Scientific Investigations Report 2006–5209
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2006–5209
To this point the presentation of the data has focused on three aspects of water quality that contribute to chronic stress on endangered suckers when they exceed or fall below certain limits—dissolved oxygen, pH, and temperature. An analysis of all possible causes of mortality during fish die-offs in the mid-1990s concluded, however, that mortality was ultimately the result of hypoxia, even considering that substantial mortality occurred after hypoxic conditions had abated, because hypoxia had compromised the immune systems of the fish and left them vulnerable to infection (Perkins and others, 2000). Observations of the fish die-off in 2003 were consistent with this conclusion, in that they indicated that the role of hypoxia was to force fish into crowded conditions in the vicinity of Pelican Bay, where many eventually succumbed to disease (B.J. Adams, U.S. Geological Survey, oral commun., 2005). For that reason, the following discussion focuses on the probable causes for the occurrence of a low dissolved oxygen event (LDOE) in the study area, which encompasses the area of the lake that is the preferred habitat of adult suckers.
For the purposes of discussion it is useful to define an LDOE quantitatively. There are several ways to do this; four were chosen, and the time periods that emerge in each year as a result of those four definitions are compiled in table 6. A few dates consistently emerge in each year when all but the most stringent definition is used—August 1–2 in 2002, July 23 to August 2 in 2003, October 1–8 in 2003, and August 13 in 2004. When an LDOE is defined as 50 percent or more of site-hours having a dissolved oxygen concentration less than 4 mg/L, then only the period July 23–30 in 2003 meets the definition. By combining the various datasets collected during this study, including weekly chlorophyll a samples, meteorological data, water-quality profiles from the profiling buoys, and, in 2003 and 2004, the ADCP current data, it has been possible to evaluate three of the factors that contribute to an LDOE—thermally stable conditions, a bloom decline, and the wind-driven circulation.
The profiling buoys collected observations of dissolved oxygen concentration over the water column between 1 m off the bottom and 1 m from the surface (figs. 23–25). Even though the degree of stratification in Upper Klamath Lake is moderate compared with that of deeper lakes, it has implications for water quality. Stratification isolates a near-bottom layer of water, allowing oxygen demanding processes to deplete this layer of oxygen and potentially creating unsuitable conditions for endangered suckers. These oxygen depleted conditions do not usually extend throughout the water column, however, because stratification also tends to create favorable conditions for AFA in the upper layer. The AFA cells can regulate their buoyancy and thereby maintain a position close to the surface; consequently, the upper water column tends to become supersaturated with respect to dissolved oxygen during periods of stratification.
Because stratification rarely persists from day to day, the relatively oxygen-rich upper water column and the relatively oxygen depleted lower water column are usually mixed together on a daily basis. That is why the maximum concentration of the day at the profiling buoys was usually greater than the value of concern of 4 mg/L (figs. 23–25) even when the minimum of the day was less than this value, and why the concentration in the mixed water column (indicated by the daily minimum difference between the upper and lower monitors) usually was greater than 4 mg/L even on days when the minimum of the day was less than this value. The daily maximum range in dissolved oxygen concentration between the upper and lower water column generally was such that the upper water column was characterized by concentrations tolerable to fish (greater than 4 mg/L) even when the lower water column was characterized by concentrations severely stressful to fish or even lethal. These observations from the profiling buoys suggest that periods when habitat is limited by water quality to only part of the water column and to less than a full day are common. Observations of fish behavior during 2002–04 suggest that the endangered suckers are, for the most part, able to survive through these periods, although the effects of the chronic stress that accompanies these conditions cannot be determined from tracking studies.
The periods of increased surface-to-bottom difference in dissolved oxygen were related to periods of enhanced stratification, as indicated by an increased surface-to-bottom difference in temperature (fig. 26). These two quantities were highly correlated on a daily basis in all 3 years (table 7). This did not, however, translate into a high correlation between enhanced stratification, as measured by either increased surface-to-bottom difference in temperature or increased surface-to-bottom difference in dissolved oxygen concentration, and the minimum daily dissolved oxygen (fig. 26). These quantities were only significantly and negatively correlated in 2002 (table 7). Thus, the occurrence of the lowest minimum dissolved oxygen concentrations is not necessarily associated with a high degree of water column stratification, although some LDOE dates appear to coincide with a stratified water column—July 10, July 23, July 30, August 8–13 in 2002, September 25–29 in 2003, and August 11–12 and August 19 in 2004.
