Review

for the Oregon Chapter of the American Fisheries Society
of Agency  Pre-decisional Draft Professional Scientific Review Copy
 of 

Biological Opinion and Conference Report for the Continued Operation of the Bureau of Reclamation’s Klamath Project as it Effects Endangered Lost River Sucker (Deltistes luxatus), Endangered Shortnose Sucker(Chasmistes brevirostris), Threatened Bald Eagle (Haliaeetus leucocephalus), and Proposed Critical Habitat for the Suckers

(dated February 5, 2001)
 by

Douglas F. Markle, David Simon, Michael S. Cooperman, and Mark Terwilliger
Dept. of Fisheries and Wildlife, Oregon State University, Corvallis, OR 97331-3803
6 March, 2001

The single most important management action ever taken for the recovery of Klamath Basin suckers was the cessation of the snag fishery on spawning adults by the Klamath Tribe and subsequently by the State of Oregon. Highly fecund, long-lived species experience periodic recruitment failure compensated by diverse age classes of spawners, especially older fish. The positive impact of this important action has been detected but will require at least another 10-20 years to be fully functional. The primary role of subsequent management actions is to insure periodic, not necessarily continuous, annual recruitment and to insure that some, as yet undefined proportion, of females live 20 –30+ years. This Biological Opinion (BO) seeks to address both issues and although it might be reasonable to support a minimum Upper Klamath Lake level of 4139 or 4140 ft, this document provides little basis for choosing any level. The linkages among lake level, water quality and sucker benefits remain unclear.

 This review of the BO will address both the key scientific issues related to the opinion and editorial problems with the document. The editorial problems are of such magnitude that they severely influenced this review. The misspelled words, incomplete sentences, apparent word omissions, missing or incomplete citations, repetitious statements, vagueness, illogical conclusions, inconsistent and contradictory statements (often back to back), factual inaccuracies, lack of rigor, rampant speculation, format, content, and organizational structure make it very difficult to evaluate this BO.

 We urge, in the strongest possible way, that the Service re-visit every single sentence for importance, applicability, grammar, spelling, content and internal consistency with other parts of the document.  The document is excessively long. The problems are not “window dressing”, rather they obscure the data and make it very difficult to find validity in claims. This document has the potential to have a severe negative impact on the Service’s public credibility.

             In the following we present three sections: Overall evaluation of Opinion (p. 2), Major editorial comments (p. 8), Minor editorial comments (p. 18).

 Overall evaluation of opinion

  The BO primarily seeks to relate lake levels in Upper Klamath Lake to 1) water quality, 2) habitat for larvae, 3) habitat for juveniles, and 4) movement corridors for adults. The logical assumption is that somewhere between bone dry and full pool there is a negative impact on suckers. The analytical problem with the system is that lake level is a seasonally monotonous function of date, so that sequential observations are serially auto-correlated and variables of interest are cross-correlated. For example, low lake level and low temperature do not co-occur because low lake level happens in late summer or fall and low temperature happens in winter. An important consequence is that lake level cannot be easily separated from cross-correlated physical variables or from seasonal behavioral patterns of the fish. Fish responses that are temperature related cannot be easily separated from lake level. A further consequence is that an entire year’s worth of observations become a statistical sample of one. The BO does not seem to appreciate this fundamental analytical problem.

 Water Quality

 The BO argues that lake elevation is related to water quality and was responsible, in part, for fish kills such as those observed in 1995, 1996, and 1997. The case for a fish kill – lake level relationship rests on weak or inappropriate data, such as the following:

· Pg. 27.  “In contrast, suckers captured in 1994-1996 (years with better water quality and higher lake levels) were substantially more robust”.

-         This is an instance where thin fish are used as evidence of poor water quality when no such evidence is presented, not even a correlation coefficient. Further, one of the years, 1994, had the lowest lake level on record, and directly challenges the premise.

· Pg 63. “Studies yielded variable LC-50 values for different species and life stages” and “LC-50s for the four water quality variables did not vary significantly between species”

-         The BO frequently uses the words “variable” and “different” inappropriately, sometimes implicating statistical significance when there is none. Variation happens around a central tendency such as a mean and statistics usually test differences in central tendencies.  LC-50’s therefore did not differ significantly between species.

 · Pg 67. “Nevertheless, OSU biologists noted a substantial drop in age 0 sucker cast net catches in September and October suggesting these fish were affected by the die-off.” 

-         This is an instance where a general seasonal pattern described elsewhere is used to support an argument that it does not support. The decline in catches is as much associated with behavior as with mortality, and has happened every single year from 1995-2000, regardless of whether adults experienced a fish kill or not. Fish behavior is often seasonal and temperature related. Here and elsewhere in the BO it is important to establish the “typical’ behavior patterns and then evaluate if a change in a variable of interest changes that behavior.

·  Pp. 68-70. Discussion of climate influence on sucker fish kills.

-         This provides a strong review of existing climate data and linkage to water quality for one of three fish kill years. What happened during the other two years? What was the distribution of variables each month? What data (such as cloud cover) was ignored?

·  Pg. 74. “Lowered lake elevations may increase AFA production and worsen water quality.”

-         Again, the two lowest water years, 1992 and 1994, are not explained. This discussion describes a complex, non-linear system that either implicates intermediate lake levels or suggests that almost any lake level can be associated with poor water quality. The data implicate intermediate, not lower, lake levels because 1) historical data have been interpreted to indicate that fish kills were common prior to Link River Dam, 2) the pre-dam minimum elevation was 4139.93 and therefore all historical fish kills took place at higher lake elevations, and 3) no die-off has ever been documented when elevations were below the historical minimum (pg. 46),

· Pg. 101 and vicinity.  In the discussion of water quality and sucker status in Gerber Reservoir, several lines of evidence are given suggesting habitat quality is stressful, yet “SNS in Gerber Reservoir are believed to be doing well.”

-         This conclusion appears to be inconsistent with the data and BO and would not seem to implicate water quality and low population numbers.

