Calibration and Validation of the BMWP Index for the Assessment of Fluvial Systems in High Andean Mining Areas of Peru

Abstract

The High Andean region of Peru, characterized by a complex orography, has unique and highly biodiverse ecosystems. This region has several headwater basins that play a critical role in the hydrological cycle, providing diverse ecosystem services essential to sustain biodiversity and supply water to human communities. Despite the importance of this region, it faces significant human intervention, particularly mining activities, which affect basin headwaters and jeopardize water security. This study aimed to calibrate the Biological Monitoring Working Party (BMWP) index to evaluate water quality in High Andean rivers in Peru affected by mining activities, using aquatic macroinvertebrates as bioindicators. We used a 15-year dataset (2008 to 2023) from three headwater basins in the High Andean region; this dataset included physicochemical water quality parameters, trace metals, and aquatic macroinvertebrates. The BMWP was calibrated for the High Andean region of Peru with this dataset (BMWP/PeIAZIM); afterward, it was validated to assess water quality in an area influenced by mining activities in this region. The results allowed us to differentiate between aquatic macroinvertebrate families tolerant to mining pollution and highly sensitive families. The sites heavily affected by mining activity returned very low BMWP/PeIAZIM scores; sites with no mining impact had the highest scores. These findings indicate that the calibrated index can be used for water resource management in the High Andean region, contributing to the conservation of its ecosystems.

1. Introduction

High Andean regions are characterized by their altitude, from 1982 to 4082 m a.s.l., associated with the complex physiography of the Andean Cordillera. This territory is located in the westernmost area of South America and is one of the largest mountain ranges worldwide [1]. The topographic complexity of this region influences wind circulation, climatic conditions, and the hydrological regime, causing significant environmental heterogeneity and high biodiversity, for which it has been considered a biodiversity hotspot [2,3]. These High Andean areas include the headwaters of several basins, which play a central role in water regulation, as they are essential to supply water to local natural communities and sustain their high biodiversity [4]. Furthermore, streams associated with High Andean wetlands flow through environments with extreme conditions, highly variable temperatures, low atmospheric pressure, intense solar radiation, and low productivity [5]. In Peru, High Andean basins have a vast and varied geography and a complex diversity of natural resources and ecosystems, which provide several ecosystem services [6]. These basins function as ecological corridors that house a conglomerate of biotic and abiotic resources with complex interactions, traveling through hydrographic streams to their mouths [7].

This High Andean region has faced intense anthropogenic alterations, including mining activities, particularly gold mining, which has severely affected aquatic ecosystems [8]. Large-scale mining, initially restricted to southern Peru, has spread northward to regions with high mineral potential concentrated in the Andean territory, causing considerable damage to biodiversity and natural resources [8]. Guerra and Lozano [9] demonstrated how gold mining is linked to water pollution, especially in basin headwaters, affecting the quality and amount of water resources and jeopardizing the water security for vulnerable communities in areas of direct influence. Indigenous ancestral knowledge highlights the conceptualization of Peruvian basin headwaters as high-altitude zones that receive water input from rainfall, hail, groundwater, and especially from wetlands locally known as “bofedales”, which play a vital role in storing water resources and distributing them during the dry season. These highly vulnerable ecosystems must be protected due to their fragility and critical role in the hydrological cycle and water conservation [10].

The use of bioindicators to evaluate aquatic ecosystems has become relevant worldwide [11]. Aquatic bioindicators include macroinvertebrates, i.e., invertebrates larger than 0.2 mm (worms, crustaceans, and insects); they are critical components of aquatic ecosystems that play a central role in the food chain and organic matter breakdown [12]. Macroinvertebrates are highly specialized organisms adapted to very specific environmental conditions. They are frequently considered important biological indicators of the ecological integrity of aquatic systems because they are sensitive to various anthropogenic alterations, both physicochemical and hydromorphological. The absence of certain species due to such alterations, the increased abundance of tolerant species, or the lower taxonomic and functional diversity are usually reliable indicators of the loss of environmental quality [13].

Several studies have been conducted around the world using benthic macroinvertebrates as water quality indicators. One of the first studies proposed the saprobic index [14], which set the basis for the development of other widely used indices. The Biological Monitoring Working Party (BMWP) index was established in the United Kingdom in 1970 as a simple and rapid tool to assess water quality using macroinvertebrates as bioindicators. The BMWP is a biotic index based on bioindication values (which reflect pollution tolerance) of macroinvertebrate families in samples collected from rivers. The bioindication score ranges from 1 to 10, according to the tolerance to organic contamination of the different macroinvertebrate families. The most sensitive families are assigned a score of 10, while those most tolerant to pollution receive a score of 1 [15]; the total BMWP score is the sum of bioindication values for all families in a sample [15]. Other studies [16] have highlighted the benefits of using macroinvertebrates as water quality bioindicators. In the 1980s, Spain and Portugal adapted the BMWP to develop another index known as the IBMWP (Iberian Biomonitoring Working Party) [17]. However, this adaptation was considered an open proposal due to the limited taxonomic and ecological information on Mediterranean river macroinvertebrates available at the time [18]. More recent research studies have consolidated advances in this field [19,20,21] by providing feedback to the Water Framework Directive of the European Parliament (WFD 2000/60/EC: Directive 2000/60/EC), which uses, among other elements, benthic macroinvertebrates as bioindicators to guide the management of water resources.

