A Review of Riverbank Filtration with a Focus on Tropical Agriculture for Irrigation Water Supply

Abstract

Tropical agriculture requires sustainable irrigation solutions that balance water availability with quality and environmental protection. This review synthesizes current knowledge on riverbank filtration (RBF)—a nature-based technology for improving agricultural water quality—with objectives to elucidate design principles, water quality performance, and operational challenges specific to tropical contexts. Through systematic analysis of 128 peer-reviewed articles across topics including RBF hydrogeology, contaminant removal mechanisms, sediment transport, pathogen reduction, site selection criteria, and monitoring strategies, this work consolidates interdisciplinary evidence on RBF effectiveness for irrigation water supply. The Roldanillo–Unión–Toro (RUT) district in Valle del Cauca, Colombia, serves as a case study illustrating RBF application to sediment-rich, pathogen-prone rivers typical of tropical agricultural regions. While RBF is established for drinking water supply in temperate zones, its adaptation to tropical irrigation remains underexplored. This review identifies critical hydrogeological, environmental, and operational considerations for implementing RBF systems in tropical agricultural settings characterized by high water demand, seasonal variability, and challenging water quality conditions. Key findings are synthesized into a practitioner-oriented framework—covering site selection, design optimization, and adaptive management—intended to guide deployment of RBF for irrigation in tropical agricultural settings.

1. Introduction

Modern agriculture sits at the intersection of intensifying resource demands and declining water security. Population growth and economic development amplify pressures on the water–food–energy nexus, while groundwater—already the backbone of global irrigation—faces rapid depletion in many regions [1,2]. In tropical production systems, the quality and quantity of available water frequently fall short of what pressurized irrigation technologies require to operate reliably, leading to recurrent clogging, higher maintenance costs, shortened system lifespans, and health risks where vegetables and leafy crops are irrigated with contaminated sources [3,4,5]. These constraints motivate nature-based pretreatment strategies that can transform variable-quality surface water into irrigation-fit supplies while protecting infrastructure and public health.

Riverbank filtration (RBF) is a mature, low-energy, nature-based technology that improves surface water quality by inducing lateral flow through riverine sediments toward strategically placed wells, where physical straining, sorption, redox transformations, and microbial processes attenuate particulates and contaminants [6]. Developed primarily for drinking-water pretreatment, RBF reduces dependence on chemical dosing and advanced membranes while providing a stable raw-water quality for downstream treatment trains [7]. Despite extensive potable-water experience, its agricultural potential remains comparatively underexplored, particularly in settings with high turbidity, sediment loads, and microbial contamination that challenge drip and sprinkler systems [8]. RBF performance depends on sediment grain size distributions, aquifer permeability, well configuration, and surface-water hydroperiod, with the hyporheic zone acting as a reactive biogeochemical interface that enhances contaminant removal (Figure 1). The approach has been operated for more than two centuries and today contributes substantially to national supplies in parts of Europe within multi-barrier frameworks [9,10,11,12].

When adapted for irrigation, the performance targets shift from potability to preventing emitter clogging, reducing microbial risk at the crop–water interface, and extending system longevity [13]. Target thresholds emphasize low turbidity and minimized concentrations of Fe, Mn, and ammonium, alongside reduced microbial indicators, in line with national regulations and international agricultural-water guidelines aimed at crop safety and infrastructure protection [14,15]. Crop specificity further conditions design goals: microbiologically sensitive horticultural commodities demand higher-quality water than broad-acre cereals, and salinity or metal accumulation constraints vary by crop and soil context [16,17,18]. Accordingly, irrigation-oriented RBF can operate at shorter residence times and higher flow rates than potable systems while maintaining acceptable quality, supporting cost-effective deployment in rural and peri-urban agricultural regions and strengthening resilience under climatic variability [19,20].

Operational evidence underscores these functions. Large well fields in high-permeability alluvial settings have demonstrated sustained induced infiltration, high well yields, and strong particulate and pathogen attenuation along flow paths from river to production wells [21]. Reported outcomes include non-detects for protozoan pathogens in groundwater despite riverine occurrence, multi-log reductions of particulate surrogates, and consistently low turbidity in abstracted water—conditions that directly mitigate clogging risks in pressurized irrigation and lower microbial hazards for fresh-produce systems [22]. Beyond microbial safety, RBF can attenuate an array of organic and inorganic contaminants—including pesticides, pharmaceuticals, and select metals—an advantage in basins receiving domestic effluents and diffuse agrochemical inputs [23,24,25]. Tropical applications introduce predictable but manageable challenges. Elevated temperatures can compress oxic zones and, under slow infiltration, promote transitions to suboxic or reducing conditions that mobilize Mn²⁺, Fe²⁺, NH₄⁺, and NO₂⁻—necessitating careful design, monitoring, and, where needed, simple polishing steps [12,26] (Figure 2).

Site selection and design are therefore paramount: aquifer texture, organic-matter content, and redox regime govern permeability and reaction windows, with well-sorted sandy alluvium generally most favorable for sustained performance [27]. Riverbed hydraulic conductivity and its temporal stability—strongly influenced by grain size, sediment thickness, and episodic clogging—control induced infiltration and groundwater–surface-water exchange; sensitivity analyses in large alluvial systems show riverbed properties can dominate groundwater-level dynamics and RBF sustainability over other recharge components [28].

This review synthesizes current knowledge on RBF principles, design parameters, water quality benefits, operational challenges, and case studies, with specific objectives to (1) establish a comprehensive understanding of RBF effectiveness and applicability in tropical agricultural contexts, and (2) identify critical site selection, hydrogeological, and design considerations for implementing RBF systems in water-scarce agricultural regions. The Roldanillo–Unión–Toro (RUT) district in Valle del Cauca, Colombia, serves as a case study illustrating RBF adaptation to sediment-rich, pathogen-prone rivers typical of tropical agricultural zones.

2. Methodology

To ensure methodological rigor and comprehensive coverage, this review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework [29]. A bibliometric search strategy was implemented across four major scientific databases—Scopus, ScienceDirect, Web of Science, and SpringerLink—supplemented by targeted searches in specialized repositories. Boolean operators (AND, OR, NOT) were applied to a set of predefined keywords, which included “riverbank filtration,” “bank filtration,” “geochemical conditions,” “geohydraulic conditions,” “transport modeling,” “seasonal variation in riverbank filtration,” “performance of riverbank filtration,” “aquifer transmissivity,” “colmation layer,” “clogging issues,” “spatiotemporal variations,” “redox conditions,” “surface water–groundwater interactions,” “removal efficiency,” “micropollutants,” “bacteriological parameters,” “water use efficiency,” and “drip irrigation systems.” To expand coverage, forward and backward citation tracking was carried out using ResearchRabbit, generating additional relevant studies. In total, 1077 records published between 2000 and 2025 were initially retrieved. After removing duplicates and screening titles and abstracts, 164 articles were retained for eligibility assessment. The full text of these articles was then reviewed for relevance and applicability to the scope of this study. Ultimately, 128 peer-reviewed publications met the inclusion criteria and were synthesized to provide comprehensive insights into the applications of RBF in tropical agricultural irrigation systems.

This review is structured to address the stated objectives systematically. Section 3 establishes the technological foundations of RBF, covering induced hydraulics, collector well configurations, infiltration performance, and natural attenuation mechanisms active along subsurface flow paths—providing the principles necessary to understand RBF effectiveness globally. Section 4 examines site-specific design parameters, including aquifer transmissivity, hydraulic conductivity, redox conditions, and well placement strategies, which are critical for optimizing RBF performance in diverse hydrogeological contexts. Section 5 synthesizes water quality improvements documented across RBF installations worldwide, with emphasis on turbidity reduction, pathogen attenuation, and contaminant removal under varying climatic and operational conditions. Section 6 applies these principles to the Colombian RUT district, presenting hydrogeological characterization, water quality challenges, and site suitability assessment as a case study illustrating RBF adaptation to sediment-rich, pathogen-prone tropical rivers. Section 7 identifies operational and environmental challenges—including clogging, redox shifts, and seasonal variability—that constrain RBF sustainability in tropical settings. Section 8 outlines a research agenda for pilot-scale validation, monitoring protocols, and knowledge gaps specific to tropical agricultural irrigation. Collectively, these sections provide a practical framework for site selection, design optimization, and adaptive management to enhance irrigation system resilience and advance sustainable water resource management in vulnerable agricultural zones, directly addressing the objectives articulated in Section 1.

3. RBF Technology Background and Principles

3.1. Induced Hydraulics, Collector Wells, and Infiltration Performance

Riverbank filtration operates on hydrodynamic principles whereby hydraulic gradients—generated through continuous groundwater extraction by wells adjacent to rivers—induce surface water infiltration through riverbed sediments. As water moves through porous media, it undergoes natural attenuation processes including physical filtration, chemical adsorption, ion exchange, and microbial degradation [30,31]. Field-scale systems in alluvial aquifers with transmissivity ≥ 1500 m² d⁻¹ and saturated thickness > 6 m consistently achieve high infiltration rates [32]. Horizontal collector wells—often installed with 150–300 m of screen length, and in some cases up to 750 m—expose larger inflow areas, reducing entrance velocities, minimizing clogging, and maintaining low head loss between river and well. In European applications, such designs yield between 900 and 3000 m³ h⁻¹ per well, with >50 collector wells along the Rhine and >200 in the Danube basin [32]. Infiltration velocities are typically low enough to maximize particulate and microbial removal, with observed turbidity reductions and pathogen attenuation exceeding 2–3 log units in many installations [33]. Proper sitting in coarse-grained, hydraulically connected deposits, combined with optimal pumping rates, prevents excessive clogging and sustains both water quantity and quality over decades of operation.

Hydraulic residence time—the period infiltrated water remains subsurface—critically determines RBF performance. Longer residence times enhance contact with mineral surfaces, microbial biomass, and SOM, enabling more effective contaminant removal. Key variables controlling residence time include well-to-riverbank distance, pumping rate, sediment permeability, and aquifer porosity (typically 0.32–0.38) [12]. Under optimal conditions, residence time spans several hours to months, allowing effluent quality stabilization and contamination pulse attenuation [34]. At the Jenderam Hilir RBF site in Selangor, Malaysia, travel distances from river to wells of ~20–30 m combined with moderate pumping rates produced marked improvements in water quality: turbidity was reduced from 230.7 NTU in the Langat River to 0.01 NTU post-pumping, COD decreased from 11.67 to 3.67 mg L⁻¹, and suspended solids from 89.3 to 6.33 mg L⁻¹, with the Water Quality Index improving from Class IV (50.16) to Class III (60.44) [35]. These outcomes reflect subsurface residence times sufficient for physical filtration, ion exchange, and microbial degradation to act synergistically. In well-designed systems with aquifer porosities of 0.32–0.38 and hydraulic conductivities exceeding 10⁻⁴ m s⁻¹, residence times can range from several hours to multiple months depending on well placement and abstraction rates, enabling effective attenuation of turbidity, pathogens, nutrients, and trace organics [36]. Optimal configurations balance travel time and yield, as excessive drawdown can shorten residence time, reducing contaminant removal efficiency, while overly long residence times may risk anoxic conditions and associated geochemical mobilization.

Surface–groundwater mixing ratios depend on well placement, aquifer morphology, and pumping regime. In highly connected alluvial settings, wells within 20–100 m of rivers may capture up to 90% surface water, as observed in Düsseldorf, Germany [11]. Isotopic and geochemical tracer analyses in the Saint-Charles River watershed (Québec, Canada) showed that vertical wells 30–50 m from the riverbank intercepted 53–78% surface water under moderate pumping, whereas wells > 150 m away captured <35%, with the balance derived from native groundwater [37]. In the Nakdong River system (South Korea), Ref. [38] found that bank filtrate comprised 62–88% of pumped water at wells 40–70 m from the channel, decreasing to 41–57% at 120 m distance, with higher groundwater fractions correlating with lower DOC and microbial counts but also reduced nitrate removal efficiency. These results highlight that both water yield and treatment performance depend on the proportion of infiltrated river water, which in turn is governed by hydraulic connectivity, sediment heterogeneity, and abstraction rates.

Well configuration directly influences induced flow rates and water composition. Vertical wells are most common due to construction simplicity, while horizontal wells and infiltration galleries offer higher contact surface areas and improved residence times in shallow aquifers. Egyptian systems using wells 20–100 m from the Nile achieved residence times of 1–7 days with significant fecal coliform and nutrient reductions. However, capture efficiency declined during droughts, with surface water fractions dropping from 70% to 40% [39]. In the Kamphaengphet RBF pilot in Thailand, horizontal collector wells placed along the Ping River intercepted bank filtrate from unconsolidated aquifers with permeability > 10 m/day and thickness ≥ 20 m, achieving sustained yields suitable for provincial water supply while maintaining groundwater TDS ≤ 300 mg/L [40]. In another study in Egypt, design trials with vertical wells located 20–50 m from the Nile demonstrated that increasing lateral distances and optimizing screen placement improved infiltration rates and pathogen removal efficiency, with wells in higher-permeability zones producing greater yields and better water quality [41].

