**2.1. Study area**

**1. Introduction**

40 Water Quality

graphical differences.

resources and bring them a sustainable management.

In the framework of the World Economic Forum, the "Water Crisis" is positioned as the highest concern global risk for the next 10 years [1]. In this sense, water quality and management of freshwater ecosystems are one of the main challenges worldwide [2]. However, these ecosystems face impacts and degradation that are result of human population increase and agricultural and industrial development [3]. Consequently, freshwater ecosystems and their biota are considered as the most endangered and threatened worldwide [4]. In developing countries, there is an extremely high population growth, increasing industrialization and urbanization processes, with severe and constant changes in land use, whereby the freshwater ecosystems are highly impaired [5]. Rivers crossing different land uses (urban, industrial and agricultural) are the most threatened by anthropogenic activities [5]. The threat to freshwater systems, in particular in developing countries, make evident an urgent need for developing tools for the assessment and classification of aquatic conditions in order to manage water

Biomonitoring is considered as the most appropriate method for environmental studies and for the control of water quality, due to that living organisms are excellent biosensors of the physicochemical and biological characteristics of water [6]. The aquatic macroinvertebrates have been used as bioindicators because they have a wide range of habitats and sensitivity to environmental pollution and other types of stressors, including sediment [7]. Thus, the macroinvertebrate assemblages change in response to environmental disturbances in predict‐ able ways including a strong reduction in species and abundance in impacted areas and more tolerant species predominate; whereas, sensitive species are only present in environments with the least impact or un‐impaired conditions. Moreover, biomonitoring integrate information over longer periods of time and better represent the responses of aquatic habitats providing information concerning the present state and the past trends in environmental conditions [8]. The Biological Monitoring Working Party (BMWP) is among the most used bioassesment index in Europe, which was originally developed in the UK in 1976 [9] and it has been used by the regulatory authorities in the UK as the basis of their river invertebrate status classification system since 1980. This index assigns scores to each macroinvertebrate taxon according to their responses to oxygen deficits caused by organic pollution. The analysis of these pollution‐ induced responses allows the calculation of sensitivity values by the different groups of organisms. Because of its ease of use and low cost, the BMWP index has been used in many other countries in Africa, Asia, Oceania and Latin America [10]. Nevertheless, the BMWP scores for each taxon must be calibrated to each ecological region since the taxonomic com‐ position, ecological, zoogeographic and anthropogenic conditions promote important geo‐

Additionally, the scale to ranking water conditions must be adapted for each particular condition. In Latin America, attempts have been made to develop regional indices [5]. In México, the water quality indicators used by the National Commission of Water are fecal coliforms, biochemical oxygen demand, chemical oxygen demand and total dissolved solids [11]; unfortunately, biomonitoring is not included in the current legislation, while information Apatlaco and Chalma‐Tembembe rivers are located in the Balsas Basin (**Figure 1a**), one of the largest catchment areas in Mexico (area of 117,405 km2 ) [15]. Both rivers are in the same zoogeographic region (Neotropical), which belongs to the Ecoregion Balsas Complex and belongs to the Biogeographic Province "Depresión del Balsas," particularly to the "Alto Balsas" [15]. The Chalma‐Tembembe River is formed by the Chalma and Tembembe rivers; the first has a length of 70 km and the second of 50.72 km. The Chalma River joins the Tembembe River at its lower reaches. The Chalma‐Tembembe subbasin has a mean annual rainfall around 600 mm. The area consists of tropical deciduous forest (≈47% of landcover), with some areas of rain‐dependent and irrigated agricultural use. The strong pressure from agricultural activities has favored land use changes and the loss of the original vegetation [15]. Six study sites were selected in Chalma‐Tembembe: El Arco (I), La Loma (II), El Platanar (III), Casa de la Escuela (IV), Coatlán (V) and Hacienda de Cuautlita (VI) (**Figure 1a**). The Apatlaco River has a length of 63 km, with annual rainfall from 850 to 1500 mm. The natural vegetation has been highly fragmented and transformed, with only 27% of the original area of tropical deciduous, coniferous and oak forest remaining. Moreover, an important urban‐industrial corridor (Cuernavaca corridor), runs alongside the river. The activities in the vicinity of the river include agriculture, lumber forestry, hunting and fishing [15]. Nine study sites along the main river channel (Apatlaco River) were selected: Las Truchas (VII), El Pollo (VIII), El Rayo (IX), El Encanto (X), Salida Panochera (XI), Xochitepec (XII), Alpuyeca (XIII), Xoxocotla (XIV) and Zacatepec (XV) and two more along the westerly tributary: Buenavista 1 (XVI) and Buenavista 2 (XVII), study sites located before and after the effluent of a wastewater treatment plant and three more along the easterly tributaries: El Texcal (XVIII), La Gachupina (XIX) and Las Juntas (XX), resulting in a total of 14 sampling sites (**Figure 1a**). In both rivers, four sampling campaigns were undertaken (the dry season in December 2012 and February‐March 2013 and the rainy season in August‐September 2012 and June 2013).

