The littoral platform culminates in the base of the so-called marginal relief (Araújo et al. 2003), which corresponds to a steep relief and is roughly aligned along sub-meridian orientation, varying between NNW and NNE (Araújo et al. 2003;Afonso et al. 2004). The local geomorphology of the study area is characterized by alternating flattened areas (mostly 80 and 125 m a.s.l) and the bottom of the small valleys (Fig. 1b). The flattened areas vary in altitude from west to east and south to north. At the highest altitude, the flattened areas are found in the northeast of the city (between 130 and 160 m a.s.l), and the lowest is in the west, with altitudes below 25 m a.s.l .
The drainage network reflects the regional fracturing and morphotectonics whose main orientations are NW-SE to NNW-SSE, NE-SW to NNE-SSW and W-E. Most of the streams drain into the Douro river. In the studied urban areas, most of the streams that have been channelled are almost buried in the ground (Carríngton da Costa 1938;COBA 2003).
The NW part of mainland Portugal is an Atlantic type of climate, greatly affected by the direction of the relief (Alcoforado et al. 1982). The climatic classification of Köppen (Peel et al. 2007) pointed out that the urban region studied presents a type Csb (temperate climate with dry and mild summer). Rainfall is irregularly distributed throughout the year, with the dry season occurring between July and August (Fig. 1d). The rainy season is shaped by the winter months and late autumn. The average annual precipitation corresponds to 1236.8 mm. The months with the highest precipitation values correspond to December (181 mm) and November (172 mm).
The urban geology is characterised by a diverse granitic and metasedimentary substratum (mainly granite, gneiss, migmatite, micaschist, schist, and metagraywacke) and cover deposits (mostly sandy, gravely, silty and clayey sediments with a heterogeneous sorting) (e.g., Carríngton da Costa 1938;Almeida 2006;Chaminé et al. 2010). The bedrock is largely weathered to various grades, from fresh granitic rock to residual soil (saprolite mass), developing arenisation and kaolinisation profiles with depths of over 30 m (e.g., Begonha and Sequeira Braga 2002;Gaj et al. 2003;Afonso et al. 2007). The groundwater field inventory noticed that most of the springs and springs/fountains occur in fissured media and some in cover deposits (Fig. 1c). Furthermore, the groundwater is nearly neutral (pH 6.7), nitrate enriched (NO 3 > 50 mg/L), SO 4 -Ca to HCO 3 -Ca hydrochemical facies, average values of − 5.3‰ for δ 18 O and − 33.2‰ for δ 2 H, which shows these springs are shallow groundwater that reacts quickly to precipitation occurrences (Afonso et al. , 2020. In addition, the permeability is low to very low (< 1 m/day) in fissured rock media and moderate (< 4 m/day) in porous media (Afonso et al. , 2020. The works of Afonso et al. (2007Afonso et al. ( , 2016Afonso et al. ( , 2019Afonso et al. ( , 2020 described and presented in detail the urban groundwater of Porto city. Figure 2 highlights the Urban Potential Infiltration Index (IPI-URBAN) mapping developed by Freitas (2019). The IPI-URBAN defines the areas where, according to a set of factors (particularly, geology and morphotectonics; climate and hydrology; urban hydrogeology and hydrogeomorphology; urban hydraulics and sanitation), there are favourable conditions for water infiltration in the ground (details in Freitas et al. 2019a, b). Therefore, it can be concluded that most of the study region (89.2%) is characterised by a very low to moderate potential infiltration index ( Fig. 2A). This can be justified by the degree of urbanisation that promotes the imperviousness of the ground and the local geomorphology, which is characterised by higher slopes that do not favour infiltration. When crossed with precipitation, the hydrogeomorphology and IPI-URBAN index in each area allows the estimation of the urban groundwater recharge. Analysing Fig. 2B, half of the study area has recharge ranging between 30 and 45 mm/year. The areas where a higher recharge represent a very small area, mainly characterised by green urban areas and small bottom valleys where alluvial deposits outcrops (details in Freitas et al. 2019a, b).

