The Yakima River Basin, in central Washington State (fig. 1), is critical habitat for spawning, rearing, and migration of multiple species of native salmonids. Of these species, Pacific salmon (Oncorhynchus spp.) receive significant attention because they rely on cold water to survive. Concerted efforts in the Yakima River Basin have improved lower Yakima River (downstream from Union Gap) water quality, habitat, and flow to support native salmonid species. However, the lower river currently (2023) experiences low dissolved oxygen and high pH values from aquatic plant production, often exceeding water-quality criteria during summer months (June–August). In addition, the lowest portion of the Yakima River Basin (downstream from Prosser) undergoes seasonal increases in water temperatures during summer months that threaten the existence of native salmon (Appel and others, 2011; Gendaszek and Appel, 2021). Specifically, warmer river water temperatures during base-flow conditions are a limiting factor for adult migration of steelhead (Oncorhynchus mykiss), sockeye salmon (Oncorhynchus nerka), spring and summer Chinook salmon (Oncorhynchus tshawytscha), Pacific lamprey (Entosphenus tridentatus), and to a lesser amount fall Chinook salmon (Oncorhynchus tshawytscha). Rapidly warming late spring and summer water temperatures can limit success of late season juvenile out-migrants (Kock and others, 2020). These challenges are compounded by warm water temperatures providing more favorable conditions for invasive predatory piscine populations compared to native salmon stocks (McMichael, 2017).

The lower Yakima River is a critical section of the Yakima River for native salmonids. It is a migration corridor for Yakima River salmonid species, and sections of the lower river provide key spawning and rearing habitat for fall and summer Chinook. In summer months, main-stem river temperatures warm from Sunnyside Dam to the confluence. At base-flow conditions, river temperatures are driven primarily by solar radiation (Stanford and others, 2002; Voss and others, 2008). As such, cooler Yakima River temperatures rapidly warm in late spring and early summer (late May–June) and cool with the onset of autumn sometime between late August and early September (Appel and others, 2011). The seasonal timing of river warming coincides with critical migration periods for several Yakima River salmonids. River temperature provides cues to salmonoid species for migration timing and impacts the amount of available dissolved oxygen in the water. Pacific salmon species have maximum temperature thresholds ranging from 22 to 26 degrees Celsius (°C), but even sub-lethal temperatures have negative effects on fish health and behavior (Brett 1952, 1971). Warm water temperatures increase physiological stress, pre-spawn mortality, susceptibility to disease, and likelihood of straying, while decreasing their swimming performance, growth, and overall survival of salmon (Coutant, 1977; Brett, 1979; Li and others, 1994; Carter, 2005). Although the Yakima Basin Steelhead Recovery Plan (Conley and others, 2009) names the six focal fish species as bull trout, spring Chinook, fall Chinook, sockeye, steelhead, and Pacific lamprey, river management decisions are often based on the needs of steelhead because they are Endangered Species Act (16 U.S.C. 1531 et seq.) listed, and benefits to steelhead are assumed to benefit other native species. In a review of numerous studies, the Washington State Department of Ecology (Ecology, 2012) concluded that daily maximum water temperatures should not exceed 21 °C to fully protect adult steelhead during migration (Carter, 2005). The main-stem water-temperature standard of 21 °C is the thermal threshold for salmonids in the Yakima River Basin downstream Cle Elum, even though there might be slight adaptive differences to the local watershed conditions because of species differences and life stage.

