A Spatially Explicit Physically Based Modeling Framework for BOD Dynamics in Urbanizing River Basins: A Case Study of the Chao Phraya River—Tha Chin River

Abstract

Biochemical Oxygen Demand (BOD) is a key indicator of organic pollution and a proxy indicator reflecting organic loading that can indirectly influence eutrophication processes in aquatic systems. This study presents a spatially explicit, physically based modeling framework for simulating BOD dynamics in the urbanizing Chao Phraya and Tha Chin Rivers Basin in central Thailand. The framework integrates the Personal Computer Storm Water Management Model (PCSWMM) with GIS-based datasets to represent pollutant sources, hydraulic flow, and land use. The model was calibrated and validated using data from 36 monitoring stations (2021–2022), achieving strong performance: an NSE of 0.72 and an MAE of 0.35 mg/L for the Chao Phraya River, and an NSE of 0.88 and an MAE of 0.12 mg/L for the Tha Chin River. Scenario simulations for 2032 projected BOD concentrations exceeded 4 mg/L in several downstream segments under the baseline (no-intervention) scenario, indicating elevated organic pollution and potential oxygen depletion that may indirectly exacerbate eutrophication risk in the Upper Gulf of Thailand, particularly in tidal zones with low dilution and nutrient accumulation. Model projections suggest that effective mitigation would require a 20–30% reduction in BOD loads, achievable through enhanced wastewater treatment and stricter pollution controls. Although BOD reduction alone cannot eliminate eutrophication, it supports broader nutrient management efforts by improving baseline water quality conditions. The proposed model offers a robust tool for identifying pollution hotspots, evaluating management strategies, and informing integrated river basin policies under continued urban growth.

1. Introduction

Eutrophication of coastal estuaries, embayments, and nearshore areas is a water quality challenge experienced globally [1,2,3], including Southeast Asia [4,5,6,7,8]. Eutrophication drivers can result in a complex interaction of nutrient inputs from urban, agricultural (and aquaculture) activities, atmospheric deposition, groundwater, seasonal and event-based hydroclimatology, and coastal hydrodynamics [6,9,10,11,12,13].

The Upper Gulf of Thailand also experiences water quality challenges, including eutrophication [14,15,16,17], which has raised concern with regulatory and research communities [18,19,20]. The upper gulf receives discharge from four large rivers [21] along the northern coast (Chao Phraya, Tha Chin, Maeklong, and Bangpakong Rivers) and, as such, the development of a water quality management plan, focusing on these source waters would be of benefit to improve water quality conditions in the upper gulf. As such, the objective of this study was to pilot the development of a spatially explicit, physically based framework to simulate Biochemical Oxygen Demand (BOD) dynamics focusing on the Chao Phraya and Tha Chin Rivers as a proof of concept. As will be shown, due to modeling assumptions and data availability, the model framework is most appropriately applied at a planning level to support policy decision-making rather than at a design scale.

The model integrates the Personal Computer Storm Water Management Model (PCSWMM) with GIS datasets to characterize pollutant sources, land use patterns, and hydraulic conditions across 13 provinces in central Thailand. Water quality data from 36 monitoring stations, collected between 2021 and 2022, were employed for model calibration and validation. Scenario-based simulations were conducted to assess the implications of current and enhanced wastewater management practices on downstream BOD levels and the potential risks of eutrophication in the Upper Gulf of Thailand. The results contribute to the development of data-driven tools for integrated water quality planning and adaptive river basin management under continued urbanization. Physically based hydrodynamic models, when integrated with high-resolution Geographic Information System (GIS) data, offer a robust platform for representing the generation, transport, and transformation of organic pollutants such as BOD across heterogeneous, rapidly urbanizing watersheds. These models support scenario-based planning, facilitate the identification of pollution hotspots, and inform evidence-based strategies for basin-scale water quality management [22,23].

