Instream Flows: New Tools to Quantify Water Quality Conditions for Returning Adult Chinook Salmon

Returning adult fall-run Chinook salmon (Oncorhynchus tshawytscha) face potential obstacles in basins where their natal streams have reduced flows due to land use and associated water demands (Sheer and Steel 2006; Waples et al. 2009; Keefer et al. 2010). Specifically, where irrigation demands coincide with aquatic habitat needs, water shortages may result for both demands, particularly during dry years (Downard and Endter-Wada 2013). Whereas seasonal low-flow conditions and periodic drought are natural occurrences in freshwater ecosystems, extended drought-like conditions can seriously impair ecosystem function (Gasith and Resh 1999; Lake 2003; Lindley et al. 2007; Arthington et al. 2010).

Securing water for environmental outcomes is currently gaining momentum in the United States and in Australia and South Africa (Postel and Richter 2003). Flow restoration programs based on environmental flow prescriptions are currently under way in many basins in the United States, including the Colorado River in the Grand Canyon (Meretsky et al. 2000), the Trinity River in California (Stalnaker and Wick 2000), and the Kissimmee River in Florida (Toth 1995). Large-scale, multiagency flow restoration programs such as these illustrate a generally top-down approach to managing watershed-scale resources that involve coordination or reoperation of large infrastructure (Hillman and Brierley 2005). Despite these well-known efforts, in general, very few environmental flow programs have been implemented, and even fewer in a research and management context (Davies et al. 2013). Davies et al. (2013) noted that society’s willingness to allocate water to sustain the environment is increasingly competitive. Where environmental flow programs have been proposed or developed, several challenges have been identified, particularly with regard to transaction quantification.

New types of environmental flow programs aimed at securing water voluntarily from private water rights holders are currently gaining momentum in the West. In the past decade, environmental flow programs such as the Columbia Basin Water Transaction Program and Scott River Water Trust have been created to provide water instream while working collaboratively with willing water right holders (Garrick et al. 2009; Garrick and Aylward 2012; Wheeler et al. 2013). These programs follow the broader trend of providing environmental flows that have been designed to focus on fish species with significant recreational or commercial value, particularly salmonids in the cool-water streams of North America (Jensen and Johnson 1999; Tharme 2003). Cooperative agreements with water right holders that augment instream flows are also being negotiated and implemented in the Shasta Basin and are the subject of this study.

The problem of securing water supply for environmental purposes is twofold. First, in much of the West, conservation and regulatory policies focus on aspects of aquatic ecosystems (such as water temperature and sediment) rather than developing mechanisms to secure the water supplies that are crucial to ecosystem function (Soule et al. 2005; Fisher et al. 2008). Secondly, environmental objectives may not be considered a “beneficial use” of a water right. Because environmental water uses were not legally recognized by most states until the late 1970s, water rights for many of these uses have far less secure, more “junior” water rights (Getches 2009; Hillman et al. 2012; Downard and Endter-Wada 2013). Therefore, securing water supplies for environmental objectives may be challenging given the existing “first in time, first in right” water rights framework that is commonly implemented throughout the arid West (Downard and Endter-Wada 2013). Alternative approaches are providing flexibility to access water supplies independent of seniority-based frameworks (Baron et al. 2002; Grey and Sadoff 2007). During times of drought, cooperation between neighboring landowners outside the formal requirements of water law can alleviate negative impacts of water scarcity (Endter-Wada et al. 2009).

