Previous Investigations

Several early studies characterized water quality, algal taxonomy, and primary productivity in Blue Mesa after algae and cyanobacteria were observed in the reservoir soon after impoundment (Blackwell and Boland, 1979; Britton and Wentz, 1980; Cudlip and others, 1987; Long and others, 1996; Bauch and Malick, 2003). In 1999, the USGS and the NPS completed an intensive study of nutrient status and productivity in Blue Mesa Reservoir (Bauch and Malick, 2003). They reported that nitrogen and phosphorus concentrations were low in Blue Mesa, and conditions were likely nitrogen limited during summer and fall. Diatoms (primarily Asterionella and Melosira sp.) dominated the phytoplankton community composition during spring and early winter; cyanobacteria (primarily Aphanizomenon) dominated during summer and fall when temperatures were warmer. Phytoplankton density was highest in Iola basin compared to the downstream basins. Bauch and Malick (2003) also computed trends using historical nutrient and chlorophyll-a data during 1992–99. They reported there were no statistically significant trends in concentrations of nutrients or chlorophyll-a within Blue Mesa except for a small downward trend in phosphorus at the west end of the reservoir, suggesting that the reservoir had not experienced increased nutrient enrichment and productivity prior to 1999 (Bauch and Malick, 2003).

Previous studies have also evaluated nutrient inputs to Blue Mesa from tributary streams. Brown and Thoma (2012) compiled nutrient data for streams in the Blue Mesa watershed for 1974–2007. They compared concentrations to EPA-recommended nutrient criteria for forested mountain and xeric west ecoregions (EPA, 2000) and evaluated trends for selected sites during varying periods of record between 1995 and 2007. More recently, Weissinger and Gibney (2019) analyzed water-quality data collected by the NPS at Blue Mesa during 2010–18. They reported elevated total phosphorus in several tributaries and in Blue Mesa that sometimes exceeded the State of Colorado interim aquatic life standard of 0.11 milligram per liter (mg/L) (Colorado Department of Public Health and Environment, 2023). Trend analysis by Weissinger and Gibney (2019) for 2001–14 determined a few statistically significant trends in pH and major-ion concentrations at tributary sites but no trends in nutrients for the tributary or reservoir sites.

Concurrent Investigation

A companion investigation completed concurrently with this study assessed the potential use of remote-sensing data to map chlorophyll-a concentrations and water temperature to support algal bloom monitoring in Blue Mesa (King and others, 2024, 2025). In situ chlorophyll-a measurements collected during this study and multispectral satellite imagery were used to train a machine learning model that reconstructed summertime chlorophyll-a concentrations throughout all basins in Blue Mesa from 2016 to 2023. Historical water surface temperature measurements and a remote-sensing temperature product were used to reconstruct summertime water surface temperatures in Blue Mesa from 2000 through 2023. The chlorophyll-a results indicated that the algal blooms (as mapped by chlorophyll-a concentration) exhibit a consistent spatial pattern across multiple years, initiating in Iola basin and spreading west to Cebolla and Sapinero basins through the season. The temperature records indicate that water temperatures in Blue Mesa increased linearly at a rate of 0.3 degree Celsius (°C) per decade for 2000–23. This increase was 0.04, 0.03, and 0.02 °C per year for Iola, Cebolla, and Sapinero basins, respectively. Further, the temperature records showed that, in general, years with below average reservoir levels had above average water temperatures. Finally, there was a moderate positive correlation between summertime remotely sensed chlorophyll-a concentrations and summertime remotely sensed surface water temperature (King and others, 2025).

Continuous Water-Quality Measurements

Multiparameter continuous water-quality monitors (EXO-2) were installed at 3.3-ft depth at Iola basin at site 2 (USGS site 382852107054801) and at Blue Mesa Reservoir west of Dry Gulch near Sapinero, Colo. (USGS site 382847107120401; site 25, table 1, fig. 1). The Iola basin monitor at site 2 (Iola monitor) operated from July 25 through September 28, 2022, and the Cebolla basin monitor at site 25 (Cebolla monitor) operated from September 1 through September 24, 2022. Each monitor conducted field water-quality measurements using wiped sensors that measured water temperature (degrees Celsius), specific conductance (microsiemens per centimeter at 25 °C), pH (standard units), DO (milligrams per liter), turbidity (formazin nephelometric units), and total algae in relative fluorescence units for chlorophyll-a and phycocyanin (fPC). The EXO-2 monitors were wrapped in protective sleeves, and each sensor was wrapped in copper antifouling tape. Monitor calibrations at both sites were checked before and after deployment according to USGS guidelines (Wagner and others, 2006). At the Iola monitor, a cleaning and fouling check was performed 6 weeks after deployment. The monitor recorded measurements every 15 minutes and data are published in NWIS database using USGS site numbers (table 1; USGS, 2024a).

During September 2023, two boat surveys were conducted in Iola basin using an EXO-2 monitor that was continuously deployed to obtain additional data to aid in constructing the satellite-derived chlorophyll-a maps (King and others, 2025). The EXO-2 monitor was configured and calibrated as described in the previous paragraph. The monitor data are available in Qi and others (2025), and results are briefly discussed in appendix 3.

Quality Assurance

Quality-control samples for nutrients, chlorophyll-a, and major and selected trace elements included 14 field blanks and 18 sequential replicates collected during the study period (USGS, 2024a). Field blanks were used to evaluate the potential for sample contamination from sample collection, processing, and analysis, and replicate samples were used to evaluate sampling and analytical variability (Mueller and others, 2015). For the main constituents discussed in this report (ammonia, nitrate plus nitrite, organic nitrogen plus ammonia, orthophosphate, total phosphorus, and chlorophyll-a), detections in field blanks were no more than two times the laboratory reporting limits except for one detection of total phosphorus and two detections of chlorophyll-a (table 2.1). However, concentrations in the associated environmental samples (USGS, 2024a) were well above (more than five times) the amount detected in the blanks, indicating that the analytical results for the environmental samples are not affected by contamination. Detections of total nitrogen occurred in 4 of the 8 blanks indicating possible contamination. Therefore, the measured values in 2022 were not used in this report and instead total nitrogen for reservoir samples was computed as the sum of nitrite plus nitrate and organic nitrogen plus ammonia from unfiltered samples.

For the replicate samples, relative percent difference (RPD) was computed for each replicate pair as the absolute difference between replicate analyses divided by the average of the analyses and expressed as percent (table 2.2) (Mueller and others, 2015). The RPD was not computed if the pair included censored values. Most of the RPD values were less than 10 percent, indicating analytical results were reproducible for the constituents of interest. Values of RPD were most often above 10 percent for nutrient species (for example, total phosphorus and orthophosphate). Nutrient concentrations tended to be closer to the reporting levels, and their concentrations may have a higher degree of variability and uncertainty compared to other constituents (Mueller and others, 2015).