**Figure 1.** Study area location. (a) Study sites. (b) Validation and range extension study sites.

### **2.2. Water quality measurements and water quality index**

For each site and sampling period, some variables were recorded: altitude (masl), water temperature (°C), conductivity (µS/cm), pH, salinity (PSU), dissolved oxygen (DO mg/L O2) and turbidity (NTU), using a Quanta® multiparameter probe. Air temperature was recorded with a 45118 EXTECH anemometer. Water samples were transported to the laboratory in refrigerated and dark conditions. Biochemical oxygen demand (BOD5 mg/L O2), chlorides (mg/ L Cl), alkalinity (mg/L CaCO3) and total and fecal coliforms (MPN/100 mL) were determined according to [16]. Nitrite (mg/L NO2), nitrate (mg/L NO3), ammonia (mg/L NH3), total nitrogen (mg/L TN), orthophosphate (mg/L PO4), total phosphorus (mg/L TP), hardness (mg/L CaCO3), sulfates (mg/L SO4 2−) and color (U Pt‐Co) were analyzed using a Hach DR 2500 spectropho‐ tometer. Additionally, the DO saturation (%) was computed. The water quality index (WQI) proposed by Dinius [17] was calculated, which range from 0 to 100; 100 is excellent and 0 is strongly polluted.

### **2.3. Macroinvertebrate sampling**

Aquatic macroinvertebrates were sampled at each sampling site and season with a multi‐ habitat monitoring system [18], along a section of 100 m length to incorporate local habitat variation (fast flowing riffles, pools, submerged vegetation and riparian vegetation). An area of 1 m2 was sampled for each type of habitat. Triplicate samples for each type of habitat, with a 10‐min collecting effort, were collected and pooled for analysis. The samples were taken using a kick net for fast‐flowing riffles and pools, while type‐D nets for submerged and riverine vegetation, all nets with a mesh size of 500 µm. Organisms collected were preserved with 70% alcohol. Taxonomic identification at the family level was conducted using stereomicroscopes (Nikon C‐Leds) and with the use of keys [19, 20].

### **2.4. BMWP index calibration**

### *2.4.1. Data processing*

The BMWP index is calculated by adding up the individual tolerance scores of aquatic macroinvertebrates at family taxonomic level present at a sample site. We calibrate the

tolerance scores of the aquatic macroinvertebrates in several steps [14]: (1) obtaining a physicochemical quality index (*Pcq*) for all the study sites; (2) assessing the bioindication values to each macroinvertebrate family according to the *Pcq* and their abundance class; and (3) with the scores calibrated for each family of macroinvertebrate from Apatlaco and Chalma Tembembe rivers, we assessed the BMWP. Additionally, we define the water quality categories following the procedure of Ref. [21].