 In summary, the argument for a fish kill – lake level relationship is complex, but does not account for the observation that extremely low lake elevations in 1992 and 1994 did not produce fish kills. Further, the BO suggests that 1995-1999, the most heavily managed years in the lake’s history, were higher water years, yet fish kills occurred in three of the five years. The data presented give little support for the contention that low summer lake level is related to fish kills. If anything, the data support the notion that intermediate summer levels are dangerous.

 Larval habitat

 The BO argues that lake elevations provide habitat for larval suckers.  Curiously, life history stages are not defined and it is clear in places throughout the document that “larvae” is sometimes used loosely for all young-of-the-year, including juveniles. For practical purposes, we consider 20 mm as the upper limit of larval stages and lower limit of juvenile stages. It should be noted that the ecological and behavioral juvenile transitions take place below this size.

At higher lake elevations, more inundated marsh and shoreline vegetation would be available for larval rearing.  This habitat appears critical for larval suckers because age 0 juvenile sucker indices approached zero in the two low water years, 1992 and 1994 (Simon et al., 1996), but were higher in all other years from 1991-2000 (a critical observation the BO failed to notice). The processes responsible for this pattern are claimed to be food and predation.  

The BO repeatedly states that zooplankton abundance and diversity is greater in vegetated areas and that low water levels force larval suckers out of vegetation into habitats with less food. We found no support in the document demonstrating that zooplankton abundance and diversity in Upper Klamath Lake is greater in vegetated habitats.

The BO makes a confused case for predation noting on p. 18, “Sucker predation by fathead minnows (Pimephales promelas) in the laboratory was greatest when larvae lacked cover (Dunsmoor 1993, Klamath Tribes 1995)” while on p. 94, “Presence of vegetation structure significantly influenced survival rates in none of the trials with SNS and in two of seven trials with LRS. Interaction of depth and structure factors was not significant in any of the trials”. Despite these data, “The Tribes submits (sic) that as water depth increases to about 0.6 m, the surface orientation of the sucker larvae and the bottom orientation of the fathead minnows results in enough separation to almost eliminate predation, even when the fathead minnows are hungry.” This conclusion is not logical. As lake elevation changes, sucker larvae and the fathead minnows move together. The lake has a gradual slope and is not a steep-sided aquarium. Higher lake elevation will not lead to segregation of the two species.

As noted under editorial comments, the BO consistently misinterprets larval sucker data, noting on pg 14 “Larval production appeared relatively low in 1999”, when both Simon et al. (2000) and Cooperman’s unpublished 1998-2000 work suggest 1999 was the best year of larval and juvenile production of the late 90’s. It engages in speculation, for example, on pg. 114, “Channelization and diking of the lower Williamson and Wood rivers has shortened and widened both rivers....Slower velocities may delay emigration of larval suckers to UKL.” Cooperman (unpublished) has demonstrated that most larval suckers move through the river quickly so that the role of channelization in recruitment failure is highly speculative. It selectively ignores data, for example on pg 127, in reference to dry years... “July elevations range from 4139.1 to 4140.9 ft (average 4140.2 ft).  None of the emergent vegetation is inundated at the lowest elevation and 2-28% at the highest elevation ... So little emergent vegetation habitat is available that it is highly unlikely that a year-class could develop.” The 1991 year class (widely acknowledged as a strong year class in the BO) was produced with a July 15 elevation of 4140.81. It, therefore, does not follow that it is “highly unlikely that a year-class could develop”, in fact, one of the best, recent year classes was produced in this range. Furthermore, higher lake elevations could have the unintended consequence of being a benefit to recruitment of cyprinids, especially exotic fathead minnows, which need submerged substrates for egg deposition. The BO fails, here and elsewhere, to consider unintended consequences of the actions.

 Juvenile habitat

 The BO argues that lake elevations provide habitat for juvenile suckers. Although the BO says, “It is unclear what habitat and substrate types are preferred” by juveniles (pg. 20), it notes that “Highest age 0 sucker densities were found on small mix and gravels and lowest densities were found on fines, sand, and boulders  (Simon et al. 2000a).” It seems clear that the daytime habitat of juveniles is clean gravel and small rocky particles and that type of habitat does not extend far from shore in Upper Klamath Lake. Consequently, reduction in lake elevation will reduce the available habitat. But why say the habitat and substrate preference is unclear when it is some of the most robust data available?

The BO does not seem to recognize that juvenile ecologies are often very different among closely related fish species. It is to be expected that juvenile KLS will differ in their ecological requirements from SNS and LRS. The BO discusses juvenile habitats as if there are two rather than three (or four) species in UKL and Lost River subbasins. The anecdotal stories of “stream-resident SNS” and juvenile suckers in vegetated areas are based on unidentified specimens that may or may not be listed suckers.

The BO argues for canal screening in the southern end of Upper Klamath Lake.  Substantial numbers of juvenile suckers leave Upper Klamath Lake in late summer (Gutermuth et al., 2000) and in-lake estimates of age 0 juvenile suckers decline during this same period (Simon et al., 2000). The movement from the shoreline is a seasonal behavioral pattern. The movement out of the lake may be part of this pattern and possibly related, in part, to water quality.

Although the argument for screening may be valid, the system is too poorly understood to predict un-intended consequences and, therefore, actions should be cautious. There are two fundamentally different population models applicable to Klamath Basin suckers – either a metapopulation or a member-vagrant system. A key difference between the two is that life history stages that are not retained within the hydrological system of the adults are recruits for other populations in a metapopulation but are lost to the species in member-vagrant systems. In member-vagrant systems, a management option that does not let negative selection operate on vagrants could have the unintended consequence of selecting for vagrants. Elsewhere the BO seems to imply that this is a metapopulation. UKL vagrants clearly seem to be the metapopulation source for downstream reservoir sinks, especially Copco, but this is almost certainly a recent process. However, historical connections between Lost River and UKL and genetic similarities suggest that vagrants from both subbasins may have been part of a natural metapopulation. In contrast, the genetic uniqueness of some populations and the generation-long absence of re-establishment of the Harriman springs populations suggest a member-vagrant model.