For the 2022 analytical results produced by the NWQL, a large percentage of nutrient analyses were flagged by the NWQL for hold-time exceedance owing to delays in analysis at the laboratory (Struzeski, 2025). Of the constituents discussed in this report, dissolved nitrate plus nitrite and total phosphorous concentrations were most affected by this laboratory issue. A recent hold-time study conducted by the NWQL determined that total phosphorus was stable for as many as 180 days and dissolved nitrite plus nitrate was stable for as many as 60 days (Struzeski, 2025). Because holding times were only greater than 60 days for 5 out of 233 analyses, hold-time exceedances likely had a minimal effect on the data interpretations presented in this report.

Algal taxonomy samples were analyzed using Imaging FlowCytobot (IFCB), which is a somewhat new semi-automated procedure (PhycoTech, Inc., 2024a). Seven sequential field replicates were collected and analyzed by IFCB, and results are summarized in table 2.3. The replicate pairs compared fairly well (median RPD of 23 percent) for taxa with high cell counts (above 1,000 cells per milliliter), but comparisons were poor for taxa with low cell counts. Although the variability for replicates was less than ideal, this report focuses on algal blooms when cell counts are high, and the IFCB reliably detected dominant taxa in replicate samples. A subset of algal taxonomy samples also was analyzed using traditional microscopy wherein phytoplankton are identified and counted manually using optically clear slide mounts to compare with IFCB counts, and results are presented in appendix 3.

Hydrology and Air Temperature

Hydrologic data for Blue Mesa were retrieved from the Reclamation website (Bureau of Reclamation, 2024) to characterize trends in select metrics from water years (WYs) 1970 through 2023 (a water year is defined as a 12-month period beginning October 1 and ending September 30 of the following year and is designated as the year in which it ends). Hydrologic data included daily reservoir level (in feet), storage volume (in acre-feet), total volume released (in acre-feet per day), and residence time (in days). Residence time is calculated as storage volume divided by total volume released. Mean, minimum, and maximum air temperature data for 1970–2023 were retrieved from the PRISM database (PRISM Climate Group, 2024), for three 2.5-mile grids, one centered across each of the three reservoir basins in Blue Mesa.

Trends in streamflow were analyzed for the Gunnison River site 13 from WY 1980 to 2023. This site only had streamflow data from 2018 to 2023; therefore, a record of daily streamflow was estimated by adding together daily streamflow records from the two upstream contributing streams: Gunnison River near Gunnison, Colo. (USGS site 09114500) and Tomichi Creek at Gunnison, Colo. (USGS site 09119000) (fig. 1; table 2; USGS, 2024a). For the 6 years with streamflow at all three streamgages, the sum of Gunnison River near Gunnison and Tomichi Creek at Gunnison was within 5 percent of the measured streamflow at Gunnison River site 13, indicating this was a reasonable approach for estimating the long-term daily streamflow record for the Gunnison River site 13.

Water-Quality Data

Trends were assessed in water temperature, Secchi disk depth, chlorophyll-a, total phosphorus concentrations, and select trophic state indices in the Iola, Cebolla, and Sapinero basins (sites 1–4), during the growing season (June–September) from WY 2001 to 2023. Trophic state indices based on Secchi disk (TSI–SD), chlorophyll-a (TSI–CHLa), and total phosphorus (TSI–TP) measurements were calculated using the following formulas from Carlson (1977):

Trophic state index transforms chlorophyll-a, Secchi disk depth, and total phosphorus measurements to a common scale ranging from 0 to 100, with higher TSI values indicating greater productivity.

Trends also were analyzed for 2001–23 for dissolved concentrations and loads of nitrite plus nitrate, total nitrogen, organic nitrogen plus ammonia, and total phosphorus at the Gunnison River site 13. Loads were computed by multiplying the concentration in milligrams per liter by the daily streamflow in cubic feet per second times a conversion factor of 5.39 to obtain loads in pounds per day. Trends in dissolved concentrations of nitrite plus nitrate and total phosphorus were analyzed at Beaver Creek site 14 and Steuben Creek site 15. The water-quality data analyzed included samples collected by USGS and NPS (table 2). The NPS utilized USGS collection and processing techniques, and samples were analyzed at the NWQL (Weissinger and Gibney, 2019). The consistency of data collection and analytical techniques justified merging USGS data with data collected by NPS.

Statistical Methods

Trends in hydrologic and air temperature data were computed using the nonparametric Mann-Kendall trend test (Helsel and others, 2020) using the “trend” R package (R Core Team, 2021; Pohlert, 2023). Trend slopes were calculated using the Sens slope estimate (Sen, 1968) and represent the median change per year. The statistical strength of the trends was assessed using the Mann-Kendall trend test (Helsel and others, 2020), and 90-percent confidence intervals were used to determine significance. These nonparametric statistical methods were chosen because the methods require no assumptions of sample distribution, trend shape, or data continuity when measuring the strength of trends. Trends were calculated for growing season means for select reservoir hydrologic data and air temperature, and for annual (WY) means for streamflow. Growing season includes the months of June to September when most water-quality sampling occurred in the reservoir and when cyanoHABs were most active.

Trends in water-quality constituents were analyzed using the nonparametric seasonal Kendall test with the “restrend” R package (Lorenz, 2014; R Core Team, 2021). Trends were computed for 2001–23 using four seasonal periods per year (June, July, August, and September) at the reservoir sites and Beaver Creek site 14 and Steuben Creek site 15, and six seasonal periods (January–February, March–April, May–June, July–August, September–October, and November–December) for the Gunnison River site 13. The magnitude of the trend slope was determined using the Theil-Sen slope estimator (Helsel and others, 2020). Trends were not computed for constituents with greater than 40 percent of the results below the laboratory reporting limit. Trends were considered strongly statistically significant for a probability value (p-value) less than or equal to 0.05 and weakly significant for a p-value greater than 0.05 and less than or equal to 0.10.

Reservoir Water-Level Elevation

Reservoir levels in Blue Mesa generally increase during spring and early summer in response to snowmelt runoff from the surrounding mountains, peak in June or July, and then decline through the late summer and fall months (fig. 5). In addition to the timing and amount of snowmelt runoff, reservoir operations also affect reservoir level, particularly during the summer (Bureau of Reclamation, 2012). Snowpack, represented by snow-water equivalent (table 3; Natural Resources Conservation Service, 2024) was above average in 2019 and 2023, and the resulting snowmelt runoff (“Annual inflow” in table 3; Natural Resources Conservation Service, 2024) supported high reservoir levels during the summer growing seasons in 2019 and 2023 (fig. 5). In contrast, reservoir levels during 2018, 2021, and 2022 were below the 30-year growing season mean of 7,492.8 ft (fig. 5). The April 1 snow-water equivalent for the upper Gunnison River watershed (table 3) indicates below mean snowpack occurred in 2018 (66 percent), 2021 (88 percent), and 2022 (95 percent), with 2022 being close to mean (Natural Resources Conservation Service, 2024). Gunnison County was in drought conditions for those 3 years (National Oceanic and Atmospheric Administration, 2024). Summer reservoir levels were low in 2018, 2021, and 2022 (fig. 5) in part because of lower snowpack and drought conditions, but also because reservoir outflows exceeded inflows in those 3 years (table 3). This result reflects reservoir operations because additional releases from Blue Mesa were necessary during late summer to provide water for downstream uses as well as “calls on the river” for Lake Powell, which began in 2021 (Bureau of Reclamation, 2021).

