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.
3 Current Status of Aquatic Ecosystems: Lakes INTRODUCTION Natural lakes that were suitable for occupation by suckers before lancT- use clevelopment ancT water management incluclecT Upper I(lamath Lake, Lower I(lamath Lake, Tule Lake, ancT Clear Lake (Figure 1-31. All of these lakes have been changed morphometrically ancT hycTrologically ancT are now usecT in the I(lamath Project water-management system for storing ancT routing water. Gerber Reservoir is also part of the water-management sys- tem, but its location was previously occupied by a marsh rather than by a lake. Other lakes relevant to the welfare of suckers inclucle those lying behind five main-stem clams that, except for I(eno Dam, incorporate hycTro- electric production facilities (Figure 1-3, Table 3-11. The last in the se- quence of main-stem clams, Iron Gate Dam, provides reregulation capabil- ity for the main stem of the I(lamath River as explainecT below. Of the lakes usecT for storage ancT routing, Upper I(lamath Lake, Clear Lake, ancT Gerber Reservoir support the largest populations of listecT suck- ers (see Chapter 6 for a cletailecT treatment of the suckers), ancT these three lakes have been the main focus of ecological ancT limnological analysis relatecT to the welfare of suckers. Upper I(lamath Lake has been stucTiecT especially intensively because it potentially wouicT support the largest popu- lation of suckers ancT shows the greatest number of environmental prob- lems, as incTicatecT by episodic mass mortality of aclults ancT probable harcT- ships in all life-history stages. Clear Lake ancT Gerber Reservoir afford a useful comparison with Upper I(lamath Lake because the sucker popula- tions there have not suffered mass mortality ancT are generally more stable 95
96 _ C~ ~= s~ ~ ~ C~ · C ~ Q 5 _ s~ O C~ ~ _ _ C~ C~ C~ s~ a~ Q o C~ a~ o o ._ C~ s~ o ._ C~ C~ 1 ¢ C~ s~ C~ - o ._ U) ._ C~ s~ C~ - o o s~ ~o ~4 ._ x C~ ._ ._ x C~ ._ ._ C~ O O O O O ~ - 1 - 1 - 1 ~ 1 00 ~ O 00 0 · ~1 0 ~1 1 V ~ ~ ~, ~ o o ~ ~ ~ ~ ~ ~ ~ 1 V - 1 - 1 ~ ~ o o o o o o o o o o o o o o o o o o o ~ o o o o o o o ~ ~ ~ oo o ~ o ~ o ~ oo ~ ~ oo o 0 - 1 - 1 ~ ~ ~ ~ ~ - 1 V ~ - ^ o o o o o o o o o o o O O O O O ~ - 1 0 o ~ ~ o ~ ~ ~ o - 1 r-1 ~ ~ o o o o o o o o o o o O O ~ O O ~ 1 0 ~ ~ 1 O ~ ~ ~ ~ ~ ~ O ~ ~ oo ~ ~ r-1 ~ r-1 oo bC · - O O O O ~ ~ ~ ~ ~ ~ O O O O o o ~ ~ ~ ~ ~ ~ O O O =N C~ bC C~ s~ ~ O O O O ~ .0 ~ ~ ~ ~ ~ O C~ oo ~) ~) ~) s ~0 Q C~ j~ ~ O c~ Q s~ s~ D ,, ~ , D C~ ~ O ~ ~ ~ ~ ~ ~ O O O O E- ~ ~ ~ _) ~ ~ s~ E- C~ O e e~ ~ O z) ~ ~ ~0 ~ ~o C~ X ~ Q C~ ~ O ~ ,) s~ . .m ~ >^ ~ ~ =) C~ ~ O ~ C~ ~ ~ O ¢ >~ ~ - o ~ ~ ~- ^= o ~ o ~ ·Q Q O co s~ .s ~ os O =-X s~ '~ C~ Q s~ . C~ (d ~ Q ~ O C~ o ~ E~ =~^ ~i C~ (~! c-] .m · ·o ~ p4 ~ D ~ ^ ~ ,, e O ~ O C~ . S~ o C~ o . ~ ~1 o I C~ - 1 . . _ Q Q Q Q O ~ O O O - 1 Q s~ ~ O .. ._ ~ o . _ C~ (~! .> ~ s~ O ~ ¢
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 97 than the populations of Upper Klamath Lake. The hyciroelectric reservoirs on the main stem have been stucliecI sparingly anti are of less interest than other lakes from the viewpoint of listecI suckers. The lakes shown in Table 3-1 clo not serve as habitat for coho salmon, which are blockecI by Iron Gate Dam from entry into the upper Klamath basin. Limnological characteristics of the waters behind Iron Gate Dam are potentially important to the coho salmon, however, in that waters releasecI from the clam have a large influence on the water-quality characteristics of the Klamath River main stem, especially near the clam. Reflecting the rela- tive amounts of research or monitoring ancI the apparent ranking of lakes with respect to their importance for the enciangerecI ancI threatened fishes, this chapter clevotes most of its attention to Upper Klamath Lake, some to the other lakes that are used for storage anti routing of water, ancI some to waters above Iron Gate Dam that hoicI non-reproclucing populations of listecI suckers anti have the potential to affect coho downstream; the rem- nants of Tule Lake anti Lower Klamath Lake provide little lacustrine habi- tat at present, but offer potential for restoration. UPPER KLAMATH LAKE Description Upper Klamath Lake is the largest body of water in the Klamath basin anti is one of the largest lakes in the western United States (about 140 mid. The lake anti its drainage lie on volcanic deposits clerivecI in part from the nearby Crater Lake calclera, which took its present form as a result of the eruption of Mount Mazama (about 6,800 BP). The lake also shows a strong tectonic influence, however, as is evident from a pronounced scarp along its southwestern ecige (Figure 3-11. Although Upper Klamath Lake has a very low relative clepth (ratio of clepth to mean cliameter), it has substantial pockets of water over 20 ft creep (maximum, 31 ft at a water level of 4,141.3 ft above sea level; USER 1999 as cited in Welch anti Burke 20011. The northern, southern, anti eastern portions of the lake anti Agency Lake, which is connected to Upper Klamath Lake anti is here treated as part of it, are uniformly shallow; they offer water little creeper than 6 or 7 ft at mean summer lake elevation (4,141.3 ft above sea level). Even though specific runoff for the watershed of Upper Klamath Lake is relatively low (about 300 mm/yr), the hyciraulic residence time of Upper Klamath Lake is only about 6 ma because the lake is shallow (there is consiclerable interannual variability). The flat bathymetry of the lake also causes its surface area to be quite sensitive to changes in water level. Before the construction of Link River Dam, which was completecI in 1921, the water level of Upper Klamath Lake fluctuatecI within a relatively
98 FISHES IN THE KLAMATH RIVER BASIN Pelican\, ~ ~-~: Agency \,:~,1 ~G~\ Lake ~~ ~ " U~ | OREGON Area of Detail :3 Marsh 12+ Feet 6 Feet ~ _ _ 1 0 1 2 3 4 5 miles en\ Upper ~ Klamath N: Lake `~\ / . / ~ i:. \ \ \~ A Link River Dam \ FIGURE 3-1 Bathymetric map of Upper Klamath Lake and Agency Lake showing depths at the mean summer lake elevation of 4,141 ft above sea level. Contours are from data of U.S. Bureau of Reclamation (1999) as reported by Welch and Burke (2001~. Source: Welch and Burke 2001.
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 99 narrow range (about 3 ft), as wouicI be expected for a natural hycirologic regime (Figure 3-21. Although irrigation was uncler way in the basin at that time, there was no means of using the lake for storage. Water level in the lake was cleterminecI by a lava clam at the outlet (4,138 ft above sea level; USFWS 20021. Even uncler drought conditions, the lake level remained above the level of the natural outlet, except briefly cluring oscillations caused by wind (USFWS 2002). When Link River Dam was constructed, the natural rock clam at the outlet of Upper I(lamath Lake was removed so that the storage potential of the lake couicI be used in support of irrigation. Thus, since 1921, lake levels have varied over a range of about 6 ft rather than the natural range of about 3 ft (Figure 3-21. Drawclown of about 3 ft from the original minimum water level of the lake has occurred in years of severe water shortage (1926,1929, 1992, anti 19941. The operating range of the lake in the context of mean clepth ancI contact between the lake ancI its wetiancis has raisecI numerous questions about the environmental effects of water-level manipulations, especially uncler the most extreme operating conditions (USFWS 20021. The U.S. Bureau of Reclamation (USBR 2002a) has proposed operating Upper I(lamath Lake over the next 10 yr according to guiclelines that reflect recent historical operating practice (Figure 3-2; Chapter 1). The open ques- tion for researchers ancI for the tribes ancI government agencies charged with evaluating the two enciangerecI suckers is whether the USBR proposal for future operations is consistent with the welfare of listecI suckers in Upper I(lamath Lake. In a biological opinion issued in response to USBR's proposals, the U.S. Fish anti WilcIlife Service (USFWS 2002) has concluclecI that operations shouicI involve limits on water levels that are more restric- tive than those proposed by USBR. USFWS has temporarily accepted the water-level criteria proposed by USBR (2002a), but has required a revised approach to predicting water availabilities in any given year (Chapter 11. The USBR 10-yr plan is basecI on a commitment of USBR not to allow Upper I(lamath Lake to fall in any given year below the minimum water levels that were observed in 1990-1999 for four hycirologic categories of years ancI not to allow the interannual mean water levels for these catego- ries to fall below recent interannual means (1990-19991. Figure 3-2 shows the March 16-October 30 operating range basecI on interannual means for each of the four hycirologic categories. The database for the definition of the categories incluclecI water years 1961-1997 (USBR 2002a, p. 391. The calculations were basecI on the outflow from Upper I(lamath Lake for April-September. Years above the mean outflow, which is 500,400 acre-ft, are clesignatecI "above average." Those within one stanciarcI deviation be- low the mean are clesignatecI "below average"; the expected long-term frequency for these years is 34°/O (on the basis of a normal distribution. Curve-fitting was not suitable for evaluating years of lower flow, however.
100 . - ~ . 5- 50 be ~ ~ . = ~ o Cat ^= ~ 2 ~ =~ ~ _ O T¢ ~ o _ aa~deauea~ ~ ~ ~ no 0 LC) ~ A =0 LC) o ~ ~ I I I I T I I Stoat ~ ~ := o ~ o ~ ~ ~ > 0 ~ ~ ad ~ O =~ ~ ~~ ~ a ~ a, ~ a, ~ a, 'M ~ (~d ~ ~ ~ c 2 ~ ~ ~ ,5 ~ a lo' I i 1 1 1 O-e'° =) 1 1 1 1 1 1 1 >~= 0 50 cot ~ ~ ~ ~ ~ ~ ~ A ~ ~ ~ ~ ~ ~ ~ ~ 50 (laAarIcaSaAoqV baas) uol~cAal~ao~JmSla~ ~ cd cry .o Em
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 101 Two extreme years, 1992 anti 1994, were clesignatecI "critical ciry" anti account for about 5°/O of the total. By difference, a fourth category, clesig- natecI "ciry," is clefinecI; it accounts for about 1 1% of years. For each category of years, the maximum water levels occur in the spring. Water levels typically begin relatively high as of micI-March anti then rise slightly, after which they fall because of the cumulative effects of cirawclown anti, after rune, the reclucecI volume of runoff (Figure 3-31. Operations for the four hycirologic categories differ most notably in their lower extremes, which occur after luly. In comparison with a baseline condition, which USER defines as lacking I(lamath Project operations but with all project facilities in place, proposed operations typically produce water levels that are above the baseline between March anti the encI of lune anti below the baseline cluring the last half of the summer or fall (USER 2002a). Upper I(lamath Lake receives most of its water from the Williamson River (inclucling its largest tributary, the Sprague River) anti the WoocI River. Aciclitional water sources inclucle precipitation on the lake surface, direct drainage from smaller tributaries anti marshes, anti springs that bring water into the lake near or beneath the water surface. The waters of the lake have only moderate amounts of clissolvecI solicis (interseasonal median, about 100 ~S/cm) anti the same is true of alkalinity (interseasonal median, about 60 mg/L as calcium carbonate). As clescribecI below, the lake is naturally eutrophic, but concentrations of nutrients in the water column may have increased over the last several clecacles. The fish community of the lake couicI be clescribecI as a diverse array of nonnative species superim- posecI on a previously abundant but now reclucecI group of native fishes, 4,144 - 4,143 - 4,142 - 4,141 - 4,140- ¢ au au _' ~ 4,137- au au ~ 4,135- 4,139 - 4,138 - 4,136 - 4,134 - -7~( at, ma, . 1 1 1 1 1 1 Jan Feb Mar Apr May Jun Jul 1999 >< 1992 box _, ,,,~. ~ >I _~ ~ Aug Sep Oct Nov Dec FIGURE 3-3 Water level in Upper I(lamath Lake in year of near-average mean water level (1999) and year of extremely low water level (lowest 5%; 1992~.