Methodological approach
The methodology performed in this study is characterised by multidisciplinarity and multi-technical applications (Fig. 3). The research is carried out in several phases: urban data collection, geographical information systems (GIS) mapping, GIS modelling, geovisualisation techniques and groundwater resources sustainability studies. The research began with data collection using field and desk-based techniques. Next, the information collected was processed and stored using interactive geographical databases and GIS (Kresik and Mikszewski 2013). Finally, GIS overlay methods were used to analyse and interpret the input layers related to topography, climatology, geology, hydrogeomorphology, hydrogeology, urban hydraulics, and sanitation.
Using GIS modelling and geovisualisation techniques, urban groundwater mapping was carried out. The intrinsic vulnerability to groundwater contamination was reached by applying four indexes: GODS (Foster et al. 2002), DRAS-TIC-Fm (Denny et al. 2007), SI (Ribeiro et al. 2017), and the adapted DRASTIC-Fm-URBAN (present study). The methodology used in the DRASTIC-Fm-URBAN index will be detailed in the following section. The overlap between the vulnerability mapping has enriched this analysis, and the inventory of potential contamination activities carried out as part of doctoral research took place in the study area between Page 5 of 24 156 2005 and 2022 (Freitas 2019). The potential contamination activities mapped were based on the guidelines proposed by Vrba and Zaporozec (1994) and Zaporozec (2004). As this is a GIS multi-parametric analysis, the layers were converted into a geometric matrix whose pixels have an area of 25 m 2 . The vulnerability maps used the international colour scheme for the DRASTIC index. This facilitated the comparison and analysis of the four indexes, enabling a robust assessment of the intrinsic vulnerability to groundwater contamination.
The assessment of recharge and discharge was carried out using the Urban Potential Infiltration Index [IPI-URBAN] (Freitas 2019;Freitas et al. 2019aFreitas et al. , b, 2021. It corresponds to a quantitative multi-parametric index based on a weighted sum of eight parameters: geology and hydrogeology, slope, land use, tectonic lineaments density, drainage density, and urban hydraulics and sanitation. In addition, the safeguard zones of historic springs in the study area were delineated using the multi-parametric index DISCO-URBAN (Meerkhan et al. 2021(Meerkhan et al. , 2022. This index corresponds to an adaptation for urban areas of the DISCO method developed for general purposes by Pochon et al. (2008). DISCO-URBAN considers four parameters: discontinuities, protective cover, land use and cover, urban hydraulics, and sanitary features (Meerkhan et al. 2021).
This integrative approach provides a comprehensive understanding of groundwater resource management. Furthermore, the methodology is intended to be replicable in other urban areas with a strong historical connection to groundwater resources. This work locates the areas vulnerable to contamination and those favourable to groundwater infiltration/recharge and presents the delimitation of protection areas for historical springs.
A GIS-based software was applied for map data visualisation, overlay analysis and layout design. All overlay analysis was performed using a raster file format with a pixel resolution of 5 × 5 m. Finally, the 3D urban hydrogeological site conceptual model was created using Arc-GIS Pro.

DRASTIC-Fm-URBAN index: a proposal for urban areas
DRASTIC-Fm-URBAN index is a method for identifying vulnerability to urban groundwater contamination. The DRASTIC-Fm-URBAN index adapts the DRASTIC index (Aller et al. 1987a, b) and the DRASTIC-Fm index (Denny et al. 2007) and consists of a critical and updated review for better application in urban areas. The DRASTIC-Fm-URBAN index considers all the parameters of DRASTIC and DRASTIC-Fm, but with the addition of the two parameters related to the urban component (Table 1 and Fig. 4): LUC (Land Use/Cover parameter) and HS (Urban Hydraulics and sanitary features including water supply, stormwater, and sewer networks). To determine the DRASTIC-Fm-URBAN index, every parameter is assigned a numerical score (1-10, i.e., lowest to highest contamination potential). In addition, each parameter is assigned a weighting factor (1-5) based on its comparative impact on contamination potential (Table 1). Finally, the analytic hierarchy process (AHP) was used to define the weight of every parameter. Using the AHP technique, all parameters are compared and ranked according to their importance (e.g., Saaty 2008Saaty , 2012Brunelli 2015;Mu and Pereyra-Rojas 2017). Then, a proper Consistency Ratio (CR) is calculated by dividing the Consistency Index for the set of judgments by the index for the corresponding random matrix (Goepel 2013). Saaty (2012) proposes that the judgment set is inconsistent if that ratio exceeds 0.1. In practice, when CR < 0.1, the consistency of the judgment matrix is acceptable; when CR ≥ 0.1, the judgment matrix needs to be tested again (e.g., Goepel 2013;Brunelli 2015;Mu and Pereyra-Rojas 2017). The equation to calculate DRASTIC-Fm-URBAN can be found in Table 1 and Fig. 4. Table 2 presents all the parameters used in DRASTIC-Fm-URBAN and describes each parameter. The approach used to achieve the vulnerability parameters are comparable to the DRASTIC index (Aller et al. 1987a, b), and the "Fm" parameter is adapted from Denny et al. (2007) and recently updated inputs from Erlandson (2014).