Infiltration rates are governed by vertical hydraulic conductivity (K_x) of sediments. Under saturated flow conditions, K_x controls infiltration rates. When river stages exceed local groundwater levels, unsaturated flow predominates, significantly lowering K_x and impairing residence time and contaminant removal [42]. These effects intensify during droughts and floods, destabilizing redox conditions and altering flow paths.

3.2. Sediment-Rich Tropical Rivers: Challenges and Adaptations for RBF Systems

Tropical regions face significant challenges in implementing sustainable RBF systems due to extreme turbidity conditions that accelerate sediment clogging and compromise system longevity. Indian case studies along the Yamuna and Ganga rivers demonstrate these constraints, where turbidity values frequently exceed 100–300 NTU during monsoon floods, with extreme events surpassing 1000 NTU [43]. These conditions severely impact riverbed infiltration and well performance, with operational data from Uttarakhand and Uttar Pradesh revealing infiltration rate reductions exceeding 50% during prolonged flood events, necessitating temporary system shutdowns or costly well rehabilitation [43]. The combination of high sediment loads and monsoon-driven flow variations accelerates bio-physical clogging processes, reducing expected well lifespans from 20–30 years to less than 10 years without proper management interventions. Indian RBF adaptations include elevated wellheads positioned above flood levels, distributed shallow well networks to reduce hydraulic stress, and pre-treatment systems incorporating gravel filter drains for coarse sediment interception [43]. These design modifications highlight the critical need for flood-resilient RBF configurations in tropical environments where seasonal hydrology drives extreme turbidity dynamics.

The Wakaf Bunut water treatment plant in Kelantan, Malaysia, exemplifies effective integration of RBF with ultrafiltration technology to address high-turbidity challenges characteristic of tropical rivers. Commissioned in 2013, this hybrid system delivers 14 million liters per day through an RBF pre-treatment stage followed by ultrafiltration processing via 120 Dizzer XL 0.9 MB 60 W modules [44]. Despite raw river water turbidity ranging from 11.8 to 51.4 NTU, the integrated system consistently achieves effluent turbidity below 0.3 NTU, meeting Malaysia’s potable water standards. The RBF component alone reduces turbidity from 30.4 to 18.2 NTU (40% reduction), typically delivering water at 3–7 NTU and substantially reducing particle loading on downstream membrane systems. Additional water quality improvements include iron reduction from 1.5 mg L⁻¹ to below 0.05 mg L⁻¹ and pH stabilization within 7.5–8.4 [44]. This hybrid approach eliminates pre-coagulation and flocculation requirements while maintaining exceptionally low operational costs of RM 0.13 per m³ (approximately USD 0.04 per m³), establishing it among Southeast Asia’s most cost-efficient municipal-scale treatment plants [44]. The case demonstrates that RBF-membrane integration provides resilient solutions for high-turbidity tropical rivers by minimizing membrane fouling, ensuring consistent water quality, and reducing operational expenses.

Northern Thailand’s Ping River system illustrates how extreme turbidity in tropical monsoon environments accelerates riverbed clogging and limits RBF efficiency. Seasonal turbidity commonly exceeds 100 NTU, with experimental studies at Chiang Mai sites (Mae Rim and San Pa Tong) documenting peaks of 200–950 NTU during wet season conditions [45]. Channel and column experiments revealed that low flow velocities (<0.05 m s⁻¹) combined with high turbidity (>200 NTU) caused K_x of the upper 3 cm sediment layer at Mae Rim to decrease nearly one order of magnitude, from 5.6 × 10⁻⁶ to 7.3 × 10⁻⁷ m s⁻¹ within 22 days. Although increasing flow velocity to 0.38 m s⁻¹ temporarily restored K_x values, conductivity declined again due to persistent fine particle intrusion [45]. At San Pa Tong, where turbidity ranged from 33 to 145 NTU, clogging extended to greater depths (0–40 cm), and manual removal of external filter cake achieved only partial K_x recovery because internal clogging dominated, reducing conductivity by over 60% in affected layers [45]. These findings underscore that tropical rivers experiencing monsoon-driven floods that mobilize massive-suspended loads require RBF designs accounting for both external and internal clogging mechanisms. Failure to address these dynamics leads to overestimated abstraction rates and premature system underperformance.

The Beberibe River in Pernambuco State, Brazil, demonstrates RBF capabilities for addressing high turbidity and eutrophication challenges in tropical catchments. The pilot system consisted of one production well positioned 65 m from the river and seven observation wells, operating at an average abstraction rate of 12.6 m³ h⁻¹ (3.5 L s⁻¹) [46]. Raw river water with mean turbidity of 34 NTU underwent 97.7% turbidity reduction, producing bank filtrate averaging 0.8 NTU—well below the 5 NTU drinking water standard. Chemical improvements were substantial: ammonia decreased from 4.1 to 0.45 mg L⁻¹, nitrite from 0.46 to 0.005 mg L⁻¹, nitrate from 1.3 to 0.5 mg L⁻¹, biochemical oxygen demand (BOD) from 6.2 to 1.7 mg L⁻¹, and chemical oxygen demand (COD) from 26.8 to 9.2 mg L⁻¹. Heavy metal attenuation was equally effective, with Fe reduced from 2.0 to 0.09 mg L⁻¹ and Mn from 0.06 to 0.04 mg L⁻¹, both achieving Brazilian and WHO potability standards. Microbial removal was complete, with total coliforms (1500–31,000 MPN/100 mL) and fecal coliforms (300–3400 MPN/100 mL) absent in filtrate samples. Notably, cyanobacteria populations in river water (1678–15,059 cells mL⁻¹), including toxin-producing genera such as Oscillatoria and Aphanizomenon, were entirely eliminated through RBF treatment, removing significant health risks [46]. These results demonstrate that RBF systems can reliably produce potable-quality water even under stress conditions typical of tropical rivers characterized by high suspended loads and recurrent algal blooms, highlighting RBF potential as a pre-treatment technology for sustainable water supply in warm, turbidity-prone environments.

Colombian rivers present extreme turbidity challenges, with the Cauca and Magdalena rivers recording turbidity peaks up to 10,000 NTU that force frequent shutdowns of major treatment facilities such as Puerto Mallarino in Cali [47]. Colombian Pacific basin sediment yields reach 1150–1714 t km⁻² yr⁻¹, while the Magdalena–Cauca system averages 560 t km⁻² yr⁻¹—values comparable to or exceeding major Asian river systems. These sediments carry substantial contaminant loads, including an estimated 122 kg d⁻¹ mercury, 2600 kg d⁻¹ lead, 3300 kg d⁻¹ Cd, and 490 kg d⁻¹ Cr, plus over one million kilograms daily of N and P [47]. RBF systems have been evaluated as sustainable alternatives under these extreme conditions due to their capacity for turbidity attenuation, pathogen removal, and selective heavy metal reduction while simultaneously decreasing chemical input requirements and operational complexity compared to conventional treatment plants. However, the intensive suspended loads accelerate physical clogging at the river–aquifer interface, reducing infiltration capacity unless counterbalanced by natural scouring during high-flow events [47]. Under Colombian conditions, RBF could significantly dampen shock loads and stabilize water quality for downstream treatment processes, but careful design must address the trade-off between system productivity and long-term clogging risks imposed by the extreme sediment dynamics characteristic of these tropical river systems.

3.3. Attenuation Mechanisms Along the Subsurface Flow Path

RBF activates synergistic physical, chemical, and biological mechanisms as water infiltrates. Physical filtration retains suspended solids and pathogens within the first few dm of riverbed sediments, where the colmation layer functions as a primary barrier (Figure 3). Field-scale investigations along the River Elbe in Germany demonstrated that suspended solids and microbial contaminants are substantially reduced within the upper 20–30 cm of the sediment bed, where the colmation layer promotes pore clogging and particle entrapment [7]. This fine-grained, biofilm-rich layer contributed to a 2–3 log unit reduction of Escherichia coli and enterococci within infiltration distances as short as 0.5 m from the riverbed surface, corresponding to travel times of less than 24 h under hydraulic gradients of 0.02–0.04 [7]. Similarly, in a Hungarian RBF system along the Danube, pilot column experiments confirmed that >90% of suspended particulate matter and associated organic-bound micropollutants were removed within the first 25 cm of infiltration, with removal efficiencies remaining stable over seasonal temperature variations between 5 °C and 20 °C [48]. Moreover, mesocosm studies simulating bank filtration under controlled flow conditions revealed that the synergy between fine-particle trapping, extracellular polymeric substances from biofilms, and sediment surface charge enhanced the capture of viruses and protozoan cysts, with retention efficiencies > 99% for Cryptosporidium oocysts within the first 10 cm of travel [49].

Chemical attenuation in RBF systems primarily occurs through adsorption onto mineral surfaces such as Fe and Mn oxides, clay minerals, and soil organic matter (SOM). These reactive surfaces effectively bind a range of contaminants, including Fe²⁺, Mn²⁺, Zn²⁺, NH₄⁺, NO₃⁻, dissolved organic matter (DOM), and various emerging pollutants. Adsorption efficiency is strongly influenced by environmental conditions, particularly pH—optimal in the range of 7.2–8.1—along with temperature and the heterogeneity of sediment composition [27]. In addition to adsorption, ion exchange processes facilitate the removal of charged species such as Ca²⁺, Mg²⁺, NH₄⁺, and other inorganic and organic contaminants, further improving water quality [33,50,51,52].

Microbial biotransformation is central to RBF functionality. Indigenous microorganisms in hyporheic zones degrade organic contaminants, facilitate denitrification, and metabolize trace organics including pharmaceuticals. Redox gradients follow typical sequences: aerobic respiration, nitrate reduction, and Mn/Fe oxide reduction under anoxic conditions [53]. Evidence from [54] shows that targeted manipulation of microbial communities in sediment systems can markedly enhance these biotransformation processes. In a CSO-impacted urban river, remediation combining calcium nitrate, composite functional microorganisms, and low-DO aeration increased denitrification pathway enzyme genes by 233.98% and nitrification-related genes by 9.32%, while suppressing dissimilatory nitrate reduction to ammonium (DNRA) by 65.27%. These interventions improved redox conditions (DO up to +900%, ORP +198.75%) and promoted removal of NH₃–N (82.11% reduction) and acid volatile sulfide (94.46% reduction) within 30 days [54]. Microbial community shifts favored Proteobacteria and Chloroflexi as functional hubs, while enhanced negative co-occurrence patterns stabilized network structure, reducing vulnerability by 54.11% and increasing compositional stability by 86.12%. These results demonstrate that optimizing redox gradients and microbial network stability directly supports the sequential aerobic–denitrifying–metal-reducing transformations essential to contaminant attenuation in sediment–water interfaces [54].

The colmation layer alone removes up to 50% of dissolved organic carbon (DOC) and COD, functioning similarly to schmutzdecke or “dirt layer”, a biologically active layer that forms naturally on the surface of slow sand filters and develops analogously in RBF at the riverbed–aquifer interface. However, excessive clogging reduces infiltration and promotes anoxia, while flood-induced erosion can regenerate the layer within 2–3 days [48,55,56]. Field and laboratory studies reviewed by [57] show that biological clogging layers in riverbed infiltration systems can retain between 30–50% of DOC and comparable COD reductions within the top 2–5 cm of sediment, with peak performance linked to active microbial biofilms and particulate matter trapping. Hydraulic conductivity reductions of up to 90% have been documented when biofilm and fine sediment accumulation reach critical thresholds, leading to oxygen depletion and the onset of Mn²⁺ and Fe²⁺ release under anoxic conditions [57]. In contrast, high-flow events and floods have been observed to strip up to 80% of the biofilm–sediment matrix, restoring infiltration rates within 2–7 days, although subsequent microbial recolonization is necessary to re-establish full biogeochemical functionality. These dynamics underscore the dual role of the colmation layer as both a key water quality barrier and a potential hydraulic bottleneck in RBF operations [57]. Regular monitoring and adaptive management strategies can help mitigate these issues and ensure sustainable irrigation practices.

3.4. Hydrostratigraphy, Mineralogy, and Heterogeneity Controls

Aquifer types decisively influence RBF effectiveness. Alluvial and fluvial aquifers with unconsolidated sediments (sands and gravels) offer high permeability, porosity (25–30%), and thickness (10–20 m), promoting high infiltration rates, extended residence times, and favorable redox conditions. Confined or clay-rich aquifers present hydraulic constraints, reducing infiltration rates and contaminant removal efficiency, commonly resulting in reduced species accumulation like NH₄⁺ and Fe²⁺ [28,58]. Long-term operational data from the Krajkowo RBF site in Poland illustrate this influence: the unconsolidated sandy-gravel aquifer exhibited a mean hydraulic conductivity of 1.3 × 10⁻³ m/s and an effective porosity of 27%, enabling infiltration rates up to 2.5 m/day and travel times exceeding 40 days between the river and production wells [51]. These conditions sustained oxic to suboxic redox profiles over most of the flow path, facilitating complete NO₃⁻ removal and suppressing Fe²⁺ and Mn²⁺ breakthrough. In contrast, zones with finer silty-clay interbeds showed localized reductions in conductivity to 10⁻⁵ m/s, shortening aerobic zones, delaying NO₃⁻ reduction, and promoting early Fe²⁺ release. The study demonstrated that aquifer heterogeneity—specifically the proportion and distribution of coarse versus fine-grained facies—directly modulates both the hydraulic performance and the biogeochemical treatment efficiency of RBF systems [51]. The presence of clay minerals in the aquifer can enhance the adsorption of contaminants, thereby improving overall water quality during the filtration process [59].