In this study, we used BOD as an indicator of eutrophication risk. Frequently, analytes such as total nitrogen, total phosphorus, and chlorophyll a are used as indicators of trophic status, although for certain ecosystem health assessments, detailed quantitation of algal species distribution is required [24,25,26,27]. However, we used BOD in this study because it is an analyte for which there is relatively more data available from the Thai Pollution Control Department (PCD) to support the modeling effort. While BOD, like other water quality parameters, exhibits spatial and seasonal variability, its long-term and spatially extensive monitoring coverage across the basin enables consistent calibration, validation, and scenario comparison at the basin scale. In contrast, nutrient species such as nitrogen, phosphorus, and chlorophyll a are monitored less consistently in space and time, limiting their direct application for basin-scale scenario modeling under current data availability. BOD is frequently used as an indicator of organic pollutants in receiving waters, including wastewater [28,29], and the PCD has established clear standards for BOD levels, categorizing the health of water bodies. Under the PCD system, Class 1 rivers are excellent quality with BOD < 1.5 mg/L and are suitable for consumption, aquatic life conservation, fisheries, and recreation; Class 2 rivers are good quality with BOD < 2.0 mg/L and are suitable for consumption after ordinary treatment, aquatic life conservation, fisheries, and recreation; Class 3 rivers are medium clean with BOD < 4 mg/L, and are suitable for consumption with ordinary treatment processes, agriculture, and aquatic life conservation; Class 4 rivers are fairly clean with BOD < 5 mg/L, and are suitable for consumption after special treatment processes and for industry. Class 5 rivers are those not included in Classes 1–4 and are typically considered severely polluted, and are only suitable for navigation. Studies have shown that BOD can be associated with eutrophic conditions and other trophic status indicators such as chlorophyll a, nitrate, and total phosphorus [30,31,32,33]. In selecting BOD as a key indicator, this study focuses on organic pollution that indirectly relates to eutrophication processes. While eutrophication is typically assessed using nutrients (N and P) and chlorophyll-a, BOD serves here as a practical proxy indicator due to data availability and its strong correlation with organic matter that influences oxygen dynamics in receiving waters. This approach aligns with PCD’s monitoring framework and supports a proof-of-concept modeling application under data-limited conditions.

The TMDL program was introduced in the U.S. under Section 303(d) of the Clean Water Act. As noted by Birkeland (2001) [34], “In order to meet water quality standards under the new TMDL program, states must allocate pollutant load reductions among sources in a watershed. The new regulations compel states to address nonpoint source pollution in order to achieve the pollutant load reductions necessary to meet TMDLs because, in many instances, technological fixes for reducing point source discharges have approached their cost-effective limit”. US EPA (2008) [35] notes that, most simply, the TMDL to meet water quality standards can be calculated as follows:

TMDL = Σ(WLA) + Σ(LA) + MOS

(1)

where WLA is the sum of waste load allocations for point sources; LA is the sum of waste load allocations for nonpoint sources; and MOS is a margin of safety that accounts for uncertainty. The TMDL calculation can be accomplished through physically based modeling, non-modeling, empirical and data-driven approaches, or a combination of the two [36]. US EPA (1995) [37] notes that different models may be employed for development of a TMDL, and guidance can be provided on model selection, depending on the conditions of the watershed and the analytes of interest. The general TMDL approach has been applied to address different water quality challenges throughout the United States and other areas of the world [38,39,40,41,42,43,44] and is well-suited for application in this study.

The use of a one-dimensional dynamic wave formulation within PCSWMM is consistent with this planning-level objective, enabling basin-scale screening, identification of pollution hotspots, and comparative scenario analysis under limited data availability.

2. Materials and Methods

2.1. Study Area

The Chao Phraya and Tha Chin Rivers form an interconnected riverine system that serves as the hydrological backbone of central Thailand. The Chao Phraya River originates at the confluence of the Ping and Nan Rivers in Nakhon Sawan Province and flows southward through several densely populated provinces, including Chainat, Ayutthaya, and Bangkok, before discharging into the Upper Gulf of Thailand. The Tha Chin River branches off from the Chao Phraya in Chainat Province and follows a more westerly course through Suphan Buri and Nakhon Pathom, ultimately reaching the sea at Samut Sakhon [45,46,47]. Together, these rivers define an extensive watershed covering approximately 160,000 square kilometers and encompass 13 provinces: Nakhon Sawan, Uthai Thani, Chainat, Singburi, Lopburi, Ang Thong, Suphan Buri, Ayutthaya, Pathum Thani, Nonthaburi, Bangkok, Nakhon Pathom, and Samut Sakhon (Figure 1) [20].

This region exhibits a diverse range of land uses, transitioning from mountainous, forested upstream areas in the north to intensively urbanized and industrialized floodplains in the central and lower reaches. The lower Chao Phraya Basin, in particular, supports a population of over 20 million and hosts key economic corridors [48], export-oriented industries, and large-scale agricultural zones. Major industrial estates located in Pathum Thani, Nakhon Pathom, and Samut Sakhon contribute significant volumes of treated and untreated wastewater. Similarly, dense urban areas such as Bangkok and its surrounding metropolitan region, exert heavy domestic sewage loads that often exceed the treatment capacity of existing municipal infrastructure [19,49].

