USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2008–5201

Back to Table of Contents

Dissolved Oxygen Production and Consumption Experiments

Depth of Photic Zone

A general inverse relation was observed between photic zone depth and chlorophyll a concentrations during the light and dark bottle experiments (fig. 15). A Spearman’s rank correlation coefficient between chlorophyll a concentrations and depth of photic zone resulted in a significant negative correlation for pre- and post-incubation measurements, –0.916 (p < 0.05, n=111) and –0.877 (p < 0.05, n=110), respectively. The inverse relation between depth of photic zone and chlorophyll a was indicative of the effect that blooms of AFA can have on the transmittance of radiation through the water column. The most extensive occurrence of shading by the AFA bloom was at the HDB site on July 24, when high chlorophyll a concentrations blocked light to the water column to cause net respiration in light bottles on both racks. Other than this occurrence at site HDB, no other occurrence of a substantial negative production rate was measured from light bottles positioned on the upper rack. During some experiments, production rates from light bottles on the lower rack were negative, indicating that the lower rack was sometimes below the photic zone. The median photic depth of all measurements in 2006 was 1.58 m, 0.08 m deeper than the position of the lower rack in the water column.

Dissolved Oxygen Production and Consumption

The rate of change in dissolved oxygen concentration in light and dark incubation bottles at two levels in the water column are shown as time series in figure 16. Rates obtained from the two light bottles on each rack were averaged to get the net production rate at each depth. Oxygen consumption rates were based on the value from a single dark bottle at each depth. Typically, rates from light bottles were positive, representing net dissolved oxygen production, and rates from dark bottles were negative, representing net dissolved oxygen consumption. The experiments showed that the water column in Upper Klamath Lake was characterized by high rates of both dissolved oxygen production and consumption, with a maximum production rate of 2.79 (mg/L)/h of oxygen recorded at HDB on August 7 and a maximum consumption rate of –2.14 (mg/L)/h of oxygen measured at site HDB on July 24. Outside of site HDB, which was something of an outlier, the maximum consumption rate measured was –0.98 (mg/L)/h of oxygen at site MDN on August 22, and the maximum production rate was 2.77 (mg/L)/h of oxygen at site MDN on July 11. Because the bottles were filled with unaltered lake water, the experiments measured oxygen production and consumption of the entire planktonic community (phytoplankton, zooplankton, and bacteria). Because AFA typically constituted more than 90 percent of the phytoplankton by weight during midsummer (Kann, 1997; Perkins and others, 2000), dissolved oxygen production was attributable mostly to photosynthesis by AFA. However, the abundance of oxygen consuming components of the planktonic community relative to AFA is less well known, particularly bacterial biomass. AFA likely is a major contributor to oxygen consumption observed in the experiments, but without knowledge of the bacterial biomass relative to that of AFA it cannot be conclusively stated that AFA dominates dissolved oxygen consumption processes.

The rate of change of dissolved oxygen measured in dark bottles was sometimes positive (fig. 16), suggesting the unlikely occurrence of oxygen production in absence of light. Most positive rates of change in oxygen of dark bottles occurred in the beginning of the sampling season, coincident with low chlorophyll a concentrations, when oxygen production due to algae was minimal. The positive change in dissolved oxygen measured in dark bottles was within the precision of the dissolved oxygen meter (± 0.1 mg/L or ± 2 percent of measurement). The minimum and maximum positive change in dissolved oxygen measured in dark bottles was 0.01 and 0.12 mg/L, within the range of error associated with measurements. Another positive rate of change in dissolved oxygen in dark bottles was measured on August 22 at site WMR in the upper and lower racks (0.28 and 0.33 (mg/L)/h of oxygen, respectively). This rate of change was too large to be explained by measurement error of the dissolved oxygen meter. The only plausible explanation for production in a dark bottle would be that light leaked into the bottle.