 Adult corridors

 The BO concludes that an Upper Klamath Lake elevation of 4140.0 ft. is critical for persistence of endangered LRS and SNS because it provides sufficient water quality refugia, while below 4140.0 ft adult sucker access to refugia is impaired. The primary data in support of this notion is Peck (2000) and pp 107-111.

Although disorganized, the presentation starts with an analysis comparable to one saying that people prefer Calcutta ghettos because that’s where you find the highest density of humans. Previously un-reported analyses on pg. 108 purport to show an adult preference index of 4 (meaning the fish were 4 times as likely to be in this depth range than would be predicted by random distribution across all depths, not “4 times as many fish were observed”) for the 6-9 ft depth and “strong avoidance” for depths < 3 ft. When the data show a seasonal movement into deeper water in fall, a common behavioral pattern in many fish, a “This is so” statement makes it a response to poor water quality. The presentation fails to point out that all of the telemetry depth data are daytime observations so that nighttime patterns are unknown. A behavior that might not happen in daylight might be very common at night, especially a movement into shallow water.

The critical analysis of August – September 1994 data needs scrutiny for, in no way, does it show a “critical minimum acceptable depth is 3-4 ft”.

·  Pg. 109. “During 1994, a low water year, many of the radio-tagged suckers were concentrated in Ball Bay in early summer, most likely feeding (Hazel [1969] found that the north end of the lake had the highest biomass and variety of benthic organisms, some of which are eaten by suckers).  In July, as DO levels declined below 4.0 mg/l the fish had moved from the Ball Bay area closer to the entrance to Pelican Bay where water depths were >3 ft.  In August and September when lake levels dropped below 4138 ft, radio-tagged suckers moved from areas of good water quality into other areas where the quality was poorer.  In 1996, when water levels were higher, radio-tagged sucker movement in summer was often from areas of poor water quality to higher quality.  Data from these two years indicates that suckers move in response to water quality changes but may avoid areas of higher water quality when such areas become too shallow.  Based on telemetry, the critical minimum acceptable depth is 3-4 ft; however, >4 ft would be much safer because this is the depth preferred by suckers.”

-         “many … concentrated in Ball Bay”

o       9 of 31 suckers located in June were in Ball Bay. There were 8 offshore Pelican Bay and 7 in Shoalwater Bay. They were no more “concentrated” in Ball Bay than in the other two areas (table 4 in Peck, 2000). 

-         “In July … to entrance of Pelican Bay”

o       15 fish off Pelican Bay (an increase of 7 fish from June) and 0 fish in Ball and Shoalwater bays. There was a drop in sightings from 31 to 22 so 9 fish seen in June were unaccounted in July (table 4 in Peck, 2000). 

-         “In August and September when lake levels dropped below 4138 ft, radio-tagged suckers moved from areas of good water quality into other areas where the quality was poorer.”

o       23 sightings offshore of Pelican Bay in August (increase of 8 from July), 26 in Mid North in August (increase of 21 from July) and 6 LRS back into Ball Bay (table 4 in Peck, 2000).  

o       So, in August, there were 29 fish moving to “good water” and 6 LRS to “poor water”.

o       These patterns are complicated by an apparent increase in number of tagged fish (a maximum of 31 fish sighted prior to August and 64 sighted in August). It is not clear if these are newly tagged fish or just more observations. It would be instructive to know if the 6 LRS that moved back into Ball Bay had recently been tagged and might be demonstrating abnormal behavior.

o       The August to September sightings are comparable (64 each month but the numbers per species differ) and the patterns follow. From August to September there is an increase or no change in SNS sightings in all areas with Ball Bay going from 0 to 11. Sightings of LRS decline by 18 offshore Pelican Bay, by 11 at Mid North, and by 4 in Ball Bay. Sightings of LRS increase by 5 in Ball Point Trench, by 4 offshore Williamson, and by 11 in Mid Lake (table 4 in Peck, 2000).  

§         Part of the decline in LRS sightings is due to 13 fewer sightings, so all of the LRS patterns as well as the SNS pattern are most simply explained as haphazard movements over the whole area. The data most certainly do not demonstrate that fish move to poor water areas when lake level declines. 

§         This, in fact, is the conclusion reached earlier, on pg. 71 of this document, “In August and September when lake levels dropped below 4138 ft, radio-tagged suckers moved further offshore even though water quality remained good.” 

Assuming Peck’s (2000) data were converted from metric to English using 3.28 ft/m, his data show nothing remotely approaching a demonstration of “critical minimum acceptable depth”. Peck’s (2000) Figure 2 shows that 50% of all tagged fish from 1993-1998 were observed at daytime depths of about 4.26 – 11.15 ft. In other words, 50% were found at depths either less than 4.26 ft or greater than 11.15 ft. Since the document states that “Ninety-five percent of the observations in spring and summer months were at depths from 3 ft to 15 ft (pp. 108-109)”, an observation which can not be obtained from Peck (2000), and assuming normality of the data, 22.5% of the observations were made between 3-4.26 ft.  There is nothing in the data to say that 3-4 ft is critical and no demonstration of cause and effect. Finally, it should be pointed out that adult spring-spawning suckers are often seen in water less than 3 ft.

Dams

Considering its negative impacts, it is incongruous that “Chiloquin Dam is the only major dam affecting suckers that is not part of the proposed action.” (pg.93). Why? Siltation? It should be removed and any negative consequences (likely to be short-lived and probably unmeasurable) of reservoir siltation monitored.

There is no evidence that listed suckers had viable populations in the Klamath River prior to construction of the Pacificorp dams. With the exception of the Link River Dam, the Klamath River dams created habitat for listed suckers.  They are artificial, vagrant-populated populations. These dams certainly have altered the river ecosystem, but to claim they have played a role in the decline of listed suckers is unfounded.

Major editorial comments

 · General. Be consistent. For example, SNS, shortnose, shortnose sucker, Chasmistes brevirostris, and C. brevirostris are all used within a few pages of each other.  Define terms, and do it once, then be consistent. 