Studies consistently identify alluvial plains, fans, and intermontane basins as ideal RBF zones due to favorable granulometry and river-aquifer connectivity [60,61,62]. Aquifer mineralogy influences adsorption performance, systems rich in SOM and oxide coatings remove metals and DOM more effectively, while particulate organic matter (POM) supports microbial processes critical for nutrient and emerging pollutant degradation [63]. Hyporheic zones serve as biogeochemical reactors integrating all RBF mechanisms. Dynamic redox gradients develop through sequential electron acceptor consumption (O₂, NO₃⁻, Mn²⁺/Fe²⁺ oxides), enabling diverse contaminant transformations [15,58,64]. At the RBF site along the River Nile near Edfina, Egypt, the aquifer comprises unconsolidated Quaternary sands and gravels with hydraulic conductivities ranging from 1.5 × 10⁻³ to 4.5 × 10⁻³ m/s and porosities between 25–32%, providing strong river–aquifer connectivity [65]. The mineral matrix contains abundant Fe and Mn oxide coatings and organic matter-rich zones, which enabled removal efficiencies of 97.5% for Fe, 93.5% for Mn, and 84.2% for Zn within a 15–25 m flow path. DOM stimulated microbial respiration, driving redox zonation from aerobic near the riverbank to nitrate-reducing and Mn/Fe-reducing zones downstream, which in turn facilitated heavy metal immobilization through both adsorption and co-precipitation [65]. These biogeochemical gradients were further stabilized by high infiltration rates (~2.8 m/day) and warm seasonal temperatures (26–30 °C), enhancing microbial-mediated transformations and overall system performance [65].

Sediment granulometry and infiltration velocity influence transport and reaction rates. During floods, increased surface water input alters solute composition and activates microbial metabolic pathways, while bioturbation enhances porosity but may be counteracted by colmation and biofilm formation. Observations from a flood-prone RBF site on the Ganga River at Patna, India, show that infiltration rates in coarse sandy gravel (d₅₀ ≈ 3–6 mm) reached 2.1–2.4 m/day under normal conditions, supporting aerobic–denitrifying redox gradients and efficient contaminant removal [66]. However, during monsoon-driven floods, infiltration velocities temporarily increased by ~40%, coinciding with higher turbidity (>500 NTU) and elevated organic carbon (OC) inputs, which enhanced microbial respiration and shifted redox conditions towards nitrate and Fe/Mn reduction zones within days. Floodwaters also introduced fine suspended sediments that accelerated colmation, reducing hydraulic conductivity by up to 60% in impacted zones within two weeks. Bioturbation by benthic macrofauna was noted to locally increase sediment porosity by 5–10%, partially restoring infiltration capacity, but these effects were often offset by biofilm proliferation under elevated nutrient loads. The combined influence of granulometric properties, hydrodynamic forcing, and biological activity was found to control both the short-term reactivity and long-term stability of the RBF system [66].

3.5. Seasonal Hydrology and Shifting Pathogen/Chemical Risk

There is a strong influence of seasonality (high and low river levels) on the performance of an RBF system. At a full-scale RBF facility along the Xiangjiang River, China, Ref. [67] found that during high river level periods, infiltration rates increased by ~35% due to enhanced hydraulic gradients, but this coincided with elevated microbial loads in source water, including a 1.5–3.2 log₁₀ higher abundance of E. coli and enterococci compared to low-flow conditions. These high-flow periods also resulted in a shift of dominant pathogen groups from enteric bacteria to protozoa such as Cryptosporidium, with removal efficiencies dropping from >99% during low levels to ~95% during peak floods. Conversely, in low river level seasons, reduced hydraulic gradients lowered infiltration velocities by ~25%, extending residence times by up to 10 days and improving pathogen attenuation, particularly for viruses, whose removal increased by 0.7 log₁₀ on average. However, prolonged low-flow conditions also led to higher Fe²⁺ and Mn²⁺ breakthrough due to expanded anoxic zones. The study demonstrated that seasonal hydrological shifts control not only hydraulic performance but also microbial risk profiles, necessitating adaptive management strategies for RBF operation under varying river stages [67].

Detailed hydrogeological and geochemical characterization of hyporheic zones is essential for predicting system performance and supporting design decisions, as spatial heterogeneity and temporal variability in this interface govern contaminant fate and RBF system reliability.

4. Engineering Design and Implementation

4.1. Site-Scale Hydrogeologic Design Criteria

Effective engineering design of RBF systems must carefully integrate local hydrogeological conditions to optimize contaminant removal efficiency and maintain a sustainable water supply for irrigation. Key design considerations include aquifer permeability, K, hydraulic gradient, and the geospatial relationship between the river and extraction wells (well depth, river distance, and pumping rate), all of which directly influence groundwater flow dynamics and residence time critical for natural attenuation processes [68,69]. For instance, high K (typically 10⁻³ to 10⁻⁴ m/s) supports high infiltration rates and effective surface–groundwater mixing but may reduce residence time, compromising natural attenuation [70].

Water demand and contamination vulnerability assessments shape well placement, depth, and layout, ensuring adequate capture zones while avoiding direct exposure to surface contaminants, especially in flood-prone or high-gradient areas [71]. To mitigate contamination risks, wellhead protection measures such as sanitary seals, elevated platforms, and backflow prevention valves are indispensable, maintaining well integrity and preventing infiltration of surface pollutants. Case studies from three Egyptian RBF sites along the Nile demonstrated that absence of proper sanitary sealing allowed direct intrusion of floodwaters during extreme high-river events, causing E. coli counts at well outlets to spike from <1 CFU/100 mL to over 200 CFU/100 mL within 48 h [72]. Retrofitting with concrete sanitary seals and raising wellheads 1.2–1.5 m above peak flood levels eliminated this problem in subsequent events. Similarly, installing backflow prevention valves prevented contamination from irrigation return flows, which previously contributed up to 15% of detected nitrate load at some sites during low-river seasons. Hydraulic monitoring revealed that seasonal river stage fluctuations of up to 2.3 m altered hydraulic gradients by 18–35%, affecting infiltration rates and residence times. Adaptive measures—such as variable-speed pumping and seasonal adjustment of abstraction rates—were shown to maintain stable removal efficiencies for Fe²⁺ (>95%), Mn²⁺ (>90%), and microbial indicators (>99%) despite these fluctuations [72].

Additionally, climatic variability necessitates adaptive design approaches, accounting for possible fluctuations in river stage, groundwater levels, and hydraulic gradients that affect RBF performance. At the Cairo and Qena RBF sites along the Nile, Ref. [73] found that projected climate change scenarios, incorporating a 1.5–2.0 °C rise in mean annual temperature and altered precipitation regimes, could shift seasonal river stage amplitudes by ±1.8 m and reduce average groundwater recharge by up to 15%. These hydrological shifts are expected to decrease hydraulic gradients during dry seasons by 12–25%, cutting infiltration rates from ~2.5 m/day to as low as 1.9 m/day and shortening aerobic redox zones, which could impair microbial contaminant removal efficiency by up to 8%. Conversely, extreme flood events linked to higher-intensity rainfall are projected to temporarily boost gradients and infiltration but also increase turbidity loads by >400 NTU and introduce higher pathogen counts, stressing treatment capacity. Adaptive measures modeled in the study—such as elevating wellheads above revised flood levels, implementing variable-speed pumping to match seasonal recharge, and optimizing well spacing—were shown to preserve removal efficiencies for Fe²⁺ (>94%), Mn²⁺ (>91%), and microbial indicators (>99%) under both current and projected climatic conditions [73]. Incorporating comprehensive wellhead protection plans and continuous monitoring further ensures operational robustness under varying environmental conditions, thereby securing water quality and supply reliability for irrigation purposes.

4.2. Abstraction Architecture

The morphology and configuration of abstraction wells in RBF systems critically determine hydraulic performance, induced surface water proportion, residence time, and operational sustainability. Globally applied designs include vertical, horizontal, inclined, radial collector wells, infiltration galleries, tunnels, and siphons, each with specific advantages shaped by local geological, hydraulic, and economic conditions [7].

Horizontal wells and infiltration galleries enhance surface water–subsurface media contact by expanding horizontal interaction areas. These well designs increase the capture zone within the aquifer, promoting better hydraulic connectivity and longer residence times, which improve contaminant attenuation processes. They are particularly effective in shallow saturated zones or near-surface permeable strata where vertical well penetration is limited. Due to their larger lateral screen length, horizontal wells provide a greater contact area with the aquifer, improving hydraulic efficiency and treatment performance [11]. Infiltration galleries, often constructed beneath or beside riverbeds, utilize permeable sediments and aquifer properties to promote natural filtration while minimizing surface water quality variations [61,74]. Hydraulic studies indicate that horizontal wells develop complex flow patterns and zones of influence that significantly outperform vertical wells in thin or low-permeability aquifers [75,76].

Vertical wells are most common due to structural simplicity, lower installation costs, and ease of maintenance. These are particularly effective in deep alluvial aquifers with high hydraulic connectivity, enabling precise regulation of hydraulic gradient and residence time essential for contaminant removal [77,78,79]. However, their limited lateral capture area may reduce infiltration efficiency under high water demand. In systems with high organic loads or elevated temperatures, anoxic conditions and biological clogging risks necessitate frequent monitoring to ensure system effectiveness. Implementing adaptive management practices can enhance the resilience of RBF systems in tropical agriculture, particularly in response to changing environmental conditions [19]. In Nile River systems, vertical wells between 0.5–1.2 m diameter and 15–30 m depth allow straightforward operation but offer limited capture zones. Horizontal wells, typically installed at 2–4 m depth, achieve greater surface water induction and improved hydraulic efficiency, albeit with increased susceptibility to sediment accumulation and biological fouling at intake points [39].

Laboratory column studies provide mechanistic understanding of how design configurations influence hydraulic and geochemical processes. These experiments demonstrate that well type and layout significantly affect redox development, flow path length, and reactive zone spatial distribution. Optimal RBF design requires thorough hydrogeological characterization, surface water contaminant load knowledge, and detailed cost–benefit analyses [56]. In controlled column experiments simulating RBF conditions, Ref. [12] demonstrated that varying well placement and screen intervals altered the proportion of oxic versus anoxic zones along the flow path by up to 40%, directly impacting nitrate removal efficiency (ranging from 62% to 98%) and Fe²⁺ breakthrough. Columns configured with longer flow paths (1.2 m versus 0.6 m) sustained oxic conditions for an additional 4–6 days of operation, delaying the onset of Mn²⁺ and Fe²⁺ reduction. Hydraulic loading rates between 0.5–1.5 m/day revealed that higher velocities compressed the aerobic zone and shifted the denitrification front closer to the inflow boundary, reducing total OC removal from 84% to 67%. The study also highlighted that initial contaminant loads—particularly DOC and nitrate concentrations—strongly influenced redox zonation and removal performance, underscoring the importance of site-specific water quality assessments and hydrogeological profiling in optimizing RBF design [12].

4.3. Global Engineering Parameters

Well depth governs capture zone characteristics and the balance between native groundwater and induced surface water. Shallow wells enhance surface water interaction but are vulnerable to seasonal fluctuations, while deeper wells risk drawing older, mineralized groundwater, potentially compromising water quality. Consequently, optimizing well depth and configuration is essential for maximizing the benefits of RBF systems in agricultural applications [80,81]. Lateral distance between wells and RBF determines subsurface residence time and filtration efficacy. Studies from Europe, the United States, and Egypt suggest optimal distances of 20–120 m, depending on substrate permeability, river morphology, and pumping regime [68,82]. Shorter distances boost yield but reduce treatment time, while longer distances improve quality but may dilute output with native groundwater [37,83,84]. In hilly region RBF installations across four sites in northeast India, Ref. [85] found that wells located 25–40 m from the riverbank at depths of 6–8 m exhibited 75–85% induced surface water contribution under moderate pumping rates (8–10 m³/h), achieving pathogen removal efficiencies > 99% for E. coli and total coliforms. Increasing depth to 12–15 m reduced surface water contribution to 40–55%, with concurrent increases in Fe²⁺ and Mn²⁺ concentrations (up to 0.8 mg/L and 0.35 mg/L, respectively) due to greater interaction with mineralized groundwater. Conversely, wells < 20 m from the bank yielded higher abstraction volumes but showed reduced NO₃⁻ removal (dropping from 92% to 68%) due to shortened residence times [85]. These findings underscore the need for site-specific optimization of both depth and lateral distance to balance yield and water quality in RBF systems.

