Scientific Investigations Report 2008–5201
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2008–5201
Through their studies in Upper Klamath Lake, fish biologists have formed the hypothesis that poor water quality may not be the sole factor causing fish kills in Upper Klamath Lake, it is a compounding factor in which fish are stressed to the point that diseases may gain a foothold in the population. Perkins and others (2000), suggested that even though infection may be the clinical cause of death, poor water quality should be considered as the primary cause of death. With this hypothesis in mind, different potentially stressful water quality scenarios were investigated to gain a broader understanding of the different ways water quality conditions may cause harm to fish in the study area.
To identify potentially harmful low dissolved oxygen and high pH conditions in the study area, the hours when dissolved oxygen concentration was less than 4 mg/L, temperature was greater than 28 ºC, or pH was greater than 9.7 were enumerated at each site. These values were based on high stress thresholds for Upper Klamath Lake suckers (Loftus, 2001). Instances meeting a higher tier of potentially harmful conditions (dissolved oxygen concentration less than 2 mg/L and pH greater than 10; no higher tier for temperature) also were investigated. These higher tiers for dissolved oxygen concentration and pH were based on 96-hour median lethal values for Lost River juvenile suckers (Saiki and others, 1999). Together, these two tiers of dissolved oxygen, pH, and temperature conditions indicated when conditions may have been severe enough to cause stress and when conditions were potentially lethal.
Time series graphs (“bubble plots”) of these conditions are shown in figure 26 for representative near-bottom monitors (positioned 1 m from the lake bottom) in deep and shallow areas, as well as surface monitors (positioned above deep monitors at 1 m below the surface), in the open-waters of Upper Klamath Lake. Graphs for the two Agency Lake sites are shown in figure 27. Each instance when conditions exceeded the criterion appears as a “bubble” on the graph, centered on the midpoint of the time span of the event. The duration of the event is represented by the size of the bubble. Temperatures exceeded 28ºC only at the MDT-Upper site for 2 hours or less on 4 individual days in late July, so a plot of potentially harmful temperature conditions for open-water sites is not shown.
Occurrences of potentially harmful pH and dissolved oxygen conditions were markedly different between deep sites and shallow sites in Upper Klamath Lake. Potentially harmful pH conditions were more likely to occur in shallower water, whereas potentially harmful low dissolved oxygen concentrations were more likely to occur in deeper water (fig. 26). Monitors at the shallow sites commonly recorded periods of pH greater than 9.7 over multiple days before the AFA bloom decline in late July and for much of the season thereafter. The process of photosynthesis elevated the pH at these sites by consuming carbon dioxide (CO2) and reducing the concentration of carbonic acid in the water (Wetzel, 2001). However, few potentially harmful dissolved oxygen concentrations were recorded at these sites at any time during the field season, 2006. Conditions at the relatively shallow Agency Lake sites followed this trend, with more frequent occurrences of pH greater than 10 (fig. 27). However, potentially harmful pH conditions were not present in Agency Lake after the bloom recovered in late July, corresponding to the lack of a vigorous bloom later in the season.
The near-bottom monitors at the deep sites recorded instances of low dissolved oxygen concentrations exceeding the less than 2 and 4 mg/L criteria of potentially harmful conditions over multiple days during the period of seasonally low lakewide median dissolved oxygen. Occurrences of potentially harmful low dissolved oxygen continued at these sites even after lakewide average dissolved oxygen had recovered from seasonal lows in the rest of the lake (fig. 26). These post-recovery episodes of potentially harmful low dissolved oxygen most frequently met the less than 4 mg/L criterion and were more numerous at sites EPT and EBB. A plausible explanation, as demonstrated with the light and dark bottle incubation experiments, is that there was a net consumption of oxygen at the deep sites in the trench, resulting in greater oxygen depletion and increased frequency of potentially harmful low dissolved oxygen at sites EPT and EBB than shallow sites (Wood and others, 2006). Potentially harmful pH conditions were rare during the field season at near-bottom and near-surface monitors at the deep sites.
