Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
7 Fishes of the Lower I(lamath Basin Native fishes of the lower I(lamath basin are mainly anacTromous spe- cies that use productive flowing-water habitats ancT a few nonmigratory stream fishes typical of cool-water environments. Because the watershed has been cTrastically alterecT by human activities, it has become progressively less favorable for anacTromous fishes, inclucTing coho salmon. Given that the native anacTromous fishes support important tribal, sport, ancT commer- cial fisheries ancT have high iconic value, there is wiclespreacT support among stakehoiclers, both insicle ancT outside the basin, for restoration of these fishes to their earlier abundances. Restoration efforts wouicT most rationally apply to all native fishes, not just those listecT or proposed for listing uncler the fecleral EnciangerecT Species Act (ESA). If broacTly basecT restoration cloes not occur, acicTitional anacTromous species are likely to be listecT uncler state ancT fecleral enciangerecT species acts. Furthermore, because actions that are . . . . . . . perceived to benefit one species may clo harm to another, the species cannot be treated as isolatecT units. The lower I(lamath basin supports 19 species of native fishes (Table 7- 11. Thirteen (68%) of the 19 are anacTromous, ancT two are amphicTromous (larval stages in salt water); thus, 80% of the fishes require salt water to complete their life histories. The remaining four species spencT their life entirely in freshwater ancT show close taxonomic ties to fishes in the upper basin or adjacent basins. The species composition of native fishes supports geologic evidence that the I(lamath River in its present form is of relatively recent origin. One of the resident fishes (the lower I(lamath marblecT scul- pin), however, is distinctive enough to be recognized as a subspecies ancT 250
FISHES OF THE LOWER KLAMATH BASIN TABLE 7-1 Native Fishes of the Lower I(lamath River anti Its T. ret Starves 251 Namea Life History Status in Lower Klamath and Trinity Riversa Comments Pacific lamprey, Lampetra A tridentata River lamprey, L. ayersi A Uncommon I(lamath River lamprey, N Common L. similis Green sturgeon, Acipenser A med. irostris White sturgeon, A. A transmontan?vs I(lamath speckled dace, Rlvinicl~tl~ys osc?vl?vs klamatlvensis I(lamath smallscale sucker, Catostom?vs rimic?vl?vs Eulachon, Tlvaleicl~tl~ys . ,. packs Declining State special concern, proposed for listing Uncommon N Common, widespread N Common, widespread A State special concern Longfin smelt, Spirinclv?vs A tlvaleicl~tl~ys Prickly sculpin, Cott?vs asper Am Common Coastrange sculpin, Am Common C. ale?vtic?vs Lower I(lamath marbled N Common ? sculpin, C. klamatlvensis polypor?vs Threespine stickleback, Gas teros tens a cat lea this Coho salmon, Oncorl~ynclv?vs kis?vtclv Southern Oregon-Northern California ESU Chinook salmon, O. tslvawytsclva Southern Oregon-Northern A California ESU Upper I(lamath and Trinity rivers ESU Fall run TTS, probably multiple runs Poorly known Poorly known TTS, important fishery May not spawn in river Most widespread fish in basin State special concern A/N Common A Federally threatened Commonest salmon below mouth of Trinity River A Commonest salmon in both rivers Found also in Smith and Rogue rivers TTS, huge runs now gone, lowermost river only Small population . . . mainly In estuary Larvae wash into estuary Larvae wash into estuary Endemic Migratory close to ocean, anadromous; upstream forms nonmigratory Being considered for . . , - , - ~ state 1st1ng, TTS Much reduced in numbers Much reduced, focus of hatcheries (continued on next page)
252 TABLE 7-1 continued FISHES IN THE KLAMATH RIVER BASIN Status in Lower Life Klamath and Namea History Trinity Riversa Comments Late fall run A Possibly extinct Presence uncertain Spring run A Endangered but not Distinct life history, recognized as ESU adults require cold water in summer Chum salmon, O. keta A Rare, state special Southernmost run of concern species, TTS Pink salmon, O. gorb?vsclva A Extinct Breeding in basin poorly documented, TTS Steelhead (rainbow trout), A, N Common but declining; Resident populations 0. mykiss proposed for listing above barriers, TTS I(lamath Mountains Province ESU Winter run A Most abundant Distinct life history Summer run A Endangered but not Distinct life history, recognized as separate adults require cold ESU water in summer Coastal cutthroat trout, A, N State special concern Only in lower river and 0. clarki clarki tributaries, resident populations above barriers, TTS aEvolutionarily significant unit. Abbreviations: A, anadromous; Am, amphidromous; N. non-migratory; TTS, tribal trust species. several of the anaciromous species have distinct forms aciaptecI to the special conditions of the I(lamath basin. In aciclition, 17 nonnative species of fishes have been recorclecI in the basin (Table 7-21; only two of these are anaciromous. For the most part, these fishes are confined to human-createcI environments such as reser- voirs, poncis, anti clitches although inclivicluals constantly escape into the streams, where they may take advantage of favorable habitats created by human activity. In aciclition, nonnative fishes come clown continually from the upper I(lamath basin. COHO SALMON The coho salmon (Figure 7-1) once was an abundant ancI wiclely clis- tributecI species in the I(lamath River ancI its tributaries, although its his- torical numbers are poorly known because of the dominance of Chinook salmon. Snyder (1931) reported that coho were abundant in the I(lamath
FISHES OF THE LOWER KLAMATH BASIN TABLE 7-2 Nonnative Fishes of the Lower I(lamath ancI Trinity Rivers 253 Life Name History Status Comments American shad, Alosa A Uncommon . 7. Sap1d1SS1ma Goldfish, Carassius auratus Fathead minnow, Pimephales promelas N Uncommon N Uncommon Small annual run in lowermost reach of river Ponds and reservoirs Invading from upper basin where extremely abundant Golden shiner, N Uncommon Important bait fish in Notemigonus chrysoleucas California Brown bullhead, N Locally abundant Ponds and reservoirs, Ameiurus nebulosus especially Shasta River; some in main stem Wakasagi, N Locally abundant In Shastina Reservoir Hypomesus nipponensis but a few downstream records Kokanee, N Locally abundant Reservoirs Oncorhynchus nerka Brown trout, Salmo trutta N. A Common in some Sea-run adults rare streams Brook trout, N Common Only in headwater Salvelinus fontinalis streams and lakes Brook stickleback, N Locally abundant, Recent introduction Culea inconstant spreading into Scott River Green sunfish, N Common Warm streams, Lepomis cyanellus ditches, and ponds Bluegill, L. macrochirus N Common Ponds and reservoirs Pumpkinseed, L. gibbosus N Uncommon Abundant in upper basin Largemouth bass, N Common Ponds and reservoirs Micropterus salmoides Spotted bass, N Locally common Only in Trinity River M. punctulatus reservoirs Smallmouth bass, N Locally common Only in Trinity River M. dolomieui reservoirs Yellow perch, N Locally common Abundant in upper Perca flavescens basin, including Iron Gate Reservoir Abbreviations: A, anadromous; N. non-migratory. River but also incTicatecT that reports of the salmon catch probably lumpecT coho ancT Chinook. Historically, coho salmon occurrecT throughout the I(lamath River ancT its tributaries, at least to a point as high up in the system as the California-Oregon border. It is possible that they once migrated well into the upper I(lamath basin (above I(lamath Falls), as click Chinook ancT
FISHES IN THE KLAMATH RIVER BASIN FIGURE 7-1 Coho salmon male (topl, female (head), and parr. Source: Moyle 2002. Drawing by Chris M. Van Dyck. Reprinted with permission; copyright 2002, University of California Press. steelheacT, but there are no records of this, perhaps because most people are unable to distinguish them (Snycler 19311. Tociay coho salmon occupy remnants of their original range wherever suitable habitat exists ancT wherever access is not prevented by clams ancT diversions (Brown et al. 1994, Moyle 20021. Because the coho salmon is clearly in a long-term severe clecline throughout its range in California, all populations in the state have been listecT as threatened uncler both state ancT fecleral enciangerecT species acts (CDFG 20021. Life History Coho salmon in the I(lamath basin have a 3-yr life cycle (3 yr is the time from spawning of a parent to spawning of its progeny), about the first 14- 18 ma of which is spent in freshwater, after which the fish live in the ocean until they return to freshwater to spawn at the age of 3 yr. The main variation in the cycle is that a small percentage of the males return to freshwater to spawn early (in their second year, before spending a winter at sea) as " jacks." A few juveniles may also remain in freshwater for 2 yr (e.g., Bell et al. 2001), although this has not been clocumentecT for I(lamath basin coho. Aclults typically start to enter the river for spawning in late Septem-
FISHES OF THE LOWER KLAMATH BASIN 255 her. They reach peak migration strength between late October anti the micicIle of November. A few fish enter the river through the micicIle of December (USFWS, unpublishecI material, 19981. Adult coho generally enter streams when water temperatures are uncler 16°C anti rains have increased flows (Sanclercock 19911. The presence, however, of small numbers of aclult coho in the fish kill of September 2002, indicates that some coho begin migration without these stimuli. Most spawning takes place in trib- utaries, especially those with forested watersheds, but some main-stem spawning has been recorclecI (Trihey ancI Associates 19961. Spawning usu- ally takes place within a few weeks of the arrival of fish in the spawning grouncis. Females clig recicis (nests) in coarse grave! ancI spawn repeatecIly with large, hooknose males anti with small jacks over a period of a week or more. The fertilizecI eggs are covered with grave! after each spawning event. Aclults clie after spawning. Embryos clevelop anti hatch in 8-12 wk. clepencling on temperature. Alevins (hatchlings with yolk sacs attached remain in the grave! for an- other 4-10 wk (Sanclercock 19911. In forested watersheds with relatively stable slopes anti stream channels, mortality is lower for embryos anti alevins than it is in clisturbecI watersheds (Sanclercock 19911. Major sources of mortality inclucle scouring of recicis by episodes of exceptionally high flow anti smothering of embryos by silt. When most of the yolk sac is absorbed, the alevins emerge from the grave! as fry (30-35 mm) anti seek the shallow stream margins, where velocities are low anti small inverte- brates are abundant. Fry start emerging in late February anti typically reach peak abundance in March anti April, although fry-sizecI fish (up to about 50 mm) appear into June anti early July (CDFG, unpublishecI ciata, 2000, 2001,20021. Fry are nonterritorial and have a tendency to move around (I(ahier et al. 20011; this allows them to disperse. Thus, some fry are captured in outmigrant traps at the mouths of the Shasta anti Scott rivers from May to early July (CDFG, unpublishecI data, 2000, 2001, 2002), although most probably stay in the tributaries close to the areas in which they were spawned. There is no sharp separation between fry anti juvenile (Parr) stages; juveniles are typically over about 50-60 mm anti partition available habitat among themselves through aggressive behavior (Sanclercock 19911. luve- niles clevelop in streams for a year. Typical juvenile habitat consists of pools anti runs in forested streams where there is clense cover in the form of logs anti other large, woocly clebris. They require clear, well-oxygenatecI water anti low temperatures. Preferred temperatures are 12-14°C, although juve- nile coho can uncler some conditions live at 18-29°C for short periods (McCullough 1999, Moyle 20021. For example, Bisson et al. (1988) plantecI juvenile hatchery coho in streams that hacI been clevastatecI by the eruption of Mount St. Helens 3-4 yr earlier anti founcI that they showed high rates of
