Scientific Investigations Report 2006-5111
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2006-5111
Human actions and their impacts on streams are well-documented by numerous authors (Heede and Rinne, 1990; Bayley, 1991; Gilvear and Winterbottom, 1992; Gilvear, 1993; Baker, 1994; Brookes, 1996; Stanford and others, 1996; Bravard and others, 1999; Schick and others, 1999; and McDowell, 2000). River alterations include the acute impacts of dams, channelization, water pollution, and long-term hydrologic and sediment modifications that result from these activities. The natural disturbance regimes that maintain habitats and biological communities are lost (Stanford and others, 1996). These changes can dramatically affect many aspects of aquatic ecosystems, including the habitat structure and the water quality necessary to maintain a viable fish population. To fully comprehend and appreciate changes to aquatic ecosystems, and to develop appropriate restoration plans, the condition of a river must be viewed as the result of a complex history of alterations and not just the result of current watershed conditions.
The habitat of the lower Boise River has changed dramatically over the past century, as indicated by comparison of recent (1994–2002) and historic (1867 and 1868) habitat features (MacCoy and Blew, 2005, table 3). Qualitative measures of embeddedness and substrate size collected by the USGS from 1994 to 2002, summarized in table 3, indicate an “armoring” of substrate throughout the lower Boise River. In its investigation of the availability of habitat for salmonid spawning in the lower Boise River, the IDFG also noted armoring of the bottom substrate and a lack of spawning-sized gravels (Asbridge and Bjornn, 1988). When gravels were suitable, the IDFG reported embeddedness from 25 to 49 percent (Asbridge and Bjornn, 1988). U.S. Geological Survey embeddeddness measures increased in a downstream direction with the highest embeddedness (75 percent) measured in reach 7, upstream of Middleton (Mullins, 1999a). Both historic and recent data were available for average bankfull width, channel forms, and number of sloughs in the basin (table 3). The average bankfull widths measured in most reaches in recent years have decreased to less than one-half of the historic width. For example, the historic average bankfull width of 900 ft upstream of Eagle Island has decreased to 140 ft (table 3). Historic channel forms and parafluvial surfaces (coarse sediments within the active channel and outside the wetted stream) have almost disappeared from all reaches of the lower Boise River. Gravel and sand bars were dominant downstream of Eagle Island, but these habitat features either have been stabilized or have been exposed (table 3). Historically, sloughs were abundant in the lower Boise River downstream of Eagle Island. Recently, the sloughs either have been filled in or have been converted to irrigation drains (table 3).
Cottonwood stands are considered to be major components to large gravel-bed alluvial systems (Merigliano. 1996; Poff and others, 1997) and are native to the lower Boise River. Historically, the lower Boise River’s riparian vegetation was dominated by willows and cottonwoods, owing to the dynamic flows and spring flooding that occurred. In 2002, cottonwood stands were confined to a narrow corridor at the river margins (table 3). Rood and Mahoney (1993) list several impacts on riparian cottonwood forests on dammed rivers in North America, including the lack of extreme flows that reduce forest abundance and seedling production. Today’s (2002) absence of parafluvial surfaces and the limited recruitment of new cottonwood or willow trees are largely due to the lack of extreme flows to recruit and move instream and riparian substrate. The extent to which the lower Boise River’s riparian vegetation has been affected by alteration in the natural flow regime is still unknown.
Higher than normal flows on the lower Boise River, resulting from flood-control releases and springtime irrigation returns, can last from January through June and persist all the way to the Snake River. The highest instantaneous discharge recorded between 1994 and 2002 was greater than 8,000 ft3/s measured at Glenwood in the spring of 1998. In years of severe and (or) consecutive drought, late winter and spring discharge remains low. Irrigation releases typically begin in mid-April (or following flood releases from Lucky Peak Dam during high-flow years) and continue through mid-October. Recent annual and mean monthly discharges for the lower Boise River at Glenwood and Parma illustrate the wide variation between water years and the regulated monthly discharge in the river (fig. 2). Water is diverted from the lower Boise River at several locations, and 12 major irrigation tributary/drains discharge to the lower Boise River between Lucky Peak Lake and the mouth (fig. 3).
Recent annual mean flow in the lower Boise at Diversion is less than one-half of the calculated unregulated flow. Regression equations were used to estimate unregulated flow calculated from basin characteristics at Diversion (Hortness and Berenbrock, 2001; USGS Streamstats online report, accessed June 16, 2006, at http://streamstats.usgs.gov/html/idaho.html). The estimated annual mean flow for Diversion based on unregulated flow was about 1,870 ft3/s (average standard error of 33 percent; Hortness and Berenbrock, 2001), and the regulated annual mean flow for the period of record (1987–1993) was about 830 ft3/s (U.S. Geological Survey National Water Information System Web site, accessed August 30, 2005, at http://nwis.waterdata.usgs.gov/id/nwis/qwdata).