The study area is dominantly in a fissured media and is intended to identify its impact on groundwater vulnerability. The fractured media considers three fundamental geostructural characteristics that determine the influence of the rock discontinuities network (Denny et al. 2007): orientation, length, and fracture density (Table 3). Denny et al. (2007) suggest that the fractured media shall be classified to the average rate value of each category. Therefore, the Fm parameter was achieved from the tectonic lineaments density map and then classified into six rated classes, as shown in Table 3 (Denny et al. 2007;Shirazi et al. 2012;Teixeira et al. 2015). In addition, structural geology and hydrogeotechnical field rock mass data available at a local scale should also be integrated into the "Fm" parameter assessment, as pointed out by Freitas (2019), Meirinhos et al. (2016), and Meerkhan et al. (2016). The main criteria are geostructural studies developed into the investigation site, i.e., from megascale to local scale geostructures assessment and analysis processed by remote sensing techniques and field structural geology mapping. In addition, hydrogeotechnical rock mass surveys were performed using, for example, the suggested methods highlighted by CFCFF (1996) and Chaminé and Fernandes (2023).
The Land use/cover parameter has been integrated into this new index approach, as it helps define the degree of imperviousness of the surface under study. Table 4 indicates the classification of the LUC parameter according to the impervious surface grade and the corresponding rates. It is important to note that it is in the most urbanised areas (LUC 0-3), where the activities with the most significant potential for contamination of water resources are located. In the less urbanised classes (LUC 4-8), impervious activities   are lower, but the activities there have less potential for contamination. Therefore, they receive a lower rate, reflecting the lower propensity to contamination and the possibility of filtering the water that infiltrates. As stated before, the HS parameter considers water supply, stormwater, and sewer networks. HS is a density of hydraulic and sanitary features calculated in a GIS environment according to the description made in Table 2. Areas with very high density (> 200 km/km 2 ) have a rating of 10, meaning a higher impact on groundwater vulnerability (Table 5). On the other hand, areas with very low to null density (< 50 km/km 2 ) have a rating of 2. This parameter is related to the impact that could potentially occur in case of leakages in these infrastructures. Leakage from public water supplies can contribute to increased aquifer recharge but will impact the hydrogeochemical quality of groundwater. Stormwater network leaks often occur in urban areas, which may increase surface runoff and infiltration, carrying  (Aller et al. 1987a, b). Therefore, to obtain the water table, the elevations extracted from the Digital Elevation Model (DEM) and the depths of the water wells were compared. In addition, hydrodynamic data derived from hydrogeological units are cross-checked. The parameter "D" depths vary from 0 to 10 m R net recharge The primary groundwater source is precipitation which infiltrates through the surface of the ground and percolates to the water table. This recharge is thus available to transport a contaminant (solid or liquid) to the water table. Net recharge is the total amount of water reaching the soil surface and infiltrating. For the urban area of Porto, the collected data were from the meteorological station of "Serra do Pilar" (located in the Vila Nova de Gaia), and the average annual precipitation corresponds to 1236.8 mm. Previous studies in the study area (e.g., Afonso et al. 2007Afonso et al. , 2016Afonso et al. , 2019Afonso et al. , 2020Freitas et al. 2019a, b) suggest an initial recharge rate of 8%, corresponding to an average recharge of 98.94 mm/year A aquifer media Aquifer media refers to the consolidated or unconsolidated geological material which serves as an aquifer.
