Anthropogenic River Segmentation Case Study: Bahlui River from Romania

Abstract

This manuscript introduces a river segmentation method and explores the impact of human interventions through a long-term study of total nitrogen, total phosphorus, chemical oxygen demand, and biochemical oxygen demand. An indicator linking parameter concentrations to the river’s flow rate was used to assess the development of the examined parameters. The analysis spanned from 2011 to 2022, considering both seasonal and yearly variations. Normal probability plots served as statistical tools to evaluate whether the data followed normal distributions and identify outliers. The proposed segmentation divided the Bahlui River into four segments, each defined by anthropogenic stressors. It was found that, due to human activity, each river segment could be viewed as an “independent” river. This supports the idea that river segments can be analyzed separately as distinct components. The proposed segmentation approach represents an alternative approach in river water quality research, moving from traditional continuous system models to fragmented system analysis, which better reflects the reality of heavily modified river systems. The study’s findings are important for understanding how anthropogenic modifications affect river ecosystem functioning in the long term.

1. Introduction

Building dams, regardless of their purpose, will inevitably affect in-stream ecological processes, as well as downstream natural hydrology and hydrochemistry. According to a 2016 study by Van Cappellen and Maavara, more than half of the world’s streams and rivers are crossed by dams, and they estimated that by 2030, this figure could increase to 90% [1]. A few years later, in 2020, 1.2 million barriers, including weirs, dams, barrages, sluice gates, ramps, bed sills, culverts, and fords, were counted only for Europe’s rivers and streams [2]. Globally, the number of free-flowing rivers is steadily decreasing due to land use, climate change, and increasing water and hydropower demands [3]. River damming creates physical barriers that fragment rivers and obstruct the natural movement of matter and energy [4,5,6]. The presence of multiple dams on a single river stream or its main tributaries leads to river fragmentation and, consequently, to consistent flow alteration [7,8]. Numerous studies discuss how dams affect the hydrology and hydrochemistry of rivers [4,6,7,8,9,10,11,12,13,14,15]. The hydro-chemical regime is greatly affected by natural fluctuations in river flow, such as seasonal changes, and by human activities that typically involve discharge control [8]. Unfortunately, dam construction is not the only anthropogenic stressor on rivers; urban runoff, agricultural practices, WWTP plants, and industrial discharges all significantly contribute to changes in nutrient loads and COD and BOD dynamics [16,17,18,19,20]. A free-flowing river has a natural flow regime that allows for the continuous transport of nutrients, organic carbon, and sediments [3]. In contrast, an anthropogenically stressed (fragmented) river [16] has transport discontinuities caused by abrupt changes in flow patterns, sediment retention [21], organic matter accumulation, nutrient cycling [14,22], WWTP discharge [23], and so on.

Damming’s main effect on nutrients and organic matter is to increase their hydraulic retention time (HRT). The longer the HRT time, the higher the chance of biogeochemical and physical transformations that lead to nutrient consumption [11]. Barriers affect water quality monitoring by creating spatial differences in COD and BOD measurements. Due to pollutant buildup or flow changes, upstream areas may exhibit different features than downstream areas [20]. Due to fragmentation, different river segments process organic matter and nutrients in various ways. The different segments of a river may experience different levels and types of pollution because of land use, urbanization, and industrial activities, and each segment may have unique hydrological features that affect its biogeochemical cycling [7]. Therefore, analyzing independent river segments can show how fragmentation affects nutrient transport, COD and BOD dynamics, and other factors. Such an approach, which involves examining individual segments and comparing them with the entire river system, might provide a more accurate assessment of human impacts. There are very few works that assume that each segment of a fragmented river can behave independently [24,25,26], even though concepts like river continuum [27] and serial discontinuity [28] are not new and have been thoroughly defined in the literature. Dam-fragmented rivers are thoroughly examined throughout their whole length, and various analysis have been made regarding sediment transport [14,15,21,29,30,31,32,33], nutrient fluxes [1,4,14,15,33,34,35], water quality [20,36,37,38], and ecosystem impact [28,39,40].

With 17 dams in its hydrographic basin, the Bahlui River is one of Romania’s most hydro-technically engineered rivers. Our previous study highlighted the effectiveness of these dams in managing streamflow and reducing the impacts of droughts and floods [41]. However, we could not find a statistical relationship between the river’s nutrient loads and discharge. The data did not support the initial hypothesis that there would be at least a seasonal association between river discharge and nutrient flow, except for a few weak correlations in specific sampling areas. In this work, an equation that links river flow rate to nutrient loads and two relevant water quality chemical indicators (COD and BOD) was employed for this new approach of examining river segments as “independent” watercourses. This study considered the following four river segments separated by a series of anthropogenic stressors: Pârcovaci Dam, Tansa Dam, Podu Iloaiei Dam, and the Iasi Waste Water Treatment Plant (WWTP).

The following aspects underline the novelty of this work:

-

Unlike previous studies that focus on water quality and typically examine the Bahlui River as a unified hydrological system, this work introduces a novel conceptual framework that treats river segments as individual watercourses.

-

Anthropogenic stressors delimit the analyzed river segments. This segmentation approach allows for understanding how human interventions create localized impacts that do not necessarily propagate throughout the entire river system.

-

For each segment, water quality indicators are linked with the river’s discharge in an integrated flow–nutrient–pollution analysis.

-

The seasonal and annual variations across each segment are presented for twelve years, offering a detailed view of how fragmentation affects long-term water quality patterns.

This research contributes to a better understanding of anthropogenic impacts by providing a comprehensive viewpoint on the transport of nutrients and dynamics of COD and BOD in a fragmented river. Examining the seasonal and yearly changes in nutrients and organic pollutants in separate river segments offers specific benefits, especially in understanding hydrological effects and the ecological impacts of river fragmentation.