Hydrologically, the river system is characterized by low longitudinal gradients (e.g., approximately 0.0003 m/m in the lower Chao Phraya), extensive floodplains, and dense irrigation canal networks that create extensive hydraulic connectivity. Tidal influence from the Gulf of Thailand extends up to 350 km inland, inducing flow reversals during high tides and contributing to backwater effects. These conditions result in prolonged pollutant residence times and limited downstream flushing, thereby facilitating the accumulation of organic pollutants and increasing the likelihood of hypoxic events and eutrophication, particularly in the estuarine transition zones [20,50].

The downstream discharge from both rivers enters the Upper Gulf of Thailand, a semi-enclosed coastal zone characterized by weak tidal mixing and limited oceanic exchange. This area has been increasingly affected by eutrophication, primarily driven by high nutrient and organic matter inputs from upstream sources. Water quality assessments over the past 20 years have reported elevated levels of total nitrogen, phosphorus, and BOD, particularly during the dry season when river flows are minimal [15,51]. Consequences include oxygen depletion (hypoxia), fish kills, and degradation of sensitive coastal habitats such as seagrass beds and mangrove ecosystems. The Tha Chin Estuary, in particular, has been consistently classified among Thailand’s most polluted coastal water bodies, often exhibiting dissolved oxygen levels below 2 mg/L [19].

The combination of high ecological sensitivity and intensifying anthropogenic pressures underscores the need for watershed estuarine management [52,53,54,55]. A comprehensive understanding of BOD transport dynamics at the basin scale is therefore essential for developing effective pollution control strategies and protecting downstream coastal ecosystems.

2.2. Data Collection and Processing

Hydrological, water quality, and spatial data were collected from multiple government agencies to support model construction, calibration, and scenario-based analysis. Time-series observations from 2021 to 2022 were obtained from 36 water quality monitoring stations along the Chao Phraya and Tha Chin Rivers (Figure 1), operated by the Pollution Control Department (PCD) and Hydro-Informatics Institute [19,56]. Key parameters included BOD, flow rate, water level, dissolved oxygen (DO), and temperature. These datasets served as the basis for both calibration and validation of the model, as well as for scenario simulations.

Spatial data were sourced from national databases and processed for integration into the modeling platform. Land use data were obtained from the Land Development Department (LDD), using high-resolution raster maps classified into urban, industrial, agricultural, and forested zones. Digital elevation models (DEMs) at 30 m × 30 m resolution were retrieved from the Shuttle Radar Topography Mission (SRTM) and used to delineate catchment area boundaries, sub-catchments, and overland flow paths [57]. The river network and canal infrastructure shapefiles were acquired from the Royal Irrigation Department and processed in ArcGIS 10.8 to construct the base hydrologic geometry of the model [20]. Pollution source data were categorized into three major point sources, including the following:

• Domestic wastewater loads were estimated using district-level population data combined with per capita wastewater generation of approximately 200 L/person/day. Total wastewater volume for each district was then converted to pollutant loadings using standard BOD and nutrient generation factors, producing spatially aggregated domestic loads for model input [19].
• Industrial loads were derived from the Department of Industrial Works (DIW) factory registry, using reported discharge points, industry types, and wastewater characteristics. Sector-specific pollutant coefficients were applied to estimate loads, with adjustments based on compliance records where available. All discharge locations were spatially aligned with the river and canal network for model routing [58].
• Agricultural runoff was estimated based on production volumes from key agricultural sectors, including aquaculture, marine aquaculture, livestock (e.g., swine), and rice cultivation. Pollutant loads were calculated using sector-specific nutrient and organic matter generation factors, combined with information on feed usage and waste management practices. These loads were then spatially distributed across relevant agricultural zones and aggregated at the sub-catchment level for model input [19].

All datasets were georeferenced and standardized to a common coordinate system (UTM Zone 47N, WGS 84) to ensure spatial consistency and seamless integration into the modeling environment.

2.3. Model Description and Setup

This study is explicitly designed as a planning-level, proof-of-concept modeling application aimed at supporting basin-scale assessment and strategic decision-making under data- and resource-constrained conditions. The modeling framework is intentionally configured to elucidate system-level hydrodynamic responses and relative spatial patterns of BOD transport across alternative management scenarios. Accordingly, the PCSWMM is applied to represent first-order behavior and comparative dynamics that are most relevant for scenario evaluation, prioritization of interventions, and policy-relevant insights. Model assumptions and simplifications are selected to ensure internal consistency, computational efficiency, and interpretability, with results intended to inform strategic planning and adaptive management at the basin scale.