Algal Toxins

Algal toxin data were collected by the NPS during 2018–23 at shore locations in Blue Mesa that were easily accessible to the public. These data are presented to identify the years and time periods during which toxic cyanoHABs occurred in Blue Mesa (Qi and others, 2025). Microcystin was the dominant toxin in Blue Mesa and was detected in 2018 through 2023 (fig. 6) with a single detection of anatoxin-a in 2022. Most of the 91 microcystin detections were in Iola basin, although there were 11 detections in Cebolla basin, including 4 that exceeded the concentration of 8 µg/L in 2022 mainly at a site on the far east end of the basin. Microcystin concentrations in Iola and Cebolla basins occurred at concentrations of greater than 8 µg/L in 2018 and 2020–22, and at concentrations up to 8 µg/L in 2019 and 2023. The warning concentration of 8 µg/L for microcystin is the threshold concentration to protect human health and can be used as the basis for recreational contact (table 4; EPA, 2024a). The caution concentration is a lower-level recreation threshold and occurs from detection up to 8 µg/L for microcystin (EPA, 2024a). The duration of toxin detections above the warning concentration varied among years but typically started in early to mid-September and continued through the fall months, ending on September 26 in 2018, October 11 in 2020, September 14 in 2021, and November 1 in 2022 (fig. 6).

Algal Taxonomy

Phytoplankton abundance and community composition varied by time, basin, and depth (fig. 7). Phytoplankton abundance based on biovolume (cells per milliliter) was greatest in Iola basin and generally peaked in mid- to late summer (fig. 7). Cyanobacteria were the dominant taxa in Iola basin with Aphanizomenon being the most abundant although some early and late-season samples had greater proportions of Gloeotrichia, Microcystis, and Woronichinia (fig. 7). This result is generally consistent with results from Bauch and Malick (2003) that identified Aphanizomenon in samples composited across the 0–16-ft depth interval in August, October, and November 1999 in Iola basin, indicating that cyanobacteria have existed in Iola basin for many decades. Phytoplankton abundance was smaller in Cebolla and Sapinero basins compared to Iola basin (fig. 7), and the community composition in Cebolla and Sapinero basin was different, being mostly dominated by nontoxin-producing cyanobacteria, although toxin-producing cyanobacteria were dominant in some near-surface samples (surface and 3.3 ft) collected in late summer. Aphanizomenon generally was the most abundant toxin-producing cyanobacteria in Cebolla and Sapinero basins, followed by Dolichospermum and Woronichinia. Middle and deep samples mostly contained nontoxin-producing cyanobacteria taxa such as Aphanocapsa-Aphanothece and Merismopedia (Qi and others, 2025).

Some cyanobacteria, including Aphanizomenon, possess certain traits that allow them to dominate phytoplankton assemblages, which may partly explain the dominance of Aphanizomenon in Iola basin (Patiño and others, 2023). Along with Dolichospermum, Microcystis, and others, Aphanizomenon species regulate buoyancy allowing them to remain near the surface of the water where light is abundant and they dominate the phytoplankton assemblage (Walsby, 1994; Jöhnk and others, 2008; Paerl and Huisman, 2008; Yang and others, 2016). Furthermore, Aphanizomenon and Dolichospermum can fix nitrogen from the atmosphere, which allows them to thrive even in water bodies with low levels of inorganic nitrogen provided there is sufficient phosphorus (Chorus and Welker, 2021b).

Of the cyanobacteria taxa detected in Blue Mesa, Aphanizomenon, Dolichospermum, Gloeotrichia, Microcystis, and Woronichinia, are potentially capable of producing the toxin microcystin (Carey and others, 2012; Bernard and others, 2016). Aphanizomenon and Dolichospermum also can potentially produce anatoxin-a (Bernard and others, 2016), which was only detected once in Blue Mesa in 2022. Because Aphanizomenon is the dominant taxa in Blue Mesa, it is likely the main source of algal toxins detected in the reservoir.

Depth Profiles of the Reservoir

Depth profiles of discrete field water-quality measurements including water temperature, DO, and pH illustrate degree of stratification of Blue Mesa during the growing season and show distinct differences among the basins (note that in 2022, the Iola basin site was moved about 1 mile to the west from site 1 to site 2 [fig. 1] to a deeper part of the reservoir). Profiles generally show temperature maxima at the surface and decreasing temperature with depth (fig. 8A). Iola basin was the warmest at all depths, and Cebolla and Sapinero basins were the coolest. Because of its shallow depth, Iola basin showed minimal mid-summer temperature stratification, whereas Sapinero basin was strongly stratified by mid-summer, with temperature gradients of nearly 20 °C. Cebolla basin was also stratified, but the temperature gradients were not as steep or well defined as in Sapinero basin. Across all basins, water temperatures increased following spring runoff, peaked between mid-July and early August (fig. 8A), and declined during late summer and fall. Surface water temperatures in the reservoir in 2021 and 2022 were higher compared to other years, especially at Iola basin (site 1, 2021; site 2, 2022) where temperature at 3.3-foot depth reached 22.8 °C on July 15, 2021 (fig. 9A).

Dissolved oxygen profiles (fig. 8B) display several characteristics typical of productive (biologically active) water bodies. The DO patterns in Cebolla and Sapinero basins parallel the stratification observed in the water temperature profiles. At the start of each season, DO showed minimal variation with depth. As the seasons progressed, DO decreased with depth in Iola basin, and Cebolla and Sapinero basins developed a metalimnion (middle layer of a thermally stratified water column) DO minima, particularly in 2018. Cebolla and Sapinero basins, and to a lesser extent Iola basin, also had some oxygen depletion near the sediment water interface at the reservoir bottom. Microbial degradation of organic matter consumes oxygen and likely causes the DO minima in the metalimnion and hypolimnion (below the thermocline) (Marcé and others, 2024). Metalimnion oxygen minima dropped below the hypoxic threshold of 2 mg/L on at least one date in Cebolla basin every year (fig. 8B). Maximum DO occurred near the water surface in all basins, with the highest DO recorded in Iola basin in 2022 (DO greater than 11.5 mg/L in 2021 and 2022; fig. 9B). The elevated DO values in Iola basin indicate the water column was supersaturated with respect to oxygen. For example, at Iola basin (site 2) on September 8, 2022, DO values at the surface, 3.3-ft, and 9.8-ft depths were greater than 11 mg/L with saturation greater than 120 percent. These elevated, supersaturated DO values result from abundant algae and photosynthesis (Ignjatovic, 1968; Marcé and others, 2024). Photosynthesis (eq. 4) uses carbon dioxide (CO₂) plus water to create chemical energy from light energy and produces oxygen (O₂):

During photosynthesis (Marcé and others, 2024), the loss of CO₂ from the water column increases pH while the production of O₂ increases DO faster than it can equilibrate with the overlying atmosphere. Maximum DO values at 3.3-ft depth in Iola basin were generally greater than those in Cebolla and Sapinero basins, particularly from 2019 to 2022 (fig. 9B), and indicate the dominance of photosynthesis in Iola basin relative to Cebolla and Sapinero basins.

