Interannual Variability, 2007–09


Seasonal and interannual fluctuations in Upper Klamath Lake water quality and associated environmental parameters frequently reflect changes in A. flos-aquae-dominated phytoplankton biomass as measured by the photosynthetic pigment, chlorophyll a, and may be important for understanding the occurrence of cyanotoxins in the lake. Indicators of bloom decline and cell senescence include decreases in chlorophyll a concentrations, pH, and dissolved oxygen concentrations that accompany decreased photosynthetic activity (and active decomposition), and an increase in dissolved nutrient concentrations as cells lyse in the water column. In addition, year-to-year variability in the severity of these changes may be strongly related to weather (Perkins and others, 2000; Kann and Welch, 2005). Periods of bloom decline often coincide with periods of sustained high temperature and low wind, which increases water column stability (Kann and Welch, 2005). In years when a severe mid-season bloom decline is observed, the decline is accompanied by the seasonal maximum water temperature (Wood and others, 2006; Hoilman and others, 2008; Lindenberg and others, 2009; Kannarr and others, 2010). 


Chlorophyll a and Nutrients


Water quality and meteorological data collected in 2007 and 2008 and used in this study are discussed in Kannarr and others (2010). Temporal changes in the median values of a subset of water quality parameters that are hypothesized to influence the presence of, or short-term changes in the occurrence of, cyanotoxins are summarized in figure 2 of the current report. Comparison of median values show that chlorophyll a concentrations were generally higher from July to September in 2008 than at the same sites sampled in 2009. Although the difference in seasonal median values was not large, 2008 samples collected between June 16 and September 15 exhibited a wider chlorophyll a concentration range (median values from 206 to 27 µg/L in 2008 and from 164 to 3.4 µg/L in 2009), but did not decrease sharply mid-season, as in 2009, signifying a major bloom decline. Chlorophyll a data analyzed in 2007 were not reported because the quality of the data was poor (Kannarr and others, 2010). Median values of total nitrogen and phosphorus were highest in 2007 and comparable between 2008 and 2009. The annual median ammonia concentration measured in 2009 was more than double that of 2007 and more than 10 times the median value for 2008. This large difference in overall ammonia concentrations between years, in contrast to the smaller difference in orthophosphate, accounts for the higher seasonal median ratio of dissolved inorganic nitrogen to dissolved inorganic phosphorus (DIN:DIP) observed in 2009. Between July and September, the 2009 median DIN:DIP (3.30) was nearly three times the median value observed in 2007 (1.19) and six times the 2008 median value (0.55). 


Total nitrogen to total phosphorus (TN:TP) ratios indicate potential for phytoplankton nutrient limitation, and low TN:TP values (less than 29 by weight; Smith, 1983) are often associated with cyanobacterial success (Smith, 1983; Paerl, 1988; Jacoby and others, 2000). TN:TP ratios greater than 17, by weight, and chlorophyll a to total phosphorus ratios (chlorophyll a:TP) greater than 1, by weight, have been measured in phosphorus-limited lake systems (Forsberg and Ryding, 1980; White, 1989; Graham and others, 2004), including Upper Klamath Lake, where previous data generally show potentially phosphorus-limiting conditions early in the sampling season (late June and early July; Lindenberg and others 2009). In a study of 30 waste-receiving lakes sampled with high frequency over a 6-year period (4,500 samples), nitrogen was the most growth-limiting nutrient at TN:TP ratios less than 10 (Forsberg and Ryding, 1980). In addition, laboratory experiments have shown decreased amounts of microcystins under low phosphorus conditions (Sivonen and Jones, 1999), and recent studies using environmental samples show strong correlations between microcystin concentrations or M. aeruginosa cell densities and levels of total phosphorus (microcystin-LR in particulate form; Kotak and others, 2000; Downing and others, 2001; Chen and others, 2009) or total nitrogen (Downing and others, 2001). Seasonal median values of TN:TP ratios in Upper Klamath Lake between July and September were near 10 in all years (fig. 2), and chlorophyll a:TP ratios during this time do not support phosphorus (P)-limitation, because most values were less than 1 (fig. 3). However, consistent with results from previous years (Lindenberg and others, 2009), ratios of TN:TP were higher between mid- to late-June in all years (fig. 2), suggesting that, if P-limitation does occur, it is most likely during the first bloom of predominantly A. flos-aquae. 


Dissolved Oxygen, pH, and Temperature


Dissolved oxygen and pH data collected from continuous monitors provide high-resolution measurements of water quality resulting from growth and decline of the A. flos-aquae-dominated bloom. Rapid growth phases were indicated in 2007 by supersaturated dissolved oxygen concentrations prior to the first week in July, between the last week in July and first week in August, and after the first week in September. Undersaturated dissolved oxygen concentrations (less than about 90 percent) and slightly lower pH signified bloom declines for 2 weeks in mid-July, at the end of August, and into early September (fig. 4; Kannarr and others, 2010). In 2008, undersaturated dissolved oxygen concentrations and lower pH in late May indicated a minor bloom decline, although the first major decline did not occur until mid-August and for a shorter time period in mid-September (fig. 4; Kannarr and others, 2010). Based on the degree of undersaturation and the relatively small decrease in pH, bloom declines in 2007 and 2008 were less severe than in 2009, when a large, mid-season decline occurred during the last 2 weeks in July (fig. 4), coinciding with the seasonal minimum in dissolved oxygen (2.45 mg/L) and pH (7.30; fig. 4). As previously noted, when a large mid-season bloom decline occurs, minimum dissolved oxygen concentrations and pH values are observed within a few days of the seasonal maximum water temperature (Hoilman and others, 2008; Lindenberg and others, 2009; Kannarr and others, 2010). Therefore, the bloom cycle in each year of this study can be qualitatively summarized between July and September as: two declines in 2007 that culminated in mid-July and early September, a relatively extended, mild bloom decline in 2008, which occurred during the last 3 weeks of August, and a major bloom decline in 2009 that progressed throughout July and was most severe in the second half of that month. This description is broadly consistent with results of chlorophyll a analysis in samples collected once every 2 weeks from a larger set of eight sites in Upper Klamath Lake by the Klamath Tribes, Chiloquin, Oregon (Kann, 2010). 