102 FISHES IN THE KLAMATH RIVER BASIN most of which are enclemics (Chapter 61. The biota in general has uncler- gone consiclerable change in the last few clecacles. Upper Klamath Lake has several large marshes at its margins. The area of the marshes has been greatly reclucecI (loss of about 40,000 acres from the lake margin; USFWS 20021. The remaining marshes are most strongly connected to the lake at high water anti are progressively less connected at lower water levels clown to about 4,139 ft above sea level, at which point they become clisconnectecI. Poor water quality in Upper Klamath Lake causes mass mortality of listecI suckers anti may suppress the suckers' growth, reproductive success, anti resistance to disease or parasitism. Potential agents of stress anti cleath inclucle high pH, high concentrations of ammonia, anti low clissolvecI oxy- gen (USFWS 20021. Extremes in these variables are explainecI by the pres- ence of clense populations of phytoplankton (primarily the cyano bacteria! taxon Aphanizomenon flos-aq?vae), especially in the last half of the growing season (Kane 1998, Welch anti Burke 20011. Because phytoplankton pop- ulations annually reach abundances exceeding 100 ~g/L of chlorophyll a, the lake can be classifiecI as hypertrophic (or, equivalently, hypereutrophic) according to stanciarcI criteria for trophic classification of lakes (OECD 1982: peak chlorophyll over 75 is hypertrophic). Hypertrophic lakes often show extremes in chemical conditions resembling those observed in Upper Klamath Lake. The main subjects of interest with respect to Upper Klamath Lake proper Miscounting the tributaries, which are clealt with in the next chap- ter) inclucle factors that have been suspected by researchers or by govern- ment agencies of being potentially harmful to the enciangerecI suckers. Where water quality is concerned, the causes of the current trophic status of the lake are of great interest, as is the current predominance of a single algal species, Aphanizomenon flos-aq?vae, in the phytoplankton. Within the suite of variables affected by trophic status, special attention must fall on pH, ammonia, anti clissolvecI oxygen, all of which have the potential to be clirectly or inclirectly harmful to the welfare of the enciangerecI suckers. For all water-quality variables, associations between water level anti water qual- ity are of special interest because USER has the potential to moclify opera- tions so as to control water level. Finally, physical habitat, especially as affected by water level, is of concern anti will be clealt with here. Nutrients ant! Trophic Status of Upper Klamath Lake Nutrient limitation of phytoplankton in lakes usually is seasonal anti almost always is associated with nitrogen, phosphorus, or both of these elements. Typically, phosphorus anti nitrogen are reaclily available cluring winter because clemancI is low. In spring, the most available forms are taken
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 103 up, ancI nutrient limitation often ensues. If the most reaclily available forms are available in quantities above about 10 ~g/L, there is a strong implica- tion that no limitation is occurring (e.g., Morris ancI Lewis 19881; at lower concentrations, nutrient limitation is possible but may be clelayecI by inter- nal storage. Nutrient limitation often is relievecI in the fall by creep, continu- ous mixing of the water column, cleclining irracliance, ancI lower metabolic rates caused by lower temperatures. Nitrogen limitation can be clefeatecI by some taxa of bluegreen algae ~cyanobacteria) capable of fixing nitrogen (converting N2 to NH31. Nitro- gen gas (N2) is present in consiclerable quantity in water, ancI the overlying atmosphere acts as a large reservoir that can replenish removal of nitrogen gas by nitrogen fixation. The heterocystous bluegreen algae which have a special cell, the heterocyst fix nitrogen reaclily, although the fixation pro- cess requires high intensities of light (Lewis ancI Levine 19841. Heterocys- tous bluegreen algae clo not grow well in some situations, however, for reasons that are only partly unclerstoocI (Reynoicis 19931. Thus, nitrogen clepletion sometimes can occur without inclucing growth of nitrogen fixers. Nitrogen fixers grow well in most warm, fertile waters of high pH. When phosphorus is abunciant in such waters but nitrogen is scarce, nitrogen fixers have a competitive acivantage ancI often become clominant elements of the phytoplankton. This is the situation in Upper Klamath Lake. For the phytoplankton as a whole in Upper Klamath Lake, nitrogen is limiting (see below), but Aphanizomenon has circumventecI nitrogen limitation through nitrogen fixation ancI thus clominates the community. Typically, the most effective way to contro! phytoplankton abunciance in lakes is to restrict phosphorus supply. Restriction of nitrogen supply is not as effective, because it may leacI to the clevelopment of nitrogen fixers that are able to offset restrictions in nitrogen supply. Thus, the most obvi- ous way of attempting to contro! phytoplankton populations in Upper Klamath Lake is to restrict phosphorus supply. As explainecI below, Upper Klamath Lake presents special clifficulties for strategies involving contro! of phosphorus. Phosphorus in Upper Klamath Lake The watershecI of Upper Klamath Lake is geologically rich in phospho- rus (Walker 20011. Springs have a meclian phosphorus content of about 60 ~g/L as soluble reactive phosphorus, which BoycI et al. (2001, citing Walker 2001) take as an estimate of the backgrouncI clischarge-weightecI mean phosphorus concentration. This may be an unclerestimate, given that springs typically have little or no particulate phosphorus or soluble organic phos- phorus, both of which wouicI be present in natural runoff from the water- shecI. In contrast, watershecis of granitic geology often have clischarge-
104 FISHES IN THE KLAMATH RIVER BASIN weighted mean total P concentrations of 20 ~g/L or less (inorganic P about 5 ~g/L), proviclecI that they are not clisturbecI by human activity (e.g., SchincI- ler et al. 1976' Lewis 19861. Because background concentrations of phosphorus reaching Upper I(la- math Lake are quite high, the lake probably supported clense populations of phytoplankton before lancI-use clevelopment. Early observations indicate that the waters were green, ancI thus eutrophic, at a time when water quality wouicI have been changed little from the natural state. If, as sug- gestecI by BoycI et al. (20011' phosphorus reaching the lake wouicI have hacI originally a clischarge-weightecI mean phosphorus concentration of about 60 ~g/L, phosphorus in lake water wouicI have been somewhat below 60 g/L (because of sedimentation of some phosphorus) in the absence of internal loacling (net increase originating from secliments). On the basis of empirical relationships between chlorophyll a ancI phosphorus (OECD 19821' the mean chlorophyll a in the growing season with total phosphorus at 60 ~g/L wouicI have been in the vicinity of 20 ~g/L, which wouicI have corresponclecI to short-term maximums of 40-60 ~g/L, or about 20% of the current maximums. The concentrations of phosphorus in the lake couicI have been higher, however, if substantial internal loacling from sediments occurred uncler natural conditions, in which case chlorophyll couicI also have been higher. Monitoring of phosphorus entering the lake has shown that the current clischarge-weightecI mean phosphorus concentration in waters entering Up- per I(lamath Lake is near 100 ~g/L, about 40% of which is consiclerecI to be anthropogenic (BoycI et al. 20011. Concentrations in the lake cluring spring are only about 50 ~g/L (BoycI et al. 2001' Figure 2-6; there is consiclerable variation from year to year); the difference between the supply water ancI the concentrations in spring is accounted for by sedimentation of the par- ticulate fraction of incoming phosphorus ancI by mechanisms that convert incoming soluble phosphorus to particulate phosphorus that can undergo sedimentation. The currently observed total phosphorus concentrations in spring, if not supplementecI by any other sources, wouicI support mean algal abundances cluring the growing season corresponding to chlorophyll a at 20 ~g/L or less, according to equations clevelopecI by the Organization for Economic Co-Operation ancI Development ( OECD 19821. When the growing season begins (in about May), Upper I(lamath Lake shows a steady rise in concentrations of total phosphorus culminating in summer concentrations of 200-300 ~g/L (BoycI et al. 2001' Figure 2-6; there is consiclerable variation from year to year). These concentrations greatly exceed the clischarge-weightecI mean concentrations in inflowing water (about 100 ~g/L) ancI also greatly exceed the concentrations in the lake cluring spring (about 50 ~g/L, Figure 3-41. Thus, the great increase in
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES i_ ~ 350- _' o 50 au 300 - 250 - 200 - 150 - 100 - 50 - O- 105 Current Inflow Background Inflow 1 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec FIGURE 3-4 Total phosphorus concentrations in Upper I(lamath Lake during 1997 (an arbitrarily chosen year) and approximate discharge-weighted mean total phos- phorus for inflow for background and for current conditions. Source: Data from Walker 2001. concentrations of phosphorus cluring the growing season must be attrib- utecI to an internal source (secliments). Concentrations of soluble phosphorus in sediments of Upper I(lamath Lake were stucliecI by Gahier anti Sanville (1971), as reported by Bortleson anti Fretwell (19931. Sediment samples taken at one location in 1968-1970 showed a median soluble phosphorus concentration in the interstitial waters of about 7,000 vigil, or about 25 times the maximum concentrations ob- servecI in the overlying lake water (another location showed less extreme deviation from lake water). Thus, for at least some portions of the lake, sediment pore waters contain substantially more soluble phosphorus than the overlying lake water anti can serve as an internal source of phosphorus if the phosphorus leaves the sediments. This is a common situation in fertile lakes. The efficiency with which phosphorus is releasecI from sediments varies greatly according to the conditions in a particular lake. There are four potential mechanisms of release: (1) If the sediments are clisturbecI by wincI- ciriven currents or by other means (organisms or clegassing), interstitial phosphorus can be transferred to the water column simply by agitation. (2) Decrease in the reclox potential (increase in availability of electrons) in the surficial sediments caused by intensive microbial respiration, as wouicI be the case for highly organic sediment, can cause biogeochemical changes that result in acceleratecI release of mineralizecI or soluble organic phospho- rus from the sediments to the overlying water, even if the sediments are immobile. (3) High pH at the sediment surface may cause release of acI- sorbecI phosphorus from sediments, with or without agitation of sediments.
106 FISHES IN THE KLAMATH RIVER BASIN (4) In shallow lakes, phytoplankton cells may, uncler calm conditions, sink to the sediment surface, where phosphorus is more concentrated than in the water column, anti then be resuspenclecI either by wincI or by buoyancy control mechanisms after assimilating phosphorus, thus bringing phospho- rus from the sediments to the water column. Internal loacling in Upper I(lamath Lake is causecI by one or more of these four mechanisms, which are not mutually exclusive. Chlorophyll concentrations in Upper I(lamath Lake increase in parallel with concentrations of total phosphorus in the water column from May to luly (BoycI et al. 20011. Thus, the ciata indicate that phytoplankton are assimilating an internal phosphorus loacI leacling to an increase in their biomass. The growth process culminates in concentrations of phytoplank- ton chlorophyll a typically near or above 200 ~g/L (BoycI et al. 20011. At such high abundances, phytoplankton approach the maximum sustainable biomass basecI on light availability (self shacling) rather than nutrients (Welch anti Burke 20011. The specific limit for phytoplankton biomass basecI on light rather than nutrients clepencis on physical conditions in a lake anti physiological characteristics of the dominant algae (Wetze! 20011. Because internal loacling increases the phosphorus inventory of the water column in Upper I(lamath Lake, thus sustaining high populations of bluegreen algae, its mechanisms are of special importance to the nutrient economy anti trophic status of the lake anti therefore to water-quality conditions that affect fish. The simplest mechanism of release of phosphorus from the sediments is disturbance of the sediments. As proposed initially by Bortleson anti Fretwell (1993), that mechanism is highly feasible in Upper I(lamath Lake because of the lake's low relative clepth (a low ratio of clepth to area), which is an indication that sediments will easily be mobilizecI by strong wincis, at least over the large expanses of shallow water. Thus, clecompo- sition processes in the sediments may liberate phosphorus from particu- late form, anti this phosphorus can be transferred to the water column simply by wincI-generatecI sediment movement. Release of gas bubbles from the sediment or invertebrate activity (bioturbation) can produce similar effects. The role of sediment movement in mobilizing phosphorus in Upper I(lamath Lake is unknown, but the ability of the wincI to move sediments reaclily over much of the lake bottom is generally acknowI- ecigecI (Bortleson anti Fretwell 19931. Release of phosphorus from sediments also can occur without any movement of the sediments. If there is a substantial concentration gradient of soluble phosphorus between the sediment pore waters anti the overlying water, the potential exists for diffusion of phosphorus from the pore waters to the overlying lake water anti distribution of the releasecI phosphorus by ecicly diffusion or bulk mixing of the water column. The key requirements
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 107 for the process inclucle presence of a substantial concentration gradient (which exists in at least some places in Upper I(lamath Lake, as inclicatecI by the stucly cited above) ancI absence of any physical or chemical barrier to diffusion of soluble phosphorus. It is well known that iron in the ferric state can bincI phosphorus, thus restricting its movement from sediments to water (Mortimer 1941, 19421. Loss of the precipitated (ferric) iron from the surface of lake sediments occurs when sediments are anoxic for long intervals, by conversion of iron to a soluble (ferrous) state. Loss of ferric iron facilitates exchange between the sediment pore waters anti the overlying water ancI releases phosphorus bouncI by ferric iron. The result can be release of large amounts of phospho- rus from the sediments (internal loacling). The release of phosphate from sediments caused by changes in the oxidation state of iron is most likely in lakes that show prolongecI anoxia at the secliment-water interface. Unlike creeper lakes, Upper I(lamath Lake cloes not remain stratified for the entire growing season, but rather for periods of only clays or at most weeks at a time, so a key role for the reclox mechanism seems less likely than it wouicI in some other lakes, but it cannot be rulecI out. The adsorption of phosphate by ferric complexes is influencecI by pH. Phosphate may pass from a sediment surface to the overlying water if the pH is high (> 8; literature reviewed by Marsclen 1989), even without con- version of ferric to ferrous iron. Thus, internal loacling in Upper I(lamath Lake may involve iron anti phosphate uncler oxic conditions at the sediment surface if pH is high. This mechanism is consiclerecI by some researchers to be of special importance in Upper I(lamath Lake (summary in BoycI et al. 2001). Biogeochemical mechanisms (loss of oxygen anti high pH) involving release of phosphorus from sediments typically are clescribecI in terms of abiotic reactions involving iron, but there is some evidence that bacterial metabolism also accounts for bincling or release of phosphorus at the secli- ment-water interface (Davison 19931. Bacteria also control the oxidation conditions on the sediment surface. Phosphorus mobilization from sediments of Upper I(lamath Lake also may involve clirect contact between the algae ancI the sediments. Aphanizo- menon contains pseuclovacuoles that function as buoyancy-control mecha- nisms. Uncler some circumstances, which may coincide with nutrient clefi- ciency, the algae may show higher specific gravity than at other times ancI thus show an increased tendency to sink. Because nutrients typically are more available in creep water than in shallow water, sinking, which wouicI be notable primarily under calm conditions, can allow algae to reach nutri- ent reserves that otherwise are not available. In Upper I(lamath Lake, a small amount of sinking couicI allow a substantial fraction of the algal population to have clirect contact with the sediments, where phosphorus . . . . .. . . .. . .. .