Examination of the long-term flow record from the Boise River near Boise gaging station (USGS station 13202000) just downstream of Lucky Peak Dam shows a change in the magnitude and variability of seasonal flow following dam construction (fig. 4). Median mean monthly discharge for December and August prior to 1915 were about 1,090 and 1,200 ft3/s, respectively, with standard deviations near 460 ft3/s. In comparison, median discharge after dam construction (post-1957) for December and August were 350 and 4,020 ft3/s, respectively, with standard deviations of 350 and 640 ft3/s, respectively (U.S. Geological Survey National Water Information System Web site, accessed August 30, 2005, at http://nwis.waterdata.usgs.gov/id/nwis/qwdata). In fact, the flow regime in 2002 is opposite of pre-dam flows in December and August (fig. 4). The mean December post-dam flows are significantly lower than those in pre-dam years (P<0.001, Wilcoxon rank sum test with α=0.05); and the mean August post-dam flows are significantly higher (P<0.001, Wilcoxon rank sum test with α=0.05) than those recorded during pre-dam years.
Little information is available on the effect of flow alteration on the lower Boise River fishery, although most of the lower Boise River fish investigations have indicated that low winter flows were the reason for the decrease in the fish community (Idaho Department of Fish and Game, 1975; 1988; 2000; Mullins, 1999a). Altering the flow regime affects not only the fish community, but the entire aquatic environment. Several studies have shown that altering the natural river flow regime affects fish community biodiversity, food availability, habitat complexity, life history patterns, and connectivity (the ability of an organism to move freely through the stream hierarchy) (Ward and Stanford, 1983; Collier and others, 1996; Poff and others, 1997; Bunn and Arthington, 2002; Postel and Richter, 2003).
Recent water-quality data revealed longitudinal increases in constituent concentrations in the lower Boise (Mullins, 1998; MacCoy, 2004). Nitrogen, phosphorus, and suspended‑sediment concentrations increased between Diversion and Parma (table 4). Increased agricultural activity in the lower basin appears to increase nutrient and sediment concentrations and is directly correlated with specific conductance. Urban land use also appears to increase nutrient concentrations in the lower Boise River. Maret (1997) found that specific conductance and percentage of fine sediment in streams and rivers of southern Idaho were highly correlated with agricultural land use. The suspended-sediment criterion of 80 mg/L for no more than 14 days (Rowe and others, 2003) was exceeded most frequently at the downstream-most site at Parma. Total nitrogen concentrations at Glenwood, Middleton, and Parma exceeded National background concentrations of 1.0 mg/L (U.S. Geological Survey, 1999). Middleton and Parma had more than twice the median flow-adjusted total nitrogen concentrations compared to undeveloped basins across the country (0.26 mg/L; Clark and others, 2000). Glenwood, Middleton, and Parma also exceeded the flow-adjusted total phosphorus concentrations for undeveloped basins (0.02 mg/L; Clark and others, 2000).
In natural stream environments, the temperature regime varies longitudinally and can be modified by land-management activities that influence channel width, riparian canopy cover, pool volume, runoff timing, and instream flow. Temperature has been an influential parameter in determining fish community structure (Poole and Berman, 2001; Poole and others, 2004), and it is vital to the understanding of the fish community in the lower Boise River. The State’s daily maximum temperature standards of 22°C and 13°C (Idaho Department of Environmental Quality, 2001) to protect cold water biota and salmonid spawning, respectively, were exceeded most frequently at Middleton and Parma (MacCoy, 2004). Continuous long-term monitoring of temperature is needed in the lower Boise River to monitor compliance with these standards. The City of Boise and the USGS began continuous temperature monitoring at selected sites on the lower Boise River in 2004 as part of a modeling effort; data from those monitoring efforts have not yet been published. For more information on the Idaho State water-quality standards for temperature refer to http://www.deq.idaho.gov/water/data_reports/surface_water/monitoring/temperature-index.cfm.