The relation between wind speed, water column stratification, and daily minimum dissolved oxygen was not straightforward. On a daily basis, wind speed was significantly and negatively correlated with water column stratification, as measured by the surface-to-bottom difference in temperature, in 2002 and 2004, but not in 2003. The correlation between wind speed and daily surface-to-bottom difference in dissolved oxygen concentration was significant in all 3 years (fig. 26 and table 7). Kann and Welch (2005) noted an inverse correlation between wind speed and the surface-to-bottom difference in dissolved oxygen when profile data collected from 1990 to 2000 were averaged over the entire July–August period; in this study the relation appears to hold on a daily basis as well. The correlation between daily wind speed and daily minimum dissolved oxygen, however, was insignificant in 2002, weakly positive in 2003, and weakly negative in 2004 (table 7). So, although wind speed clearly influences the degree of stratification in the water column, it has less influence over when the lowest dissolved oxygen concentrations occur. Some of the LDOE dates were associated with low wind speeds; many were not.
Significantly, the most severe LDOE of the 3 study years, which occurred in July 2003, was not associated with stratification of the water column. Concentrations during the last week in July were about 3 mg/L or less during the entire day and throughout the water column at site UKL07 (fig. 24). Data from the other monitoring sites showed that low concentrations were widespread during a period of several days (fig. 19), meaning that habitat in the entire southwestern part of the study area was unsuitable throughout the water column and throughout the day for several days. The distinction between the severity of the LDOE of July 2003 and times during 2002–04 when water quality was perhaps chronically stressful but not lethal appears to have been important to the endangered suckers in the lake. The suckers responded to the 2003 event by leaving the area (B.J. Adams, U.S. Geological Survey, oral commun., 2004), whereas they generally remained in place in other years during times when the lowest dissolved oxygen concentrations were less than 4 mg/L near the bottom but moderated slightly both closer to the surface and at intervals during the day.
Sites UKL07 and UKL08 were in similar water depths (table 1), and the 2002 data from these sites show that the periods when very low dissolved oxygen was observed and the periods of maximum vertical variability in dissolved oxygen over the water column were similar as well (fig. 23). Data collected by profiling buoys in 2003 from sites UKL07 and UKL13, sites that were in similar water depths, show a corresponding similarity in periods when low dissolved oxygen was observed in the lower water column (fig. 24). In 2004, however, one of the profiling buoys was moved to site UKL16, a site in the middle of the trench on the western side of the lake (fig. 1). This site was much deeper than any of the sites where profiling buoys were located previously, and the data show that very low near-bottom dissolved oxygen developed more readily at this site than at any of the shallower sites. There was very little evidence of low near-bottom dissolved oxygen at site UKL07 in 2004, but there were several periods when low concentrations developed in the trench at UKL16 (July 12–20, July 30–August 3, and August 13–19, fig. 25). The causes and implications of the development of these conditions in the trench area are discussed below in the context of the wind-driven circulation.
In a hypereutrophic water body like Upper Klamath Lake, a shutdown or sharp decline in the rate of photosynthesis can cause a rapid decrease in dissolved oxygen concentration. This decrease occurs because the rate of oxygen demanding processes, which can be high in lakes with large amounts of labile organic matter in the water column and the surficial sediments, continue unabated while the rate of oxygen production ceases or slows markedly. Senescing cells also add to the oxygen demand in the water column. During the 3 years of this study, bloom declines seem to have contributed to several LDOEs.
In 2002, the bloom decline, as estimated from weekly chlorophyll a, occurred between about July 19 and August 1, so it preceded and then overlapped the LDOE between July 31 and August 3 (table 6). By one definition, this LDOE lasted until August 13, but chlorophyll a concentrations largely recovered, although not to prebloom concentrations, during the first 2 weeks in August, so it seems unlikely that the bloom decline could be the primary cause of the undersaturated dissolved oxygen conditions that lasted until August 13. The timing of the LDOE was coincident with elevated particulates in the air caused by the smoke from forest fires in 2002, most notably the Biscuit Fire, which began in the Siskiyou National Forest west of Upper Klamath Lake on July 13 and ultimately burned nearly 500,000 acres. The first burst of 10- and 2.5-µm particulates entered the area on July 25, and the particulates remained elevated in concentration through August (U.S. Environmental Protection Agency, 2006). Solar radiation data collected during the same period at the Bureau of Reclamation Agrimet site on Agency Lake does not show a reduction in incident radiation associated with the elevated concentration of particulates in the air, however, so the evidence linking the LDOE between August 2 and August 13, 2002, directly to reduced photosynthetic activity caused by smoke in the air from the Biscuit fire is not conclusive.
