A Framework to Evaluate Groundwater Quality and the Relationship between Rock Weathering and Groundwater Hydrogeochemistry in the Tropical Zone: A Case Study of Coastal Aquifer Arroyo Grande, in the Caribbean Region of Colombia

: Hydrochemical analysis is crucial for understanding soil and water composition dynamics in coastal aquifers. This study presents a novel framework for the comprehensive assessment of groundwater quality, integrating multivariate analysis and hydrochemical techniques. It comprises seven stages aimed at characterizing physicochemical properties, identifying water constituents, elucidating dominant mechanisms in water composition, evaluating ion exchange processes, analyzing spatial distribution of components, identifying impacting processes, and assessing drinking water quality. The framework was applied to the coastal unconfined Arroyo Grande aquifer in Cartagena, Colombia. Fifteen points were sampled, assessing physicochemical parameters such as total hardness, alkalinity, pH, temperature, electrical conductivity, anions, cations, among others. Findings reveal the presence of dominant anions including bicarbonate, chloride, and sulfate, with relevant variations observed between the dry and wet season, with manganese and iron surpassing WHO drinking water standards. The prevalence of these constituents has been attributed to mineral dissolution, ion exchange, salinization due to seawater intrusion, and anthropogenic contamination. Over 50% of samples in both seasons fail to meet freshwater drinking standards due to elevated dissolved mineral concentrations in groundwater. These findings provide insights for sustainable management and mitigation strategies, and the systematic approach enables researchers to identify key factors influencing water composition.


Introduction
Groundwater is a vital resource necessary to meet the increasing demand for freshwater worldwide.Nevertheless, knowledge of the subterranean systems and their proper management is usually limited.Regions with groundwater governance key deficiencies, such as, low political commitment, limited integration of science-based groundwater understanding, and limited awareness of long-term groundwater risks, usually have significant gaps in information about the resource [1,2].Furthermore, the pressures of human activities and climate change continue to impact groundwater quality, posing a threat to its sustainability [3,4].
In recent decades, research has increasingly focused on improving the study of unconfined coastal aquifers, particularly in regions where the effects of climate change are more pronounced, influencing water availability and quality [5].These aquifers exhibit unique

Review of Hydrochemical Studies of Coastal Aquifers
A review was conducted to assess recent advances in hydrochemical analyses of coastal aquifers, highlighting several common aspects.These include the aquifer's location, geological characteristics, seasonal climate conditions during the study period, classification of water types based on predominant anion and cations, temporary ion analyses identifying dominant ions, prevalent hydrochemical processes, and applied hydrochemical analyses.The selected studies encompassed coastal aquifers situated in various regions: the Arabian Sea southeast of India in Udupi district [20]; the Mediterranean Sea in the Gaza Strip, Palestine [21]; the South China Sea, Malaysia [22]; the Gulf of Guinea in Lagos, Nigeria [23]; the Mediterranean Sea in various locations including Tunisia (Jerba Island), Libya (Tripoli), Algeria (Nador plain), Morocco (Bou Areg), Italy (Sardinia), and Egypt (Baghoush area) [24]; the Caribbean Sea coast in Quintana Roo, Mexico [25]; the Recife Sea coast, Brazil [17]; and the Caribbean Sea coast in the Gulf of Morrosquillo, Colombia [6].Relevant findings are summarized in Table S1 (Supplementary Material).
A comparison was made among parameters considered crucial for identifying geogenic and anthropogenic contamination processes, such as chlorides (Cl − ), total dissolved solids (TDS), and nitrates (NO 3 − ).The majority of the reviewed aquifers in Appendix A Table A1 consisted of sedimentary rocks and deposits.Ions such as Ca 2+ , Na + , HCO 3 − , and SO 4  2− show high significance to the groundwater composition due to rock-water interactions, mineral dissolution and ion exchange, and seawater intrusion [21,25].
According to the data shown in Appendix A Table A1, the prevalence of chloride anion (Cl − ) is linked to systems experiencing marine intrusion, which is increasingly relevant in coastal areas.By 2019, at least 501 coastal cities worldwide reported marine intrusion crises, with population density being a primary contributing factor [26].For example, in the coastal aquifer of southwestern India [20], the highest chloride anion concentration observed was 17,402 mg/L during the pre-monsoon period characterized by low rainfall.Similarly, coastal aquifers along the Mediterranean Sea, in both in Europe and Africa, exhibit elevated average chloride concentrations (Tunisia: 1951 mg/L; Libya: 577 mg/L; Algeria: 770 mg/L; Morocco: 1828 mg/L; Italy: 595 mg/L; Egypt: 1987 mg/L) [24], highlighting marine intrusion as the primary agent of salinization in these systems.
Elevated nitrate anion levels can be attributed to anthropogenic activities such as wastewater and industrial discharge, as well as fertilizer and pesticide use [17,27].Concentrations exceeding 10 mg/L are considered indicative of anthropogenic contamination in groundwater [20].The World Health Organization (WHO) has established a threshold of 50 mg/L for nitrate (NO 3 − ) content as a guideline of groundwater quality [28].Analyses conducted during different climatic seasons showed higher nitrate (NO 3 − ) concentrations during the rainy season.For instance, in the aquifer of the Recife Metropolitan Region (Brazil), maximum recorded values reached up to 207.6 mg/L [17].In the Lagos Aquifer (Nigeria), maximum values of 108.7 mg/L were reported during the rainy season [23], while the aquifer in the southern zone of Quintana Roo (Mexico) [25] recorded a maximum value of 7.93 mg/L during the same period.Observations from the Gaza aquifer (Palestine) and Mediterranean coastal aquifers [24] recorded maximum values above 100 mg/L, with the aquifer in the Baghoush area (Egypt) reaching the highest value at 7142 mg/L.
Despite the varying factors affecting these systems, common aspects can be identified in the processes determining groundwater composition and the pollution issues they face.Table 1 summarized the majority ions, the dominant hydrochemical processes, the type of aquifers, and the applied hydrochemical analysis for the studies revised in Table A1.From the review and Table 1, it is evidenced that the geology and hydraulic connectivity with the sea in coastal aquifer systems have a significant impact on groundwater composition.

