USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5117

Back to Table of Contents

Temporal and Spatial Variability of Poor Water-Quality Conditions

For this report, water-quality conditions are considered poor when they have the potential to be harmful to the endangered sucker species. Based on high-stress thresholds established to calculate stress indices for Upper Klamath Lake suckers (Reiser and others, 2000), the values used in this study to delimit poor water-quality conditions were water-temperature values greater than 28°C, dissolved-oxygen concentrations less than 4 mg/L, and pH values greater than 9.7. Additionally, the USEPA (U.S. Environmental Protection Agency) criteria (U.S. Environmental Protection Agency, 1999) were used to screen total ammonia concentrations, and un-ionized ammonia concentrations were evaluated against LC₅₀ (median lethal concentrations) values of 530 µg/L for Lost River suckers and 780 µg/L for shortnose suckers (Saiki and others, 1999).

Water Temperature

Of the 7,090 water temperatures recorded during May–October 1990–2006, only 10 values were greater than 28°C. These temperatures were all measured at the surface in the afternoon, and most measurements were in the latter part of July, when air temperatures are expected to be elevated. The two warm water temperatures were recorded in June 1997 and 2000, years with warmer springs and earlier blooms (table 6). High pH values also were recorded with these June warm water temperatures, which fit with the early onset of the bloom. The period when the bloom is growing generally is associated with high dissolved-oxygen concentrations and high pH values, as photosynthesizing algae consume carbon dioxide and produce oxygen. One-half of measurements with water-temperature values greater than 28°C also were associated with dissolved-oxygen concentrations greater than 12 mg/L and pH values greater than 9.7.

Wood and others (2006) determined the NCDC wind-speed dataset to be somewhat troublesome and suggested that these data should be interpreted qualitatively. They noted evidence of “binning” of the data (rather than being continuous, data values tended to fall on discrete values) and numerous occurrences of zero wind speeds, perhaps suggesting the use of a threshold value for reporting the data. As a way of examining the quality of the NCDC data, the median values were compared to the median monthly wind speeds (calculated in the same manner as the NCDC medians) for July and August from the Agrimet weather station located adjacent to Agency Lake. This comparison was repeated in this report including data through 2006 (table 5). Many differences, not only in values between the years, but also in the ranking of the years based on median wind speed, exist between these two datasets. One notable discrepancy in the NCDC dataset is the large difference between the 2001 median values for July and August and the rest of the years. These discrepancies further hint at the questionable quality of these data; therefore, an overall ranking of the years based on wind speed was not pursued. Because wind speed was suspected to be a potentially important climatic factor, and because the Agrimet dataset only covered 2000–2006, the longer-term NCDC wind-speed data were used in the information-theoretic approach discussed later in this report.

Dissolved Oxygen

By filtering the 1990–2006 dataset using the dissolved-oxygen guideline for delimiting poor water-quality conditions, the mid- to late-season pattern of excursions into the less than 4 mg/L range can be observed (table 7). The highest frequency of concentrations less than 4 mg/L occurred in August, then in July, and then in September. Low dissolved oxygen was measured during 3 years (1998, 2000, 2002) as early as June.

Further examining the dissolved-oxygen data by site and year (table 8) revealed that the highest frequencies of low dissolved-oxygen concentrations were observed at the Eagle Ridge site. These increased frequencies are due to the larger number of measurements made at this deeper site (typically 7 to 8 measurements per visit at this deeper site and 3 to 4 measurements per visit at the other sites) and the effect of deep circulation through the trench. Light does not penetrate this deeper water enough to drive photosynthesis, yet respiration continues, depleting the water column of dissolved oxygen.

pH

Unlike the patterns of poor water-quality conditions observed in the dissolved-oxygen data, high pH values were more frequent and occurred earlier in the season (table 9). This offset in the occurrence of low dissolved-oxygen concentrations and high pH values is better for overall water-quality conditions than if these stressors were occurring together more frequently. The highest frequency of pH values greater than 9.7 occurred in June and July, with high values persisting into August and September for some years.

Spatial patterns show that when high pH values were measured, they were often observed at sites all around the lake (table 10). During most years, there are occurrences of high pH values at most sites; however, during years with a lower overall frequency of pH values greater than 9.7, not all sites experience these poor water-quality conditions. The Agency Lake sites seem to have a higher frequency of high pH values, and in June, often had higher pH values than Upper Klamath Lake sites. The depth profiles also showed that there was little vertical variation in pH values throughout the water column. High pH values tend to occur either at the surface or throughout the water column, as can be observed by examining the depth profile pH data for 2000, a year with a high frequency of pH values greater than 9.7 (fig. 5).

Ammonia

Prior to 1999, the USEPA criteria for ammonia were established to evaluate un-ionized ammonia concentrations, because it was believed that the un-ionized part of total ammonia was much more toxic to aquatic life. It has been recognized, however, that the ammonium ion (the other part that makes up total ammonia) also may contribute significantly to ammonia toxicity under certain conditions (U.S. Environmental Protection Agency, 1999). Therefore, the current criteria are based on total-ammonia concentrations. These criteria are based on the water temperature and pH of the samples and are applicable for waters with a pH between 6.5 and 9. The guidelines used to evaluate total-ammonia concentrations for this analysis are based on the acute criterion (based on a 1-hour average concentration) for when salmonids are present and the chronic criterion (based on a 30-day average concentration) for when fish early life stages are present. For example, the criterion is 2.2 mg/L at pH 8.5 when adult salmonids are present and 0.8 mg/L at pH 8.5 and 18°C when early life stages are present.