· General. Check citations; for example, under Longevity and Growth Rates, Simon et al. 2000 is incorrectly cited for observations made by Houde 1987.

· General. Restructure document to avoid repetition.

· General. Use a spellchecker, make verb tenses agree with subjects, write complete sentences.

·  Pg 7. The descriptions of ranges for lateral line scales and gillrakers is incorrect, based on the papers cited and Markle and Simon (1993).

·  Pg 8. “Currently it is uncertain if this phenomena is due to a lack of clearly defined characters for some fish is due to a natural tendency for extreme phenotypic plasticity, close genetic relationships, hybridization, or a combination of these factors.”

-         Phenomenon is, phenomena are.

-         What phenomena are being referred to?

·  Pg 9. Last sentence, “LRS have retained substantial genetic variation, suggestion continued reproduction isolation.’

     - The sentence does not make sense. If it means to say that presence of genetic variation indicates reproductive isolation, the logic of the statement makes no sense. 

· Pg 10.  Timing of Spawning, “LRS, possibly because they are the largest Klamath Basin suckers, also tend to spawn earliest.”

-         There is a body of evidence to show that larger fish within a species tend to spawn first in spring spawners and last in fall spawners. That phenomenon is demonstrated on Pg 13; why do conclusions precede evidence?

-         Further, it was documented that KLS was the earliest spawner (Buettner and Scoppettone 1990).

· Pg 10.  “in years of low flow there may not be any significant spawning”

-         No evidence is presented to support this statement in general or for specific areas. It would not be expected to be true for in-lake, spring-spawning fish.

· Pg. 12.  Paragraph starting:  Historically, sucker spawning occurred at Barkley Springs, Odessa Springs, Harriman Springs, and other lake spring areas (USFWS 1993a).  Reclamation has made infrequent daytime visual observations at these locations during the spring months since 1993.  No suckers have been observed.

-         Observations of sucker spawning activity at other lake spring areas show suckers spawn at night and are often absent during the day.  Infrequent daytime observations may not be sufficient to determine use.

· Pg.13. A paragraph indicates that USGS set trammel nets at 31 shoreline spawning sites “in the southern portion of the lake (R. Shively, USGS, per. com.)” and that most fish were captured at Modoc Point and Howard Bay.

-         Does this mean “southern 2/3rd of the lake”? What was caught in the southern 1/3? 

· Pg.13. “12 Klamath largescale/largescale hybrids”

-         This may relate to the general comment about consistent terminology, though, in this case, the issue would seem to be the definition of “hybrid”. Typically, a Klamath largescale sucker crossed with another Klamath largescale sucker would not be considered a hybrid.

· Pg.13. Referring to Wood River spawning it should be noted that larval suckers were captured in downstream drift samples at Weed Road in Wood River in 1991 (Markle and Simon 1993) and also collected with dip nets in the lower Wood River in 1996 (Simon and Markle 1997). Reference is made without proper citation of the evidence on Pg. 18.

· Pg. 14. Larval Biology –“Very little is known about the early life history of stream-resident SNS and Klamath largescale suckers.”

-         We are not aware of stream resident SNS, particularly early life stages. All evidence we have suggests that SNS larvae do very poorly in streams and need to get to Upper Klamath Lake quickly. If both SNS and KLS are in Lost River subbasin, then the simplest explanation is that stream resident juveniles are KLS not SNS.

-         Also, use out-migration or emigration not both; pick one term.

· Pg. 14. “In the Williamson River, larval sucker out-migration from spawning sites can begin in May or June and is generally completed by the end of July.”

-         Begin in June????? Where is evidence for that? Late April, early May is beginning.

· Pg 14 “Larval production appeared relatively low in 1999”;

     - Both Simon et al 2000 and Cooperman’s unpublished 1998-2000 work suggests 1999 was the best year of larval production of the late 90’s.

· Pg. 15  “Larvae caught in the Williamson River were younger than those in the lake, suggesting that larvae quickly leave the river and move into the lake (Cooperman and Markle 2000).”

     - As a point of clarification, >50% of larvae exit the river without significant delay (>1+days) within the river. 

·  Pg 15.  “Cooperman and Markle (2000) documented substantial numbers of sucker larvae in the area west of the Williamson River mouth.  It was previously assumed that few larvae occurred in this area because the Williamson River typically flows east towards Goose Bay.  In fact, pop net catches were several times higher for this site than Goose Bay in June 1998.”

-         The statement is accurate, but it should also be noted that by late in the growing season, greater numbers of more advanced larval and juvenile suckers were found in Goose Bay.

·  Pg 15.  After stating that collection of larvae at Link River suggests nearby spawning at undetected sites or, more likely, that wind-driven currents transport larvae down the east side of the lake, they note Simon et al.’s (2000a) larval catches at Stone House in 1999 (on the west side of the lake). 

-         Either Stone House is the undetected site or the wind driven currents operate in a manner other than described.

·  Pg 16.  “Researchers have consistently observed sucker larvae in the lower Williamson River with empty guts (L. Dunsmoor, Klamath Tribes, per. com.).  This appears to be related to morphological changes in the lower river, i.e., channelization, that altered flows so that larvae do not move downstream to the lake, where zooplankton number are high, sufficiently fast.”

     - Cooperman (unpublished) has demonstrated that most sucker larvae exit the river within 1 day.  The connection of empty guts to morphological changes of the river channel is speculative at best.

· Pg 18. “Sucker predation by fathead minnows (Pimephales promelas) in the laboratory was greatest when larvae lacked cover (Dunsmoor 1993, Klamath Tribes 1995).”

-         This sentence is hard to square with the following description of the same studies from pg. 94 –“Presence of vegetation structure significantly influenced survival rates in none of the trials with SNS and in two of seven trials with LRS. Interaction of depth and structure factors was not significant in any of the trials”. (emphasis added)

· Pg. 19. burreed, not burr weed.

· Pg. 19. 61 larval suckers were caught on June 30, 1999 at pH 10.31, so the implication that they are not found in water with pH>10.2 is incorrect..