256 FISHES IN THE KLAMATH RIVER BASIN growth anti survival in areas where maximum ciaily temperatures regularly exceeclecI 20°C anti occasionally reachecI 29°C. Early laboratory studies in which juvenile coho were rearecI uncler constant temperatures inclicatecI that exposure to temperatures over 25°C, even for short periods, shouicI be lethal (Brett 19521. But laboratory studies in which temperatures were increased graclually (for example, 1°C/hr) suggest that lethal temperatures range from 24 to 30°C, clepencling on other conditions ancI the temperature to which the fish were originally acclimatecI (McCullough 19991. In the laboratory, juvenile coho can be rearecI at constant temperatures of 20- 23°C if foocI is unlimitecI (McCullough 19991; but in hatcheries, they typi- cally are rearecI at lower temperatures because of their reclucecI growth anti increased mortality from disease at higher temperatures. Coho at Iron Gate Hatchery are rearecI at summer temperatures near 13-15°C (Bartho- low 19951. Consistent with the experiences of hatcheries, most coho clevelop ancI grow where water temperatures are at or near the preferred temperatures for much of each 24-hr cycle. For example, in tributaries to the Matolle River, California, Welsh et al. (2001) founcI that juveniles persisted through the summer only in tributaries where the ciaily maximum temperature never exceeclecI 18°C for more than a week. In the I(lamath basin, such suitable conditions exist today mainly in portions of tributaries that are not yet excessively clisturbecI (Figure 1-11. NMFS (2002) has iclentifiecI, in aciclition to the Shasta, Scott, Salmon, anti Trinity rivers, six creeks between Iron Gate Dam anti SeiacI Valley, 13 creeks between SeiacI Valley ancI Orieans, anti 27 creeks between Orieans anti the mouth of the I(lamath as important coho habitat in the I(lamath basin. The explanation of seemingly contradictory information on tempera- ture tolerance lies in the realm of bioenergetics. luvenile coho can survive anti grow at high ciaily maximum temperatures proviclecI that (1) foocI of high quality is abundant so that foraging uses little energy anti maximum energy can be clivertecI to the high metabolic rates that accompany high temperatures, (2) refuge areas of low temperature are available so that exposure to high temperatures is not constant, anti (3) competitors or preciators are largely absent so that the fish are not forced into physiologi- cally unfavorable conditions or energetically expensive behavior (such as aggressive interactions). Thus, in the streams around Mount St. Helens cited above, foocI was abundant ancI temperatures were low much of the time. Temperatures ciroppecI well below 15°C at night even after the hottest summer clays, were below 16°C for 65-80% of the time, anti rarely ex- ceeclecI 25°C (Bisson et al. 19881. There were also areas of coo! grounc~wa- ter inflow that servecI as refuges on hot clays, although the extent of their use by coho was not clocumentecI. AncI coho were the only species present. In some rivers, however, interactions of coho with juvenile Chinook ancI
FISHES OF THE LOWER KLAMATH BASIN 257 steelheacI cause shifts of coho into energetically less favorable conditions (Healey 1991, Harvey anti Nakamoto 19961. For example, coho juveniles occupying tributaries at the Matolle River faced not only limited foocI supplies but also energetically expensive interactions with juvenile steel- heacI (Welsh et al. 2001) anti so were restricted to cool water. Observations of juvenile coho in the main-stem I(lamath River cluring summer suggest that juvenile coho live in the main stem despite tempera- tures that regularly exceed 24°C anti are usually over 20°C for much of the clay from late lune through the micicile of September (M. Rocle, CDFG, personal communication, USFWS, unpublished ciata, 20021. Temperatures at night typically cirop to 18-20°C cluring the warmest period. The coho occupy mainly pools at the mouths of inflowing streams where tempera- tures are usually 2-6°C lower than the water in the main river. The pools apparently are the only cool-water refugia in the river anti occupy only a small area (B. A. McIntosh anti H. W. Li, unpublished report, 19981. The coho in the pools appear to move into warmer water to forage on the abundant aquatic insects (D. Hillemeier, Yurok Tribe, personal communi- cation). Thus, it is at least possible that coho could, from a bioenergetic perspective, occupy the main stem. Snorkel surveys of mouth pools in 2001 show, however, that juvenile coho, in contrast with Chinook anti steelheacI, occupied 16% of the tributary-mouth pools in lune but only a single pool in August ancI September (T. Shaw, USFWS, unpublished material, 2002; Table 7-31. Most of the tributary mouth pools contain juvenile Chinook salmon, steelheacI, or both (Table 7-31. These fishes can compete with anti prey on juvenile coho (ancI each other) ancI are somewhat more tolerant of high temperatures than coho. While many of these juveniles resulted partly from natural spawning, many of them likely came from Iron Gate Hatchery. TABLE 7-3 Pools Containing luvenile Coho Salmon, Chinook Salmon, anti SteelheacI Along Main Stem of I(lamath River, 2001' as Determined in Snorkeling Surveysa No. of Mouth No. (%) of Pools with Juvenile Fish Month of Survey Pools Surveyed Coho Chinook Steelhead June 31 5 (16) 26 (84) 26 (84) July 46 7 (15) 41 (89) 43 (93) August 39 1 (3) 26 (67) 34 (87) September 32 1 (3) 13 (41) 28 (88) aThe data are comprehensive in that they include all tributaries large enough to form a cool pool, and include some tributaries below the Trinity River (e.g., Blue Creek). Source: T. Shaw, USFWS, unpublished material, 2002.
258 FISHES IN THE KLAMATH RIVER BASIN Many large (70-90 mm) juvenile Chinook from the hatchery move clown the river from late May through luly, as clo large numbers of hatchery steelheacI smolts in March ancI April. Interactions among hatchery ancI wilcI fish of all species may cause wilcI fish, which are smaller, to move clown- stream prematurely when cool-water habitat becomes limiting in summer, although this possibility has not been clocumentecI for the I(lamath River. The number of pools occupied by Chinook salmon cleclines by August ancI September, as cloes the number of Chinook present in each pool that has fish (T. Shaw, USFWS, unpublishecI material, 20021; this reflects the nor- mal outmigration of both wilcI ancI hatchery juvenile Chinook. SteelheacI remain in most pools throughout the summer. Although 2001 was a year of exceptionally low flows, Table 7-3 sug- gests that coho juveniles are uncommon in the main stem in early summer ancI become progressively less common as the season progresses. luvenile coho are virtually absent from the main stem, inclucling pools at tributary mouths, by late summer, even though juvenile Chinook ancI steelheacI per- sist in these habitats. Although the overall rarity of coho in the I(lamath basin may contribute to their absence from the mouth pools, their presence early in the summer anti the reclucecI densities of juvenile Chinook salmon as summer progresses suggest that juvenile coho wouicI be noticed by ob- servers in late summer if they were present. In one respect, the near absence of coho by late summer is surprising because juvenile coho clo move about anti shouicI be continually ciropping into the pools from tributaries (I(ahier et al. 20011. Movement of coho juveniles may be prevented by the warming or cirying of the lower reaches of tributaries in late summer. Overall, it appears that the bioenergetic clemancis of juvenile coho pre- vent them from occupying the main stem. Even with abundant foocI, the thermal refugia (the pools at mouths of tributaries) are inadequate: night- time temperatures stay too high for them, ancI the energy costs of interac- tions with Chinook ancI steelheacI, both of which are much more abundant in the pools, are probably high. Coho juveniles in the pools cluring lune ancI July may clie by late summer. Alternatively, they couicI be moving back into tributary streams, but temperatures in the lower reaches of the tributaries are similar to those of the mouth pools by late summer, ancI barriers to reentry (such as grave! bars) are often present. It is also possible that coho juveniles move to the estuary, perhaps traveling at night, when tempera- tures are lowest. Estuarine rearing of juvenile coho has been clocumentecI in other systems (Moyle 20021. A rotary-screw trap set near Orieans on the lower river for 10 yr (1991-2001) caught juvenile coho from April through luly, after which the trap was taken from the river; peak numbers were observed in May ancI lune 5 times higher than in luly (T. Shaw, USFWS, unpublishecI data, 20021. Annual seining data from the estuary (1993- 2001) indicate, however, that coho are absent from the estuary or are very
FISHES OF THE LOWER KLAMATH BASIN 259 rare from July through September, when temperatures often exceed 18°C (M. Wallace, CDFG, unpublishecI memorandum, 20021. Thus, the evidence points to the conclusion that juvenile coho are not occupying either the estuary or the main stem through the summer. One proposal for increasing the survival of juvenile coho in the main stem in summer has been to release more water from Iron Gate Reservoir to increase the habitat for juvenile coho, as clefinecI by analogy with habitat used by juvenile Chinook salmon, ancI to recluce ciaily temperature fluctua- tions in the river, thus removing the potentially lethal temperature peaks (Chapter 41. The water available from Iron Gate Reservoir, however, is quite warm in summer (18-22°C or more) anti, because it is increasingly warm as it moves downstream, is unlikely to ameliorate high temperatures very much. Mocleling suggests that aciclitional flows may incleecI recluce maximum temperatures some distance downstream but that they will also increase minimum temperatures (Chapter 41. From a bioenergetic perspec- tive, increasing minimum temperatures may be especially unfavorable for coho in the main stem because nocturnal relief from high temperatures wouicI be reclucecI. The low abundance of juvenile coho in the main stem in summer, the known thermal regimes of the main stem, ancI the bioenergetic require- ments of coho together suggest that the most crucial rearing habitat for juveniles is that of coo! tributaries. Today, coo! tributaries are mainly small streams that flow clirectly into the I(lamath or into the Shasta, Scott, Salmon, ancI Trinity rivers. With its large, coicI springs in the heac~waters, the entire Shasta River was probably once favorable habitat for coho juveniles in most years, but diversions ancI removal of riparian vegetation have macle it generally lethal thermally for salmonicis in summer. If warming occurs with future climate change, it wouicI likely exacerbate other factors that have lecI to warming of the tributaries (see Chapter 81. Even a stream that has suitable summer habitat for juvenile coho may be unsuitable in winter. Studies in Oregon ancI elsewhere indicate that overwintering habitat is a major limiting factor where summer conditions are favorable (Nickelson et al. 1992a, b). luveniles neecI refuges from win- ter peak flows. The refuges are sicle channels, small clear seasonal tributar- ies, logjams, ancI other similar areas. Simplification of channel structure through removal of woocly clebris or channelization eliminates much of the overwintering habitat. The condition of winter habitat for coho in the I(la- math basin has not been evaluatecI. Barred juveniles (Parr) transform into silvery smolts ancI begin migrat- ing downstream in the I(lamath basin between February ancI the micicIle of lune (USFWS, unpublishecI material, 1998) when they are about 10-12 cm long. Most smolts captured in the Orieans screw trap are taken in April ancI May (T. Shaw, USFWS, unpublishecI material, 2002) ancI appear in the