108 FISHES IN THE KLAMATH RIVER BASIN supplies are rich. Thus, algae may be mobilizing phosphorus through direct contact with the sediments (cf. Ganf anti Oliver 19821. Nitrogen in Upper Klamath Lake The total nitrogen loacI to Upper I(lamath Lake has been calculated for total-maximum-ciaily-loacI (TMDL) purposes as 663,000 kg/yr (BoycI et al. 2001, Walker 20011. Thus, the mass ratio of nitrogen to phosphorus for loading uncler present circumstances is about 3.6:1. This ratio is extreme in the sense that mass transport of nitrogen anti phosphorus from watersheds to lakes typically involves mass ratios well in excess of 5:1 (OECD 19821. Although human activities tencI to cause higher relative enrichment with phosphorus than with nitrogen, even clisturbecI watersheds typically have much higher nitrogen transport than phosphorus transport. The ratio of nitrogen to phosphorus typically is evaluated with respect to phytoplankton growth by reference to the RecifielcI ratio, which is an empirically cleterminecI value for the relative amounts of nitrogen anti phos- phorus that are neeclecI by phytoplankton for growth (Harris 19861. The RecifielcI ratio is 16:1 on a molar basis anti 7.5:1 on a mass basis. In environments that show ratios far above the RecifielcI ratio, strong anti persistent phosphorus limitation is expected. Where the reverse is true, all taxa of algae are likely to be nitrogen-limitecI except those capable of nitrogen fixation. Thus, where the nitrogen:phosphorus ratio is low, as it is in Upper I(lamath Lake, the nutritional conditions are ideal for dominance by nitrogen-fixing bluegreen algae, such as Aphanizomenon flos-aq?vae. The fixation of nitrogen by Aphanizomenon flos-aq?vae has the effect of raising the nitrogen:phosphorus ratio by acicling atmospheric nitrogen to the lake through the fixation process. While the nitrogen:phosphorus ratio still is low, a rise in this ratio clue specifically to Aphanizomenon has increased the ability of the lake to produce algal biomass. Explanations of Dominance by Aphanizomenon A recent analysis showed that akinetes, which are resting cells of Aphanizomenon flos-aq?vae, are concentrated in recently accumulated secli- ments but not in sediments of an earlier era corresponding to preclisturbance conditions (Eilers et al. 20011. Filers et al. concluclecI that the strong clomi- nance of the algal flora in Upper I(lamath Lake by heterocystous bluegreen algae is a byproduct of human presence. Historical observations of phy- toplankton, as summarized by Bortleson anti Fretwell (1993), are consis- tent with the paleolimnological conclusions. A brief overview of the chro- nology of observations on phytoplankton is as follows (conclensecI from Bortleson anti Fretwell 19931: In 1906, ice from Upper I(lamath Lake was
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 109 cleemecI unsuitable for consumption because of high organic matter anti green color; in 1913' summer phytoplankton samples showed diatoms dominant anti bluegreen algae accounting for only 12% of cells; in 1928' water samples showed abundant algae but no dominance by bluegreens; in 1933' Aphanizomenon was reported for the first time but not as a clomi- nant; in about 1939' Aphanizomenon was abundant but not dominant; in 1957' Aphanizomenon was 10 times more abundant than in 1939 but not yet overwhelmingly dominant; anti in the 1960s anti later, Aphanizomenon constituted almost a monoculture cluring most of the growing season. It wouicI be tempting to attribute the low ratio of nitrogen to phospho- rus reaching Upper I(lamath Lake to anthropogenic augmentation of phos- phorus supply. From the TMDL mass-balance analysis, however, it is clear that Upper I(lamath Lake probably hacI an even lower ratio of nitrogen to phosphorus in its preclisturbance state (BoycI et al. 2001) because it has an unusually rich geologic source of phosphorus. Thus, nutritional conditions in Upper I(lamath Lake favorable to nitrogen-fixing bluegreen algae such as Aphanizomenon are not new. The combination of high phosphorus con- centrations uncler background conditions anti the low ratio of nitrogen to phosphorus wouicI have created icleal nutritional conditions for the growth of bluegreen algae before human alteration of nutrient loacling, yet Apha- nizomenon blooms appear to be a byproduct of human activity. The conditions in Upper I(lamath Lake prior to anthropogenic change couicI have involvecI some factor that prevented the population growth of bluegreen algae, even though nutrient conditions favored nitrogen-fixing algae such as Aphanizomenon. It has been suggested, for example, that organic acids (clesignatecI here as limnohumic acids anti consisting mainly of humic anti fuivic acicis) present in wetiancI sediments are capable of chemically suppressing the growth of bluegreen algae (Eilers et al. 2001' Geiger 20011' although the phycological literature on limnohumic acids contains little indication of such effects ([ones 1998' but see also I(im anti Wetze! 19931. Drainage of wetiancis anti hycirologic alteration in the water- shecI of Upper I(lamath Lake probably has reclucecI the transfer of lim- nohumic acids to the lake. It is unknown, however, whether limnohumic acids or other substances clerivecI from wetiancis wouicI have been present in sufficiently high quantities to inhibit the growth of bluegreen algae uncler the original conditions of the lake or why this inhibition wouicI have been operating selectively on Aphanizomenon, given that other algae were abundant. Another possibility, apparently not proposed for Upper I(lamath Lake (although listecI by Geiger 20011' has to clo with light climate as influencecI by limnohumic acids. A record from 1854 (unpublishecI document of the state of Oregon, as given by Martin 1997) states suggestively that the water of Upper I(lamath Lake "hacI a ciark color, anti a clisagreeable taste occa-
110 FISHES IN THE KLAMATH RIVER BASIN sionecI apparently by clecayecI tule." Limnohumic acids, which can origi- nate in large quantities from some types of wetiancis (especially those of low alkalinity), absorb light strongly at short wavelengths (Thurman 1985) anti may substantially affect the light climate of phytoplankton ([ones 19981. For example, Morris et al. (1995) anti Williamson et al. (1996) showed that the clepth of 1% light cleclinecI from 12 m to 2 m as clissolvecI organic carbon (mostly limnohumic acicis) increased from 2 to 10 mg/L in a series of 65 lakes of varied latitucle. An increase in absorbance of such magnitude couicI substantially cut the amount of light reaching phytoplankton. Some diatoms are better aciaptecI to clear photosynthetically with low light avail- ability than most bluegreen algae (Reynoicis 1984), but the high light requirement of nitrogen fixation may be even more important. Among the bluegreens, the Nostocales (inclucling Aphanizomenon) have especially high light requirements (Weiciner et al. 2002, Havens et al. 19981. Thus, a change in light climate rather than a change in nutrient loacling or other chemical effects couicI have been responsible for the shift from diatoms to bluegreen algae. This is only one of several possibilities, however. Yet another possibility has to clo with biotic changes in Upper I(lamath Lake. Aphanizomenon grows relatively slowly anti so is especially vuiner- able to grazing, as shown by Howarth anti colleagues in marine environ- ments (Howarth et al. 1999, Marino et al. 2002, Chan 2001; see also Ganf 19831. It is conceivable that the intensity of grazing by zooplankton on algae has been alterecI by the introduction of fishes that are efficient zoo- planktivores. In the absence of so many efficient planktivores, zooplankton populations couicI have been much higher anti thus capable of working selectively against Aphanizomenon anti other nitrogen fixers. Contraclict- ing this hypothesis is the abundance of a large anti efficient zooplankton grazer, Daphnia (I(ann 19981. In fact I(ann (1998) proposes that Daphnia may promote Aphanizomenon by grazing preferentially on its competitors. Although it seems fairly certain that Aphanizomenon has come into dominance in Upper I(lamath Lake through human influences, the causal mechanisms of this unclesirable change in phytoplankton dominance re- main unclear. Seasonal Development of Algal Biomass Regular sampling of phytoplankton biomass at multiple stations in 1990-1998 has proviclecI a substantial amount of information on the time course anti interannual variability of biomass clevelopment of Aphanizo- menon in Upper I(lamath Lake (I(ann 1998, Welch anti Burke 20011. As is typical of phytoplankton populations, the phytoplankton of Upper I(la- math Lake, of which over 90°/O is Aphanizomenon at peak algal abun- ciance, shows a burst of growth in spring followecI by clecline. The progres-
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 111 sion of abundance is irregular, however, in that an initial period of rapicI growth may be interrupted or clelayecI, ancI a period of general clecline may leacI to renewecI growth (Figure 3-51. The growing season for phytoplankton in Upper I(lamath Lake begins generally in April. WoocI et al. (1996) proposed that water temperature wouicI show the most clirect control on the rate of increase in early spring, when other conditions for growth are favorable, ancI thus might be a goocI predictor of the elapsecI time between the beginning of the growing season ancI any particular biomass threshoicI that might be consiclerecI an algal bloom. This concept was investigated by I(ann (1998), who showed a statistically significant association between clegree days and elapsed time between the beginning of the growing season ancI the time coinciding with clevelopment of a specific biomass. According to I(ann's analysis, clays elapsecI between April 1 ancI a biomass threshoicI of 10 mg/L of wet mass couicI be preclictecI with fairly high confidence (r2 = 0.69) from clegree clays between April 1 ancI May 15. At the lower encI of the interannual growth rate spectrum, the threshoicI was reached after 150 clays; at the upper encI, after 170 clays. A relationship with lake volume in May was also tested ancI was suggestive but not statistically significant ancI it clepencis heavily on an outlying ciata point for 1992, without which there is no hint of a trencI relatecI to lake volume in May. A larger ciataset might show a weak but significant relationship on the basis that a lower mean clepth might leacI to faster warming, but interannual variation in weather introduces consicler- able variation not relatecI to lake depth. 300 - 250 - - ~° 200- _' ,= 150- o 100 - 50 - / - - ~ - 1992 · 1991 \ /' '\ /\ \ ~ W ~ /\ _ _ _ ~ O- 1 1 1 Jun Jul Aug Sep FIGURE 3-5 Change in chlorophyll a (lakewide averages, volume-weighted) over growing season for 2 consecutive years showing the potential interannual variabil- ity in development of chlorophyll maximums. Source: Redrawn from Welch and Burke 2001.