Adult and juvenile salmonids exploit temperature variation in river systems to thermoregulate their core body temperatures by seeking and remaining in cooler water areas when average river temperatures are warm, enabling species to survive at their tolerance limits (Neill, 1979; Torgersen and others, 1999; Breau and others, 2007). These cooler areas are thermal refuges that enable fish to metabolically recuperate as they migrate through reaches of a river, which are warm enough to cause physiological stress. Fish can detect water temperature differences to within less than 0.1 °C (McCullough, 1999) and respond to these temperature changes by moving to areas that are more favorable. Thermal heterogeneity in rivers and streams is formed from a variety of features and includes groundwater seeps, cold-water tributaries, riparian shading, deep pools, and side channels (Ebersole and others, 2003; Vaccaro, 2011; Kurylyk and others, 2015; Gendaszek and Appel, 2021). In addition, groundwater-surface water exchange and hyporheic flow can cause cold-water anomalies in rivers (Vaccaro, 2011). Features of a river that are known to contribute to hyporheic exchange include stream meanders, riffle-pool sequences, and complex instream habitats. In the Yakima River, the cold-water features are thought to provide refuge for some migrating salmon species during times when ambient river temperatures are otherwise too warm. In contrast, during winter and spring, these areas might be warmer than the main stem, likely providing rearing and growth opportunities for out-migrating juvenile salmon and other native fish species.

Torgersen and others (2012) summarized the literature on thermal refuge use by fish and discussed the hierarchical river structures that contribute to thermal refuges in basins. On a basin and subbasin level, cold-water refuges are driven by elevation, topography, geology, channel slope, and interactions between the surface and sub-surface hydrology (Torgersen and others, 2012). However, use of thermal refuges by fish is complex; there are often physiological and biological tradeoffs for fish that move to cold-water refuges. For example, fish use certain areas of rivers for feeding and shelter even if these areas provide sub-optimal temperatures (Torgersen and others, 2012). In addition, although temperatures might be more favorable in thermal refuges, the conditions of the refuge (depth, cover from predators, dissolved oxygen, connectivity, and habitat space) might not be optimal (Torgersen and others, 2012).

Thermal variability in the lower Yakima River has been studied over the past several decades, with evolving methods. In 1997, the Bureau of Reclamation (Reclamation) conducted a Forward Looking Infrared (FLIR) flight over the lowest 108 miles of the Yakima River to identify cool-water inputs and document longitudinal river temperatures (Holroyd, 1998). Quantum Spatial (2020) completed a second FLIR survey of the river from Union Gap to the mouth. FLIR imaging only documents surface water temperatures, but both surveys indicated numerous cold-water surface features (such as inflow from tributaries, springs, and hyporheic inflow), ranging from 1 to 4 °C cooler than the main stem, which could be used by salmonids. In 2008 and 2009, Benton Conservation District (BCD) documented several locations of cooler water patches on the Yakima River downstream from Prosser Dam during a longitudinal thermal survey of the lower river. Appel and others (2011) reported that daily maximum temperatures in the river exceeded 21 °C for the 2008 and 2009 summertime floats; however, thermal heterogeneity in the lower reach was reported because of cooler areas resulting from non-point source seeps, irrigation wasteways and tributaries, and deeper pools. These cooler areas are often along the riparian area and sometimes in side channels (for example, the side channel at Interstate 182 [I-182] bridge). Though the cooler areas often exceeded 21 °C, these locations ranged from 0.5 to 2.0 °C cooler than the surrounding main-stem Yakima River water temperatures (Appel and others, 2011). This temperature difference is important because a water temperature difference of 1 °C can be sufficient to reduce physiological stress in salmonids, significantly reducing metabolic costs in elevated but sub-lethal temperature conditions (Berman and Quinn, 1991). Vaccaro (2011) summarized a variety of thermal data for the Yakima River and confirmed that incoming shallow groundwater and subsurface flows (likely enhanced or driven by applied irrigation water and return flows from fields) buffered the daily rise in summer water temperatures in the Prosser Reach. Another thermal longitudinal survey during summer 2018 covering more than 100 river miles (RM), from Wapato to Richland, Washington, identified numerous in-stream cold-water patch locations from groundwater, basin drainages, and springs and irrigation returns (Gendaszek and others, 2020; Gendaszek and Appel, 2021). Longitudinal temperature data were collected simultaneously from three boats along the right bank, center, and left bank of the river and covered the same river reach studies by Appel and others (2011). In addition to these large-scale temperature surveys, continuous temperature was monitored at three fixed locations on the lower Yakima River by the U.S. Geological Survey (USGS) and BCD at Prosser (USGS site 12509489), Kiona (USGS site 12510500), and West Richland (USGS site 12511800). These data were collected from June 2018 through September 2020, and monitoring at Kiona and West Richland continued from March through October in 2021 and 2022. Data from these fixed stations show that the temperature of the main stem of the river exceeded 21 °C throughout the summer months and approached 30 °C during the warmest periods, with temperatures increasing from Prosser downstream to West Richland (U.S. Geological Survey, 2021).