For this parameter, the study area's geological and hydrogeological background and features are used to assign a DRASTIC rating S soil Soil media refers to the weathered upper zone of the ground. The study area's geological background and/ or hydropedological data were assigned a DRASTIC rating for this parameter T topography Topography refers to the surface's slope and the slope variability (Aller et al. 1987a, b). The topography parameter is measured in a per cent slope. To obtain the topography, the slopes were extracted from the DEM. The hillslope was classified into five classes, namely, 0-2, 2-6, 6-12, 12-18, and > 18 I impact of vadose zone The DRASTIC methodology measures the impact of the vadose zone parameter by the velocity (m/s) that water moves through the zone. The ranking of this parameter was determined by extracting the lithologies of the material encountered above the water table from the water well database. The "I" parameter ranges between 3 and 8 C hydraulic conductivity Based on local and regional studies performed in the study area (e.g., Afonso et al. 2007Afonso et al. , 2016Afonso et al. , 2019Afonso et al. , 2020Freitas et al. 2014Freitas et al. , 2019a, hydraulic conductivity values were assigned to the geological background as a reference to assign DRASTIC. In the study area, the hydraulic conductivity ranges between 1 and 2 Fm fracture media A density map was created to achieve the fracture media (Fm) parameter. The density increases with proximity to geological discontinuities at several scales, such as tectonic lineaments, faults, fractures, geological contacts, and veins. In the study area, the density varies between 0 and 268 km/km 2 . The density map was classified into five classes. The ratings range from 2 (< 50 km/km 2 ) to 10 (> 200 km/km 2 ) LUC land use/cover The official land cover cartography was used for the LUC parameter and also the suggested guidelines by DGT (2019). However, to improve the high-resolution mapping and maintain the COS2018 baseline structure, this was cross-checked and updated using Maxar satellite imagery from January 2022 (0.5 m resolution, available through the ArcGIS base cartography service) and fieldwork surveys. This parameter is of significant importance as it defines the imperviousness of the ground surface. According to the impervious surface grade, the "LUC" parameter varies between 0 to 10 m HS hydraulics and sanitary features HS considers water supply, stormwater and sewer networks. The calculation of the HS parameter was performed with GIS technology, the line density tool was used, and different weights were considered for each network: 1 × water supply network, 2 × stormwater network, and 4 × sewer network. The rating ranges from 2 to 10. Areas with very high density (> 200 km/km 2 ) have a rating of 10, meaning the highest contamination potential. In contrast, areas with very low to no density (< 50 km/km 2 ) have a rating of 2 a potential contamination load to groundwater resources. Finally, a leak in the sewage network and the direct impact on recharge will add to the chemical and biological contamination transported to groundwater.
To compute the urban hydraulics and sanitary network, the line density was calculated as the density of linear features in the neighbourhood of each output raster cell. Density is calculated in units of length per unit of area (Silverman 1986). Table 6 outlines a comparison of the DRASTIC-Fm and DRASTIC-Fm-URBAN indexes. Figure 4 shows the methodological approach to the DRASTIC-Fm-URBAN Index.

Results and discussion
Delimiting the vulnerability of groundwater to contamination in urban areas is essential, for example, to delineate the safeguard of historic springs, fountains or underground aqueducts. Therefore, an integrative vulnerability indexes study, GIS modelling and geovisualisation techniques were  Denny et al. 2007) a Fracture orientation: determines whether a fault or fracture acts as a hydraulic conduit or barrier to groundwater contamination b Fracture length: determines whether it is a regional or discrete structure. Regional structures often include several fracture intersections, which can increase the hydraulic conductivity of a fault or fracture to a large extent c Fracture density: fracture density may increase with proximity to known faults   Regional to local-scale: Small-scale maps to illustrate regional groundwater vulnerability conditions for national or regional summary maps Large-scale maps to high-resolution groundwater vulnerability map in an urban context for planning and management purposes The maps use diverse legends and explanatory notes to promote a uniform mapping methodology (e.g., Aller et al. 1987a, b) Local-scale: Large-scale maps to high-resolution vulnerability maps for delineating safeguard zones (local framework: 1:1000 to 1:10,000; detailed surveys: ranging 1:100 to 1:250). The maps use diverse legends and explanatory notes to promote a uniform mapping methodology (e.g., Aller et al. 1987a, b) Vulnerability Take into account the land use/cover for general purposes, but do not take into account impervious areas Take into account the land use/ cover and impervious areas Do not take into account the hydraulic and sanitary features Take into account the hydraulic and sanitary features: water supply, stormwater, sewer Do not take into account the hydraulic and sanitary features Take into account the hydraulic and sanitary features: water supply, stormwater, sewer performed to assess groundwater vulnerability and protect key areas.

