In the semi-arid southwest, rapid urbanization and population growth have led to increased use of treated effluent to augment and maintain hydrologic conditions in the watershed resulting in both positive and negative consequences in terms of overall watershed quality [1,2]. Planned water reuse is a common occurrence globally and began as early as 1918 in California and Arizona in order to provide irrigation water for crops [3]. Discharge of treated effluent into stream channels recharges the groundwater aquifers, supports riparian habitation, enhances ecosystem services, and is commonly implemented by state agencies for these reasons [4,5]. For example, natural perennial and ephemeral flows in the Upper Santa Cruz River are artificially augmented by treated effluent from the cities of Nogales and Tucson where, historically, portions of the Santa Cruz River near the city of Tucson were pumped dry as early as 1910 [6].

However, reliance on treated effluent for perennial streamflow potentially endangers human health due to recreational exposure and possible contamination of domestic water supplies by increased microbial pathogen concentrations in surface and ground waters [4,7,8,9]. Common sources of potential pathogenic contamination in surface waters include storm runoff from urban and agricultural landscapes, wild animal wastes, wastewater treatment plant discharges, and failing septic system drainage [8,10,11]. Monitoring river networks for all potential pathogenic agents is expensive and not feasible; therefore, methodologies for monitoring fecal indicator bacteria (FIB) and determining acceptable risk have been established [12,13,14,15]. Current ambient water quality criteria for FIB in fresh waters are aimed to protect human health from gastroenteritis due to pathogenic exposure based on the estimated relative risk of 8 cases of gastroenteritis per 1000 swimmers [12]. The appropriateness of the methods used and FIB capability for correlating and identifying human health risk from pathogens has been debated in the literature [16,17,18,19]. Despite the ongoing debate, most states monitor for total coliforms, fecal coliforms, Escherichia coli (E. coli), fecal streptococci, or enterococci as indicators of potential pathogens in water resources. In Arizona, E. coli has replaced fecal coliform as the preferred FIB in stream networks [20,21].

To minimize the potential risk of wastewater to public health and the environment, state agencies regulate and permit planned wastewater reclamation and reuse facilities [3]. In many cases, these facilities, regardless if the intended reuse is for recharge or irrigation, achieve a high degree of consistent water quality, and the removal of microbial and other contaminants associated with human waste are of paramount concern [22,23]. As this case study will show, additional research and assessment of the fate and transport of pollutants released indirectly into effluent-dominated and/or effluent dependent stream networks are critical to controlling overall FIB loading in the watershed. The objectives of this study are (1) to assess the water quality in the Upper Santa Cruz Watershed in terms of FIB by comparing the available data to the water quality criteria established by Arizona, (2) to provide insights into FIB response to the hydrology of a semi-arid watershed and (3) to identify major FIB contributors (point sources versus nonpoint sources) to the stream.

The entire Santa Cruz Watershed is composed of approximately 28,749 km², roughly 10% of the state of Arizona; land ownership is approximately 40% tribal, 25% federal, 20% private and 15% state [20]. The Santa Cruz River has its headwaters in Arizona’s San Rafael Valley, which is in the southeast/central part of the state. The river flows south and makes a 40 km loop through Mexico before returning to the United States (U.S.) about eight kilometers east of Nogales, Arizona. The river then flows north from the U.S.-Mexico border and converges with the Gila River, just southwest of Phoenix. According to the Arizona Department of Environmental Quality (ADEQ), grazing is the dominant land use while irrigated crop production is limited to areas near streams, but restricted land uses have been established near several wilderness areas, national forests, and national monuments. In addition, mining operations, both active and abandoned, are located throughout the watershed [20]. Annual precipitation ranges from 280 to 860 mm (valley to mountain, respectively). This study focuses on the sub watersheds containing the Santa Cruz River south of Tucson, Arizona.

Most of the population in the Upper Santa Cruz Watershed is found in the city of Tucson (population 530,000), the state’s second largest city after Phoenix [24]. There is also a population of 370,000 located on the U.S.-Mexico border in the sister border cities of Heroica Nogales, Sonora, Mexico and Nogales, Arizona, U.S. According to the U.S. Census Bureau (2005), the population in the state of Arizona is projected to increase by approximately 52% over 30 years from 2000 to 2030 which is expected to increase the urban water demand by approximately 45% despite sustainable development efforts [25,26]. The growth in Sonora, Mexico is expected to increase at an even higher rate which is anticipated to increase the urban water demand by 18% by 2030 [25]. As more demand from urban growth and land use is placed on the system, understanding the fate and transport of pollutants released and how treated effluent impacts the overall water quality, especially water supplies designated for human consumption, is necessary.

