USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2008–5026

Back to Table of Contents

Nutrients and Chlorophyll a

Bloom Dynamics of Aphanizomenon flos-aquae

Photosynthetic pigments, like chlorophyll a, are measured as a surrogate for algal biomass because the cost and time required to collect and analyze chlorophyll a is less than that for measurements of algal biomass. In this report, algal biomass will be represented by chlorophyll a concentration. Between May and November, 90–100 percent of the total phytoplankton biomass in Upper Klamath Lake consists of AFA (Kann, 1997). The weekly water samples, although collected at only a subset of the sites in the study area, provided valuable context for understanding the data recorded by the continuous water quality monitors.

Trends and fluctuations in water quality parameters commonly were associated with algal trends and fluctuations, as reflected in the chlorophyll a data. Maximum concentrations of chlorophyll a, coinciding with similar maximum concentrations of dissolved oxygen and pH values, indicated an algal bloom. Week-to-week variation in chlorophyll a concentrations indicated that either populations were periodically blooming and declining (as a result of algal cell death) or that there was an inherent patchiness of algal growth and that the patches were moving around the lake. Indicators of cell death are low dissolved oxygen concentration and increased dissolved nutrient concentration. During an event of massive cell death, the magnitude of the increase in dissolved nutrient and decrease in dissolved oxygen concentrations can be variable and is related to the chlorophyll a concentrations before the event. Variation in chlorophyll a concentration caused by patchiness is not associated with increased nutrient or decreased dissolved oxygen concentration. Many blue-green algae, such as AFA, contain gas vesicles that allow cells to regulate their position in the water column by floating to the surface or sinking to the bottom. The sampling protocols used in this study may be insufficient to detect chlorophyll a very near the lake bottom, thereby falsely indicating a lack of chlorophyll a at a site. In addition, algae floating to the surface can form mats that are moved around by the wind, possibly resulting in a low chlorophyll a concentration at the site of measurement when in fact a bloom is in progress elsewhere in the lake.

A bloom decline is characterized by a sharp reduction in oxygen production through photosynthesis and is manifested as a decrease in dissolved oxygen concentration, generally to values potentially harmful to fish, as oxygen demand in the water column or sediments caused by decomposition of organic material continues. In contrast, periods of growth in the bloom are generally manifested as supersaturated dissolved oxygen concentrations and high pH values, as photosynthesizing algae consume carbon dioxide and produce oxygen.

At site MDN, data were collected from 2002 through 2005, allowing direct interyear comparison. In each year, the algal bloom expanded rapidly to an initial peak sometime between mid-June to mid-July, as measured by a maximum in chlorophyll a concentration (fig. 9). Between late July and mid-August in each year, chlorophyll a concentrations decreased concurrent with increased dissolved nutrient and decreased dissolved oxygen concentrations, indicating an algal bloom decline associated with large-scale cell death. Dissolved oxygen concentrations of less than 4 mg/L lasting between several hours and several days were measured at site MDN in association with this type of bloom decline from as early as late July to as late as mid-October during 2002–05 (fig. 9). The most severe (longest duration) low dissolved oxygen events occurred mid-season in 2003 and in 2005, and were associated with the largest decreases in chlorophyll a from the previous peak and the highest peaks in nutrient concentrations, particularly ammonia (note difference in the vertical scales among the years). The occurrence of a low dissolved oxygen event in October 2003, however, demonstrates that this type of event is not limited to occurring at mid-season, and can occur more than once in a single season.