Temperature levels fluctuate over the day and night from solar radiation, and air temperature increases during the daytime. Because the health of aquatic species is predominantly a function of maximum temperatures, most Washington State temperature criteria are the largest 7-day average of the daily maximum temperatures in a waterbody. However, the Washington Administrative Code 173-201A-602 (Ecology, 2011) provides a special criterion for the Yakima River from the confluence of the Cle Elum River (RM 186) to the mouth that temperature should not exceed a 1-day maximum value of 21.0 °C because of human activities, and if natural conditions exceed this value, temperature increases should not be greater than 0.3 °C. This criterion recognizes that not all waters can naturally stay below the fully protective temperature criteria. When a waterbody is naturally warmer than the forementioned criteria, the standards allow additional warming because of human activities. In this case, the combined effects of all human activities should not increase 0.3 °C more than a the naturally warmer temperature condition.

Temperature criteria for the lower Yakima River attempts to protect migrating adult salmonids by setting a maximum threshold criterion of 21 °C (Washington Administrative Code 173-201A-200; Ecology, 2020). Though thresholds for adult survivability of salmon species are unknown for Yakima River salmon stocks, work on the Willamette River Basin by White and others (2022) indicated that adult Chinook salmon migration is impaired between 20 and 23 °C, with mortality occurring above 24 °C. Because the entire lower Yakima River is not able to meet State temperature criteria during base-flow conditions, thermal heterogeneity and enhancing cold areas on the lower river are being investigated to support salmonid spawning, rearing, and migration during inhospitable temperatures. Thermal variability within the lower 100 miles of the Yakima River might provide areas of the river that meet the temperature criteria for salmonoids. Therefore, identification of in-stream thermal variability is the first step for development of future projects geared toward meeting lower Yakima River temperature criteria and regulatory standards. This study expands on the previous lower Yakima River work by collecting continuous temperature data during a two-year period (October 2018 to October 2020) to document the duration and extent of known thermal refuges in the lower Yakima River.

The lower Yakima River flows through arid south-central Washington State. It is part of the Yakima River Basin that drains approximately 6,000 square miles on the east side of the Cascade Range (fig. 1). The study area encompasses the lowest part of the Yakima River Basin, which flows through south Yakima and Benton Counties. The lower Yakima River Basin is separated from the upper Yakima River Basin by a natural break at Union Gap (Wise and others, 2009; Gendaszek and Appel, 2021). The hydrologic and geomorphic conditions of the Yakima River Basin downstream from Union Gap contribute to the formation, maintenance, and distribution of cool water anomalies on the lower Yakima River (Gendaszek and Appel, 2021). For this study, the lower Yakima River is divided into three reaches based on basin hydrology, counties, and distinct reach geomorphology. The upper-most reach extends downstream from Union Gap (RM 103.8) to Mabton, Washington (RM 60). This stretch of river flows through south Yakima County. The lowest reaches for this study, Prosser and Kiona Reaches, are in Benton County. The Prosser Reach begins upstream from Prosser, Washington (RM 47), and ends at Benton City near the USGS gaging station at Kiona near RM 30. The Kiona Reach is the lowest stretch of river and flows from Benton City to the river’s confluence (RM 0) with the Columbia River in Richland, Washington.