U.S. EPA (1995) [37] reviewed the application of five different physically based models for use in TMDL studies and assessed them based on level of complexity, sources, support, and documentation: HSP-F, WASP5, CE-QUAL-RIV1, QUAL2E, and Multi-SMP. Since then, the SWAT and SWMM models have also been employed in TMDL studies [41,59]. In Southeast Asia, MIKE products are popular and might be employed. An advantage of the SWMM model is that it is a fully dynamic hydrologic and hydraulic model that seamlessly represents runoff quantity and quality from urban and non-urban land uses and can explicitly consider the subsurface sewer network in urban areas; an option generally not available in the models noted above (with the exception of some MIKE products). Based on these considerations, and because our team has extensive experience in applying the PCSWMM version of the SWMM model [60,61,62,63,64], we employed PCSWMM in this study. Unlike more data-intensive models such as MIKE11 [65] and WASP [66], which typically require extensive bathymetric inputs and are often tailored for deep-water systems or reservoirs, PCSWMM provides a flexible, modular architecture that supports dynamic hydraulic routing, BOD transport from both point and non-point sources, and the integration of land use and wastewater treatment scenarios through GIS. These features make it particularly appropriate for river basins characterized by mixed urban–agricultural landscapes, tidal backwater effects, and limitations in monitoring data, thereby enabling practical, spatially explicit decision-making under future urban growth trajectories.

2.3.1. Model Structure and Hydrologic Representation

The PCSWMM framework was configured to simulate in-stream pollutant transport, focusing on key hydraulic control points and pollutant sources. The hydraulic network consisted of 179 conduits and 54 internal hydraulic nodes, excluding designated pollutant source nodes. These nodes represented junctions, outfalls, and control structures throughout the river systems (Figure 2). Importantly, the spatial configuration and number of conduits and nodes were not arbitrarily selected. Rather, they were defined directly based on the officially surveyed river cross-sections and hydraulic control points provided by the Royal Irrigation Department. The model resolution, therefore, reflects the spatial detail available from these authoritative surveys. Three primary loading nodes were designated for domestic, industrial, and agricultural discharges, based on empirical data and national databases. Pollutant routing was simulated under dynamic wave conditions to accurately reflect the prevailing flow regimes, including tidal fluctuations, backwater effects, and unsteady flow patterns that are characteristic of the Chao Phraya and Tha Chin Rivers. These flow regimes influence the transport and dilution of pollutants, particularly in low-gradient river sections and estuarine zones where water movement is highly variable [67]. For simplification in this study, land use-specific rainfall–runoff mechanisms were intentionally omitted to focus on in-stream transport processes; as such, parameters related to surface hydrology (e.g., curve number, imperviousness) were not considered but could be in a more detailed modeling scenario. This simplification aligns with the study’s objective of modeling BOD dynamics primarily driven by point-source loading and hydrodynamic conditions.

For this planning-level application, the following modeling assumptions and simplifications were adopted:

• Explicit simulation of land use–specific rainfall–runoff generation, pollutant buildup, and wash-off processes was intentionally excluded to maintain a focus on basin-scale in-stream transport processes.
• The model objective is to evaluate relative management scenario performance driven by aggregated source loading and hydrodynamic behavior, rather than event-scale runoff dynamics.
• Incorporating detailed buildup and wash-off processes would require extensive site-specific monitoring and parameterization that are not currently available at the scale of the Chao Phraya–Tha Chin River system.
• Diffuse pollutant contributions are therefore implicitly represented through calibration against observed BOD concentrations, consistent with screening-level and TMDL-oriented modeling practice.

Boundary conditions at the downstream estuarine nodes were defined using observed tidal water levels provided by the Hydro-Informatics Institute [56]. These data captured both tidal variations and seasonal fluctuations. The model employed a 5 min reporting time step, selected to maintain numerical stability while ensuring temporal resolution of pollutant movement and hydraulic responses.

2.3.2. BOD Simulation and Pollutant Loading

BOD transport and transformation were simulated based on direct pollutant loading from identified sources and in-stream decay processes, following standard PCSWMM water quality modeling procedures [68,69,70]. The model setup and parameterization were defined as follows:

• Point sources, including municipal outfalls, industrial discharges, and agricultural wastewater, were represented as direct loads using location-specific data obtained from the Department of Industrial Works and municipal wastewater survey reports [19,58].
• Non-point BOD loads were incorporated indirectly by calibrating initial estimates against observed monitoring data. When simulated BOD was consistently lower than observations, additional diffuse loads were added and distributed across sub-catchments until model outputs aligned with measured concentrations.
• In-stream BOD decay was represented using a first-order kinetic formulation, which assumes that the degradation rate is proportional to the existing organic concentration. This formulation is widely applied in riverine and urban water-quality models because of its computational efficiency and its ability to approximate microbial oxidation and biodegradation processes occurring during transport [71,72]. The governing equation is expressed as follows:

C_t = C₀ e^{−k t}

(2)

where C_t is the BOD at time (t), C₀ is the initial concentration, and k is the decay coefficient (day⁻¹), typically ranging from 0.1 to 0.3 depending on temperature, flow velocity, and land use [69,73]. Calibration was performed separately for dry and wet seasons to capture seasonal variability in decay rates and loading behavior.