In Latin America, several studies have been carried out to adapt the BMWP to particular regions, including Costa Rica, Colombia, Venezuela, and Chile [22,23,24,25]. These have been adaptations of proposals based on several rounds of questionnaires sent to a panel of experts on the bioindication values that should be assigned to different families of aquatic macroinvertebrates. Unlike those adapted indices, the calibration procedure proposed by Picos et al. [26] clearly establishes how bioindication values are obtained using a formal and repeatable mathematical method. These authors propose that bioindication values assigned to macroinvertebrate families derive from analyzing the abundances of macroinvertebrates exposed to different physicochemical quality conditions along a basin or in a region. Following this method, the BMWP has been calibrated to particular areas, such as the Apatlaco River (Mexico) [26] and surface streams in Panama [27]. In these studies, the calibration process used large datasets of physical and chemical variables matched with aquatic macroinvertebrate datasets, reflecting the tolerance of aquatic macroinvertebrates to their particular physicochemical conditions.

Aquatic macroinvertebrate assemblages in High Andean streams in Peru have been scarcely studied, contrasting with those that thrive at lower altitudes. Furthermore, it has been pointed out that the heterogeneous conditions in these streams produce differences in the richness and structure of macroinvertebrate assemblages, even more so when faced with anthropogenic activities such as open-pit mining [5]. Some studies have put forward proposals to bring the use of these biological indicators closer to environments where pollution is primarily chemical, such as research that evaluated the effects of mining on macroinvertebrate communities in the Sierra Gorda Biosphere Reserve (Mexico) [28] and Ciénaga Plaza Seca (Colombia) [29].

Peru lacks a calibrated and validated index to monitor water quality in its territory. For this reason, the studies carried out in Peruvian rivers have used BMWP indices from other countries, mainly IBMWP from Colombia, as is the case of the study on water quality conducted in the Ayacucho [5], Aguaytía [30], Cañipia [31], and Oxampampa [32] rivers. Medina [19] proposed a biotic index for rivers on the northern coast of Peru (nPeBMWP) at altitudes below 2000 m a.s.l. This index is an approximation of indices from neighboring countries with conditions similar to the Peruvian reality; it has been used in several studies [33,34,35,36] but does not take into consideration the mining influence.

Despite advances in the potential of macroinvertebrates as bioindicators in the High Andean region of Peru, there is no suitable index to assess the effects of mining on the ecological status of rivers in this region. There is no biotic index to assess water quality and support the management of these critical areas in the High Andean rivers for the protection, conservation, and sustainability of local ecosystems. In this sense, Picos et al. [26] highlighted the need to calibrate the index for each ecological region as there may be important differences in the taxonomic composition of the bioindicators used, the ecological, zoogeographical, and geological history, and the conditions associated with anthropogenic impacts. Therefore, the present study aimed to develop a BMWP version calibrated to environments influenced by mining activities following the method proposed by [26] and integrating heavy metals as evidence of pollution from mining activities in water bodies. We hypothesized that if mining pollution affects stream water quality, a biotic index such as BMWP should detect changes in water quality through variations in the presence/absence of the bioindicators used. The BMWP has been used assuming the presence of organic pollution [37]; however, its calibration incorporating the influence of mining operations can broaden its scope to other regions and basins where this condition occurs. Mining is considered one of the most polluting anthropogenic activities [38], affecting many streams worldwide.

This study aimed to calibrate the BMWP index (based on aquatic macroinvertebrates) for the headwaters of three High Andean basins located in mining areas of the La Libertad mountain range in northern Peru. For calibration, we included variables associated with organic matter and metal concentrations in water since the study focused on an area dedicated to open-pit gold and silver extraction since 2005. The study covered a period of 15 years (2008 to 2023). We hypothesized that the assessment of the BMWP calibrated for High Andean basins would differentiate between sites with mining impact, which would return low BMWP scores, and sites with no mining activities, which would produce high BMWP scores. This research addressed the following aspects: (1) calibrate the BMWP for High Andean mining environments using water quality variables such as factors associated with organic matter, metals as indicators of mining activity, and aquatic macroinvertebrates; (2) validate the BMWP index; and (3) compare the BMWP values of areas with and without mining influence to test the effectiveness of the calibrated BMWP index and identify impacts associated with mining activities on the ecological condition of the rivers studied. Finally, based on the results from the BMWP index developed here, we propose future directions for managing hydrological ecosystems affected by mining operations.

2. Materials and Methods

2.1. Study Area

The study zones are located in the upper portion of three basins that converge within a mountain system. Two basins, the Santa Basin (Caballo Moro study zone) and the Chicama Basin (Perejil study zone), discharge into the Pacific Ocean; the other, the Crisnejas Basin (Chuyugual study zone, belonging to the Amazon region), discharges toward the Atlantic Ocean. The study area comprises an ecosystem ranging between 1982 m and 4085 m a.s.l., characterized by undulating hills and rugged mountains with rocky outcrops and shallow soils covered by grasslands with remnant patches of native shrubland and forest [39]. The climate between 1000 m and 3000 m a.s.l. is temperate subhumid, with temperatures around 20 °C and precipitation between 500 mm/year and 1200 mm/year. At altitudes between 3000 m and 4000 m a.s.l., the climate is cold, with a mean precipitation of 700 mm/year and a mean temperature around 12 °C, with frost events in winter.

The Lagunas Norte mining company started operations in 2005 in the confluence area of the three basins mentioned above. It is dedicated to open-pit gold extraction using Merrill–Crowe heap leaching technology and columnar coal. It is currently managed by Boroo [40] (Figure 1).