Fish species that have been collected in the lower Boise River from the studies listed in table 1 are summarized in appendix A. Table 5 summarizes the occurrence of each species by river mile from USGS and IDFG sampling events. All USGS data and only that from the first pass of the IDFG depletion sampling are summarized in appendix A and used in summary statistics. Water-quality and habitat conditions have changed in the basin since the first study was completed in 1974 by the IDFG. Therefore, the species listed for a given location in table 5 may not occur at that location today. For example, in 1988 and 1992, IDFG sampled common carp near Glenwood Bridge. Studies since that time have not found common carp at that site. The occurrence of fish species found near Middleton and near the mouth also is a result of only one sampling event at each USGS sampling site between 1995 and 1996 (table 1); a different community may occur at these sites today (2004).
Twenty-two species of fish distributed among 7 families have been identified in the lower Boise River: 3 Salmonidae (2 trout and 1 whitefish), 2 Cottidae (sculpins), 3 Catostomidae (suckers), 7 Cyprinidae (minnows), 4 Centrarchidae (sunfishes), 2 Ictaluridae (catfishes), and 1 Cobitidae (loach) (table 5).
Mountain whitefish (fig. 5) is the most widely distributed salmonid, having been collected from downstream of Barber Dam to the mouth. Brown trout (Salmo trutta, fig. 6) have been collected downstream of Barber Dam to Eagle Road Bridge; although rainbow trout (fig. 7) are the least distributed of the salmonids, having been collected downstream of Barber Dam to Middleton.
Both mottled (Cottus bairdi) and shorthead (Cottus confusus) sculpin (fig. 8) have been found only downstream of Barber Dam to Glenwood Bridge. Both largescale sucker (Catostomus macrocheilus, fig. 9) and bridgelip sucker (Catostomus columbianus, fig. 10) were found at all locations sampled, whereas mountain suckers (Catostomus platyrhynchus) were collected only from Glenwood Bridge to Middleton.
Most minnow species were widely distributed. Common carp, northern pikeminnow (Ptychocheilus oregonensis, fig. 11), redside shiner (Richardsonius balteatus), longnose dace (Rhinichthys cataractae) and Umatilla dace (Rhinichthys Umatilla) were found at all sampling sites from Glenwood Bridge to the mouth. One minnow, chiselmouth (Acrocheilus alutaceus), was collected at all sampling sites downstream of Barber Dam.
Of the sunfish, bluegill were found from Glenwood Bridge to Star, smallmouth bass (Micropterus dolomieu) (fig. 12) were found from Middleton to the mouth, and largemouth bass were found from Glenwood Bridge to the mouth. Catfish were found in the lower reaches: channel catfish from Caldwell to the mouth, and tadpole madtom (Noturus gyrinus) from Star to Middleton.
The oriental weather fish (loach) is an invasive species that probably was introduced in the drains of the lower Boise River from tropical fish aquariums; it has been found from Glenwood to Middleton. The species is native to northeastern Asia and central China. They prefer still or slow-moving shallow waters in which they can burrow into the mud. They are tolerant to a wide range of water temperatures and conditions. For example, their ability to absorb oxygen from the air allows them to survive in water that is low in oxygen content. Moreover, the species has few predators and high production rates (Gulf States Marine Fisheries Commission: accessed March 2005, at http://nis.gsmfc.org/nis_factsheet.php?toc_id=192).
Fish metrics and the associated IBI scores have been summarized in table 6 for reaches in the lower Boise River sampled during low-flow periods (November to April) and from least-disturbed rivers in southern Idaho. Sampling events for low-flow periods were chosen for IBI comparison to eliminate any sampling bias due to the inability to capture fish during high flows. Although all 10 metrics are similarly weighted, the occurrence of cold water species, sculpin, and common carp, or the occurrence of tolerant species, tended to drive the IBI scores either lower or higher. For example, the lack of sculpin species, as well as the decrease in cold water and sensitive species, decreased the overall IBI score. The occurrence of common carp at reach 3 in 1988 and 1992, and reaches 7 and 9 in 1996, also decreased the IBI scores. The percentage of anomalies also were highest at reach 3 in 2003, but this metric did not appear to have a profound effect on the IBI score.
IBI scores were higher for reaches 1 through 5 in 2003 (average IBI score of near 81) than in 1988 (average IBI score of near 62), which may indicate improved water-quality conditions in the upper reaches. The IBI scores from four least-disturbed sites ranged from 65 (intermediate biotic integrity) at the South Fork Snake River near Heise in 2003 to 99 (high biotic integrity) at the South Fork Payette River near Lowman in 2001. The median IBI score for all four least-disturbed sites was 81, indicating high biotic integrity (table 6). The dominant metrics driving the IBI scores at the least-disturbed sites were percentage of sculpin and cold water species. No carp were found at the least-disturbed sites, although tolerant individuals were found at all sites except the South Fork Payette River near Lowman.