Study Area
The Arroyo Grande hydrogeological unit and its corresponding basin are situated in the Colombian Caribbean region, north of Cartagena.The aquifer lies along a coastal strip and encompasses towns within the tourist district of Cartagena, as well as the municipalities of Santa Catalina and Clemencia.It is part of the coastal hydrogeological province PC1-Sinú San Jacinto.The study area covers approximately 143 km 2 (see Figure 1), bounded by coordinates X(m): 454,559.081 to 468,442.630 and coordinates Y(m): 1,181,790.930and 1,165,642.759 in zone UTM 18N.This hydrogeological formation is of critical importance to the coastal region, serving as the primary water source for 36,255 inhabitants from the nearby communities.According to records from meteorological stations in the area, the multiannual average rainfall corresponds to 1235.37 mm/year, with an estimated average maximum temperature of 30.26 °C [29].The precipitation is distributed throughout the year in two climatic periods: a dry period with average monthly rainfall ranging from 0 to 60 mm, considered to be the dry season, and a rainy period with average monthly rainfall ranging from 120 to 250 mm, considered the wet season [8,30].The dry and wet season usually last According to records from meteorological stations in the area, the multiannual average rainfall corresponds to 1235.37 mm/year, with an estimated average maximum temperature of 30.26 • C [29].The precipitation is distributed throughout the year in two climatic periods: a dry period with average monthly rainfall ranging from 0 to 60 mm, considered to be the dry season, and a rainy period with average monthly rainfall ranging from 120 to 250 mm, considered the wet season [8,30].The dry and wet season usually last six months each; however, during periods of El Niño Southern Oscillation (ENSO), the seasons can last longer than normal and generate drastic consequences to the water quality of surface and subterranean water resources [3,31].
The hydrography of the area includes various water sources such as the Caribbean Sea, lakes, springs, and streams.The Arroyo Grande basin, of the same name as the aquifer, has an extension of 96 km 2 , and it is the most important basin in the area [32].

Geology and Hydrogeology
The area is composed of sedimentary rocks that are part of the Sinú belt or Turbaco tectonic block.These rocks include the Arjona Formation from the Oligocene-Miocene Epoch of the Cenozoic, the Bayunca Formation from the Miocene-Pliocene Epoch of the Cenozoic, and the Arroyo Grande and La Popa Formations from the Pleistocene Epoch of the Cenozoic [33].The Arroyo Grande Formation (Q1-Sm) (refer to Figure 2) is considered a young unit due to the limited consolidation of its deposits, primarily consisting of gravel and sand deposits found in Arroyo Grande township, located northwest of the Cartagena district [34].The stratigraphic columns of greatest reference correspond to the Balastrera section of Arroyo Grande, featuring a base composed of pebble gravel and imbricated grains, with conglomeratic sandstones and sandy conglomerates.The upper part consists of slightly coarse-grained conglomeratic sandstones in thick layers, with the presence of lenticular layers of mudstones in the lower part.The Turbaná-Turbaco section comprises reddish The stratigraphic columns of greatest reference correspond to the Balastrera section of Arroyo Grande, featuring a base composed of pebble gravel and imbricated grains, with conglomeratic sandstones and sandy conglomerates.The upper part consists of slightly coarse-grained conglomeratic sandstones in thick layers, with the presence of lenticular layers of mudstones in the lower part.The Turbaná-Turbaco section comprises reddish and yellowish claystone in the lower part, interspersed with fine-grained quartz sandstones.The Pasacaballos section is mainly characterized by fine-grained, reddish quartz sandstones alternating with intercalations of gray claystone [34].
The Arroyo Grande aquifer (aT) hydrogeological unit belongs to the PCA1 Sinú San Jacinto hydrogeological province [32].It is an unconfined regional aquifer with continuous extension.The hydraulic conductivity values range between 1.0 m/day and 3.5 m/day, and the coefficient of storage ranges from 5.0 × 10 −1 to 3.0 × 10 −7 .Extraction flow rates can vary between 2 and 35 L/s [8].Vertical Electrical Sounding (VES) previously conducted in the aquifer exploitation zones allowed the depiction of the lithological columns presented in Figure 3.The lithology describes layers of sediment that vary from sand and gravel from the topsoil to approximately 20 m in both VES.VES 13 shows a large layer of sand from 2 m to 12 m followed by sand and gravel.Layers saturated with fresh water lie at depths ranging from 20 to 100 m (see Figure 3) [36].These layers consist predominantly of coarse sands exhibiting resistivities exceeding 100 Ω.m, while the surface layers are characterized by granular sand type material and saturated silt as detailed in Appendix A Tables A2 and A3.

Proposed Integrated Framework
Table 2 presents the proposed integrated framework for analyzing groundwater quality and the relationship between rock dewatering and groundwater composition.This framework outlines the necessary steps for a comprehensive analysis of the physicochemical composition of groundwater and the associated processes, with a specific focus on rock-water interactions.Based on the findings from Appendix A Table  3. Lithological columns derived from Vertical Electrical Soundings (VESs) in the area of study [37].The resistivity of each stratum is presented, alongside the VES location coordinates and the location of the water table.
These layers consist predominantly of coarse sands exhibiting resistivities exceeding 100 Ω.m, while the surface layers are characterized by granular sand type material and saturated silt as detailed in Appendix A Tables A2 and A3.