The study included sites or stations without mining influence (PC) and sites impacted by mining (PP) for each study zone (Caballo Moro, Perejil, and Chuyugual zones) (Table S1). A total of 23 monitoring stations were established in these zones from 2008 to 2023 (Figure 1). Physical, chemical, and metal parameters were recorded at 19 monitoring stations (P20, P21, P22, P23, P24, P25, P1, P3, P4, P5, P6, P31, P33, P34, P35, P36, P37, P38, and P39); macroinvertebrate families were determined at 21 monitoring stations (P1, PS1, PS2, P3, P4, P5, and P6 for Perejil zone; P21, PS21, P22, P23, P24, and P25 for Caballo Moro zone; P31, P33, P34, P35, P36, P37, PS37, and P39 for Chuyugual zone).

2.1.1. Caballo Moro Zone

Seven monitoring stations (five PP and two PC) were located in this zone at altitudes from 4017 m to 4085 m a.s.l. In addition, occasional macroinvertebrate monitoring was carried out at one site (PS21). The altitude and characteristics of each monitoring station are presented in Table S1.

2.1.2. Perejil Zone

A total of seven monitoring sites (five PP and two PC) were located in this zone. Four (P1, P4, P5, and P6) remain active, while station P3 was decommissioned in 2011. Artisanal coal mining is carried out in this area. In addition, macroinvertebrate assessments were occasionally conducted at stations PS1 and PS2. The details of the study monitoring stations are shown in Table S2.

2.1.3. Chuyugual Zone

Nine monitoring sites (seven PP and two PC) were located in this zone. The Chuyugual zone belongs to the Crisnejas basin, which drains into and eventually becomes part of the Amazon basin. The characteristics of each station are detailed in Table S3.

2.2. Data Collection

This study is part of a community environmental monitoring and surveillance program in which members of the local communities (civil society), organized into environmental committees, monitor and ensure proper field data and sample collection by specialized academic staff.

The dataset corresponding to the period 2008 to 2023 contains physical, chemical (9), microbiological (2), metal (28), and biological (macroinvertebrate) parameters determined and evaluated with the participation of accredited laboratories, academic entities, civil society organizations grouped into surveillance committees, community environmental monitoring, and non-governmental organizations dedicated to environmental care and protection. This 15-year dataset was used to calibrate and validate a BMWP index to assess water quality in mining contexts in the High Andean region of northern Peru.

2.3. Analysis of Physical and Chemical Parameters

The physical and chemical parameters of water, such as pH, dissolved oxygen (DO, mg/L), electrical conductivity (EC, μS/cm), and temperature (T, °C), were recorded in situ using a YSI Professional Plus multiparameter probe. Water samples for subsequent metal and metalloid testing were collected following the EMTAL Project protocols of the Ministry of Energy and Mines [41], which were later replaced by the Peruvian Protocol for Quality Monitoring in Natural Surface Water Bodies (2011–2015), updated in 2016 with the Peruvian Protocol for Surface Water Resources Quality [42]. Samples were fixed with 1 mL of HNO₃ (1:1 N) to reach a pH between 1 and 2 and transported to the laboratory under cool and dark conditions.

All laboratory tests were carried out by laboratories accredited by INDECOPI (up to 2016) and INACAL (since 2017). Samples were analyzed using standardized Environmental Protection Agency procedures [43,44,45,46], as appropriate for each method. Laboratory testing of water samples included total and fecal coliforms (MPN) and chemical parameters (mg/L), including ammonium nitrogen, sulfates, nitrates, and biochemical oxygen demand (BOD₅). The metals measured (mg/L) were aluminum (Al), arsenic (As), boron (B), barium (Ba), calcium (Ca), iron (Fe), potassium (K), magnesium (Mg), manganese (Mn), sodium (Na), strontium (Sr), zinc (Zn), beryllium (Be), cadmium (Cd), cobalt (Co), copper (Cu), chromium (Cr), tin (Sn), mercury (Hg), molybdenum (Mo), nickel (Ni), silver (Ag), lead (Pb), thallium (Tl), titanium (Ti), vanadium (V), lithium (Li), and antimony (Sb).

Between 2005 and 2020, mercury traces were determined following the procedures outlined in EPA 621-C-99-004 [47], including atomic absorption spectrometry (AAS) after sample digestion. On the other hand, arsenic was analyzed using hydride generation coupled with atomic absorption spectrometry (GH-AAS), described in Section 3114 C of the Standard Methods [44].

The remaining metals and metalloids (from 2005 to 2021) were determined using inductively coupled plasma atomic emission spectrometry (ICP-AES), a procedure standardized by the EPA [48] (Method 200.7) and its 2007 and 2014 updates (Methods 6020A and 6020B, respectively) [49,50]. This procedure allows for the simultaneous detection of multiple elements with high sensitivity and precision after sample acid digestion and subsequent reading on a mass spectrometer, where elements are first classified according to their mass-to-charge ratios and then quantified.

Finally, from 2021 to 2023, all metals and metalloids were determined by inductively coupled plasma mass spectrometry (ICP-MS) using EPA Method 200.8 (Revision 5.4) [51]. The samples were previously subjected to acid digestion to ensure the release of the elements of interest.

For all metal quantifications, the accuracy and precision of the analytical results were controlled with a high-purity Standard Reference Material (SRM) (certified analyte solution) and a blank spiked with known concentrations of each analyte.

The limits of quantification (LoQs) of the methods used were always below the values of the National Environmental Quality Standards (Estándar de Calidad Ambiental; ECA, in Spanish) established by the Peruvian National Water Authority (Asociación Nacional de Agua; ANA, in Spanish).