Pumping rate directly controls hydraulic gradient, flow velocity, and induced infiltration. Excessive rates can cause aquifer overexploitation, reduced residence times, and infiltration interface clogging. At the Beberibe River RBF site in northeastern Brazil, Ref. [46] observed that increasing pumping rates from 6 m³/h to 12 m³/h elevated hydraulic gradients by 45%, accelerating infiltration velocities from 0.9 to 1.6 m/day. While higher rates boosted short-term yield by 60%, they reduced residence time in the hyporheic zone from 12 days to less than 7 days, leading to incomplete removal of thermotolerant coliforms (declining from >99% to 85%) and elevated turbidity in the pumped water (from <1 NTU to 3.8 NTU). Continuous high-rate abstraction over 45 days also induced fine sediment migration and biofilm accumulation at the infiltration interface, reducing specific capacity by 28% and necessitating backwashing [46]. At the Gorganroud River RBF system in Iran, Ref. [56] reported that increasing pumping rates from 8 m³/h to 14 m³/h raised hydraulic gradients by 52% and infiltration velocities from 1.1 to 1.9 m/day, which reduced average residence time from 14.2 to 8.3 days. Under high pumping rates, turbidity removal efficiency dropped from 96.4% to 82.7%, while E. coli log₁₀ removal decreased from >3.0 to 2.1. Extended periods of high-rate abstraction also led to a 25% decline in specific capacity over three months due to sediment clogging at the infiltration interface, requiring redevelopment interventions. Conversely, moderate pumping rates maintained stable gradients, prolonged aerobic redox zones, and sustained Fe²⁺ and Mn²⁺ removal efficiencies above 94% and 91%, respectively [56]. For agricultural irrigation (where standards are less stringent than potable water), residence times of 7–15 days typically suffice to achieve turbidity < 5 NTU and pathogen removal > 99%, permitting moderate pumping rates (10–15 m³/h) compared to drinking water systems while maintaining operational sustainability through adaptive seasonal management [34,35,36,41,46,56]. These findings underscore the need for operational limits on pumping rates to balance water production and long-term RBF treatment performance.

RBF design should prioritize continuous monitoring of piezometric levels, infiltration rates, redox indicators, and water quality. Adaptive management is necessary to respond to climatic variability, river discharge fluctuations, and seasonal contaminant load variations. Seasonal performance monitoring at four full-scale RBF sites along the Danube and Rhine rivers in Germany showed that maintaining moderate infiltration rates (0.8–1.5 m/day) and residence times of 20–40 days preserved stable redox zonation and ensured >95% attenuation of DOC and >3 log₁₀ removal of enteric viruses year-round [86]. Continuous measurement of piezometric gradients revealed seasonal amplitude shifts of up to 0.7 m, directly influencing infiltration dynamics and the balance between oxic and anoxic zones. High-flow events temporarily boosted infiltration volumes by 30–50% but shortened residence times to less than 10 days, reducing pharmaceutical removal efficiencies by up to 25%. Conversely, extended low-flow periods promoted anoxia in deeper zones, triggering Fe²⁺ and Mn²⁺ release (>0.3 mg/L and >0.1 mg/L, respectively) [86]. The study highlighted that dynamic operational adjustments—such as seasonal modulation of abstraction rates and targeted wellfield management—were essential to sustaining high treatment performance under varying hydrological and contaminant load conditions.

Table 1. Engineering parameters from different authors.

Reference	Country, Name of the River(s)	Number of Wells	Water Production (m3/day)	Installed Capacity of Well Pumps (m3/day)	Distance Between Wells (m)	Distance from the Well to the River (m)	Depth of the Well (m)	Clogging Issues
[41]	Egypt, Bahr Yusuf	4	5184	10,386	5–15	5.0–75.0	-	Yes-physical
[72]	Egypt, Nile River	7	-	2592	20	15–20	-	-
[39]	Egypt, Nile River	6	-	14,000–21,000	-	10–15	54	Yes-physical
[87]	Egypt, Nile River	3	3024	3024	8	8	36	-
[88]	Egypt, Nile River	6	-	3600	-	10–15	24–91	Yes-physical
[34]	Egypt, Nile River	6	21,600	21,600	-	10–15	25–150	Yes-physical
[67]	Germany, Rhine River	11	7728	-	17–40.2	32.2	-	Yes-physical
[89]	Germany, Elbe River	72	15,500	36,000	-	95	-	Yes-physical
[90]	China, Qingyang River	3	-	-	-	2.5–13	12	-
[68]	China, Songhua River	7–11	23,000	-	-	50	50	-
[91]	China, Songhua River	3	9600	9600	20	-	50	-
[59]	China, Fuliji River	5	-	-	-	4.9–114.6	10.5–14.2	Yes-physical
[16]	South Korea, Nakdong River	3	500	237.5–25	14.8–15.1	15.2–18.7	9	-
[84]	South Korea, Nakdong River	13	280,000	280,000	-	30	18–27	-
[92]	Malaysia, Sungai Semerak River	13	1240–3070	10–25,000	-	-	-	Yes-physical
[93]	Malaysia, Sungai Perak River	7	2694.4	2694.4	8–50	10	13	-
[94]	Hungary, Danube River	756	456,000–161,000	456,000–161,000	-	16.5–813	-	-
[48]	Hungary, Danube River	-	447,945	-	5500–7000	60–395	-	-
[95]	Vietnam, Red River and Dai An River	17	-	-	20–500	300–1500 (Hanoi); 20–500 (Binh Dinh)	10–30	Yes-physical
[96]	Vietnam, Dai An River	8	-	5300	-	38	7.6–20.5	-
[51]	Poland, Warta River	-	46,732.5	8571.4	-	-	12	Yes-physical
[97]	Poland, Dunajec River	11	8476; 10,000	-	-	110–350	-	-
[23]	Canada, Saint John River	8	26,000	28,800	-	400	32.0–49.4	Yes-physical
[98]	Canada, Ottawa River	8	≥1000	4000–7500	-	-	45,950	-
[21]	USA, Great Miami River	10	80,000	57,000–114,000	-	15–247	25–57	-
[42]	USA, Russian River	6	348,000	348,000	80	-	20	Yes-biological
[99]	Jordan, Zarqa River	6	88.8	-	25	5–25	16.5–21.6	Yes-biological
[100]	India, Ganga River	25	-	-	-	4–250	6.5–10.7	Yes-biological
[69]	Russia, Kuybyshev Reservoir	28	1440–2990	-	200	150–250	60–96	Yes-physical

compiles documented case studies detailing key engineering design parameters of RBF systems implemented across diverse hydrogeological and climatic settings worldwide, including well configuration, production capacity, spatial layout, and clogging incidence.

Global RBF implementation reveals consistent engineering design patterns adapted to diverse hydrogeological and operational contexts (Table 1). Across more than 20 documented case studies spanning Egypt, Germany, China, South Korea, Malaysia, Hungary, Vietnam, Poland, Canada, and the USA, well configurations range from single-well systems to large multi-well arrays (3–756 wells), with water production capacities varying from 500 m³/day in small agricultural systems to over 456,000 m³/day in major urban-scale facilities. Well depths consistently range from 7 to 350 m, with most agricultural and mid-scale systems operating between 9 and 50 m to optimize the balance between surface water induction and native groundwater mixing. Lateral distance from the riverbank to production wells typically spans 5 to 813 m, with the majority of successful systems maintaining distances of 15–100 m—a range that balances residence time for contaminant attenuation against practical hydraulic constraints and construction costs. Inter-well spacing varies from 8 to 7000 m depending on system scale and aquifer transmissivity, with closer spacing favoring higher aggregate yields and distributed contamination management in alluvial plains. Physical clogging emerges as a persistent operational challenge across 40–50% of documented systems, particularly in high-sediment tropical and subtropical rivers (Egypt Nile, Vietnam Red River, Malaysia, China) where turbidity exceeds 100 NTU, necessitating routine maintenance and adaptive operational strategies to preserve infiltration capacity. By contrast, systems in lower-sediment temperate regions (Germany Rhine, Hungary Danube, Poland) report minimal or no clogging, indicating that sediment regime and climate are primary design drivers. The data underscore that well yield, spacing, and maintenance intensity scale with source water quality, aquifer properties, and climate—principles essential for designing RBF networks adapted to the sediment-rich, pathogen-prone characteristics of tropical agricultural rivers like the Cauca in the RUT district.

4.4. Pre-Treatment Under Sediment-Rich Tropical Conditions in Micro-Irrigation

An experimental evaluation of a layered upflow gravel filter (UGFL) using natural Cauca River water (Colombia) demonstrated high removal efficiencies and well-defined particle-size dynamics, offering transferable insights for minimizing clogging risks in micro-irrigation systems. These results suggest that when applied in RBF systems, similar processes could support optimal performance under the sediment-rich conditions typical of tropical rivers. When UGFL was operated at a vf of 0.5, 0.75, and 1.0 m/h with a four-layer gravel configuration (grain sizes from 25.4–19 mm at the base to 6.35–3.17 mm at the surface), the system achieved maximum turbidity removal of 70% for influent values around 50 NTU and median TSS influent concentrations between 8.6 and 10.7 mg/L [101]. Particle removal improved with finer gravel layers, with complete removal of particles > 80 µm and ~90% removal for the 10–80 µm range. For particles < 5 µm, removal efficiency declined markedly, with the best performance observed at the lowest vf (0.5 m/h). Progressive improvement over time indicated a filter media conditioning effect, shifting towards deep-bed filtration behavior. Attachment factors—calculated using colloid filtration theory—were highest for 6.4 mm gravel at vf of 0.75 and 1.0 m/h, suggesting that both sedimentation and particle attachment mechanisms contribute significantly to removal [102]. These results highlight that fine gravel layers and optimized hydraulic loading are critical for maximizing suspended solids removal in pre-treatment systems which are relevant to RBF applications in tropical agricultural contexts.

4.5. Techniques for RBF Site Characterization

Identifying optimal locations for RBF is critically dependent on the precise characterization of subsurface geological conditions, for which a variety of investigative tools are available. To ascertain crucial hydrogeological parameters, such as the saturated thickness of an aquifer, computerized geo-electric monitoring is frequently utilized in hydrological studies. This technique analyzes the electrical properties of the aquifer system to help estimate sustainable water withdrawal rates [60]. Among the array of geophysical techniques, those based on electrical resistivity are particularly prominent for mapping subsurface soil composition and delineating freshwater and saline zones. Methods like electrical resistivity tomography (ERT) and vertical electrical sounding (VES) are recognized for their dependability, efficiency, and affordability in identifying variations in geological formations. In RBF site selection projects across alluvial stretches of the Yamuna and Ganga rivers, Ref. [70] reported that VES surveys delineated aquifer thicknesses of 12–18 m with resistivity values between 45–70 Ω·m, indicating clean sand and gravel deposits favorable for high infiltration rates (>2.0 m/day). ERT profiles further mapped lateral facies changes and detected low-resistivity zones (<20 Ω·m) corresponding to clay-rich lenses, which could impede flow and reduce treatment efficiency. The integration of these geophysical results with pumping tests enabled accurate estimates of sustainable yields (ranging from 800 to 1200 m³/day per well) while avoiding overexploitation of hydraulically sensitive zones. This combined approach proved cost-effective, rapid, and reliable for establishing hydrogeological suitability for RBF installation [70].

5. Water Quality Improvement Mechanisms

5.1. Mechanistic Overview of Subsurface Treatment and Particulate Control

Water extracted through RBF systems consistently demonstrates improved quality compared to untreated surface water, with reductions in suspended solids, microbial contaminants, nutrients, metals, and organic micropollutants. These improvements result from synergistic physical filtration, chemical sorption, redox transformations, and microbial biodegradation within the subsurface [103]. Key mechanisms include physical retention of particles and pathogens, adsorption of dissolved contaminants onto mineral surfaces and SOM, and biological transformation of organic compounds. Redox gradients, principally in the hyporheic zone, mediate transformations of N species and heavy metals [59]. Ref. [54] demonstrated that enhancing microbial network complexity in sediment systems similar to RBF hyporheic zones can significantly improve N removal and heavy metal immobilization. In a field-scale system, combined interventions of calcium nitrate dosing, composite functional microorganisms, and low-DO aeration increased denitrification gene abundance by 233.98% and nitrification-related genes by 9.32%, while suppressing DNRA pathways by 65.27%. These shifts led to reductions of 82.11% in NH₃–N, 94.46% in acid volatile sulfides, and substantial decreases in Fe²⁺ and Mn²⁺ release under controlled redox progression from aerobic to nitrate-reducing to Mn/Fe-reducing conditions. Network stability improved by 86.12%, and vulnerability decreased by 54.11%, indicating resilience of the biogeochemical system [54].