Proposed Integrated Framework
Table 2 presents the proposed integrated framework for analyzing groundwater quality and the relationship between rock dewatering and groundwater composition.This framework outlines the necessary steps for a comprehensive analysis of the physicochemical composition of groundwater and the associated processes, with a specific focus on rock-water interactions.Based on the findings from Appendix A Table A1 and the conducted review, it is evident that the main characteristics of groundwater are of geogenic origin.The framework delineates essential information, applicable techniques (including statistics, graphics, and geospatial analysis), relevant considerations at each stage, and their interrelationships.Associate with spatial analysis results (step 5) and processes identified in step 6.
To apply the framework effectively, it is necessary to preliminarily identify the target aquifer for study, delineate the study area, and develop an inventory of groundwater infrastructure along with surface water sources connected to the aquifer.
The initial step involves conducting physicochemical characterization of water samples following a defined sampling program within the monitoring network.Laboratory analysis results are subjected to statistical and graphical methods to identify key components, water types, and indications of water composition origin.Graphical techniques are employed to determine relevant mechanisms influencing water composition, including rock-water interaction, rainfall, and evaporation.Given the critical role of ion exchange processes in groundwater composition, their evaluation is integral within this framework.
During the next stage, a multivariate statistical analysis is applied to identify processes with the most significant impact on water composition and their interaction with chemical parameters.
The final step entails assessing water quality and establishing the state of drinking water quality based on its composition and related processes.This framework was applied to analyze the hydrogeological characteristics of the Arroyo Grande aquifer in Colombia.

Physicochemical Characterization
A comprehensive survey of the study area was conducted, accompanied by an inventory of water points.A total of 10 groundwater points and 5 surface water points were strategically selected for further analysis based on criteria such as spatial representativeness, accessibility, availability of infrastructure, and cooperation of operators and owners [38].Georeferencing was performed for all fifteen points, and a detailed map was generated using ArcGIS 11.0 software as presented in Figure 4.
Sampling was performed following the recommendations of the standard methods for examination of water and wastewater [39] and the Colombian Technical Standard [40].Samples were collected through the well pumping system or directly in the different surface water bodies and then stored and preserved, following the recommendations of the IDEAM and the standard methods for examination of water and wastewater [39].
Temperature, pH, conductivity, salinity, dissolved oxygen, oxidation-reduction potential (ORP), and total dissolved solids (TDS) were measured in situ.Samples to determine the rest of the variables were packaged and preserved for transport to the laboratory, following the methodology of the APHA, AWWA, WEF in the standard methods for the examination of water and wastewater [39].
Two sampling campaigns were conducted: the first during the dry season (April 2022) and the second during the rainy season (July 2022), adhering to established regulations and technical guidelines for water monitoring.On-site measurements were conducted using the HI9829 multiparameter equipment from HANNA Instruments, which facilitated the measurement of various parameters including pH, temperature, electrical conductivity, salinity, dissolved oxygen, ORP (oxidation-reduction potential), and TDS (total dissolved solids).Additionally, water samples ranging from 0.25 to 2 L were collected in both plastic and glass containers.These samples were carefully preserved, stored, and transferred to certified laboratories for the analysis.
The anions were measured by ionic chromatography, following the regulations of U.S. EPA 300:0 [41], and cation concentrations were measured following the ICP/MS method (standard in water chemistry) [42].HCO 3 was calculated using the volumetric method following the standard method SM 2320 B.
spectrophotometry were employed for these analyses [39].The analytical errors a included in Appendix A Table A4.
The anions were measured by ionic chromatography, following the regulations U.S. EPA 300:0 [41], and cation concentrations were measured following the ICP/M method (standard in water chemistry) [42].HCO3 was calculated using the volumet method following the standard method SM 2320 B.

Identification of Water Constituents and Types of Water Descriptive Statistics
Upon receiving and tabulating laboratory results, statistical calculations for position, centralization, dispersion, and box-and-whisker plots were performed using JASP 0.15 software [44].Initial analysis was focused on key water components and their average concentrations across different sampling points [17,45].

Hydrochemical Diagrams and Indices
To identify water types and underlying chemical processes, hydrochemical diagrams including Piper, Gibbs, Schoeller, and scatter plots of ionic ratios along with the CAI index were generated using AQUACHEM 11.0 software [37].These diagrams utilize major ion concentrations to detect element sources and evolution patterns.

Schoeller Plot
The Schoeller plot was used to determine the order of abundance of anions and cations in groundwater, plotting average ion concentrations from the two sampling campaigns on a semi-logarithmic scale [46].

Piper Diagram
Water types and associated chemical processes were identified using the Piper diagram, which represents major anions (HCO 3 − , SO 4 2− , Cl − ) and cations (Ca 2+ , Mg 2+ , Na + , K + ) in two equilateral triangles.The intersection of plotted points within a central diamond establishes different water types based on their chemical composition [19,21].This diagram is widely used in hydrochemical analysis and represents a relatively easy and applicable tool for the study of groundwater; however, it has the disadvantage that it is difficult to show concentration values.

Identification of Dominant Mechanisms in Water Composition Gibbs Plot
The Gibbs diagram correlates total dissolved solids (TDS) with the ratio of chlorides to bicarbonates (RI) to identify dominant mechanisms influencing water composition, such as rock-water interaction, rainfall, and evaporation, as described in Equation (1) [22,47].According to this diagram, rainwater typically exhibits high concentrations of NaCl due to its marine origin and low TDS concentrations.In contrast, mineral dissolution processes resulting from rock-water contact increase the TDS content, while evaporation and runoff can lead an increase in both TDS and NaCl concentrations [17].