The variability of chlorophyll a concentrations measured in 2005 suggests the occurrence of multiple summer chlorophyll a maximums, with concentrations at some sites reaching two or more maximums during the season (fig. 10). The highest chlorophyll a concentrations occurred after the mid-season bloom decline at sites HDB, MDL, MDT, and WMR (table 4). Data collected during previous years also indicated late-summer blooms of AFA (Kann, 1997; Wood and others, 2006). Measured reductions in bloom-related chlorophyll a concentrations at all sites except HDB in late July 2005 were notable because of the associated low dissolved oxygen events (LDOE) and increased un-ionized ammonia concentrations that were potentially harmful to fish. Chlorophyll a concentrations were lowest after the June–July bloom and were measured between July 18 and August 3, 2005, with most of the lowest concentrations in early August (table 4). The decrease in chlorophyll a concentrations was measured at site HDB in early August, a week later than the onset of decline at other sites. Howard Bay is largely disconnected from the circulation in the main body of the lake, and as a result the bloom there probably developed its own seasonal dynamics in response to localized nutrient, wind, and circulation patterns within the bay. During the widespread bloom decline, measurements recorded by the continuous water quality monitors showed that dissolved oxygen concentrations decreased to less than 6 mg/L at all sites, less than 4 mg/L at site MDN, and less than 2 mg/L at sites HDB, MDT, and EPT. Dissolved oxygen concentration at sites MDL and WMR did not decrease to such low levels, indicating that oxygen consumption was not as extreme at those two sites. Decreases in chlorophyll a concentrations after late-season blooms generally were not coupled with low dissolved oxygen concentrations, except at the trench sites MDT and EPT, and to a lesser extent at site HDB in October, as discussed below in the context of the water quality monitors. Dissolved oxygen concentrations at the trench sites probably were affected more by the respiration processes associated with bloom declines than concentrations at the shallow sites because most of the water column was below the photic zone at the deep sites, whereas most of the water column was within the photic zone at the shallow sites. One notable aspect of the sudden decrease in chlorophyll a concentrations at the end of July was that it occurred nearly simultaneously at all of the water quality sampling sites. A simultaneous change in chlorophyll a concentrations at each site indicates that algal production was affected by factors extending throughout the area, such as temperature or irradiance or a combination of the two.

Although samples were not collected in Agency Lake, qualitative weekly observations of bloom conditions were made at the Agency Lake sites (fig. 11). Trends in these observations closely match trends in chlorophyll a data gathered biweekly by the Klamath Tribes at sites AGN and AGS in 2005 (Klamath Tribes, unpub. data, 2005). From mid-June through August, bloom conditions cycled through extremes (between “very light” and “very heavy”) on a weekly to biweekly basis at site AGN when bloom conditions at site AGS were observed to be less variable (between “medium” and “heavy”). Bloom conditions above “light” were not observed past mid-August at either site in Agency Lake. This is in contrast to the bloom in Upper Klamath Lake, where the late-season bloom recovered to at least predecline levels at most sites (fig. 11).

Unlike deep lakes in which summertime chlorophyll a can often be predicted based on the phosphorus available at spring turnover (Dillon and Rigler, 1974), shallow lakes are not necessarily well-described by simple relations between phosphorus and biomass (Havens and others, 2001). Total phosphorus and chlorophyll a in UKL tend to increase simultaneously in spring, as determined within the resolution of weekly or bi-weekly sampling (Kann 1997; Wood and others, 2006). Mass balance studies have confirmed that the source of phosphorus is internal loading from lake sediments (Kann and others, 2001), although the precise mechanism for phosphorus loading is as yet unknown (Jacoby and others, 1982; Barbiero and Kann, 1994; Laenen and LeTourneau, 1996; Fisher and Wood, 2004). Initial results using pore-water profilers suggest that diffusive flux in combination with bioturbation is a possible mechanism for internal phosphorus loading (Kuwabara and others, 2007). Maximums in chlorophyll a concentrations in 2005 corresponded loosely with maximums in total phosphorus concentration (fig. 10), and total phosphorus concentrations correlated positively with chlorophyll a concentrations (fig. 12). Application of the Spearman Rank Order Correlation test to the data resulted in a correlation coefficient of 0.59 (p < 0.05, n = 101) for the relation between total phosphorus and chlorophyll a concentration in samples from combined sites. Management of the lake’s watershed has focused primarily on reducing phosphorus loads to the lake (Walker, 2001), but the evidence for phosphorus limitation seems inconclusive. A ratio of chlorophyll a to total phosphorus concentration around 1 or greater indicates that phosphorus was potentially limiting, whereas ratios much less than 1 indicate that phosphorus was not a limiting nutrient (White, 1989; Graham and others, 2004). Ratios of chlorophyll a to total phosphorus in the 2005 data ranged from 0.05 to 5.28, with 63 percent of samples less than 0.8, indicating that during most of June through October, phosphorus was not potentially limiting (fig. 13). The concentration of bioavailable phosphorus was higher during late-season blooms in comparison to the early-season bloom (fig. 10), and the ratio of bioavailable nitrogen to bioavailable phosphorus decreased between the early season bloom and late-season blooms (fig. 13), both of which suggest that phosphorus limitation was not consistent from June through October. In general, it is necessary to evaluate nitrogen as a potentially limiting nutrient (for example, TN:TP less than 10 indicates potential nitrogen limitation; Forsberg and Ryding, 1980), but because AFA can fix nitrogen, it seems unlikely that nitrogen is limiting. Other potential limiting factors include micronutrients such as iron, or light energy. Kuwabara and others (2007) noted that the trends of phosphorus in UKL were not as expected for a limiting nutrient, whereas the trends in iron were. It seems prudent at this point to leave the question of what is limiting the bloom in UKL open for further research.