Vulnerability mapping assessment
The selected vulnerability indexes are related mainly to the characteristics of the study area, namely, the geological, morphotectonic and hydrogeological context and the fact that it is an urban area. Therefore, an integrative and complementary multi-parametric vulnerability approach was used to obtain a realistic urban vulnerability mapping in a studied area of 12.9 km 2 . Figure 5 summarises the results attained with the vulnerability indexes and the inventory of groundwater potential contamination activities: GODS, DRASTIC-Fm, and SI. The map explanation shows each class's areas [%], showing their representativeness.
The GODS index (Fig. 5A) reflects the local geology and hydrogeological background, with a clear distinction between sedimentary cover (porous media) and granitic and metasedimentary rocks (fissured media). Thus, moderate to high vulnerability occurs in porous media, where permeability is low to moderate (< 4 m/day), comprising alluvia and clayey arenite-conglomerate deposits. Conversely, low to moderate vulnerability, which includes 68% of the studied area, occurs in granitic rocks that are characterised by a low to very low permeability (< 1 m/day). In the metasedimentary rocks, the vulnerability is negligible. Figure 5B shows the mapping of the DRASTIC-Fm index. Parameters such as recharge, topography, and fractured media are added to its calculation, which is why the results obtained differ from the previous index. With the inclusion of these parameters, the geological boundaries are no longer so clearly reflected. For example, porous media are classified with high to very high vulnerability. In the granitic rocks, the vulnerability varies between moderate and high. Finally, the metasedimentary rocks presented a low to moderate vulnerability. Regarding the SI index (Fig. 5C), the impact of the land use parameter is visible. The degree of urbanisation and the activities related to population concentration increases the vulnerability to contamination in the study area, whether in the porous or fissured media.
In this study, the proposed DRASTIC-Fm-URBAN fits better in urban areas and could contribute to urban groundwater resource planning and sustainable management (Fig. 6). The application of DRASTIC-Fm-URBAN enables a clear distinction between the different hydrogeological media, with vulnerability in the porous media being very high to extremely high. In fissured media, the vulnerability varies from low to moderate and high to very high. In areas, where the density of tectonic lineaments is higher, an increase in vulnerability is observed. In addition, the different land use is characterised by distinct surface imperviousness levels, highlighted in the synthesis map (Fig. 6). For example, the main road network is observed under the porous media, making them more impermeable and lowering the vulnerability to contamination in these areas. On the other hand, urbanised areas, where the potential sources of contamination are located, increase the vulnerability of water resources to contamination. In addition, less impermeable areas allow water infiltration and filtration, reflecting a lower susceptibility to contamination. Finally, areas with a higher density of hydraulic and sanitary features increase the possibility of contamination due to leakages. Hydrochemical studies showed a correlation between the vulnerability indexes with contaminant parameters or land use (e.g., Rupert 2001;Javadi et al. 2011;Saida et al. 2017;Vu et al. 2019). For example, analytical NO 3 − data in the Porto region indicates values ranging from 61 to 122 mg/L (Afonso et al. 2020). In addition, there are recorded nitrate values (< 40 mg/L) in some springs in the surroundings of the studied area. Comparing the nitrates values with the regional vulnerability mapping (GOD-S, DRASTIC-Fm, SI), the higher values meet the moderate-high classes.
Comparing the results achieved by the DRASTIC-Fm and DRASTIC-Fm-URBAN in the study area, the following remarks could be highlighted: (i) DRASTIC-Fm-URBAN offers a better high-resolution in urban vulnerability mapping than DRASTIC-Fm because reflects the cross-check of land use/cover and urban hydraulics features. Furthermore, a comparative vulnerability classes analysis of both indexes (see Figs. 5 and 6) highlights that the DRASTIC-Fm-URBAN vulnerability protection output mapped is more representative in urban areas; (ii) comparing the hydrogeological media between the indexes suggests that: fissured media dominate the study region and the areas generally assessed as "moderate to high" by DRASTIC-Fm are according to DRASTIC-Fm-URBAN as "high to very high"; by contrast, the sedimentary cover mapped areas suggest that in DRASTIC-Fm-URBAN is mostly classified as "very high to extremely high". Conversely, the DRASTIC-Fm was "high to very high".