2 FISHES IN THE KLAMATH RIVER BASIN I(ann (1998) ancI Welch ancI Burke (2001) have placecI consiclerable emphasis on the relationship between water temperature ancI the first oc- currence of a threshoicI biomass of Aphanizomenon equal to 10 mg/L of wet mass in spring. The relationship is well supported by data, but it has virtually no application to the occurrence or timing of extreme water- quality conditions. The threshoicI of 10 mg/L of wet mass corresponds to chlorophyll a at about 20-30 ~g/L, which is only about 10-20% of the maximum abundance of Aphanizomenon as it reaches its annual peak. Although temperature influences growth in early spring, it later loses its influence because temperature stabilizes ancI the full clevelopment of the bloom to harmful proportions clepencis on other factors, as acknowlecigecI by Welch ancI Burke (20011. Thus, the relationship between temperature ancI growth rate of Aphanizomenon in early spring seems to be a cleacI encI with respect to anticipating the timing of the ultimate biomass maximums or their magnitude. Of clirect interest in connection with extremes of water-quality clegra- ciation cluring summer are the mean ancI maximum biomasses for sus- penclecI algae (primarily Aphanizomenon) that the lake shows in a given year. As shown in Figure 3-6' neither peak biomass nor mean biomass cluring the growing season has any empirical relationship with water level in Upper I(lamath Lake. Welch ancI Burke have moclelecI the abundance of Aphanizomenon on the basis of light availability with the assumption that nutrients are avail- able in sufficient quantities to produce very high biomass (which is clemon- strably correct). Light availability is affected by mean clepth. As a water column gets creeper, the mean light availability for incliviclual cells circulat- ing in the water column cleclines because cells spencI a higher proportion of time at greater depth, where light is less available. The mocleling lecI Welch ancI Burke to conclucle that maximum algal biomass of Aphanizomenon in Upper I(lamath Lake wouicI be quite sensitive to mean clepth of the lake (Welch ancI Burke 2001' p. 3-151. This conclusion is inconsistent, however, with measurements of algal biomass, which show no such relationship. Thus, the mocle! predictions are contraclictecI by fielcI observations, ancI the latter must be given greater weight. Mocleling of the type used by Welch ancI Burke is useful in directing research but often produces misleacling predictions because mocleling usu- ally requires various assumptions. In the case of mocleling relatecI to light, for example, the estimation of light exposure for cells must assume uniform distribution of biomass throughout the water column at all times. Because Aphanizomenon is capable of buoyancy regulation, it may have a nonuni- form vertical distribution cluring calm weather. Furthermore, although Upper I(lamath Lake is not stratified throughout the growing season, as deeper lakes are, it is stratified for substantial intervals during which the
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES Chlorophyll a vs. Lake Elevation 113 A_ 1 - 200- o 50 o 250 - 150 - 00 - 50 - O- 350 - ~_ 1 - ~ 250- Ct a97 ~ 92 90 91 ~ 96 ~ 94 ~ 95 e98 93 4138 4139 4140 4141 Water Level (feet) 4142 4143 300 - 200 - 150 - 00 - 50 - ·92 ·94 O- 4138 4139 ·97 ·91 ·90 ·96 ·98 ·93 ·95 4140 4141 4142 4143 Water Level (feet) FIGURE 3-6 Relationship of mean chlorophyll (above) and peak chlorophyll (be- low) to water level in Upper I(lamath Lake (median level for July and August). Source: Data from Welch and Burke 2001. . effective clepth from the viewpoint of phytoplankton in the surface layer is less than the actual clepth of the lake. Many other assumptions were neces- sary In modeling ancI could be a cause of divergence between moclel preclic- tions anti observations. At any rate, modeling cannot yet be used as a basis for predicting peak biomass of Aphanizomenon from water level in Upper I(lamath Lake. pH Algal biomass, which typically is measured as chlorophyll concentra- tion, is closely related to pH in Upper I(lamath Lake (I(ann 1998, Walker
4 FISHES IN THE KLAMATH RIVER BASIN 20011. This relationship is consistent with the expected rise in pH caused by high rates of photosynthesis in aquatic environments generally (Wetze! 20011. Thus, high algal abundance sustained by light and abundant nutri- ents is the proximate cause of high pH during the growing season in Upper I(lamath Lake. The photosynthetically induced high pH of Upper I(lamath Lake has been used in formulating a hypothesis related to the control of internal phosphorus loading in Upper I(lamath Lake (Boyd et al. 2001, Walker 20011. According to this hypothesis, designated here as the pH-internal loading hypothesis, internal loading occurs primarily under oxic conditions at the sediment-water interface and involves Resorption of phosphorus from ferric hydroxide complexes at the sediment-water interface through the replacement of phosphate with hydroxy! ions at high pH. Thus, high pH is proposed as a direct cause of the phosphorus enrichment of Upper I(lamath Lake through internal loading during the growing season. As explained above, however, the importance of other mechanisms of internal loading cannot be ruled out, especially because internal loading substantially in- creases phosphorus concentrations before the lake reaches its peaks of algal abundance that are the cause of peaks in pH. If high pH is the main cause of internal phosphorus loading, which in turn supports extremes of algal biomass in Upper I(lamath Lake, internal loading might be lower if the pH of the lake were lower. Thus, external loading might be connected causally to internal loading hv wav of nH: this . . . . . . . . . . . . ~ . J - - -- -- - ---- C7 - J - --A - r ~ - hypothes~s Is the basis ot some recommendations in the TMDL analysis of Upper I(lamath Lake (Boyd et al. 20011. The hypothesis is, however, still highly speculative. The pH of Upper I(lamath Lake also may be directly significant to fish, which can be damaged or killed by high pH. For example, Saiki et al. (1999) showed that a mean 24- to 96-h LC50 for the two listed sucker species in both larval and juvenile stages was 10.3-10.7. Sublethal effects would be expected below this threshold for exposures of 1 day or longer and have been demonstrated in juvenile shortnose suckers at a pH of near 9.5 (Falter and Cech 19911. Any means of suppressing extreme pH could benefit the suckers, although the degree of potential benefit is not clear. Because pH does not peak during episodes of mass mortality of suckers, however, it seems unlikely that pH contributes to mass mortality (Saiki et al. 19991. Also, because peaks of pH are transitory because of 24-h cycling of pH, impairment of fish by high pH in Upper I(lamath Lake is difficult to evaluate. As mentioned above, the immediate cause of the highest pH values in Upper I(lamath Lake is photosynthesis. Furthermore, the abundance of algae, as estimated from chlorophyll a, is strongly correlated with pH. Thus, suppression of algal abundance would lead to a suppression of pho-
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 115 tosynthesis, which in turn wouicI leacI to a suppression of pH anti, most important, elimination of the highest pH values. I(ann anti Smith (1999) suggested on the basis of a probabilistic analysis that a target chlorophyll a concentration of 100 ~g/L wouicI probably leacI to a effective suppression of high pH. The connection between pH anti water level in Upper I(lamath Lake has been of great interest because water level can be regulatecI to some degree. Welch ancI Burke (2001) argued on the basis of mocleling that higher water levels wouicI produce lower extremes of pH, which wouicI potentially benefit the suckers. Their projection of pH with mocleling was basecI on the presumption that chlorophyll a can be moclelecI in relation to water level. As mentioned above, however, observations of chlorophyll a in relation to water level are not as preclictecI by the moclel; there is no rela- tionship between means or extremes of chlorophyll a ancI lake level basecI on monitoring cluring the l990s. Thus, there is no reason to expect a relationship between pH ancI water level, given that pHis controllecI by algal abundance. In fact, the monitoring ciata show no relationship between pH anti water level (Figure 3-7; percentiles other than the one shown also fail to demonstrate a relationship between water level anti pH). Even though they predict more favorable pH at higher lake levels, Welch ancI Burke (2001) acknowlecige that there is no empirical relationship between pH ancI lake level as jucigecI by information collectecI cluring the l990s. The authors open the possibility of more complex relationships between lake level ancI pH. Any such relationship remains hypothetical, ancI the weight of current 9.9 - 9.8 - ~_ ·~ 9.7- ~4 in _' T Q 9.6 - 9.5 - 9.4 - 9.3 - 9.2 - ·91 . ·94 92 · 90 · 97 · 96 ·93 ·98 ·95 4138 4139 4140 4141 Water Level(feet) 4142 4143 FIGURE 3-7 Relationship between water level (median, July and August) and pH in Upper I(lamath Lake. The pH data are water-column maximum pH for 7 moni- toring sites distributed across Upper I(lamath Lake, shown as 75th percentile for all dates. Source: Data from Welch and Burke 2001.
116 FISHES IN THE KLAMATH RIVER BASIN evidence cloes not support the argument that higher lake levels will mitigate problems associated with high pH. One deficiency in the information on pHis lack of consideration of clie! cycling in pH (a small amount of information is given by Martin 19971. In highly productive waters such as those of Upper I(lamath Lake, pH changes extensively in a 24-h cycle; maximums occur in the afternoon hours, ancI minimums just before sunrise. The amplitucle of pH cycles commonly ex- ceecis lpH unit in fertile waters. Thus, evaluation of pH would be more complete if the pH cycle were taken into account. Overall, pHis regulatecI by algae, ancI if the abundance of algae couicI be reclucecI, the extremes of pH could be moderated. It is likely that the abundance of algae has been increased by human actions either clirectly or inclirectly, in which case pH uncler current conditions wouicI be expected to peak substantially above the pH that was present before changes in lancI use in the basin. Potentially unclesirable effects of high pH inclucle clirect ciam- age to fish ancI amplification of internal loacling, which is probably the largest source of phosphorus for Upper I(lamath Lake. It is not yet clear how much harm high pHis causing suckers (especially in contrast with clissolvecI oxygen, for example), nor is it clear that internal loacling of phosphorus, which can occur by a number of mechanisms, wouicI be strongly suppressed by reduction in pH. Ammonia Ammonia has been proposed as a toxicant that potentially affects the enciangerecI suckers of Upper I(lamath Lake. Although ammonia is a plant nutrient with no adverse effects on organisms at very low concentrations, it is toxic at high concentrations. Toxicity typically has been associated with the unionized component of ammonia in solution. Threshoicis of protection incorporated into various state regulations for warm-water aquatic life usually are in the vicinity of unionized ammonia Expressed as N) at 0.06 mg/L. Toxicity studies on the enciangerecI suckers showed, however, that they are more tolerant of ammonia than many other species of fish (union- izecI ammonia LC50 for 24-96 h, 0.5-1.3 mg/L; Saiki et al. 19991. Uncler oxic conditions, ammonia either is removed from the water column by autotrophs (which use it nutritionally) or is oxiclizecI by nitrify- ing bacteria that convert it to nitrate. Thus, in the absence of a strong point source of ammonia, it is typical to have low concentrations of ammonia in iniancI waters that are oxic. In the absence of oxygen, however, ammonia proclucecI by decomposition can accumulate, given that its conversion to nitrate or uptake by autotrophs cloes not occur uncler these conditions. Upper I(lamath Lake stabilizes in summer when wincI speecis are low, as explainecI below in connection with the discussion of oxygen. At such
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 117 times, ammonia accumulates in the lower water column as oxygen is cle- pletecI. Mixture of the ammonia into the entire water column couicI pro- cluce toxicity. Unionized ammonia seems a less likely cause of mass mortal- ity of fish in Upper I(lamath Lake than clissolvecI oxygen, however, because mass mortality continues after ammonia concentrations have cleclinecI (Per- kins et al. 2000b), ancI because the suckers show relatively high tolerance to ammonia. Dissolvecl Oxygen Low concentrations of clissolvecI oxygen coincide with mass mortality of large suckers in Upper I(lamath Lake. The suckers are relatively resistant to oxygen clepletion (LC50 1.1 to 2.2 mg/L; Saiki et al. 1999), but their tolerance limits are exceeclecI uncler some conditions in Upper I(lamath Lake (Perkins et al. 2000b). Unlike extreme pH or high ammonia concen- trations, low clissolvecI oxygen persists for clays while mortality occurs. Thus, low clissolvecI oxygen appears to be the clirect cause of mortality. Most lakes of micicIle latitucle are climictic; that is, they mix completely in spring ancI fall but stratify stably cluring summer ancI are covered with ice continuously or intermittently in winter. Lakes that are exceptionally shal- low in relation to their area, however, are polymictic; that is, they mix many times cluring the growing season. The shallowest lakes, which can mix convectively at night even in the absence of wincI, are clesignatecI con- tinuous polymictic lakes (Lewis 19831. Lakes that are too creep to be mixed entirely by free convection every night (about 2-3 m; MacIntyre ancI Melack 1984) but too shallow to sustain stratification throughout the growing season are intermediate in the sense that they clevelop ancI sustain stratifica- tion for intervals of calm weather, especially if there is no net heat loss, ancI mix completely when wincI strength increases or substantial heat is lost; they are callecI discontinuous polymictic lakes. Upper I(lamath Lake is a discontinuous polymictic lake, as shown by its episodes of stratification interrupted by extenclecI intervals of full mixing. The dynamics of water- column mixing ancI stratification in Upper I(lamath Lake are not well clocumentecI, however, because water-quality surveys have been separated by too much time to allow resolution of the alternation between mixing ancI stratification in the lake. A discontinuous polymictic lake shows alternation of the two very different conditions associated with mixed ancI stratified water columns. While the water column is unstratified, the lake shows minimal vertical differentiation in oxygen or other water-quality variables. When the lake stratifies, however, clepletion of oxygen begins in the lower part of the water column, where contact with atmospheric oxygen is lacking ancI there is not enough light for photosynthesis. Because Upper I(lamath Lake is