The pH in all three basins in Blue Mesa was generally greatest at the surface, decreasing with depth except for the last measurement period in Iola basin in 2019 when pH was almost constant with depth (fig. 8C). Near the surface, pH generally increased during the season, peaking usually after July to values as great about 9.5 in Iola basins in late August and early September 2019 (fig. 8C). In 2018, Iola basin had a different pattern with pH near the surface, peaking early in the season, decreasing, and then increasing toward the end of the season. Iola basin consistently had the greatest pH values each year (figs. 8C and 9C). The pattern of elevated pH at the surface with decreasing pH at depth in all three basins in Blue Mesa is typical of productive reservoirs where the increased pH at the surface is caused by photosynthesis that removes CO₂, a weak acid, thus increasing pH. With depth, organic degradation and respiration produce CO₂ and decrease pH (eq. 4) (Ibelings and others, 2021; Cole and Prairie, 2024; Marcé and others, 2024). The highest pH values in Iola basin are evidence that photosynthesis was more active in that basin than in Cebolla and Sapinero basins. Some cyanobacteria outcompete other algae at elevated pH levels because they are more effective at maintaining photosynthesis at elevated pH when other algae may become carbon-limited or otherwise stressed owing to the elevated pH (Glibert and Burkholder, 2018). Thus, it may be possible that the elevated pH during blooms in Blue Mesa and particularly in Iola basins is potentially exacerbating the dominance of HAB species.

In summary, water-quality depth profiles collected from 2018 to 2023 indicated that Iola basin behaved differently than Cebolla and Sapinero basins. Iola basin was shallower and sometimes weakly stratified, whereas the deeper Cebolla and Sapinero basins became stratified during summer with Sapinero basin generally exhibiting stronger and steeper metalimnion water temperature and DO gradients than Cebolla basin. Iola basin had greater maximum temperatures, pH, and DO concentrations near the surface than Cebolla and Sapinero basins, with the elevated pH and DO indicating enhanced photosynthesis in Iola basin.

Continuous Water-Quality Measurements

The continuous water-quality monitor data collected during July–September 2022 in Iola and Cebolla basins (fig. 1) were examined to help understand and potentially estimate the timing of algal blooms. One of the most evident patterns was diurnal fluctuations in DO and pH (fig. 10), where DO and pH reach maximum levels during daytime and minimum levels at night, indicative of active photosynthesis by algae and cyanobacteria at the Iola monitor. Photosynthesis (eq. 4) uses CO₂ to create chemical energy from light energy and produce oxygen. During photosynthesis, the loss of CO₂ from the water column increases pH via the carbonic acid buffer while the production of O₂ increases DO. At night, photosynthesis ceases, and cellular respiration dominates, which consumes oxygen and produces CO₂, decreasing the DO and pH of the water column (Marcé and others, 2024).

The continuous water-quality monitor data also were examined graphically for patterns that might help predict the timing of cyanoHABs. In Iola basin, discrete dissolved oxygen data and continuous monitor data were collected at the same Iola site (site 2). In Cebolla basin, discrete dissolved oxygen data were collected at site 3 and the continuous monitor was at site 25 (fig. 1). At the Iola monitor, the fPC signal was stronger than at the Cebolla monitor (fig. 11), which was parallel to the greater abundance of cyanobacteria in Iola basin (fig. 7). In addition, the fPC at the Iola monitor reached its maximum peak on August 28, 2022, about 2 weeks before toxin samples started to exceed the warning concentration for microcystin (September 14, 2022). Values of DO (fig. 11) and pH (fig. 8C) also peaked approximately 2 weeks prior to the detection of toxins greater than the warning concentration. The observed patterns in continuously monitored fPC, DO, and pH have potential to be used as an early warning monitoring strategy for future response to toxic cyanoHABs. Stackpoole and others (2024) also examined the utility of continuous chlorophyll measurements in freshwater as an early warning indicator of harmful algal blooms. More data collection of this type, along with collocated algal taxonomy and toxin sample collection, could help clarify whether continuous water-quality measurements could be used as an early indication of toxic cyanoHABs.

Spatial and Temporal Variability

Selected constituents from discrete samples collected at 3.3-ft depth are displayed graphically and discussed to potentially identify factors that may drive cyanoHABs and the release of algal toxins. These constituents are compared to water-quality standard concentrations and across reservoir basins and years to explore how conditions change from late spring into fall. Data presented include water temperature, chlorophyll-a concentrations, three trophic state indices (TSI–CHLa, TSI–SD, TSI–TP), and nutrient concentrations. The discussion incorporates variations in trophic status in each basin (Iola, Cebolla, and Sapinero) and for Blue Mesa. Trophic status is a measure of the biological productivity of a water body and is typically measured using chlorophyll-a concentrations with generally accepted divisions being oligotrophic (less than 2 µg/L), mesotrophic (2–7 µg/L), eutrophic (7–30 µg/L), and hypereutrophic (greater than 30 µg/L) (EPA, 2024b). The Colorado Department of Public Health and Environment (CDPHE) water-quality standard for chlorophyll-a is 8 µg/L for aquatic life and recreation (table 4). Trophic state indices (TSI) are another measure of biological productivity, and three independent estimates of biomass were calculated from chlorophyll-a, Secchi depth, and total phosphorus using equations 1–3 (Carlson, 1977). Productivity divisions for these metrics are oligotrophic (TSI less than 30), mesotrophic (TSI between 30 and 50), eutrophic (TSI between 50 and 70); and hypereutrophic (TSI greater than 70) (Carlson and Simpson, 1996; Dillon and Molot, 2024). The TSI–CHLa is considered a better predictor of trophic status than TSI–SD or TSI–TP (North American Lake Management Society, 2024), but relations between them can be used to understand different drivers for lake productivity (Carlson and Havens, 2005).