In the 3 years of this study, the period when the connection between the bloom decline and the occurrence of a LDOE was most direct was between July 21 and August 2, 2003. This severe LDOE, which culminated in a fish die-off, was coincident with a sharp decrease in chlorophyll a concentrations, indicating that a major decline in the bloom had occurred (fig. 9); the dates estimated to bracket the decline were July 19 and August 1, 2003. Even so, the bloom decline does not appear to be the entire explanation for the LDOE because dissolved oxygen began to trend downward early in July 2003, well before the marked decrease in the weekly chlorophyll a concentrations. This apparent discrepancy suggests that although the bloom decline may have exacerbated and perhaps accelerated the downward trend in dissolved oxygen, a larger set of circumstances also contributed to the event.
In 2004, an LDOE occurred between August 11 and August 13, which coincides with the estimates of the dates of a bloom decline in that year (fig. 9). This bloom decline was less definitive than in the other 2 years, and the dates were difficult to estimate, but chlorophyll a concentrations on August 11 were very low at sites UKL07 and UKL02 (5.8 µg/L at UKL07 and 1 µg/L at UKL02, respectively). It seems likely, therefore, that a bloom decline, perhaps more localized than in the other 2 years, accompanied by a reduction in photosynthesis, is part of the explanation for the LDOE during mid-August in 2004.
Under prevailing wind conditions, the wind-driven circulation moves water northward in the lake through the trench along the western shoreline (fig. 8). The immediate source of water to the study area, therefore, is the cross-section between Bare Island and Eagle Ridge (A-B in fig. 8). Direct observations of the water leaving the trench were made as part of this study starting in 2003, when sites UKL13 and UKL14 were established. Upon leaving the trench, water circulating in the prevailing clockwise direction first passes through site UKL14, then UKL13. Site UKL07, centrally located in the study area, can be considered “downstream” of site UKL13. The traveltime between UKL14 and UKL07 is short—about 1.4 days, on the basis of a current velocity of 5 cm/s as representative of the currents measured at ADCP3 (fig. 1).
From these observations, it is apparent that the water that enters the area from the trench often is depleted of dissolved oxygen compared to other parts of the lake (figs. 27 and 28). Under some circumstances there is very little change in the dissolved oxygen concentration of the water as it progresses from sites UKL14 to UKL07, as during the entire month of July 2003 (fig. 27). At other times, however, nonconservative processes alter the dissolved oxygen concentration substantially between the time that the water leaves the trench and when it passes through site UKL07. This was the situation through much of July and August 2004, when concentrations at site UKL07 were consistently higher than those at sites UKL14 and UKL13 by several milligrams per liter (fig. 28). In mid-August in particular, a pulse of water low in dissolved oxygen was observed at sites UKL14 and UKL13, but by the time this pulse of water passed through site UKL07, the dissolved oxygen concentration had increased substantially. Whether nonconservative processes can alter the characteristics of the water leaving the trench substantially as it passes through the study area depends on the actual traveltime, which varies with wind speed and direction, and the intrinsic oxygen-producing and oxygen-consuming capacities of the water. Neither of these can be estimated with the datasets presented in this report, but a better understanding of both in the future may be key to understanding the development of a severe LDOE in the northern part of the lake.
The wind-driven circulation must be considered when evaluating the causes of long-term trends (on weekly to monthly time scales) in water quality, such as the steady decrease in dissolved oxygen in July 2003 that culminated in a severe LDOE. Because the wind-driven circulation moves much of the volume of the lake around a closed loop in about 10 days, trends in water quality on time scales that long or longer must be understood as the cumulative effect of nonconservative processes as water moves in its flow path around the lake. A steady decrease in dissolved oxygen, such as was seen in July 2003, therefore, indicates that, when integrated over the entire flow path, the oxygen demands exceeded oxygen production through photosynthesis. Given that is the case, further investigation of oxygen consumption and production rates around the lake, particularly in areas not monitored in this study, will be helpful. It also is important to understand how changes in wind speed and direction are manifested as changes in the wind-driven circulation, and how changes in circulation around the lake interact with the spatially varying oxygen consumption and production rates. Changes in the circulation pattern could be part of the explanation for why severe LDOEs develop in some years but not in others.