260 FISHES IN THE KLAMATH RIVER BASIN estuary at about the same time (M. Wallace, CDFG, unpublishecI memo- ranclum, 20021. Typically, coho smolts migrate downstream on the cleclin- ing encI of the spring hycirograph. About 60-70% of the smolts are of hatchery origin (M. Wallace, CDFG, unpublishecI memorandum, 20021. They are largely gone from the estuary by luly. The transformation of juveniles into smolts appears to be triggered by light (perhaps moonlight) ancI other changing environmental conditions. Smoltification results in pro- founcI physiological ancI morphological changes in the fish. Smolts are compellecI to move to the marine environment ancI will actively swim clown- stream to clo so, especially at night. Exact timing of the downstream move- ment appears to be affected by flow, temperature, ancI other factors (Sancler- cock 19911. Higher flows in the river in April ancI May probably decrease transit time of the smolts. Low transit time couicI recluce preciation rates ancI recluce energy consumption in swimming, although this has not been clemonstratecI in the I(lamath River. Smolts may feecI ancI grow in the estuary for a month or so before entering the ocean (e.g., Miller ancI Saciro 20031. Coho entering the ocean generally have their highest mortality rates in their first few months at sea (Pearcy 19921. The first month or so after entry may be especially impor- tant clue to preciation, which suggests that smolts will have higher survival rates if they are large before going out to sea (C. Lawrence, UCD, personal communication, 20021. Once at sea, they spencI the next 18 ma or so as immature fish that feecI voraciously on shrimp ancI small fish, ancI grow rapicIly. Ocean survival clepencis on a number of interacting factors, inclucling the abundance of prey, density of preciators, the clegree of intraspecific competition (inclucling that from hatchery fish), ancI fisheries (NRC 19961. The importance of these factors in turn clepencis on ocean conditions (pro- cluctivity, preciation, ancI other factors), which vary wiclely on both spatial ancI temporal scales. Even relatively small changes in local ancI annual fluctuations in temperature, for example, can be relatecI to changes in salmon survival rates (Downton ancI Miller 19981. Even more important are multiclecacial (20-50 yr) fluctuations in ocean conditions, which can result in drastic changes in ocean productivity for extenclecI periods of time (Hare et al. 1999 Chavez et al. 20031. ProlongecI climatic shifts have causecI significant shifts in salmonicI populations to the north or south through modification of ocean temperatures (Ishicia et al. 20011. Global warming thus couicI result in a shift in salmonicI distribution to the north, ancI cause an overall decrease in abundance of salmonicis (Ishicia et al. 2001 ). When the ocean is in a period of low productivity, survival rates may be low, ancI thus result in reclucecI runs coming into the streams. Commercial fishing is most likely to affect salmon populations cluring periods of natu-
FISHES OF THE LOWER KLAMATH BASIN 26 rally low ocean survival, but the fishery for wilcI coho salmon has been banned in California since 1997 anti the fishery for Chinook has been greatly reclucecI (Boycistun et al. 20011. A fishery for coho still exists off the Oregon coast, but only hatchery fish, which are marked, can be retained. Historically, the abundance of coho spawners reflectecI a balance be- tween ocean survival anti freshwater survival (Figure 7-21. A year of espe- cially poor conditions for survival in freshwater (e.g., created by cirought) couicI be compensated for if conditions in the ocean (e.g., high regional productivity: HoLciay anti Boehiert 2001) enhanced survival there. Persis- tently poor conditions in freshwater, such as exist throughout the Klamath basin today, make the recovery of populations clifficult, however, even when ocean conditions are favorable anti fisheries have been shut clown or reclucecI. When ocean conditions are poor, the positive effects of restoring of salmonicI habitat in streams may be masked (Lawson 1993, NRC 19961. Thus, only long-term monitoring can reveal effects of restoration. ~ ~ ~ ~~ .~ _ \ ~ / ~ / ~ rREAM Mort~lilv B ~ ~ '-~ ~ in ~ .. . . . . . . ~ Good Conditions \ ~ _ ,/ Poor cOnMdojtjtality' OCEAN Mortal V I _ ~ Good Conditions /\ ~ a, I_ TIC ~ \ ~~ / _ of\ \ /STREAM Mortality ~ \ \ <=~ ~ Poor Conditions ( E STREAM Mortality \ \ / \~_ Poor Conditions \ \ L rat E:~:~l it L ( OCEAN Mortality ~ Poor Conditions ) + Fishing _/ 1 -3 , Embryos Summer Juveniles Winter Juveniles , Smolts · Subadults [I Adults ~ Ocean Lou Entry FIGURE 7-2 Population cycles of coho salmon in California. If conditions are favorable in spawning and rearing streams (A) and conditions are also favorable for high survival rates in the ocean, large populations of salmon will result. Even if conditions for survival are relatively poor in the ocean All, large populations of coho may be maintained (although not as well as in cycle A) as long as production of coho in freshwater is high. Likewise, poor conditions in freshwater from natural causes (C) can be partially compensated for if ocean survival rates are high. If coho streams are degraded by human activity (D) and ocean conditions are poor, com- bined mortality may result in downward spiral of population size. If conditions in both fresh and salt water result in low survival (El, extinction may occur. Source: Based on information in Moyle 2002.
262 FISHES IN THE KLAMATH RIVER BASIN Hatcheries Coho salmon have been an important part of the I(lamath basin fish fauna since prehistoric times (CDFG 20021' anti many attempts have been macle to augment their populations in the I(lamath basin. The first attempt occurred in 1895' when 460~000 fish from Rec~woocI Creek part of the same evolutionarily significant unit (ESU) as I(lamath River coho were stocked in the Trinity River. It is not known whether these fish, which were taken from a small stream, survived anti contributed to later populations. Hatchery production of coho salmon in the I(lamath basin began in the ~ ~ ~ A ~ ~ ~ ~ 1 · 1 ~ v ~ ~ ~ u- ~ ~ ~ ~ season and continued ror another 5 yr. From 1919 to 1942' six aciclitional plants of hatchery-rearecI fish, all apparently of local origin, were concluctecI (CDFG 20021. The principal hatcheries today are the Iron Gate Hatchery (operating since 1966) on the I(lamath anti the Trinity River Hatchery (operating since 1963) on the Trinity River. Faced with a cleclin- ing egg supply, operators of the two hatcheries at various times brought in fertilizecI eggs from the Eel anti Noyo rivers in California anti the Cascade anti Alsea rivers in Oregon (CDFG 20021. Thus, present hatchery stocks probably are of mixed origin. Although a few hatchery fish have been plantecI in tributaries, hatchery fish are for the most part releasecI as smolts into the main stem on the assumption that they will heacI clirectly to the sea. Genetic studies of the contribution of hatchery coho to wilcI popula- tions in the I(lamath basin are not available. Brown et al. (1994) inferred that most wilcI coho stocks in the basin were partially mixed with hatchery stocks because the two hatcheries are at the far upstream encI of coho distribution anti produce large numbers of fish. In recent years, the Trinity River Hatchery has releasecI an average of 525~000 coho per year anti the Iron Gate Hatchery about 71~000 per year (CDFG 20021' although histori- cally the Iron Gate Hatchery has releasecI about 500,000 coho per year (CDFG, unpublishecI ciata, 20021. The coho typically are rearecI to the smolt stage anti marked with a maxillary clip before release, which occurs between March 15 anti May 1. They reach the estuary in concert with wilcI smolts, which peak in late May anti early rune, but typically are longer than the wilcI fish about 170-185 mm vs 135-145 mm (M. Wallace, CDFG, unpublishecI data, 20021. Although the effect of large numbers of hatchery coho on wilcI coho is not known for the I(lamath, hatchery fish may clomi- nate wilcI fish when the two are together (Rhocles anti Quinn 19981. In any event, hatchery fish are apparently more numerous than their wilcI counter- parts. In 2000 and 2001' 61% and 73%' respectively, of the smolts cap- turecI in the estuary were of hatchery origin (M. Wallace, CDFG, unpub- lished data, 20021. The percentage of hatchery fish in the spawning population has not been estimated clirectly, but Brown et al. (1994) estimated that 90°/O of the
FISHES OF THE LOWER KLAMATH BASIN 263 aclult coho in the system returned clirectly to the hatcheries or spawned in the rivers in their immediate vicinity. Other hatchery coho no cloubt stray into other streams, but the percentage is not known (CDFG 20021. In a survey of spawning coho in the Shasta River in 2001' inclivicluals from the Iron Gate ancI Trinity River hatcheries were iclentifiecI; seven of 23 car- casses examined were hatchery fish (CDFG, unpublishecI data, 20011. Re- garcIless of origin, natural-spawning coho in the basin's tributaries have managed to maintain timing of runs ancI other life-history features that fit the basin's hycirologic cycle well. Status Coho salmon populations in California in general ancT in the Klamath basin specifically have cleclinecT cTramatically in the last 50 yr (Brown et al. 1994' Weitkamp et al. 1995, CDFG 20021. The Southern Oregon-Northern California Coast (SONCC) ESU, of which Klamath stocks are part, was listecT as threatened by the National Marine Fisheries Service (NMFS) as a consequence. The California Department of Fish ancT Game (CDFG 2002) recommenclecT listing the ESU as threatened uncler the California state encian- gerecT species act, ancT the recommendation was acloptecT by the Fish ancT Game Commission as official state policy. Analysis by CDFG (2002) suggests that SONCC populations have stabilizecT at a low level since the late 1980s but couicT easily clecline again if stream conditions change. Surveys in 2001 incTicatecT that 17 (6g%) of 25 historical coho streams in the Klamath basin contained small numbers of juvenile coho (CDFG 20021. In the Trinity River, wilcT coho stocks have experienced reduction of about 96% (USFWS/HVT 19991. The role of coho spawners of hatchery origin in maintaining these populations is not known, but marked fish of hatchery origin have been founcT among the spawners. CHINOOK SALMON Chinook salmon were ancI continue to be the most abundant anaciro- mous fish in the Klamath basin, ancI their management potentially influ- ences the abundance of coho in the basin ancI vice versa. They support important commercial, sport, ancI tribal fisheries. Annual runs have ranged from about 30~000 to 240~000 fish in the last 25 yr (CDFG, unpublishecI ciata, 20021' although runs were much larger historically (Snycler 19311. Chinook salmon spawn ancI grow primarily in the main stem of the Kla- math River, in the larger tributaries (such as the Salmon, Scott, Shasta, ancI Trinity rivers), Bogus, Indian, Elk, ancI Blue creeks, ancI also in some smaller tributaries. Large numbers once spawned in the Williamson, Sprague, ancI
264 FISHES IN THE KLAMATH RIVER BASIN Wood rivers above Upper I(lamath Lake, but these runs were eliminatecI by the construction of Copco Dam in 1917 (Snycler 19311. Two ESUs are recognized for I(lamath basin Chinook: the Southern Oregon ancI Coastal (SOCC) ESU ancI the Upper I(lamath ancI Trinity rivers ESU (Myers et al. 19981. The SOCC ESU consists only of fall-run Chinook that spawn in the main-stem I(lamath roughly from the mouth of the Trinity River to the estuary ancI is tiecI to other runs in coastal streams from Cape Blanco, Oregon, to San Francisco Bay. The Upper I(lamath ancI Trinity rivers ESU encompasses the rest of the Chinook in the basin, inclucI- ing Trinity River fish. It consists of three runs (fall, late fall, ancI spring). Runs are named for the season of entry ancI migration up the river, which is not necessarily the same as the spawning time. Thus, spring-run Chinook migrate upriver cluring the spring, but spawn in the fall. The spring run differs in its life history from other runs ancI diverges slightly from them genetically as well; it may merit status as a separate ESU (Myers et al. 19981. Because studies of Chinook salmon ancI fisheries in the I(lamath basin clo not separate fish from the two ESUs (e.g., Hopelain 2001, Prager ancI Mohr 2001), Chinook salmon are treated here as either fall-run or spring-run. The late fall-run Chinook in the basin is either extinct or poorly clocumentecI (Moyle 20021. The vast majority of the fish today are fall-run fish of both wilcI ancI hatchery origin. Fall-Run Chinook Salmon Life History of Fall-Run Chinook Salmon Fall-run Chinook in the I(lamath have the classic ocean type of life- history pattern: juveniles spencI less than a year in freshwater (Healey 19911. This pattern allows the salmon to take advantage of streams in which conditions may become unfavorable by late summer (Moyle 20021. Aclult Chinook salmon that have the ocean type of life-history pattern also typi- cally spawn in the main channels of large rivers ancI their major tributaries. Historically, the fall run in the I(lamath was known as a summer run because fish started entering the estuary ancI lower river in luly, peaked in August, ancI were largely finished by late September (Snycler 19311. Today, the run peaks in early September ancI continues through late October (Trihey ancI Associates 1996; USFWS, unpublishecI material, 19981. The 2- to 4-wk shift in run timing suggests that the main-stem I(lamath ancI Trinity rivers have become less favorable to aclult salmon in summer, presumably because of high temperature (Bartholow 1995), or perhaps because of excessive harvest of early run fish. Even with the shift in timing, temperature cluring the time of the spawning run probably is stressful to the migrating salmon ancI may result in increased mortality of spawning aclults. Literature re-