Despite the lack of environmental flow augmentation programs, researchers have advanced a variety of methods to develop flow prescriptions aimed at improving ecological health over the past two decades (Postel and Richter 2003). Many methods primarily focus on water quantity without explicitly considering water quality (Olden and Naiman 2010; Null and Lund 2012). Nonflow, spatiotemporal variables such as water quality (Jager and Smith 2008; Null et al. 2010) and geomorphology (Null and Lund 2012) have been identified as critical to identifying effective water management actions with environmental objectives. When nonflow variables are considered, another challenge is trying to translate ecological principles into hydrologic management recommendations for particular river basins (Poff et al. 2003). For example, approaches such as DRIFT (King et al. 2003) and ELOHA (Poff et al. 2009) have been developed to identify environmental flow regimes; however, these approaches involve substantial resources to quantify variables across multiple disciplines (King et al. 2003; Pahl-Wostl 2007) or assess the water system at global (Pahl-Wostl 2007) or regional scales (Poff et al. 2009). When environmental flow experiments are developed to address specific questions in particular reaches, critical information is provided but rarely within short management timeframes (Arthington et al. 2006). Regardless of the timing, scale, or recommended breadth of human, physical, or biological variables, effective environmental flow recommendations include two factors: the spatiotemporal variability of water quality conditions and their relationship to ecological function in targeted stream reaches.

The objective of this study was to develop a method to quantify the effects of flow transactions on physical habitat and water quality. Specifically, this study focused on the relationship between stream flows, habitat capacity, and dissolved oxygen conditions in holding habitat for adult fall-run Chinook in the Shasta River. The results of this study help quantify the direct benefit of individual, small-scale flow transactions on ecological conditions. Also, the findings help improve understanding of the processes through which these benefits are provided. The tools and methods developed during this study can be broadly applied to basins in which environmental flows have been proposed to improve ecological conditions by providing quantification and transparency to the flow transaction process.

The Shasta River is a tributary to the Klamath River located in Siskiyou County, California. The stream supports an array of anadromous salmonids, including fall-run Chinook and coho salmon, and steelhead trout. Agricultural and municipal water resources development has occurred over the past century. In this highly managed, spring-fed and snowmelt-runoff basin, low flows during the end of the irrigation season (April 1 through September 30) result from a combination of minimal seasonal base flows and water diversions (Nichols et al. 2014). Prior to recent restoration activities and changes to water management practices in the Shasta Basin, decades of poor grazing practices and water diversions degraded water quality and physical habitat conditions for multiple life stages, including migration (MTI and HSU 2014). As a result, the estimated returns of Southern Oregon-Northern California Coastal Evolutionary Significant Unit (SONCC-ESU) adult fall-run Chinook salmon (FRC) have declined from over peak numbers in the early twentieth century of 80,000 to generally less than 10,000 in more recent years.

Although FRC in the SONCC-ESU are not listed as threatened, Chinook salmon generally have considerable commercial and recreational value in California (Yoshiyama et al. 1998; Weise and Harvey 2005); in 2013, Klamath fall-run Chinook accounted for approximately 29% of California’s commercial and recreational fishery (PFMC 2014). It is estimated that 30% of Klamath River salmonids originated in the Shasta River (Moyle et al. 2008). Because of voluminous spring-fed sources that provide desirable physical and water quality conditions for salmonids, and support a robust food web, the Shasta River is the focus of extensive conservation activities that focus on restoring and maintaining the aquatic ecosystem (Jeffres et al. 2010; Willis et al. 2012).

Throughout the basin, water resource users and managers are modifying water use activities to include environmental objectives; however, competing demands remain a challenge. A flow transaction program is one proposed management strategy to add flexibility to the existing water right allocation system that defines water use in the Shasta Basin. Water rights in the Shasta Basin are a mix of riparian and appropriative water rights that were adjudicated as per the 1932 Court Decree (DWR 1932). The Shasta River Water Transaction Program, developed in 2009, provides a mechanism for willing individual water right holders to contribute water to instream flows for targeted, short-term periods. Generally, these contributions focus on specific spatial, temporal, quantity, and quality objectives.

Although the primary limitation to salmonid recovery in the Shasta Basin is insufficient oversummering habitat caused by elevated water temperature conditions (Jeffres et al. 2008, 2009), agricultural water use also coincides with adult FRC migration. Agricultural diversions that occur during September coincide with adult FRC staging period, when the fish hold in downstream pools (in a reach commonly called the Canyon Reach) before migrating to upstream spawning habitat. Currently, seasonal low flows during this staging period create the potential for impaired holding habitat conditions (Morgan Knechtle, personal communication, 2014). Flows increase after irrigation activities cease on October 1. Within days of the end of seasonal irrigation activities, base flows at the mouth of the Shasta River nearly double, increasing an average of $1.8{\mathrm{m}}^3/\mathrm{s}$ and largely ameliorating adverse conditions in the Canyon Reach.