2.4. Macroinvertebrate Sampling

Aquatic macroinvertebrates were collected according to multi-habitat protocols [23,52,53]. Multi-habitat sampling, which consists of recognizing and selecting identifiable predominant habitats, was performed.

Samples were collected using a 300 µm pore-size net, traversing the river until the entire sagittal section of the channel was covered, including the selected habitats: hard substrates, plant debris, sand, vegetation, submerged macrophytes, and fine sediments. Samples were poured onto white trays to visualize macroinvertebrate specimens, which were placed separately in airtight containers using forceps and droppers. Specimens were preserved by immersion in 70% alcohol spiked with glycerin drops. All containers were labeled with the monitoring station code, date, and time.

The taxonomic determination of macroinvertebrate specimens was carried out at the SAM 302 Invertebrate Zoology Laboratory of the Faculty of Biological Sciences, National University of Trujillo. Specimens were identified to family using taxonomic keys [13,23,54,55].

2.5. BMWP Calibration

The BMWP for High Andean rivers in Peru was calibrated following the protocol of Ruiz-Picos et al. [26], which was also used in a historical study of the Duero River in the Mexican Plateau [56]. Two major modifications were made to the protocol for this study. First, a time series with annual average data from 2008 to 2023 was used for each monitoring station in the study zones. The second modification consisted of incorporating heavy metal and metalloid concentrations as indicators of mining activity to generate the Physicochemical Quality Index.

Briefly, the modified protocol consists of four steps (Figure 2): (1) Determine the Physicochemical Quality Index (PQI), which describes water quality at each monitoring station using a 1-to-10 scale. The calculation includes the following: the curation of physical, chemical, and metal data for all stations during the study period; data are standardized (log x + 1) before applying factor analysis, in which parameters with a significant positive or negative correlation (p < 0.05) in the first two factors are considered qualifying variables. Once obtained, each qualifying variable is given a 0-to-1 score using maximum and minimum values established in the Peruvian regulations; then, the score assigned to each qualifying variable is re-scaled to values from 1 to 10 to match the scale used in the BMWP for macroinvertebrate families. All scores are averaged for each monitoring station to derive the PQI, which range from 1 to 10 classes. (2) Assign abundance classes to macroinvertebrate families in each station. Each macroinvertebrate family in each monitoring station is assigned to one of six classes based on its abundance, as follows: Class 0, no organisms recorded; Class 1, 1 to 3 organisms; Class 2, 4 to 10 organisms; Class 3, 11 to 33 organisms; Class 4, 34 to 100 organisms; and Class 5, more than 100 organisms. This step aims to reduce the effect of local dominance of some macroinvertebrate families. (3) Determine the bioindication value. The abundance classes were pooled within each PQI classes; if the abundance class of a family appears in more than one monitoring stations within the same PQI class, the abundance classes of such monitoring station are averaged to obtain a single abundance value per family for each PQI class. In this way, a distribution of abundance classes by family in PQI classes is obtained. The bioindication value is obtained from the fifth percentile of this distribution, see [26]. (4) Determine BMWP categories (water quality classes). The bioindication value is applied to the original macroinvertebrate matrix, and the BMWP is calculated by adding the bioindication values of each aquatic macroinvertebrate family at each monitoring station. Afterward, the BMWP scale is established using the median of the 5% monitoring stations with the highest BMWP values to make the first division of water quality categories (values equal to or higher than the median value are assigned to the Excellent water quality category). The second category (Good water quality) is established for BMWP scores ranging from below the median just mentioned to the tenth percentile of the distribution of the highest BMWP values. The other four categories (Fair, Poor, Poor/Highly Polluted, and Extremely Poor/Extremely Polluted water quality) were set by dividing the tenth percentile value by four.

2.6. BMWP Validation

The BMWP index validation process consisted of performing a multiple linear regression between the qualifying variables of the PQI and the BMWP, thereby obtaining a calculated BMWP value using the model. The observed values (BMWP derived from the calibration process) and the calculated values using the model were plotted to obtain the best-fit curve and 95% confidence intervals.

2.7. Statistical Analysis

Normality and homoscedasticity tests (Shapiro–Wilk and Levene, respectively) were performed. Analysis of variance (ANOVA) and post hoc multiple comparison tests (Tukey and Duncan) were applied in each study zone to identify significant differences between monitoring stations with and without mining influence. Statistical tests were performed using the XLSTAT software V. 2020.4.1.

3. Results and Discussion

The results obtained in this study reflect a mixed process that employed regulatory and validated techniques, as well as methodologies combined with community environmental surveillance and monitoring. The latter is an active and participatory process where the organized civil society observes, records, and responds to situations that may affect their environment or territory, specifically their water bodies. It can sometimes be used as an early warning tool for pollution [57]. The continuity of the participatory monitoring program allowed us to follow up on long-term monitoring, obtaining a database spanning more than 15 years [58].

3.1. Acidification Evidence

The first monitoring event was conducted in June 2005, one month before the Lagunas Norte mine started operations. The physicochemical results in the three zones studied showed that the characteristics of water, including the presence of fecal coliforms, were consistent with organic pollution derived from the daily activities of local inhabitants (extensive agriculture and small-scale livestock farming). Since the mining company started operations, changes in water quality have been observed. Station P39 (PP), in the Chuyugual zone, receives direct industrial wastewater discharges that affect downstream monitoring stations (P35 and P36, both PP). Water quality degradation was observed as an increase in the values of several environmental variables (Table S4).