Evaluation of Ion Exchange Processes Scatter Diagrams of Ionic Ratios
Ion exchange processes (IEPs) were evaluated using dispersion diagrams of ionic relations.One such diagram plots the sum of cations (Ca 2+ and Mg 2+ ) against the sum of anions (HCO 3 − and SO 4 2− ), with cations on the X-axis and anions on the Y-axis.Points near a 1:1 slope indicate mineral dissolution processes, while deviations above and below the line suggest direct and reverse ion exchange processes, respectively [46,47].

CAI Index
The chloro-alkaline indices (CAI) were used to identify softening or hardening processes in groundwater based on specific variables defined in Equations ( 2) and (3) [2,23].

Identification of Processes Impacting Water Composition and Interrelationship between Chemical Parameters Multivariate Statistical Analysis
The Statgraphics 19 software was employed to perform the principal component analysis (PCA).Prior to conducting the analysis, the variables were standardized to achieve a normal distribution, using Equation (4) [37,48].
where Z ij , X ij , and σ j are the standardized value of sample j for parameter i, original value of sample i in parameter j, and standard deviation of parameter j, respectively.

Drinking Water Quality Assessment
The water quality index (WQI) was applied to assess the quality of groundwater based on the concentration of certain physicochemical parameters [49][50][51].Such parameters were identified and assigned weights based on their potential impact on human health, particularly in the context of drinking water [52,53].Table 3 presents the assigned weights (w i ) to physicochemical parameters ranging between 1 and 5.The relative weight (W i ) was calculated using Equation (5), where n is the total number of parameters.Subsequently, the quality rating (q i ) is calculated based on the concentration of each parameter (C i ) and its standard (S i ) as established by the drinking water standards of the World Health Organization (WHO) [54] (Equation ( 6)).Finally, the WQI is determined by summing the sub-index (SI) calculated for each parameter using Equations ( 7) and (8).Based on the results of the WQI, the quality status is assessed for each sample according to the ranges specified in Table 4 [52].

Results and Discussion
The data obtained from the field measurements provided a preliminary identification and general overview of the environmental conditions and physical characteristics of the water in the study area.The descriptive statistics presented in Table 5 show that the average water temperature at the sampled points during both campaigns correlates with the average maximum ambient temperature of 30.26 • C, with an average temperature of 29.6 • C for the first campaign and 29.36 • C for the second campaign.The highest temperature values recorded correspond to surface waters, while groundwater shows similar temperature during dry and wet seasons.These conditions can be attributed to the shallow nature of the aquifer.In studies of aquifers deeper than 700 m, the temperature tends to increase with depth, with gradients of up to 0.06 • C/m [23].Furthermore, it can be stated that the aquifer under study is not a geothermal aquifer and is not affected by magmatic activity that could influence water temperature [55].Recent water measurements in the study area over the last five years show comparable results to those obtained in the present study, suggesting that this parameter has remained relatively stable during this period.
The monitored waters show a slightly acidic condition based on the recorded pH values, with the dry season showing more pronounced acidity, with a mean value of 6.8, a minimum value of 6.2, and a maximum of 7.74.This acidic condition favors the occurrence of mineral dissolution processes, which ultimately affect the composition of the water [56].In the case of the results of the second sampling campaign corresponding to the rainy season, the average pH value increased to 6.91, with a minimum value of 6.3 and a maximum of 8.57.
For both climatic periods, it was found that there is a high variability of mineralization levels in the water bodies of the studied system, based on the behavior of parameters such as electrical conductivity (EC) and total dissolved solids (TDS) [25].In the case of EC in the first sampling campaign, the calculated standard deviation was 333.98 µS/cm with a minimum of 510 µS/cm and a maximum of 1842 µS/cm, while for the second sampling campaign, the standard deviation obtained was 531.44 µS/cm with a minimum of 438 µS/cm and a maximum of 2518 µS/cm.Aquifers under similar environmental conditions, such as the multilayered aquifer systems in the Recife metropolitan region of Brazil, attributed the present variation of EC in the different sampling points to the chemical heterogeneity of the minerals present in the geology, the water residence times, and the different processes present [17].
Considering the EC values, water samples can be classified as freshwater (<1500 µS/cm), brackish (1500-3000 µS/cm), and saline (>3000 µS/cm) [20].In coastal areas, brackish and saline groundwater are usually found due to salinization processes caused by saline intrusion, as has been observed on the Mediterranean coast [24].In the case of the results obtained in the Arroyo Grande aquifer, fresh and brackish waters are identified.
The TDS showed a behavior proportional to that of the EC.The surface water points and the groundwater points close to the coastal strip were the ones that presented the highest EC and TDS value.The first were subjected to a saline environment and direct reception of rainfall and the second with hydraulic connection to the sea, which favors the possible occurrence of saline intrusion processes.Studies carried in the coastal aquifer of Morrosquillo, also located in the Caribbean coast of Colombia [6], show that the mul-tilayered nature of the formations in the area causes the groundwater to have different degrees of mineralization and therefore compositional conditions that vary according to space and depth.