Preliminary instream flow recommendations have been developed for the early (September 15–30) adult Chinook migration in the Canyon Reach of the Shasta River but have not been formally implemented (MTI and HSU 2014). Further research into the ecological relationship between stream flows and migratory cues are important for long-term management planning in the Shasta Basin (Willis et al. 2013). However, until those relationships are clarified, an assessment of existing conditions and potential short-term management strategies has been explored through the use of voluntary flow transactions during this critical fall period.

Recent returns of adult FRC to the Shasta River have been well above average. Fish counts of adults entering the Shasta River indicate that, until 2011, the number of returning FRC before October 1 was 173–3,481 fish (Chesney and Knechtle 2014). The number of FRC that returned to the Shasta River before October 1 increased to more than 5,600 fish in 2011 and more than 17,500 fish returned in late September in 2012. The higher returns have prompted resource managers to consider the implications of potential impairments associated with limited habitat and water quality conditions.

Previous surveys of spring-run Chinook in holding habitat suggest that pools have a limited holding capacity, characterized by an asymptotic relationship between specific pools and number of fish (Campbell and Moyle 1992). High fish densities (i.e., near the carrying capacity of the surrounding water) can rapidly deplete dissolved oxygen, particularly if fish are stressed (Portz et al. 2006) because of crowding, elevated temperatures, limited or suboptimal holding habitat, or other conditions. Given these chronic stressors, fish experience more severe effects (Pickering 1981; Barton 2002) from which they are less likely to recover (Portz et al. 2006; Newcomb and Pierce 2010). Previous studies have identified relationships between water quality conditions experienced by spawning adults and the fitness and development of larval and juvenile offspring (Burt et al. 2011); generally, such studies have focused on water temperature (EPA 2003).

Dissolved oxygen (DO) is another critical, although less studied, component of holding habitat. Although much uncertainty remains around the relationship between low DO and adult Chinook fecundity (Newcomb and Pierce 2010; Strange 2010), minimum DO concentrations of $5\mathrm{mg}/\mathrm{L}$ and at least 80% saturation are generally recommended for migrating adult Chinook (Westers and Pratt 1977; Bjornn and Reiser 1991; Newcomb and Pierce 2010; Strange 2010). Aquaculture experiments indicate that during periods of peak activity by migrating adults in holding ponds, DO concentrations can become less than $6\mathrm{mg}/\mathrm{L}$, even given inflows near saturation (McLean et al. 1993). Few studies have examined DO concentrations and potential effects on Chinook salmon holding in natural, riverine habitats.

Given the potential impairment suggested by previous studies, and the need to support the ongoing recovery of this sensitive species in the Shasta River, an experiment was designed to quantify potential DO conditions that could result from a combination of larger returns of adult FRC, low water conditions, and spatiotemporal water quality of holding habitat in the Canyon Reach of the Shasta River. This study was designed in two phases to quantify several factors and inform management decisions of cold-water species during a period of competing water use.

Phase 1 occurred in 2012 and tested quantification methods to evaluate impacts of water transactions on physical habitat. Positive outcomes from Phase 1 activities led to Phase 2 activities in 2013. During Phase 2, the Phase 1 quantification methodology was applied, and the analysis was extended to include the water quality elements. By using the data collected during flow transactions that occurred in 2012 and 2013, a methodology was developed and tested to evaluate water transaction effects on physical habitat in a representative pool. A spreadsheet model was developed to quantify the potential DO changes in holding habitat caused by the presence of adult FRC and assess the extrapolation of predicted outcomes to larger study reaches.

The Phase 2 hypothesis was that increasing discharge would mitigate potentially low DO conditions caused by high fish densities through increasing pool volume and decreasing pool water residence time in these principal holding habitat areas. Phase 2 was based on previous studies that identified relationships between adult Chinook, water temperature, and oxygen consumption (respiration) (Farrell et al. 2003; Clark et al. 2008).