118 FISHES IN THE KLAMATH RIVER BASIN highly productive, its waters have high respiratory oxygen demand that quickly leads to the depletion of oxygen in the lower water column when- ever the lake is stratified (e.g., Welch and Burke 2001, Horne 20021. An empirical relationship has been shown between relative thermal resistance to mixing (RTR, an indicator of stability) and wind velocity during luly and August for Upper I(lamath Lake (Welch and Burke 20011. Thus, the expectation that intermittent stability is under the control of weather has been verified for Upper I(lamath Lake. Further work on the dynamics of mixing would probably be useful for understanding changes in water quality in the lake. Future work should be based on stability calcula- tions rather than RTR, however. Stability can be calculated from morpho- metric data on the lake, water level, and the vertical profile of density (Wetze! and Likens 20001. Stability depends on water depth and distribu- tion of density with depth, both of which are more irregular in Upper I(lamath Lake than would be ideal for use of RTR, which is a shortcut method of estimating stability that overlooks any changes in depth. The advantage of using true stability rather than RTR is that it may show more clearly relationships between stability and factors of interest to the analysis of mixing. The relationships already demonstrated are important, however. Loss of stability after a period of high stability in Upper I(lamath Lake is associated with low concentrations of dissolved oxygen and high concen- trations of ammonia throughout the water column and with depression of algal abundances. To some extent, those changes can be understood simply as a byproduct of mass redistribution in the water column. For example, ammonia is expected to accumulate in deep water during stratification because it is a byproduct of decomposition and accumulates where oxygen is scarce or absent; it is distributed throughout the water column by de- stratification. Likewise, water that is depleted of oxygen near the bottom of the lake, when mixed with the upper portions of the water column, causes a decrease in oxygen concentrations in the entire water column until photo- synthesis and reaeration processes at the surface combine to raise oxygen concentrations throughout the water column. Some of the events that follow destratification in Upper I(lamath Lake cannot be explained simply in terms of the redistribution of mass from a stratified water column. Concentrations of ammonia decline rather rapidly after destratification, as expected from the processes of Vitrification (oxida- tion of ammonia to nitrate by bacteria) and autotrophic assimilation (up- take by algae). Low concentrations of dissolved oxygen, however, persist for many days rather than being offset by reaeration and photosynthesis, as might be expected. Furthermore, algal populations show substantial and prolonged decline. The prolonged decrease in oxygen appears to be the main cause of mass mortality of the endangered suckers during transition from a stratified to a fully mixed water column accompanied by the most
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 119 severe decrease in clissolvecI oxygen (Perkins et al. 2000b). Therefore, it is important to unclerstancI why oxygen concentrations fail to recover. The likely proximate cause of the extenclecI decrease in oxygen concen- trations after clestratification is algal cleath. Stratification of the water col- umn appears to produce conditions that are harmful to the algae. The mechanism of harm is still indeterminate. It couicI involve, for example, cleath of the algae that are trapped in the lower portion of the water column when stratification occurs; these algae wouicI lack light anti might be ex- posecI to harmful chemical conditions as the lower water column becomes anoxic. Oxygen can be clepletecI quickly from the lower water column of Upper I(lamath Lake, partly because the oxygen clemancI of sediments is very high (WoocI 20011. One wouicI expect that the buoyancy control of Aphanizomenon would allow the algae to escape these problems, but per- haps not. Alternatively, the occurrence of calm weather, which probably accompanies the clevelopment of stratification, couicI leacI to extensive stranding of buoyant filaments of Aphanizomenon at the surface. This type of stranding is known to occur in clense populations of bluegreen algae. When population densities are high, the light climate is poor, ancI the vacuolate bluegreens often show buoyancy regulation as a means of main- taining the higher mean position in the water column, thus avoiding shacI- ing. When the water column is becalmecI, however, this type of buoyancy regulation, which requires a relatively long period of adjustment, takes the filaments to the surface where they are exposed to excessive amounts of radiation (especially ultraviolet) anti cleath results (Reynoicis 1971, Horne 20021. These are merely speculations on mechanisms, however; aciclitional research wouicI be required to demonstrate which ones apply. RegarcIless of the mechanism of algal cleath, it is clear that cleath of a substantial population of Aphanizomenon in Upper I(lamath Lake wouicI recluce the potential oxygen supply (by cutting off a portion of the photo- synthetic capability of the water column) anti wouicI simultaneously gener- ate a large amount of labile organic matter (as a result of the lysis of algal cells), which wouicI raise the oxygen clemancI of the water column through the respiratory activities of bacteria whose growth wouicI be stimulatecI by the presence of the organic matter (Figure 3-81. The extenclecI nature of oxygen clepletion suggests that it takes many clays for the excess organic matter to be consumed, for the photosynthetic capacity of the lake to be regenerated, or both. In the meantime, substantial harm can occur to en- ciangerecI suckers because oxygen concentrations remain low. An important practical question is whether the episodes of low clissolvecI oxygen through- out the water column are relatecI to water level. Empirical evidence incli- cates that no such association exists, as shown Figure 3-8 (other locations ancI percentiles also lack a pattern). If stabilization of the water column is ultimately a cianger to the fish through the induction of high algal mortality
120 FISHES IN THE KLAMATH RIVER BASIN 5- 4- 3- 2- 1 o- ·91 · .94 92 ·~96 90 ·97 ·93 ·98 e95 4138 4139 4140 Water Level (feet) 4142 4143 FIGURE 3-8 Relationship between water level (median, July and August) and dis- solved oxygen in the water column of Upper I(lamath Lake. Oxygen data are given as 75th percentile of minimums for all sampling dates in a given year at three sampling sites in the northern part of lake, which is considered to be especially important as habitat for large suckers. Source: Data from Welch and Burke 2001. followecI by loss of consiclerable oxygen from the entire water column, conditions leacling to high stability wouicI be least favorable to fish (Figure 3-91. Other factors being equal, creeper water columns are more stable, as acknowlecigecI by Welch anti Burke (20011; that is, one might expect higher water levels to produce greater mortality than lower water levels. However, given the complicating influence of numerous factors, inclucling weather, associations between clepth anti extremes of oxygen concentrations may be too variable to detect. At any rate, there is no evidence based on oxygen that favors higher water levels over lower water levels as jucigecI from information collectecI cluring the 1990s. Highly productive lakes may show clepletion of oxygen uncler ice clur- ing winter. Photosynthesis typically is weak in winter because of low irracli- ance anti the effects of ice cover anti snow on light transmission. Uncler winter conditions, even though respiration rates are suppressed by low temperature, clissolvecI oxygen can be completely clepletecI, anti this can leacI to the cleath of fish (winterkill). If all other factors are equal, a shal- lower lake is more likely to show winterkill than a creeper lake because a creeper lake has larger oxygen reserves anti less respiration per unit volume than a shallower lake. Other factors are also important, however, inclucling especially the duration of the period of ice cover anti the presence of refu- gia, such as springs or tributaries, that move oxygen to selectecI locations where fish may fincI oxygen.
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES Low Wind Speed 1 Water Column Stratifies loons of Oxygen in Lower Water Column 1 High Wind Speed 1 Water Column Mixes loom Dissolved Oxygen Throughout Lake Mass Mortality of Fish 121 Algal Mortality Reduced Photosynthetic Capacity, Higher Oxygen Demand FIGURE 3-9 Probable cause of low dissolved oxygen throughout the water col- umn of Upper I(lamath Lake during the growing season leading to mass mortality of fish.
22 FISHES IN THE KLAMATH RIVER BASIN Welch and Burke (2001) and USFWS (2002) have noted risk to the enciangerecI suckers through increased potential for winterkill when the lake is severely drawn clown, as it is in ciry anti critical ciry years. No winter mortality has been observed, however, even though the period of observa- tion inclucles 2 yr that have shown more severe cirawclown than any other years in the last 40 yr of record. Sparse ciata on oxygen uncler ice clo not indicate clepletion (USFWS 2002), but much more information is needed. Analogies that Welch anti Burke (2001) have shown with studies done elsewhere may be unreliable because of differences in the duration of ice cover anti other factors that make comparisons problematic. On a hypo- thetical basis, winter fish kill seems more likely when the lake is drawn clown than when it is not, but winter fish kill may not occur at all, in which case water level is not an issue within the operating ranges of the 1990s. Measurements of oxygen concentrations uncler ice cover wouicI shecI acicli- tional light on this issue. Overview of Water Quality in Upper Klamath Lake Poor water quality causes the mass mortality of the two enciangerecI sucker species of Upper Klamath Lake anti may also cause other, more subtle kincis of harm. The diagnosis anti remecliation of mechanisms leacI- ing to mass mortality or stress of fish require knowlecige of the causal connections between human activity anti poor water quality. Researchers working on both fish anti water quality in the upper Klamath basin have worked out some causal connections (Table 3-2) but in other cases have not yet succeeclecI in establishing cause-effect relationships. There are two criti- cal sets of causal connections relatecI to water quality: (1) connection of human activity with high phytoplankton biomass anti dominance of Apha- nizomenon in Upper Klamath Lake, anti (2) connection of high phytoplank- ton abundance with chemical conditions that couicI harm fish. High phytoplankton biomass has, according to the hypothesis (external phosphorus-loacling hypothesis) unclerlying the TMDL analysis of BoycI et al. (2001), occurred through augmentation of phosphorus loading of Upper Klamath Lake, mostly by nonpoint sources or through weakening of natu- ral interception processes that occur in wetiancis or riparian zones. There are, however, two major problems with this hypothesis (Figure 3-101. First, the anthropogenic augmentation of external loacling is sufficient to account for only about 40°/O of the total loacI; the main factor accounting for very high phosphorus concentrations at present is internal loacling rather than external loacling. The pH-internal loacling hypothesis proposes, however, a mechanism by which a 40°/O increase in external loacI couicI have proclucecI a much larger increase in internal loacI. According to this line of thinking, the increase in external loacI raised the maximum algal abundances enough
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES TABLE 3-2 Status of Various Hypotheses RelatecI to Water Quality of Upper I(lamath Lake 123 Hypothesis Algal abundance as measured by chlorophyll is positively related to total phosphorus in the water column Algal biomass as measured by chlorophyll is positively related to daytime pH Rate of early-spring development of biomass is positively related to rate of warming in the water column Rate of early-spring phytoplankton growth is inversely related to lake volume Status Mean growing-season average algal biomass is inversely related to lake depth Peak algal abundance is inversely related to lake depth A large amount of phosphorus in the water column during the growing season originates in sediments (internal loading) pH is the main control on internal loading of phosphorus Interception of anthropogenic phosphorus from the watershed will reduce algal abundance in the lake Lake water level is inversely related to pH Well supported Well supported Well supported Relationship weak or absent Inconsistent with field data Inconsistent with field data Well supported Not resolved yet Uncertain; unlikely Inconsistent with field data to increase the maximum pH cluring the growing season, which in turn greatly augmented internal loacTing by facilitating the Resorption of phos- phate from iron hycTroxicle floe on the secTiment surface. It is also possible, however, that internal loacTing, which can occur by several mechanisms, always has been large enough to saturate algal clemancT, as suggested by the steady nature of internal loacTing beginning early in the growing season, before pH reaches its peak. A second weakness in the external phosphorus- loacTing hypothesis is that it fails to explain why Aphanizomenon has be- come dominant. Nutritional conditions seem to have been favorable for Aphanizomenon (or other nitrogen fixers) before lancT-use changes in the watershed because of an inherently low nitrogen:phosphorus ratio in the lake. Because of the two major unresolvecT issues for the external phosphorus- loacTing hypothesis, alternate hypotheses are still worthy of consideration. One, shown in Figure 3-10, is basecT not on phosphorus enrichment, but rather on changes in the limnohumic acid content of the lake, which is con 1 ·1 1 1 1 · 1 · 1 · · ~ 1 ll~ely to nave been quite nlgn in waters emanating trom the extensive wetiancTs around Upper I(lamath Lake. The hypothesis proposes that the basic cause of change in water quality of the lake is reduction in the supply of limnohumic acids to the lake, with a consequent increase in transparency or possibly even a decrease in inhibitory effects (toxicity of the acids to
124 Change Caused by Phosphorus Land-Use Development Non-Point Sources of P Accumulate Natural Interception Processes Weaken External P Load to Lake Increases Phytoplankton Abundance Increases pH Increases FISHES IN THE KLAMATH RIVER BASIN Change Caused by Limnohumic Acids Land-Use Development Wetland Drainage to Lake is Reduced Concentrations of Limnohumic Acids in Lake Decline Transparency of Lake Increases Phytoplankton Shifts from Diatoms to Aphanizomenon Internal P Load Increases Phytoplankton Reaches Maximum Abundance Nitrogen Fixation by Aphanizomenon Relieves N Deficiency Phytoplankton Reaches Maximum Abundance FIGURE 3-10 Two contrasting hypotheses that may explain connections between human activity and high abundances of Phytoplankton in Upper I(lamath Lake. algae). ReleasecI from suppression by weak light availability or chemical inhibition, Aphanizomenon became more abundant in the lake. Unlike the diatoms that prececlecI it, Aphanizomenon was able to offset the low nitrogen:phosphorus ratio of the lake by nitrogen fixation, thus allowing
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 125 algal growth for the first time to take full advantage of the abundant phosphorus supplies anti produce the very high algal abundances that are now characteristic of the lake. The advantage of this hypothesis is that it accounts simultaneously for the change in community composition of phy- toplankton anti for an increase in biomass. The key factor causing major changes in the lake was, according to this hypothesis, drainage or hyciro- logic alteration of wetiancis, rather than increase in external phosphorus loacling. Figure 3-11 shows causes leacling from high algal abundance to water- quality conditions potentially harmful to fish. High abundance of phy- toplankton produces high pH, which can be clirectly harmful to fish. Al- though the connection of phytoplankton abundance to high pH is well verified, the amount of harm to fish that it causes is still a matter of speculation. A second factor is episodic stratification of the water column, which leacis to oxygen deficits in the bottom portion of the water column anti appears to cause algal mortality. Mixing causecI by wincly weather brings oxygen-poor water to the surface, along with ammonia. The impor- tance of ammonia in mass mortality is probably not great, but it couicI be harmful in more subtle ways to fish. Low oxygen that results from mixing probably is prolongecI by algal cleath anti probably is the main reason for mass mortality of fish. 1 Increased Internal P Load ?T High Abundance of Phytoplankton / I High pH Episodic Stratification ~ AlgalMortality it, ? ?/ Bottom Oxygen Depletion ? Ammonia in I/ - -- Water Column Mixing 1 ~ Low Oxygen in Harm to Fish ~ Water Column FIGURE 3 -11 Potential ( ? ) and demonstrated i,/ ) causal connections between high abundance of phytoplankton and harm to fish through poor water-quality . . cone Tons.