Initial decay coefficients for each land use type (residential, industrial, and agricultural) were selected based on these published ranges and are summarized in

Table 1. BOD decay parameters by land use type.

Land Use Type	Decay Coefficient
Residential	0.18 day−1
Industrial	0.22 day−1
Agriculture	0.12 day−1

. Final parameter values were refined through model calibration to ensure consistency between simulated and observed BOD levels along the river system. Calibration was performed separately for the dry and wet seasons to account for seasonal variations in temperature, flow velocity, and pollutant loading patterns.

The selected BOD decay coefficients were chosen within ranges widely reported for large lowland rivers and estuarine systems in contemporary water quality modeling studies, reflecting basin-average behavior rather than site-specific reaction kinetics (e.g., [74,75]). Consistent with the planning-level objective of this study, the decay coefficient was treated as a basin-scale parameter to support robust comparative evaluation of alternative management scenarios.

Both exploratory testing and published sensitivity analyses in water quality modeling indicate that variation in first-order decay coefficients primarily influences the magnitude of simulated BOD concentrations, while having a comparatively limited effect on relative spatial patterns and scenario-to-scenario contrasts at the basin scale. Sensitivity assessments have shown that although degradation and decay parameters are influential, model responses in comparative applications are more strongly governed by differences in source loading, hydraulic residence time, and dilution processes. This behavior is consistent with findings from systematic parameter sensitivity studies, which demonstrate that moderate uncertainty in decay rate selection does not substantially alter the relative performance ranking of management scenarios [76]. Accordingly, the primary conclusions regarding scenario effectiveness in this study are considered robust to plausible uncertainty in decay coefficient specification.

2.3.3. Estuarine Influence and Tidal Dynamics

Tidal boundary conditions were defined based on observed water level and BOD dynamics in the Chao Phraya and Tha Chin estuaries. As illustrated in Figure 3, tidal fluctuations exert a substantial influence on river hydrodynamics and pollutant behavior, particularly in terms of pollutant accumulation and dispersion across the tidal cycle.

In the Chao Phraya Estuary (Figure 3a), a sharp decrease in water level during early tidal phases corresponds to a pronounced BOD peak exceeding 6 mg/L, followed by a gradual decline. This trend indicates pollutant retention under low-flow conditions, likely caused by limited dilution and stagnation resulting from tidal backwater effects [77]. Conversely, the Tha Chin Estuary (Figure 3b) exhibits a more gradual water level rise, accompanied by a moderate BOD increase [78]. This suggests improved flushing and dilution of organic matter under tidal inflow conditions.

The PCSWMM simulation engine was able to represent bidirectional flow behavior induced by tidal forcing, capturing periods of pollutant accumulation during flood tides and enhanced downstream transport during ebb tide conditions. These dynamics are important for characterizing longitudinal variations in organic matter transport and residence time within the estuarine and nearshore reaches of the Upper Gulf of Thailand.

The estuarine reaches were represented using a one-dimensional dynamic wave formulation under the assumption of longitudinally well-mixed conditions. This simplification is appropriate for the present planning-level analysis, which focuses on first-order effects of tidal backwater, flow reversal, and hydraulic residence time on BOD accumulation and transport. While this approach does not explicitly resolve vertical stratification or lateral gradients characteristic of partially mixed estuaries, it enables consistent basin-scale comparison of management scenarios under limited data availability. More detailed two- or three-dimensional hydrodynamic–water quality models would be required to explicitly represent stratification and complex estuarine processes; however, such applications are more suitable for design-oriented or process-based investigations beyond the scope of this study.

2.4. Model Calibration and Validation

Model calibration and validation were conducted to assess the performance and reliability of BOD simulations across the Chao Phraya and Tha Chin River systems. The process was conducted in two distinct phases:

• Calibration phase: January 2022—December 2022;
• Validation phase: January 2021—December 2021.

Both phases utilized observed BOD concentration data collected from 36 water quality monitoring stations, strategically distributed across upstream, midstream, and downstream locations in each river system. The calibration focused on adjusting hydrologic parameters (e.g., Manning’s coefficients and flow routing time) and BOD-specific parameters (e.g., BOD initial, decay rates), with the objective of minimizing the difference between simulated and observed values.