FISHES OF THE LOWER KLAMATH BASIN 265 viewed by Bartholow (1995) suggests that temperatures under 14°C are optimal for aclult migration ancI that chronic exposure of migrating aclults to 17-20°C can be lethal, although they can enclure temperatures as high as 24°C for short periods. McCullough (1999, p.75), commenting on adult migration primarily with ciata from the Columbia River, conclucles that spring Chinook migrate at 3.3-13.3°C, summer Chinook migrate at 13.9- 20.0°C, and fall Chinook migrate at 10.6-19.4°C. Fall-run Chinook reach upstream spawning grouncis 2-4 wk after they enter the river; they then spawn ancI clie (USGS 19981. In 2001, adult Chinook were first recorclecI entering the Shasta River on September 11; the run peaked on October 1, ancI 95°/O of the run hacI entered the system by October 27 (CDFG, unpublished data, 20011. In 1993-1996, spawning in the reach between SeiacI Creek ancI within 40 mi of Iron Gate Dam on the main stem began in the second week of October, peaked in the last week of October, ancI was completecI by the micicIle of November (USGS 19981. This spawning period coincides with cleclining temperatures, which by early November are within the optimal range for incubation of cleveloping em- bryos (4-12°C); 2-16°C is the range for 50°/O mortality (Healey 1991, Myrick and Cech 20011. Time to emergence from the grave! varies with the temperature regime to which the embryos are exposed. In the main-stem I(lamath River, alevins can emerge from early February through early April, but peak times vary from year to year (USGS 19981. In the Shasta River, newly emerged fry have been captured as early as the micicIle of lanuary (USGS 19981. After they emerge, fry disperse downstream, ancI many then take up residence in shallow water on the stream eciges, often in flooclecI vegetation, where they may remain for various periods. As they grow larger, they move into faster water. Some fry, however, keep moving after emergence ancI reach the estuary for rearing (Healey 19911. This pattern seems to be common in the I(lamath River, although the small juveniles in the estuary leave, apparently for the ocean, after only a few weeks (Wallace 20001. The time that juve- niles spencI in the estuary may clepencI on upstream conditions (Wallace ancI Collins 19971. When river conditions are relatively poor (for example, warm), the juveniles move into the estuary when smaller ancI stay there longer. In other systems, juveniles may live in the estuary through the smolt stage ancI this can be important for allowing juvenile Chinook of the ocean life-history pattern to grow to larger sizes before entering the ocean (Healey 19911. luveniles are found in the I(lamath estuary from March through September (the sampling season), over which time new fish constantly enter ancI oicler fish leave (Wallace 2000; unpublishecI ciata 20021. Other juvenile fall-run Chinook rear in the river or large tributaries for 3-9 ma, but downstream movement is fairly continuous. During lune ancI luly, movement of wilcI fish may be stimulatecI by the release of millions of
266 FISHES IN THE KLAMATH RIVER BASIN juvenile salmon from Iron Gate Hatchery; the hatchery fish probably com- pete with wilcI fish for space. An outmigrant trap set at Big Bar, near Orieans, for 10 yr (1991-2001) captured juveniles from late February through late August, although the trap was usually set only from early April through July (T. Shaw, USFWS, unpublishecI material, 2002~. Time of peak catch varied from year to year but usually was between late May ancI the micicIle of luly. Outmigrant traps on the Scott anti Shasta rivers catch Chinook fry, part, ancI smolts from early February through luly in most years. Peak numbers occur in March or early April for the Shasta River ancI from the micicIle of April to the micicIle of May in the Scott River. A survey of main-stem pools at the mouths of creeks in 2001 indicates that juveniles can be founcI in the main stem from lanuary through September, but abundances are consiclerably reclucecI by August ancI September (T. Shaw, USFWS, unpublishecI material, 2002~. Thus, there appears to be a steady movement of fish clown the main stem throughout summer; the fry stay for various periods in the main stem at temperatures of 19-24°C. That pattern is consistent with the thermal tolerances of juvenile Chinook salmon, which can feecI ancI grow at continuous temperatures up to 24°C when foocI is abundant anti other conditions are not stressful (Myrick anti Cech 2001~. Uncler constant laboratory conditions, optimal temperatures for growth are around 13-16°C. Continuous exposure to 25°C or higher is invariably lethal (McCullough 19991. luveniles can, however, tolerate higher tempera- tures (28-29°C) for short periods. Depending on their thermal history, fish in wilcI populations may experience high mortality at temperatures as low as about 22-23°C (McCullough 19991. In the lower I(lamath River, the presence in late summer of refuges that are 1-4°C cooler than the main stem ancI lower temperatures at night may increase the ability of the fry to grow. The abundance of invertebrate foocI also makes the environment bioenergetically favorable, although intense intraspecific competition may occur around the refuge pools. What limits the survival of Chinook fry in the main stem is not known. FoocI is apparently abundant, anti summer temperatures, although poten- tially stressful, are rarely lethal. It is possible that shallow-water rearing habitat is limiting for fry, especially if there is competition for space with other salmonicis, inclucling hatchery-rearecI Chinook ancI steelheacI (e.g., I(elsey et al. 2002~. Fry (under 50 mm) require shallow edge habitat for feeding ancI protection from predators. Thus, increasing flows to increase ecige habitat may be clesirable for as long as small fish are present. Some fall-run Chinook apparently remain in the river long enough to become smolts before they migrate to the sea; the rest clo not (migration to the estuary is known to occur without smoltification in some cases). Timing of migration may be critical. Baker et al. (1995) indicated that prolonged exposure of outmigrating smolts to temperatures of 22-24°C in the Sacra-
FISHES OF THE LOWER KLAMATH BASIN 267 mento River resultecI in high mortality. luvenile Chinook salmon that trans- form into smolts at temperatures over 18°C may have low ability to survive in seawater (Myrick anti Cech 20011. Once the Chinook are at sea, they grow rapicIly on a cliet of shrimp ancI small fish (Healey 19911. They can move wiclely through the ocean but typically are most abundant in coastal waters, where growth ancI survival are strongly influencecI by ocean conditions. They return to the I(lamath mainly as 3-yr-oicI fish, but jacks (2-yr-oicI males) ancI 4-yr-oicI fish also are common. Hatcheries Hatcheries for Chinook salmon have been operating continuously since 1917. Both the Iron Gate Hatchery ancI the Trinity River Hatchery produce large numbers of spring-run ~ 13 % ~ ancI fall-run ~ 87% ~ juvenile Chinook of native stock (Myers et al. 19981. The hatcheries release 7-12 million juve- niles into the river each year (about 70°/O from the Iron Gate Hatchery, all fall-run). The fish generally have been releasecI over 2-3 clays in late May or early lune ancI take 1-2 ma (mean, 31 clays) to reach the estuary (M. Wallace, CDFG, unpublishecI data, 2002), although some fish probably remain in pools for most of summer. Smaller fish take longer than larger fish to reach the estuary, but because they are feeding ancI growing on the way downstream, all juveniles are about the same size when they reach it. About 40°/O of the juvenile fish in the estuary in 2000 were of hatchery origin (CDFG, unpublishecI ciata, 20001; this is presumably a fairly typical figure. Aclult Chinook returning to the hatcheries are roughly one-thircI of the total run 30°/O in 1999' 44% in 2000, ancI 28% 2001 (CDFG, unpub- lishecI ciata,20011. There has been an increase in the percentage of hatchery fish in the run in recent years up from 18% in 1978-1982, anti 26% in 1991-1995 (Myers et al. 19981. Their contribution to natural spawning is not known, but estimates for the Trinity River suggest that it is roughly the same as the percentage of hatchery returns (Myers et al. 19981. Status The fall-run Chinook salmon in the I(lamath basin overall probably has cleclinecI in abundance, but it is still the most abundant salmonicI in the basin. In the first major stucly of I(lamath salmon, Snyder (1931) statecI that "the actual contribution of the river to the entire salmon catch of the state is not known, nor can it be known.... The fishery of the I(lamath is particularly important, however, because of the possibility of maintaining it, while that of the Sacramento probably is cloomecI to even greater clepletion than now appears." Snyder clicI not provide estimates of run sizes, but the river harvest
268 FISHES IN THE KLAMATH RIVER BASIN alone in 1916-1927 was 35~000-70~000 fish (as estimated from Snyder's ciata showing an average weight of 14 Ib/fish ancI a harvest of 500,000- 1,000,000 Ib each year). If, as Snycler's ciata suggest, the river harvest was roughly 25% of the ocean harvest in this period, annual total catches were probably 120~000-250~000 fish. This in turn suggests that the number of potential spawners in the river was consiclerably higher than the number spawning in the river today. Since 1978' annual escapement has varied from 30~000 to 230~000 adults. In both 2000 and 2001' runs were over 200~000 fish. If it is assumed that fish returning to the hatcheries are, on the average, 30% of the population ancI that 30% of the natural spawners are also hatchery fish, then roughly half the run consists of salmon of natural origin (inclucling progeny of hatchery fish that spawned in the wilcI). Aciclitional evidence of clecline is the exclusion of salmon from the river ancI its tributaries above Iron Gate Dam in Oregon, where fairly large numbers spawned, ancI the clocumentecI clecline of the runs in the Shasta River. The Shasta River once was one of the most productive salmon streams in California because of its combination of continuous flows of coicI water from springs, low gradients, ancI naturally productive waters. The run was probably aireacly in clecline by the 1930s' when as many as 80~000 spawn- ers were observed. By 1948' the all-time low of 37 fish was reachecI. Since then, run sizes have been variable but have mostly been well below 10,000. Wales (1951) noted that the clecline hacI multiple causes, most relatecI to fisheries ancI lancI use in the basin, but laicI much of the blame on I(lamath River lampreys: the lampreys preyecI extensively on the salmon in the main stem when low flows clelayecI their entry into the Shasta River. In some respect, it is remarkable that fall-run Chinook salmon in the I(lamath River are cloing as well as they seem to be. Both aclults migrating upstream ancI juveniles moving downstream face water temperatures that are bioenergetically unsuitable or even lethal. As explainecI later in this chapter, the vulnerability of the run to stressful conditions was ciramatically clemonstratecI by the mortality of thousands of aclult Chinook in the lower river in late September 2002. Spring-Run Chinook Life History Like coho, spring-run Chinook have a stream type of life history, which means that juveniles remain in streams for a year or more before moving to the sea (Healey 19911. In aciclition, the aclults typically enter freshwater before their gonacis are fully clevelopecI ancI hoicI in creep pools for 2-4 ma before spawning. In California, this strategy allows salmon to spawn ancI clevelop in upstream reaches of tributaries that often are inaccessible to fall-