· Pg. 20. In describing juvenile habitat, “It is unclear what habitat and substrate types are preferred.”

-         In the same paragraph, “Highest age 0 sucker densities were found on small mix and gravels and lowest densities were found on fines, sand, and boulders  (Simon et al. 2000a).” It seems clear that the daytime habitat is clean gravel.

-         The remainder of the discussion of this subject is a bit confounded. It seems to be saying that the suckers are found in both places at the same time. Do you mean to suggest that un-vegetated gravel sites only have “overflow” from the vegetated areas? Why is a clear, and repeated pattern that shortnose and Lost River sucker juveniles are found on gravel substrates during the day and that those sites are not typically associated with vegetation questioned based on observations of unidentified suckers of undisclosed age from haphazard sampling? 

· Pg. 21. “Tributary-resident juveniles, mostly SNS, generally are associated with pools having gravel and cobble bottoms (Buettner and Scoppettone 1990).  Reid and Larson (unpub. data) observed age 0 SNS in Willow Creek, a Clear Lake tributary, in sandy pools where they schooled with dace.

-         We can find no evidence of the first statement in Buettner and Scoppettone (1990).

-         Of 31 fish collected for the Klamath sucker genetics study from Willow Creek, 14 are SNS and 17 are KLS. We are unaware of any criteria by which Reid and Larson could identify age 0 SNS “by observation”.

-         It is likely that the confounded juvenile habitat descriptions result from describing KLS juvenile habitats.

· Pg. 24. “The current status of suckers in Clear Lake appears good, although there have been no detailed studies to support this claim.”

-         maybe status is “uncertain but presumed good”

· Pg. 25. “Spawning runs at Clear Lake are more flow dependent”

-         meaning unclear and data not shown – does this mean they like low flows? what cfs?

· Pg. 26. “Preliminary analysis suggests that suckers in Clear Lake are ­relatively young.”

-         the rest of the paragraph does not support this, in fact dwells more on the old age.

· Pg. 26. “In Clear Lake, most SNS mature at age 5 (CDFG 1993).  Growth rates have been assumed to be similar for SNS in Clear Lake…”

-         Similar to what? The paragraph construction suggests the meaning is that they are similar to themselves; probably not necessary to point that out.

· Pg. 27.  “In contrast, suckers captured in 1994-1996 (years with better water quality and higher lake levels)”

-         Where is the evidence to suggest that fish were thin because of water quality?

-         1994 had the lowest lake level on record, so how can that be considered evidence of better anything?

· Pg. 27.  “Gerber Reservoir and its major tributaries have a distinct population of SNS”

-         Evidence?

-         What is point of this sentence relative to a paragraph on spawning?

· Pg. 28. “SNS have been documented in several Gerber Reservoir tributaries including Ben Hall, Barnes Valley, Long Branch and Lapham creeks”

-         Are you sure these aren’t KLS?

· Pg. 29. “This has likely affected wetlands and wet meadows and may have resulted in lowered water tables, increasing the need for irrigation.”

-         Surely there is data on the water table that support this speculation?

· Pg. 35. The unfounded speculation that suckers may migrate out of Wilson Res. into LRDC and the abominable way of saying that you don’t know, a) if it’s true or b) how many are there, is not helpful. The tagging studies in fact suggest that it did not happen last year.

· Pg. 37. “invertebrates identified by MacCoy”

-         is this a citation?

· Pg. 37. The repetition about historical sucker abundance in Lost River, canneries, etc. is unnecessary unless you want to make a new point. Here is one to consider.

-         A. F. McEvoy (1986. The fisherman’s problem, ecology and law in California fisheries, 1850-1980. Cambridge University Press, NY, 368p.) makes a case for the point that “historical” high abundance of certain fish in the late 1800’s early 1900’s was due to ecological release from their primary predator, native Americans who had been decimated by disease and systematically slaughtered in the century before. We’ll expand on this below.

· Pg. 37+. The sections on Historical Account of Sucker Populations and Sucker Movements in Tule Lake Sumps have many repeated paragraphs.

· Pp. 38-39. Paragraphs are repeated and Scoppettone et al. (1985) are variously reported to have caught only 67 suckers or approximately 60 of each species and the population is either unmeasurable or measured with 95% confidence intervals.  What is the elevated condition factor (data?) and what is the elevated productivity?

· Pg. 40. What depth is too shallow? What is the actual depth range and what is the evidence to show that the mouth is too shallow?

· Pp. 40-41. Why did Reclamation reduce flows to 30 cfs in 1999 from 50 cfs in previous years?  Why did a reduction in flows benefit sucker migration and spawning? Is this part of the evidence of “flow dependence” from pg. 25?

· Pg. 42.  “Specific conductance in Tule Lake is high, up to about 1,000 umhos/cm, compared to UKL (120 umhos/cm).  This increase is due to salts leached from soils in agricultural return flows and from bottom sediments of Tule Lake.  High rates of evaporation (over 3 ft per year) in the shallow and warm sumps also increase salt concentrations.   This salt concentration, however, does not appear to be an immediate threat to LRS and SNS.  In Pyramid Lake where specific conductance is nearly 6.5 times higher than Tule Lake, another lake sucker, the cui-ui, Chasmistes cujus, thrives.”

-         This is probably safe extrapolation across species but it is speculation and should be clearly indicated as such.

· Pg. 43. “This area apparently has better water quality due to a lack of submerged aquatic plants and filamentous green algae.“

-         Since emergent vegetation is elsewhere promoted as important for better water quality, this curious relationship should be explained.

· Pg. 44. “In 1999, another water quality monitoring and sucker telemetry study was conducted at Tule Lake (Hicks et al. 2000).”

-         Prior to this paragraph, 2 studies were discussed -  “In 1999, the Refuge began a study of sucker habitat use and water quality in Tule Lake sumps as part of a proposed wetland enhancement project (Hicks et al.  2000).  This was an extension of earlier radio tracking studies done by Reclamation staff from 1992-1995”. So is this a third? Is Reclamation’s not counted?