Roughly one-half of samples collected in Upper Klamath and Agency Lakes between 1990 and 2006 had pH values of less than 9 and therefore were compared against the USEPA criteria (table 11). For those samples with exceedances, the chronic (30-day average) criteria values generally were about 70 percent lower than the acute (1-hour) criteria values. A small number of samples (less than 1 percent) exceeded the acute guideline, and one-half of these were measured in Shoalwater Bay. There were a few more exceedances (less than 10 percent) of the chronic guideline, but the data were too sparse to calculate representative 30-day environmental concentrations. Almost one-half of these exceedances were measured in the bay areas (Coon Point, Shoalwater Bay, and Wocus Bay).

The part of the dataset that could not be evaluated against the USEPA criteria because the pH was greater than 9 may be the part that signals more detrimental conditions for the suckers. When high ammonia concentrations occur with high pH and water temperatures, a significant fraction of the total concentration of ammonia will be present in the un-ionized form, which is much more toxic to aquatic life. By using the USEPA formulas (U.S. Environmental Protection Agency, 1999), un-ionized ammonia concentrations were calculated from total ammonia concentrations, pH values, and water temperatures for the entire 1990–2006 dataset. These concentrations were evaluated against the LC₅₀ value of 530 µg/L for Lost River suckers, recognizing that although the LC₅₀ value for shortnose suckers is higher, concentrations greater than 530 µg/L would be stressful for shortnose suckers as well. Although the acute and chronic total-ammonia USEPA criteria represent “acceptable no-effect levels” of total ammonia, the LC₅₀ values for un-ionized ammonia concentrations represent “unacceptable severe-effect levels” for the health of the suckers. About 3 percent of the 2,023 un-ionized ammonia concentrations exceeded 530 µg/L, and there was at least one exceedance at each of the 10 sites (table 12). One-third of these exceedances occurred in the bay areas. All except one of these exceedances occurred between 1996 and 2003, when total ammonia concentrations were the highest of the 17-year period (Wood and others, 2006).

Interrelatedness of Water-Quality Variables

It is useful to examine how these water-quality variables that are used to delimit poor water-quality conditions are interrelated, and how they relate to the algal blooms occurring in Upper Klamath Lake. According to Wood and others (2006), bloom dynamics, as observed in the chlorophyll‑a data, are associated with the trends and fluctuations observed in water-quality parameters. A crash in the algal bloom results in a sharp reduction in chlorophyll‑a concentrations. As the oxygen production from photosynthesis stops or decreases, and the sediment and water-column oxygen demands continue, there is a decrease in dissolved oxygen, generally to values below saturation. In contrast, when the bloom is growing, photosynthesizing algae consume carbon dioxide and produce oxygen, resulting in supersaturated dissolved oxygen and high pH.

Four years (1992, 1991, 2006, and 1997) were used to explore how mean water temperatures, minimum dissolved-oxygen concentrations, maximum pH values, median ammonia concentrations, and median chlorophyll‑a concentrations are related under different combinations of conditions. As a reminder, 1992 was ranked as the lowest lake level, warmest spring, and earliest bloom onset. The year 1991 also was ranked as a low lake-level year, but perhaps as a result of the cool spring, it had the second latest bloom onset. The second earliest bloom onset occurred in 1997, a year with a lake level in the middle to high end of the rankings but a warm spring. In contrast, 2006 had a middle level ranking for both lake level and spring temperatures, but had the latest bloom onset.

In every year, the dissolved oxygen decreases during July–August, but for years with large chlorophyll‑a concentrations in June (namely 1992 and 1997), the dissolved-oxygen concentration is more likely to fall below 4 mg/L. The large early bloom in 1992 was most noticeable in the bay and outflow areas, although the dissolved-oxygen concentration did not decline as low in the outflow areas as it did in the bay areas. The later bloom in 1991 was much more apparent in the bay areas than in any other area. The dissolved-oxygen concentrations in 1991 did not fall below 4 mg/L until September-October. The pH values correspond very well with the chlorophyll‑a concentrations—pH increasing with the increases in chlorophyll‑a—which fits with the growth of the bloom. Ammonia concentrations in 1997 were notably larger than in any other year. This was determined by Wood and others (2006) to be correlated with discharge from the Williamson River. The fact that the ammonia concentrations were elevated in 1997 in Agency Lake also indicates that the discharge in the Wood River, which flows into Agency Lake, was probably higher during those years as well.

When evaluating dissolved-oxygen concentrations, pH values, and ammonia concentrations against the criterion defined to delimit poor water-quality conditions, one year (2000) was in the top three highest frequencies of poor conditions for all three parameters. The interrelatedness of these water-quality variables was further explored in figure 6. Similar to the conditions discussed by year, the areas with the higher chlorophyll‑a concentrations earlier in 2000 (namely the bay and outflow areas) experienced the lowest dissolved-oxygen plunges in June and July. Dissolved-oxygen concentrations measured in all areas were low in late June and early July, but the bay areas experienced low dissolved-oxygen conditions through September and into October in some years. All areas experienced elevated pH values with the onset of the bloom. The highest ammonia concentrations coincided with the elevated chlorophyll‑a concentrations associated with the growth of the bloom. These patterns illustrate the importance of the bloom (both the magnitude and timing of the onset, growth, and decline) on water-quality conditions and why it is necessary to further explore potential correlations with lake level in an attempt to find a management tool to influence water-quality conditions.

Back to Table of Contents