FISHES OF THE LOWER KLAMATH BASIN 269 run Chinook because of low flows and high temperatures in the lower reaches during fall (Moyle 20021. Major disadvantages of such a life-history pattern in the present system are that low flows and high temperatures during the adult and smolt migration periods can prevent the fish from reaching their destinations or greatly increase mortality during migration (Moyle et al. 1995, Trihey and Associates 19961. Spring-run Chinook enter the I(lamath system from April to luly, al- though the fish that appear later apparently are mainly of hatchery origin (Barnhart 19941. The Chinook aggregate in deep pools, where they hold through September. Temperatures below 16°C generally are regarded as necessary for spring-run Chinook because susceptibility to disease and other sources of mortality and loss of viability of eggs increase as temperature increases (McCullough 19991. In the Salmon River, temperatures of pools holding spring-run Chinook often exceed 20°C (West 1991, Moyle et al. 19951. Spawning peaks in October. Fry emerge from the reads from March to early Tune; the fish reside through the summer in the coo! headwaters . ~ , ~ (West 19911. Because most of the streams in which they reside also are likely to be used by juvenile coho salmon, interactions between the two species are likely (see O'Neal 2002 for information specific to the I(la- math). Some juveniles may move down to the estuary as temperatures decline in October, although most do not move out until the following spring (Trihey and Associates 19961; they spend summer in the same reaches as the holding adults. More precise details of the life history of spring-run Chinook in the I(lamath basin are unavailable. Status Spring-run Chinook may once have been nearly as abundant as fall-run Chinook in the I(lamath basin. Perhaps 100,000 fish spread into tributaries throughout the basin, including the Sprague and Williamson rivers in Or- egon (Moyle 20021. The Shasta, Scott, and Salmon rivers all supported large runs. Spring-run Chinook suffered precipitous decline in the 19th century caused by hydraulic mining, dams, diversions, and fishing (Snyder 19311. The large run in the Shasta River disappeared coincidentally with the construction of Dwinnell Dam in 1926 (Moyle et al. 19951. In the middle to late 20th century, the decline of the depleted populations contin- ued as a result of further dam construction (for example, of Trinity and Iron Gate Dams) and, in 1964, heavy sedimentation of habitat that resulted from catastrophic landslides due to heavy rains on soils denuded by logging (Campbell and Moyle 19911. By the 1980s, spring-run Chinook had been largely eliminated from much of their former habitats because the cold, clear water and deep pools that they require were either absent or inacces- sible. In the I(lamath River drainage above the Trinity, only the population
270 FISHES IN THE KLAMATH RIVER BASIN in the Salmon River ancI Wooley Creek remains; it has annual runs of 150- 1,500 fish (Campbell ancI Moyle 1991, Barnhart 19941. Numbers of fish in the area continue to clecline (Moyle 20021. Because the Trinity River run of several thousand fish per year is apparently sustained largely by the Trinity River Hatchery, the Salmon River population may be the last wilcI (natu- rally spawning) population in the basin. The Trinity River Hatchery re- leases over 1 million juvenile spring-run Chinook every year, usually in the first week of rune. Apparently, all spawners in the main-stem Trinity River below Lewiston Dam are of hatchery origin. NMFS clebatecI designation of the I(lamath spring-run Chinook as a distinct ESU, but cleciclecI that it was too closely relatecI to fall-run Chinook to justify separation (Myers et al. 19981. Nevertheless, the presence of genetic differences ancI of great differences in life history suggest that it shouicI be managed as a distinct ESU (as was clone for the Sacramento River spring-run Chinook) or as a distinct population segment. Protection ancI restoration of streams used by spring-run Chinook salmon wouicI provide aciclitional protection for coho salmon because the two salmon have similar temperature ancI habitat requirements. STEELHEAD SteelheacI (anaciromous rainbow trout) are wiclely clistributecI ancI com- mon in the I(lamath basin. They consistently co-occur with coho salmon in streams, ancI the juveniles of the two species can have strong interactions (e.g., Harvey ancI Nakamoto 19961. All populations are consiclerecI by NMFS to be part of the I(lamath Mountains Province ESU. Besicles having genetic traits in common, the populations share a life-history stage callecI the half-pouncler, which is an immature fish that migrates to the sea in spring but returns to spencI the next winter in freshwater (Busby et al.1994, Moyle 20021. Two basic life-history strategies are recognized in the basin: summer steelheacI (stream-maturing) ancI winter steelheacI (ocean-maturing). Barnhart (1994) ancI Hopelain (1998) clivicle the winter steelheacI further into early (fall-run) ancI late (winter-run), but the two forms have similar life histories ancI will be treated together here as winter steelheacI. Winter Steelheac! Life History Winter steelheacI are the most wiclely clistributecI anaciromous salmo- nicis in North America. I(ey factors in their success in a wicle variety of habitats inclucle an aciaptable life history, higher physiological tolerances than those of other salmonicis, ancI ability to spawn more than once (Moyle
FISHES OF THE LOWER KLAMATH BASIN 27 20021. The flexibility in life-history pattern is reflectecI in the fact that most populations have juveniles that spencI 1, 2, or 3 yr in freshwater ancI aclults that spencI 2-4 yr in the ocean ancI return one to four times to spawn. This variability virtually ensures that runs can continue through periods of acI- verse conditions unless the stream habitat becomes chronically unfavorable to survival of steelheacI. Winter steelheacI enter the I(lamath River from late August to February (Barnhart 19941. They disperse throughout the lower basin ancI spawn mainly in tributaries but also show some main-stem spawning. Snyder (1933) noted that fish entering the Shasta River in 1932 came in bursts of 2-3 clays over a 7-wk period. Spawning, which can take place any time from January through April, apparently peaks in February ancI March. Mature fish first return to spawn after a year, at 40-65 cm; the smallest fish are those that spent a winter in freshwater as half-pounclers (Hopelain 19981. Up to 30°/O of the mature fish spawn a second time, after another year at sea; up to 20% spawn a thircI time; ancI a very few a fourth time (Hopelain 1998). Fry emerge from the grave! in spring ancI most (80-90%) spencI 2 yr in freshwater before going to sea. The rest spencI either 1 or 3 yr in freshwater (I(esner ancI Barnhart 1972, Hopelain 19981. The juveniles occupy virtually all habitats in the basin in which conditions are physiologically suitable. They can tolerate minimal depths ancI flows ancI so can be founcI in the smallest accessible tributaries ancI in the main river channels. Although spawning occurs mainly in tributaries, the juveniles distribute themselves wiclely, ancI many move into the main stem. For example, large numbers of Parr have been observed moving out of the Scott ancI Shasta rivers in early luly (W.R. Chesney, CDFG, unpublishecI reports, 2000, 20021. Habitat preferences change with size: bigger fish are more inclinecI to use pools or creep runs ancI riffles, ancI the larger juveniles prefer water at least about 50- 100 cm creep with water-column velocities of 10-30 cm/s ancI creep cover (Moyle 20021. luveniles feecI primarily on invertebrates, especially drifting aquatic ancI terrestrial insects, but fish (inclucling small salmon) can be an important part of the cliet of larger inclivicluals. Aggressive 2-yr-oicI steel- heacI ~ 14-17 cm) often dominate pools. A key to the success of steelheacI in freshwater is their thermal toler- ance, which is higher than that of most other salmonicis. Preferred tempera- tures in the fielcI are usually 15-18°C, but juveniles regularly persist in water where daytime temperatures reach 26-27°C (Moyle 20021. Long- term exposure to temperatures continuously above 24°C, however, is usu- ally lethal. SteelheacI cope with high temperatures by finding thermal ref- uges (springs, stratified pools, ancI so on) or by living in areas where nocturnal temperatures drop below the threshoicI of stress. Persistence in thermally stressful areas requires abundant foocI, which steelheacI will shift
272 FISHES IN THE KLAMATH RIVER BASIN their behavior to fincI. Thus, Smith ancI Li (1983) found that juvenile steel- heacI persisted in a small California stream in which daytime temperatures sometimes reachecI 27°C for short periods by moving into riffles where foocI was most abundant; these fish, however, were at their bioenergetic limits for survival. Overall, the ability of steelheacI to thrive uncler the summer temperatures experienced in the lower I(lamath ancI the different habitat requirements of juvenile steelheacI of different sizes indicate that they will benefit from the expansion of habitat created by increased flows in the main-stem I(lamath ancI tributaries, as long as water quality, especially temperature, remains suitable for them. SteelheacI juveniles become smolts ancI move into the estuary from early April to the micicIle of May (I(esner ancI Barnhart 19721. Small numbers continue to trickle into the estuary all summer (M. Wallace, CDFG, unpub- lishecI ciata, 20021. A majority of the early fish that return each year to the river in September are immature (half-pounclers, 25-35 cm). These fish usually stay in the lower main stem of the I(lamath through March before returning to the sea. This life-history trait allows the steelheacI to consume eggs of the large numbers of Chinook salmon that enter the river at the same time (USGS 19981. Half-pounclers that return to spawn in the follow- ing winter are much smaller (40-50 cm), however, than the first-time spawn- ers that skipped the half-pouncler stage (55-65 cm) (Hopelain 19981. Hatcheries The Iron Gate Hatchery produces about 200,000 ancI the Trinity River Hatchery about 800~000 winter steelheacI smolts per year (Busby et al. 19941. The fish are releasecI into the rivers in the last 2 wk of March, ancI most reach the estuary about a month later (M. Wallace, CDFG, personal communication, 2002), coincident with the emigration of wilcI smolts. Di- ets of outmigrating smolts are similar to those of wilcI smolts, although the consumption of a greater variety of taxa ancI fewer organisms by the hatch- ery fish than by wilcI fish suggests that they have lower feeding efficiency than wilcI fish (Boles 19901. Otherwise, the interactions between hatchery ancI wilcI fish in the I(lamath are not known, although hatchery steelheacI releasecI into a stream will dominate the wilcI steelheacI (McMichae! et al. 1999), potentially increasing the mortality in wilcI fish from predation, injury, or reclucecI feeding. Hatchery steelheacI also can have adverse effects on juveniles of other salmonicis, especially Chinook ancI coho salmon, through aggressive behavior ancI predation (I(elsey et al. 20021. In the 1970s ancI early 1980s' aclults of hatchery origin macle up about 8% of the run of I(lamath River steelheacI ancI 20-34% of the run in the Trinity River (Busby et al. 19941. As numbers of wilcI steelheacI clecline, the percentage of hatchery fish in the population presumably will increase.