Groundwater Hydrochemistry
According to the average concentrations of the analyzed parameters in the sampled groundwater, bicarbonate (HCO 3 − ), chloride (Cl − ), and sulfate (SO 4 2− ) are identified as the most abundant anions.Among the cations, sodium (Na + ), calcium (Ca 2+ ), and magnesium (Mg 2+ ) were identified (Tables 6 and 7).These elements and their order of abundance can be observed in the Schoeller diagram in Figure 5.This result is consistent with the geology of the study area, where sandstonetype sedimentary rocks predominate.On the other hand, the dispersion statistics show the spatial heterogeneity of the system, in terms of the concentrations of the elements that make up the groundwater, a condition generated by the different hydrogeochemical processes that govern the hydrogeological cycle.
These dominant element results are similar to those obtained in the Morrosquillo aquifer, which, like the system under study, is located in the Caribbean coast of Colombian on a geology composed of sedimentary rocks [6].Similarly, there is agreement with the results of other studies carried out in the Gaza coastal aquifer, Palestine [21], the Kelantan aquifer, Malaysia [22], and the aquifer of the southern zone of Quintana Roo, Mexico [25].These systems are located in other areas of the world but possess similar geological characteristics.
Based on the criteria for determining the type of water, considering the dominant anion and cation, six types of water were identified in each of the points sampled in the two campaigns conducted.These correspond to calcium bicarbonate, sodium chloride, sodium sulfate, sodium bicarbonate, magnesium chloride, and magnesium sulfate.
These elements and their order of abundance can be observed in the Schoeller diagram in Figure 5.This result is consistent with the geology of the study area, where sandstone-type sedimentary rocks predominate.On the other hand, the dispersion statistics show the spatial heterogeneity of the system, in terms of the concentrations of the elements that make up the groundwater, a condition generated by the different hydrogeochemical processes that govern the hydrogeological cycle.These dominant element results are similar to those obtained in the Morrosquillo aquifer, which, like the system under study, is located in the Caribbean coast of Colombian on a geology composed of sedimentary rocks [6].Similarly, there is agreement with the results of other studies carried out in the Gaza coastal aquifer, Palestine [21], the Kelantan aquifer, Malaysia [22], and the aquifer of the southern zone of Quintana Roo, Mexico [25].These systems are located in other areas of the world but possess similar geological characteristics.
Based on the criteria for determining the type of water, considering the dominant anion and cation, six types of water were identified in each of the points sampled in the two campaigns conducted.These correspond to calcium bicarbonate, sodium chloride, sodium sulfate, sodium bicarbonate, magnesium chloride, and magnesium sulfate.
In six of the ten groundwater points, there was a difference in the percentage of the dominant ions between the dry and the wet season.In the remaining four points, there was no difference.For the rainy season, there was an increase in the concentrations of sodium and chlorides with respect to the values obtained in the dry season, which caused them to lead the order of abundance in several sampling points.Flash floods and runoff during intense rainfall can increase the mineralization of surface and shallow groundwater in the rainy season [23], thereby changing the composition of the water.In six of the ten groundwater points, there was a difference in the percentage of the dominant ions between the dry and the wet season.In the remaining four points, there was no difference.For the rainy season, there was an increase in the concentrations of sodium and chlorides with respect to the values obtained in the dry season, which caused them to lead the order of abundance in several sampling points.Flash floods and runoff during intense rainfall can increase the mineralization of surface and shallow groundwater in the rainy season [23], thereby changing the composition of the water.
The Piper diagram (Figure 6) prepared from the results of the sampling campaigns shows us that the most representative hydrochemical facies in the groundwater correspond to calcium bicarbonate, sodium bicarbonate, and sodium chloride.This shows a mixture of different types of water due to mineral dissolution processes and ionic exchange between calcium, sodium, and magnesium [47].Figure 5 shows that the monitoring points are concentrated in an intermediate zone of the central diamond of the diagram, which corresponds to a mixing zone.Thus, there is no evidence of a marked dominance of a single water type.The Piper diagram (Figure 6) prepared from the results of the sampling campaigns shows us that the most representative hydrochemical facies in the groundwater correspond to calcium bicarbonate, sodium bicarbonate, and sodium chloride.This shows a mixture of different types of water due to mineral dissolution processes and ionic exchange between calcium, sodium, and magnesium [47].Figure 5 shows that the monitoring points are concentrated in an intermediate zone of the central diamond of the diagram, which corresponds to a mixing zone.Thus, there is no evidence of a marked dominance of a single water type.As shown in Figure 7, Gibbs diagram confirms rock-water interaction processes to be most dominant in groundwater chemical composition, considering both dry and wet season results.A case in point is the presence of the bicarbonate anion.Its origin is related to the dissolution of carbonate in the geology of the study area, a process generated by the infiltration of rainwater and the slightly acidic conditions of the environment [57].The presence of calcium is related to minerals such as calcite, dolomite, gypsum anhydrite, As shown in Figure 7, Gibbs diagram confirms rock-water interaction processes to be most dominant in groundwater chemical composition, considering both dry and wet season results.A case in point is the presence of the bicarbonate anion.Its origin is related to the dissolution of carbonate in the geology of the study area, a process generated by the infiltration of rainwater and the slightly acidic conditions of the environment [57].The presence of calcium is related to minerals such as calcite, dolomite, gypsum anhydrite, and fluorite, which are sources of this cation and are present in the sedimentary rocks of the studied formation.In addition, calcium is considered to be the most dominant alkaline earth metal [58].The ionic relationship graph Ca 2+ + Mg 2+ versus HCO3 − + SO4 2− , prepared with data from the dry season (Figure 8a), shows that in addition to the mineral dissolution process already mentioned, reverse ion exchange processes (IEPs) are present, in which Ca 2+ + Mg 2+ are released into groundwater and Na + and K + are fixed to the aquifer material, causing water hardening [49].The graphical data of the wet season (Figure 8b) show that the direct ion exchange process (IEP) together with mineralization are the most relevant processes during the wet season.This was also evidenced by the increase in sodium concentrations in the Arroyo Grande in the same season.The occurrence of IEP was also verified with the application of chloro-alkali indices (CAI).Iron (Fe 2+ ) and manganese (Mn + ) have lower concentrations than the ions, but the values recorded are above those normally found in surface and underground waters, whose presence is evident in the samples collected, as they give a reddish color to the water.The presence of these minerals in groundwater is mainly due to the dissolution of rocks and minerals such as siderite, which are found in the ferrous state (Fe 2+ ); given the anoxic conditions of the aquifer environment, once in contact with the atmosphere it oxidizes, changing to the ferric state (Fe 3+ ), obtaining the appearance described above [53,59,60].The identification of the exact minerals contributing to the presence of iron in the Arroyo Grande aquifer recures further mineralogic studies.
The ionic relationship graph Ca 2+ + Mg 2+ versus HCO 3 − + SO 4 2− , prepared with data from the dry season (Figure 8a), shows that in addition to the mineral dissolution process already mentioned, reverse ion exchange processes (IEPs) are present, in which Ca 2+ + Mg 2+ are released into groundwater and Na + and K + are fixed to the aquifer material, causing water hardening [49].The graphical data of the wet season (Figure 8b) show that the direct ion exchange process (IEP) together with mineralization are the most relevant processes during the wet season.This was also evidenced by the increase in sodium concentrations in the Arroyo Grande in the same season.The occurrence of IEP was also verified with the application of chloro-alkali indices (CAI).
Figure 9 shows that the index values are mostly higher than zero for the data obtained from the first sampling campaign, confirming the occurrence of reverse ion exchange processes in the dry season.This changes for the data of the wet season, where the points of the graph show a tendency towards zero, indicating the occurrence of the direct IEP.
Mg 2+ are released into groundwater and Na + and K + are fixed to the aquifer material, causing water hardening [49].The graphical data of the wet season (Figure 8b) show that the direct ion exchange process (IEP) together with mineralization are the most relevant processes during the wet season.This was also evidenced by the increase in sodium concentrations in the Arroyo Grande in the same season.The occurrence of IEP was also verified with the application of chloro-alkali indices (CAI).Figure 9 shows that the index values are mostly higher than zero for the data obtained from the first sampling campaign, confirming the occurrence of reverse ion exchange processes in the dry season.This changes for the data of the wet season, where the points of the graph show a tendency towards zero, indicating the occurrence of the direct IEP.Salinization is another process evidenced by the chemical composition of the sampled water.This can occur in onshore areas due to the dumping of untreated wastewater or due to the intrusion of seawater.Due to their hydraulic connection with the sea, coastal aquifers are subject to the intrusion of seawater with a high salinity content, a process that occurs with the rupture of the equilibrium or transition zone, generated by the increase in sea levels and the overexploitation of the aquifer [61].From the Cl anion, it was possible to analyze the changes to which the groundwater is exposed due to salinization processes.
The sampling point PZ N°030, which corresponds to a deep well (80 m) located 680 m from the coast, showed Cl -concentrations of 234 mg/L in the dry season and 226.9 mg/L in the wet season.In the case of water with neutral and natural pH, the expected concentration range of this anion is from 1.8 to 71 mg/L [42].Therefore, an increase in chloride concentrations can be detected, which decreases inland from the coast as shown in the isoline map of Figure 10, except for sampling point AJ N°023, which corresponds to a cistern.Point AJ N°023 draws water from the shallow aquifer and shows higher chloride concentrations than the deep wells in the sector due to the infiltration of wastewater from septic systems.Salinization is another process evidenced by the chemical composition of the sampled water.This can occur in onshore areas due to the dumping of untreated wastewater or due to the intrusion of seawater.Due to their hydraulic connection with the sea, coastal aquifers are subject to the intrusion of seawater with a high salinity content, a process that occurs with the rupture of the equilibrium or transition zone, generated by the increase in sea levels and the overexploitation of the aquifer [61].From the Cl anion, it was possible to analyze the changes to which the groundwater is exposed due to salinization processes.
The sampling point PZ N • 030, which corresponds to a deep well (80 m) located 680 m from the coast, showed Cl − concentrations of 234 mg/L in the dry season and 226.9 mg/L in the wet season.In the case of water with neutral and natural pH, the expected concentration range of this anion is from 1.8 to 71 mg/L [42].Therefore, an increase in chloride concentrations can be detected, which decreases inland from the coast as shown in the isoline map of Figure 10, except for sampling point AJ N • 023, which corresponds to a cistern.Point AJ N • 023 draws water from the shallow aquifer and shows higher chloride concentrations than the deep wells in the sector due to the infiltration of wastewater from septic systems.Similarly, the dispersion graph (Figure 11), constructed from the parameters of electrical conductivity (EC) and chlorides (Cl − ), depicts that the wells near the coast are located in an area of mixing of fresh water with seawater.Near the coast, Cl -exceed concentrations of 200 mg/L and EC greater than 1000 mS/cm [24], occurring both in dry and wet seasons.Similarly, the dispersion graph (Figure 11), constructed from the parameters of electrical conductivity (EC) and chlorides (Cl − ), depicts that the wells near the coast are located in an area of mixing of fresh water with seawater.Near the coast, Cl − exceed concentrations of 200 mg/L and EC greater than 1000 mS/cm [24], occurring both in dry and wet seasons.Similarly, the dispersion graph (Figure 11), constructed from the parameters of electrical conductivity (EC) and chlorides (Cl − ), depicts that the wells near the coast are located in an area of mixing of fresh water with seawater.Near the coast, Cl -exceed concentrations of 200 mg/L and EC greater than 1000 mS/cm [24], occurring both in dry and wet seasons.