The differences in patterns of potentially harmful pH and dissolved oxygen conditions between deep and shallow water are a result of the depth of light penetration. At deep sites, the near-bottom monitor is constantly in the aphotic zone, where photosynthetic activity, and therefore photosynthetically elevated pH, does not occur. The lack of photosynthesis in deep water results in decreased dissolved oxygen production and greater potential for metabolic processes to cause a net consumption of oxygen.
The near-surface monitors at the deep sites recorded more occurrences of potentially harmful dissolved oxygen concentration, and fewer occurrences of potentially harmful pH conditions than the monitors at the shallow sites, even though, at 1 m from the water surface, these monitors were usually closer to the water surface than the monitors at the shallow sites (fig. 26). The water column, even at the deepest sites, typically mixes daily (Hoilman and others, 2008). This mixing has the effect of moderating pH conditions and decreasing dissolved oxygen concentrations near the surface.
In addition to low dissolved oxygen concentrations during episodes of bloom decline, high concentrations of un-ionized ammonia were present in trench areas (fig. 28). When high ammonia concentrations occur coincident with high pH and high temperatures, a significant fraction of the concentration is present in the un-ionized form, which is particularly toxic to aquatic life. The percentage of un-ionized ammonia relative to ammonia concentrations throughout the sampling period ranged from 0 to 87 percent. The dependence on pH is stronger than the dependence on temperature; at 22ºC and a pH of 9, for example, 31 percent of the ammonia will be in un-ionized form, but at a temperature of 22ºC and a pH of 9.5, the fraction jumps to 59 percent (U.S. Environmental Protection Agency, 1998). Mean un-ionized ammonia concentrations lethal to suckers (Saiki and others [1999]) ranged between 480 and 1,290 µg/L. Concentrations of un-ionized ammonia measured in the Upper Klamath Lake did not reach lethal concentrations. Ammonia was released into the water column concurrent with AFA bloom decline and the peaks of un-ionized ammonia, which occurred simultaneously with high pH values, were offset in time from ammonia peaks because as ammonia was released into the water column, pH values were decreasing.
Although un-ionized ammonia concentrations did not reach lethal levels, the combination of low dissolved oxygen and high un-ionized ammonia would cause stress to fish, making them more susceptible to disease. Lethal concentrations of un-ionized ammonia might occur for short periods or in localized, unmonitored areas more susceptible to extremes in temperature and pH, such as shallow water.
Unlike other shallow sites in Upper Klamath Lake, instances of potentially harmful dissolved oxygen concentrations were measured at site HDB (fig. 26). However, water quality dynamics at HDB are known to be somewhat disconnected from the rest of the lake (Hoilman and others, 2008). In early October at site HDB, dissolved oxygen remained less than 2 mg/L for more than 5 consecutive days. During this time, several dead fish (mostly fathead minnows and chubs) were observed in Howard Bay. This event, which was not observed anywhere else in Upper Klamath Lake, demonstrates that the water quality dynamics in Howard Bay are somewhat disconnected from the rest of the lake. The dissolved oxygen plots for HDB (fig. 16), together with the results of dissolved oxygen production and consumption experiments there, give insight into metabolic processes behind the LDOEs that occurred at HDB. The smaller LDOE in late July (fig. 26) coincided with net oxygen consumption measured in light bottles throughout the water column (fig. 16), indicating that oxygen consuming processes in the water column played a part in this LDOE. Water column oxygen demand was not as strong during the larger LDOE at site HDB in early October, when dark bottle oxygen rates decreased only slightly and light bottle oxygen rates (at 0.5 m depth) showed net oxygen production. This indicates that SOD processes, which can influence water column dissolved oxygen concentration, but operate separately from water column processes, probably caused the early October LDOE at site HDB. SOD can be significant in Upper Klamath Lake, especially when large mats of AFA sink to the bottom and decompose. This process was thought to be the cause of an exceptionally large SOD measured in Ball Bay in 1999 (Wood, 2001), and it is likely that the same process occurred at site HDB in the current study.