Water temperature in the three reservoir basins closely paralleled one another increasing through the season to maximum values in mid-July to late August (fig. 12A). Chlorophyll-a concentrations generally peaked near or after the temperature peaked (fig. 12B). On average, chlorophyll-a concentrations in Iola, Cebolla, and Sapinero basins indicated mesotrophic conditions in most years except 2021–22, when conditions were primarily eutrophic. It is worth noting that chlorophyll-a sampling in 2018 and 2020 did not extend to the period when algal toxins were detected in late summer. Chlorophyll-a concentrations in Iola basin were almost always greater than Cebolla and Sapinero basins (fig. 12B), exceeded 8 µg/L on at least one sampling date each year from 2018 to 2023, and reached hypereutrophic conditions in 2021 and 2022. Concentrations in Cebolla basin exceeded 8 µg/L and reached eutrophic conditions in 2021 and 2022 as did concentrations in Sapinero basin in 2019 and 2021, although concentrations in both basins were generally lower than in Iola basin. In 2021 and 2022, chlorophyll-a maxima generally coincided with algal biomass maxima at 3.3 ft in all three basins (fig. 7). In 4 of the 6 years shown in figure 12, algal toxins were detected in late summer and seemed to follow spikes in chlorophyll-a by several weeks particularly in 2021 and 2022 (years with higher sampling frequency). In years with late September sampling events, chlorophyll-a concentrations typically showed another increase following the August decline, which sometimes coincided with toxin detections in those years (fig. 12B).

Trophic state indices for chlorophyll-a, Secchi depth, and total phosphorus generally increased through the seasons in all three basins (fig. 12C, D, E). Decreases in TSI–SD and TSI–TP during 2019 and 2023 may indicate early season turbidity and sediment-phosphorus input from elevated runoff (increased inflow to Blue Mesa) those years (fig. 12D, E; table 3). Iola basin generally exhibited a higher trophic state and more productive conditions than Cebolla and Sapinero basins for most sampling dates and most years. In 2021 and 2022, conditions were generally eutrophic to hypereutrophic in Iola basin and mesotrophic to eutrophic in Cebolla and Sapinero basins (fig. 12C, D, E).

Bauch and Malick (2003) reported trophic state indices by the same three methods and reported values from 34 to 55 for Blue Mesa Reservoir during 1999. In this study, values varied from about 27 to 73 (fig. 12C, D, E) during 2018–23 indicating a wider variation in trophic states, and some instances of hypereutrophic conditions not previously reported. Previous work characterized the overall reservoir as mesotrophic (EPA, 1977; Blackwell and Boland, 1979; Long and others, 1996; Bauch and Malick, 2003). Consistent with these previous results, the median trophic state of Blue Mesa for chlorophyll-a, TSI–CHLa, TSI–SD, and TSI–TP was mesotrophic from 2018 to 2023 for each metric (fig. 12B–E).

Because the three index variables (TSI–CHLa, TSI–SD, and TSI–TP) are interrelated by linear regression models developed in Carlson (1977), they are expected to produce approximately the same index value for a given combination of chlorophyll-a, Secchi disk, and total phosphorus measurements. According to Carlson and Havens (2005), differences between the values for the same samples may provide additional insights into limnological processes, especially those not controlled by nutrients. The TSI–CHLa values for Blue Mesa are plotted against TSI–SD and TSI–TP in 2022 to illustrate how they deviate in time and space within Blue Mesa. Where TSI–SD is greater than TSI–CHLa (above the 1:1 line in fig. 13A), nonalgal particles may be limiting Secchi disk depths (table 1 in Carlson and Havens, 2005). The three highest values of TSI–SD occurred in Iola basin during May when the inflowing Gunnison River and Iola basin had higher levels of turbidity (USGS, 2024a), indicating suspended inorganic (nonalgal) sediment likely contributed to the Secchi disk reading in Iola basin during the spring months. When TSI–CHLa is greater than TSI–SD (below the 1:1 line in fig. 13A), large chlorophyll-containing particles, such as Aphanizomenon, which tends to aggregate into clumps, may be dominating light attenuation (Carlson and Havens, 2005), allowing more light penetration than from more dispersed algal particles. This result is indicated for most samples from Iola and Cebolla basins, especially surface samples (fig. 13A), and is consistent with the algal taxonomy showing a dominance of Aphanizomenon (fig. 7). When TSI–TP is greater than TSI–CHLa, there is excess phosphorus in the system, and some other factor may be limiting growth of algae (Carlson and Havens, 2005). Many of the samples with elevated TSI–TP relative to TSI–CHLa in figure 13B were collected at Iola basin in May or September 2022. A possible explanation is algal growth during the early and late part of the growing season may be limited by light availability and (or) cooler water temperature rather than phosphorus. Together, the information indicates that large chlorophyll-containing particles (likely from Aphanizomenon), rather than inorganic sediment, dominate light attenuation in Iola and Cebolla basins, and that something other than phosphorus limits algal growth in the early and late season.

Total phosphorus concentrations were generally greater in Iola basin than in other basins, except on a few occasions, and were greater than 0.021 mg/L (the CDPHE standard which equals 21 µg/L, table 4) at least twice each year (fig. 12F). There were no consistent seasonal patterns in total phosphorus concentrations among years. Concentrations increased with increasing chlorophyll-a in 2021 and 2022 (fig. 12B, F), and total phosphorus continued to increase following chlorophyll-a declines at 3.3-ft depth in late September 2022. Phosphorus concentrations in Iola basin were often greater than 0.021 mg/L (fig. 12F), with the highest observed concentrations in 2021 and 2022. Total phosphorus concentrations in Cebolla and Sapinero basins were greater than 0.021 mg/L at least twice from 2018 to 2023 with more exceedances in Cebolla basin than Sapinero basin.

Nutrient concentrations (total phosphorus, orthophosphate, and inorganic nitrogen) were examined with depth and among the three reservoir basins for surface, 3.3 ft, middle, and deep samples in each basin collected in 2021–22, when samples collected at multiple depths. Patterns in Cebolla and Sapinero basins were similar where total phosphorus, orthophosphate, and inorganic nitrogen (nitrate plus nitrite plus ammonia) increase with depth (fig. 14A–C). In contrast, patterns among constituents differ for Iola basin. Total phosphorus concentrations are similar at all depths, orthophosphate is greater at depth than the near surface or middle depths (somewhat like Cebolla and Sapinero basins), and inorganic nitrogen is similar at all depths having many values less than laboratory reporting limit (fig. 14A–C). Censored concentrations of nitrogen species were set at the laboratory reporting limit, which is indicated on figure 14. Depletion of nutrients at the surface is likely caused by phytoplankton usage and is consistent with previous patterns reported in Blue Mesa by Bauch and Malick (2003). Depth enrichment is typical for stratified lakes (Dillon and Molot, 2024). Total phosphorus is enriched in Iola basin at almost all depths compared to Cebolla and Sapinero basins, likely reflecting phosphorus inputs from the Gunnison River into Iola basin and progressive removal by algae moving downstream through the reservoir. In contrast, orthophosphate and inorganic nitrogen concentrations were lower in Iola basin, likely because of their utilization by the more abundant algal biomass in Iola basin (fig. 7).