· Pg. 45. “The lake has six major tributaries, including springs.”

-         Vague. Do all the springs count as “one tributary”? What are the five other tributaries?

·  Pg 46.  “Prior to construction of the dam, the lake level varied from about 4139.9 to 4143.1 ft., with a mean annual variation of about two ft (USBR data).  However, the range may have been even greater from 4140.3 to 4145 ft (USBR 2000b).  Since 1921, water levels have varied from 4136.8 to 4143.3 ft., a range of about 6 ft (USBR data).”

-         The pre-dam high was 4143.08 in March and April, 1906 and the low 4139.93 in June, 1917. The estimated pre-dam range must include the observed minimum, thus 4139.9 to 4145 ft.

· Pg. 47. “Approximately 35 species of fish are known from UKL, 20 are exotic; most of the native species are endemic (Logan and Markle 1993).”

-         This is wrongly cited; there are not 35 fish species in UKL. That count includes the whole basin plus all exotics that have been unsuccessfully introduced.

· Pg. 49. These numbers for population sizes don’t mesh with the cited source - Simon et al. (2000).

· Pg. 50. “Juvenile suckers were collected from fixed sites throughout UKL using beach seines, cast nets, and otter trawls.”

-         Beach seines and larval trawls were at fixed sites. Cast nets and otter trawls were stratified random.

· Pg. 50. Results need to be presented as gear specific results. Age 0 suckers were most common in southern 1/3 not 2/3 of UKL if based on cast net sampling.

· Pg. 50. The seasonal decline in age 0 abundances may also be attributed to losses out the A-canal and canals along either side of the Link River Dam. This is important and is noted elsewhere. One of the problems with so much repetition is that the same data and ideas are not being repeated.

· Pg. 50. Spring catch rates of juveniles are, by definition, of age 1+.

· Pg. 52. Crooked Creek is described as no longer having a spawning run or having a small run on the same page. Which is it?

· Pg. 54. Give the adult index value for 2000 – not just that it went up.

· Pg. 55. The cautionary statements about adult abundance indices need to precede not conclude the discussion of these data. What, if any, conclusions should be drawn from the observations?

· Pg. 57. “Habitat in the lower Link River is crucial to sucker and other fish since it may be the best now available in the reach upstream of Keno.  The lower Link River probably serves as a critical refuge for fish during periods of low DO.”

-         Where is the data to support these statements?

· Pg. 57. There is one report of seasonal “mullet” runs up the Link River prior to construction of the Link River Dam (L. Dunsmoor, Klamath Tribes, per. com.).

-         This is not personal knowledge of Dunsmoor. Identify the source.

· Pg. 57 “Limited endangered sucker sampling has occurred since 1992 in the Link River, Lake Ewauna, the Klamath River, and PacifiCorp reservoirs downstream to Iron Gate Dam.  Few juvenile and adult suckers were collected during this period.  Efforts to collect larval stages in this area have not been attempted since 1992 (M. Buettner, USBR, per.  com.).”

-         on page 59 the report of 1998-1999 work by Desjardins and Markle (2000) contradicts these sentences.

· Pg 58. “Four species of suckers are known from the upper Klamath River, LRS, SNS, Klamath largescale, and the Klamath smallscale sucker, Catostomus rimiculus.”

-         Klamath River? upper?

-         Isn’t this a little late in the document to be describing the fauna? 

· Pg 59. Why is Pacificorp’s salvage effort in Link R. thrown into the middle of a section called “Keno Dam to Iron Gate Dam”?

· Pg 61. CHU 5:  extends from the mouth of the Williamson River at UKL, upstream to the confluence of the Sprague River....

-         In recounting threats to young suckers in this unit, no mention is made of established populations of exotic species.

· Pg 63. “Studies yielded variable LC-50 values for different species and life stages” and “LC-50s for the four water quality variables did not vary significantly between species”

-         These statements are in the same paragraph.

· Pg 67. “Nevertheless, OSU biologists noted a substantial drop in age 0 sucker cast net catches in September and October suggesting these fish were affected by the die-off.” 

-         As noted elsewhere, the seasonal drop is an annual phenomenon, as much associated with behavior as with mortality.

·  Pg. 68. define acronym, APS.

·  Pp. 68-70 Discussion of climate influence on sucker fish kills.

-         A strong review of existing climate data and linkage to water quality is provided for the first year of fish kills but not for other years. Why provide an analysis with N=1 when data for N=3 are available?

-         This one-year analysis assumes normality in the distribution of variables. Are they? For example, a monthly average temperature of 50 can happen from a wide range of biologically meaningful daily averages.

-         Isn’t cloud cover relevant?

·  Pg. 69. AMean lake-wide ammonia concentrations…..”

-         unionized?

·  Pg. 70. Are fish using Pelican Bay during periods of bad water or all the time?  If it and other areas are refugia, the numbers of suckers using the area should increase when water quality is bad relative to their abundance when water quality is good. Merely looking for them when water quality is bad, and finding them, does not establish that the areas are refugia. This should not be separated from the analysis of tagging data below.

·  Pg. 71. If fish are found dead in an area, the parsimonious explanation is that there is something bad about that area. The logical problem occurs when the area is proclaimed a refugium. Now you are forced to make up stories about a dying ground and claim that the fish that die there were sick before they got there and moved from the place that made them sick.  Your evidence is contrary to your story because, “Two and four radio-tagged SNS remained in close proximity to the Wood River in Agency Lake throughout the summers of 1996 and 1999, respectively (USBR, unpub. data).” And from pg. 109, “In August and September when lake levels dropped below 4138 ft, radio-tagged suckers moved from areas of good water quality into other areas where the quality was poorer.” No water quality data is presented with this statement, but assuming one agrees with the evaluation of water quality, the movement is opposite to the “dying ground” story. The story-line “twist” added on pg. 109 is that the contrary movement to areas of poor water quality is based on water depth. Peck (2000) says, “there is a general decrease in depth as mort date approaches”, but his Figure 6 only shows a slight trend and one that is certainly not significant. Peck (2000) also says, “ Sick and dying suckers have been noted to move towards areas of tributary influence, mainly Pelican Bay”, but provides no basis for accepting the validity of the observation.  You have not shown that sick fish move.