2. Materials and Methods

2.1. Research Area, Anthropogenic Amendments, and Segmentation Strategy

2.1.1. Research Area

The main characteristics of the Bahlui River are well described in the literature [41,42,43,44,45,46]. In brief, the Bahlui River, which is 119 km long, has a catchment area of 2025 km² and a hydrographical network of 3100 km. It is the main tributary of the Jijia River. The primary water source of the Bahlui River is runoff from precipitation; in the upstream area, some small tributaries are fed by groundwater. Therefore, as is common for rain-fed rivers, the river’s discharge varies significantly throughout the year (seasonal variation) [42]. The annual precipitation in the river’s basin is approximately 500 mm. Therefore, it is important to note that only 30% of the hydrographic network has a permanent flow [42,45]. The average discharge ranges from 2.8 m³/s to 4 m³/s, but it can reach up to 600 m³/s (with the highest recorded during floods in 1932) and fall to complete depletion during extended drought periods [47]. To manage these significant fluctuations in discharge flow, often accompanied by heavy floods, 17 dams have been constructed in the Bahlui River basin over the past century. More information about the river’s tributaries and the hydrotechnical infrastructure in the Bahlui River system can be found in our previous work [41].

2.1.2. Anthropogenic Amendments

The hydrographic basin of the Bahlui River has experienced significant human interference over the past century. The 17 dams built between 1964 and 1988 provide hydrotechnical management for about 70% of the river basin [41,48]. Between 2010 and 2013, during the European Project “Water-course regulation of Bahlui River, County of Iaşi”, the riverbed was regularized for a length of approximately 11 km, which crosses the city of Iaşi [43]. Furthermore, over recent decades, the possibility of implementing new projects related to the development of recreation areas along the urban section of the Bahlui River has been heavily debated [49]. Therefore, the Bahlui River can be considered as one of the most anthropized rivers in Romania.

Table 1. The reservoirs considered in this study, with some constructive details.

Reservoir Name	Coordinates	Watercourse	Dam Elevation, m	Canopy Length, m	Normal Storage Capacity, m3∙106	Water Surface, m2∙104	Year of Completion
Accumulation Pârcovaci (A1)	47°27′19″ N 26°48′47″ E	Bahlui	25	290	2.75	48	1984
Tansa Lake (A2)	47°17′30″ N 27°4′49″ E	Bahlui	14.2	4890	6.79	289	1974
Accumulation Podu Iloaiei (A3)	47.196842° N 27.191161° E	Bahlueț	14.1	640	3.699	240.9	1964

provides detailed information about the anthropogenic barriers directly related to this study.

The Pârcovaci accumulation (denoted A1) is situated on the main course of the river, in the upper part, in a highly forested hilly area. From a geological point of view, the primary materials are clays with intercalations of sands, sandstones, and limestones [50]. The Pârcovaci dam (Figure S1) is an earth dam built between 1978 and 1984 and put into operation in 1984 [51]. It is classified as a homogeneous earth embankment; the construction materials are fine, dusty, alluvial clayey, and clayey deluvials [52]. Initially, the main functions of this accumulation were the flood protection of downstream localities and the water supply of the town of Hârlău. The average discharge of the Pârcovaci reservoir is 0.424 m³/s [42]. Since 2000, according to Law no. 5 of 6 March 2000 regarding the approval of the National Territorial Planning Plan-Section III-protected areas [53], modified by the Emergency Ordinance no. 49 of 31 August 2016 [54], the Pârcovaci accumulation has been a protected area for several species of local ichthyofauna.

The Tansa dam (denoted A2) is classified as a homogeneous embankment made from local materials (clay soils) [52]. The Tansa Lake (Figure S2) sedimentary blanket is composed of a layer of brown and grey clay, underlain by Sarmatian deposits consisting of alternating marls and fine sands [55]. The Tansa dam was built on the Bahlui River between 1971 and 1976, becoming operational in 1975. The lake is a multi-purpose reservoir, being used for flood protection, as drinking water supply for the Belceşti village, as an irrigation source for farms in the area, and for fish farming [56]. The average discharge of the Tansa dam is 0.839 m³/s [42].

The Podu Iloaiei Accumulation Lake (denoted A3 and presented in Figure S3) is located on the Bahlueț River, which is the main tributary of the Bahlui River, at a distance of 2.5 km upstream of the confluence with the Bahlui River and 400 m upstream from the town of Podu Iloaiei [57]. The dam has operated since 1964 and is made of local materials, predominantly yellow dusty clays of a homogeneous type [52]. Alluvial soils, carbonated alluviums, soils rich in sodium sulphates, meadow clayey, maroon substrates, and gray soils represent the hydrographic basin [57]. The main functions of Podu Iloaiei Dam Lake include (i) flood protection by controlling the Bahlueț River water flow rate; (ii) supplying the water volume required for fish farming; and (iii) acting as a water source for irrigating an agricultural area of 526∙10⁴ m² [58]. The lake water is evacuated in the winter time [57,59], hence the area and volume of the lake can vary considerably throughout the year. The average discharge of the Podu Iloaiei dam ranges from 0 (in winter) to 1.02 m³/s [58].

According to the Romanian Government Decision no. 663 of 14 September 2016 on establishing the regimes of protected natural areas and designating special avifauna protection areas as an integral part of the European ecological network Natura 2000 in Romania, the Sârca-Podu Iloaiei accumulation was declared a special avifauna protection area [60].

Additionally, on the fourth section of the river, there is also the Iași Waste Water Treatment Plant (WWTP), which significantly impacts the river discharge after crossing the Iași municipal area. The Iași WWTP (Figure S4) has been developed in stages since 1968 [61]. It is located in the Holboca commune, in the area of the Dancu district, and uses the Bahlui River as effluent. The WWTP daily discharge varies between 2.2 m³/s and 4.033 m³/s [62] and can reach up to 8 m³/s and above during heavy rains [61]. Since the daily amount of urban wastewater remains relatively constant—mainly depending on the number of residents and not on rain or groundwater sources—the daily discharge capacity of the Iasi WWTP is medium-high and stays fairly steady compared to the river’s flow rate. During heavy rain, when urban runoff is also collected, the WWTP’s daily discharge capacity doubles. In some situations, especially during extended droughts, the Iasi WWTP’s maximum discharge capacity can surpass the average flow rate of the Bahlui River [47].