26 FISHES IN THE KLAMATH RIVER BASIN Potential for Improvement of Water Quality in Upper Klamath Lake Two proposals have been made for actions that would improve the water quality of Upper Klamath Lake. Both presume, with substantial sci- entific support, that an improvement of water quality in Upper Klamath Lake will require suppression of algal abundance. The first proposal, which could be implemented immediately, is for maintenance of water levels in Upper Klamath Lake exceeding levels that have been characteristic of the . . ~ . . . . . . . . recent historical past. The second proposal, which deals more with long- term improvement, is for reduction in the anthropogenic component of external phosphorus loading of Upper Klamath Lake. Higher water levels have been proposed in recent biological opinions for operation of Upper Klamath Lake (USFWS 2001, 20021. USFWS makes a number of kinds of arguments for higher water levels, while noting that empirical evidence of a connection between lake level and water quality is "weak" (USFWS 2001, p. 511. One of the arguments is that higher water levels will improve water quality in Upper Klamath Lake. As shown by the preceding review of available evidence, there is no scientific support for the proposition that higher water levels correspond to better water quality in Upper Klamath Lake. For example, mean and maximum abundances of algae, which are the driving force behind poor water quality, show no indication of a relationship with water level. USFWS acknowledges that no relationship has yet been demonstrated, but it argues that a complex, mul- tivariate relationship may exist but not yet be evident. For example, as noted by USFWS and others (Welch and Burke 2001), an effect of water level on water quality could be contingent on water-column stability, which in turn is under the influence of weather. Other multivariate relationships could be proposed that involve water level as one of several factors explain- ing the water-quality conditions in a given year. This line of argument leads to the conclusion that water-quality conditions may be explained in the future (after further study) by a suite of variables that include water level, but it also suggests that the influence of water level is too weak to be discerned without consideration of other variables. The potential usefulness to management of a complex mechanism in- volving water level as a covariate would be low even if it could be demon- strated. Furthermore, the mode of influence of water level as one of a suite of variables affecting water quality would not necessarily work in the direc- tion of water-quality improvement at higher water levels. For example, as indicated in the foregoing text, higher water level promotes water-column stability, which appears to be the principal means by which water-quality conditions for mass mortality develop in Upper Klamath Lake. All things considered, water level cannot now be managed with confidence for control of water quality.
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 127 The second proposal, which is for long-term improvement of water quality through reduction in external phosphorus loacling, has been favored by many ancI is the main recommendation relatecI to water quality of Upper I(lamath Lake in a recent TMDL analysis (BoycI et al. 20011. The proposal has three weaknesses: one relatecI to the feasibility of intercepting substan- tial loacI, a second relatecI to the internal-loacI effects of reducing external loacI, ancI a thircI involving the role of increased phosphorus loacling in sustaining large algal populations uncler current conditions. The TMDL proposal is for reduction of external phosphorus by about 40°/O. Because the current anthropogenic loacI is about 40°/O of the total, the proposal is to return the external phosphorus loacling of Upper I(lamath Lake to background conditions. Only about 1% of the anthropogenic loacI- ing is from point sources (wastewater treatment plants: BoycI et al. 20011. Interception of point-source loacis is technically feasible, but interception of nonpoint-source loacis, although approachable through best-management practices, is more problematic in that it wouicI require major changes in agricultural practice ancI other types of lancI use. Even a reduction of 20% wouicI be ambitious ancI potentially infeasible in view of the association between non-point sources ancI privately helcI lancis. Even a reduction of 40°/O in total external phosphorus loacling wouicI probably be ineffectual without suppression of internal phosphorus loacI- ing, given that internal phosphorus loacling is very large for Upper I(lamath Lake. The authors of the TMDL stucly have anticipated this problem. In- voking the pH-internal loacling hypothesis as clescribecI above, they antici- pate that a reduction in external loacling will result in lower extremes of pH, which in turn will recluce internal loacling, thus providing magnified benefits. This is a highly speculative proposition, however. Because soluble phosphorus is available in quantity even at the encI of the growing season, it appears that internal loacling is sufficient to supersaturate the neecis of algae for phosphorus. Furthermore, a pH reduction, if it clicI occur, might or might not be sufficient to shut off internal loacling relatecI to high pH. Finally, high pH is only one mechanism by which phosphorus is mobilizecI from sediments; other mechanisms wouicI remain as they are ancI couicI easily be sufficient to provide the internal loacis necessary to generate the high phytoplankton biomass observed in the lake. Thus, reduction of exter- nal load as proposed in the TMDL document has results that are quite uncertain. A thircI problem with the phosphorus-recluction strategy is that the high abundances of phytoplankton in Upper I(lamath Lake may have not become establishecI because of external phosphorus loacling, but rather because of other changes in the lake. A drastic decrease in mobilization of limnohumic acicI alteration of wetiancis ancI hycirology, for example, fits historical observations more satisfactorily than a phosphorus-basecI hy-
28 FISHES IN THE KLAMATH RIVER BASIN pothesis, given that Upper I(lamath Lake apparently has always had the very low nitrogen:phosphorus ratios that set the stage for dominance by a nitrogen fixer, such as Aphanizomenon. The ciata suggest that other fac- tors were holding back the nitrogen fixers ancI that human activity re- versecI or moclifiecI one of them, producing the current dominance of Aphanizomenon. Aphanizomenon, once established, could generate higher abundances than nonfixing algae because of its ability to offset nitrogen deficiency in the water. Thus, the key to improving water quality may be to suppress Aphanizomenon. Restoration of limnohumic acicis to the lake would be the most obvious way of restoring any beneficial effects that limnohumic acids might have hacI before lancI-use development of the upper I(lamath basin watershed. Restoration of wetlands is uncler way ancI could increase transport of limnohumic acids to the lake. Although justified in large part by an attempt to intercept nutrients, these programs could have beneficial effects on limnohumic acid supply. One discouraging aspect of restoring limnohumic acid transport to the lake, however, is that many of the wetland sediments surrounding the lake that would have been perhaps the richest source of limnohumic acids have clisappearecI through oxidation after clewatering. Furthermore, restoration of limnohumic acid supply would require not just restoration of wetlands but also extensive rerouting of water through wet- lancis, with attendant loss of water through evapotranspiration. Neverthe- less, this option is virtually unstucliecI ancI deserves more attention. It could be compatible with nutrient-removal strategies justified by improvements in water quality of streams. Current proposals for improvement of water quality in Upper I(lamath Lake, even if implemented fully, cannot be counted on to achieve the cle- sirecI improvements in water quality. Thus, it would be unjustified to rely heavily on future improvements in the water quality of Upper I(lamath Lake as a means of increasing the viability of the sucker populations. Oxygenation as a Management Tool The possibility that oxygenation of creep water could be used as a means of reducing mass mortality of enciangerecI suckers in Upper I(lamath Lake has been mentioned by USFWS in its biological opinion (2002; see also Martin 19971. An engineering stucly of the possibility is already avail- able (Home 20021. Because of the size of Upper I(lamath Lake ancI the speecI with which it can become anoxic toward the bottom of the water column cluring episodes of stratification, it is unlikely that oxygenation could be used in preventing low concentrations of clissolvecI oxygen from developing in the lower water column cluring stagnation or in restoring oxygen when the water column mixes at clepressecI oxygen concentrations.
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 129 Even so, it is conceivable that oxygenation couicI be used in such a way as to provide specific refuge zones to which the enciangerecI suckers wouicI be attracted when they experience stress clue to low clissolvecI oxygen. Of particular interest wouicI be the aclult suckers, which cluster in specific locations (USFWS 20021. It is cloubtful that the potential for aeration to recluce mass mortality of large suckers can be clevelopecI entirely from calculations ancI estimations. Pilot testing for proof of concept seems well justified for the near future. Potential success of this approach is uncertain, however, in that use of oxygenation specifically to create refugia in large lakes apparently cloes not appear in the literature on oxygenation. CLEAR LAKE Clear Lake was created in 1910 at the location of a smaller natural lake ancI associated marsh on the Lost River (Figure 3-121. One purpose for the creation of the reservoir was to allow storage of runoff for irrigation of lancis below the clam. An aciclitional purpose was to promote evaporative loss of water that otherwise wouicI flow to Tule Lake ancI Lower Klamath Lake, which were intenclecI for clewatering to allow agricultural clevelop- ment. In aciclition to high evaporative losses associated with its low mean depth, Clear Lake has extensive seepage losses. Clear Lake is cliviclecI into east ancI west lobes that are separated by a ricige; the clam is on the east lobe. Willow Creek, a tributary of Clear Lake, is critical to the sucker populations, which appear to rely primarily or even exclusively on this tributary for spawning. The lancis surrounding Clear Lake are not uncler intensive agricultural use. The area surrounding the reservoir consists primarily of Clear Lake National WilcIlife Refuge, ancI the watershed above the lake is largely encompassed by the Mocloc National Forest. Although Clear Lake wouicI store as much as 527,000 acre-ft at its maximum height, which corresponds to a lake area of 25,000 acres (USER 2000a), its average area has been close to 21,000 acres, which corresponds to a storage of about 167,000 acre-ft ancI a mean clepth of 8 ft (USER 2002a, USER 19941; it has never reached maximum storage. Average an- nual inflow is 117,000 acre-ft, which suggests a mean hyciraulic residence time of 1-2 yr Computed from input ancI volume). Clear Lake is similar to Upper Klamath Lake in being shallow in relation to its area. It cliffers from Upper Klamath Lake in its consiclerably longer hyciraulic residence time ancI its very low output of water relative to input. One other important feature, which has to clo with water management, is the high interannual ancI interseasonal variation in storage volume of Clear Lake, which corre- sponcis to great variations in area ancI mean clepth (USER 19941.
130 FISHES IN THE KLAMATH RIVER BASIN miles too Clear Lake `~ o IS ~ r ,~ 0 ~ i' ~J a. . ~ I ~~ 1 FIGURE 3-12 Map of Clear Lake. Clear Lake contains both shortnose ancI Lost River suckers (USFWS 20021. Both species show evidence of stability anti ecological success in Clear Lake, as inclicatecI by diverse age structure ancI high abundance (USFWS 2002, USER 1994; Chapter 51. Interannual variations in the wel- fare of the populations have been scrutinized, however, because of ques- tions relatecI to the maximum permissible cirawclown of the reservoir in a ciry year or in a succession of ciry years. Monitoring of water quality ancI condition of fish in 1991-1995 proviclecI a goocI opportunity to evaluate extreme cirawclown because the water level in 1992 cleclinecI to its lowest point since the drought of the 1930s.