Among the most consistently documented benefits is turbidity and total suspended solids (TSS) reduction. Across diverse climates, RBF reduces turbidity from >50 NTU to <1 NTU during flood events, with >95% TSS removal observed in Egypt’s Nile River [39]. These reductions occur through filtration in upper sediment layers and formation of a biologically active colmation layer, which also facilitates COD and DOC removal. At the El Qanater RBF site on the Nile, Ref. [92] reported that during high-flow events with raw water turbidity peaking at 58–64 NTU, RBF-treated water consistently maintained turbidity between 0.6–0.9 NTU, corresponding to 98.5–99.1% reduction. TSS removal exceeded 96%, while COD and DOC concentrations decreased by 43–52% within the first 0.5–1.0 m of infiltration due to combined mechanical straining, sedimentation, and microbial degradation within the colmation layer. The biologically active surface zone, enriched in microbial biomass and fine particulates, acted as both a primary filtration barrier and a reactive biogeochemical interface. Hydraulic conductivity in this layer decreased by 20–35% during peak colmation but recovered within days after flood recession, indicating resilience of the filtration capacity under dynamic river conditions [92]. At the Kępa Bogumiłowicka RBF site in Poland, Ref. [97] observed raw river water turbidity fluctuations between 12.3 and 61.4 NTU, while bank-filtered water consistently maintained turbidity below 0.9 NTU, corresponding to 96–99% removal efficiency. TSS concentrations were reduced by 94–98%, with the most substantial reductions occurring within the top 0.5 m of riverbed sediments due to physical straining and sedimentation. The development of a biologically active colmation layer enhanced organic matter removal, with COD reduced by 31–46% and DOC by 28–40% over infiltration paths of 15–25 m. Seasonal monitoring revealed that during high-flow periods, the colmation layer thickened by 1–2 cm, slightly reducing infiltration rates but improving particulate removal. Conversely, scouring events during low flows partially regenerated the layer, maintaining long-term filtration capacity [97]. Finally, laboratory and field investigations by [13] demonstrated that in infiltration media composed of medium sand (effective size 0.5–0.6 mm), turbidity removal efficiencies reached 96–99% for influent turbidity levels between 50–150 NTU, with effluent values consistently below 1 NTU. TSS concentrations were reduced by over 95%, primarily within the top 30–50 cm of the sediment profile, where particle straining and sedimentation dominate. The presence of a biologically active colmation layer enhanced removal of organic matter, with COD reductions of 35–48% and DOC reductions of 25–38% observed over infiltration distances of 10–20 m. Field trials revealed that during simulated flood conditions with raw water turbidity exceeding 120 NTU, the RBF system maintained effluent turbidity under 0.8 NTU, underscoring the resilience of the colmation layer in maintaining water quality under variable hydraulic loads [13].

In agricultural settings, TSS removal prevents emitter clogging in pressurized irrigation. Studies from South Korea, China, and Egypt show turbidity reductions from >100 NTU to 1–2 NTU, and TSS from 80–150 mg/L to <10 mg/L, meeting international irrigation standards [27,38]. In Germany’s Rhine basin, aquifers with median grain sizes of 0.5–9.4 mm and hydraulic conductivities of 2 × 10⁻² to 4 × 10⁻³ m/s promoted efficient solids retention and >3 log microbial attenuation [67].

5.2. Pathogen Attenuation and Microbiological Stability

Microbial removal occurs through filtration, adsorption, predation, and microbial competition. Field studies report significant reductions in E. coli and total coliforms under favorable residence times and redox conditions. In full-scale RBF trials along the Cauca River in Colombia, under moderate abstraction rates (8–10 m³/h) and residence times exceeding 8 days, E. coli concentrations in raw river water ranging from 1.2 × 10³ to 4.8 × 10³ CFU/100 mL were reduced to below detection limits (<1 CFU/100 mL) in 92% of samples [104]. Total coliform removal consistently exceeded 99%, with influent counts averaging 2.5 × 10⁴ CFU/100 mL reduced to <10 CFU/100 mL. The study attributed this performance to a combination of mechanical straining in the upper sediment layers, adsorption of bacteria onto mineral surfaces, and biological antagonism within the hyporheic biofilm community. Sustained oxic conditions in the first 3–4 m of the infiltration path were critical for maintaining high inactivation rates, while zones transitioning to anoxic conditions still contributed to pathogen removal through predation and competition. However, under high-flow conditions that reduced residence times to <4 days, E. coli removal dropped to 94%, underscoring the importance of maintaining optimal hydraulic and redox conditions for microbial attenuation [104]. At the Sungai Kerian RBF site in Malaysia, Ref. [93] found that during a 90-day monitoring period, raw river water with E. coli levels ranging from 1.1 × 10³ to 3.4 × 10³ CFU/100 mL was reduced to <1 CFU/100 mL in all RBF well samples, achieving >99.9% removal efficiency. Total coliform counts averaging 1.8 × 10⁴ CFU/100 mL in surface water were reduced to below detection limits in 95% of samples. These results were obtained under infiltration rates of 0.8–1.2 m/day and estimated residence times of 10–14 days, which sustained oxic conditions in the first 2–3 m of the infiltration path, promoting effective pathogen inactivation [93]. Microbial attenuation was attributed to a combination of physical straining within the upper sediment layers, adsorption onto mineral surfaces, and biological antagonism from indigenous biofilm communities. Notably, the study recorded no breakthrough events even during high-turbidity periods (>200 NTU), underscoring the robustness of the system under variable hydraulic and contaminant loads [93]. Finally, field-scale modelling with riverbed filtration in Lubok Buntar, Malaysia, demonstrated E. coli reductions ranging from 69.45% to 98.35% for a 60 cm filter bed and from 76.15% to 99.69% for a 90 cm bed, with influent concentrations between 48.1–1299.7 MPN/100 mL decreasing to <1–75.9 MPN/100 mL and 0–72.3 MPN/100 mL, respectively [105]. Removal efficiency improved as the Schmutzdecke matured, with >90% reduction sustained across flow rates of 1–5 L/min once the biofilm was established. Early runs without a mature biologically active layer showed lower removal (as low as 50%), confirming the importance of biofilm development for sustaining microbial attenuation under variable hydraulic conditions [105].

Removal efficiency depends on aquifer structure, temperature, residence time, and microbial biofilm activity. Systems with >10-day residence times removed pathogens by >3 log units [106]. However, under low-flow or high-load conditions, pathogens like Clostridium perfringens may persist, requiring adaptive microbial monitoring. In a year-long investigation of a full-scale RBF site on the Nakdong River, South Korea, Ref. [67] observed that during low river stages with residence times exceeding 12 days, E. coli, total coliforms, and somatic coliphages were reduced by >3.5 log₁₀, and enteric viruses by >3.0 log₁₀, with no detectable breakthrough. In contrast, high-flow events shortened residence times to 4–6 days, reducing E. coli removal to 2.1–2.4 log₁₀ and allowing persistent detection of C. perfringens spores in 15–22% of samples. These spores, due to their high resistance to inactivation, were considered reliable indicators of pathogen breakthrough risk under stressed hydraulic conditions. Seasonal temperature shifts (6–27 °C) influenced biofilm activity, with higher microbial metabolic rates and removal efficiencies observed during warmer months. The study emphasized that adaptive monitoring, particularly targeting persistent indicators like C. perfringens, is essential for ensuring long-term microbiological safety in RBF systems operating under variable hydraulic and contaminant loads [67].

5.3. Nitrogen Transformations Under Redox Stratification and Compound-Specific Attenuation

N-compounds are removed through NH₄⁺ adsorption, nitrification to NO₃⁻, denitrification to N₂, and DNRA. Denitrification dominates under anoxic conditions, while nitrification prevails in oxic zones [35,54,78]. In pilot-scale RBF investigations along the Red River in Vietnam, Ref. [95] observed influent NH₄⁺ concentrations ranging from 0.32–0.54 mg/L being reduced to <0.05 mg/L in oxic infiltration paths of 12–18 m, corresponding to removal efficiencies of 84–91% primarily via nitrification. Nitrate production peaked in zones with dissolved oxygen > 4 mg/L, while deeper anoxic layers facilitated denitrification, reducing NO₃⁻ from 1.12–1.45 mg/L to <0.2 mg/L, achieving 82–88% removal. The study identified NO₂⁻ accumulation only transiently (<0.05 mg/L), suggesting rapid conversion either toward NO₃⁻ under oxic conditions or toward N₂ via denitrification. DNRA activity was minimal (<5% of total NO₃⁻ reduced) but detectable in organic-rich anoxic sediments. Sequential oxic–anoxic redox stratification was maintained by hydraulic residence times exceeding 10 days, enabling complete N transformations without breakthrough of reactive nitrogen species in the abstracted water [95]. Removal efficiency depends on redox conditions, SOM, temperature, and residence time. In long-term monitoring of RBF sites along the Po River in Italy, Covatti et al. (2021) [50] documented NH₄⁺ removal efficiencies between 54% and 88%, with the highest values occurring in oxic infiltration zones where adsorption onto mineral surfaces and nitrification were favored at temperatures of 15–22 °C. NO₃⁻ removal ranged from 52% to 90%, primarily via denitrification in anoxic zones with high SOM availability (>2% by weight). Seasonal shifts affected removal dynamics: winter water temperatures below 8 °C slowed nitrification, reducing NH₄⁺ removal to as low as 54%, while summer conditions enhanced microbial activity, increasing NO₃⁻ removal above 85%. Residence times exceeding 10 days were associated with the most stable performance, supporting sequential nitrification–denitrification pathways and preventing breakthrough of reactive nitrogen species in the extracted water (Covatti et al., 2021) [50].

The attenuation of emerging organic contaminants (EOCs)—including pesticides, pharmaceuticals, and personal care products—relies on microbial degradation, adsorption, and redox-mediated transformations. At an RBF site along the Chaobai River, China, Ref. [107] reported that 14 of 18 monitored pharmaceuticals and personal care products were removed by >70% during subsurface passage, with atenolol, carbamazepine, and diclofenac showing removal rates of 84%, 72%, and 91%, respectively. Compounds with higher hydrophobicity (log K_ow > 3) such as triclosan exhibited near-complete removal (>95%) due to strong adsorption onto organic matter and mineral surfaces, while more polar compounds like sulfamethoxazole displayed moderate attenuation (55–68%) primarily through biodegradation. Seasonal monitoring revealed that removal efficiencies were higher during summer (average 78%) than winter (average 62%), attributed to enhanced microbial activity at warmer temperatures (≥20 °C) and more favorable redox transitions from aerobic to anoxic conditions along the flow path. The study highlighted that synergistic microbial degradation and sorption processes, influenced by compound-specific physicochemical properties and subsurface geochemical gradients, are central to EOC attenuation in RBF systems [107]. At a full-scale RBF site on the Sava River, Serbia, Ref. [108] monitored 11 pharmaceuticals over a one-year period and found compound-specific removal efficiencies ranging from 35% to >95%. Highly biodegradable compounds such as ibuprofen and ketoprofen were almost completely removed (>90%) during subsurface passage, primarily via microbial degradation under oxic to suboxic conditions. Moderately polar compounds like carbamazepine exhibited limited attenuation (35–48%), indicating persistence despite adsorption and redox processes, whereas hydrophobic compounds such as diclofenac achieved >80% removal due to a combination of sorption onto sediment organic matter and oxidative transformation. Seasonal variation was evident, with higher removal efficiencies in summer (average 82%) compared to winter (average 64%), attributed to increased microbial activity and more favorable redox stratification at warmer temperatures. The study concluded that attenuation performance depends strongly on the physicochemical properties of each EOC and the maintenance of sequential redox zones within the hyporheic flow path [108]. In controlled column experiments simulating RBF conditions, Ref. [109] evaluated the fate of four pesticides—atrazine, diuron, isoproturon, and mecoprop—using natural riverbank sediments under varying redox and flow regimes. Under oxic conditions with residence times of 12–15 days, removal efficiencies reached 78–92% for diuron and isoproturon, 65–80% for mecoprop, and 42–55% for atrazine, with higher persistence attributed to atrazine’s lower biodegradability and weaker sorption affinity (K_oc < 100 L/kg). Anoxic zones enhanced the removal of moderately sorptive compounds via reductive degradation, particularly for diuron, which achieved >95% attenuation. Breakthrough curves indicated that sorption dominated removal during early infiltration stages, while microbial degradation became the primary pathway after biofilm establishment, as confirmed by increased biomass and enzymatic activity within the sediment profile. The study demonstrated that combined sorption–biodegradation–redox transformations, tailored by hydraulic residence time and geochemical gradients, are key to maximizing pesticide removal in RBF systems [109].

Heavy metal attenuation—mostly Fe²⁺, Mn²⁺, As⁵⁺, Pb²⁺, and Zn²⁺—occurs through adsorption, precipitation, and co-precipitation. Oxidation of Fe and Mn under oxic conditions produces insoluble oxides retained in sediments [46,67]. As and Pb are removed by adsorption to Fe²⁺/Mn²⁺ oxides or co-precipitation. Removal efficiencies range from 50–95%, depending on redox potential, pH, and SOM [108]. Field studies in Egypt, China, and South Korea report Fe²⁺ and Mn reductions > 80%, with filtered water concentrations below 0.1 mg/L and 0.05 mg/L, respectively [27,38,39]. However, Mn exceeded 0.4 mg/L at some sites under reducing conditions, emphasizing the need for redox management and monitoring [72]. Metal mobilization under anoxia—including Fe²⁺, Mn²⁺, and NH₄⁺—poses risks when SOM is high or residence time is insufficient. Ref. [106] reported that during induced bank filtration at sites with elevated SOM (>2.5% by weight), transition to anoxic conditions within 3–5 m of infiltration led to Fe²⁺ concentrations increasing from <0.05 mg/L in oxic zones to 0.8–1.6 mg/L, and Mn²⁺ from <0.01 mg/L to 0.2–0.45 mg/L. NH₄⁺ release was also significant, rising from <0.05 mg/L in the source water to 0.4–0.9 mg/L in recovered bank filtrate when residence times dropped below 5 days. The study found that these mobilization events were most pronounced during warm periods (>18 °C) when microbial respiration accelerated organic matter degradation, depleting oxygen and triggering reductive dissolution of Fe/Mn oxides. Seasonal high-flow events further reduced residence time and enhanced mobilization risks, underscoring the importance of maintaining adequate infiltration distances and monitoring redox-sensitive parameters to prevent deterioration of water quality [106]. Despite these challenges, RBF’s passive nature and consistent performance across diverse settings confirm its suitability as a natural pretreatment barrier for irrigation applications.