Principal Component Analysis (PCA)
For the multivariate analysis, a total of 12 parameters were considered in 14 sampling points, pH, CE, TDS, Cl − , HCO 3 − , NO 3 − , SO 4 2− , Fe 2+ , Na + , Ca 2+ , Mg 2+ , and Mn + .The application of the principal component analysis to the standardized data of sampling campaign No. 1 shows that three components explain 83.5% of the total variance of this group of data, with three components obtaining eigenvalues greater than or equal to one, a condition that can be observed in the sedimentation plot of Figure 12a.
For the multivariate analysis, a total of 12 parameters were considered in 14 sampling points, pH, CE, TDS, Cl − , HCO3 − , NO3 − , SO4 2− , Fe 2+ , Na + , Ca 2+ , Mg 2+ , and Mn + .The application of the principal component analysis to the standardized data of sampling campaign No. 1 shows that three components explain 83.5% of the total variance of this group of data, with three components obtaining eigenvalues greater than or equal to one, a condition that can be observed in the sedimentation plot of Figure 12a.
In the case of the data from the second sampling campaign, the sedimentation graph suggests taking four components, but considering the variance provided by each of them, the first three are the most representative and provide an accumulated variance of 76.36% (Figure 12b).The first component of the dry season explains 39.61% of the variance (Table 8), where the variables EC, TDS, HCO3 − , Ca 2+ , and Mg 2+ are the ones with the greatest loadings, indicating mineral dissolution processes such as calcite and dolomite [25,49].In the case of the data from the second sampling campaign, the sedimentation graph suggests taking four components, but considering the variance provided by each of them, the first three are the most representative and provide an accumulated variance of 76.36% (Figure 12b).
The first component of the dry season explains 39.61% of the variance (Table 8), where the variables EC, TDS, HCO 3 − , Ca 2+ , and Mg 2+ are the ones with the greatest loadings, indicating mineral dissolution processes such as calcite and dolomite [25,49].The variables with the greatest loadings in component two are Na + and NO 3 − , with 28.85% of the total variance, which owe their presence in the chemical composition of the water to ion exchange processes in the case of sodium and anthropogenic contamination.Chloride (Cl − ) is the variable with the greatest loadings in component three, which explains 15.03% of the total variance.This anion is mainly related to salinization processes, either by seawater intrusion or by liquid waste dumping [55,62].
The PCA results for the wet season show that the first component explains 42.07% of the total variance.The variables with the largest loadings were CE, TDS, Cl − , and Na + and to a lesser extent Ca 2+ and Mg 2+ .It can be noted that the direct ion exchange process, with a contribution of sodium to the composition of the water, is the factor that determines this component, along with salinization processes.Component two accounts for 21.27% of the total variance for this data group, with Fe 2+ , Mn + , Ca 2+ , and Mg 2+ as the variables with the greatest weight.In addition to mineral dissolution processes, the increase in iron and manganese concentrations can be related to sulfate and nitrate reduction processes [56].
Finally, component three has a variance of 21.27% of the total variance; the variables with the greatest loadings are HCO 3 − , NO 3 − , and Ca 2+ , related to mineral dissolution processes and anthropogenic activity.
In both the dry season and wet seasons, mineral dissolution processes and ion exchange processes influence the water quality, with the latter becoming more relevant during the wet season.This is evidenced by the types of water obtained according to the data of the second sampling campaign, where sodium (Na + ) becomes one of the predominant cations in groundwater and the variable with the greatest weight in component number one for this group of data.