The Yakima River Basin headwaters are in the Cascade Range of Washington State. Reservoirs capture snowpack and spring melt. The spring freshet typically occurs between April and May. Low flows on the lower Yakima River coincide with warm temperatures during June through August. River temperatures rapidly cool between late August and early September (Reclamation, 2008). The irrigation season, when water is withdrawn from the Yakima River in Yakima and Benton Counties, runs from mid-March to mid-October. Central Washington receives an average of 7–9 inches of rain annually and relies heavily on irrigation to maintain agricultural crops and livestock, which influences water quality and seasonal flow of the Yakima River.

For decades, warm temperatures, suspended solids, turbidity, Dichlorodiphenyltrichloroethane, and other pesticides have been documented in the lower Yakima River. By the mid-1990s, water-quality evaluations by the USGS indicated that some improvements had been made, but water-quality issues related to sediment and sediment-borne pollutants remained (Rinella and others, 1999). As a result, several reaches of the lower Yakima River and several of its tributaries did not meet numerous State water-quality criteria and Federal guidelines. Consequently, Ecology (2012) placed these water bodies on Washington State’s 303(d) list (https://www.epa.gov/tmdl/approval-washington-2012-303d-list). Water-quality issues of concern in the entire Yakima River Basin include fecal coliform bacteria, suspended sediments, turbidity, toxic chemicals, pH, nutrients, dissolved oxygen, and temperature. The water-quality issues in the basin impact the beneficial uses of the water, potentially making it unsafe for drinking or recreation and threatening the health of aquatic animals and fish living in it. Currently (2023), there are two Yakima River fish species listed as “threatened” under the Endangered Species Act: mid-Columbia bull trout (U.S. Fish and Wildlife Service, 1998) and mid-Columbia steelhead (National Oceanic and Atmospheric Administration, 2014). Conley and others (2009) summarized studies in the upper and middle Yakima River that indicated elevated temperatures, toxic chemicals, lack of foraging habitat, and predation were creating obstacles for survival of these species.

The Wapato Reach has a broad alluvial floodplain with a dynamic river channel and extensive riparian forests; however, construction of bridges, dikes, and roads in this stretch have constricted or cut off portions of the floodplain (Stanford and others, 2002). The Wapato Reach floodplain supports a mixture of agricultural, conservation, and small urban areas, and the Confederated Tribes and Bands of the Yakama Nation has maintained expansive amounts of floodplains and vegetated growth along the river corridor. Significant tributaries entering the Wapato reach include Toppenish Creek, Satus Creek, and Sulphur Creek Wasteway.

Columbia River basalt dominates the Prosser reach between Prosser and Benton City restricting the river channel. There is minimal floodplain, braiding, or channel meander in this stretch of the river. This portion of the river is known as the “bedrock reach” and has minimal opportunities for shading and bank vegetation. The river is wide and shallow during base-flow conditions. Several irrigation wasteways drain into the river in the Prosser Reach (Spring/Snipes, Corral, and Knox Creeks). These wasteways are a source of cold water because of groundwater infiltration from heavy irrigation and act like natural tributaries for the lower Yakima River. Downstream from Benton City, the reach is less confined because the basalt transitions to more alluvial deposits. Alluvial islands formed by historical floods in this reach help mediate changes in channel morphology. Only one tributary drains into the river in the Kiona Reach. Amon Wasteway, at RM 2.7, is the lowest tributary to the Yakima River Basin.

The lower Yakima River supports anadromous runs of steelhead; spring, summer, and fall Chinook salmon; Coho salmon; and sockeye salmon. Juvenile salmon and Steelhead out-migrate through the lower Yakima River to the Columbia River, and adult salmonids migrate from the Columbia through the lower Yakima River. Historically, most of the fall Chinook spawned downstream from Prosser, but recent changes in aquatic vegetation have resulted in a shift of fall Chinook spawning to upstream from Prosser Dam because of decreased spawning gravel quality (Appel and others, 2011).