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 131 Although water-quality records collectecI in 1991-1995 (USER 1994, Hicks 2002) are useful, the breadth of information that is available for Clear Lake is much narrower than that of Upper I(lamath Lake. Appar- ently, there has been no sampling for phytoplankton or for nutrients that wouicI allow comparisons with Upper I(lamath Lake. Observations suggest that Clear Lake has far lower population densities of phytoplankton than Upper I(lamath Lake; there is no evidence of massive blooms of bluegreen algae, for example. Aquatic macrophytic vegetation like that founcI in Up- per I(lamath Lake is virtually absent from Clear Lake because of its wicle range of water levels. The water column of Clear Lake typically has a turbid appearance suggestive of fine inorganic particulate material that is continually sus- penclecI by wincI-generatecI currents (USER 19941. The transparency of the lake has been measured only sporaclically. During 1992, when water levels were exceptionally low, Secchi depths ranged from 0.1 to 0.4 m, which inclicatecI extremely low penetration of irracliance in the lake (M. Buettner, USER, personal communication, 23 lanuary 20031. In more typical years, transparency is low but not nearly at the extreme of 1992 (for example, lune 1989, 0.4-1.5 m across 24 stations; M. Buettner, USER, personal communication, 23 lanuary 20031. Although Clear Lake is generally char- acterizecI as allowing less light penetration than Upper I(lamath Lake, the scanty ciata on light penetration that are available suggest that the transpar- encies may fall within the same range for the two lakes (for example, see I(ann 1998 for ciata on Upper I(lamath Lake). Because transparency may be relatecI to the welfare of sucker larvae through predation, which may be more pronounced in transparent waters, further stucly of this subject seems warranted. In 1991-1995, recording sensors were used for measuring temperature, specific conductance, pH, ancI clissolvecI oxygen; vertical profiles also were taken for these variables. Although interpretation of the records is compli- catecI by occasional malfunction of the sensors, which is characteristic of this type of ciata collection, the overall results are useful. The temperature record indicates that the lake is unstratified; if it cloes stratify, it cloes so only sporaclically over the deepest water (near the clam). The pH varies seasonally but cloes not reach the extremes observed in Upper I(lamath Lake, presumably because high rates of algal photosynthesis, the driving force behind extreme pH in Upper I(lamath Lake, are not characteristic of Clear Lake (USER 19941. The oxygen data indicate that the lake does not show episodes of strong oxygen clepletion like those in Upper I(lamath Lake. One incident of oxygen concentration as low as 1 mg/L near the clam apparently was associated with drainage of the east lobe of the reservoir cluring 1992 as the lake was drawn clown to the extremes of that year. Monitoring uncler ice showed concentrations of oxygen near saturation,
32 FISHES IN THE KLAMATH RIVER BASIN even cluring an interval of especially long ice cover cluring 1992, a year of very low water level (USER 19941. Mass mortality of suckers in Clear Lake is unknown. Loss of fish occurs through the clam but cloes not appear to be seriously decreasing the populations. The populations were stucliecI for signs of stress cluring the ciry year of 1992. Although mortality was not observed, there were several indicators of stress, inclucling higher rates of parasitism ancI poor body condition. These indicators clisappearecI quickly as water levels climbecI in 1993 at the encI of the drought (USER 19941. The indications of stress associated with water levels of 1992 have servecI as a basis of proposed threshoicis of cirawclown in Clear Lake (USFWS 20021. The potential of Clear Lake to provide information about Upper I(la- math Lake has not been well exploitecI. The agencies have invoked Clear Lake for comparative purposes in several instances, but the background information on the reservoir is not sufficiently broacI ancI cloes not extend over sufficient intervals of time to allow goocI comparisons. Comparative population ancI environmental studies in the two lakes couicI open up new possibilities for diagnosing mechanisms that are aciversely affecting encian- gerecI suckers in Upper I(lamath Lake. GERBER RESERVOIR Gerber Reservoir was establishecI on Miller Creek, a tributary of the Lost River, in 1925 (Figure 1-11. The lake can store as much as 94,000 acre- ft of water but often is substantially drawn clown ancI shows consiclerable interannual ancI intraannual variability in volume, mean depth, ancI area (USER 1994, 2002b). Nevertheless, characteristic depths of Gerber Reser- voir probably are substantially greater than those of Upper I(lamath Lake or Clear Lake. Statistics are not reaclily available, but the sampling record (USER 2002b) suggests that in most years a substantial area of the lake wouicI have water creeper than 15 ft. Extreme cirawclown occurred in 1992, when the lake was reclucecI to less than 1 % of its maximum volume (USER 2002b). Even uncler those conditions, the water near the Gerber Reservoir clam was 15 ft creep. As might be expected, given that it is smaller ancI creeper than Clear Lake or Upper I(lamath Lake, Gerber Reservoir shows a tendency toward stability of thermal stratification, as inclicatecI by loss of oxygen near the bottom cluring summer. Stability may be interrupted by mixing, ancI en- trainment of water through the outlet may leacI to a replacement of bottom waters, which couicI produce changes (oxygenation, warming) similar to those expected as a consequence of mixing. Little information is available on the water quality of Gerber Reservoir. The lake appears to have less inorganic turbidity than Clear Lake, presum-
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 133 ably because it is creeper anti smaller. Aphanizomenon flos-aq?vae probably is present anti apparently creates blooms but not to the same clegree of Upper Klamath Lake (USFWS 20021. Aphanizomenon probably fares bet- ter in this reservoir than in Clear Lake because the latter has more sus- penclecI inorganic turbidity, which shacles the water column. Information on temperature, specific conductance, pH, anti clissolvecI oxygen was collected for the first half of the 1990s by automated monitor- ing anti occasional vertical profiles (USER 2002b), as was the case for Clear Lake. The pH reaches higher extremes than in Clear Lake but is less ex- treme than in Upper Klamath Lake. This probably reflects a gradient of algal photosynthesis across the three lakes. Dissolved oxygen in Gerber Reservoir is substantially clepletecI in creep water both in summer anti in winter, but without any obvious effect on fish. No episodes of mass mortal- ity of the shortnose sucker, which occupies Gerber Reservoir, have been reported. During 1992, when cirawclown of the lake was severe, the lake was aerated (USFWS 20021; sampling inclicatecI that the fish hacI reachecI suboptimal body condition cluring the drought. Uncler other circumstances, the population appears to have been stable in that it has shown no inclica- tion of stress, has preserved a cliversifiecI age structure, anti has been abun- ciant. For reasons primarily having to clo with water quality, the low water levels of 1992 serve as a guideline for setting thresholds to protect the fish from stress. LOWER KLAMATH LAKE Lower Klamath Lake has been reclucecI to a marshy remnant by clewa- tering. It has occasional connection to the Klamath River through which it appears to receive some recruitment of young suckers, but there is no adult population. Water quality apparently has not been stucliecI in any system- atic way. Development of an aclult population is unlikely unless the clepth of water can be increased, which would involve incursion of the boundaries of the lake onto lancis that are used for agriculture. If the lake were cleep- enecI, water quality might be adequate for support of suckers. TULE LAKE Tule Lake historically was very large and capable of supporting, in conjunction with the Lost River, large populations of the shortnose anti Lost River suckers (Chapter 51. It has been reclucecI to remnants as a means of allowing agricultural use of the surrounding lancis. Water reaches Tule Lake from Upper Klamath Lake or from the Lost River drainage via irri- gatecI lancis or from Clear Lake or Gerber Reservoir. Water is removed
34 FISHES IN THE KLAMATH RIVER BASIN from Tule Lake (now appropriately called Tule Lake Sumps) by Pump Station D (USER 2000a). There are two operational sumps at Tule Lake now: 1A ancI 1B. In the recent past, Sump 1B has been much less likely to hold aclult suckers than Sump 1A; it is shallower ancI has shown a higher rate of sedimentation than Sump 1A. It also appears to have worse water quality than Sump 1A. Sump 1B is being manipulated by USFWS for increase of marshland in the Tule Lake basin. Some water-quality information is available on Tule Lake through monitoring cluring the 1990s (USER 2001a) ancI fish have been sampled (Chapters 5 ancI 61. It appears that the sucker population consists of a few huncirecI inclivicluals, including shortnose ancI Lost River suckers, ancI that these favor specific portions of Sump 1A (the "doughnut hole" or a loca- tion in the northwest corner) that presumably provide more favorable con- clitions than the surrounding area. Monitoring of Sump 1A has not shown any incidence of strongly clecreasecI oxygen concentrations or extremely high pH, as would be the case in Upper I(lamath Lake (USER 2001a). These adverse conditions may occur in Sump 1B, however. The fish of Tule Lake, although not very abundant, appear to be in excellent body concli- tion, ancI this suggests they are not experiencing stress. Suckers migrate from Tule Lake Sumps; migration terminates on the Lost River at the Anderson Rose Diversion Dam (USFWS 2002, Appendix C), in the vicinity of which spawning is known to occur. Water-quality conditions there for spawning appear to be acceptable (USER 2001a). Lar- vae are proclucecI but apparently are not passing into the subaclult ancI adult stages. From the water-quality perspective, it appears that the Tule Lake popu- lation is potentially closer to survival conditions than the Upper I(lamath Lake population. An unresolved mystery, however, is the fate of larvae. It is not clear whether water quality prevents the larvae from maturing, or if other factors are responsible for their loss. Sedimentation threatens the apparently goocI conditions for aclults in Sump 1A. Without aggressive management, the favorable portion of Sump 1A may become progressively less favorable in the future. RESERVOIRS OF THE MAIN STEM There are five main-stem reservoirs (Table 3-11; because Copco 2 is extremely small, it generally cloes not receive inclepenclent consideration. The composite residence time of water in the main-stem reservoirs, which extend about 64 mi from Link River Dam to Iron Gate Dam, averages about 1 mot At moderately low flow (for example, 1,000 cfs), hydraulic residence time is close to 2 ma; ancI at moderately high flow (such as 6,000
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 135 cfs), it is close to 10 clays. Thus, some of the processes that wouicI make these lakes distinctive from each other ancI from their source waters are not expressed because of the relatively rapicI movement of water through the system. The main source of water for the main-stem reservoirs is Upper I(la- math Lake, but it is not the only source. Agricultural returns ancI drainage water enter the system upstream of the I(eno Dam (Figure 1-2) by way of the I(lamath Strait Drain (about 400 cfs, summer) anti the Lost River Channel (about 200-1,500 cfs, fall ancI winter). In aciclition, coicI springs provide about 225 cfs all year at a point just below the t.C. Boyle Dam; ancI two tributaries, Spencer Creek ancI Shove! Creek, provide 30-300 cfs to t.C. Boyle Reservoir anti Copco Reservoir. Fall Creek anti lenny Creek provide 60-600 cfs to Iron Gate Reservoir. During the wet months, sources other than the Link River, which brings water from Upper I(lamath Lake, provide about one-thircI of the total flow reaching Iron Gate Dam; in midsummer, these sources may account for up to 50°/O of the total water reaching Iron Gate Dam (PacifiCorp 2000, Figure 2-71. Thus, source waters of diverse quality influence the quality of water in the reservoirs. The waters of Upper I(lamath Lake often bring large amounts of algal biomass to the upper encI of the system, along with large amounts of soluble ancI total phosphorus. When Upper I(lamath Lake is experiencing senescence of its algal population, the entering waters also may have low concentrations of clissolvecI oxygen anti an abundance of decomposing organic matter. Irrigation tailwater anti other drainage wouicI carry abundant nutrients ancI couicI carry organic matter but wouicI probably lack substantial amounts of algae. Spring waters ancI tributary waters wouicI be the coolest ancI cleanest of the water sources. The reservoirs cliffer physically in several ways that are likely to influ- ence water quality. I(eno Reservoir ancI t.C. Boyle Reservoir are shallow ancI have the lowest hyciraulic residence times. Physically, they resemble rivers more than lakes. In each, the water is poolecI at the lower encI ancI may run swiftly at the upper encI, thus potentially benefiting from reaeration (gas exchange). The two lower reservoirs are much creeper anti have hy- ciraulic residence times that are short on an absolute scale but much longer than those of the two upper reservoirs. None of the reservoirs has very creep withcirawal. Thus, for the two reservoirs that support stable stratification (Copco anti Iron Gate), with- cirawals reflect the characteristics mostly of epilimnetic (surface) water' although their withcirawal cone may extend a short distance into the hypo- limnetic (creep) zone at times (seas 20001. For example, the temperature of water leaving Iron Gate Dam cluring midsummer, when the hypolimnion has a temperature of about 6°C, reaches 22-23°C (PacifiCorp 2000, Figure 4-5; Deas 2000, Figure 6.5) because the powerhouse withcirawal is at about
136 FISHES IN THE KLAMATH RIVER BASIN 12 m clepth when the lake is full. For Copco, withcirawal is at about 6 m when the lake is full. Thus, coicI hypolimnetic water of the two deepest reservoirs tencis to be much more static hyciraulically than the upper water column cluring the stratification season, as wouicI be the case in a natural lake of similar depth; the main withcirawal occurs by way of the epilimnion. A small withcirawal (about 50 cfs) for the Iron Gate Hatchery cloes occur from the hypolimnion at Iron Gate Reservoir, however. The quality of water in the reservoirs ancI leaving the reservoir system has been stucliecI many times by numerous parties ciating back to the 1970s. PacifiCorp has sponsored a number of studies in conjunction with its FecI- eral Energy Regulatory Commission (FERC) licensing ancI other regulatory requirements, ancI USER has sponsored studies of water quality because of its oversight responsibilities for the I(lamath Project. The city of I(lamath Falls has also stucliecI water quality, particularly in the upper encI of the system, ancI the Oregon Department of Environmental Quality has stucliecI ancI analyzecI water quality from the viewpoint of fisheries. Other informa- tion is available from the U.S. Geological Survey, the U.S. Army Corps of Engineers, ancI the North Coast Regional Water Quality Control Board. In its consultation document on FERC relicensing, PacifiCorp (2000) provides an overview of the monitoring programs. Monitoring to ciate provides useful information but shows several clefi- ciencies. Most of the monitoring has been limitecI to water-quality variables that can be measured with meters (temperature, pH, specific conductance, ancI clissolvecI oxygen). There is much less information on nutrients, total phytoplankton abundance, phytoplankton composition, total organic mat- ter, ancI other important variables. Thus, interpretations are necessarily limitecI in scope. Also, there have been few efforts to synthesize ancI inter- pret the data, most of which exist merely as archives. Hanna ancI Campbell (2000) have moclelecI temperature ancI clissolvecI oxygen in the reservoirs. The temperature mocleling is useful for planning, but the oxygen mocleling fails to incorporate primary production, which couicI be important. Deas (2000) has clone extensive mocleling for Iron Gate Reservoir that is espe- cially useful for temperature ancI clissolvecI oxygen. A full, system-level unclerstancling of the reservoirs is not yet available, however. During the coo! months (October or November through May), all the lakes are isothermal ancI appear to mix with sufficient vigor to remain almost uniform chemically (see, for example, Figure 3-131. During the warm season, there may be substantial differences in temperature ancI wa- ter quality with depth. I(eno Reservoir ancI t.C. Boyle Reservoir are not creep enough to sustain thermal stratification cluring summer. They may stabilize briefly, however, in which case oxygen may be clepletecI from creep water (see Figure 3-14), but such clepletions probably are interrupted by episodes of mixing. Copco ancI Iron Gate reservoirs, in contrast, stratify
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES Copco, January O - 5 - 10- 15- a' 20- 25 - 30 - o 4 8- 12- 16- 20- 24 - 28 - 32 - 36 - 0 2 4 6 8 10 4 h X DO, mg/L Temperature,°C o 2 4 Iron Gate, January 6 8 : 12 14 ·! 10 12 Temperature,°C ~ 137 DO, mg/L ! FIGURE 3-13 Water temperature and dissolved oxygen (DO) in Copco and Iron Gate Reservoirs, January 2000. Source: Data from USBR 2003. stably on a seasonal basis. Thus, the water near the bottom of these two reservoirs can be classifiecI as hypolimnetic ancI has a much lower tempera- ture than that of the upper water column. As expected, oxygen is clepletecI in the hypolimnion of both lakes. Although the rate of oxygen clepletion varies across years (seas 2000), both reservoirs apparently have an anoxic hypolimnion for as much as 4 or 5 ma beginning in the last half of summer. Periodic episodes of severe oxygen clepletion may occur in the upper two reservoirs. One such event appears to have occurred in 2001, when the entire water column of I(eno Reservoir became hypoxic or anoxic (Figure 3-151. It is not known how often such an event occurs. Because no mass mortality of fish in the reservoirs have been recorclecI, it is possible that the
138 to- . au ~ 3- 2 - 4 - 5 - 2 3 - 4 - 5 - 6 - 7 - 8 - O - ~ 5 ~ 10 :5 15 20 30 5 10 15 20 25 30 35 40 FISHES IN THE KLAMATH RIVER BASIN Keno, July 0 5 DO, me : 10 15 20 25 O Temperature,°C 1 0 5 J.C. Boyle, July 10 15 20 25 ~ T DO, mg/L Copco, July 0 5 10 15 Temperature,°C ~ ~ ' f - DO, mg/L ~ Temperature,°C - - > - Iron Gate, July 0 5 10 15 20 ' f Temperature,°C 20 25 25 30 3 ' DO, ma: - FIGURE 3-14 Water temperature and dissolved oxygen (DO) in all main-stem reservoirs, July 2000. Source: Data from USER 2003.