At the monitoring station P21 (PP), in the Caballo Moro zone, the pH was 6.7 in 2005 and dropped to 3.7 in 2013. The Perejil zone also experienced acidification at station P4 (PC), with pH values of 5.7, 6.4, 6.0, 6.4, 4.86, and 6.47, outside of the water quality standards, in 2010, 2012, 2018, 2019, 2022, and 2023, respectively. Sulfates, mining byproducts that are part of acid mine drainage [59], were recorded each year at concentrations that generally exceeded Environmental Quality Standards (ECAs). In 2023, concentrations of 2079 mg/L (station P39), 2054 mg/L (P35), and 2068 mg/L (P36) were recorded, which exceeded the Environmental Quality Standards for water [60] by 902.8%, 821.6%, and 827.2% with respect to Category 1: A1 (Table S4). This increase was not observed at P31, a station not exposed to industrial discharges that showed adequate concentrations of this ion during all monitoring periods.

Some heavy metals also exceeded the values of the current Peruvian regulations [60,61]. In 2010, monitoring stations P36 and P39 in the Chuyugual zone showed arsenic concentrations of 0.03 mg/L, exceeding Category 1: A1. Furthermore, mercury exceeded the ECAs thresholds considering Category 1: A1, Category 3: D1/D2, and Category 4, reaching 0.0056 mg/L (P36) and 0.0066 mg/L (P39). For its part, P35 yielded a mercury concentration of 0.0045 mg/L; similar events have occurred sporadically in the Perejil and Caballo Moro zones (Table S5). Similarly, studies conducted by [62] in areas close to the study zones have shown the impact of mining operations on water resources over time. Several studies conducted around the world in areas impacted by mining zones have reported the acidification of water bodies due to the influence of mining activities. This severe impact on water bodies is considered one of the most serious environmental problems [63].

In the study area, the mining infrastructure is located primarily in the Chuyugual zone [40], where the discharge of effluents from mining wastewater treatment is critical and severely impacts the ecosystem. Vázquez and Ruiz [64] and Delgado et al. [65] have evidenced the impact of mining operations on water resources over time in areas close to the study area. In a review work, González et al. [66] examined the impact of intensive human activities in several basins in Latin America; in the present study, the results show that mining operations in the High Andean region cause severe damage to aquatic ecosystems, as was detected by González et al. [66].

3.2. Physicochemical Quality Index

According to the factor analysis, the physicochemical qualifying variables were water temperature, conductivity, sulfates, ammoniacal nitrogen, and nitrates, and the qualified metals were Ba, Fe, K, Mg, Mn, Na, and Sr.

Regarding the mean PQI scores, the minimum score was 4.8 for station P39 in the Chuyugual zone, and the maximum was 8.95 for station P24 in the Caballo Moro zone (Figure 3).

For the three study zones, it was found that the reference or control monitoring station (PC) achieved higher PQI scores than the monitoring stations impacted by mining activities (PP). In the case of the Chuyugual zone, monitoring stations under mining influence reached the lowest PQI scores since the mining facility is in this zone (Figure 4).

Since the Physicochemical Quality Index is based on regulatory limits and resulted from a factor analysis, it provided a comprehensive view of the physicochemical quality status of the monitoring stations, showing that PC stations returned the highest physicochemical quality scores, supporting their status as reference sites; likewise, it showed that the Chuyugual zone has the lowest scores since the mining infrastructure has been there [40].

The calibration process ensures that the final result (bioindication values for macroinvertebrate families and BMWP water quality classes) is defined according to the environmental conditions of the streams studied. The modification of the calibration protocol by Ruiz-Picos et al. [26] included a set of seven heavy metal concentrations as qualifying variables.

3.3. Bioindication Values

Between 2008 and 2023, 46 families of aquatic macroinvertebrates in 14 orders, six classes, and four phyla were identified in the Chuyugual zone. A total of 8011 specimens were collected; Baetidae was the most predominant family, with 1794 specimens. Regarding the number of specimens per monitoring station in this zone, a total of 2234 specimens were found at station P31 and only 33 at station P39. For the Perejil zone, 49 families belonging to 14 orders, six classes, and five phyla were collected. A total of 4491 specimens were identified; again, the family Baetidae predominated, with 1699 specimens. The number of specimens per monitoring station showed that 1887 specimens were collected at station P6, while only 192 were found at station P4. Finally, in the Caballo Moro zone, 51 families were recorded, distributed in 19 orders, 10 classes, and five phyla; 7931 specimens were found, and Leptophlebiidae was the most abundant family, with 1525 specimens collected. A total of 2615 specimens were found at station P25 and 937 at station PS21. Our results show the highest richness of aquatic macroinvertebrate families in the Caballo Moro zone, where the mean PQI score was high; furthermore, the lowest richness of aquatic macroinvertebrate families was detected in Chuyugual (where the PQI was the lowest). Studies conducted in Panama [67], Colombia [68], and Peru [69] found that streams that have been restored, are protected, or are subjected to low human intervention exhibit a higher macroinvertebrate diversity than streams affected by human activities, such as fish farming and agriculture.

Bioindication values were calculated from the abundance classes of macroinvertebrate families collected at the different monitoring stations and were determined as the fifth percentile of the distribution of abundance classes. Bioindication values ranged from 5 to 9 (

Table 1. Bioindication values of macroinvertebrate families for the High Andean zones studied.

