Remotely Sensed Early Warning of Algal Blooms in an Eastern Nebraska Reservoir: A Comparison of Temporal and Spatial Indicators

Authors

Mercy Kipenda, University of Nebraska-LincolnFollow

First Advisor

Daniel R. Uden

Committee Members

Jessica Corman, Elizabeth VanWormer, Brian Wardlow

Date of this Version

8-2024

Document Type

Thesis

Citation

A thesis presented to the faculty of the Graduate College at the University of Nebraska in partial fulfillment of requirements for the degree of Master of Science

Major: Natural Resource Sciences

Under the supervision of Professor Daniel R. Uden

Lincoln, Nebraska, August 2024

Comments

Copyright 2024, Mercy Kipenda. Used by permission

Abstract

Cyanobacterial harmful algal blooms (CyanoHABs) detrimentally affect human, animal, and ecosystem health. Remotely sensed early warning systems for cyanoHABs in inland lakes could contribute to more proactive water quality monitoring and help mitigate negative impacts. Advances in freely available remote sensing imagery, with finer spatial, temporal, and spectral resolutions, present new opportunities for the development and comparative analysis of methods to detect sudden deterioration in lake water quality. In this thesis, I compared and tested for temporal and spatial early warning signals of cyanoHABs in field-based and remotely sensed datasets from 2019 to 2023 in Pawnee Lake in southeast Nebraska, United States of America. Field data consisted of biweekly microcystin (MC) levels from the Nebraska Department of Environment and Energy’s Beach Watch Dataset and remotely sensed data consisted of two-week Normalized Difference Chlorophyll Index (NDCI) composites from the Sentinel 2B surface reflectance satellite. In Chapter 1, I tested for rising variance in biweekly MC and NDCI time series from May-September of each year at three rolling window sizes. I also computed the correlation between MC and mean lake wide NDCI and examined within-year trends in each variable. Both MC and NDCI tended to increase from May-September of each year and the relationship between MC and NDCI approached statistical significance (p = 0.06) but rising variance did not provide early warning of documented cyanoHAB events for either variable. In the second chapter, with a landscape ecology-based approach, I computed the number of high-NDCI patches (i.e., contiguous pixels with elevated NDCI values) within Pawnee Lake, computed the correlation between MC and the high-NDCI patch count, and tested for rising variance in high-NDCI patch count at three rolling window sizes. Although both MC and high-NDCI patch count tended to increase from May – September of each year, I found no relationship between MC and high-NDCI patch count and no evidence of early warning of documented CyanoHABs. Reasons for the lack of advanced warning could include small seasonal sample sizes and insufficient temporal resolution in both the field- and remotely sensed observations, examination of only a subset of temporal and spatial early warning indicators and limited geographic scope. This study provides a baseline for guiding future analyses with higher-resolution observations and alternative metrics and locations.

Advisor: Daniel R. Uden