Overall, the IBI scores for the lower Boise River (calculated since 1988, n=26, median of 67, intermediate biotic integrity) were lower than those for the least-disturbed sites. The IBI scores of 11 and 40 (indicating poor biotic integrity) at the Mouth and Middleton, respectively, were much lower than the scores at the least-disturbed sites. Additional sampling in the lower Boise River would be required to better characterize the biotic integrity of the system, particularly in the lower reaches where findings are based on a small number of samples.
Most IBI score classifications for sites with multiple years of data remained similar except for reach 3 (fig. 13). In this reach, IBI scores increased from 36 (poor biotic integrity) in 1988 to 73 (intermediate biotic integrity) in 2003. Median IBI scores calculated for reaches 1 and 2 sampled between 1988 and 2003 were greater than 75, indicating high biotic integrity, and for reaches 4 and 5 were near 60, indicating intermediate biotic integrity.
IBI scores were compared to some water-quality and habitat features measured between 1994 and 2002 (tables 3 and 4). Maximum instantaneous values of water temperature were negatively correlated with IBI scores (Spearman’s rank correlation coefficient >0.5, n=10, α=0.1). A significantly negative correlation was found between IBI scores and maximum instantaneous values of specific conductance and suspended sediment (Spearman’s rank correlation coefficient >0.80, n=10, α <0.5). Recent habitat measures such as embeddedness and bankful width were similar throughout the lower Boise River (table 3) and did not correlate well with IBI scores.
A dramatic longitudinal shift occurs in feeding groups from upstream to downstream in the lower Boise River (fig. 14, appendix A). Communities numerically dominated by piscivores (fish feeding on other fish) and invertivores (fish feeding on invertebrates), and those dominated by omnivores (fish feeding on both plant and animals) and herbivores (fish feeding on primarily plants) were compared between reaches for the samples collected in 1996. There was a decrease in piscivores and invertivores in reaches 3, 7, and 9. Only 20 percent of the fish species in reach 9 fed on macroinvertebrates or other fish. MacCoy (2004) found little difference in macroinvertebrate abundance from reach 1 to reach 9, but did identify differences in macroinvertebrate tolerance levels and feeding habits. In reach 1, the macroinvertebrate community consisted of only 2 percent tolerant species and was primarily filterers. Filterers spend most of their life cycle on the surface of coarse substrate, using specialized web and filtering mechanisms to feed on suspended detritus, and they are readily available for fish to consume (Voshell, 2002). The macroinvertebrate community in reach 9 consisted of near 50 percent tolerant species, and the community was primarily gatherers. Gatherers eat fine detritus that has fallen out of suspension and is located either on bottom sediment or between coarse substrate (Voshell, 2002). These species may not be readily available for fish to consume because they usually reside between course substrate or burrow into bottom sediment.
Subbasin areas were similar except for the upstream of Mouth subbasin that included an additional 1,000 mi2 of land (table 7). No significant correlation was found between basin area and IBI scores. However, specific land-use metrics calculated for each lower Boise River subbasin appear to correlate with IBI scores. Area of developed land, impervious surface area, and number of major diversions increased in a downstream direction (table 7) and had a significant negative correlation with IBI scores (Spearman rank correlation coefficient >0.9, n=7, α<0.05). Additional fish-community data in the downstream reaches would help to test the relation between land use and fish community.
Most of the fish-community data have been collected upstream and downstream of the Lander and West Boise WTFs (reaches 2 through 5; fig. 1). Data available for reaches 1 through 5 allowed for a more in-depth analysis of fish-community responses to water-quality and habitat variables. Table 6 provides individual metrics and IBI scores for sites upstream and downstream of WTFs.
