Environmental factors regulating toxic blooms of golden alga (Prymnesium parvum) and their effects on fisheries resources

The chemical, physical, and biological properties of water, collectively known as water quality, are critical aspects of aquatic habitats and are necessary considerations for anthropogenic water supplies. Water quality impairment is a major cause of aquatic ecosystem degradation. For example, nutrient enrichment can lead to the formation of noxious algal blooms, diminished habitat for native species, reduced biodiversity, and economic losses. Perturbations to water quality can also provide a pathway for invasive species, which may further impair freshwater resources. Some algal blooms can have negative effects on aquatic ecosystems, impact human health, or cause socioeconomic losses; such blooms are called “harmful algal blooms” or HABs. Toxic blooms of the golden alga (Prymnesium parvum) have caused substantial ecological and economic harm in inland and marine systems throughout the world. Golden alga has invaded North America, where its negative effects have been most apparent in the state of Texas. Golden alga-related fish kills have claimed millions of fish worth several million dollars, and have also impacted the recreational value of affected systems. Habitats affected include hatcheries, small ponds, large reservoirs, and numerous river stretches located in five major river basins. Whereas golden alga has been successfully controlled in small systems (e.g. hatcheries), mitigation has been a major challenge in reservoirs because of the logistic and economic difficulties of management at such a large scale. Alternative strategies for golden alga control in reservoirs are therefore needed. One proposed strategy is to stock species that are resistant to golden alga toxins or that can quickly recover. Information about species-specific tolerance to or recovery from toxic golden alga blooms, however, is relatively limited. Another potential strategy for control in large reservoirs is the management of land and/or water uses in a manner that creates environmental conditions unfavorable for golden alga growth and/or toxicity. Because the ion composition and nutrient loading of surface waters is highly influenced by underlying geology and land use within the basin, it is difficult to determine whether established associations between environmental conditions and golden alga are system-specific or can be applied at broader spatial scales. The overall goal of this dissertation is to determine the environmental factors that regulate or influence golden alga presence, bloom formation, and toxicity in surface waters, and the long-term impacts of toxic blooms on fish populations. The immediate purpose of this research is to provide information to resource managers that will assist in the design of new or better strategies to control or mitigate golden alga. Several reservoirs in the Upper Colorado River (UCR) and Brazos River basins in Texas have experienced toxic blooms of golden alga and associated fish kills since 2001. I used Multiple Before-After-Control-Impact analysis on 21 years of fish population data to determine whether repeated toxic blooms in these basins have led to declines in the relative abundance and size structure of fish populations (Chapter 2). Sustained declines were noted for nine of twelve fish species surveyed in the UCR, whereas only one of eight species was impacted by golden alga in the Brazos River. Common carp (Cyprinus carpio) and longnose gar (Lepisosteus osseus) appear to be relatively resistant to golden alga toxins, whereas gizzard shad (Dorosoma cepedianum) and white crappie (Pomoxis annularis) appear to be relatively resilient. Overall, toxic golden alga blooms negatively impacted fish populations over the long term, but patterns of impact varied considerably between river basins and among species (Chapter 2). Because of the much greater impacts on fish populations in the UCR relative to the Brazos River, environmental conditions in the UCR may be relatively favorable for golden alga. The UCR is therefore an ideal location to assess the relationships between water quality characteristics and golden alga presence, abundance, and toxicity. Golden alga abundance (hemocytometer counts), ichthyotoxicity (bioassay), and water quality (surface grab samples) were quantified in three golden alga-impacted reservoirs from the Colorado River or its tributaries; two reference reservoirs located on the Concho River, which has no history of toxic golden alga blooms; and three sites at the confluence of the Colorado and Concho rivers (Chapters 3 and 4). Sampling for basic water quality variables (temperature, dissolved oxygen, specific conductance, pH, oxidation-reduction potential, hardness, fluoride) occurred monthly from January 2010 to July 2011 (Chapter 3). Sampling of nutrient variables (orthophosphate, total phosphorus, nitrate+nitrite, total Kjeldahl nitrogen, ammonia) occurred monthly from December 2010 to July 2011 (Chapter 4). Golden alga abundance and toxicity were highest at high salinity (≥ 3761 µS/cm), low temperature (< 21.7°C), high orthophosphate (≥ 0.00229 mg/L), and under nitrogen-limited conditions (nitrate < 0.0529 mg/L, DIN:DIP < 20.8). Impacted sites were characterized by higher specific conductance (salinity), hardness, fluoride, and organic nitrogen and lower inorganic nitrogen relative to reference and confluence sites. Within impacted reservoirs, bloom termination coincided with increases in inorganic nitrogen and DIN:DIP in late-spring (Chapter 4); golden alga abundance and ichthyotoxicity were also strongly associated with water temperature. Specifically, blooms peaked at ~10°C and generally did not occur above ~20°C (Chapter 3). In spring 2013, golden alga was absent in E. V. Spence, Lake Colorado City, and Moss Creek City Reservoir despite being present at these reservoirs in low-to-high abundance every spring over an eleven-year period between 2002 and 2012 (Chapter 6). Basic water quality and nutrient characteristics were quantified in April 2013 and compared to results from earlier studies (Chapters 3 and 4) and interpreted in the context of a major inflow event that occurred in September 2012 (Chapter 6). Following the inflow, salinity in E. V. Spence Reservoir declined ~80% to levels previously associated with low abundance and toxicity; salinity in Lake Colorado City and Moss Creek City Reservoir, however, remained at high levels. High nitrate and high ratios of inorganic nitrogen to inorganic phosphorus at Lake Colorado City and E. V. Spence Reservoir were outside the range that favors toxicity, whereas nutrients were favorable at Moss Creek City (Chapter 6). Favorable salinity and nutrients but no bloom at Moss Creek City suggests other factors, perhaps the high turbidity observed at this site, may have inhibited golden alga growth. Overall, management of land and water use to reduce salinity or modify nutrient profiles could produce unfavorable conditions for golden alga. Some studies have indicated that, in golden alga cultures, toxicity is poorly correlated with the abundance of golden alga cells. This relationship, however, has not been fully evaluated in the field where the mixture of toxic compounds produced by golden alga may differ from those produced in the laboratory. If a strong relationship can be established between golden alga abundance and toxicity, this would allow managers rapid assessments of the potential for ichthyotoxicity without bioassay confirmation, which requires additional resources to accomplish. I therefore evaluated the influence of water quality, reservoir characteristics, and weather variables on the relationship between golden alga abundance and ichthyotoxicity (Chapter 5). Several a priori models relating lethal levels of ichthyotoxicity to abundance and environmental covariates were constructed then estimated using archived data from four river basins in Texas and New Mexico (Colorado, Brazos, Red, Pecos). Overall, golden alga abundance was a generally good predictor of ichthyotoxicity as cross validation of abundance-only models ranged from ~80% to ~90% (Chapter 5). Environmental covariates slightly improved predictions; specific conductance, temperature, and pH were generally among the top-ranked environmental variables, as determined by Akaike Information Criterion (AIC) values and Akaike weights, but top variables differed among the four basins. These associations may be useful for monitoring as well as understanding the abiotic factors that influence toxicity during blooms (Chapter 5). In conclusion, fish kills associated with toxic golden alga blooms can lead to sustained declines in fish populations, but impacts differed between the UCR and Brazos basins and among fish species. Favorable reservoir golden alga habitat had high levels of salinity and associated ions (calcium, magnesium, fluoride), high organic nitrogen, and was nitrogen-limited (low nitrate, low DIN:DIP). Inflows can create unfavorable golden alga habitat, but are irregular and of low magnitude in the UCR. Golden alga abundance was a generally good predictor of toxicity, and inclusion of environmental covariates slightly improved predictions. Strategies aimed at reducing favorable golden alga habitat, such as reducing salinity by increasing flows in streams feeding reservoirs or altering nutrient conditions by dredging, probiotic application, or forced vertical mixing, could provide long-term benefits to impacted systems.

This dissertation won 2nd Place in the Texas Tech University Outstanding Thesis and Dissertation Award, Biological Life Sciences, 2015.

Embargo status: Restricted from online display. To request an access exception from the author, please click on the PDF to the left.