FISHES OF THE LOWER KLAMATH BASIN 273 There is some indication that the runs most heavily influencecI by hatchery steelheacI in the Trinity River have a lower frequency of half-pounclers in the population than clo wilcI populations (Hopelain 19981. Status Historical numbers of winter steelheacI in the I(lamath River are not known, but total run sizes in the 1960s were estimated at about 170,000 for the I(lamath ancI 50,000 for the Trinity (Busby et al. 19941. Historical numbers for the I(lamath River above the Trinity uncloubtecIly were much higher because by 1917 all access to the upper basin was eliminatecI ancI habitat in the tributaries was greatly clegraclecI or blockecI. In the 1970s, I(lamath River runs were estimated to average around 129,000; by the 1980s, they hacI ciroppecI to around 100~000 (Bushy et al. 1994). Similar trencis were noted for the Trinity River. Numbers presumably are still cleclining, although all estimates of abundance, past anti present, are very shaky. NMFS consiclerecI winter steelheacI in the I(lamath to be in low abundance ancI to be at some risk of extinction (Busby et al. 1994) but has not listecI them uncler the ESA. ~ ~ J / Summer Steelheac! Life History Summer (spring-run) steelheacI have the same relationship to winter steelheacI that spring-run Chinook salmon have to fall-run Chinook salmon in the I(lamath River. They are closely relatecI but have different life histo- ries. Summer steelheacI enter the I(lamath River as immature fish from May to luly ancI migrate upstream to the coo! waters of the larger tributaries (Barnhart 1994, Moyle 20021. They hoicI in creep pools roughly until De- cember, when they spawn. Temperature requirements of aclult summer steel- heacI are not well clocumentecI, but maximum daytime temperatures of less than 16°C seem to be optimal, ancI temperatures above 20°C increase stress substantially (Moyle et al. 1995) through susceptibility to starvation (they clo not feecI much while hoicling) ancI disease. High temperatures also cle- crease viability of eggs insicle the females. luveniles probably occupy mainly the same upper stream reaches in which they were spawned, that is, above the areas in which most winter steelheacI spawn ancI rear but where coho are likely to be present. Other aspects of their life history are similar to those of winter steelheacI, inclucling a predominance of 2-yr-oicI smolts ancI the presence of half-pounclers (Hopelain 19981. There is some evidence, however, that summer steelheacI have higher repeat spawning rates ancI grow larger in the ocean (Hopelain 19981. As is the case with spring-run
274 FISHES IN THE KLAMATH RIVER BASIN Chinook salmon, major clisacivantages of the summer steelheacl's life-his- tory pattern in the present system are that reclucecI flows ancI increased temperatures cluring the aclult ancI smolt migration periods prevent the fish from reaching their destinations or greatly increase their mortality cluring migration (Moyle et al. 1995, Trihey ancI Associates 19961. Status Summer steelheacI once were wiclely clistributecI in the I(lamath ancI Trinity basins ancI were present in most heac~waters of the larger tributaries (Barnhart 19941. In the 1990s, estimated numbers were 1,000-1,500 aclults cliviclecI among eight populations; the largest numbers were in Dillon ancI Clear creeks (Barnhart 1994, Moyle et al. 1995, Moyle 20021. Numbers presumably are still cleclining because of loss of habitat, poaching in sum- mer, ancI reclucecI access to upstream areas cluring migration periods as a result of diversions. Summer steelheacI ancI winter steelheacI probably are different ESUs. NMFS considers the stocks clepressecI ancI in cianger of extinction (Busby et al. 19941. Summer steelheacI are not proclucecI by I(lamath basin hatcheries. OTHER FISHES Pink Salmon Small runs of pink salmon probably once existed in the I(lamath River ancI elsewhere on the coast. The pink salmon now appears to be extirpated as a breeding species in California, although inclivicluals stray occasionally into coastal streams (Moyle et al. 1995, Moyle 20021. Chum Salmon Periodic observations of aclult chum salmon ancI the regular collection of small numbers of young suggest that this species continues to maintain a small population in both the I(lamath ancI Trinity rivers (Moyle 20021. It was more abundant in the past ancI occasionally was harvested, but it has never been present in large numbers. The run in the I(lamath basin is the southernmost of the species. The life history of this species in the I(lamath basin, inclucling timing of spawning runs ancI outmigration of juveniles, is probably similar to that of fall-run Chinook salmon. Coastal Cutthroat Trout Because of their similarity to the more abundant steelheacI, coastal cutthroat trout have been largely overiookecI in the I(lamath basin. They
FISHES OF THE LOWER KLAMATH BASIN 275 occur mainly in the smaller tributaries to the main stem within about 22 mi of the estuary. They also have been observed further upstream in tributaries to the Trinity River (Moyle et al. 19951. Their life history in the I(lamath River is poorly clocumentecI but is apparently similar to that of winter steelheacI. Aclults enter the river for spawning in September ancI October, ancI juveniles grow in the streams for 1-3 yr before going to sea. Cutthroat trout can spawn two to four times. Competition for space by spawners ancI juveniles with the dominant steelheacI is reclucecI by the ability of cutthroat to use habitats higher in the watersheds than are typically used by steelheacI (Moyle 2002~. Voight and Gale (1998) suggest that in small tributaries in the lower 22 mi of the I(lamath River, cutthroat may actually be more abundant in heac~water streams than they were historically because they have become resident above migration barriers created by human activities, such as log jams ancI clebris flows. The life history of one such population on the nearby Smith River is clocumentecI by Rails back ancI Harvey (2001~. The general absence of cutthroat trout from streams higher in the I(lamath basin presumably results from their general intolerance of water that exceeds 18°C (Moyle 2002) and from competition with the more tolerant steelheacI ancI perhaps other salmonicis. luveniles move downstream when they reach 12-20 cm cluring April through rune, coinciclentally with the outmigration of juvenile Chinook salmon, a major prey (Hayclen ancI Gale 1999, Moyle 2002~. Adults apparently do not move far once they reach salt water ancI some may return to overwinter in freshwater; others may move up in summer. Movements into freshwater by nonbreecling fish may be triggered by abundance of juvenile salmon, which are prey; the timing of such movements into the lower I(lamath appears to vary greatly from year to year (Gale et al. 19981. Large numbers of aclult cutthroat are observed every summer in lower Blue Creek, where they seek refuge from poor conditions in the main-stem I(lamath (Gale et al. 19981. Eulachon The eulachon or cancIlefish is a smelt (Osmericiae) that reaches the southern extent of its range in the MacI River, Rec~woocI Creek, ancI the I(lamath River (Moyle 2002~. Historically, large numbers entered the river to spawn in March ancI April, but they rarely moved more than 8 mi iniancI. Spawning occurs in grave! riffles, ancI the embryos take about a month to clevelop before hatching ancI being washed into the estuary as larvae. The eulachon in the I(lamath River once was an important foocI of the Ameri- can Indians in the region (Trihey ancI Associates 19961. Since the 1970s, their numbers have been too low in most years to support a fishery. The causes of the clecline are not known but probably are tiecI to changing ocean conditions ancI poor habitat ancI water quality in their historical spawning areas (Moyle 2002~.