The median IBI scores calculated from 1988 through 2003 at reaches 3 and 5 downstream of WTFs (53 and 67, respectively) were lower than reaches 2 and 4 upstream of the WTFs (78 and 65, respectively). Of the sites upstream and downstream of WTFs, the average IBI score upstream of Lander WTF, reach 2, was the only site with high biotic integrity (fig. 13). More tolerant fish species were found downstream of WTFs, with the average percentage of tolerant species increasing from reaches 2 to 3 (11 to 43 percent), and reaches 4 to 5 (27 to 29 percent). The difference in tolerant species between reaches 4 and 5 was small and these sites are considered to support similar fish communities. The average percentage of cold water species decreased from reach 2 to 3 (from 59 to 38 percent) and from reach 4 to 5 (from 71 to 50 percent). Percentage of sculpin had the most dramatic longitudinal decrease for an individual metric. The average (1988–2003) percentage of sculpin in reach 1 was close to 50 percent of the fish community and decreased to an average of 33 percent in reach 2 and decreased even further to an average of 4 percent in reach 3. Sculpin have not been found since 1988 at reaches 4 and 5, and were found in very low numbers. The decrease in the sculpin population in the downstream reaches may be a factor of a combination of habitat loss, predation, and water-quality degradation. There was an increase in percentage of sculpin between 1988 and 2003 in reach 3 that suggests a slight improved water quality (fig. 15). There also was an increase in the IBI score in reach 3 from 36 (poor biotic integrity) in 1988 to 73 (intermediate biotic integrity) in 2003. The City of Boise changed from chlorination/declorination treatment of wastewater to ultraviolet treatment at the Lander WTF in 1995 (Robbin Finch, City of Boise, written commun., 2003), which may have had an impact on the sculpin population. Sculpin were not found in the downstream Lander WTF sample in August 2004 (appendix A). Previous samples from this reach were taken during low-flow periods (December to March) suggesting that sculpin may have flow or seasonal preferences in habitat. Although similar changes to wastewater treatment were made at the West Boise WTF in 2000, sculpin have not been found upstream or downstream of this facility.
Sculpin and other benthic small-bodied fish have been used as sentinel species for environmental monitoring because they may be sensitive to chemical or other stressors. They are not subject to fishing harvest and stocking which confound the analyses of game fish. Sculpin have a limited home range (less than 50 m and respond to local conditions (Brown and Downhower, 1982; Gray, 2004; Petty and Grossman, 2004). In contrast, the home range of large-bodied fish such as suckers or mountain whitefish can be more than 50 km making it difficult to relate exposure to responses (Pettit and Wallace, 1975; Baxter, 2002).
Mountain whitefish comprise a large portion of the fish biomass in the lower Boise River. The species has been classified as a cold water native, but it is intermediate in its sensitivity to degraded water-quality conditions such as siltation, elevated temperatures, and low dissolved-oxygen concentrations (Zaroban and others, 1999). Several studies have documented the adverse effects of elevated summer temperatures and suspended sediment on mountain whitefish. For example, in laboratory experiments with adult fish, increasing lethality was observed at temperatures greater than about 23ºC (Ihnat and Buckley, 1984). In July 2002, a large kill of mountain whitefish occurred in the Snake River near Waters Ferry, Idaho, following several days during which the maximum daily temperatures in the river exceeded 26ºC (Idaho Department of Environmental Quality, 2003). Mountain whitefish also showed avoidance behavior and gill damage following exposures to elevated suspended sediments during a sluicing operation in Wyoming in which sediment concentrations increased to 500 times the background concentration (Bergsted and Bergersen, 1997). In contrast, when exposed to elevated concentrations of suspended sediment that were within the natural range in an Alberta River, mountain whitefish showed no avoidance behavior (Reid and others, 2002).
Because lower Boise River mountain whitefish are somewhat sensitive
to degraded water quality and account for a large portion of the fish biomass,
additional analyses of length and weight data were made to determine the relative
“health” of the population. The relation of mountain whitefish length and weight
observed in the lower Boise River from sites upstream of Middleton is compared
to those from least-disturbed rivers in southern Idaho and to mountain whitefish
collected from throughout their natural range in the northern United States
and Canada (Rogers and others, 1996) to give a relative “condition” of individual
fish. Condition indexes or standard weight equations are considered most representative
if they are developed by using a large portion of the population across the
geographical range (Murphy and others, 1991). Rogers and others (1996) developed
a North American standard weight equation for mountain whitefish using more
than 13,000 fish from their range in the northern United States and two Canadian
provinces. Condition equations were developed from the lower Boise River and
least-disturbed rivers using regression analysis of log-transformed mountain
whitefish total lengths and weights (fig. 16). The
lower Boise River equation (
)was
very similar to both the North American standard equation (
)
and the least-disturbed site equation (
).
Lower Boise River mountain whitefish appear to be slightly smaller than those
from the least-disturbed sites in southern Idaho (fig.
17). Based on their abundance and their relative weight, the general condition
of the mountain whitefish in the lower Boise River appears to be high, although
mountain whitefish are uncommon downstream of Middleton. Possible factors that
may limit mountain whitefish populations in the downstream reaches include water-quality
conditions such as high temperatures and increased siltation.
U.S.
Department of the Interior | U.S. Geological
Survey
Persistent URL: https://pubs.water.usgs.gov/sir20065111
Page Contact Information: Publications Team
Page Last Modified: Thursday, 01-Dec-2016 19:08:35 EST