2.1.3. Segmentation Strategy

The accuracy of the study heavily depends on segment selection, but identifying relevant local boundaries is not always easy. For example, Zhang et al. divided the Shaying River into the following three segments: upstream, middle stream, and downstream [26]. Shehab et al. used a Digital Elevation Model to identify the boundaries required to divide the Tigris River into eight zones [63]. Nardini et al. described four approaches for river segmentation, including (i) manual segmentation based on expert judgement; (ii) segmentation based on image recognition using artificial intelligence and machine learning algorithms; (iii) segmentation based on statistical analysis; and (iv) segmentation based on logical or heuristic algorithms [64]. An equivalent for the term “segment” is the concept of “reach”, described and critically analyzed by Parker et al. [65]. A reach is commonly defined as a convenient subdivision, which may be any length of river with fairly uniform characteristics, the length between gauging stations, or the length of a watercourse between any two defined points [66].

In this study, river segmentation was based on expert judgment, considering the main anthropogenic stressors (previously described), how they interfere with the river’s natural course, and land use. Manual dam-based segmentation establishes clearly defined boundaries that match actual physical and hydrological breaks in the river system. Unlike arbitrary geometric divisions, dams cause real discontinuities in flow, sediment transport, and water quality. This method aligns segmentation with the physical processes that govern river behavior.

Accordingly, four reaches were designated for further analysis (Figure 1,

Table 2. Brief characterization of the Bahlui River selected segments.

Segment No.		1st Segment	2nd Segment	3rd Segment	4th Segment
Range		S1–S2	S2–S3	S3–S4	S4–S6
Length (km)		11.5	37.5	40	30
Anthropogenic elements		A1	A1, A2	A2, A3	A3, A4
Reach stream	inlet	Natural	Direct dam controlled	Direct dam controlled	Indirect dam controlled
	outlet	Direct dam controlled	Direct dam controlled	Indirect dam controlled	Natural, influenced by the Iasi WWTP discharge
Land use		Forest, aquaculture	Vineyards, orchards, agriculture, and pastures	Agriculture, aquaculture, pastures, animal farms	Agriculture, highly urbanized area

). Due to anthropogenic impact, each segment can be treated separately as an “independent” river. The selected river segments and the decision support (the form of human control) for each choice are briefly presented in the following paragraphs.

In Figure 1, A1–A4 represent the considered anthropogenic modifications, as follows: A1 = Pârcovaci Dam; A2 = Tansa Dam; A3 = Podu Iloaiei Dam; and A4 = Iasi WWTP. The green dots, S1 and S6, represent the river’s spring and the river’s end at the confluence with the Jijia River, resepectively. The red dots S2–S5 indicate the sampling locations.

River Spring-Pârcovaci Dam (S1–S2). The first reach is between the river spring (S1) and sampling site S2, including the Pârcovaci accumulation (Figure 1), and features a largely undisturbed riparian ecosystem. The anthropogenic element of control is the Pârcovaci dam, which manages the river’s discharge and supplies inflow to the second segment (S2–S3). This entire reach is located in a forested area with minimal to no agricultural or urban disturbances. While aquaculture may influence the nutrient levels, this is well managed, as the Pârcovaci accumulation has been designated as a protected area.

Pârcovaci Dam-Tansa Dam (S2–S3). The stream flow in the second segment is fully controlled, with the inlet by the Pârcovaci dam and the outlet by the Tansa dam. There is a relatively low urban impact through the Hârlău WWTP that services the small town of Hârlău, which is crossed by the Bahlui River. The rural land use includes vineyards, orchards, non-irrigated arable fields, and pastures [67].

Tansa Dam-Podu Iloaiei Dam (S3–S4). The Tansa dam discharge directly controls the entrance to the third reach, while the exit is indirectly regulated by the Podu Iloaiei dam, which manages the flow rate of the Bahlueț River, the main tributary of Bahlui. Sampling site S4 is located downstream of the river’s confluence. Before entering the Podu Iloaiei accumulation, the Bahlueț River passes through two towns, Târgu Frumos and Podu Iloaiei. Land use in this section of the Bahlui River primarily consists of pastures and arable fields, both irrigated and non-irrigated. Aquaculture, along with some pig and poultry farms (which operate periodically), are potential sources of nutrients and organic matter. Both the Belcesti WWTP and Podu Iloaiei WWTP utilize this river segment, directly or indirectly, as their effluent source.

Podu Iloaiei Dam-Iasi WWTP (S4–S6). The Podu Iloaiei dam indirectly controls the inlet flow of the fourth river segment. Since the dam is drained during winter, the river flow rate varies significantly, impacting the transport of nutrients and organic matter. Sampling site S5 is located downstream of the Iasi WWTP. The outlet flow is heavily influenced by the Iasi WWTP, which handles municipal wastewater and urban runoff during rainfall. This last segment of the river passes through a relatively small agricultural area, a region of pastures, and the rapidly expanding metropolitan area of Iasi.

2.2. Sampling Sites, Monitored Parameters, Analyzed Period

The four sampling sites, denoted S2, S3, S4, and S5 in Figure 1, were extensively described in our previous work [41]. Water samples were collected according to a specific seasonal or annual schedule. The National Administration of Romanian Waters provided all the data analyzed during this study. The monitored parameters included the river flow rate and the nutrient flow indicated by total nitrogen, total phosphorus, and the level of organic pollution, as shown by chemical oxygen demand (COD) and biochemical oxygen demand (BOD). The analyzed period ranged from 2011 to 2022 for the S2, S4, and S5 sampling sites and from 2015 to 2022 for S3.