Excluding the positive oxygen change rates in the dark bottles, consumption rates measured in dark bottles were comparable to the overnight consumption rates measured in Upper Klamath Lake in 2002 by Lieberman and others (2003). Consumption rates ranged from 0 to -2.14 (mg/L)/h of oxygen in this study, whereas consumption rates in the 2002 study ranged from -0.05 to -0.49 (mg/L)/h of oxygen. The larger range of consumption rates determined in this study is attributable to annual variability and because experiments in this study spanned 6 months, whereas the 2002 study comprised 2 consecutive days in late July and 2 consecutive days in mid-September.

Rates of oxygen production were determined by the rate of photosynthesis. Rates of net oxygen production were most variable at site HDB and least variable at site EPT. Net production rates measured on the top rack (at 0.5 m depth) generally increased in late June and decreased in late July and early August at most sites. At sites WMR and EPT, oxygen production decreased to rates measured during pre-bloom conditions. The decreased net oxygen production rates at the 0.5 m rack is consistent with a decrease in chlorophyll a concentrations measured during the same time. At four sites, MDN, WMR, RPT, and EPT, the net oxygen production rate was correlated positively and significantly with chlorophyll a concentration. The Spearman’s rank correlation coefficient between oxygen production rates and chlorophyll a concentrations at sites MDN, WMR, RPT, and EPT was 0.549, 0.900, 0.672, and 0.852, respectively (p < 0.05, n=18, 19, 19, and 19, respectively) (fig. 17). However, there was no significant correlation between the two variables at sites HDB and MDT, possibly because at chlorophyll a concentrations greater than about 300 µg/L, self-shading can limit the transmittance of light through the water column. Thick mats of algae have been observed in the areas of sites HDB and MDT (fig. 18). The most extreme case of water column shading was seen at site HDB on July 24, 2006, where high respiration rates were measured in dark and light bottles on both racks, indicating negligible photosynthetic production of oxygen even 0.5 m below the surface.

The potential change in dissolved oxygen over a 24-hour period was calculated from oxygen production and consumption rates measured in light and dark bottles. The relation of chlorophyll a concentration to 24-hour change in dissolved oxygen varied among sites (fig. 19). The 24-hour change in dissolved oxygen at sites in the trench, sites MDT and EPT, was negatively correlated with chlorophyll a concentration, indicating that as chlorophyll a concentrations increased, oxygen consumption increased. Oxygen-consuming processes dominated at the trench sites because most of the water column is below the photic zone. Most of the 24-hour changes in dissolved oxygen calculated for sites MDT and EPT were negative, indicating that oxygen consumption processes predominated in the deeper trench areas. In contrast, 24-hour change in dissolved oxygen was positively correlated and predominantly positive at sites WMR, RPT, and HDB, indicating that oxygen production processes were dominant at the shallow sites.

Site MDN was more variable in the amount of oxygen consumption or production in a 24-hour period, and no significant correlation was determined between chlorophyll a concentration and 24-hour change in dissolved oxygen. The water at site MDN is not as shallow as at sites WMR, RPT, and HDB, but not as deep as at sites MDT and EPT; therefore, both oxygen production and consumption processes are equally probable.

During the period of bloom decline, the 24-h change in dissolved oxygen concentrations decreased in magnitude at every site except HDB (fig. 20), indicating that this period was characterized by lower production and lower consumption at those sites. This also was evident in the light and dark bottle incubations (fig. 16). Because this period also was characterized by decreasing dissolved oxygen concentrations (fig. 21), the reduced water-column oxygen demand suggests that much of the oxygen consumption was taking place at the sediment-water interface in the form of sediment oxygen demand (SOD). SOD was measured in Upper Klamath and Agency Lakes in spring (May 18–May 27) and late summer (August 24–September 1) 1999 (Wood, 2001). The potential 24-hour reduction in dissolved oxygen obtained from the measurements at nine sites around Upper Klamath Lake ranged from 0.2 to 0.6 mg/L in spring, and from 0.4 to 1.3 mg/L in late summer (excluding one uncertain value of greater than 3.7 mg/L in Ball Bay). At these rates, processes at the sediment-water interface would be secondary to water-column processes (fig. 20) during most of June through October; during a bloom decline, however, photosynthetic production of oxygen is so diminished that this magnitude of SOD is more important in the oxygen mass balance than during periods of bloom growth.

Back to Table of Contents