Model performance was evaluated using standard statistical metrics, including Nash–Sutcliffe Efficiency (NSE), Mean Absolute Error (MAE), and Percent Bias (PBIAS), to ensure that the calibrated parameters could be reliably applied for predictive scenario analyses.

2.5. Scenario Development and Analysis

To evaluate long-term riverine water quality trajectories and assess the effectiveness of alternative management options, a set of predictive scenarios was developed using the calibrated PCSWMM. The scenarios were constructed to reflect plausible future conditions in terms of land use change, population growth, and pollution control efforts. All assumptions were aligned with the policy directions and intervention pathways outlined in the 10-year national surface water quality improvement plan of the Pollution Control Department (PCD), which establishes strategic targets for BOD reduction and wastewater management by 2032.

Scenario Design

To assess projected water quality conditions under differing levels of management intervention, three distinct scenarios were formulated for each target year. These include:

• Baseline Scenario: Assumes no additional interventions beyond existing policies, with projected demographic and wastewater growth.
• Moderate Reduction (MR-20%): Envisions modest improvements through expanded treatment coverage and partial implementation of pollution control measures.
• Enhanced Reduction (ER-30%): Represents comprehensive mitigation efforts, aligned with national water quality goals, involving full-scale infrastructure deployment and strict enforcement mechanisms.

Each scenario was applied basin-wide, incorporating corresponding adjustments in pollutant generation rates, treatment infrastructure, and regulatory enforcement assumptions. A summary of key assumptions and strategic actions under each scenario is presented in

Table 2. Assumptions and Strategic Actions under Different Pollution Management Scenarios.

Scenario	Key Assumptions	Actions/Interventions	Reference Guidelines
1. Baseline	- No changes in land use or wastewater treatment practices	- Status quo maintained	[19,58]
	- Population growth at 1.2% annually	- Existing infrastructure only
	- No new industrial treatment plants
2. Moderate Reduction (MR-20%)	- 20% BOD load reduction	- Decentralized wastewater systems in peri-urban areas	[19]
	- Targeting all major sources: domestic, industrial, agricultural	- Enforcement of effluent standards
3. Enhanced Reduction (ER-30%)	- 30% BOD load reduction	- Expansion of centralized wastewater treatment (38% → 60%)	[19,77]
	- In line with National Strategy targets	- Green infrastructure (wetlands, bioswales) promotion

.

3. Results and Discussion

3.1. Calibration and Validation Results

BOD concentrations in the Chao Phraya and Tha Chin River systems were simulated and calibrated using water quality monitoring data from 2022. The calibration results indicated that the model captured observed BOD trends with reasonable accuracy. In the Chao Phraya River, the model reproduced BOD trends in the upper and middle reaches reasonably well, although it slightly overestimated concentrations in the downstream section, achieving a Nash–Sutcliffe efficiency of 0.72, a mean absolute error of 0.35 mg/L, and a percent bias of +18% (Figure 4a). By contrast, the Tha Chin River showed consistently strong agreement between observed and simulated BOD values across all reaches, with a Nash–Sutcliffe efficiency of 0.88, a mean absolute error of 0.12 mg/L, and a percent bias of +4% (Figure 4b), indicating higher calibration performance [78].

Validation with 2021 data further confirmed the model’s robustness and transferability (Figure 5). In the Chao Phraya River (Figure 5a), the model adequately replicated BOD distribution patterns along all river sections, although minor overestimations were observed in the downstream segment. The performance was quantitatively evaluated, yielding a Nash–Sutcliffe Efficiency (NSE) of 0.76, a Mean Absolute Error (MAE) of 0.30 mg/L, and a Percent Bias (PBIAS) of +12%, indicating acceptable predictive accuracy for scenario-based simulations. In contrast, the Tha Chin River (Figure 5b) exhibited superior model performance, with a high degree of correspondence between observed and simulated BOD values across the entire river profile. The statistical metrics NSE = 0.91, MAE = 0.10 mg/L, and PBIAS = +2% demonstrate excellent predictive capability with minimal deviation. These validation outcomes reinforce the model’s applicability for water quality forecasting, particularly under varying management scenarios [78].

These results demonstrate the model’s ability to simulate spatial and temporal variability in BOD under diverse hydrological conditions. The comparatively higher accuracy in the Tha Chin River likely reflects more uniform land use and discharge patterns, whereas the downstream complexity of the Chao Phraya River, driven by tidal influence, multiple discharge sources, and canal inflows, may have contributed to localized overprediction. Despite these challenges, the model performance is comparable to that reported in previous large-river water quality studies and is considered appropriate for scenario-based forecasting and regulatory screening applications.