2.3. River Load Calculation Approach Methods: Theoretical Background, Selected Equation, Statistical Analysis

A river’s “instantaneous load” with a particular constituent is commonly defined as the result of the water discharge and the concentration of that constituent as they pass through the stream cross section [68]. The total amount and/or the flow rate of a particular component (pollutant, nutrient) that is transported by the river in a given period can be calculated by knowing its “instantaneous load.” These definitions can be expressed using Equation (1) [68,69,70], as follows:

$$F_i\cong {\int }_{t_i}^{t_f}{L}_i\left(t\right)·dt =k·{\int }_{t_i}^{t_f}{C}_i\left(t\right)·Q\left(t\right)·dt$$

(1)

where F_i is the amount/flow rate of the component “I” transported during the monitored time interval; t_i and t_f represent the beginning and ending times of the measurements; C_i represents the concentration of the “i” component at the time t, while Q_i represents the river’s discharge at the time t; and the constant k is used to adjust the measurement units (when necessary).

Naturally, the length of the monitored period, the sampling frequency, and accuracy are critical to the precision of the load estimation [71]. There are additional factors that can influence the accuracy of the calculation; some of these are fairly predictable, like the succession of seasons with rainy and dry periods, while others are less or not at all predictable, like heavy rains accompanied by floods or prolonged drought periods, which are frequently associated with climate change phenomena [72]. In addition to these natural interferences, human activity can also have a direct impact on the river loads through damming, aquaculture, discharges from WWTPs, land use, and other actions [7,8,9,16,41,73,74].

Over the years, several methods have been developed to estimate the contaminant loads in rivers and solve Equation (1) [69,75,76]. In a recent study, Zhang and coworkers [76] identified the following three categories of methods for estimating river loads: (i) interpolation algorithms, which use interpolations between measured concentrations; (ii) ratio methods, using the ratio of the average load to the average flow; and (iii) regression methods, which establish a regression relationship between concentration and flow to estimate loads.

In this work, the method of weighted averages [77,78] was applied to estimate the annual discharge of nutrients, as well as COD and BOD fluctuations. The approach falls under the category of interpolation algorithms, and Equation (2) describes the yearly load of a particular component.

$$F_i=K_x·\overline{Q_i}·{\displaystyle \frac{\sum _{i=1}^n{C}_i·Q_i}{\sum _{i=1}^n{Q}_i}}$$

(2)

where the constant K_x is used to correlate the measure units and the considered period, e.g., for a year, K_x considers a value of 366 days, for a month, 30 days, and for a season, 91 days.

To visually represent the data and assess normal distributions, normal probability plots (NPPs) were used. When such graphs display a nearly straight line, it is relatively safe to assume that the data follow a normal distribution. Additionally, it is easy to visually identify the presence of skewness, outliers, or other nonlinear anomalies that suggest unexpected events. In addition to the visual inspection, NPPs provide specific parameters such as the p-value, mean, and standard deviation, with the p-value being the most commonly used in actual statistical analysis. To determine whether the data are normally distributed, a significance level of 0.05 is used. A p-value less than or equal to this significance level suggests that the data do not follow a normal distribution. Conversely, a p-value greater than the significance level indicates a high probability that the data do follow a normal distribution [79].

The calculations were performed using Microsoft Excel, while the NPPs were generated using the Minitab^® Statistical Software Version 21.2 (Minitab, LLC, 2022, State College, PA, USA). The overall workflow of the steps performed in the current work is presented in Figure 2.

3. Results and Discussion

3.1. Fragmentation Impact on Nutrient Dynamics and Organic Pollution Load: Seasonal Variation

Seasonal monitoring can reveal short-term changes in nutrient levels and spikes in organic pollution caused by factors such as weather events, agricultural runoff, and accidental spills. These may be missed in annual reports due to data averaging. Additionally, seasonal analysis helps to distinguish between natural cycles and human impacts, providing a comprehensive view of water quality dynamics.

3.1.1. River Spring-Pârcovaci Dam (S1–S2)

The first river segment (Figure S5) runs through a forested area; the main controlling factor is the discharge from the Pârcovaci dam. The main tributaries along this part of the river are Bahluețul Mic, Valea Mare, and Valea Cetațuiei.

The presence of nitrogen and phosphorus in this initial segment is attributed to natural sources. This aligns with the findings in the literature, where Duchesne et al. studied how nutrients transfer seasonally from forest canopy leaves to lake water [80].

Table 3. Nutrient dynamics for the first river segment.

S1–S2	NT (t/Season)				PT (t/Season)
Year	1. Winter	2. Spring	3. Summer	4. Autumn	1. Winter	2. Spring	3. Summer	4. Autumn
2011	4.06	0.12	4.00	3.81	0.14	0.00	0.27	2.16
2012	-	0.27	0.15	5.03	-	0.03	0.01	0.52
2013	7.31	1.59	2.23	0.52	1.47	0.02	0.05	0.03
2014	3.69	0.20	0.39	0.31	0.08	0.04	0.04	0.01
2015	0.24	0.20	0.19	1.32	0.01	0.01	0.01	0.05
2016	0.13	2.29	0.43	0.31	0.01	0.19	0.04	0.03
2017	1.02	0.71	0.62	0.63	0.25	0.07	0.01	0.02
2018	0.61	0.38	0.39	0.22	0.04	0.13	0.06	0.13
2019	0.26	0.23	0.79	0.29	0.19	0.08	0.64	0.04
2020	0.35	0.32	0.33	0.36	0.09	0.05	0.08	0.06
2021	0.30	-	6.15	0.14	0.04	-	0.20	0.01
2022	0.14	0.15	0.29	0.15	0.01	0.01	0.06	0.00

presents the seasonal evolution of total nitrogen (N_T) and total phosphorus (P_T) for the considered interval (2011–2022). The seasons are represented by numerical designations and their corresponding names to maintain the standard sequence of seasons in all tables and figures. The river annually carries an average of 4.47 tons of N_T. This amount varies greatly each year, ranging from 0.74 tons to 12 tons. Notable extreme loads are recorded in 2011 and 2013, affecting the overall average. The large standard deviation of 3.86 tons indicates that this segment’s total nitrogen is mainly affected by a few high-load years or seasons.

The average annual P_T load is 0.63 tons. The variability is significant, ranging from 0.08 tons to 2.5 tons, with a standard deviation of 0.74 tons, which is above the mean. This variation is affected by notable seasonal events, including the heavy autumn load in 2011 (2.16 tons) and the high winter load in 2013 (1.47 tons).