When the combined pressure of dissolved gases in the water column is greater than the combined local barometric and hydrostatic pressures, it is possible for gas bubbles to form in the tissues of fish. This condition, called gas bubble disease, can cause direct mortality to fish through bubbles in gill tissue or chambers of the heart that restrict oxygen uptake and (or) blood flow in the fish. Sublethal effects caused by the disease, such as lesions and blindness, can make fish more susceptible to mortality (Weitkamp and Katz, 1980). Gas supersaturation typically is associated with physical processes that drive atmospheric gases into solution, such as water crashing into a deep pool after flowing over the spillway of a dam. However, supersaturation caused solely by oxygen production through photosynthesis also has been known to cause gas bubble disease (Weitkamp and Katz, 1980). Gas bubble disease is probably not a major cause of mortality in fishes of Upper Klamath Lake, but it may be another source of stress that could act synergistically with other factors to cause harm (Scott Foote, U.S. Fish and Wildlife Service, oral commun., 2006).
Photosynthetic production created supersaturated dissolved oxygen conditions during much of 2006 in Upper Klamath Lake (fig. 21). However, supersaturation of dissolved oxygen alone may not always provide the necessary conditions to produce the disease. Waters must be supersaturated with respect to total atmospheric gases to produce gas bubble disease. Additionally, gases must be supersaturated enough to come out of solution to form bubbles. The hydrostatic pressure of the depth of the water column helps keep dissolved gases in solution. Equations given by Colt (1984) take these factors into account and were adapted to estimate when total dissolved gases may come out of solution given the percent saturation of dissolved oxygen at the depth where measurements were made (fig. 29). These instances were defined by times when the total dissolved gas pressure (Pd) was equal to or greater than the atmospheric and hydrostatic pressures combined (Pa+h). Nitrogen gas was assumed to be fully saturated throughout the water column for all calculations.
Sites with full pool depths deeper than 4 m (table 1) did not have occurrences of potential gas bubble formation, so data for these sites are not shown. The graphs in figure 29 are ordered according to increasing site depth. The shallowest sites, all with full-pool depths of 2.8 m or less, had the most potential for bubble formation, reflecting the lessened hydrostatic pressure relative to the deeper sites, where the greater pressure kept gases dissolved. At SET-U, a site in deep water where the monitor was placed 1 m from the water surface, the nearly daily mixing with water lower in oxygen from large aphotic zones below made instances of possible bubble formation less frequent and of shorter duration than at shallow sites where conditions were monitored at a similar depth (such as sites FBS and NBI). When lakewide average dissolved oxygen concentrations were high (fig. 23) the shallow areas of Upper Klamath Lake were likely to have conditions that could potentially lead to bubble formation, often lasting for several days at a time. These occurrences were most frequent in August and September, when potential for dissolved oxygen production was still high and the depth of the water column was decreasing.
As previously mentioned, potentially harmful low dissolved oxygen conditions occurred in deeper waters of Upper Klamath Lake during August and September. This coincidence with increased instances of possible gas bubble formation in shallow water could further limit available refuge for fish from potentially harmful water quality conditions: shallow waters tend to have more dissolved oxygen, but in concentrations that could potentially lead to gas bubble disease. Gas bubble disease is unlikely to occur in deep waters, but these areas also are more likely to have potentially harmful low dissolved oxygen concentrations.
Because of the direct correlation between dissolved oxygen concentration and pH, potentially harmful dissolved oxygen and pH conditions are unlikely to occur simultaneously. Previous research has shown that, in shallow waters of Upper Klamath and Agency lakes, daily minimum dissolved oxygen concentrations tend to occur early in the morning, and maximum pH conditions are most likely to occur during the late afternoon. In areas of deeper water, the timing of these daily extremes becomes less definite due to diel stratification of the water column (Wood and others, 2006; Hoilman and others, 2008). Data from 2006 are consistent with these findings. Although potentially harmful pH and dissolved oxygen did not occur simultaneously at any time during 2006, instances of these events taking place closely in time (the same day) did occur at some locations in Upper Klamath and Agency lakes (table 6). The occurrence of these potentially harmful dissolved oxygen and pH conditions in close succession may present additional harm to fish beyond what is experienced by either of these conditions alone (Power, 1997). However, because only 25 events were recorded during May through October, these conditions probably were not a significant threat to fish during 2006.