Water Quality Index (WQI)
The results obtained through the application of the WQI show that the water in the study area is not suitable for drinking purposes according to the water quality standards.For the two sampling campaigns, more than 50% of the surface and groundwater points were found to be unsuitable for drinking according to the classification based on the WQI value (Table 9).Less than 30% of the sampling points have a good water quality status.In the specific case of groundwater, which shows the most unfavorable quality conditions, the concentrations of iron and manganese impacted the calculation of the WQI.Other elements, such as calcium, magnesium, sodium, and bicarbonates, which are introduced into the water through mineral dissolution processes, increase the mineralization state of the groundwater and affect the quality of the water [63].Regarding the high presence of iron and manganese in groundwater, it should be noted that prolonged consumption of high concentrations of these minerals can cause diseases such as Alzheimer's, Huntington's, Parkinson's, and other cardiovascular and respiratory diseases, as well as cognitive impairments [53].

Conclusions
The proposed framework offered a systematic and robust approach to groundwater quality assessment, addressing key challenges in understanding and managing complex hydrochemical systems.By integrating advanced analytical tools and data-driven methodologies, this framework contributes to advancing scientific knowledge and informs evidence-based decision-making for sustainable water resource management.Some of the key findings of the application of the framework in the Arroyo Grande aquifer include the following: • The groundwater and surface waters of the aquifer evidenced distinct levels of mineralization due to the occurrence of mechanisms such as mineral dissolution, ion exchange, seawater intrusion, and anthropogenic contamination.• The characteristics acquired by the water affected the availability of freshwater suitable for drinking for the communities near the area, to the point that in the dry and wet seasons, at least 50% of the sampled groundwater points are inadequate for drinking and fish farming.

Water 2024 , 27 Figure 1 .
Figure 1.Location of study area in the northern area of Colombia.The area in pink represents the Arroyo Grande aquifer, and the zone delineated in black is the area of this study.The blue line (Arroyo Grande stream) and the pink line (Bonga stream) show the two main streams that serve the aquifer, and the red diamonds show the location of the vertical electrical soundings (VESs).

Figure 1 .
Figure 1.Location of study area in the northern area of Colombia.The area in pink represents the Arroyo Grande aquifer, and the zone delineated in black is the area of this study.The blue line (Arroyo Grande stream) and the pink line (Bonga stream) show the two main streams that serve the aquifer, and the red diamonds show the location of the vertical electrical soundings (VESs).
Figure 2 was developed by the authors using ArcGisPro 11.0.The layers of geology and the geological data used for this map are open data from the Geologic map of Colombia 2023, available at the Colombian Geologic Survey [35].Water 2024, 16, x FOR PEER REVIEW 6 of 27 geology and the geological data used for this map are open data from the Geologic map of Colombia 2023, available at the Colombian Geologic Survey [35].

Figure 2 .
Figure 2. Chronostratigraphic geological units in the study area surrounding the Arroyo Grande aquifer.The area delineated in green refers to the area selected for this study, and the area delineated in blue is the Arroyo Grande formation.