The NPP for nutrient dynamics in the River Spring-Pârcovaci dam segment is shown in Figure 3. Across all seasons, the p-values and NPP indicate that the N_T does not follow a normal distribution. The highest and most variable measurements are recorded in winter, while the lowest and most consistent levels occur in spring. The outliers correspond to N_T values from 2011, 2013, and 2014, which are linked to the drought period from 2011 to 2013 [41].

For P_T, the calculated values are several orders of magnitude lower than those for N_T, and the NPP exhibits different seasonal trends, with autumn showing the highest average load and the greatest variability. The fall peak illustrates a “first flush” phenomenon, where phosphorus bound to soil particles, stored during the dry summer, is released into the river by the first substantial autumn rainfall. This suggests that, for this section, soil erosion is a primary pathway for phosphorus delivery.

The p-value for spring is slightly above the 0.05 threshold, suggesting that the data mostly conforms to the normal distribution. This small load contradicts several studies that predict a surge after spring snowmelt; however, it could indicate a phase of significant biological uptake or resource depletion following winter runoff [81]. The outliers observed in all seasons except spring correspond to the same 2011–2013 drought period when N_T values are also recorded.

Table 4. Organic pollution load for the first river segment.

S1–S2	COD (t/Season)				BOD (t/Season)
Year	1. Winter	2. Spring	3. Summer	4. Autumn	1. Winter	2. Spring	3. Summer	4. Autumn
2011	83.38	4.12	99.90	84.46	32.81	1.55	49.75	34.68
2012	-	5.92	8.00	321.17		2.09	2.96	115.76
2013	200.52	12.55	17.95	10.91	57.35	2.92	5.38	2.43
2014	204.48	17.19	15.29	12.32	38.01	2.84	2.21	0.78
2015	3.66	4.85	8.88	32.20	0.57	1.01	1.25	5.96
2016	6.77	208.91	11.49	14.55	2.19	46.53	3.69	4.57
2017	26.62	16.81	14.63	14.75	8.47	5.38	4.46	5.15
2018	5.37	9.30	12.00	6.02	1.69	3.68	4.35	2.16
2019	4.38	10.94	67.21	10.72	1.51	3.59	23.56	3.47
2020	11.22	13.82	11.28	14.64	3.52	4.48	4.82	6.93
2021	10.29	-	220.26	11.81	4.82	-	105.65	5.98
2022	15.65	13.73	13.23	13.69	4.94	2.11	1.86	1.65

presents the seasonal evolution of COD and BOD for the considered interval (2011–2022). The examination of the COD data emphasizes the event-driven characteristic of this indicator. The mean yearly load is 47.7 tons, and the standard deviation is significantly higher at 75.7 tons. This indicates that this segment’s total COD budget is governed by several years of significantly elevated values, rather than a typical “average” year. The highest yearly loads occur during intense seasonal events, specifically in 2012 and 2014. The average annual BOD load is 13.0 tons, with a standard deviation of 18.9 tons. Similar to COD, the variability surpasses the mean, demonstrating that outlier events impact the annual load.

The NPP showing the trends observed during each season is presented in Figure 4. In terms of COD, although winter has the highest average load, each season has experienced at least one major pollution event, leading to substantial standard deviations that often exceed the seasonal average. Winter also displays the highest average BOD load, with a standard deviation greater than the mean, indicating a high variability.

3.1.2. Pârcovaci Dam-Tansa Dam (S2–S3)

A landscape featuring vineyards, orchards, arable (non-irrigated) fields, and grasslands is crossed by the second segment of the river (Figure S6). The Hârlău WWTP, situated at the start of this section, has a relatively minor urban impact. The flow from the Pârcovaci dam controls the river’s intake, while the flow from the Tansa dam manages the river’s outflow. The main tributaries along this part of the river are Buhalnița, Magura, and Vulpoiul.

Table 5. Seasonal evolution of N_T and P_T for the S2–S3 sector.

S2–S3	NT (t/Season)				PT (t/Season)
Year	1. Winter	2. Spring	3. Summer	4. Autumn	1. Winter	2. Spring	3. Summer	4. Autumn
2015	78.77	4.96	0.20	1.87	0.74	0.09	0.02	0.04
2016	2.22	0.46	13.89	43.28	0.03	0.01	3.27	11.77
2017	2.87	38.48	1.06	1.32	0.04	1.86	0.07	0.18
2018	8.14	4.46	3.06	27.03	0.16	0.24	0.19	3.90
2019	6.03	6.71	4.76	4.22	0.30	0.39	0.28	0.28
2020	9.83	9.51	5.40	7.13	0.15	0.17	0.29	0.45
2021	12.25	16.54	21.80	1.82	0.21	0.86	4.69	0.36
2022	13.45	3.31	1.99	12.51	0.24	3.31	0.62	0.94

presents the seasonal evolution of N_T and P_T for the considered interval. For N_T, it can be observed that there is extreme inter-annual variability. For example, in 2015, the annual value is dominated by a single large winter event, whereas other years tend to be more evenly distributed or feature multiple seasonal peaks. The fact that most of the annual load in some years occurs in just one season underscores the significance of episodic events like heavy rain, snowmelt, or pulses of agricultural runoff. Similarly, for P_T, there are highly variable annual loads, mainly driven by single events.

When considering the S2–S3 vs. S1–S2 segments, the ratio between nutrient values in similar seasons is bigger than 1, ranging from 1.04 in summer 2015 to 626.97 in autumn 2022. There are only a few exceptions: spring 2016 for N_T and autumn 2015, spring 2016, winter 2017, and summer 2019 for P_T.

Figure 5 shows the NPP for nutrients in the S2–S3 sector. In all cases, the p-value is less than 0.05, indicating that the data does not follow a normal distribution. Additionally, compared to the mean, the standard deviations are higher in all cases, reflecting a high variability. Similar to the S1–S2 sector, for N_T, the highest means are observed in winter, while for P_T, they occur in autumn.

Table 6. Organic pollution load for the S2–S3 sector.