The total nitrogen to total phosphorus ratio (TN:TP) can be used to understand whether nitrogen or phosphorus is a limiting nutrient for productivity in aquatic systems (Bergström, 2010). The TN:TP mass ratios were calculated using Iola basin samples collected at all depths in 2022 (fig. 15), which was the year that exhibited the most severe algal bloom of the study period (fig. 6). It is notable that the largest fraction of the total nitrogen concentration was organic nitrogen because inorganic nitrogen species (ammonia and nitrate) were often below the laboratory reporting level (USGS, 2024a). Thus, TN:TP ratios were larger than they would be if only total inorganic nitrogen was included. Nonetheless, the 2022 data show TN:TP ratios generally less than 19, indicating that Iola basin was almost always nitrogen limited during summer of 2022 according to the predicted thresholds from Bergström (2010). Low TN:TP ratios are characteristic of conditions that allow cyanobacteria to dominate (Smith, 1983) because cyanobacteria can fix atmospheric nitrogen (Beversdorf and others, 2013; Willis and others, 2016).

In summary, chlorophyll-a data and TSI values indicated increased trophic status of Iola basin during 2018–23 relative to previous reports, whereas Blue Mesa remained overall mesotrophic consistent with previous reports. In Iola basin, the longest periods of toxicity warnings from cyanoHABs blooms and the lowest recorded reservoir level since 1984 occurred during 2021–22 (Bureau of Reclamation, 2024). During this time, chlorophyll-a and TSI–TP indicated hypereutrophic conditions at some points in each year. The relation between TSI–CHLa and TSI–SD values and the low TN:TP ratio indicate that the reservoir is nitrogen limited and that turbidity is primarily from algae and cyanobacteria, rather than suspended sediment. Total phosphorus is enriched in Iola basin at almost all depths compared to Cebolla and Sapinero basins. In contrast, orthophosphate and inorganic nitrogen concentrations were lower in Iola basin, likely because of their utilization by the more abundant algal biomass in Iola basin. The TN:TP ratios are consistent with conditions that allow cyanobacteria to dominate. Nitrogen limitation is consistent with the ability of cyanobacteria to fix nitrogen from the atmosphere and thus the cyanobacteria do not require exogenous nitrogen. The relation between TSI–CHLa and TSI–TP during the early and late part of the growing season indicates that something other than nutrients (possibly water temperature or light availability) limits algal growth in Blue Mesa during these times of the year.

Reservoir Hydrology, Air Temperature, and Streamflow

Change in climate has emerged as a potentially important factor affecting the frequency and severity of algal blooms in lakes and reservoirs (Paerl and Huisman, 2009; Ho and Michalak, 2020). In the Colorado River watershed, there have been recent periods of extended drought (Udall and Overpeck, 2017) and unprecedented increases in air temperatures (Milly and Dunne, 2020) with model-based analyses indicating that warming and reduced streamflows will continue in the region (Udall and Overpeck, 2017). In anticipation of changes in regional climate and in particular reduced wintertime precipitation (Woodhouse and others, 2021), management strategies in western reservoirs are adapting to best maintain or augment current water resource availability (Bureau of Reclamation, 2025). In Blue Mesa, the reservoir level is predominately controlled by snowmelt runoff and releases of water related to hydropower generation and downstream water needs (Bureau of Reclamation, 2012). To evaluate long-term changes in Blue Mesa that may be driven by climate or reservoir management, trends in reservoir hydrologic conditions, including reservoir level, release volume, residence time, streamflow inputs from the Gunnison River (the primary tributary), and air temperature were explored. Trends in growing season (June–September) and annual means are presented graphically (fig. 16).

No trend was detected in the growing season mean for reservoir level during 1970–2023 (fig. 16A). Mean reservoir levels during the growing season were below the long-term mean in 4 out of the last 10 years (2014–23) and included the 2 lowest years since 1970 when mean reservoir levels were 42 ft (2021) and 39 ft (2022) below mean reservoir level (7,493 ft), and 69 ft and 66 ft below the maximum reservoir level of 7,519 ft. A statistically significant upward trend in mean release volumes of 16.6 acre-feet per day per year was detected during the growing season from 1970 to 2023 (fig. 16B). As a result, a downward trend in growing season residence time of 1.75 days per year was detected from 1970 to 2023 (fig. 16C). Decreased residence time (in the range of days) may dilute cyanobacteria faster than they can multiply and (or) alter other growth conditions like nutrient concentrations (Ibelings and others, 2021); however, it is not known how changes in Blue Mesa residence time may affect cyanobacteria growth. Residence times in Blue Mesa are long (200 days) for the whole reservoir (Bureau of Reclamation, 2024) but they may be considerably shorter for the smaller and shallower Iola basin.

Statistically significant upward trends in air temperature were observed for growing season mean temperatures (0.05 °C per year) from 1970 to 2023. For the Gunnison River site 13, statistically significant downward trends were detected in annual streamflow (31 percent during the 43-year period or −0.0087 cubic foot per second [ft³/s] per year from 1980 to 2023 [fig. 16E]). Because the Gunnison River is unregulated upstream from the reservoir, these declines in streamflow could reflect decreases in precipitation and (or) changes in snowmelt timing in addition to air temperature increases and withdrawals for agricultural, municipal, and industrial uses.

Water Quality and Trophic Status

In addition to the long-term hydrologic datasets, water-quality data have been collected at Blue Mesa during the growing season since 2001 (table 2). The seasonal Kendall test was used to evaluate trends in Secchi disk depth, water temperature, chlorophyll-a, total phosphorus concentrations, and the three trophic state indices (TSI–SD, TSI–CHLa, and TSI–TP) for the period 2001–23 at 3.3-ft depth at reservoir sites, and results are summarized in table 5. Statistically significant trends (p-value less than 0.05) were detected for water temperature at Cebolla basin and for chlorophyll-a and TSI–CHLa in all three reservoir basins. Slightly less significant trends (p-value between 0.05 and 0.10) were also detected for water temperature in Iola and Sapinero basins and total phosphorus and TSI–TP in Iola and Cebolla basins. The chlorophyll-a trends were upward in all three basins and were greatest in Iola basin (0.218 µg/L per year) and less than half that in Cebolla (0.106 µg/L per year) and Sapinero (0.069 µg/L per year) basins (table 5). Despite the upward trends in chlorophyll-a, Secchi disk depth showed no statistically significant trends from 2001 to 2023 although the slopes were all negative. This result indicates that the chlorophyll-a trends, less than 5 µg/L across the 23-year period, did not substantially affect reservoir clarity as measured with Secchi disk readings. Significant upward trends in TSI–CHLa values indicate that all basins are shifting to higher trophic states through time (table 5), corresponding to a change in trophic status of 10–15 units across the 23-year period of record.

Surface water temperatures during the growing season showed statistically significant upward trends in Cebolla basin (0.031 °C per year) with less significant upward trends in Iola (0.024 °C per year) and Sapinero (0.030 °C per year) basins. These increases are similar to those derived from the 2000–23 satellite modelling (King and others, 2025) of 0.04, 0.03, and 0.02 °C per year for Iola, Cebolla, and Sapinero basins, respectively. These results are also consistent with the trend of increasing growing-season air temperatures of 0.05 °C per year from 1970 to 2023 reported in the “Long-term trends in hydrologic and limnological conditions” section. Together, these different temperature records provide evidence that water temperatures in Blue Mesa have risen during the past several decades.