Consistent with findings from published parameter sensitivity analyses in basin-scale water quality models, variation in decay coefficients primarily affects the magnitude of simulated concentrations, while relative spatial patterns and scenario-to-scenario contrasts are governed more strongly by differences in pollutant loading and hydrodynamic retention. Accordingly, the comparative evaluation of management scenarios presented in this study is considered robust to plausible uncertainty in decay coefficient specification [76].

3.2. Scenario-Based BOD Simulation and Implications

Three scenarios, including Baseline, Moderate Reduction (MR-20%), and Enhanced Reduction (ER-30%), were simulated to assess projected BOD conditions in the Chao Phraya and Tha Chin Rivers by 2032.

Under baseline conditions, which assume no changes in land use, infrastructure, or regulatory enforcement, BOD concentrations exceeded national thresholds in multiple river segments. In the Chao Phraya River, concentrations ranged from 1.94 mg/L in the upper section to 5.44 mg/L downstream, surpassing the 4.0 mg/L limit, which is the maximum allowable concentration under Thailand’s Surface Water Quality Standard for Class 4 surface water and represents the upper threshold for acceptable water quality in the country. (Figure 6a). The Tha Chin River exhibited similar degradation, with exceedances in the upper and middle sections, although the lower reach remained within acceptable limits (Figure 6b). Spatial mapping (Figure 7a) revealed widespread non-compliance, particularly in urban-industrial corridors, driven by untreated discharges and limited treatment capacity [79].

The MR-20% scenario, which incorporated decentralized treatment and moderate source control. In this scenario, decentralized treatment refers to introducing small-scale or community-level wastewater treatment systems in areas not connected to central WWTPs, allowing a portion of domestic wastewater to be partially treated before entering the river. Moderate source control refers to implementing practical, medium-intensity measures that reduce pollution at the source, such as improving industrial pretreatment, better management of livestock and agricultural runoff, and reducing organic loads from urban areas without major infrastructure expansion. This resulted in moderate improvements. BOD levels in the Tha Chin River fell within regulatory limits (0.81–1.54 mg/L), but exceedances persisted in the lower Chao Phraya, indicating that modest load reductions are insufficient for basin-wide compliance (Figure 7b).

In contrast, the ER-30% scenario, combining an expansion of centralized wastewater treatment coverage from 38% to 60% with the introduction of nature-based infrastructure such as constructed wetlands and bioswales, assumed a corresponding reduction in BOD loads at the source, driven by improved treatment efficiency and enhanced pollutant retention prior to discharge. Under this integrated approach, system-wide compliance was achieved. Simulated BOD concentrations ranged from 1.47 to 3.93 mg/L in the Chao Phraya River and 0.71 to 1.35 mg/L in the Tha Chin River (Figure 6), with all river segments remaining below national water-quality standards (water surface standard class 4). Spatial outputs (Figure 7c) further illustrated widespread improvement, reflecting substantial reductions in both point and non-point loads across the basin. Collectively, these measures increased assimilative capacity, lowered stream pollutant retention, and provided additional hydrological buffering particularly in tidally influenced estuarine zones prone to backflow and contaminant accumulation.

Cumulative exceedance analysis reinforced the spatial trends. Under baseline conditions, BOD concentrations exceeded the 2.0 mg/L (For water surface standard class 3) threshold over approximately 245 km of river length. This was reduced to 156 km under MR-20% and further to 84 km under ER-30%, underscoring the basin-wide effectiveness of integrated infrastructure and green interventions. These results provide quantitative spatial evidence for prioritizing wastewater infrastructure and regulatory planning under continued urbanization.

Projections further suggest that, without intervention, organic loading to the Upper Gulf of Thailand may increase by 18%, significantly raising the risk of eutrophication, hypoxia, and algal blooms, especially near Samut Sakhon and the Chao Phraya delta. The ER-30% scenario would mitigate this risk, reducing coastal BOD loading by approximately 25%, in line with national marine water quality targets. These findings emphasize the urgency of integrated basin-to-coast governance, sector-specific pollution source control, and climate-resilient planning to sustain water quality and ecological health in both riverine and coastal systems.

3.3. Spatial and Temporal Patterns of BOD Concentration

Model simulations revealed distinct spatial gradients in BOD concentrations along both river systems (Figure 6). Upstream reaches, such as those in Uthai Thani, maintained BOD levels below 2.0 mg/L under baseline conditions, primarily due to forest cover, limited development, and relatively high flow velocities that promoted effective dilution [57,80]. In midstream areas, including Ayutthaya and Pathum Thani, concentrations ranged from 2.1 to 2.9 mg/L, reflecting increasing urbanization, agro-industrial inputs, and insufficient treatment infrastructure.