Concentrations of the dissolved nutrients orthophosphate and ammonia were slightly negatively correlated with chlorophyll a concentrations (fig. 14). Correlations using the Spearman Rank Order Correlation test were –0.21 (p < 0.05, n = 101) and –0.31 (p < 0.05, n = 101) for orthophosphate and ammonia, respectively. The greatest maximums of orthophosphate and ammonia coincided with minimums in chlorophyll a concentrations at the end of July into early August. Low dissolved oxygen concentrations coincided with the decrease in chlorophyll a concentrations and the increase in dissolved nutrient concentrations in late July through early August, indicating that the nutrient maximums resulted from decomposition of senescing cells, which converts organic nutrients to inorganic form while consuming dissolved oxygen. Nitrite-plus-nitrate concentrations were low in comparison to ammonia concentrations, indicating that nitrification is not a rapid or effective process for removal of ammonia from the water column in Upper Klamath Lake.

Spatial Variability in Nutrient and Chlorophyll a Concentrations

As with dissolved oxygen concentrations, the bathymetry of the lake appears to influence dissolved nutrient concentrations. Median concentrations of ammonia and orthophosphate were highest at the deep trench sites MDT and EPT (fig. 15), particularly during the period of bloom decline when dissolved nutrient concentrations were the highest of the season. Median concentrations of dissolved nutrients were lowest at shallow sites MDN, WMR, MDL, and HDB, outside of the trench. This contrast indicates a greater degree of mineralization at the deep sites. The oxygen consumption rates obtained from dark bottles (discussed in section, “Dissolved Oxygen Production and Consumption Experiments”) do not indicate a large oxygen demand from the water column during the period of bloom decline. Therefore, oxygen consumption and mineralization processes likely occurred primarily at the sediment–water interface. This idea is supported by the downward vertical velocities measured near-bottom in the trench with ADCPs, in conjunction with sediments containing a large amount of organic matter (Gartner and others, 2007), which together indicate a great deal of settling of organic matter at the deep trench sites.

Chlorophyll a and total phosphorus concentrations at sites HDB and MDT were higher than at the other sites (table 4). The high concentration of chlorophyll a and total phosphorus at site HDB may be explained by the fact that Howard Bay is largely isolated from the flow regime in the rest of the lake, and the bloom in this bay tends to develop and evolve somewhat independently of the bloom in the rest of the lake. The high chlorophyll a concentrations in samples collected at site MDT coincide with observations of the field crews of thick mats of AFA close to the water surface. Additionally, the primary location for companies collecting AFA for commercial production is near site MDT, indicating easy access to high concentrations of AFA near the surface at that site. These observations are consistent with the measurements made with the water quality monitors at MDT that show the water column there has a tendency to stratify more strongly than in the shallow parts of the lake (as discussed in the section on stratification and wind speed), particularly when air temperature is increasing, creating conditions that allow AFA colonies to make use of their buoyancy-regulating mechanism. On the basis of the circulation model, a counterclockwise flow cell forms east of Eagle Ridge and south of Bare Island that may capture large algal mats in a continuous loop (fig. 7). In addition, particle-tracking studies with the model have shown that water can be trapped in the cell for days, particularly when winds are weak (Cheng and others, 2005; Wood and Cheng, 2006). The combination of surface accumulation when conditions are stratified and relatively limited horizontal transport creates ideal conditions for AFA to form very dense blooms at this site.

Back to Table of Contents