The seasonal duration and spatial extent of thermal refuges in the study area were determined during a 1–2-year period (October 2018–20) at seven locations. Six of these locations were between the mouth of the Yakima River and Prosser Dam, and one was upstream from the other six, near Zillah, Washington. At each monitoring location, multiple internally logging temperature sensors (Hobo Water Temperature Pro V2, Onset Computer Corp.) were installed to record the thermal characteristics of the cold-water source and to determine if the main-stem temperature near the input was influenced by this cooler input. To examine these thermal dynamics, the temperature sensors were deployed at each site in the cold-water source and upstream and downstream from the cold-water source to bracket influence on ambient river temperatures. Temperature sensors were attached to a steel spike or housed in a small PVC (Polyvinyl Chloride) casing and anchored to the riverbed (fig. 10). Temperature sensors were positioned within 1 foot of the river bottom.

Thirty temperature sensors were deployed for this study, but data from only 27 are in table 1. Three of the sensors were either lost in the field because of high flows or the sensor malfunctioned resulting in data loss. Temperature data were recorded at 30–60-minute intervals for 1–2 years. Most deployments took place in September 2018 or 2019 during lower river stage to facilitate access to the streambed.

Each monitoring location had a unique set of conditions to assess the duration and extent of cold-water areas. Results are presented individually from the most downstream location at Amon Wasteway to the most upstream location at Zillah, Washington.

Continuous temperature monitoring at seven locations in the lower Yakima River revealed consistent patterns during the September 2018 to September 2020 period. The main stem of the Yakima River exceeded 21 °C almost every day in the summer months at all locations. On average, the main˗stem maximum daily temperatures exceeded 21 °C for more for 92 days during summer months (June through August, table 2); some locations exceeded 21 °C for over 110 days. The extended time where main˗stem temperatures are above 21 °C is problematic because summer months are when summer Chinook salmon and Sockeye salmon attempt to migrate through the lower Yakima River (Kock and others, 2020). Salmonids have a preferred temperature of less than 20 °C (Neill, 1979), and water temperatures of 22 °C exceed their thermal tolerance and can be lethal (Brett, 1971). Therefore, current thermal conditions in summer months in the lower main˗stem Yakima River are detrimental to the survival of these migratory species.

Data from this study show that cold˗water inputs provide water that is substantially cooler than the main˗stem Yakima River during warm summer months (June through August), so enhancement of these sites to maximize the spatial extent of cooling could benefit salmon during these periods. Even though many tributaries exceed the maximum daily temperature of 21 °C, these water sources are often at least 2 °C cooler compared to the main stem (table 3). For example, in 2019, Amon Wasteway was at least 2 °C cooler than the main stem for 72 days from June through September, despite being warmer than 21 °C during this same period. This temperature difference might explain why adult Sockeye have been observed holding near the Amon Wasteway mouth in temperatures otherwise too warm for these species. However, the exceedingly warm temperatures in the lower Yakima River during summer months have been shown to deter adult Sockeye salmon from entering and migrating upstream (Kock and others, 2020).

Cold˗water sources from tributaries (Amon Wasteway, I˗182 groundwater spring, Corral Creek, and Snipes Creek) showed the same patterns across sites. Tributary inputs remained cooler than the main stem of the Yakima River during summer months, indicating these tributaries could act as thermal refuges for salmonids where the cooler water enters the main stem of the river. The difference between the tributary input temperatures and the main˗stem river temperatures varied across sites, with Amon Wasteway showing the smallest difference (0–4 °C), and the I˗182 spring input showing the greatest difference (0–8 °C). Corral and Spring/Snipes Creeks showed similar differences, ranging from 0 to 6 °C compared to the upstream main˗stem temperatures. However, these cooler water conditions were diminished by early to mid˗September, when the main stem of the Yakima River began to cool and the potential benefit from these tributaries decreases (table 3). Therefore, development of these cool˗water patches into potential thermal habitat would benefit the earliest adult salmonid migrants, namely Sockeye, and spring and summer Chinook salmon. Across all sites, and both water years, the main stem of the Yakima River consistently exceeded 21 °C from late May˗early June until early to mid˗September (table 2).