### *2.4.2. Mathematical formulation*

**Figure 1.** Study area location. (a) Study sites. (b) Validation and range extension study sites.

For each site and sampling period, some variables were recorded: altitude (masl), water temperature (°C), conductivity (µS/cm), pH, salinity (PSU), dissolved oxygen (DO mg/L O2) and turbidity (NTU), using a Quanta® multiparameter probe. Air temperature was recorded with a 45118 EXTECH anemometer. Water samples were transported to the laboratory in refrigerated and dark conditions. Biochemical oxygen demand (BOD5 mg/L O2), chlorides (mg/ L Cl), alkalinity (mg/L CaCO3) and total and fecal coliforms (MPN/100 mL) were determined according to [16]. Nitrite (mg/L NO2), nitrate (mg/L NO3), ammonia (mg/L NH3), total nitrogen (mg/L TN), orthophosphate (mg/L PO4), total phosphorus (mg/L TP), hardness (mg/L CaCO3),

tometer. Additionally, the DO saturation (%) was computed. The water quality index (WQI) proposed by Dinius [17] was calculated, which range from 0 to 100; 100 is excellent and 0 is

Aquatic macroinvertebrates were sampled at each sampling site and season with a multi‐ habitat monitoring system [18], along a section of 100 m length to incorporate local habitat variation (fast flowing riffles, pools, submerged vegetation and riparian vegetation). An area of 1 m2 was sampled for each type of habitat. Triplicate samples for each type of habitat, with a 10‐min collecting effort, were collected and pooled for analysis. The samples were taken using a kick net for fast‐flowing riffles and pools, while type‐D nets for submerged and riverine vegetation, all nets with a mesh size of 500 µm. Organisms collected were preserved with 70% alcohol. Taxonomic identification at the family level was conducted using stereomicroscopes

The BMWP index is calculated by adding up the individual tolerance scores of aquatic macroinvertebrates at family taxonomic level present at a sample site. We calibrate the

2−) and color (U Pt‐Co) were analyzed using a Hach DR 2500 spectropho‐

**2.2. Water quality measurements and water quality index**

sulfates (mg/L SO4

42 Water Quality

strongly polluted.

**2.3. Macroinvertebrate sampling**

**2.4. BMWP index calibration**

*2.4.1. Data processing*

(Nikon C‐Leds) and with the use of keys [19, 20].

The physicochemical quality index (*Pcq*) is a value that describes water quality at each sample site on a scale from 0 to 10; 0 corresponds to a highly impacted site and 10 to a site with excellent water quality. Its calculation utilizes a data matrix with the mean values of each physicochem‐ ical parameters obtained from the four study periods for each sample site. Parameters recorded in situ, as well as quantified in the laboratory, were included. Values for each parameter were normalized: *Ci* = Ln (*i* +1) and for percentages *Ci* = (2/π)\* arcsin . Where *Ci* is a physico‐ chemical variable and *i* is the mean value of a physicochemical variable. The data matrix was subjected to a factor analysis with the software XLstat version 2013. Parameters that showed a significant correlation (*p* < 0.05), either positive or negative, in the first two factors were considered as qualifying variables or *Ci***2**; these variables were arranged in a new data matrix of sampling sites *vs* the *Ci***2**. The maximum (*Ci***2max**) and minimum (*Ci***2min**) values for each qualifying variable were determined and assigned taking into account the environmental legislation for the water management in México (Mexican Official Standards: [22, 23]), the USA [24], Canada [25], Central and South American countries [26, 27] and a worldwide level [28]. With *Ci***2max** and *Ci***2min** values for each qualifying variable the *Ci***2** were standardized:

$$Ci\_3 = \frac{Ci\_2 - Ci\_{2\text{min}}}{Ci\_{2\text{max}} - Ci\_{2\text{min}}}.\tag{1}$$

Each *Ci***3** value was adjusted to fit the BMWP scale that goes from 0 to 10, using the following formula: *Ci***4** = (1 − *Ci***3**)\*10. In the case of the qualifying variables associated with good water quality (DO and DO% saturation), the inverse procedure was followed by applying the formula: *Ci***4** = *Ci***3**\* 10. For each study site, an average was assessed with the *Ci***4** values for the selected variables to determine the physicochemical quality:

<sup>4</sup> <sup>1</sup> . *n i Ci ni* <sup>=</sup> <sup>=</sup> å *Pcq* (2)

The *Pcq* values fluctuate from 0 to 10, for all 20 sampling sites and were incorporated into 10 categories of quality or *Pcq* intervals.