In addition to heavy metal dynamics, RBF performance is strongly influenced by interactions among microbial contaminants, turbidity, nitrogen species, and other redox-sensitive constituents. Field studies across diverse climatic and hydrogeological settings highlight substantial variability in removal efficiencies, mechanisms, and seasonal responses.

Table 2. Microbial, turbidity, N species, and contaminant removal in RBF systems under various redox conditions.

Reference	Total Coliforms and/or Fecal Coliforms	Turbidity	NH4+, NO3− or Other N Species	Removal Rates	Redox Conditions
[93]	E. coli and total coliforms monitored; effective E. coli removal (~100%) and likely similar for total coliforms; 72 h monitoring.	Removal 98.78%; pumped water 0.3–0.4 NTU; Columbia River water 1.0–5.0 NTU.	-	~100% E. coli removal	Fe oxides mobilized under reducing conditions, adsorbed/precipitated under oxidizing; Fe3+ in alluvium formed in reduced state; higher Fe in river water linked to anthropogenic sources; pH and ionic strength influence removal.
[48]	-	Southern wells 0.38–5.1 NTU; Northern wells 0.1 NTU; p < 0.001.	NO3− highest in NW1; NO3− and Cl− higher in Northern wells; NH4+, NO2−, Mn correlated with high redox potential; NO3− correlated with atrazine in NW1–NW2.	Organic micropollutants (OMP) removal 4–97% depending on compound; NO3− correlated with atrazine presence.	High Fe, Mn, NH4+, NO2− at high redox potential; Northern wells had higher redox potential than Southern; OMP removal higher in oxidative conditions.
[55]	-	-	NO3− in river avg. 5.2 mg/L; seasonal peak 9.4 mg/L (winter), low 4.1 mg/L (summer); NO3− removal 6–89% between river and well; seasonal temperature explained 81% variance in NO3−.	OMP removal varied up to 3 orders of magnitude; diclofenac and sulfamethoxazole removal strongly temp- and redox-dependent; residence time and microbial dynamics influenced removal.	O2 fully depleted in summer; Mn2+ detected in autumn with O2 depletion and NO3− = 0; Mn2+ and Fe2+ detected during anoxic/nitrate-depleted periods (June 2016–January 2017); SO42− ≈ 187–194 mg/L in river.
[56]	150–1100 MPN/100 mL in river; ~98% removal in RBF.	25.4 NTU (river) to 0.58 NTU (RBF); 97.7% removal.	NH3 removal 39.6%, NO2− 12%, NO3− 5.4%; NH3 oxidation rate higher than NO2−, NO3−.	Fe 19.2%, Mn 22.2%, NH3 39.6%, NO2− 12%, NO3− 5.4%.	Fe and Mn decreased from river to well; Fe-reducing/Fe-oxidizing bacteria noted; redox changes linked to ion reduction and color variation.
[110]	-	RBF average turbidity < 1.0 NTU with occasional spikes up to 6.0 NTU; some wells show sporadic variations over two orders of magnitude; production wells generally < 0.3 NTU with occasional spikes; higher turbidity observed in new wells due to incomplete development.	-	Pathogen surrogate reduction > 3.5 log.	-
[21]	None detected in RBF; indicates effective pathogen removal.	Reduced below 1 NTU with 2 ft filtration; maintained <0.3 NTU; ≥4-log reduction.	-	≥4-log turbidity reduction; 100% removal of Giardia/Cryptosporidium in groundwater.	-
[36]	-	<0.5 NTU post-peak flow; 4.5–6.5 log removal over 30 days.	NO3− reduction to NH4+ under anoxic conditions.	TOC > 50% removal; DOC reduced; pathogens removed 4.5–6.5 logs (30 days), ~2 logs (24 h), ~8 logs (25 days).	Fe & Mn sensitive to redox changes; sulfate reduction observed under strong reducing conditions; NO3− acts as electron acceptor; anoxic conditions reported.
[78]	-	2–165 NTU; >20 NTU reduced injectivity; guideline 1–5 NTU; automated flushing kept <20 NTU.	NH4+ oxidizers (Nitrosoarchaeum); NO3− as main N source; denitrifiers, nitrite oxidizers; Sulfurimonas, Sulfuricurvum indicated NO3− reduction with S and H oxidation.	-	Strongly anoxic; Fe2+ = 3.5 mg/L, Mn2+ = 0.58 mg/L during standstill; native groundwater Fe2+ = 9.5–39 mg/L; SO4-reducing and methanogenic (CH4 = 11–40 mg/L).
[97]	-	Turbidity up to 25.5 NTU (2012–2022); high values due to corrosion/encrustation; RBF turbidity removal up to 100%.	NH4+ in surface water: <0.05–7.25 mg/L; groundwater < 0.05–0.26 mg/L; NO3− in surface water: 0.22–11.2 mg/L; groundwater NO3− reduced in warmer months via denitrification.	Mn removal ~40%; color removal ~80%; NO3− removal up to 97% in summer; stable groundwater composition maintained.	Fe and Mn levels influenced by pH and redox; periodic elevated Fe/Mn in river; redox-driven (re)oxidation, sorption, precipitation control mobility.
[111]	-	-	Groundwater NO3− removal mainly via DNF; DNRA causes NH4+ retention; seasonal NO3− reduction affects NH4+ enrichment; NO3−, NO2−, NH4+ analyzed by continuous flow.	NH4+-N enrichment higher in wet season; DOC decay rate: 0.1996 mmol L−1 m−1 (wet) vs. 0.022 mmol L−1 m−1 (dry); more effective DOC removal in wet season.	Sequential Mn4+ and Fe3+ oxide reduction indicates specific redox zonation influenced by OC and nutrient fluxes; anaerobic reduction environments favor DNRA/DNF processes.
[112]	-	Turbidity ratio decreased 59.4–96.9% after 96 h; highest at 80 cm (79.6–96.9%), lowest at >10 cm (59.4–78.6%); r = −0.95 (p < 0.01) correlation between turbidity and permeability at 10 cm.	NH4+ decreased 60.6–89.1% in first 120 h; additional 1.7–2.0% drop between 144–168 h; steady state at 216 h; COD, NO3−, Mn2+ also analyzed; chemical clogging indicated by NH4+ increases during intervals.	-	Oxygenated water infiltration altered Fe and Mn concentrations; higher inflow reduced Mn4+, Mn2+, Fe2+, total Fe via oxidation; precipitation of Fe hydroxides linked to chemical clogging; redox reactions key for Fe/Mn removal/transformation.
[113]	-	-	NH4+: 0.4–3.2 mg/kg; NO2−: 0–6.88 mg/kg; NO3−: 1.32–42.8 mg/kg; Nitrospirae oxidizes NO2− to NO3− aiding NH4+ biotransformation; ammonium/nitrates impact microbial structure and biogeochemical processes.	-	Redox conditions affected microbial diversity and distribution; mineral dissolution increasing Fe (6.9–31.2 g/kg) and Mn (163–678 mg/kg) in groundwater; favorable for Fe-oxidizing and Mn-oxidizing bacteria; sediment pH weakly alkaline.
[114]	-	-	River NH4+: 5.8 ± 1.4 mg/L (winter), 1.0 ± 0.6 mg/L (summer); RBF 90% NH4+: 0.6 ± 0.4 mg/L, 0.4 ± 0.1 mg/L; NO3− decreased via denitrification (reducing conditions); no difference in NO3− removal between 2009 and 2012; NO3− attenuation sustained over 7 years.	DOC removal was 42 ± 8% when RBF contributed 90% of the water, and 54 ± 3% when RBF contributed 50%, with the higher apparent removal at 50% explained by dilution with landside groundwater. Removal of sulfamethoxazole, gemfibrozil, and diclofenac—was controlled by temperature and residence time, with higher efficiencies observed in summer.	Mn < 0.2 mg/L in RBF 50% wells (oxic); Mn in RBF 90%: 0.61 ± 0.24 mg/L (2009), 0.89 ± 0.13 mg/L (2012): reducing conditions; Fe < 1 mg/L in all wells (no strong reduction); redox influenced by microbial activity/electron acceptors.
[115]	River total coliforms: 2.3 × 103–1.5 × 106 MPN/100 mL; RBF filtrate: 4.3 × 104–7.5 × 104 MPN/100 mL (~2 log reduction).	River turbidity: 3.83–13.6 NTU (pre/post-monsoon), 70–180 NTU (monsoon); RBF bed filtration reduced turbidity.	-	DOC removal during pre-chlorination: 7–18%; RBF reduced DOC, color, UV-absorbance, and fecal coliforms by ~50% vs. direct pumping.	-
[116]	River: 660 CFU/100 mL; RBF: 0 CFU/100 mL.	River: 13.47 NTU; RBF: 0.3 NTU.	NO3− 1.38 mg/L (safe for human consumption).	-	-
[72]	Total coliform removal ~99%, meeting microbiological standards.	Turbidity reduced ~90% to safe levels for human consumption.	-	Microbiological ~99%.	Fe2+, Mn2+, Fe3+, Mn4+; Fe oxidized by atmospheric O2; Mn requires aeration; moderate redox potential: higher Fe/Mn; redox/pH changes affect oxidation rate.
[117]	Total/thermotolerant coliforms (MPN/100 mL, Colilert test); river: variable counts; pumped groundwater: complete removal.	River turbidity: 1.3–156 NTU; well water: 0.2–1.6 NTU; Brazilian limit: 5 NTU; long-term drop from 10–300 NTU to 0 NTU in groundwater.	NH4+ from fertilizers/domestic effluents; NO3− increased in aquifers (electron acceptor role); NO2− transitional N species; linked to OM degradation and nitrification.	-	-
[118]	Beberibe River: total coliforms ≥ 1600–≥160,000 NMP/100 mL; E. coli 280–≥160,000 NMP/100 mL; RBF removal: 99.9% in Ganges River, 95% in Nile River; production wells (e.g., Well A) 100% removal.	Turbidity reduction 99–99.9% (Haridwar, India); Beberibe River 16.3–158.0 NTU; wells much lower; suspended particulates reduced by porous medium filtration.	-	-	-
[119]	River water TC: 4240.8–7258.8 MPN/100 mL; E. coli: 23.8–42 MPN/100 mL; sediment removal at 0–20 cm: 85% TC, 56% E. coli; Citrobacter & Enterobacter = 10.57% soil bacteria; E. coli undetectable in groundwater.	-	NH4+-N in groundwater: 2.11 mg/L (dry), 2.66 mg/L (wet); NO3−-N in river: 1.74 mg/L (dry), 2.37 mg/L (wet); groundwater NO3−-N ≈ 0.6 mg/L; at 80 cm depth: NO3−-N = 0.13 mg/L (>94.5% removal).	94.5% NO3− and 94% E. coli removed within 1 m of riverbed surface; 85% TC and 56% E. coli removed at 0–20 cm; Fe-Mn dissolution contributes to high groundwater Fe, Mn.	Shift from aerobic to anoxic; Mn reduction precedes Fe; Fe2+ increases from riverbed to aquifer; high NH4+ and DNRA observed; redox favors Fe/Mn release.

summarizes representative case studies, detailing contaminant levels, removal rates, dominant attenuation processes, and associated redox conditions.

Across all the studies synthetized in Table 2, total/fecal coliform removal is typically near complete (often ≥99.9%), with surrogate/pathogen indicators showing >3.5–6.5 log attenuation, and multiple entries report non-detectable E. coli in the pumped water. Turbidity routinely drops from double- to triple-digit NTU in source rivers to <1 NTU in bank filtrate—frequently <0.3 NTU in production wells—with individual cases documenting ~98–99% removal and stable filtrate even during source-water surges. Where brief spikes occur (e.g., up to ~6 NTU), they are linked to new or recently redeveloped wells rather than steady-state operation, reinforcing that well development and early-phase management are operationally critical.