276 FISHES IN THE KLAMATH RIVER BASIN Green Sturgeon Probably 70-80% of all green sturgeon are proclucecI in the lower I(lamath River ancI Trinity River, where several huncirecI are taken every year in the tribal fishery, which is the principal source of life-history infor- mation on this species (Moyle 20021. Green sturgeon enter the I(lamath River to spawn from March to luly; most spawning occurs from the micicIle of April to the micicIle of lune at temperatures below 14°C. Spawning takes place in the lower main stems of the I(lamath ancI Trinity rivers in creep pools with strong bottom currents. luveniles occupy the river until they are 1-3 yr oicI, when they move into the estuary ancI then to the ocean. Optimal temperatures for juvenile growth in the river appear to be 15-19°C. Tem- peratures above 25°C are lethal (MayfielcI 20021. After leaving the river, green sturgeon spencI 3-13 yr at sea before returning to spawn ancI often move long distances along the coast. They reach maturity at 130-150 cm ancI are repeat spawners. Large aclults (250-270 cm) typically are females that are 40-70 yr oicI (Moyle 20021. There is some evidence that green sturgeon populations are in clecline, but recluction of the marine commer- cial fishery for them may have alleviatecI the clecline somewhat (Moyle 20021. In 2003' NMFS rejectecI a petition to have them listecI as a threat- enecI species. Pacific Lamprey Lampreys once were so abunciant in the coastal rivers of California that they inspirecI the name Ee! River for the thircI largest river in the state. They supportecI important tribal fisheries. Tociay, their numbers are low ancI cleclining (Close et al. 2002' Moyle 20021. Their biology is poorly clocu- mentecI, but they probably have multiple runs in the I(lamath basin. Most aclults (30-76 cm) enter the river from lanuary through March to spawn from March to rune, although movement has also been observecI in most other months (Moyle 20021. How far upstream lampreys movecI histori- cally is not known, but it is certain, as shown by the genetics of resiclent lampreys, that they enterecI the upper basin above I(lamath Falis at least occasionally. Most spawning appears to take place in the main stem or larger tributaries. Like salmon, lampreys construct recicis for spawning in grave! riffles, although the tiny larvae emerge from the grave! in just 2-3 wk. They are washecI clownstream once they emerge, ancI they settle in sancI ancI mucI at the river's ecige. The larvae (ammocoetes) live in burrows in these quiet areas for probably 5-7 yr ancI feecI on algae ancI other organic matter. During the larval stage, they move about frequently, so they are commonly capturecI in salmon outmigrant traps. Factors limiting the sur- vival of ammocoetes are not known, but it is likely that rapicI or frequent
FISHES OF THE LOWER KLAMATH BASIN 277 cirops in flow deprive them of habitat ancI force them to move into open water, where they are vulnerable to preciation. They clo not appear to be limitecI by temperatures in the basin, but anything that makes their shallow- water habitat less favorable (such as pollution ancI trampling by cattle) is likely to increase mortality. The blincI, worm-like ammocoetes undergo a ciramatic transformation into eyecI, silvery aclults when they reach 14-16 cm, after which they mi- grate to the sea (Moyle 20021. Downstream migration usually is coinciclen- tal with high flows in the spring, but movement has also been observed cluring summer ancI fall (Trihey ancI Associates 19961. In the ocean ancI estuary, they prey on salmonicis ancI other fish for 1-2 yr before returning to spawn. The Pacific lamprey is a tribal trust species with a high priority for recovery to fishable populations (Trihey anti Associates 19961. Its cul- tural importance to American Indians is largely unappreciated (Close et al. 2002). Native Nonanaciromous Species SpecklecI ciace, I(lamath smaliscale sucker, lower I(lamath marblecI scuipin, threespine stickleback (some of which are anaciromous), anti I(la- math River lamprey are quite common in the lower river ancI its tributaries of low gradient. With the possible exception of the scuipin, these species probably all have fairly high thermal tolerances (Moyle 20021. In the reaches within 30 mi or so of the ocean, marblecI scuipin apparently are replacecI by the two amphiciromous species, prickly scuipin ancI coastrange scuipin. With the exception of the lamprey, which feecis on fish, all the resident fishes feecI mainly on aquatic invertebrates. The relationship between the native nonanaciromous ancI anaciromous species has not been worked out in the I(lamath, but the ciace, stickleback, scuipins, ancI suckers are prob- ably subsiclizecI by nutrients brought into the streams by the anaciromous fish ancI may suffer heavy predation, especially in the larval stages, by juvenile salmon ancI steelheacI. Nonnative Species The lower I(lamath basin is still clominatecI by native fishes, but other species have a strong presence in highly alterecI habitats, such as reservoirs ancI poncis. The Shasta River, once a coicI-water river, now supports large populations of brown buliheacis ancI other warm-water, nonnative species because summer flows consist largely of warm irrigation-return water. There also is a continuous influx of nonnative fishes from the upper I(lamath basin, where they are extremely abundant. Because there is a positive relationship between clegree of habitat disturbance ancI abundance of nonnative fishes
278 FISHES IN THE KLAMATH RIVER BASIN (Moyle ancI Light 1996), improving habitat for native fishes shouicI recluce the likelihoocI that nonnative species will become more abundant. MASS MORTALITY OF FISH IN THE LOWER KLAMATH RIVER IN 2002 During the last half of September 2002, mass mortality of fish (fish kill or fish clie-off) occurred in a reach of the Lower Klamath River extending about 30 mi up from the confluence of the river with its estuary (Figure 1-11. In responding to the general neecI for a timely assessment of the conditions leacling to this mortality, CDFG releasecI in lanuary 2003 a report that describes the extent of the mortality ancI its distribution among species, hycirologic ancI meteorological conditions that accompanied the mortality, some aspects of water quality, ancI the results of physical exami- nation of both living ancI cleacI fish. A second CDFG report will clear with long-term consequences of the mortality. Also cluring 2003, USGS releasecI a report clearing with the mortality of September 2002 (Lynch ancI Risley 2003~. The USGS report documents environmental conditions that coin- ciclecI with the mortality, but cloes not attempt to reach conclusions as to its cause. The sponsors of the NRC stucly on enciangerecI ancI threatened fishes asked the NRC committee to stucly information on the fish kill of 2002 ancI inclucle the analysis in its final report. While it is reasonable that this issue be covered in the committee's report, it is also important to note that the fish kill primarily affected Chinook salmon, for reasons that are explainecI below, ancI not the threatened coho salmon that is the focus of attention for the NRC committee in its work on the lower Klamath basin. Furthermore, the NRC committee was only able to consider the two reports cited above ancI unpublishecI records on weather ancI temperature; other reports to be issued in the future might provide aciclitional information that wouicI influ- ence conclusions about the cause of the fish kill. The fish kill of 2002 in the Klamath lower main stem is unprececlentecI in magnitude. It raises ques- tions as to whether human manipulation of the Klamath River or the adjoining estuary was clirectly or inclirectly responsible anti, if so, what might be clone to prevent its recurrence. A full ancI final explanation of mortality probably is not possible, however, given that the fish kill was not anticipated ancI therefore the conditions leacling to it were not well clocu- mentecI. Extent of Mortality CDFG, quoting USFWS, has estimated the total mortality of fish in the last half of September 2002 at about 33,000. This estimate, which is subject
FISHES OF THE LOWER KLAMATH BASIN 279 to revision, is likely to be conservative. The projected run size of fall-run Chinook salmon, which was the most abundant of the fish that cliecI, was estimated at 132,000. Thus, regarcIless of any adjustments that might be macle in the final estimate of mortality, a substantial portion of the Chi- nook salmon run was lost before spawning. Both CDFG ancI USFWS estimated the species composition of the fish kill, which extenclecI beyond salmonicis to other taxa, inclucling the I(la- math River smaliscale sucker, but percentage estimates from CDFG are limitecI to the salmonicis. A sample of 631 cleacI fish collectecI uncler the supervision of CDFG showed 95.2% Chinook salmon, 4.3% steelheacI trout, ancI 0.5°/O coho salmon. These estimates differ only slightly from the USFWS estimates. Further cletails may appear in reports yet to be issued. Among both Chinook ancI steelheacI, nonhatchery fish appeared to have cliecI in greater numbers than fish of hatchery origin. A similar cletermina- tion for coho salmon is complicatecI by the fact that only small numbers of coho were found. If the coho hacI been in peak migration at the time when mortality occurred, more cleacI coho probably wouicI have been founcI. The coho migration occurs later than the Chinook migration, which probably explains why few coho were affected. Direct Causes of Mortality CDFG has given infection as the clirect cause of cleath of the fish. Both living ancI cleacI fish were infected with Ichthyopthiri?vs m?~Itifilis, a proto- zoan, ancI Flavobacter col?vmnare, a bacterium . As inclicatecI by CD F G. these two pathogens are wiclespreacI anti, when they become lethal to fish, typically are associated with high degrees of stress. Crowding may be con- siclerecI an aciclitive agent to stress in that it facilitates efficient transmission of pathogens from one fish to another. A combination of crowding ancI stress thus wouicI be especially favorable for the clevelopment of these pathogens in sufficient strength to cause mortality of fish. Potential agents of stress, which may have acted in combination rather than alone, inclucle high temperature, inadequate concentrations of clissolvecI oxygen (unclocu- mentecI), ancI high concentrations of unionized ammonia (unclocumentecI). Indirect Causes of Mortality Low flow in the I(lamath River main stem is the most obvious possible cause of stress leacling to the lethal infections of fish in the lower I(lamath River cluring 2002. Low flow can cause crowding of the fish in their hoicI- ing areas as they await favorable conditions for upstream migration ancI can be associated with high water temperature ancI with lower than normal
280 FISHES IN THE KLAMATH RIVER BASIN concentrations of clissolvecI oxygen. CDFG therefore reviewed information on flow in the main stem, as clicI USGS (Lynch anti Risley 20031. The flow of the I(lamath River at I(lamath, which is just a few miles above the estuary, is shown in Figure 7-3 for ciry yrs used by CDFG in its overview of low flows in the river. The flows at Iron Gate Dam, about 185 mi upstream, are given for comparison. For an extenclecI span of years not restricted to drought, September flow at Iron Gate Dam is about one-thircI of the flow at I(lamath. For example, mean September discharge at I(la- math was 2~973 cfs for 1988 through 2001 (excluding 1996' 1997) and the same statistic for the I(lamath River at Iron Gate Dam is 1~130 cfs, as cleter- minecI from USGS gage records. The USGS electecI not to use ciata for the I(lamath gage because the accuracy of the gage at low flow is subject to errors greater than 15%. Figure 7-3 shows the sum of the gages at Orieans (main stem above the Trinity) anti at Hoopa (on the lower Trinity), both of which produce clis- charge readings within 10% of the true value, for comparison with the flows in the main stem at I(lamath. The two sets of values are separated by some aciclitional discharge (unclocumentecI) that accumulates below the 3,000 2,500 2,000 us ~ 1,500 ·_. 1,000 500 o I Cal Klamath | ~ Iron Gate 2026 77 1489 ~ 11894 ~- 1857 1988 1991 1992 1994 2001 2002 FIGURE 7-3 Mean flows of the I(lamath main stem at I(lamath (near the site of the 2002 fish kill) and at Iron Gate Dam (about 185 mi upstream) in September for 6 low-flow years considered by CDFG in its analysis of the fish kill. The asterisk shows the sum of flows for the I(lamath at Orleans and the Trinity at Hoopa, as a check on the I(lamath gage (this sum omits small tributaries below the Trinity). Sources: Data from CDFG 2003 and USGS gages.