S2–S3	COD (t/Season)				BOD (t/Season)
Year	1. Winter	2. Spring	3. Summer	4. Autumn	1. Winter	2. Spring	3. Summer	4. Autumn
2015	289.55	35.97	3.47	3.67	26.50	2.44	1.22	2.22
2016	3.08	2.46	341.28	732.34	0.98	0.88	125.31	257.05
2017	9.72	333.19	9.05	18.67	2.94	113.55	2.66	6.44
2018	12.26	29.30	15.36	532.69	4.38	9.75	4.71	198.62
2019	18.44	63.82	41.60	36.04	5.92	20.63	12.78	12.70
2020	29.63	18.94	38.60	92.18	13.32	8.33	19.31	40.88
2021	37.08	363.15	194.21	59.25	15.36	141.95	75.91	27.60
2022	36.38	49.47	53.08	183.19	16.99	23.63	20.07	53.25

presents the seasonal trends of COD and BOD over the period from 2015 to 2022. The ratio of organic pollution between similar periods in S2–S3 and S1–S2 is greater than 1, ranging from 1.01 in Spring 2015 (for BOD) to 220.26 in summer 2021 (for COD). There is only one exception: winter 201r for BOD, when the ratio is 0.57. This suggests that a “self-purification” process occurred, with natural conditions supporting the growth of microorganisms that degrade organic matter.

Similar to the variation observed in the S1–S2 sector, some years show seasonal sharp increases in both COD and BOD, suggesting that episodic events impact them. This pattern is also visible in the NPP (Figure 6). For COD, the lowest average value occurs in winter, while the highest is in autumn. However, across all seasons, the standard deviation exceeds the mean, indicating that outliers significantly influence the data mean. For BOD, the only case where the p-value is greater than 0.05 is during winter, suggesting that in this case, the data follows a normal distribution. Aside from winter, the seasonal patterns observed for COD are similar to those for BOD. This indicates that the river’s organic pollution mainly comes from sources that contribute both biodegradable and non-biodegradable matter, which are influenced by temperature and flow. The difference in winter highlights the greater sensitivity of BOD to temperature-driven biological processes.

3.1.3. Tansa Dam-Podu Iloaiei Dam (S3–S4)

The third segment’s land use mainly includes arable fields and pastures. Here, the river is joined by its main tributary, the Bahluieț River. Since the Bahluieț River flows through two cities, Târgu Frumos and Podu Iloaiei, it indirectly contributes to an urban influence. The Tansa dam directly regulates the flow in this section of the river, while the outflow is indirectly managed by the Podu Iloaiei dam, located on the Bahlueț River. The confluence of the two rivers occurs downstream of the Podu Iloaiei dam (S4 point in Figure S7).

Table 7. Seasonal evolution of N_T and P_T for the S3–S4 sector.

S3–S4	NT (t/Season)				PT (t/Season)
Year	1. Winter	2. Spring	3. Summer	4. Autumn	1. Winter	2. Spring	3. Summer	4. Autumn
2011	40.37	44.56	4.50	18.02	2.82	6.95	1.96	0.60
2012	9.58	11.18	2.34		0.48	1.09	0.22
2013	8.34	23.07	11.84	24.36	0.52	4.52	2.13	2.66
2014	19.57	28.68	2.89		1.12	7.50	0.17
2015	56.92	42.83	1.77	16.32	18.91	4.48	0.19	1.39
2016	4.86	17.47	1.71	11.57	0.19	1.14	0.18	1.57
2017	3.77	0.94	4.38		0.11	0.11	0.17	-
2018	9.94	7.49	5.95	24.35	0.50	1.76	0.52	8.42
2019	13.73	30.61	7.34	10.85	0.75	3.69	1.52	1.32
2020	11.36	7.51	4.42	14.90	0.39	0.50	0.44	1.65
2021	9.45	19.11	11.63	9.96	0.20	1.45	2.60	0.83
2022	11.02	2.31	0.55	1.77	0.20	0.32	0.12	0.23

shows the seasonal changes in total nitrogen and total phosphorus over the studied period (2011–2022). Compared with similar seasons in the previous segment, there is some uneven variation, which supports the idea that the river segments evolve independently. Since data from 2011 to 2015 are missing for the second segment, the comparison only includes the period from 2015 to 2022. For N_T, the ratio of reported values during the same seasons for the S3–S4 and S2–S3 segments ranges from 0.02 in spring 2017 to 38.33 in spring 2016. For P_T, the variation ranges from 0.05 in summer 2016 to 78.84 in spring 2016.

The NPP of nutrient distribution for the third river segment is presented in Figure 7. For N_T, unlike the two previous sectors, the standard deviation is lower than the mean, indicating that although there is high variability in the data, the influence of outliers on the mean is lower. The lowest mean occurs in summer, while the highest occurs in spring. Except for in winter, the p-value is greater than 0.05, indicating that the data follow a normal distribution. For P_T, the data shows a normal distribution only for spring. Distinctively from N_T, but in line with the previous sectors, the standard deviation is higher than the mean, indicating that there are still high yearly variations per season, underlining once again the impact of natural external factors on water quality.

Table 8. Seasonal evolution of COD and BOD for the S3–S4 sector.

S3–S4	COD (t/Season)				BOD (t/Season)
Year	1. Winter	2. Spring	3. Summer	4. Autumn	1. Winter	2. Spring	3. Summer	4. Autumn
2011	151.02	513.27	111.29	138.87	74.63	166.96	37.49	49.50
2012	59.60	190.84	33.63		20.84	61.09	10.86
2013	39.05	371.92	346.83	587.53	13.71	227.03	79.08	76.77
2014	132.12	795.45	25.63		18.33	145.08	5.48
2015	533.40	534.64	58.89	237.90	101.11	135.66	13.17	38.98
2016	62.94	229.82	15.47	192.64	21.58	84.24	8.84	56.53
2017	19.66	14.96	32.45		6.33	4.54	9.73
2018	29.15	60.16	70.32	485.28	9.38	20.59	24.12	186.86
2019	75.08	505.94	91.69	145.75	24.52	183.80	31.56	48.42
2020	78.79	56.73	47.05	176.99	31.65	18.95	25.74	83.96
2021	37.31	89.44	222.22	132.05	16.73	31.23	99.22	57.58
2022	37.67	25.81	7.58	37.61	15.57	3.36	0.94	12.03

presents the seasonal evolution of COD and BOD for the considered interval (2011–2022). Similar to N_T and P_T, the ratio of seasonal values for segments S3–S4 and S2–S3 fluctuates unevenly for COD and BOD. The range for COD is between 93.36 in spring 2017 and 0.04 in spring 2017, while for BOD, it ranges from 96.03 in spring 2016 to 0.04 in spring 2017. When the results for two consecutive seasons fluctuate this much, it suggests an event that directly affects the river’s organic pollution.