Family	Bioindication Score
Ancylidae, Collembola, Curculionidae, Deuterophlebiidae, Dixidae, Ephidridae, Gammaridae, Haliplidae, Libellulidae, Lymnaetidae, Oligoneuridae, Veliidae.	9
Aeshnidae, Hyalellidae, Coenagrioenidae, Hirudinea, Nematoda, Ostracoda, Pisauridae, Stratiomyidae.	8
Ameletidae, Blephariceridae, Chrysomelidae, Corydalidae, Empididae, Glossosomatidae, Gyrinidae, Hydropsychidae, Hydroptilidae, Leptoceridae, Leptophlebiidae, Philopotamidae, Planariidae, Psephenidae, Tricorytidae.	7
Baetidae, Chironomidae, Corixidae, Culicidae, Helicopsychidae, Hydrachnidae, Hydrobiosidae, Muscidae, Odontoceridae, Perlidae, Psychodidae, Pyralidae, Scirtidae, Sthaphylinidae, Tabanidae.	6
Calamoceratidae, Ceratopogonidae, Dytiscidae, Dolichopodidae, Elmidae, Hydrophilidae, Mesovelliidae, Oligochaeta, Simuliidae, Tipulidae.	5

). Since the PQI did not return low values (between 1 and 4), no abundance classes of aquatic macroinvertebrates were distributed within these physicochemical quality scores. Similarly, a PQI value of 10 was not observed; therefore, no macroinvertebrate abundance classes reached this value.

Within the bioindication values calibrated for this index, the families Dytiscidae, Hydrophilidae, Mesoveliidae, Simulidae, and Tipulidae reached a bioindication value of 5; this value matches the score assigned in the original BMWP established in the United Kingdom for the same families [15,37]. Similarly, the bioindication value for the family Aeshnidae was 8 in this study, which also matched the original BMWP score. On the other hand, the family Libellulidae reached a score of 9 in this study (the maximum bioindication value), while the original BMWP score was 8. Other indices adapted for Latin American countries include the index of Roldán [24], which was proposed for Colombia. In the present study, the calibrated score for the family Mesoveliidae matched the score assigned to this family in Colombia. The family Oligoneuriidae returned a score of 10 (the highest bioindication score) in Colombia; in this study, the score for this family (9) was also the highest bioindication score recorded in the study. However, bioindication scores for the families Ancylidae and Libellulidae in the present study differ from those assigned in Colombia; in that country, the score assigned to these families was 6; in the present study, both families showed greater sensitivity to chemical pollution, reaching bioindication values of 9. In several BMWP proposals, including the original BMWP for the United Kingdom, the family Perlidae scored 10 [15,26]; but in this study, it reached a bioindication score of 6, equal to scores calibrated for superficial affluents in Panama [27].

3.4. BMWP Quality Categories

Once the bioindication scores were obtained, the BMWP was calculated for all study zones during the monitoring period (2008–2023). From the spatial–temporal universe of the 21 monitoring stations, 5% of the stations with the highest BMWP scores were selected. From this set of BMWP values, the median (108) and the tenth percentile (102) were determined and used to establish the water quality categories.

The new BMWP index was named Biological Monitoring Working Party for the High Andean Zones of Peru Influenced by Mining (BMWP/PeIAZIM, in Spanish). BMWP/PeIAZIM values above the median calculated for the sites with the highest BMWP values are considered in the Excellent water quality category; those between the median value and above the tenth percentile are considered Good quality. Values between zero and the tenth percentile were divided by four to obtain the following water quality categories: Fair, Poor, Very Poor (Highly Polluted), and Extremely Poor (Extremely Polluted). Additionally, each water quality category was assigned a color class to present a traffic-light-style color code.

Table 2. Water quality categories for the BMWP/PeIAZIM index.

BMWP/PeIAZIM Score Range	Water Quality Category
>115	Excellent
>106 and ≤115	Good
>79.5 and ≤106	Fair
>53 and ≤79.5	Poor
>26.5 and ≤53	Very Poor (Highly Polluted)
≤26.5	Extremely Poor (Extremely Polluted)

shows the BMWP/PeIAZIM ranges for each category and the corresponding color-coded representation.

3.5. Validation

The regression model used to validate the effectiveness of the BMWP/PeIAZIM index based on the qualifying variables of the Physicochemical Quality Index is as follows (R² = 0.55; p < 0.05):

BMWP/PeIAZIM = 87.45 − 3.22T + 0.67EC − 0.64SO₄²⁻ + 23.02NH₄ + 21.72NO₃⁻ + 200.42Ba + 4.62Fe + 1.22K − 0.44Na − 6.32Mg − 56.20Mn − 829.41Sr

where T is water temperature, and EC is electrical conductivity.

The agreement index [70] was applied to validate the trend between observed and calculated BMWP/PeIAZIM values. The calculated agreement index was 0.84.

Figure 5 shows the regression model for calculated and observed BMWP/PeIAZIM values. Five sites fell outside the 95% confidence interval.

3.6. BMWP/PeIAZIM for the Chuyugual, Perejil, and Caballo Moro Zones

Once the bioindication values were obtained, the BMWP/PeIAZIM index was applied to all monitoring stations during the study period. The mean BMWP/PeIAZIM scores for the period 2008–2023 ranged from 15.16 (Extremely Poor quality; site P39, Chuyugual zone) to 97.5 (Fair quality; site PS37, Chuyugual zone). The median value was 51.03, equivalent to a Very Poor (High Polluted) overall water quality category. The maximum score for the entire study period corresponds to site P25 (Caballo Moro zone, 2008, Excellent quality), and a minimum of zero was obtained for several stations in each study zone. This minimum value implies that no organisms were collected at these stations despite having conducted monitoring: P1 in 2017, 2018, and 2021 (Perejil zone); P39 in 2011, 2014–2018, and 2022; P35 in 2017; and P36 in 2017 and 2018 (Chuyugual zone); PS21 in 2008 and P24 in 2010 (Caballo Moro zone). According to the global average, four monitoring stations, representing 19.00% of all stations, correspond to the Fair water quality category; six (28.60%) are in the Poor category; six (28.60%), in the Very Poor (Highly Polluted) category; and five (23.80%), in the Extremely Poor (Extremely Polluted) quality category (Figure 6).