In contrast to summer month conditions, tributary inputs were warmer than the main stem during late autumn and winter. Amon Wasteway was about 4–6 °C warmer than the main˗stem Yakima River. Spring/Snipes and Corral Creeks were 2–4 °C warmer, and the I˗182 spring input was 8–10 °C warmer during the winter months. The warmer temperatures in winter at these sites could benefit native fish species, and the relationships between fish use in winter at these locations could be the focus of new research.

The data show that cooler water entering the main stem from tributaries results in cooler conditions downstream. Two monitoring locations, downstream from Amon Wasteway, indicated that main˗stem temperatures along the bank decreased at least 2–4 °C to a downstream distance of 175 feet from the tributary input. At the Corral Creek monitoring location, late summer temperatures in 2020 decreased by 2–4 °C at the monitoring location 100 feet downstream from the tributary input. Main˗stem temperatures downstream from Snipes Creek were the least reduced but still showed about a 2-°C difference compared to the upstream monitoring location.

Although not as pronounced as a tributary flow, cold water from non˗point sources (such as groundwater seeps) shows potential for cooling. For example, the Benton side channel showed some cooling at the downstream end (fig. 26). In addition, temperatures at the downstream end of the side channel at Spring/Snipes Creek were cooler compared to the main˗stem temperatures upstream.

Cold˗water heterogeneity in the lower Yakima River has been documented (Appel and others, 2011; Gendaszek and Appel, 2021), but how these areas are used by native resident and migratory fish species is not as well understood. Berman and Quinn (1991) studied thermoregulation by spring Chinook salmon in the upper Yakima River. They reported that adult Chinook were associated mostly with islands, pools, and rock outcroppings along the stream banks. In particular, streamflow associated with islands could be diverted through loose gravels at the head of the island and emerge downstream where it can cool as it flows through the subsurface. Kock and others (2020) monitored tagged Sockeye salmon from June through October 2019, showing that out˗migrating salmonids left the lower Yakima River during a period of increasing water temperature, and the tagged Sockeye salmon did not reenter the Yakima River until September, when water temperatures cooled. Tagged fish in that study used thermal refuges near cooler tributaries for about 30 minutes on average before leaving the lower river in June, and about 2–3 hours, on average, during migration upstream to Prosser Dam in September 2019. There is some evidence that fish use some of the cooler water inputs. Enhancing the cold˗water inputs to provide more functional thermal refuge habitat (for example, increased depth, enhanced shading, or decreased mixing with warmer river water) might help adult salmon migrate, especially if the main˗stem Yakima River continues to warm as a function of climate change.

Kurylyk and others (2015) reviewed management strategies for preserving and enhancing existing thermal refuges, and for creating new thermal refuges. For example, groundwater and surface˗water interactions and hyporheic exchange can result in thermal anomalies in rivers and streams (Brunke and Gonser, 1997). Characteristics of the river corridor that enhance groundwater and surface˗water exchange include complex instream habitats, stream meanders, side channels, and riffle˗pool sequences. Maintaining or enhancing these characteristics, when feasible, could result in lower water temperature or reduced stress to fish. In addition, reconnection of disconnected cold˗water side channels (alcoves) might provide additional thermal refuge habitat for migratory fish. Furthermore, installation of deflection barriers where cold tributaries join a warm main stem can reduce the hydraulic and thermal mixing between cold˗water plumes and the main stem, resulting in an increased area of cold water (Bilby, 1984; Kurylyk and others, 2015). Experimental riparian shading reduces surface water temperatures up to 4 °C (Ebersole and others, 2003) in small streams. Increasing shading with native plantings along the banks and confluence of tributaries (such as Spring/Snipes or Corral Creeks) could result in substantive thermal benefits. The Benton side channel is mostly shaded throughout, which might contribute to some of the cooling measured at the location. In contrast, Amon Wasteway is heavily shaded in the lower reaches prior to entering the Yakima River, yet still exceeded 21 °C for most of the summer months. These differences in how water temperatures respond to shading suggest that although shading in tributary basins could mitigate warming, it might require multiple interventions to improve cold˗water conditions.