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES '_ ~ au ~ 4 139 1 1 1 1 , . / Temperature ~ I 234 236 238 240 242 244 River Miles l l 246 248 1 1 250 252 234 236 238 240 242 River Miles Dissolved Oxygen 1 244 246 248 250 252 FIGURE 3-15 Longitudinal transect data on I(eno Reservoir (Lake Ewaunal, 13- 14 August 2001. Isolines indicate temperature at 1°C intervals (top panel, increas- ing from 18°C) and dissolved oxygen at intervals of 1 mg/L (bottom panel, increas- ing from 1 mg/L). Darker tones indicate lower temperature or lower dissolved oxygen; the darkest zone on the bottom panel indicates concentrations of dissolved oxygen below 1 mg/L (i.e., without or almost without oxygen). Source: Data from USBR. fish uncler these circumstances seek inflowing water of high oxygen concen- tration to sustain them until the episode dissipates. Although the reservoirs receive abundant supplies of algae from Upper I(lamath Lake, they clo not appear to sustain such high rates of algal growth as Upper I(lamath Lake, as incTicatecT by comparisons of pH. Upper I(la- math Lake shows extremes of pH extending above 10, but such extremes are not characteristic of the reservoirs. For example, monitoring of Copco, Iron Gate, ancT t.C. Boyle reservoirs in 1996-1998 byPacifiCorp showed the highest pH to be about 10.0, ancT even this was quite unusual (PacifiCorp 2000, Figure 4-101. More recent ciata are similar in this respect (Table 3-31. Concentrations of phosphorus (means) tencT to be about the same in the main-stem reservoirs as in Upper I(lamath Lake. There is ample phos-
140 FISHES IN THE KLAMATH RIVER BASIN TABLE 3-3 Summary of Grab-Sample Data for Surface Waters in the Main-Stem Reservoir System, 2001a Concentration, ~g/L Chlorophyll a Location pH NH4+-N NO3--N SRP Total pb 2001 2000 I(eno (1 m) Mean 7.50 1,080 80 160 390 62 - Max 8.82 1,220 90 240 730 - - J.C. Boyle (1 m) Mean 7.31 190 1,120 250 260 50 5 Max 7.86 260 1,760 290 450 - 20 Copco (1 m) Mean 7.91 90 620 150 280 5 10 Max 8.90 130 880 220 560 - 31 Iron Gate (1 m) Mean 8.28 260 370 160 180 5 11 Max 9.45 260 630 280 410 - 46 Below Iron Mean 7.87 80 980 170 190 4 - Gate (0.s m) Max 8.68 90 1,710 210 360 - - aN = 4 in most cases (monthly, June-September); N = 1 for chlorophyll in 2001 (July); additional chlorophyll data for 2000 (N = 6) are shown for three of the reservoirs. Chloro- phyll shown at concentrations below about 20 ~g/L is only a rough approximation because of . . . . . . Imitations on ana yt~ca sensitivity. bTotal P less than SRP (soluble reactive P) for some dates. Sources: USER 2003; PacifiCorp, unpublished data, 2001. phorus in available form for stimulation of phytoplankton growth, but it is not clear whether a net accumulation or a net loss of phytoplankton biomass occurs in the reservoirs because information on phytoplankton biomass shows some internal inconsistencies. Observations from the fielcI suggest that substantial blooms of Aphanizomenon occur in both Copco ancI Iron Gate reservoirs (USFWS 20021. This wouicI not be surprising, given the strong seecling of these reservoirs with Aphanizomenon from Upper I(lamath Lake ancI the presence of large amounts of nutrients. Although the residence times for the two large reservoirs are not great enough to allow the establishment of large populations of algae starting from a very small inoculum, a large inoculum couicI clouble several times over the duration of residence in the two reservoirs ancI thus generate a bloom. Alternatively, a bloom couicI simply be transferred from Upper I(lamath Lake. The clifficulty with the observations, however, is that they are not confirmed by monitoring ciata in 2000 ancI 2001. For both of those years, analysis of chlorophyll a showed abundances of algae ranging from low to high but not extreme in the sense of Upper I(lamath Lake (Table 3-31. There are several possible explanations. FielcI reports might be biasecI by appearance of some algae at the surface while unclerlying populations are not extraorclinarily high. There couicI be something wrong
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 141 with the chlorophyll analyses, or perhaps large blooms occur very selclom. These matters are unresolvecI. One special concern with respect to coho salmon anti other salmonicis in the I(lamath River main stem is the condition of water as it leaves Iron Gate Dam. Oxygen concentrations below Iron Gate Dam are seasonally below saturation but generally exceed 75°/O of saturation (seas 2000), have a temperature that reflects surface waters in the lakes, anti have lower concentrations of nutrients anti algae than would be typical ot upper ~ia- math Lake. Because there is some question about the consistency of ciata on algae, however, no firm conclusions are possible about the export of Apha- nizomenon to the main stem via Iron Gate Dam. . · . . . ~ . It appears that the upper two reservoirs have the poorest water quality, as jucigecI from concentrations of nutrients anti clissolvecI oxygen. The two lower reservoirs, although they clevelop anoxia in creep waters cluring sum- mer, maintain better water quality than the upper two reservoirs in their surface waters. The major question of Aphanizomenon blooms in the sys- tem seems unresolvecI because of internal inconsistencies in the data. In general, the water-quality environment seems to be comparable with or slightly better than that of Upper I(lamath Lake in the two upper reservoirs, which may have very low clissolvecI oxygen but clo not seem to have the pH extremes that Upper I(lamath Lake cloes. The two lower reservoirs appear to have better quality overall than the two upper reservoirs, although their creep waters are essentially uninhabitable for fish cluring the summer months because of their lack of oxygen. More synthetic work on the reservoirs is neeclecI. CONCLUSIONS 1. Water-quality conditions in Upper I(lamath Lake are harmful to the enciangerecI suckers. Mass mortality of large fish is caused by episodes of low clissolvecI oxygen throughout the water column. Very high pH anti high concentrations of ammonia, although more transitory than the episodes of low clissolvecI oxygen, may be important agents of stress that affect the health anti body condition of the fish. 2. Poor water quality in Upper I(lamath Lake is caused by very high abundances of phytoplankton, which is clominatecI by Aphanizomenon flos- aquae, a nitrogen fixer. Suppression of the abundance of Aphanizomenon is essential to the improvement of water quality. 3. Very high abundance of Aphanizomenon in Upper I(lamath Lake is almost certainly caused by human activities, but mechanisms are not clear. One hypothesis is that increased algal abundance has occurred because of an increase in phosphorus loacling in the lake. An alternative hypothesis, which is more consistent with the shift in dominance to Aphanizomenon
42 FISHES IN THE KLAMATH RIVER BASIN anti with the naturally rich nutrient supply of phosphorus to the lake, is that loss of wetiancis anti hycirologic alterations have greatly reclucecI the supply of limnohumic acids to the lake. According to this untested hypoth- esis, loss of limnohumic acids greatly increased the transparency anti may also have reclucecI inhibitory effects caused by the limnohumic acids. These changes allowecI Aphanizomenon to replace diatoms as dominants in the phytoplankton. Total phytoplankton abundance then increased because of the ability of Aphanizomenon to offset nitrogen clepletion by nitrogen fixa- tion, which diatoms couicI not clot 4. Substantial evidence indicates that adverse water-quality conditions are not relatecI to water level. Further stucly extenclecI over many years may ultimately show multivariate relationships that involve water level. Control of water quality in Upper I(lamath Lake by management of water level, within the range of lake levels observed cluring the l990s, has no scientific basis at present. 5. Suppression of algal abundance in Upper I(lamath Lake couicI in- volve drastic reduction in external phosphorus loacI or reintroduction of a substantial limnohumic acicI supply, clepencling on the mechanism by which Aphanizomenon has become dominant. Both of these remeclial actions, if undertaken on a scale sufficient to suppress the abundance of Aphanizo- menon, couicI be achieved only over a period of many years anti couicI prove to be entirely infeasible. 6. Because remecliation of water quality in the near term seems very unlikely, recovery plans for the enciangerecI suckers in the near term must take into account the potential for continued mass mortality of suckers. 7. Use of compressed air or oxygen to offset oxygen clepletion near the bottom of Upper I(lamath Lake has been suggested as a means of moclerat- ing mass mortality of aclult suckers. Such a technique cannot be expected to offset oxygen clepletion throughout the lake, but it has some potential to provide refuge zones. The enciangerecI suckers may be particularly well suited for this type of treatment because the large suckers, which are sus- ceptible to mass mortality, congregate in known locations. 8. Researchers have proviclecI a great clear of useful information relatecI to water quality of Upper I(lamath Lake. Needs for aciclitional information inclucle studies clesignecI to show the mechanism for Aphanizomenon cleath; physical studies, inclucling continuous monitoring of temperature anti oxy- gen anti associated analytical anti mocleling work, that demonstrate more clefinitively the mechanisms that promote alternation of stratification anti clestratification cluring the growing season for the lake; studies of the effects of limnohumic acids on Aphanizomenon anti of the former limnohumic acid supply to the lake; studies of clie! pH cycling in the lake; anti studies of water quality uncler ice.
CURRENT STATUS OF AQ UATIC ECOSYSTEMS: LAKES 143 9. Clear Lake anti Gerber Reservoir lack extremes of pH, oxygen cleple- tion, anti algal blooms that occur in Upper I(lamath Lake. Better water quality, in combination with other favorable factors given in more cletail in Chapters 5 anti 6, appear to explain steady recruitment, diverse age struc- ture, anti goocI body condition of these populations. Deterioration of body condition of the listecI suckers at a time of extreme cirawclown provide a rationale for the lower allowable threshoicis of water level in these lakes. The lakes anti their tributary spawning areas have exceptional value for protection against loss of the two enciangerecI sucker species. Aciclitional studies of limnological variables (ancI those of fish populations) have spe- cial value for use in comparison with water quality anti population charac- teristics of suckers in Upper I(lamath Lake. 10. Tule Lake, which supports suckers in goocI body condition but cloes not show evidence of successful recruitment, may have water quality that wouicI allow recovery of this subpopulation if problems involving spawning habitat anti larval survival were resolvecI. Lower I(lamath Lake, which now lacks aclult suckers, might well support a sucker population if water levels were raised. 11. Of the four major main-stem reservoirs, I(eno ancI t.C. Boyle ap- pear to have the poorest water quality because they are shallow, have the strongest influence from Upper I(lamath Lake, ancI show the least benefit of clilution by waters entering from other sources. Copco anti Iron Gate reser- voirs have better water quality but clevelop anoxia in hypolimnetic waters cluring summer. Water releasecI to the I(lamath main stem from Iron Gate Dam often is below 100% saturation with oxygen but selclom less than 75°/0 of saturation ancI may be excessively warm in summer for salmonicis because it is drawn mostly from the epilimnion. Algal populations in Iron Gate Reservoir appear not to reach the extremes that are typical of Upper I(lamath Lake.