The NPP for the organic pollution load in the Tansa dam-Podu Iloaiei dam section is presented in Figure 8. For both COD and BOD, the data follow a normal distribution only in the spring, indicating that although there are variations between the years, in this case, they follow a normal pattern.

3.1.4. Podu Iloaiei Dam-Iași WWTP (S4–S6)

Although it passes through a grassland and agricultural area, the final section is significantly impacted by the 14 km it spends crossing the metropolitan area of Iaşi and the city’s WWTP. Sampling site S5 is located about 7 km from the Bahlui confluence with Jijia (the point S6 in Figure S8), immediately past the Iaşi WWTP. The dam at Podu Iloaiei indirectly controls the flow at the entrance to this last segment of the river, while the outflow is directly influenced by the WWTP of the city of Iasi, whose contribution is significant [82]. This final section of the river has a relatively high number of both temporary and permanent tributaries compared to the other three segments considered (Figure 1).

Table 9. Seasonal evolution of N_T and P_T for the S4–S6 sector.

S4–S6	NT (t/Season)				PT (t/Season)
Year	1. Winter	2. Spring	3. Summer	4. Autumn	1. Winter	2. Spring	3. Summer	4. Autumn
2011	268.70	259.70	162.87	230.16	26.32	30.72	23.11	26.08
2012	255.05	222.14	244.12	183.81	24.56	25.18	22.15	22.69
2013	284.19	286.47	240.36	322.27	12.10	31.79	27.53	29.98
2014	295.53	529.08	259.07	181.62	22.52	81.12	30.11	27.87
2015	370.32	238.23	81.92	134.19	69.44	33.20	25.14	16.40
2016	127.05	112.67	104.73	459.00	9.05	5.55	7.29	49.78
2017	108.10	205.96	83.69	168.73	4.90	14.48	12.50	14.83
2018	223.27	99.93	201.34	170.96	29.51	17.32	16.22	36.25
2019	301.61	324.64	76.03	170.95	31.72	25.54	14.68	20.28
2020	133.81	189.62	96.81	99.72	22.97	23.07	12.15	10.91
2021	157.09	142.16	144.69	210.75	12.70	18.16	19.24	13.26
2022	181.09	138.39	147.26	214.04	18.93	14.48	24.19	17.39

shows the seasonal changes in total nitrogen and total phosphorus for the period from 2011 to 2022.

Table 10. Seasonal evolution of COD and BOD for the S4–S6 sector.

S4–S6	COD (t/Season)				BOD (t/Season)
Year	1. Winter	2. Spring	3. Summer	4. Autumn	1. Winter	2. Spring	3. Summer	4. Autumn
2011	1218.74	1238.94	727.90	834.59	386.56	494.74	260.81	263.42
2012	283.39	327.71	381.10	275.71	117.13	152.66	135.13	102.24
2013	379.54	1407.91	1312.56	801.92	132.01	719.68	209.81	99.93
2014	708.19	5593.43	1138.84	1144.13	192.95	1047.26	43.23	191.13
2015	1015.08	1921.31	291.88	651.33	182.52	307.39	55.60	115.85
2016	437.05	298.12	385.25	4244.60	144.52	110.95	131.08	1456.77
2017	306.77	1779.65	336.86	1273.19	105.55	450.54	105.57	436.52
2018	595.75	990.31	805.10	797.08	191.22	314.28	285.33	310.17
2019	1007.70	1680.02	483.68	1304.99	342.61	561.07	155.59	437.19
2020	456.08	721.17	411.66	424.75	151.45	236.07	174.40	200.75
2021	439.19	555.15	870.17	657.01	189.72	214.98	361.30	294.44
2022	410.50	473.75	609.00	376.06	192.24	74.54	127.28	45.63

displays significantly higher values compared to those reported for other parts of the river. Compared to the results for the same seasons in the previous segment (autumn 2018), the N_T values are at least seven times higher, while the P_T values are at least four times higher (autumn 2018). In the summer of 2022, these ratios reach their peak: about 210 for P_T and roughly 266 for N_T. These numbers demonstrate how urbanization influences nutrient levels. The many tributaries passing through agricultural areas could also contribute to the increases in total nitrogen and phosphorus levels.

The NPP for nutrient variation in the last river segment is shown in Figure 9. Except for N_T in winter, the standard deviation is smaller than the mean, indicating that seasonal events have a lesser impact on the mean compared to the other sectors. This suggests that the presence of the WTTP in this last segment tends to lessen outliers in the data. For N_T, the lowest mean occurs in summer and the highest in spring; however, the difference between them is much smaller than in the other sectors. Except for winter, N_T data follow a normal distribution. For P_T, the normal distribution holds true for summer and autumn. The smallest mean occurs in summer and the highest in spring.

Table 10 presents the seasonal evolution of COD and BOD in the considered interval (2011–2022) for the last segment of the river. In comparison to the other sections, the final river length has substantially higher COD and BOD levels (as well as N_T and P_T values).

The NPP for the organic pollution load in the last river section is presented in Figure 10. The data follows a normal distribution for summer in the case of COD and for spring and summer in the case BOD. While the standard deviations for P_T and N_T are lower than the mean in most cases, this is not true for COD and BOD. Additionally, some years show extreme outliers, such as spring 2014 and autumn 2017.