Cell Associated Microcystins


Preliminary data suggest that exposure of juvenile suckers to microcystins in Upper Klamath Lake is through ingestion, either through the food chain or by direct consumption of toxigenic cyanobacteria strains (VanderKooi and others, 2010). Therefore, it may be important to understand how microcystin concentrations vary between the particulate (intracellular, or cell associated) and dissolved (extracellular) phases. In all years, microcystins in the large (> 63 µm) particulate fraction, representing filamentous and (or) large planktonic colonies, were more concentrated than microcystins within the small (1.5–63 µm) particulate fraction (table 2). Between July and September, large particulate microcystin concentrations, expressed volumetrically, were highest in 2007 (median = 1.87 µg/L) and were more variable with a higher maximum in 2009 (range between 0.001 and 24.4 µg/L; table 2). Large particulate microcystin concentrations were lowest in 2008, ranging between 0.01 and 1.55 µg/L (median = 0.20 µg/L). In 2009, microcystins were dominantly in the large particulate fraction on most sample dates at all sites (fig. 5), indicating the presence of a microcystin-producing bloom that year. Interestingly, 2008 samples contained the lowest microcystin concentrations in the study, but the dissolved fraction comprised a larger percentage of total microcystins that year. This may be due to the presence of fewer M. aeruginosa colonies that year (J. Kann, Aquatic Ecosystem Sciences, LLC, unpub. data, 2009). However, given the relatively low concentrations of dissolved and particulate microcystins in 2008, this dominance of dissolved microcystin concentrations may not be significant. 


Results of monthly sampling in 2007 provided only a general temporal pattern of particulate microcystin concentrations that year, but peak concentrations (> 12 µg/L) were observed at four of the six sites sampled on July 31 and August 1, at least 10 days earlier than when particulate fraction microcystin concentrations at all sites peaked in 2009 on various dates between August 10 and 31 (fig. 2). Two sites, MDN and WMR (fig. 1), were sampled for cyanotoxins in all 3 years, so the data collected from these sites can be compared explicitly. Samples from site MDN contained maximum large particulate microcystin concentrations of 17.4, 0.33, and 11.8 µg/L, measured on August 1, 2007, September 8, 2008, and August 17, 2009, respectively. Maximum concentrations measured in samples from site WMR were 7.35, 1.13, and 6.23 µg/L and occurred on September 5, 2007, September 8, 2008, and August 10, 2009, respectively (table 3). It also is noteworthy that on the last sample dates in 2007 (October 17), 2008 (September 22), and 2009 (September 14), concentrations were all less than 1 µg/L, less than or equal to 1.44 µg/L, and less than or equal to 2.26 µg/L, respectively. This indicates a significant reduction in microcystin-bearing cells in the water column by the end of September (fig. 2). 


Expressed as mass per dry weight of suspended solids, the median concentration of microcystins in the large particulate fraction between July and September was highest in 2009 (0.15 µg/mg), lowest in 2008 (0.02 µg/mg), and between these values in 2007 (0.09 µg/mg; table 2). This suggests that the contribution of toxigenic M. aeruginosa to the phytoplankton community was greatest in 2009 and least in 2008. 


Dissolved Microcystins


Samples collected in 2009 contained higher (by more than five times) and more variable concentrations of dissolved microcystins (median = 1.49 µg/L) than the 2008 samples (median = 0.27 µg/L; table 2), although 15 percent of the samples collected in 2009 contained dissolved microcystins less than the detection limit (dissolved microcystins were detected in all 2008 samples, but sampling began later that year). In 2008, microcystins were primarily in the dissolved fraction (85.8 percent), but in 2009, dissolved microcystins comprised less than one-half of the total microcystin concentration (43 percent of the total concentration was in the dissolved form, and 56 percent was in the large particulate fraction; table 2; fig. 5). Concentrations in 2008 remained less than 1 µg/L at all sites except MDN, where the concentration was just over 1 µg/L on 2 sample dates, July 14 (1.10 µg/L) and September 22 (1.4 µg/L), and, like particulate fraction microcystins that year, did not follow a recognizable trend. It is, therefore, difficult to interpret the meaning of the higher percentage of dissolved microcystins that year, particularly because the regulation of toxin production in cyanobacteria is not well understood and the microcystin content of individual cells may be highly variable. However, toxins are released into the water column primarily following death and senescence of a toxigenic bloom, so it is likely that higher dissolved microcystin concentrations in 2008 represent longer or more frequent periods of decline in the microcystin-producing population. In 2009, dissolved microcystin concentrations were less than 0.2 µg/L through July 14. After July 14, the concentrations at all sites except site RPT increased over the next 3 weeks to their seasonal maximum concentrations (fig. 2); concentrations did not peak at site RPT until September 8. Through the remaining sampling period, concentrations at all sites were highly variable. Maximum concentrations in samples from site MDN were 1.44 and 3.93 µg/L measured on September 22, 2008, and August 3, 2009, respectively. Samples from site WMR collected on June 30, 2008, and August 3, 2009, contained maximum concentrations of 0.48 and 2.49 µg/L, respectively. 


First posted May 30, 2012