FISHES OF THE LOWER KLAMATH BASIN 28 Trinity. The I(lamath gage ciata ancI the sum of the two gages above it show essentially the same picture qualitatively, as cloes the analysis by USGS basecI on the Orieans gage alone. Also, USGS restricted its analysis of flows to 1-24 September, which coincides better with observed mortality than 1- 30 September, but the mean gage readings for these differing intervals are essentially iclentical (< 1% difference at I(lamath). All ciata indicate that flows comparable with those of 2002 have occurred a number of times over the last 15 yr without causing mass mortality of salmonicis. This cloes not rule out the possibility that low flow was a factor, but it cloes suggest that the occurrence of flows similar to those of 2002 has not in the past been sufficient by itself to cause mass mortality. The USGS analysis acicis a new dimension to future concerns relatecI to flow in that it shows a substantial increase in distance to the water table over 2001 ancI 2002, both of which were ciry years. Because shallow allu- vial water reaches the tributaries ancI main-stem I(lamath as grounc~water, which supports flow in ciry weather, cirawclown of the water table by pumping shouicI be taken into account in any future evaluation of low flows, particularly if pumping is a growing response to water scarcity clur- ing drought. Flow couicI be relatecI to crowding on a conclitional basis through run size or timing of run. CDFG consiclerecI this possibility by using estimates of run timing ancI run size of Chinook salmon, which accounted for most of the fish biomass in the river cluring the last half of September. The analysis showed that the run of Chinook was only slightly larger than average ancI that it was bracketed by run sizes both smaller ancI larger for other comparably ciry years. Thus, run size cloes not show evi- clence of being a conclitional influence relatecI to flow. The August-October run of Chinook appears to have peaked earlier in 2002 than in other years of record, ancI this suggests a conclitional relation- ship with low flow in causing mortality. CDFG was reluctant, however, to attribute great significance to this possible relationship, given the small amount of information on which it is based. The ciata available to CDFG inclicatecI that air temperatures were not unusually high cluring September 2002 compared with other years of low flow when no fish kills occurred. Information on water temperature is sketchier, but also indicates that aver- age maximum water temperatures fell within the range of water tempera- tures in previous years of low water when there were no fish kills. The USGS macle comparisons of the I(lamath River with the Rogue River, which is locatecI nearby ancI has more comprehensive temperature records. Both water ancI air temperatures on the Rogue River were approximately 2°F higher in 2002 than the mean for the period of record. While the difference is small, the threshoicI for harm to salmonicis lies close to September tem- peratures, even in years of average flow. The USGS analysis, like the CDFG analysis, clicI not suggest that temperatures in 2002 were extreme by com-
282 FISHES IN THE KLAMATH RIVER BASIN parison with other years of low flow when no fish kills occurred. Thus, if temperature is a factor governing mortality, it wouicI involve coincidence of high temperatures with some other factor, the nature of which is not clear from the presently available information. Tests of water quality clicI not indicate the presence of toxicants, al- though the water was not samplecI until seven clays after the onset of the first observation of cleacI fish (CDFG 2003~. It is always possible that toxicants not tested were involvecI, but this seems unlikely, given that the fish kill occurred over an extenclecI period anti that there is no circumstan- tial evidence of the role of toxicants other than possibly ammonia generated by the fish themselves. CDFG also consiclerecI fish passage. According to CDFG, high flows in 1997 anti 1998 may have caused aggravation ancI expansion of channel bars that inhibited fish passage cluring extremes of low flow. These changes clicI not result in fish kills cluring the low-water year of 2001, but flows in 2001 were not as low as those in 2002. Thus, a current hypothesis of CDFG is that a change in channel geometry has created new conditions that are cletrimental to fish at low flows even though such flows previously clicI not leacI to high mortality. The hypothesis is speculative in that changes in channel conditions have not been establishecI by measurement, but it shouicI remain uncler consideration until further relevant evidence is collectecI. Summary of Explanations The possibility that passage is inherently more clifficult at low flows now than it was before 1997-1998 was the only explanation of unique conditions leacling to the fish kill that CDFG could not rule out in preparing its lanuary 2003 report. Because of the limitecI ciata about conditions be- fore anti cluring the kill, other hypotheses probably will emerge as other reports are prepared. One hypothesis that has not been evaluatecI by CDFG involves the effect of temperature extremes cluring the fish kill. As ex- plainecI earlier in this chapter, mean water temperature is less important for salmonicis than extremes of water temperature. Thus, for example, the failure of temperatures to clecline sufficiently at night when mean tempera- ture is high couicI place unusual stress on salmonicis but couicI be over- lookecI in a consideration of mean ancI maximum temperatures alone. Such conditions couicI occur, for example, when back radiation is so low (per- haps as a result of cloucliness or high humidity that a typical amount of cooling wouicI not occur at night. A sequence of events involving ciaily minimum temperature rather than fish passage might be a cause of mortality. A large number of salmon moved up the river coincident with a series of clays in which water tempera- tures were high enough to inhibit migration. McCullough (1999) states
FISHES OF THE LOWER KLAMATH BASIN 283 that, basecI on studies in the Columbia River, Chinook salmon cease mi- grating when maximum water temperatures exceed 21°C. Lynch ancI Risley (2003) indicate that cluring the time of the kill, maximum water tempera- tures in the river at Orieans, 30 mi upstream of the kill, averaged 20.3°C, anti that the average minimum was 19.7°C. Thus it seems likely that tem- peratures in the I(lamath River at the site of the kill reachecI or approached the inhibitory temperatures. As they commonly clo, the salmon helcI in pools when the temperatures were high, waiting for conditions to improve before continuing upstream. The temperature anti flow ciata given by Lynch anti Risley (2003) indicate, however, that conditions clicI not improve anti that nocturnal temperatures were not much lower than daytime tempera- tures. Because salmon are more vulnerable to infectious diseases at higher temperatures (McCullough 1999), crowding encouraged the disease out- break that resultecI in the kill. The fish-passage hypothesis of CDFG or the minimum temperature hypothesis given above may or may not justify aciclitional release of flow from Iron Gate Dam. It is unclear whether low flows actually blockecI upstream migration or, as suggested by the literature, that most of the fish stopped moving because of high temperature (CDFG cites evidence that at least a portion of the run was capable of moving upstream cluring these low-flow conclitions). The emergency release of 500 cfs of aciclitional water from Iron Gate Dam by USER, which arrived long after the fish kill hacI enclecI, lackecI any specific justification. For relief of physical blockage, if it occurs, only a large amount of water (e.g., 1,500 cfs) wouicI be of use. Aciclitional water from the Trinity couicI be especially valuable in that it wouicI be cooler, if releasecI in quantity. If passage is the key issue, the recurrence of low flows similar to those of 2002 will probably be accompanied by mass mortality of fish. If other explanations, inclucling minimum temperature, are the key explanation of mortality, recurrence is less likely, although higher temperatures over the long term caused by climate change couicI increase the likelihoocI that such kills wouicI occur. Aggressive pursuit of some recommendations relatecI to coho salmon (see information on augmentation of coicI-water tributary flows in Chapter 8) couicI, if successful, recluce the risk of mass mortality of Chinook salmon. In any case, it is clear that increased monitoring of water quality anti channel conditions in relation to flows in the lower main stem is neeclecI in support of measures that may be necessary to prevent loss of Chinook salmon. CONCLUSIONS The lower I(lamath basin is a geologically dynamic region that histori- cally hacI large runs of anaciromous fishes with diverse life histories. The
284 FISHES IN THE KLAMATH RIVER BASIN fishes were wiclely clistributecI in the basin; some even entered the rivers that fecI Upper I(lamath Lake. The Salmon, Scott, Shasta, ancI Trinity rivers all of which are major tributaries of the I(lamath River were major salmon ancI steelheacI producers. The Shasta River in particular, with its coicI flows ancI high productivity, was once especially productive for anaciromous fishes. In the I(lamath basin as a whole, Chinook salmon were ancI are the most abundant salmonicI, followecI by steelheacI. Coho salmon rank third, but are well below Chinook ancI steelheacI in abundance. Virtually all populations of anaciromous fishes have cleclinecI consicler- ably from their historical abundances, although documentation for some species, such as Pacific lamprey ancI green sturgeon, is poor. Three of the most distinctive forms coho salmon, spring-run Chinook, anti summer steelheacI are on the verge of extinction as naturally maintained popula- tions in the basin. It is significant that these three are the most clepenclent on summer water temperatures below 18°C ancI that they historically spawned ancI clevelopecI in tributary streams, many of which now are too warm for them. The anaciromous fishes have been in clecline since the lath century when clams, mining, ancI logging severely alterecI many important streams ancI shut off access to the upper basin. The cleclines continued through the 20th century with the clevelopment of intensive agriculture with its clams, diversions, anti excessively warm water both insicle ancI outside the basin. Continued logging in heac~water areas ancI commercial fishing also have contributed to the clecline. The main-stem I(lamath River has become a challenging environment for anaciromous fishes because of clecreasecI flows anti increased summer water temperatures. Although it is inhospitable to juvenile coho, it is still important for the rearing of juvenile Chinook salmon ancI steelheacI, but increases in temperatures in luly-September of 1-3°C may make it unsuit- able even for them in the future. Increased flows clown the river in summer are likely to benefit anaciromous fishes only if temperatures can be kept within bioenergetically favorable ranges. This is particularly true for the lowermost reach of the main stem, below the Trinity River, which may be either cooler or warmer in late summer than the main stem, clepencling on the amount of water being releasecI from Lewiston Dam. Millions of juvenile fish, inclucling Chinook salmon, steelheacI, ancI coho are releasecI into the I(lamath anti Trinity rivers each year by the Iron Gate ancI Trinity River hatcheries, which were built to mitigate salmonicI losses created by large clams. These hatcheries have helpecI to maintain fisheries for coho ancI Chinook salmon, but their effect on wilcI populations of salmonicis in the basin is not well unclerstoocI. It is likely, however, that interactions between the hatcheries ancI wilcI juveniles in the river are hav- ing an adverse effect on the survival of wilcI juveniles through competition for space ancI foocI ancI aggressive interactions (e.g., McMichae! et al. 1999,
FISHES OF THE LOWER KLAMATH BASIN 285 I(elsey et al. 2002), to the extent that the contributions of hatchery fish to fisheries are at least partially offset by the clecreasecI contribution of wilcI fish (Levin et al. 2001). A high percentage of naturally spawning aclult coho ancI Chinook salmon are of hatchery origin. Native nonanaciromous fishes are wiclespreacI ancI common in the cirain- age, but their relationships to anaciromous fishes are not known. Nonnative fishes are uncommon in the lower basin except where cirastically alterecI habitats favor them. If habitat clegraciation continues, the I(lamath River ancI its main tributaries will probably favor nonanaciromous native ancI nonnative fishes increasingly at the expense of anaciromous fishes. The hierarchical nature of watersheds assures that many environmental changes, some of which are quite small incliviclually, collectively affect fish popula- tions not only in their immediate vicinity but also both upstream ancI downstream because of the extensive movement of fishes (Fausch et al. 2002). The problems with coho salmon are a reflection of larger problems with poor habitat ancI water quality for anaciromous fishes generally in the basin. Restoration efforts that benefit coho salmon shouicI benefit most, but not necessarily all, cleclining species. Prevention of further listings uncler the ESA requires a systematic, basin-wicle approach to restoration ancI manage- ment. Some major gaps in knowlecige are as follows: 1. Information on the biology of coho ancI other salmonicis in the basin is largely unsynthesized; synthesis ancI interpretation of ciata on historical trencis ancI present conditions wouicI be especially valuable. 2. Studies on anaciromous fishes other than fall-run Chinook, winter steelheacI, ancI coho are very limitecI or lacking, particularly for summer steelheacI, spring-run Chinook, ancI Pacific lamprey. It cannot be assumed that management strategies favoring species of primary interest also favor other species. 3. The biology of nonanaciromous native fishes ancI macroinvertebrates in the basin is largely unknown, inclucling basic descriptions of life histories ancI environmental requirements ancI their relationships to coho salmon ancI other anaciromous fishes. 4. The potential effects of global climate change on the I(lamath basin ancI its fishes, especially coho, are poorly unclerstoocI, inclucling the rela- tionship between changing ocean conditions ancI the abundance of coho ancI other anaciromous fishes. Climate warming wouicI almost certainly be clisacivantageous to coho. 5. The thermal consequences of stream ancI watershed restoration ac- tions, inclucling increasing summer flows clown the main-stem I(lamath River, are not well clocumentecI, especially in relationship to coho salmon. 6. The effects of hatchery operations on wilcI populations of coho ancI
286 FISHES IN THE KLAMATH RIVER BASIN other salmonicis in the basin are not unclerstoocI, inclucling the effects of hatchery steelheacI anti Chinook on juvenile coho salmon. 7. Strategies for improving tributaries for spawning anti rearing of coho anti other anaciromous fishes are not yet well clefinecI. 8. The lower 30-40 km of the main-stem I(lamath seems to be increas- ingly unfavorable to anaciromous fishes, for reasons that are not known. The effect on the lower river of changing flows from the Trinity River neecis to be evaluatecI, as clo the potential benefits of comanaging flow releases from the clams on the Trinity anti Upper I(lamath rivers. 9. Reliable abundance estimates anti habitat affinities of juvenile coho anti other salmonicis are largely lacking.