Figure 2 .
Figure 2. Chronostratigraphic geological units in the study area surrounding the Arroyo Grande aquifer.The area delineated in green refers to the area selected for this study, and the area delineated in blue is the Arroyo Grande formation.

Water 2024 ,
16, x FOR PEER REVIEW 7 of 27

Figure 3 .
Figure 3. Lithological columns derived from Vertical Electrical Soundings (VESs) in the area of study [37].The resistivity of each stratum is presented, alongside the VES location coordinates and the location of the water table.

Figure
Figure 3. Lithological columns derived from Vertical Electrical Soundings (VESs) in the area of study [37].The resistivity of each stratum is presented, alongside the VES location coordinates and the location of the water table.

Figure 4 .
Figure 4. Location and code of sampling points.Deep well refers to a mechanically drilled structure with diameters typically ranging from 0.15 to 0.3 m and depths exceeding 30 m (P008: 77 m, P009: 80 m, P014: 80.5 m, P020: 69 m, P022: 72 m P030: 76 m, P032: 72 m).Shallow well refers to a large diameter structure built manually, which reaches the water table and is deepened below it to capture groundwater; it does not usually exceed 30 m deep (A018: 4 m, A023: 12 m A028: 23 m) [43].

Figure 5 .
Figure 5. Schoeller graph for the ten sampled groundwater points during the dry and wet season.

Figure 5 .
Figure 5. Schoeller graph for the ten sampled groundwater points during the dry and wet season.
Water 2024, 16, x FOR PEER REVIEW 15 of 27

Figure 6 .
Figure 6.Piper diagram for the ten wells sampled for groundwater quality in the dry and wet season.The legend refers to the ten groundwater sampling points assessed in this study and previously located in Figure 3.

Figure 6 .
Figure 6.Piper diagram for the ten wells sampled for groundwater quality in the dry and wet season.The legend refers to the ten groundwater sampling points assessed in this study and previously located in Figure 3.

Water 2024 , 27 Figure 7 .
Figure 7. (a) Gibbs diagram results during the dry season and (b) wet season.

Figure 7 .
Figure 7. (a) Gibbs diagram results during the dry season and (b) wet season.

Figure 9 .
Figure 9. Occurrence of direct or reverse ion exchange processes (IEPs) from the dry and wet season for both chloro-alkaline indices CAI 1 and CAI 2, as previously described in Equations (2) and (3).

Figure 10 .
Figure 10.Isoline map of chloride in groundwater-dry season.The concentration of each sampling point is presented in parentheses.

Figure 11 .Figure 10 .
Figure 11.Relationship between EC and Cl − for dry and wet season-the purple zone depicts the freshwater zone (groundwater samples with Cl − concentration less than 200 mg/L and EC~1000

Water 2024 , 27 Figure 10 .
Figure 10.Isoline map of chloride in groundwater-dry season.The concentration of each sampling point is presented in parentheses.

Figure 11 .
Figure 11.Relationship between EC and Cl − for dry and wet season-the purple zone depicts the freshwater zone (groundwater samples with Cl − concentration less than 200 mg/L and EC~1000

Figure 11 .
Figure 11.Relationship between EC and Cl − for dry and wet season-the purple zone depicts the freshwater zone (groundwater samples with Cl − concentration less than 200 mg/L and EC~1000 µS/cm) and the yellow zone represent the mixing zone between fresh water and seawater (groundwater samples with Cl − exceeding 200 mg/L and EC exceeding 1000 µS/cm).

Figure 12 .
Figure 12.(a) Sedimentation graph PCA analysis data sampling from the dry season and (b) the wet season.

Figure 12 .
Figure 12.(a) Sedimentation graph PCA analysis data sampling from the dry season and (b) the wet season.

Table 1 .
Main characteristics of the reviewed studies in coastal aquifers.

Table 2 .
Framework to evaluate groundwater quality and the relationship between rock weathering and groundwater hydrogeochemistry.

Table 3 .
Weight assignment to physicochemical parameters.

Table 5 .
Descriptive statistics of parameters measured on site at 15 sampling points-Campaign No. 1 (dry season)-Campaign No. 2 (wet season).

Table 6 .
Descriptive statistics of majority anions (mg/L) at 10 groundwater sampling points in the dry and wet season.

Table 8 .
Principal component loadings.In bold are highlighted the variables with the highest loading for each component.

Table 8 .
Principal component loadings.In bold are highlighted the variables with the highest loading for each component.

Table 9 .
Classification of water samples according to water quality status.

•
The groundwater of the Arroyo Grande hydrogeological unit exhibits high spatial and temporal variability in its main components.The heterogeneity and dynamics of this aquifer are evidenced in the several types of water found, which correspond to sodium sulfate, calcium bicarbonate, sodium chloride, sodium bicarbonate, magnesium sulfate, and magnesium chloride.•Thedominantelements in the Arroyo Grande water chemistry are the anions bicarbonate, chloride, and sulfate, in addition to cations sodium, calcium, and magnesium.The rock-water interaction mechanism is the most relevant in defining the chemical signature of the groundwater, through mineral dissolution and ion exchange processes related to the geological information of the area.The identification of the exact minerals that are contributing to the presence of Fe in the groundwater of the Arroyo Grande aquifer requires further mineralogical studies.•Otherfactors affecting the chemical composition are the intrusion of seawater and anthropogenic activities.An increase in the saline intrusion phenomenon has been observed, currently affecting the extraction points near the coastal strip, which have high values of electrical conductivity and chlorides.Concentrations of iron and manganese are higher than drinking water standards, which requires constant monitoring of these parameters.

Table A4 .
Analytical errors from the laboratory analyses.