Nitrogen and redox-sensitive species show strong dependence on temperature, residence time, and redox zonation (Table 2). Reported examples include NH₃ removal around ~40% with much lower concurrent NO₂⁻ and NO₃⁻ decreases in a single campaign, contrasted by cases with NO₃⁻ removal ranging from 6–89% between river and well, where seasonal temperature alone explains ~80% of the variance. Several entries document nitrification under oxic conditions (NH₄⁺ → NO₃⁻) transitioning to denitrification under suboxic/anoxic zones, with DNRA contributing to NH₄⁺ enrichment in wet seasons. Redox shifts also govern Fe/Mn behavior: under oxic flowpaths concentrations fall (often to <0.1–0.2 mg L⁻¹), while anoxia—especially at warm temperatures or short residence times—mobilizes Mn²⁺ and Fe²⁺. These same controls extend to DOC and trace organics: in one long-term dataset, DOC removal was 42 ± 8% when RBF supplied ~90% of the mixture and 54 ± 3% when the mixture contained ~50% RBF (the higher apparent removal at 50% reflecting dilution by landside groundwater). For trace organic compounds and EOCs, performance is compound-specific (4–97% across studies) and improves with warmer conditions and longer residence times; antibiotics and pharmaceuticals such as sulfamethoxazole, gemfibrozil, and diclofenac show distinctly better attenuation in summer and under more oxidative flowpaths.

Operationally, Table 2 underscores two tensions designers must balance. First, pushing yield (higher gradients/pumping) shortens residence time and compresses oxic zones, which can erode microbial and TOrC removal and increase the risk of Fe/Mn breakthrough; conversely, excessive colmation that boosts particulate capture can also drive anoxia and chemical mobilization. Second, seasonality matters: flood periods raise loads (particulates, microbes) but can occasionally scour/”self-clean” the riverbed, whereas dry/warm periods lengthen organic-matter residence, favoring reducing conditions and breakthrough of redox-sensitive solutes. Together, these results point to a simple prescription for irrigation-grade RBF: (i) maintain moderate, seasonally adjustable pumping to preserve oxic travel paths; (ii) ensure rigorous well development and periodic maintenance to avoid start-up spikes; and (iii) monitor a minimal sentinel set (turbidity, E. coli/indicator spores, DO/ORP, Fe, Mn, NH₄⁺/NO₃⁻, and 2–3 representative TOrCs) to track redox and treatment stability across seasons.

6. Colombian Context: The RUT District

6.1. Regional Setting and Suitability for RBF

In Colombia, RBF remains largely unexplored despite favorable alluvial settings in several regions. The RUT Irrigation and Land Reclamation District in Valle del Cauca is an appropriate test bed because of its broad alluvial plains, shallow aquifers, and strong river–aquifer connectivity—attributes consistent with successful RBF systems worldwide. The district spans 10,214 ha across three municipalities (Figure 4 and Figure 5), lies ~170 km northeast of Cali at ~930 m elevation, and occupies the left bank alluvial plain of the Cauca River adjacent to the Western Cordillera [120]. A moderate warm tropical climate (mean 24 °C; bimodal ~1015 mm precipitation) and gently sloping terrain (≈950–1050 m) interact with a seasonally variable river (≈60–1100 m³ s⁻¹) and water-transported sediments, reinforcing lateral hydraulic continuity critical for induced infiltration [121].

6.2. Source Water Quality, Irrigation Constraints, and the Role of RBF

Land use is dominated by sugarcane (5380.9 ha), followed by maize (1888.1 ha), pasture (440.4 ha), and diversified fruit crops including guava (583.9 ha), grapefruit (293.5 ha), and papaya (218.9 ha) [122]. The production mosaic combines large-scale sugarcane with smallholder horticulture (<10 ha) [120]. System-level inefficiencies (≈52% conveyance/distribution losses) and crop water needs drive an annual irrigation demand of ~59.1 million m³ (details in Section 3.1; Figure 5), supplied through two river intakes and an unlined canal network [122]. Seasonal hydrology of the Cauca River further modulates delivery reliability and water quality [121].

Untreated river water carries suspended solids, organic matter, nutrients, pathogens, and agrochemicals that accelerate clogging and reduce the service life of drip and micro-sprinkler systems [3,8]. Under high organic loads and intensive fertilization, biofouling risks and soil salinity/sodicity management become more complex [123]. Microbiological safety is a central constraint for high-value produce; pulses of E. coli and other pathogens, as well as emerging contaminants, elevate food-safety and environmental risks [4,16,22,48]. The district record documents episodic microbiological contamination, large turbidity/TSS excursions, and sporadic Fe/Mn peaks (Table S1), all of which are known triggers of filter/emitter failure and operational downtime. RBF directly targets these stressors: bioactive hyporheic sediments and subsurface residence time jointly reduce turbidity and particles, attenuate biodegradable organics, and provide a robust barrier to indicators and associated pathogens before water reaches distribution. In addition, partial stabilization of redox-sensitive metals within the bank/near-well zone lowers Fe/Mn carryover to pipelines—provided pumping maintains oxic flow paths. These improvements align with irrigation-grade targets (e.g., turbidity < 5 NTU with 2–3 log microbial reduction, Fe < 0.1 mg L⁻¹, Mn < 0.05 mg L⁻¹; Table S1) and the regulatory benchmarks cited in this manuscript (Colombian Decree 1076/2015; Resolutions 1256/2021 and 000132/2021; FAO, WHO, U.S. FDA FSMA, EU 2020/741).

6.3. Hydrogeology and Soils: River–Aquifer Connectivity and Constraints

Sediment dynamics (~560 t km⁻² yr⁻¹) and heterogeneity of water-lain deposits favor induced infiltration. Subsurface profiling identifies layered coarse gravels (≈2–5 mm) and medium sands (≈0.5–1 mm) with <10% clay lenses. Reported hydraulic conductivities range from K_x ≈ 1.2 × 10⁻⁴ m s⁻¹ to ≈3.8 × 10⁻³ m s⁻¹, enabling rapid lateral flow from river to wells [47]. Geo-electrical surveys delineate an unconfined superficial layer (0–8 m; ~25% porosity), a clayey aquitard (8–12 m), and a semi-confined unit (12–18 m) with transmissivity of 350–480 m² day⁻¹ and specific storage of 2.5 × 10⁻³–4.1 × 10⁻³ m⁻¹ [47]. Soils show localized physical degradation relevant to irrigation hydraulics and emitter performance: medium bulk density dominates (1.1–1.4 g cm⁻³), with compacted patches (1.4–1.65 g cm⁻³); available water capacity is predominantly medium (77.2%) but low in some areas associated with >40% exchangeable Mg; and saturated hydraulic conductivity is medium over most of the district but low (23.5%) where poor drainage may occur [124]. These traits increase the premium on low-sediment, chemically stable irrigation water—precisely the niche where RBF provides system-level benefits.

6.4. Conceptual Design, Scaling, and Monitoring (RBF Pilot to District Level)

Following the review’s framework, the RUT case adopts a staged pathway: (i) site screening and conceptual model grounded in geophysics, lithologic heterogeneity, and river stage dynamics; (ii) pilot design specifying well type (vertical, collector, or galleries), screen placement in the most conductive zones while avoiding reduced layers, and conservative pumping to protect oxic flow paths; (iii) performance verification via continuous field monitoring and targeted tests; and (iv) scaling guided by residence-time and yield metrics. Four sites have been identified as suitable for piloting (Figure 5).

Design-critical parameters (detailed in Section 4.3) include aquifer K and heterogeneity (controlling induced infiltration and residence time), well depth/screen length, lateral offset from the river and inter-well spacing, and adaptive pumping rates to limit sediment mobilization and biofouling. Monitoring spans turbidity, microbial indicators, nutrients, Fe/Mn, and supporting physicochemical variables (DO, redox, OC), complemented by head gradients, clogging proxies, and temperature to track biogeochemical conditions. Preliminary estimates using local K (≈1.2–3.8 × 10⁻⁴ m s⁻¹) and global analogs indicate travel velocities of ~0.5–2.0 m day⁻¹ in hyporheic sediments and 0.1–0.5 m day⁻¹ in deeper zones, yielding hydraulic residence times of ~8–25 days under moderate pumping (≈8–12 m³ h⁻¹). These ranges are consistent with RBF performance benchmarks (Table 1) for achieving <1 NTU and >3 log₁₀ pathogen reduction, and they need to be validated via step-drawdown/pumping tests and tracer studies.

At district scale, meeting the annual demand of 59,083,592 m³ (≈1873 L s⁻¹ on average; up to ~3900 L s⁻¹ at peak when accounting for losses) implies a distributed network sized by per-well yields observed in tropical alluvial settings: ~8–12 m³ h⁻¹ per vertical well and ~50–80 m³ h⁻¹ for horizontal collectors (Section 3.1; Table S1). Given the extreme variability of Cauca River quality (turbidity 20–558 NTU with flood peaks to ~10,000 NTU; total coliforms 10³–10⁶ MPN 100 mL⁻¹; Fe up to ~102 mg L⁻¹; Mn up to ~1.28 mg L⁻¹; Table S1), RBF is positioned as a pretreatment and risk-reduction step that stabilizes feedwater before any downstream polishing, reduces maintenance, and widens the feasible crop portfolio in line with international phytosanitary requirements (Figure 5).

7. Challenges

7.1. Climatic and Seasonal Variability

RBF system performance is governed by climatic and seasonal variables that influence surface water quality, aquifer dynamics, and natural attenuation processes. River discharge fluctuations represent the primary control mechanism: low flows reduce hydraulic gradients and weaken induced recharge, increasing the proportion of native groundwater (often of inferior quality) in extracted water, while high flows enhance surface water availability and oxygenation, promoting redox stability and contaminant removal efficiency [30]. Extreme events produce contrasting effects—floods transport suspended solids that clog riverbeds but simultaneously scour hyporheic zones to restore porosity, whereas droughts diminish hydraulic gradients, reduce surface water contributions, and promote anaerobic conditions with elevated salinity [123].

Flood inundation increases N, Fe, and Mn concentrations in groundwater by shifting redox interfaces and promoting reduction of Mn⁴⁺ and Fe³⁺ oxides to soluble species, while simultaneously enhancing river recharge and elevating DO, NO₃⁻, and organic carbon levels. In a study of the Yongding River floodplain, China, Ref. [125] observed that post-flood infiltration caused NO₃⁻ concentrations in bank filtrate to rise from 0.52–0.71 mg/L to 1.21–1.65 mg/L, accompanied by a 30–45% increase in dissolved organic carbon (from 2.1–2.4 mg/L to 2.8–3.5 mg/L) due to enhanced surface water recharge. Dissolved oxygen increased from <1.0 mg/L pre-flood to 3.2–4.5 mg/L immediately after inundation, reflecting a temporary re-oxidation of infiltration pathways. However, subsequent microbial respiration in organic-rich sediments drove redox potentials below +100 mV, initiating Mn⁴⁺ and Fe³⁺ reduction. This led to Fe²⁺ increasing from <0.1 mg/L to 0.84–1.27 mg/L and Mn²⁺ from <0.05 mg/L to 0.22–0.39 mg/L within 10–15 days after flooding. These dynamics highlight that flood events induce complex, time-dependent shifts in groundwater geochemistry, with an initial improvement in oxic conditions followed by mobilization of redox-sensitive metals under anoxia [125]. Continuous hydrodynamic and hydrochemical monitoring remains essential for evaluating RBF performance under variable climatic conditions.

7.2. Temperature Effects on Biotransformation and Redox Stability

Temperature directly regulates microbial metabolism within the hyporheic zone. Warmer conditions enhance oxygen consumption and accelerate biotransformation of OM and emerging contaminants. However, benefits diminish under high organic loads leading to oxygen depletion and reducing conditions. In controlled bank filtration experiments along the Erpe River, Germany, Ref. [55] demonstrated that increasing water temperature from 5 °C to 15 °C reduced the aerobic zone length from 1.6 m to 0.9 m due to a 65% increase in oxygen consumption rates, accelerating the transition to nitrate-reducing and Fe/Mn-reducing zones. At 15 °C, attenuation of the pharmaceutical diclofenac improved from 42% to 78%, and carbamazepine from 21% to 53%, reflecting enhanced microbial degradation rates. However, under conditions of elevated dissolved organic carbon (>5 mg/L), the benefit of higher temperatures diminished, as oxygen was depleted within <0.5 m of infiltration, promoting early onset of reducing conditions that favored metal mobilization (Fe²⁺ up to 0.9 mg/L, Mn²⁺ up to 0.25 mg/L). The study emphasized that while moderate warming can enhance biotransformation processes, managing organic loads is critical to avoid redox shifts that compromise water quality [55]. Column experiments simulating RBF in the Lausitz region, Germany, showed that increasing infiltration water temperature from 5 °C to 15 °C accelerated oxygen depletion rates by ~70%, reducing the oxic zone depth from 16 cm to 9 cm and triggering earlier onset of nitrate-reducing conditions [12]. At higher temperatures, DOC removal improved from 46% to 71%, and attenuation of the pharmaceutical carbamazepine increased from 18% to 49%, indicating enhanced microbial degradation kinetics. However, when influent DOC exceeded 5 mg/L, oxygen was depleted within 4 cm of infiltration regardless of temperature, leading to early Fe/Mn reduction and associated metal release (Fe²⁺ up to 0.8 mg/L, Mn²⁺ up to 0.22 mg/L). These findings demonstrate that while temperature elevation can strengthen biotransformation capacity, it must be balanced with organic matter management to prevent water quality deterioration under reducing conditions [12].