The monitoring stations classified as Fair quality were PC (without mining influence): PS1 in the Perejil zone, P24 and P25 in the Caballo Moro zone, and PS37 in the Chuyugual zone. Station P31, although not influenced by mining, obtained a mean BMWP/PeIAZIM value of 79, placing it in the Poor quality category. Its conservation status should be monitored before its deterioration worsens.

Figure 7 shows the spatial representation of mean BMWP/PeIAZIM values for the entire study period based on the BMWP traffic-light signage. This figure shows that the monitoring stations near the mining area belong to the lowest water quality categories (mainly in the Chuyugual zone). According to the BMWP/PeIAZIM, water quality improves downstream.

It is important to note that monitoring activities in the Chuyugual and Perejil zones have revealed findings that have drawn considerable attention, demonstrating the pollution and destruction levels that can occur when highly polluting activities are carried out, such as mining in any of its modalities (large-, medium-, or small-scale; formal, informal, or illegal). This is especially true considering that the Chuyugual zone, which drains into the Crisnejas basin, is part of the Amazon basin, is home to unique biodiversity, regulates the water cycle and produces 20% of the freshwater that reaches the ocean, acts as a major carbon sink, is home to countless Indigenous communities and native peoples, and is considered the largest basin worldwide, essential for global climate balance. In this regard, Salas-Mercado et al. [71], states that mining has affected several rivers in southern Peru, transforming them into “dead rivers. At station P39, macroinvertebrates were not collected in 2011, 2014 to 2018, and 2022, despite having used adequate sampling methodology; a similar case occurred at station P1 in the Perejil zone in 2016–2019, 2021, and 2023. Our results show that the Chuyugual zone has severe acidification, the lowest richness of aquatic macroinvertebrates, and the lowest PQI, and BMWP reaches the lowest water quality categories.

In Peru, several macroinvertebrate studies to assess water quality have been carried out using BMWP indices adapted from other countries. We can mention those conducted in the Huacamarcanga River [36], in Ayacucho [72], the upper basin of the Huallaga River [73], the Rímac River [74], the Chicama River [75], and the La Libertad High Andean lagoons [76]. Flores Rojas and Huamanticó Araujo [77] addressed the creation of a participatory environmental surveillance tool based on macroinvertebrates to monitor water quality in areas at risk of mining pollution. However, none of them involved formal calibration and validation of an index for the High Andean streams of Peru, nor have the indices been calibrated or validated in any of these cases considering the presence of stressors such as heavy metals, typical of mining environments, as is the case in the areas evaluated in this study.

Adapting indices from their original version is a common process that allows for a better description of the conditions under which they are applied. However, authors such as Ochieng et al. [78] point out that the adaptation of the BMWP-C (from tropical areas of Central America) failed to classify study sites according to a pollution gradient in tropical areas of Africa. These authors mention that failure is due to biogeographical and environmental differences and divergences in tolerance to pollution in macroinvertebrates. Consequently, index calibration is a fundamental process as it eliminates subjectivity and considers the actual responses of macroinvertebrates to the local environmental conditions in each system [56].

3.7. Analysis of Stations Without Mining Influence Versus Stations Influenced by Mining Activities

3.7.1. Perejil Zone

Tukey’s (HSD) and Duncan’s post hoc tests were used to assess the statistical significance of differences in BMWP/PeIAZIM index values between monitoring stations. According to these tests, stations PS1 and P6, considered reference stations (PC), had the highest mean BMWP/PeIAZIM scores (89.5 and 69.07, respectively; Group A), showing significant differences compared to the other stations in this zone. PS2, P3, and P5 were of intermediate quality (Groups B, C, and D), and P1 and P4 belonged to the lowest quality (Group D) (Figure 8a). The most frequent families in stations P1 and P4 were Elmidae, Baetidae, and Chironomidae, with the lowest bioindication values (5, 6, and 6, respectively; Table 1).

3.7.2. Chuyugual Zone

Station PS37 (not influenced by mining activities) achieved the highest mean BMWP/PeIAZIM score (94.91) and showed significant differences compared to the other monitoring stations (Figure 8b). Station P31, although not influenced by mining and having the second-highest BMWP/PeIAZIM value in this zone, did not show significant differences compared to the other monitoring stations. P33, P34, and P37 were of intermediate quality (Groups B and C); on the other hand, stations P35 and P36 were of low quality (Group D), and station P39 had the worst quality (Group E). Particularly, in the case of station P39, which showed the lowest BMWP score, the most frequent families were Calamoceratidae, Ceratopogonidae, Chironomidae, Dolichopodidae, Dytiscidae, Elmidae, Hidrophylidae, Oligochaeta, Simuliidae, and Tipulidae. These families have bioindication values of 5, except for Chironomidae, with a bioindication value of 6. These families returned the lowest bioindication values, showing the highest tolerance to mining pollution.