·  Pg. 71. Suckers are rarely observed in these areas except possibly during the spawning season.

-         Williamson and Sprague? Possibly?

·  Pg. 71. “The summers of 1995-1997 were periods of low lake mixing and very low DO levels, due largely to the DO demand of rapidly declining and dying AFA blooms and usually higher ammonia levels.”

-         low? very low? UNusually higher? UNionized?

-         data for any of this?

·  Pg. 74. “Lowered lake elevations may increase AFA production and worsen water quality.”

-         The two lowest water years 1992 and 1994 lacked the second, late-summer bloom that typically occurs.  What caused this?  What were the affects on water quality for having a relatively AFA-free lake for the latter part of summer?  Was water quality better?

-         This discussion describes a complex, non-linear system that either implicates intermediate lake levels or suggests that almost any lake level can be associated with poor water quality. Since historical data have been interpreted to indicate that fish kills were common prior to Link River Dam, since the pre-dam minimum elevation was 4139.93, and since no die-off has been documented when elevations were below the historical minimum (pg. 46), the data implicate intermediate lake levels.

· Pg. 86.  With increased human-induced changes in the watershed as a result of agriculture, grazing, nutrient input to the lake increased.  Loss of wetlands, especially those adjacent to the lake was likely also a factor affecting nutrients.

     -  “agriculture and grazing, nutrient ….” ? Did wetland loss increase or decrease nutrients?

· Pg. 87.   Risley and Laenen (1999) found changes in flows in the Williamson and Sprague rivers, when pre-1950 flow data were compared to more recent data.  These data were insufficient to allow determination of what land-use was responsible for the change.

     - Important reference not in Literature cited.

·  Pg. 92. Effects of Irrigation Diversion Dams on Suckers   

     - Have exotic species become established in the pool behind the Chiloquin dam and would they have significant impacts on success of upriver spawning runs?

·  Pg. 93. “Chiloquin Dam is the only major dam affecting suckers that is not part of the proposed action.”

-         This is remarkable, since the next paragraph details its impacts. Further, there is no evidence that listed suckers had viable populations in the Klamath River prior to construction of the Pacificorp dams. With the exception of the Link River dam, the Klamath River dams created habitat for listed suckers.  These dams certainly have altered the river ecosystem, but to claim they have played a role in the decline of listed suckers is unfounded.

·  Pg. 94.  Lab studies show that fathead minnow predation on sucker larvae is significantly influenced by water depth but usually not influenced by vegetation structure or the interaction of depth and structure. “The Tribes submits that as water depth increases to about 0.6 m, the surface orientation of the sucker larvae and the bottom orientation of the fathead minnows results in enough separation to almost eliminate predation, even when the fathead minnows are hungry.”

-         These observations do not support the discussion that follows. As lake elevation increases, the larvae and the fathead minnows move together. The lake has a gradual slope and is not a steep-sided aquarium. Higher lake elevation will not lead to segregation of the two.

·  Pg. 95. “Tule Lake sucker populations, which were perhaps as large or larger than those in UKL.”

-         Evidence?

· Pg. 101 and vicinity.  In the discussion of water quality and sucker status in Gerber Reservoir, several lines of evidence are given suggesting habitat quality is stressful, yet “SNS in Gerber Reservoir are believed to be doing well.”

This conclusion appears to be inconsistent with the data.  This argument also draws into question acceptance of 4800 ft. for Gerber level management.

·  Pg 104.  “From June until August, higher UKL levels are necessary to inundate emergent vegetation used as cover by larval suckers.”

     - Simon et al., (2000) demonstrated no larval suckers in the system by August with most of the cohort in the juvenile stage by July and using open sand and gravel bottoms during the day. 

·  Pg. 104. “Adult suckers prefer water depths of >6ft (Peck 2000).”

&

·  Pp. 107-108. “Ninety-five percent of the observations in spring and summer months were at depths from 3 ft to 15 ft.  One percent of the observations were in water of 3 ft or shallower and only 3% of LRS and 4% of SNS observations were found in water depths greater than 15 ft.  These data show that depths <3 ft and >15 ft are strongly avoided by adult suckers and depths of 6-9 ft are the preferred habitat depth.”

-         The distinction between use and preference is evaluated 3 pages after the statement with the un-reported data on pg. 108 than a preference index of 4 was found for the 6-9 ft depth and “strong avoidance” was found for depths < 3 ft. The conclusion is comparable to saying that people prefer the ghettos of Calcutta because that’s where you find the highest human densities.

·  Pg. 108.   “Adult sucker depth preference was assessed by looking at radio-tagged sucker locations in September and October 1994, when UKL levels dropped below 4137 ft (Peck 2000).  This period of minimal lake elevations provides the best basis to estimate true sucker species depth preferences.  This is so because water quality conditions improve in the fall, removing the confounding influence on fish distributions due to poor water quality avoidance.”

- “This is so” ?. This type of language has no place in a scientific document. Seasonal movement into deeper water is a common behavioral pattern in many fish and is the simplest explanation of movement into deeper water in the fall.

·  Pg. 108, para 5.  Go to Simon et al. (2000a), pages 88-91.  (document #4 at http://www.mp.usbr.gov/kbao/esa/index.html).  Distribution charts in Simon et al. (2000a) do not suggest that adult suckers use only the small portion of UKL outlined in this paragraph.  Why the discrepancy?   The USGS trammel net data were mentioned but no useful observations presented.  Extrapolation from radio-tagged suckers to all adult suckers is unwise at this point as behavioral bias in radio-tagged suckers is not understood.

·  Pp. 109-110. It is not clear from Peck (2000), but it seems that all of the telemetry depth data are daytime observations. The obvious question is what happens at night? Certainly spawners come into shallow springs at night. Is this a behavior that also happens at other times of year?