The most severe exceedances occurred in the downstream sections, notably near Samut Sakhon and the Tha Chin estuary, where projected BOD levels reached 5.44 mg/L and 3.93 mg/L, respectively, under the 2032 baseline. These zones are characterized by high pollutant loads, low assimilative capacity, and tidal backflow that limits flushing.

Seasonal variability was also evident, with broader exceedance observed during the dry season due to reduced flow and diminished dilution potential. Overall, model results aligned well with historical monitoring trends, confirming the framework’s ability to capture spatial and temporal pollutant behavior under diverse hydrological conditions (Figure 7a).

3.4. Eutrophication Risk in the Upper Gulf of Thailand

Downstream transport of BOD has direct implications for coastal water quality, particularly in estuarine and nearshore zones [81]. Although BOD is not a direct measure of nutrient enrichment or primary productivity, elevated BOD levels reflect increased organic matter loading and associated oxygen demand, which can exacerbate dissolved oxygen depletion and create conditions conducive to eutrophication-related stress. In this study, BOD is therefore applied as an indicative proxy of eutrophication potential, rather than as a direct representation of nutrient-driven processes, in light of data limitations and the planning-level scope of the analysis.

Under the Baseline 2032 scenario, modeled increases in organic loading to the Upper Gulf of Thailand suggest a heightened risk of oxygen depletion and eutrophication-related impacts, with coastal BOD inputs projected to increase by approximately 18%. These results indicate a potential increase in vulnerability to hypoxia and algal bloom development, particularly in areas where hydrodynamic conditions limit flushing and dilution. The most pronounced risk is observed near Samut Sakhon and the Chao Phraya delta, where tidal backflow, low flushing capacity, and cumulative upstream discharges converge [82,83].

In contrast, the ER-30% scenario reduces coastal BOD loading by approximately 25%, suggesting a corresponding reduction in oxygen stress and eutrophication susceptibility under this management pathway. While the present modeling framework does not explicitly simulate nutrient cycling, phytoplankton dynamics, or algal biomass, the relative reduction in organic loading under the ER-30% scenario indicates a lower likelihood of eutrophication-favorable conditions compared to the baseline case. These findings are consistent with the objectives of Thailand’s Marine and Coastal Resources Strategy (2032) [84] and highlight the importance of coordinated basin–coastal management. Future studies incorporating explicit nitrogen and phosphorus dynamics, as well as biological response indicators, would be required to quantitatively assess eutrophication processes and ecosystem responses in greater detail.

4. Conclusions

A physically based and spatially explicit modeling framework implemented using PCSWMM demonstrates robust capability for simulating basin-scale biochemical oxygen demand (BOD) transport along the river–estuary continuum. Through the integration of land use classification, hydrodynamic flow routing, aggregated point and non-point source loading, and infrastructure-based management scenarios, the framework enables systematic identification of spatial pollution hotspots and quantitative evaluation of BOD exceedance risk under alternative future conditions.

Scenario-based analysis indicates that both the spatial extent and magnitude of BOD exceedance are controlled not only by pollutant loading intensity, but also by basin hydrological response and wastewater infrastructure configuration. Downstream segments of the Chao Phraya and Tha Chin Rivers exhibit elevated vulnerability due to reduced dilution capacity, tidal backwater influence, and cumulative urban–industrial effluent inputs. Among the evaluated management options, the ER-30% scenario illustrates that basin-wide compliance with water-quality objectives can only be approached through an integrated management strategy combining expansion of centralized wastewater treatment systems with decentralized and nature-based interventions. These results indicate that uniform load reduction targets are insufficient and that effective water-quality management requires spatially differentiated strategies reflecting local land use characteristics, hydraulic conditions, and dominant pollution sources.

The present investigation is formulated as a planning-level, proof-of-concept application intended to support preliminary screening of management alternatives under data-limited conditions. Biochemical oxygen demand is applied as a proxy indicator representing organic loading and associated oxygen consumption. Extension of the framework to explicitly simulate dissolved oxygen dynamics, nutrient species including nitrogen and phosphorus, chlorophyll-a concentrations, and additional contaminants would enable more direct assessment of eutrophication processes and ecosystem response.

Future research may further enhance the framework through tighter coupling with higher-dimensional hydrodynamic and water-quality models to resolve vertical stratification, sediment–water exchange, and complex estuarine processes, particularly in design-oriented or process-focused studies. Incorporation of climate change scenarios, both independently and in conjunction with land use change, would also improve long-term planning relevance. Despite these potential extensions, the results confirm that the proposed framework constitutes a practical and transferable tool for basin-scale water-quality planning and management assessment, with applicability to other river systems discharging into the Upper Gulf of Thailand.