Water quantity and quality issues in the Upper Santa Cruz watershed are confounded by the quality of waters flowing from areas of Mexico which have less regulated infrastructure to handle wastewater treatment [27]. Continuous efforts are being made by both countries to provide wastewater service in rural areas and to enhance wastewater treatment and reclamation infrastructure to meet future needs [28]. The Groundwater Storage, Savings, and Replenishment Program managed by the Arizona Department of Water Resources (ADWR) permits groundwater and surface water recharge facilities to discharge reclaimed waters into infiltration basins and, in some cases, directly into the Santa Cruz River [5]. The ADEQ permits 22 facilities, each issued an Arizona Pollution Discharge Elimination System (AZPDES) permit, to discharge treated effluent into the Santa Cruz River and its tributaries [29]. These facilities, not all of which are actively discharging, include wastewater treatment plants (WWTP), wastewater reclamation facilities, and water pollution control facilities. The Central Arizona Project (CAP) canal allocates 563,947,056 m³ of Colorado River water per year to Pima, Pinal, and Maricopa counties to supplement domestic water supplies and also to maintain aquifer levels [30]. In 2010, Pima County, Arizona produced approximately 84,860,000 m³ of treated effluent of which about 76,720,000 m³ was discharged from facilities located in Tuscon, Arizona [31]. In Santa Cruz County, Arizona, the newly expanded Nogales International Wastewater Treatment Plant (NIWTP) (see Figure 1 Map ID C) treats more than 56,781 m³/day, approximately 20,720,000 m³ annually, of wastewater from both Nogales, Arizona and Heroica Nogales, Sonora and discharges it to the Santa Cruz River after advanced biological treatment [32].

Monthly E. coli and fecal coliform monitoring data from both point sources such as WWTP discharge pipes and nonpoint sources from numerous stream segments throughout the Upper Santa Cruz Watershed as shown in Figure 1 were used in this study. The E. coli and fecal coliform data used in this study are from numerous sampling records including ADEQ in conjunction with Friends of the Santa Cruz River (FOSCR), National Park Service at Tumacacori National Historical Park and Sonoran Desert Network, Sonoran Institute, United States Geological Survey (USGS), and U.S. Environmental Protection Agency (EPA) Envirofacts permit compliance system (PCS) database. For the point sources data, a custom search on the Envirofacts PCS database was preformed to assess indicator bacteria concentrations from WWTP monthly discharge monitoring reports (DMR) prepared by AZPDES permitted facilities which discharge treated effluent into the Santa Cruz River and surrounding washes and tributaries (Figure 1, Map ID A and B) [29]. These grab samples show a snap shot in time and space of the FIB activity for a given location and were collected to either fulfill the AZPDES monitoring requirements or for water quality assessment purposes. The available data for the watershed are organized by location and vary in regard to sample frequency, period of record, sampling method, and FIB assessed (fecal coliforms or E. coli). The WWTP DMRs data collected were summarized into a monthly report. For nonpoint source data, in-stream samples were collected primarily on a quarterly or monthly basis unless no sample could be obtained due to low or no streamflow conditions; several gaps in the sampling record exist at each location. The geometric mean and sample maximum for each WWTP DMR and each in-stream sampling location available are summarized in the results section below. Variations in the targeted FIB disallow direct comparison of each sampling location for the entirety of the sampling record and the reported concentrations have differences in terms of method quantification limits and the lab methods used. The lab method reported for E. coli samples is listed as SM9223B and fecal coliform concentrations were determined using direct plating methods (SM9222E) or the Most Probable Number (MPN) method [13,15]. For the raw in-stream sampling data, a geometric mean and maximum concentration are calculated for the FIB reported at each location. The results are presented in the Table 2, Table 3, Table 4, Table 5 below.