7.3. Sediment and Biofilm Clogging Dynamics

A persistent technical constraint is progressive riverbed and adjacent aquifer sediment clogging. This results from suspended solids, POM, chemical precipitates, and microbial biofilm accumulation, collectively reducing infiltration capacity and K. Clogging is categorized as physical (sediment deposition-driven) or biological (microbial growth biofouling), with the latter particularly active in hyporheic zones due to high organic loads fueling microbial respiration [69]. Under moderate conditions, biofilms and POM enhance removal efficiency by increasing adsorption and biodegradation potential. However, excessive accumulation reduces oxygen availability and promotes Mn²⁺, Fe²⁺, and N species mobilization. In field and mesocosm studies, Ref. [57] documented that physical clogging from fine sediment deposition can reduce hydraulic conductivity by 40–90% within weeks during high-turbidity events, while biological clogging from microbial biomass and EPS production can lower infiltration rates by up to 80% over several months in nutrient-rich waters. The formation of biofilms increased dissolved organic carbon removal efficiency by up to 35% under moderate growth, but when biomass exceeded 100 µg C/cm², oxygen penetration depth declined from 8–10 cm to <2 cm, triggering Mn²⁺ (>0.2 mg/L) and Fe²⁺ (>0.5 mg/L) release from sediments [57]. In controlled column experiments simulating RBF along the Second Songhua River, China, Ref. [112] found that fine sediment deposition reduced hydraulic conductivity by 54–78% within 20 days, while biological clogging from biofilm growth under high organic matter inputs further decreased K by up to 88%. Moderate biofilm development enhanced NH₄⁺ and DOC removal by 20–30% due to increased adsorption and microbial processing. However, once biomass exceeded 80 µg C/cm², DO penetration dropped below 1.5 cm, triggering Fe²⁺ increases from <0.1 mg/L to 0.72 mg/L and Mn²⁺ from <0.02 mg/L to 0.21 mg/L [112]. These results confirm that while biofilms and POM can initially improve contaminant attenuation, excessive accumulation shifts redox conditions and mobilizes undesirable reduced species.

Maintenance strategies include intermittent well resting, mechanical surface layer removal, and exploiting natural high-flow scouring. In high-sediment environments, clogging reduces K by up to 80% within six months, necessitating frequent maintenance and robust rehabilitation protocols [39]. Clogging management must consider seasonal dynamics. Wet seasons deliver high sediment loads accelerating clogging but potentially enable flood-induced self-cleaning through turbulence [71]. Drought periods prolong OM residence time and promote compaction, increasing anoxic breakthrough risks. Continuous observation and modeling of these dynamics are critical for sustaining tropical RBF operation. Field monitoring of an RBF site in Odisha, India, by [100] showed that during the monsoon season, turbidity in source water rose from <10 NTU to >150 NTU, leading to a 42% reduction in infiltration rate over two months due to fine sediment deposition. However, subsequent high-flow flood pulses partially restored infiltration capacity by ~25% through scouring of the colmation layer. In contrast, during the dry season, low flow conditions and elevated temperatures (>30 °C) extended OM residence times, intensifying microbial respiration and causing DO depletion within <0.5 m of infiltration. This shift promoted Fe²⁺ release up to 0.6 mg/L and Mn²⁺ up to 0.18 mg/L in recovered water [100]. The study highlighted that integrating seasonal flow and sediment load modeling with operational adjustments—such as variable pumping regimes—can sustain performance and reduce anoxic breakthrough risks in tropical RBF systems.

7.4. Extreme Turbidity Events and Pretreatment Vulnerabilities

Extreme turbidity events pose critical operational risks. Rapid well clogging under high-turbidity conditions severely limits RBF-extracted water utility for pressurized irrigation. Recommended adaptive measures include real-time turbidity monitoring, sediment loading alerts, well rotation, and strategic abstraction point relocation during high-turbidity periods [47]. Field and pilot-scale experiments with UGF units fed with raw Cauca River water revealed significant susceptibility to hydraulic and physical clogging when used as a pretreatment for micro-irrigation systems [101]. Operating at filtration velocities of 0.5–1.0 m/h and initial influent turbidities between 15 and 70 NTU, the UGF experienced a progressive head loss increase, reaching the operational limit of 15 cm within 10–20 days depending on solids load. Clogging was concentrated in the finer upper gravel layers (6.35–3.17 mm), where the highest TSS capture occurred, leading to localized reductions in porosity and infiltration rate [101]. This effect shortened filter runs and required more frequent cleaning, which, in agricultural contexts, could reduce system reliability and increase labor demand. Moreover, particle size analysis indicated that while larger fractions (>80 µm) were completely removed, fine particles (<5 µm) persisted in the effluent, posing a risk of emitter clogging in micro-irrigation applications if RBF systems are not complemented with adequate downstream filtration [101].

RBF system sustainability under tropical and high-load conditions depends on integrated management combining continuous water quality monitoring, adaptive flow and pumping strategies, predictive modeling, and proactive maintenance protocols. Long-term success requires tailoring designs to site-specific hydrogeology, contaminant profiles, and climatic regimes, with particular attention to dynamic clogging and redox process behavior.

8. Future Research Directions and Recommendations

8.1. Pilot-Scale Validation and Core Design Variables

The implementation of RBF systems in tropical agricultural contexts demands a research agenda anchored in empirical field evidence and designed to reflect complex hydroclimatic, biogeochemical, and operational dynamics of humid environments. Although RBF has been extensively studied for urban water supply in temperate regions, its adaptation for agricultural irrigation under tropical conditions—where high sediment loads, elevated temperatures, and mixed contaminant profiles are prevalent—remains a critical research frontier.

Pilot-scale studies are essential as the foundational step in validating RBF systems for tropical agriculture. These should evaluate system performance under variable hydraulic residence time, infiltration rate (<0.01 m/h), aquifer grain size distribution, and redox potential conditions. The impact of well geometry (vertical, horizontal, or gallery wells), pumping regimes, and river-aquifer connectivity must be assessed using hydrogeological profiling, permeability tests, and numerical flow modeling. Beyond classic parameters like turbidity and coliform bacteria, continuous monitoring of reduced species including NH₄⁺, NO₂⁻, and trace metals (Fe²⁺, Mn²⁺) is vital for detecting undesirable redox-driven transformations.

To address persistent contaminants, future research should evaluate hybrid systems integrating RBF with complementary treatment technologies including advanced filtration (ultrafiltration or membrane systems), chemical oxidation (ozonation), constructed wetlands, or riparian buffers [126,127]. The use of biochar could improve the operations of the RBF as well. Such configurations are particularly valuable for pharmaceutical residues, pesticides, and endocrine-disrupting compounds that often bypass conventional RBF processes. Deployment should be based on compound-specific attenuation studies quantifying removal efficiencies across different RBF depths, redox zones, and microbial communities.

8.2. Long-Term Soil and Nutrient Dynamics Under RBF Irrigation

A major knowledge gap concerns long-term impacts of RBF-treated water on soil systems under continuous irrigation. While improved water quality reduces emitter clogging and mitigates suspended solids and pathogen inputs, the fate of residual nutrients, salts, and emerging contaminants in agricultural soils remains largely unknown. Tropical soils are especially vulnerable to secondary salinization, nutrient leaching, and microbial community structure changes [25,123,128]. Long-term field experiments and coupled soil-water-crop modeling are urgently needed to evaluate potential shifts in fertility, soil structure, and nutrient cycling under RBF irrigation regimes.

Particular attention should be given to nutrient dynamics, especially filtered water interactions with PO₄³⁻, K⁺, and NO₃⁻ in the root zone. These interactions influence fertilization strategies, crop uptake efficiency, and environmental losses through percolation. Nutrient biogeochemical behavior and potential interactions with residual organic contaminants should be investigated through column studies, lysimeter trials, and soil microbial assays. Transformation pathways including DNRA, anammox, and Fe-mediated NH₄⁺ oxidation warrant deeper exploration in high-SOM tropical soils.

8.3. Geographic and Functional Gaps in Current Research

Most existing RBF studies focus on temperate climates or drinking water applications, with limited transferability to tropical irrigation contexts. This geographic and functional bias underscores the need for context-specific investigations reflecting tropical edaphoclimatic variability, high-frequency irrigation cycles, and intensive agrochemical use. Research designs must integrate spatially and temporally resolved measurements of contaminants, aquifer responses, and system resilience to extreme events such as floods or droughts that can drastically alter infiltration dynamics, contaminant loadings, and system performance.

Future research must prioritize transparency, reproducibility, and cross-regional comparability. Studies should explicitly report operational conditions—well depth, flow rates, sediment characteristics, and residence times—and include comprehensive pre- and post-filtration water quality datasets. Open-access data platforms and standardized protocols for field sampling, sensor calibration, and laboratory analysis will enable meta-analyses and accelerate scientific progress.

8.4. Monitoring and Modeling Tools for Predictive Capacity

A robust monitoring framework is central to both research and implementation. Real-time multiparameter sensors (pH, redox potential, DO, turbidity, NH₄⁺, NO₂⁻, DOC) coupled with periodic laboratory analyses (high-performance liquid chromatography for pharmaceuticals) can capture short-term variability and long-term trends. These platforms must function reliably under high-turbidity, high-organic-load conditions typical of tropical rivers.

Modeling efforts should integrate with field data to enhance predictive capacity. Coupled flow and transport models (MODFLOW, MT3DMS, Delft3D) should simulate residence time distributions, redox front propagation, and contaminant fate under various hydrological scenarios. Model calibration and validation require high-resolution datasets on RHC, aquifer structure, and flow gradients. Incorporating stochastic elements will improve model robustness under climatic uncertainty or sediment disturbance conditions.

Scaling RBF sustainably requires interdisciplinary collaboration integrating expertise in hydrogeology, soil science, environmental microbiology, water quality monitoring, agronomy, and systems engineering. This enables co-design of systems reflecting local realities and user needs, including smallholder irrigation contexts. Equally important is stakeholder engagement including water user associations, agricultural extension agents, and environmental regulators in co-developing performance benchmarks, maintenance protocols, and training programs ensuring long-term system viability and farmer adoption.

Advancing RBF application for tropical agricultural irrigation requires a multi-pronged research agenda combining field experimentation, mechanistic modeling, and participatory technology transfer. Priority areas include pilot testing under representative hydrogeological and climatic conditions; monitoring and modeling redox-sensitive contaminants and performance thresholds; investigating soil system interactions, particularly nutrient and salt dynamics under long-term irrigation; developing hybrid treatment strategies for recalcitrant contaminants; creating open science tools for transparency and cross-site learning; and building capacity for technical operators and researchers. By addressing these research priorities, the scientific community can support RBF transition from a proven potable water technology to a resilient and scalable solution for sustainable irrigation in tropical agriculture.

9. Conclusions

This review confirms RBF as a scientifically validated, operationally mature technology that reliably upgrades surface waters to irrigation-grade quality across diverse hydroclimates and aquifer types. Under well-designed residence times and moderate gradients, RBF consistently achieves strong particulate and microbial attenuation—turbidity frequently <1 NTU and >3-log pathogen reduction—via coupled physical filtration, sorption, and biologically mediated transformations in the hyporheic zone. In tropical systems, these same processes remain effective but are conditioned by warm, organic-rich waters and discharge seasonality, which accelerate biogeochemical reactions, compress oxic zones, and elevate clogging risk; these effects are predictable and manageable with conservative drawdowns, redox-aware operation, and maintenance that anticipates interface colmation during high-sediment periods.

For the RUT district, the hydrogeologic envelope reported here—transmissivity 350–480 m² d⁻¹, hydraulic conductivity 1.2 × 10⁻⁴–3.8 × 10⁻³ m s⁻¹—supports infiltration rates on the order of 0.5–2.0 m d⁻¹ and residence times of ~8–25 days under moderate pumping, aligning with global benchmarks for achieving irrigation-grade water quality. On this basis, RBF implementation in RUT is technically feasible and economically justified, with expected benefits that include substantial reductions in irrigation system maintenance (~40–60%) and multi-year extension of component lifespans, alongside the potential to shift toward higher-value crop portfolios as water quality stabilizes and particle/pathogen loads decline. A staged, distributed wellfield—validated through site-specific pumping and tracer tests—offers a pragmatic pathway to scale while preserving treatment margins across wet and dry seasons.

Before district-wide deployment, three gaps require resolution in tropical agricultural settings: long-term soil–water interactions under continuous RBF irrigation (nutrients, salts) in high-evapotranspiration environments; compound-specific behavior of priority agrochemicals under variable redox; and operational thresholds that keep systems robust through flood-pulse extremes without sacrificing removal performance. Addressing these through pilot-scale trials at the identified RUT sites—with rigorous monitoring of piezometry, redox, turbidity/microbiology, Fe–Mn–NH₄⁺ dynamics, aquifer response, and crop/soil outcomes—will yield site-specific design criteria and adaptive operating envelopes. In parallel, targeted capacity building and regulatory alignment with agricultural water standards will convert proven technical performance into durable practice, positioning RBF as a sustainable, decentralized backbone for irrigation water security in RUT and comparable tropical agricultural irrigation districts.