3.7.3. Caballo Moro Zone

Stations P24 and P25 (no influence of mining activities) achieved the highest mean BMWP/PeIAZIM scores (93.58 and 83.28, respectively, Group A), with significant differences versus the other monitoring stations. These scores result from the high frequency of families with low bioindication scores in addition to some macroinvertebrate families with high bioindication scores. Station P23 had intermediate quality (Group B), while P21, PS21, and P22 had the lowest quality (Group C) (Figure 8c). Monitoring stations P24 and P25 obtained the highest mean PQI and BMWP/PeIAZIM scores, which

These results confirm that the stations not influenced by mining activities achieved the highest BMWP/PeIAZIM scores and served as reference stations. Consequently, mining activity affects water quality and the aquatic macroinvertebrate community, which was evident in the BMWP/PeIAZIM scores. Several studies have used reference sites to highlight differences between sites unaffected and impacted by various human activities using aquatic macroinvertebrate communities as bioindicators. Examples are Granados-Martínez et al. [79], who used diversity indices in a gradient of anthropogenic disturbance in a Colombian stream, and Ochieng et al. [78], who used the original BMWP and the BMWP adapted for Costa Rica in addition to diversity indices. These authors pointed out the relevance of having reference sites to evidence environmental disturbances in impacted sites. The results of the present study show the sensitivity of the BMWP/PeIAZIM index for detecting alterations associated with mining operations in the High Andean region of Peru.

The analyses of variance applied by zone confirmed the hypothesis of lower BMWP/PeIAZIM scores at sites impacted by mining activities compared to sites with less intervention. Furthermore, records before the start of mining operations demonstrate the deterioration of water quality after the start of mining activities. Mykrä et al. [80], using replicate sampling before and during mining activities, demonstrated the effects of these activities on macroinvertebrate and algae communities in Finnish water bodies. However, they pointed out that the responses of organisms are multifactorial, indicating that the presence of metals may also be due to the geological nature of the substrate.

The strategic location of the monitoring stations allowed differentiating between those impacted by mining activities over time (PP) and those assumed not to be significantly affected by these activities due to their geographic location (PC). This network of monitoring sites remained stable throughout the monitoring period, except for station P20 in the Caballo Moro zone, which was incorporated in 2013, and station P3 in El Perejil, which was removed in 2012. The structure of this monitoring network is consistent with the recommendations of several researchers on the design of monitoring networks for water quality parameters in rivers [81,82], who emphasize the importance of a proper distribution of monitoring stations to determine the true impact of a given activity, differentiating between stations with a low probability of direct impact from point or diffuse sources and high impact monitoring stations.

In this study, the a-priori selection of monitoring stations with no mining impact allowed defining reference sites and confirmed the hypothesis regarding mining impacts on the three zones studied. Feio et al. [83] highlight the importance of having impact-free sites in addition to pre-impact historical data, which allows for a realistic view before and during impacting human activities. These data will support the implementation of restoration measures that can be monitored with tools such as the BMWP/PeIAZIM.

4. Conclusions

The historical database of physical, chemical, metal, microbiological, and macroinvertebrate parameters was essential for the BMWP calibration and the assessment of water quality before and during mining activities. This information allowed for the identification of sites with and without pollution from mining activities.

The BMWP/PeIAZIM index was successfully calibrated and validated by incorporating heavy metals into the factor analysis. This shows that macroinvertebrates are sensitive not only to organic pollution, as originally conceived in the BMWP, but also to mining pollution.

The Physicochemical Quality Index, based on locally measured values of environmental variables and regulatory thresholds, was the basis for determining bioindication values.

It is also crucial to highlight the importance of participatory monitoring. In this study, water quality monitoring between 2005 and 2023 was possible mainly due to community environmental surveillance by civil society. These initiatives were consolidated through the training of environmental monitors and watchdogs organized in community committees.

The bioindication scores obtained in the present study differed from those obtained in other studies. The present study detected that certain macroinvertebrate families with the highest bioindication values are sensitive to mining pollution. Some bioindication values coincided with those recorded using BMWP indices proposed by other authors, indicating that the calibration process places certain families at the same pollution tolerance level established by other authors. Families with higher resistance to mining pollution were also identified.

The findings of this study confirmed our hypothesis as the evidence gathered shows that mining activities lead to the extirpation (i.e., local extinction) or decline of macroinvertebrate communities, with potentially significant impacts on the balance of the aquatic ecosystem, showing that the Chuyugual zone is the most affected area, being where the mining infrastructure is located.

Therefore, the BMWP/PeIAZIM index is a valuable tool for classifying and discriminating between affected and unaffected areas. Its usefulness lies in its ability to facilitate monitoring and surveillance and define remedial measures at impacted sites, thus enabling a robust assessment of water quality in High Andean monitoring zones influenced by mining activities.

5. Future Directions

Considering that the BMWP/PeIAZIM index showed the impact of mining activities and given the geographic heterogeneity in Peru, we suggest that the methodology proposed here be replicated in other regions of the country facing similar problems, calibrating and validating the BMWP index at the basin level. Therefore, it is recommended that Peruvian environmental regulations incorporate the use of aquatic macroinvertebrates as biological indicators of water quality and the use of the calibrated BMWP index.

Promoting the sustainable use of resources in the High Andean basins is urgently needed through the planning, design, and implementation of participatory territorial management instruments that strengthen governance and involve local actors and strategies adapted to social and environmental realities that consider the restoration and conservation of High Andean freshwater habitats.

Finally, it is important to revalue ancestral knowledge such as planting, breeding, and harvesting water, promoting afforestation and reforestation with native tree species, especially in the headwaters of river basins. These actions are essential to regulate the hydrological cycle, improve water availability, and increase resilience to climate variability.