The available data at each sampling location are compared to regulatory water quality criteria for FIB established in Arizona as summarized in Table 1. According to the regulatory standards listed in Table 1, wastewater dischargers report bacteria concentrations as a geometric mean of all the test results obtained during a reporting period, which is helpful when analyzing bacteria concentrations that may vary anywhere from 10 to 10,000 fold over a given period. The single sample maximum value is also needed to ensure that public health is protected from unusually high microbial loads.

Average daily baseflow conditions were determined using the Web Based Hydrograph Analysis Tool (WHAT) and the local minimum method for daily streamflow from 1 March 1996 to 30 April 2008 at two USGS stations (09481740 and 09480500) within close proximity of the sampling locations [33]. Since the local minimum method generally overestimates baseflow during storm events, the WHAT results were compared to precipitation data for a better estimation of actual baseflow conditions. Then, the correlation between streamflow/precipitation and in-stream fecal coliform/E. coli concentrations was analyzed to identify potential factors impacting the in-stream fecal coliform/E. coli concentrations. Precipitation data was obtained from weather stations maintained by the National Oceanic and Atmospheric Administration (NOAA). Streamflow data was collected from gage stations maintained by the USGS. Finally, data collected from point source WWTPs were compared with nonpoint in-stream grab samples and statistical tests were performed to see if fecal coliform/E. coli concentrations were significantly different between WWTPs and nonpoint sources. In instances where the sample value was reported as greater than the upper method detection limit or less than the lower method detection limit, the detection limit was used in the statistical comparison.

As this study verifies, significant surface water impairment is a result of nonpoint source pollution in Arizona. In-stream concentrations of fecal coliform and E. coli are significantly greater than those concentrations discharged in the treated effluent from WWTPs, as shown in Figure 5, Figure 6. Nonpoint sources such as faulty septic systems, agricultural and urban runoff, unregulated discharges to stream washes, land use practices, and in-stream fate and transport processes contribute a significant portion of the pollution load to the Santa Cruz River; the statistical data reported in Table 6 supports this finding. According to the ADEQ 2006/2008 statewide summary report, point source contributions to stream pollution impacted 46 miles of streams while nonpoint sources contributed to pollution to 3245 miles of the statewide stream network [20]. The data presented in this study indicate all sampling locations assessed in the Upper Santa Cruz watershed, both point and nonpoint, exceed the water quality criteria established by Arizona to protect human and aquatic health. DMRs submitted to regulatory agencies have several occurrences of FIB concentrations in the treated effluent exceeding the established water quality criteria. Depending on the specifics of the facility permit and wastewater class, these exceedances may be acceptable in some cases.

Studies have shown that FIB survival in surface waters varies from hours to days or even months if protected by sediments which make identifying the source of the FIB concentrations difficult [37,38]. The decay rate of FIB in surface water is a function of many ecological influences; therefore, water quality management, best management plan (BMP) development, watershed modeling, and risk assessment practices need to incorporate better methods as to how FIB interact with the environment, and furthermore, how well FIB accurately model true pathogenic concentrations in the watershed [16,17]. Researchers and regulators continuously question which pathogen indicators are appropriate to determine safe exposure levels in recreational waters. USEPA has approved several detection tests for evaluating FIB in water samples, and comparisons of these methods indicate high variability in sample results [17]. Field et al. [19] evaluates the application of fecal source tracking as a better method for human health risk assessments and managing water quality compared to current reliance on FIB criteria. Litton et al. [39] further identifies fecal markers and source tracking tools which could vastly change the approach to FIB monitoring and regulation. These studies and the one presented here provide data on FIB concentrations in selected streams with respect to concentration, relationship to recreational water-quality standards, and influence of environmental factors such as streamflow, rainfall, sediment, and runoff [36]. Findings indicate that FIB concentrations are highly variable, especially along urban streams even in the absence of significant rainfall. Though FIB generally increase with streamflow and precipitation events as shown in Figure 2, Figure 3, Figure 4, there are occurrences of peaks in FIB concentrations during baseflow conditions.