Because photosynthetic production of dissolved oxygen also increases pH, gas bubble formation and potentially harmful pH could occur simultaneously. Water quality data indicated that these potentially harmful conditions may occur often and be widespread in Upper Klamath and Agency lakes (table 7). These conditions were not estimated to occur simultaneously at only the two deepest sites outside of the trench areas (MDN-Lower and EHB). Because almost 90 percent of the lake is less than 4 m deep, most of the lake is represented by the sites shown in table 7. The instances of possible gas bubble formation were inferred only from prevailing conditions, however. Without direct monitoring of total dissolved gases, the potential harm posed to fish by gas bubble formation and its interrelations with other aspects of water quality in these lakes cannot be known conclusively.
Graphs of low dissolved oxygen and high pH conditions potentially harmful to fish in nearshore areas (fig. 30) appear similar to those for open-water sites away from the trench (fig. 26). Few instances of dissolved oxygen concentrations met either criterion of potential harm (less than 2 or 4 mg/L) or relatively numerous instances of potentially harmful pH conditions. Occurrences of possible gas bubble formation were correspondingly numerous (fig. 31), owing to the high potential for dissolved oxygen production (and associated elevated pH) in the shallow waters of Upper Klamath Lake.
Among the nearshore sites, no clear relation of occurrences of potentially harmful pH or possible bubble formation with site depth was noted. Because these sites are similar in depth (table 1), local factors seem more likely to influence the occurrence of these conditions. For instance, occurrences of potentially harmful pH and possible gas bubble formation often were long lasting at site SSR. This likely is due to the action of prevailing northerly winds concentrating AFA along the southern shore of the lake, increasing the potential for photosynthetic oxygen production and its associated increase in pH. Temperatures exceeded 28°C at the three shallowest nearshore sites (fig. 32). Occurrences of potentially harmful temperature conditions were few and of short duration relative to potentially harmful dissolved oxygen and pH conditions, and were less likely to cause harm to fish.
Mann-Whitney rank-sum tests were conducted to provide a quantitative measure of the similarity (or difference) between nearshore and open-water occurrences of potentially harmful water quality conditions. This nonparametric test for a difference in the distribution of two variables was used because it does not require the data sets to be normally distributed. The duration of individual events at all sites, listed separately for nearshore and open-water areas, was compared, as was the total number of hours meeting the criterion at each site for nearshore and open-water areas. A quantitative comparison of potentially harmful temperature conditions was not obtained because temperature greater than 28°C in the open-water areas were few, and occurred at only one site, making the data unsuitable for the Mann-Whitney rank-sum test. As with the comparison of daily median water quality conditions between the two areas, data from the deep monitors at sites EBB, EPT, MDT, and SET were removed from the open-water group in the analysis. All p-values mentioned are two sided.
The comparisons indicated no statistically significant (p < 0.05) difference in the duration of events of potentially harmful dissolved oxygen conditions or in the total number of hours these conditions existed at sites in nearshore and open-water areas. The duration of events of pH conditions greater than 10 and the duration of events of possible gas bubble formation were determined to be significantly greater in nearshore areas (p = 0.006 and p = 0.002, respectively). No differences, however, were noted in the total hours potentially harmful pH conditions or possible gas bubble formation existed at sites between the two groups. The differences observed can be explained by the buildup of AFA at the nearshore SSR site discussed earlier, causing high pH conditions and potential gas bubble formation for considerable lengths of time. When data from this outlier were removed, no significant differences were found in any of the comparisons. The results of these comparisons indicate that potentially harmful dissolved oxygen and pH conditions occur with similar patterns near shore and in open-waters, but that areas where wind tends to cause algae to accumulate are especially prone to possible gas bubble formation and high pH.
Note that the similarities between Upper Klamath Lake nearshore areas and open-water areas away from the trench were observed during a single season when severely low dissolved oxygen was largely confined to the deep trench areas of the lake. In other years, low dissolved oxygen has been more widespread in shallower waters of the lake (Wood and others, 2006). The similarities observed between open-water and nearshore areas in 2006 might not be observed during a year when low dissolved oxygen is more widespread.
U.S. Department of the Interior | U.S. Geological Survey
URL: http://
pubsdata.usgs.gov
/pubs/sir/2008/5201/section8.html
Page Contact Information: Contact USGS
Page Last Modified: Monday, 07-Mar-2016 12:19:55 EST