Finally, contrasting the various vulnerability indexes, particularly the DRASTIC-Fm-URBAN, helps sustainable water resources management in urban areas. Its application shows the areas with a greater predisposition for potential groundwater contamination. Thus, urban planners can decide better on activities and land use in those areas, ultimately contributing to a better design with nature, as McHarg (1992) highlighted within geoethical concerns (e.g., Freitas et al. 2021;Chaminé et al. 2022;Peppoloni and Di Capua 2022).

Delineation of groundwater safeguard zones
The analysis of the urban groundwater vulnerability to contamination, particularly through DRASTIC-Fm-URBAN, has shown that some historic springs are in areas of greater vulnerability to contamination. Therefore, two historic springs were selected to delineate safeguard zones from potential contamination. The springs are located on the Douro riverside, namely, the "Caco spring" (Porto riverside) and the "Mocho spring" (Vila Nova de Gaia riverside). The delimitation of the safeguard zones was based on applying the multi-parametric index DISCO-URBAN (Meerkhan et al. 2021(Meerkhan et al. , 2022. Figures 7 and 8 show the application of the DISCO-URBAN vulnerability index. The parameters used in the calculation of DISCO-URBAN are (details in Meerkhan et al. 2021): discontinuity parameter (D), protection cover (P), land use/cover (LUC), and urban hydraulic and sanitary characteristics (HS).
The parameter discontinuities (D) of the "Caco spring" show that 84.4% of the area has low permeability, and the discontinuities with a relatively slow connection (in terms of the residence time of water) to the spring, i.e., D3 (Fig. 7). In addition, in 12.7% of the area, the highly permeable discontinuities are rather connected to the spring (i.e., D0 and D1) and 2.9% in category D2 (discontinuities with a relatively slow connection to the spring). The protective cover parameter (P) shows three classes according to its permeability, namely: class P0 (9%)consisting of alluvium with moderate to high permeability; P0 (3%) arenite and conglomerate deposits with moderate permeability and P1 (88%) granite with low to moderate permeability. The land use parameter (LUC) classifies the catchment into four classes according to its impact on safeguard zones, ranging from high to very high (continuous urban fabric-LUC0) and very low to low (forest and bushes and shrubs-LUC6 and LUC7). Finally, four classes of urban hydraulic and sanitary (HS) features were considered, of which low to moderate densities prevail. The study site is classified into three safeguard zones (S 1 -very high to S 3 -moderate). Therefore, 76% of the study area required moderate protection (S 3 ). The area with the highest vulnerability (S 1 ) corresponds to 5% of the study area and is where the "Caco spring" is located. The remaining area has a high vulnerability (19%, S 2 ) (Fig. 7B). Figure 8 exemplifies the DISCO-URBAN index in the "Mocho spring" (Fig. 8A). The parameter "D" shows that 78% of the "Quinta do Mocho" drainage basin comprises discontinuities with a slow connection to the spring (D3). The highly permeable discontinuities with preferential connection to the spring (D0) are represented by 13% of the studied area. The remaining area (9%) has discontinuities with a relatively rapid connection to the spring (D1). The parameter "P" is relatively well-distributed for four classes, namely: P0-arenite-conglomerate deposits (22%-moderate permeability); P1-granite medium to fine-grained (27%-low-to-moderate permeability); P2-granite fine-grained (16%-low permeability); P3-metasedimentary rocks (35%-very low-to-low permeability). The LUC parameter categorises the area: high to very high (continuous urban fabric-LUC0) and very low Fig. 8 "Mocho spring" site is included in a studied basin area of 0.4 km 2 : delineation of groundwater safeguard zones A location of the "Mocho" spring; B overlap between safeguard zones and groundwater potential contamination activities, C aspect of the "Mocho spring" to low (forest and bushes and shrubs-LUC6 and LUC7). Finally, the parameter HS classified the dominant area with very low to moderate densities. Four safeguard zones were mapped for the "Quinta do Mocho" watershed ( Fig. 8B), and 59% of the area was classified in class S 3 (moderate). The most vulnerable area, class S 1 , comprises 13% of the watershed and corresponds to the area immediately to the "Mocho spring". The S 2 class (high vulnerability) represents 26% of the area, and the area with lower vulnerability (S 4 ) represents 1% of the watershed.