Work to enhance cold˗water areas on the lower Yakima River is underway. Mid˗Columbia Fisheries, in coordination with Benton Conservation District, installed a thermal refuge demonstration project at RM 25 on the lower Yakima River in 2019 after monitoring of springs and seeps from 2016 to 2018 indicated cold˗water refuge potential at the site (Smith and others, 2019). At RM 25, an elevated agricultural road prevented the oxbow’s cold˗water flow from entering the Yakima River. This project replaced undersized culverts under the agricultural roads with two gated culverts to manage of the water levels. The concentrated cold˗water flows from the newly designed 520˗foot excavated oxbow into a constructed alcove on the main˗stem river. The project was designed to mimic floodplain function under normative flows by increasing the cold˗water influence of the oxbow on the river. Planting native trees and shrubs along the oxbow provides shade and decreases downstream warming in the excavated oxbow. In 2021, the project began monitoring fish use at the newly constructed alcove using passive integrated transponder tag antenna to better understand the project’s impact on migrating juvenile and adult salmon. Future thermal restoration projects are currently (2023) in consideration for the I˗182 side channel and Amon Wasteway. The I˗182 side channel showed the most potential for cold˗water refuge in this study because of the differences in cold water; however, the cold˗water source flow volume at the I˗182 site is much less than other sites investigated (Amon Wasteway, and Corral and Spring/Snipes Creeks). Furthermore, because of current hydrologic conditions (such as low flow, shallow depth, high macrophyte abundance), the site is not currently (2023) a functional habitat for migrating fish. Development of thermal refuge projects could consider additional parameters (such as flow volume, stability of the cold˗water source, and monitored temperature) for determining sites for thermal refuge habitat enhancement.

Many of the monitored cold˗water areas exceeded the 21-°C State threshold for many of the summer months and might not provide physiological refuge for all adult native species. These areas were frequently cooler than the main˗stem river temperatures during base˗flow conditions, by at least 1–2 °C, and if enhanced, might provide thermal refuge (table 3). Thermal habitat enhancement on the lower Yakima River, to insulate native adult salmonid populations against warming river temperatures driven by climate change, would likely reduce stress on salmonids.

Results from the current study provide long˗term temperature data for cold˗water areas on the lower Yakima River. Over a 2˗year period, data showed that tributary inputs are cooler in the summer months, warmer in the winter months, and these effects were translated downstream. Data from side channels indicate cooling from groundwater seeps. In general, maximum daily temperatures in the main stem exceed the 21-°C threshold for 90–110 days each summer, from late June to the end of August 2019–20. By early September, the main stem cooled appreciably, which enables the in˗migration of some Yakima River salmonids. The temperature data indicate the potential for improving habitat conditions in the future.

Future work at these cold˗water sites might include studying flow volumes and site hydrology (geometry) to ascertain usability as a thermal refuge. Management interventions (such as reconnecting floodplains and side channels, constructing deflection barriers near tributaries, enhancing shading, and maintaining river features to maximize groundwater and surface˗water exchange) could be considered to preserve, enhance, and create functional thermal refuges on the lower Yakima River. With anticipated climate change impacts in the Yakima River Basin, resulting in warmer river temperatures and lower springtime flows, and a shift in the timing of river warming to earlier in the year, thermal refuge habitat locations could become increasingly important for the basin’s migratory species.