From 2001 to 2023, there were weak (p-value greater than 0.05 and less than or equal to 0.10) upward trends in total phosphorus and TSI–TP in Iola and Cebolla basins. The magnitude of the trends is relatively small with Iola basin showing a 0.0092-mg/L increase (0.0004 mg/L slope per year for 23 years) and Cebolla basin showing a 0.0046-mg/L increase (0.0002 mg/L slope per year for 23 years). Although the increase in total phosphorus could reflect real environmental change, these detected changes are small compared to the laboratory reporting limit of 0.01 mg/L.

Nutrient Concentrations and Trends

Nutrient concentrations in 2022 are presented for Iola basin tributaries including the Gunnison River site 13, Beaver Creek 14, and Steuben Creek site 15. Data for Beaver Creek site 14 and Steuben Creek site 15 were collected as part of this study (table 1.3). Data for Gunnison River site 13 were collected as part of routine monitoring by USGS and were retrieved from NWIS database (USGS, 2024a) (refer to appendix 3 for discussion of 2022 grab samples collected at Gunnison River site 13 during this study). Inorganic nitrogen was detected in all samples from the Gunnison River site 13, and concentrations ranged from 0.01 to 0.05 mg/L (USGS, 2024a). Inorganic nitrogen (nitrate plus nitrite) was below laboratory reporting limits (0.01 mg/L) for all Steuben Creek site 15 samples and all but two Beaver Creek site 14 samples collected between July and October 2022. Thus, inorganic nitrogen concentrations generally were low or not detected in tributaries. However, a comprehensive nitrogen load analysis for all reservoir inputs and output, though beyond the scope of this study, would assist better understanding of the magnitude of tributary nitrogen input to Iola basin.

All 2022 results for total phosphorus in these tributaries were greater than the laboratory reporting limit of 0.01 mg/L (USGS, 2024a). In general, concentrations were slightly higher at Beaver Creek site 14 (mean of 0.14 mg/L) and Steuben Creek site 15 (mean of 0.086 mg/L) compared to the Gunnison River site 13 (mean of 0.034 mg/L). Two of 14 grab samples collected in 2022 at Steuben Creek site 15 and 13 of 14 grab samples collected in 2022 at Beaver Creek site 14 had phosphorus concentrations greater than the State of Colorado interim standard for phosphorus in rivers of 0.11 mg/L (or 110 µg/L, table 4; CDPHE, 2024a). None of the 2022 samples collected at the Gunnison River site 13 were greater than 0.11 mg/L. Additional discussion of phosphorus concentrations in tributaries is discussed in the “Geologic Sources of Phosphorus” section.

Changes in land use, streamflow, and watershed management can contribute to changes in nutrient concentrations and loads through time (Murphy and Sprague, 2019; Stets and others, 2020). Identifying the causes of nutrient trends can aid land managers in designing nutrient-control strategies and can serve as ongoing measures of progress toward nutrient reduction. Trend analysis was used to assess temporal changes in nitrogen and phosphorus concentrations in tributaries to Iola basin because Iola basin is primarily where toxic cyanoHABs are present (fig. 7), and increasing nutrient input can make water-quality conditions more favorable to algal growth. Trends for the Gunnison River site 13, which is downstream from a wastewater treatment plant, were computed for 2001–23 (table 5) and showed no statistically significant trends at the p-value less than 0.05 for either concentration or load. For nitrite plus nitrate and total nitrogen, less significant downward trends were detected (p-value between 0.05 and 0.10; table 5), indicating weak downward trends in inorganic nitrogen at this site. For Beaver Creek site 14 and Steuben Creek site 15, no statistically significant trends were detected in total phosphorus or organic nitrogen plus ammonia concentrations. Trends could not be computed for nitrite plus nitrate because more than 80 percent of concentrations were below the laboratory reporting limit, and total nitrogen at these sites is not available (table 5). These results likely indicate nitrogen and phosphorus inputs from major tributaries entering the Iola basin have not changed substantially during the past several decades and may have decreased slightly, indicating they are likely not the main trigger for algal blooms in this part of the reservoir. These results are consistent with previous work that did not detect phosphorus trends in Beaver Creek site 14, Steuben Creek site 15, or the Gunnison River site 13 for 2010–18 (Weissinger and Gibney, 2019). Weissinger and Gibney (2019) also did not detect trends in total ammonia and nitrite plus nitrate at Gunnison River site 13 and were unable to test for trends in nitrogen species at Beaver Creek site 14 and Steuben Creek 15 because the data were highly censored. However, Weissinger and Gibney (2019) indicated that Beaver Creek site 14 and Steuben Creek site 15 exceeded water-quality standards for total phosphorus.

Geologic Sources of Phosphorus

The elevated total phosphorus concentrations in Beaver and Steuben Creeks were noteworthy (fig. 17) and consistent with previous observations (Weissinger and Gibney, 2019). This observation was investigated further by compiling total phosphorus data collected by the NPS (and including data from this study) from 2021 to 2023 for additional tributaries flowing into Blue Mesa (Weissinger and Gibney, 2019; National Water Quality Monitoring Council, 2024). Phosphorus concentrations in most tributaries were greater than in any of the three reservoir basins and the Gunnison River site 13 (fig. 17). Total phosphorus was highest in tributaries draining the north side of the Blue Mesa particularly those located north of Iola and Cebolla basins (Beaver, Steuben, East Elk, and Red Creeks, fig. 1). Total phosphorus in tributaries, especially in East Elk and Red Creeks, frequently exceeded 0.11 mg/L (or 110 µg/L, the State of Colorado interim standard for phosphorus in rivers, table 4; CDPHE, 2024a). These two creeks are on the monitoring and evaluation list for total phosphorus in the most recent 303(d) list for Colorado (CDPHE, 2023). Previous work noted elevated phosphorus concentrations in Elk and Red Creeks and proposed that these streams flowed through volcanic rocks that might be a source for phosphorus, but no supporting evidence was provided (Brown and Thoma, 2012). The geologic map indicates that streams on the north shore of Blue Mesa drain areas that are primarily underlain by Tertiary volcanic rocks and that have no agricultural land use (fig. 4). The mean phosphorus concentration estimated for volcanic rocks of 0.11 weight percent is nearly twice that of the crustal abundance of 0.0655 weight percent for phosphorus (table 6; USGS, 2024b), indicating the volcanic rocks are enriched in phosphorus and could be a potential source to streams that drain them. The alkali intrusions on the south side of Blue Mesa (fig. 4) are also highly enriched in phosphorus having a phosphorus concentrations more than 10 times the crustal abundance (Cambrian alkalic intrusive rocks 0.99 weight percent, table 6). Total phosphorus concentrations at the Cebolla Creek at Powderhorn, Colo. (USGS site 381633107054700; 15 mi upstream from Blue Mesa, not shown on fig. 1; USGS, 2024a), which flows into Blue Mesa and drains the alkali intrusions, are elevated and concentrations are greater than the CDPHE stream standard concentration (fig. 17), likely indicating this rock type also contributes phosphorus to surface water.