The study area included the 13.7 km reach of the Shasta River from its confluence with the Klamath River upstream to the confluence with Yreka Creek, located near the Interstate 5 crossing (Fig. 1). This reach comprises the area where adult FRC hold before the end of irrigation season on October 1. Approximately one-third of the reach contains pools and nontraditional “mesohabitat” designations, including “run-pools” that function as holding habitat types for adult FRC (MTI and HSU 2014).Two pools, Hudson Pool and Weir Pool, were identified to test the quantification methodology. Both pools were located near existing fish counting facilities and monitoring stations providing flow data. Hudson Pool is located approximately 1.5 km upstream from the mouth of the Shasta River and 1.3 km upstream from the Shasta River Fish Counting Facility. Weir Pool is located approximately 90 m upstream from the mouth of the Shasta River and 100 m downstream from the Shasta River Fish Counting Facility.

The two-phase water transaction experiment presented an excellent opportunity to explore multiple elements of an environmental flow management action, from its conception and implementation to its quantification and assessment. The two most important contributions made by this experiment are the potential for flexibility in highly regulated basins and quantification of the water transaction benefit that directly relates the water transaction to a potential biological outcome. This approach also demonstrates the successful integration of spatiotemporal variability of water quantity and quality with a specified water management activity.

The implications of the transaction implementation and quantification methodology are extensive. The transaction implementation illustrated that a water transaction framework is a feasible and effective management strategy to provide environmental flows during a period of conflicting demands. The success of this management strategy provides flexibility to secure wet water during critical periods while functioning within the framework of adjudicated water rights. The voluntary participation by water rights holders also suggests that there may be potential to expand this program to include incentives to landowners.

The quantification method is a novel application of geomorphic principles to aquatic habitat evaluation (Nichols et al. 2013; Holmes et al. 2013). Whereas analytical techniques based on hydraulic geometry data are commonly used to develop minimum instream flows (e.g., Jowett 1997; Gippel and Stewardson 1998; Reinfelds et al. 2004), in this experiment, hydraulic geometry data were more generally used to quantify habitat capacity, travel time, and potential water quality conditions through discharge-aquatic habitat relationships. The strength of the discharge-pool volume relationship was tested in two experimental setting and yielded consistent results. Robust $R^2$ values (exceeding 0.999 for both Hudson Pool and Weir Pool) suggest that the methodology can be applied over the entire period of holding habitat use for specified locations. However, the actual relationships vary substantially between pools, suggesting that using a single representative pool as a surrogate for assessing reach-scale habitat may be inappropriate. The location and duration of a transaction, and channel geometry, also may affect the application of aquatic habitat rating curves. Alternative approaches may better characterize the relationship between discharge and aquatic habitat in some systems, particularly if alternative parameters (e.g., wetted perimeter) are selected to characterize aquatic habitat (Gippel and Stewardson 1998). Also, short-duration water transactions (several days to weeks) typically do not enable the generation of habitat rating curves using hydraulic rating methods, largely because of an inability to measure habitat variables across a sufficient range of flows.

The transaction quantification methodology also illustrates a cost-effective approach to assess transactions. In the absence of channel-changing flows, the same rating curve may be applicable from year to year, reducing the resources needed to assess available habitat and increasing the value of the methodology. Thus, whereas initially, regular field surveys are needed to develop the aquatic habitat curve, monitoring during subsequent years is simplified to monitoring discharge, water temperature, and dissolved oxygen in the study reach. New surveys would be recommended following channel-changing flows.

Finally, the fish habitat model provides a flexible, transparent method to assess the value of transaction flows. The approach described in this study illustrates how water transactions may be developed and quantified for environmental objectives. Discharge is directly related to potential biological benefits; this relationship provides transparency to natural resource managers and water rights holders. The fish habitat model can be applied to examine potential changes to discharge, water temperature, or return counts, allowing resource managers to assess whether alternative conditions may benefit from a water transaction. By applying the fish habitat model, stakeholders can consider the direct benefit of a proposed transaction flow and make an informed decision regarding participation in a transaction and valuation of the contributed water.

