Degree Type

Thesis

Date of Award

2019

Degree Name

Master of Science

Department

Ecology, Evolution, and Organismal Biology

Major

Environmental Science

First Advisor

Grace M. Wilkinson

Abstract

Freshwater harmful algal blooms (HABs) are increasing in frequency, scope, duration, and intensity across the globe. Aquatic ecosystems are negatively affected by HABs, which endanger ecosystem health, human health, and ecosystem services. As such, being able to predict when HABs are going to occur is necessary to prevent or mitigate the negative effects of these events. Freshwater algal blooms are considered a regime shift from clear-water to an algae dominated state. High frequency data simulation and high frequency monitoring of lakes under experimental conditions suggest that statistical early warning indicators (EWIs) such as a rise in autocorrelation (AC) and variance, which can also be measured as the standard deviation (SD) occur before a regime shift. Chapter 1 presents a post hoc analysis of temporal EWI detection of the regime shift from a clear state to an algal-dominated state in four shallow, hypereutrophic non-experimental lakes. We evaluated EWIs for four state variables (chlorophyll a, phycocyanin, dissolved oxygen, and pH) and two statistical indicators (rises in AC and SD). In addition to our investigation of temporal EWIs in impaired shallow lakes, Chapter 2 explores spatial patterns of state variables prior to, during, and after a regime shift has occurred in a lake. In order to describe these spatial patterns, we sampled state variables weekly across a shallow hypereutrophic lake and examined spatial AC and SD over time.

The four study lakes monitored for Chapter 1 vary in depth, area, shoreline development index, and nutrient concentrations. These ecosystems are in many ways the opposite of the lakes where previous EWI studies have been performed. Despite confounding factors in shallow hypereutrophic lakes such as high and stochastic nutrient loading, a preponderance of primary producers besides phytoplankton, and spatial complexity, we were still able to frequently detect temporal EWIs. We detected 86% of the total possible EWIs signals in our five lake year dataset, with an average of two weeks of warning before bloom conditions appeared. Detection of EWIs was even higher prior to the peak of the bloom, with 96% of the total possible EWIs detected before peak biomass and an average of almost four weeks of warning before bloom climaxes.

Chapter 2 focuses on a shallow, hypereutrophic Swan Lake, we sampled across the entire lake to document patterns of spatial heterogeneity in state variables before, during, and after blooms. There were consistent north-south gradients in the state variable across the lake before and after the bloom and similar variable values across the lake during bloom conditions. We detected rises in spatial AC for all state variables during both bloom events, except phycocyanin which rose ten days before the beginning of the first bloom event. Phycocyanin, dissolved oxygen, and pH rose in spatial SD before bloom conditions and chlorophyll a rose during the bloom. However, for the second bloom event chlorophyll a rose four weeks in advance and phycocyanin spatial SD rose during bloom conditions. Dissolved oxygen and pH spatial SD did not respond as expected to the second bloom and failing to indicate a regime shift due to contributions from other sources of net ecosystem production besides algae.

In summary, this is the first application of temporal and spatial indicators of a regime shift in shallow, hypereutrophic lakes. These studies provide evidence of the possible multi-pronged approach to detect early warnings to impending blooms using temporal and spatial monitoring. These studies are promising first steps in applying early warning indicator detection in impaired shallow non-experimental lakes. Our studies applied theory derived from high frequency and spatially explicit simulations and verified results from lakes under experimental conditions, by evaluating EWI detection in non-experimental lakes using both space and time monitoring methods. High frequency and spatial monitoring are proving to be efficient, reliable tools in the effort to manage HABs in surface waters.

Copyright Owner

David Adrian Ortiz

Language

en

File Format

application/pdf

File Size

69 pages

Share

COinS