·  Pg. 109. “During 1994, a low water year, many of the radio-tagged suckers were concentrated in Ball Bay in early summer, most likely feeding (Hazel [1969] found that the north end of the lake had the highest biomass and variety of benthic organisms, some of which are eaten by suckers).  In July, as DO levels declined below 4.0 mg/l the fish had moved from the Ball Bay area closer to the entrance to Pelican Bay where water depths were >3 ft.  In August and September when lake levels dropped below 4138 ft, radio-tagged suckers moved from areas of good water quality into other areas where the quality was poorer.  In 1996, when water levels were higher, radio-tagged sucker movement in summer was often from areas of poor water quality to higher quality.  Data from these two years indicates that suckers move in response to water quality changes but may avoid areas of higher water quality when such areas become too shallow.  Based on telemetry, the critical minimum acceptable depth is 3-4 ft; however, >4 ft would be much safer because this is the depth preferred by suckers.”

-         Translation from table 4 in Peck (2000):  “many … concentrated in Ball Bay” = 9 of 31 suckers located in June. There were 8 offshore Pelican Bay and 7 in Shoalwater Bay. They were no more “concentrated” in Ball Bay than in the other two areas.

-         Translation from table 4 in Peck (2000):  “In July … to entrance of Pelican Bay” = 15 fish in Pelican Bay (an increase of 7 fish from June) and 0 fish in Ball and Shoalwater bays. There was a drop in sightings from 31 to 22 so 9 fish seen in June were unaccounted in July.

-         Translations from table 4 in Peck (2000):  “In August and September when lake levels dropped below 4138 ft, radio-tagged suckers moved from areas of good water quality into other areas where the quality was poorer.” = 23 sightings offshore of Pelican Bay in August (increase of 8 from July), 26 in Mid North in August (increase of 21 from July) and 6 LRS back into Ball Bay. So, in August, there were 29 fish moving to “good water” and 6 LRS to “poor water”. These patterns are complicated by an apparent increase in number of tagged fish (a maximum of 31 fish sighted prior to August and 64 sighted in August). It would be instructive to know if the 6 LRS that moved back into Ball Bay had recently been tagged and might be demonstrating abnormal behavior. The August to September sightings are comparable (64 each month but the numbers per species differ) and the patterns follow. From August to September there is an increase or no change in SNS sightings in all areas with Ball Bay going from 0 to 11. Sightings of LRS decline by 18 offshore Pelican Bay, by 11 at Mid North, and by 4 in Ball Bay. Sightings of LRS increase by 5 in Ball Point Trench, by 4 offshore Williamson, and by 11 in Mid Lake. Part of the decline in LRS sightings is due to 13 fewer sightings, so all of the LRS patterns as well as the SNS pattern are most simply explained as haphazard movements over the whole area. The data most certainly do not demonstrate that fish move to poor water areas when lake level declines.  This, in fact, is the conclusion reached earlier, on pg. 71 of this document, “In August and September when lake levels dropped below 4138 ft, radio-tagged suckers moved further offshore even though water quality remained good.” 

-         Assuming Peck’s (2000) data were converted from metric to English using 3.28 ft/m, his data show nothing remotely approaching a demonstration of “critical minimum acceptable depth”. Peck’s (2000) Figure 2 shows that 50% of all tagged fish from 1993-1998 were observed at daytime depths of about 4.26 – 11.15 ft. In other words, 50% were found at depths either less than 4.26 ft or greater than 11.15 ft. Since the document states that “Ninety-five percent of the observations in spring and summer months were at depths from 3 ft to 15 ft (pp. 108-109)”, an observation which can not be obtained from Peck (2000), and assuming normality of the data, 22.5% of the observations were made between 3-4.26 ft.  There is nothing in the data to say that 3-4 ft is critical and no demonstration of cause and effect.

-         As is often the case, the title is incorrectly cited in the literature cited section.

· Pg. 114.  “Channelization and diking of the lower Williamson and Wood rivers has shortened and widened both rivers.... Slower velocities may delay emigration of larval suckers to UKL.”

     - Cooperman (unpublished) has demonstrated that most larval suckers move through the river quickly. The role of channelization in recruitment failure is highly speculative.

·  Pg 119. “With ice-cover stratification occurs and near bottom water may become anoxic leading to release of high levels of unionized ammonia from the sediments into the water.”

-         Is there a citation for this process? Doesn’t the fraction of unionized ammonia increase positively with temperature and pH? Wouldn’t winter conditions make this scenario unlikely?

·  Pg 127. In dry years... “July elevations range from 4139.1 to 4140.9 ft (average 4140.2 ft).  None of the emergent vegetation is inundated at the lowest elevation and 2-28% at the highest elevation ... So little emergent vegetation habitat is available that it is highly unlikely that a year-class could develop.”

     -  Flooded macrophytes are important for larvae in early July but not in late July when they are juveniles. Further, higher elevations may benefit recruitment of later-spawning cyprinids, especially exotic fathead minnows that need submerged substrates for egg deposition.

·  Pg 134.  “New Earth Corp. monitored larger suckers (> 7.5 cm FL) that were found on their debris reduction screens and on AFA harvest screens in 1996 (Gutermuth et al. 1997).  One-hundred-fifty-seven suckers were collected off the debris screens and 140 off the harvest screens.  Many of these fish were probably dead before they reached the screens.”

-         The speculation on the health of juvenile suckers at capture in the Irrigation District several miles from UKL is important to clarify. Do fish die in the lake and passively drift into the canals or actively seek the canals and then die?

· Pg 134.  “Screening of A-canal in 2002 should markedly reduce number of larger fish entrained and numbers salvaged should go down”

-         Reduction of mortality, not reduction of salvage, should be a goal. The movement out of the lake is a behavioral response, possibly related, in part, to water quality. If prevented from leaving UKL at times of bad water quality what are the consequences of screening?  If forced to remain in the lake, their densities might overrun refugia and increase mortality. The system is too poorly understood to predict un-intended consequences.

 Minor editorial comments