In Figure 2, Figure 3, Figure 4, it is important to provide insight into the data to reach sound conclusions. Overall, trends and correlations show that increased fecal coliform and E. coli concentration generally correspond to increased streamflow from rainfall and concentrations are generally higher in the summer months as shown in other similar studies [36]. However, there were instances of increased fecal coliform or E. coli concentrations observed during months of little precipitation or streamflow. The data also show that in months of little to no streamflow, several locations were noted as “no sample collected due to no-flow conditions” on the day of sampling. Opportunities for consistent sample collection are limited due to the ephemeral nature of the streamflow, especially at tributary locations. It is likely that in-stream sample collection was done during periods of higher streamflow than average during little or zero baseflow conditions; however, most sample collection was done during low flow conditions and not as a result of precipitation events. As shown in Figure 4, peaks in fecal coliform concentrations positively correlate (R² = 0.43) to months of high rainfall. The data compiled for this study provides insight into the water quality conditions related to pathogen indicators in the watershed; however, the underlying conditions, which could affect the grab sample concentrations—such as the sample collection and analysis method, agricultural activity, grazing activity, seasonal hydrology, and stream ecology—were not always clear in this assessment. The variation of the analysis methods and the FIB of interest disallow direct comparison of each sampling location for the entirety of the sampling record and may over or under estimate the actual value.

Efforts to mitigate nonpoint sources are mostly voluntary yet very active across the nation. Watershed managers encourage stakeholders to participate in watershed management groups, volunteer monitoring programs, BMP development and implementation, and education. Examples of successful BMPs for FIB mitigation in effluent dominated systems include engineered wetlands, bioretention areas, and filter strips [40,41]. In addition, improvements in watershed modeling capabilities allow better fate and transport for remediation studies and TMDL development [42,43]. In Arizona, the ADEQ adopted a suspended sediment concentration (SSC) standard of 80 mg/L in 2002 to replace its turbidity standard [20] which is closely linked to FIB concentrations released into the surface waters. Suspended sediment reduction is a priority in many watersheds in order to enhance water quality and to protect fish and aquatic communities. Hindering this progress is the lack of monitoring data in many watersheds which delays efforts to develop, implement, and assess the effectiveness of watershed control strategies such as the SSC standard.

Like much of the southwest, Arizona uses recycled waters for groundwater and surface water recharge to balance the supply and demand of a growing population. However, continuous monitoring of the fate and transport of FIB and their associated pathogens is an area needing further assessment. To fully assess the water quality in the Upper Santa Cruz watershed, a detailed analysis is needed which allows for FIB monitoring, source tracking, and reduction of nonpoint sources of pollution. This study assesses the influence of WWTP discharges and nonpoint sources on the indicator bacteria concentrations in the Santa Cruz River and surrounding tributaries. The results of this assessment find that the Upper Santa Cruz watershed is impaired with fecal coliform and E. coli at levels, which exceed the established water quality criteria in Arizona. This assessment indicates that a risk to human health exists especially during the summer months when concentration trends increase and water contact is most likely to occur. Fecal coliform and E. coli levels from the WWTP effluent assessed in this study are significantly lower than the in-stream samples assessed which indicates that nonpoint sources play a significant role in the water quality conditions. Regardless of the sample type (effluent or in-stream), all sampled locations with available data exceeded the water quality criteria for fecal coliform and E. coli indicators. Water quality issues in the Upper Santa Cruz watershed are confounded by the quality of waters flowing from urbanized areas of Mexico with less regulated infrastructure to handle wastewater treatment.

Using natural vegetation filters, stabilization of stream banks, improvement of riparian zones, and urban runoff reduction in order to reduce erosion and sedimentation, are effective watershed control strategies. Updating septic systems is another method of source reduction of potential pathogens to the aquatic environment. Sediment is linked to pollutants such as pathogens and nutrients, and suspended sediment reduction should be a priority in this watershed. Management practices aimed to reduce urban runoff and thus sediment could markedly reduce nonpoint sources of FIB in the stream network. Though likely a more expensive option, infrastructure improvements that eliminate faulty septic systems and combined sewer overflows would also reduce FIB concentrations released into the stream system. Advanced treatment of wastewater effluent and industrial discharges is another option to consider for reducing FIB concentrations within the watershed; the state of the art wastewater treatment at the Nogales plant is a good example of the current and ongoing efforts to achieve such objectives in Arizona. These recommendations could only be truly beneficial to the managers and regulators once TMDL values are established for impaired waterways and more data has been collected to assess how pathogens cycle through the entire watershed. As urbanization and population growth continues in the Santa Cruz watershed, water regulators, managers, and development planners will have to assess the impact of effluent-dominated stream sections in order to meet not only water quantity objectives, but also to maintain water quality standards.