The crossing between DRASTIC-FM-URBAN and DISCO-URBAN demonstrates the importance of adapting the parameters used in calculating vulnerability when discussing urban areas. Using parameters that bring the indexes closer to urban reality improves the results achieved and supports the sustainable management of water resources.

Conceptual hydrogeological site model: inputs from DRASTIC-Fm-URBAN
The urban conceptual hydrogeological site model for the Porto and Vila Nova de Gaia riverside area was enhanced from previous works (e.g., Afonso et al. 2016Afonso et al. , 2020Freitas et al. 2019b;Meerkhan et al. 2021), cross-checking geological, hydrogeological, groundwater vulnerability, land use/ cover, and urban features. In this region, the most representative aquifer systems are ( Fig. 9): (i) A superficial unconfined unit, corresponding to the alluvia deposits, with a low thickness (< 6 m), a shallow water table (< 4 m), a moderate transmissivity (10-15 m 2 /day), a moderate permeability (< 4 m/ day), and a Cl-Na to NO 3 -Na hydrochemical facies. Thus, the vulnerability to contamination (DRASTIC-Fm-URBAN) is very high to extremely high; (ii) An unconfined to a confined unit, corresponding to the highly weathered granite, medium to finegrained, with a thickness that may reach 40 m, a shallow water table (< 6 m), a low transmissivity (< 1.5 m 2 /day), a low permeability (< 1 m/ day), and an SO 4 -Ca to HCO 3 -Ca hydrochemical facies. Therefore, the vulnerability to contamination (DRASTIC-Fm-URBAN) is high to very high; (iii) A deep confined unit, corresponding to the fresh to low weathered granite, medium to fine-grained, with an undefined thickness, shallow to deep water table (< 40 m), a low transmissivity (< 1.5 m 2 /day), a very low to low permeability (< 0.5 m/day), and an SO 4 -Ca to HCO 3 -Ca hydrochemical facies. The vulnerability to contamination is low to very low.

Conclusions
Due to the diversity of geological and hydrogeological conditions in fractured media, it is not the best approach to delineate vulnerability indexes using a single methodology. Instead, a multi-scale approach and several vulnerability indexes (e.g., GODS, DRASTIC-Fm, SI, DISCO-URBAN) shall be applied. On the other hand, the proposal of the DRASTIC-Fm-URBAN method allows a consistent evaluation of groundwater vulnerability at a regional to local scale focused on urban environments. Thus, it is a reliable methodology for delineating urban vulnerability in fissured media. This multi-technical approach highlights the importance of groundwater vulnerability to contamination mapping to support hydrogeological conceptualisation and improve the decision-making process regarding water resources management and sustainability. Furthermore, the analytic hierarchy process is reliable for multi-criteria decision-making in applying and analysing the vulnerability indexes.
The delineation of groundwater vulnerability to contamination and mapping aquifer vulnerability is an essential component of groundwater source protection and global vulnerability assessment. In addition, these approaches provide visual analysis and a practical tool for planning and management of better sustainable groundwater systems in quality and quantity. Moreover, the GIS-based mapping techniques accurately evaluate groundwater vulnerability and support delineating the hydrogeological conceptual site model.
This study highlights and discusses the importance of the DRASTIC-Fm-URBAN methodology in contributing to sustainable groundwater resource management in urban areas. A comprehensive analysis was used, for example, field and laboratory techniques, where several data were overlapped, such as land use, morphotectonics, surface hydrology and urban groundwater. That approach was fundamental to investigating groundwater vulnerability to contamination evaluation built on cross-checking multi-parametric methods (e.g., GODS, DRASTIC-Fm, SI and DISCO-URBAN indexes). Moreover, the research also meets on a critical review, update and development of the DRASTIC-Fm methodology in the called DRASTIC-Fm-URBAN index, mainly aimed at supporting the regional to local-scale site management practice and safeguarding of historic urban spring sites. The role of this multidisciplinary methodology is to offer a comprehensive knowledge of groundwater vulnerability systemsprotection. In addition, aims to improve the hydrogeological conceptual site models focused on vulnerability issues, which is of greater importance in sustainable water resources management.