The key findings of the Shasta River water transaction study include

When evaluating holding habitat conditions, both water quantity and water quality are critical components. This study illustrates that given adequate water quality of influent water, potential impairments were mitigated by simply flushing the holding habitat to more rapidly replace water in the pool. Another notable finding is that not only were water quality conditions improved with the transaction, but more fish could be accommodated without offsetting the benefit from the transaction. When remaining dissolved oxygen conditions improved with the transaction, the “w/ transaction” simulation also included the additional fish that could hold in the pool given the additional habitat created by the transaction flow. Finally, the results illustrate that the benefits of the transaction varied throughout the study periods and generally had little to no effect on water quality conditions when optimal influent water quality conditions were met.

The results highlight the importance of determining life-stage-specific water quality thresholds, as outcomes vary considerably depending on the initial conditions. The fish habitat model helps identify these critical thresholds by linking physical aquatic habitat conditions to biological processes. For the objective of mitigating for high oxygen consumption by holding fish, the critical factors seem to be low flows, partially suboptimal water temperatures, and high returns that would put the pools at carrying capacity. Suboptimal conditions occurred only when previously identified water temperature thresholds were met (mean weekly maximum water temperatures of $>23°\mathrm{C}$). Notably, although the water temperature criteria included three critical thresholds, only one was exceeded to prompt high rates of oxygen consumption, resulting in near-critical DO conditions in the holding habitat. When these conditions are present, the transaction is an effective management strategy. However, as long as ambient water quality conditions met all recommended criteria, the remaining DO concentrations and percent saturation were above recommended levels. Refinements to relationships between water quantity, water quality, and biotic response would improve understanding of critical thresholds that could be used as criteria to implement transactions. Better estimates of holding habitat carrying capacity would also improve the quantification of transaction effects.

The fish habitat model provides a tool with which potential transaction participants (water rights holders or resource managers) directly assess the relationship between a transaction flow and water quality objectives. The model provides transparency for transaction participants, who can use the results to make informed decisions regarding participation in a transaction and the quantified benefit of contributed water. The basic data needs mean that the model can be implemented in any holding habitat setting, given appropriate geometry, discharge, water temperature, and dissolved oxygen data. It can also be applied to help manage other species, given appropriate oxygen consumption coefficients.

Potential long-term benefits include the continued recovery of FRC in the Shasta basin through the reduced exposure to suboptimal DO concentrations, limiting the cumulative effects of long-term stress. Studies suggest that wild adult salmon can make a full recovery from brief periods of moderately elevated oxygen consumption (Farrell et al. 2003); fecundity, offspring fitness, and other factors may benefit when potentially acute holding habitat conditions are moderated through water transactions. Ultimately, the fish habitat model provides a mechanism to balance resource use during periods of conflicting demands in a way that is transparent and quantified. This benefits water rights holders, who can maintain their existing water rights and cooperate with environmental resource managers within a flexible framework, and also benefits fish, for whom potentially stressful conditions can be addressed and mitigated by using targeted management actions.

Additional work includes consideration of the effects of a water transaction given more acute stressors, such as crowding or disease. The results of this study illustrated that the water transaction’s benefit varied with other physical and biological factors, and was negligible given desirable influent water quality. Similarly, the water transaction’s potential ability to mitigate for more acute conditions is unclear. Also, more information is needed to improve the scalability of this method. Whereas the aquatic habitat rating curves are robust for targeted areas, an improved understanding of reach-scale carrying capacity would help quantify the reach-scale effect of the transaction.

We would like to thank Sue Maurer for her extensive field observations that contributed valuable insight to this research. The authors are grateful for the support and information provided by Morgan Knechtle (California Department of Fish and Wildlife), Carson Jeffres (University of California, Davis), Peter Moyle (University of California, Davis), and Lisa Thompson (University of California, Davis). Development of the monitoring methods benefited from previous work completed in partnership with Farm-Stream Solutions, LLP, and the National Fish and Wildlife Foundation. The research for this project was supported by the TNC-NOAA Community Based Restoration Matching Grants Program and the Natural Resource Conservation Service Conservation Innovation Grant Program.