3.2. Fragmentation Impact on Nutrient Dynamics and Organic Pollution Load: Annual Report

The administration frequently requires annual reports, as they provide a comprehensive overview of the evolution of water quality, allowing for the identification of long-term causes such as land use changes or climate change, as well as minor variations or persistent trends. However, annual mediation can hide important seasonal or episodic events, such as pollution spikes or short-term ecological instabilities.

The annual trends of NT and PT in the considered period (2011–2022) for each river segment are shown in Figure 11. Because the values for the S4–S6 sector are much higher than the others, they were plotted on the secondary axis. Unlike the seasonal analysis, which revealed the presence of relatively high disturbances through outliers, the annual analysis tends to hide these high instabilities, especially in the NPP (Figure 12).

Although the data from Figure 10a suggests that the values are very close, this is because of the large differences between the total values recorded for each segment. Between S1–S2 and S4–S6, the order of magnitude for some of the considered years exceeds 100. The obtained p-value indicates that in the NT case, the average annual values follow a normal distribution for S3–S4 and S4–S6. For PT, only the S1–S2 segment shows data that do not follow a normal distribution.

Figure 13 presents the annual evolution of COD and BOD in the considered interval (2011–2022) for each river section. As can be observed, the yearly variation does not follow the same trend for all sectors, highlighting the fact that each sector has a different response regarding the analyzed pollutants.

The statistical analysis in this case (Figure 14) indicates that the data do not follow a normal distribution for COD only for the S1–S2 segment. This is due to the fact that there is a high order of magnitude between the values recorded in the first segment and the last segment. For all the data, the standard deviation is smaller than the mean. In contrast with the seasonal analysis, it can be observed that the outliers are hidden and that the standard deviation values are lower than the averages.

3.3. River Segmentation by Dams: Some Benefits and Drawbacks

The immediate result of anthropogenic interference on a river’s course, such as dam construction, is an increase in flood safety. On the other hand, these interferences always result in severe alterations to riparian ecosystems. Dams are massive structures that, together with the reservoirs created by their construction, irreversibly change the landscape. The natural flow of rivers, which is crucial for transporting nutrients like carbon, phosphorus, and nitrogen, is disrupted by fragmentation. Dams can trap these nutrients in reservoirs, decreasing their availability downstream. This retention modifies nutrient dynamics and can lead to hotspots of nutrient buildup in stagnant water bodies, while areas downstream may experience nutrient depletion. The main impact of damming on nutrients and organic matter is an increase in their hydraulic retention time. The longer the HRT, the greater the chance for biogeochemical and physical changes that result in nutrient consumption [11].

Dams store water during high-flow periods and release it during dry spells, ensuring a steady flow downstream. Because flow regulation maintains water supply even during droughts, it lowers the chance of depletion and protects aquatic habitats and biodiversity. Before dams were used to regulate and manage the course of the Bahlui River, complete depletion was a relatively common event, often reported during extended droughts [47].

Over time, the buildup of sediments and the natural tendency of ecosystems to reach equilibrium enable the formation of distinct habitats along each section of the river. In the case of Bahlui River, there are two examples of such habitats, as follows: the Pârcovaci dam, which serves as a protected area for local ichthyofauna, and the Sârca-Podu Iloaiei accumulation, designated as a protected area for avifauna. A mimetic effect can be observed (keeping proportions) between dams built by beavers [83] and those built by humans [84,85]. Both affect the river hydrology, geomorphology, biogeochemistry, and ecosystems, and in both cases, nature manages to restore equilibrium if it has enough time. If human influence were limited only to the construction of dams and control of water flow, this natural balance would be easier to achieve. Unfortunately, a variety of human activities, including deforestation, urbanization, and agriculture, have a significant and continuous impact on river courses (segmented or continuous) and their natural dynamics.

Overall, the analysis conducted in this study showed that the yearly and seasonal variation between the considered segments does not follow a predetermined trend. Therefore, the assumption that the river segments can be analyzed individually is valid. This is due to the presence of dams (as extensively discussed in previous paragraphs), but also because of urban areas (with WTTPs) that influence both flow and nutrient load. Furthermore, unlike continuous streams, the impact of unforeseen events, whether natural (such as severe floods) or anthropogenic (such as WWTP discharges), is limited to the specific river segment and does not affect the entire river.

4. Conclusions

This study proposed a segmentation strategy for the Bahlui River and examined the fluctuations in seasonal and annual total nutrient and organic pollution for each segment. Each river segment separated by dams and anthropogenic stressors behaves as an independent unit with distinct seasonal and yearly variations in total nitrogen, total phosphorus, chemical oxygen demand, and biochemical oxygen demand.

The Bahlui River analysis shows that dam-separated segments can be analyzed as independent watercourses, as demonstrated by the following:

• Localized impacts remain contained within specific segments rather than propagating throughout the entire river system. Statistical analyses reveal that nutrient and pollution data often do not follow typical patterns and vary greatly, highlighting the importance of considering occasional events and specific river sections when evaluating water quality. The standard deviations frequently surpass the means, indicating that extreme events have a significant influence on the overall nutrient and pollution budgets.
• Each segment exhibits different seasonal patterns and response mechanisms to anthropogenic and environmental stressors. The upstream area, mostly forested, indicates that natural nutrient sources vary seasonally, with autumn having the highest phosphorus levels due to “first flush” events and winter showing increased nitrogen levels. This baseline understanding is vital for distinguishing between natural and anthropogenic contributions in downstream sections. Furthermore, the analysis shows that urban development and wastewater treatment plant discharges elevate nutrient and organic pollution levels, highlighting significant human impacts on water quality.
• Unforeseen events (floods and WWTP discharges) affect only specific segments without impacting the entire river.

The research demonstrates that seasonal and annual analyses of the changes in nutrients and organic pollutants in different river segments provide specific benefits, especially in understanding hydrological effects and the ecological impact of river fragmentation. The statistical analysis confirms that segments behave independently, with varying seasonal patterns and response mechanisms to anthropogenic stressors. This segmented approach offers an alternative to traditional methods that treat the river as a continuous whole.