Because of the proximity of the trench to the study area, and because it is the immediate source of water to that area, the wind-driven circulation pattern, and in particular short reversals or stalls in that pattern, affect water quality in the study area on the short 1–2-day time scale that represents the residence time in the trench, as well as on longer time scales as discussed above. Passage through the trench results in a net loss of dissolved oxygen in the water column because the depth of the narrow flow through the trench is much greater than the photic zone, and the respiratory and decay processes that consume oxygen over the entire water column overwhelm the photosynthetic production in the photic zone. The magnitude of the loss will depend on the rates of oxygen demanding processes, which vary in space and time. The magnitude of the loss also depends on the traveltime through the trench, which under routine conditions probably is on the order of 1.2 days [based on a GIS estimate of the volume of the trench of 61.6 m3 (T.L. Haluska, U.S. Geological Survey, unpub. data, 2005), and an estimate of the discharge through cross-section A-B (fig. 8) based on synoptic measurements in June 2005 of 580 m3/s (R.E. Wellman, U.S. Geological Survey, unpub. data, 2005)].
There are periods lasting from one to several days when the wind reverses and comes from the east to southeast. These periods are seen in the ADCP data as times when the currents, still in alignment with the trench bathymetry, reverse direction (figs. 6 and 7). The effect of these shifts is to reverse or stall the prevailing circulation pattern. One of the consequences of a reversal in the wind, or even a period of lower-than-normal wind stress, is that the traveltime in the trench is increased. Depending on how long the reversal or low-wind period lasts, the increase in traveltime can be substantial. When this happens, oxygen demanding processes have a longer time to act, and the effect on the quality of water in the trench can be substantial. This effect can be seen in the dissolved oxygen data collected at site UKL16, which was located in the midtrench area in 2004 (fig. 29). The hourly dissolved oxygen measurements at site UKL14, which is the first study site encountered by a water parcel leaving the trench heading north, also are shown.
The wind reversed for a day or more nine times during this 3-month record. Nearly every time there was a reversal in the wind, there was a rapid decrease in the dissolved oxygen at site UKL16; sometimes the effect was limited to the 1-m off-the-bottom measurement, and sometimes the effect was seen in the mid-water-column measurement as well. The pulse of low dissolved oxygen water through site UKL14 occurred after currents returned to the prevailing direction, delayed by the traveltime between sites UKL16 and UKL14. When the pulse of low dissolved oxygen water passed by site UKL14, it generally was moderated compared to conditions at site UKL16, and it was superimposed on the trends in the data at longer time scales. Sometimes there was no discernible signal at site UKL14, even after a long period of reversal, such as occurred from July 12–18, 2004. During the reversal, the effect at site UKL14 was often to generate widely fluctuating measurements, and sometimes a rapid increase in concentration, as water farther away from the trench (the “downstream” direction under prevailing conditions) was pulled into the site.
In 2003, there was no profiling buoy at the midtrench site, but data were collected at site UKL14 (fig. 30). As in 2004, the immediate effect of the wind reversals was usually an increase in the dissolved oxygen concentration and larger fluctuations, but this was usually followed by a decrease in concentration as the pulse of low dissolved oxygen water from the trench passed by. Between August 5 and August 27, there was a series of reversals that appear to build on each other such that the concentration at site UKL14 decreased a little more with each short event, finally resulting in a period of several days between August 23 and 29 when concentrations fluctuated below 5 mg/L, often dipping below 2 mg/L. The LDOE of 2003 that led to a fish die-off, however, occurred during the last 2 weeks in July, and there were no reversals in the period leading up to it. There was a reversal event between July 24 and 26, which caused concentrations at both sites to approach zero. This effect was superimposed on a longer-term trend, however, that resulted in concentrations in the 2–3 mg/L range before the reversal occurred, so a wind reversal does not appear to be the sole explanation for the July 2003 LDOE. It does, however, appear to have caused already low dissolved oxygen concentrations to decrease even further.
U.S.
Department of the Interior | U.S. Geological
Survey
Persistent URL: https://pubs.water.usgs.gov/sir20065209
Page Contact Information: Publications Team
Page Last Modified: Thursday, 01-Dec-2016 19:13:01 EST