Geologically derived nutrients are often an overlooked source of nutrients to surface water, but a recent study in California attributed elevated concentrations of phosphorus in surface water to weathering of phosphate minerals in volcanic rocks (Deas and others, 2024). They suggested phosphorus was released from weathering of apatite but also volcanic glass, which weathers rapidly and can contain considerable amounts of phosphorus. Given the lack of agricultural development and the dominance of intrusive and volcanic rocks in the lower Blue Mesa watershed, some of which contain volcanic glass, it seems likely that phosphorous from weathering of bedrock is a potential source of elevated phosphorus in tributaries (fig. 17) that then most likely contributes phosphorus to the reservoir. This geologic source supporting elevated phosphorus concentrations in streams has probably been generally consistent through time and is likely not a trigger for the 2018–23 cyanoHABs in Blue Mesa. However, it nonetheless is an important source of phosphorus supporting algal growth in the reservoir.

Water Quality of Seeps Along the North Shore of Iola Basin

When reservoir levels were low in Blue Mesa in 2021 and 2022, several seeps appeared in the exposed reservoir bed on the north shore of Iola basin (sites 21–24 in fig. 1 and table 1). Sites 22–24 are within the drainage channel of a small tributary, North Willow Creek, where a seasonal recreational vehicle (RV) park (fig. 1) with domestic wastewater treatment facilities is located. In 2021, North Willow Seep at Blue Mesa Reservoir near Gunnison, Colorado (USGS site 382834107063501, hereafter referred to as “site 24”) had elevated concentrations of dissolved nitrogen (25.8 mg/L as nitrogen) and total phosphorus (2.48 mg/L as phosphorus, sampled September 24, 2021) (table 7) that were 1–2 orders of magnitude greater than CPDHE stream and lake water-quality standard concentrations of 0.380 mg/L for total nitrogen for lakes and 0.11 mg/L and 0.021 mg/L total phosphorus for streams and lakes, respectively (table 4). Concentrations at site 24 in 2022 were lower than in 2021 by a factor of 3 but were still greater than the water-quality standard concentrations. In general, nutrient concentrations at site 24 were much higher than at Old Stevens Seep near Gunnison, Colo. (USGS site 38285810705590, hereafter referred to as “site 21”), which was sampled as a background site in 2022 (fig. 1). The elevated nutrient data, especially the high ammonia concentration (21.3 mg/L as nitrogen) collected at site 24 in September 2021, along with a field comment noting a distinct waste-like odor when sampling in 2022 (USGS, 2024a) potentially indicate a wastewater source for water in site 24. North Willow Seep in North Willow Channel near Gunnison, Colorado (USGS site 382840107063501, hereafter referred to as “site 23”) and Middle North Willow Seep near Gunnison, Colorado (USGS site 382838107063601, hereafter referred to as “site 22”) also had elevated total nitrogen (from 0.51 to 1.1 mg/L) and total phosphorus (from 0.217 to 1.22 mg/L) concentrations relative to the CPDHE stream and lake water-quality standard concentrations of 0.380 mg/L for total nitrogen for lakes and 0.11 mg/L and 0.021 mg/L total phosphorus for streams and lakes (table 4). However, the total nitrogen concentrations at each site may have been affected by contamination (refer to “Quality Assurance” section).

Pharmaceutical compounds, nitrogen isotopes in ammonia, and concentrations of boron and lithium also were analyzed as potential wastewater indicators (Vengosh and others, 1994; Furlong and others, 2017). Three pharmaceutical compounds were detected at site 24 in 2022 (table 7) including erythromycin (antibiotic), triamterene (diuretic), and glipizide (hypoglycemic agent) (Furlong and others, 2017). Site 23 (just upstream from site 24, was sampled in 2022 before reservoir levels had fallen enough to reveal site 24) had an estimated detection (detected but concentration below lowest calibration standard; table 7) of one wastewater indicator, desmethyl diltiazem (a metabolite of a blood pressure medication) (Furlong and others, 2017). Research indicates that the presence of low levels of human pharmaceuticals in groundwater with or without elevated nitrate concentrations is unambiguous evidence that domestic wastewater contributed to the groundwater (Seiler and others, 2005).

The 2022 nitrogen isotope values for site 22 and site 24 (table 7) are within the range for septic waste and natural soils (refer to fig. 16.4 in Kendall, 1998), and thus not uniquely diagnostic of wastewater. Boron is a bleaching agent in detergents and is enriched in wastewater (Vengosh and others, 1994). Lithium is used to treat bipolar disorder (Furlong and others, 2017) and tends to remain in solution (Hem, 1985) and through sewage treatment processes (Furlong and others, 2017). Boron and lithium were detected in the seeps but at relatively low concentrations (table 7). Lithium concentrations were lower in sites 22–24 than in site 21 (background), and boron in site 23 was similar to site 21, but site 24 was higher than site 21 (table 7). Because concentrations were low overall with no consistent patterns among the sites, boron and lithium do not provide a clear signal of wastewater discharge and more likely are naturally occurring and potentially sourced from weathering of volcanic rocks as described for phosphorus in the “Geologic Sources of Phosphorus” section. Lithium data compiled for rock samples in the Blue Mesa watershed (fig. 4, table 6) indicated lithium concentrations in the volcanic rocks were similar to crustal abundances, indicating bedrock weathering could be a potential source of lithium.

Although some seeps had very high nutrient concentrations, it is worth noting that the discharge from these seeps was very low, and the load of nutrients into the reservoir from these sources is likely minor. Assuming a hypothetical discharge of 0.01 ft³/s at site 24 in September 2021, the estimated load of total phosphorus would be 0.13 pound per day (lb/day) and total nitrogen load would be 1.39 lb/day. By comparison, at the Gunnison River site 13 in August 2021, the streamflow of 635 ft³/s and total phosphorus and nitrogen concentrations of 0.030 and 0.209 mg/L (USGS, 2024a), yield loads of 101 lb/day for total phosphorus and 717 lb/day for total nitrogen, which are more than two orders of magnitude higher than the estimated seep loads.

In summary, the nitrogen isotopes in ammonia and concentrations of lithium and boron were not unambiguously diagnostic of wastewater. However, the occurrence of elevated nutrient concentrations and detection of a few pharmaceuticals provides evidence for a wastewater source to sites 23 and 24, potentially a point source like the upstream RV park (which was not in compliance with county wastewater regulations in 2023 [permit number COX634063, document identification number 20580845 and 20580890; CDPHE, 2024b]). Although it is unlikely the seeps contribute enough nutrients to affect development of algal blooms throughout the reservoir, there might be highly localized blooms occurring in the vicinity of these sites (for example, fig. 2B) and controlling nutrient sources such as these could be considered part of a management strategy.