Space-time geostatistical trend analysis and risk assessment for in-lake cyanobacterial toxicity

Session: Harmful Algal Blooms and Their Toxicity: Remote Sensing and Modeling Approaches (3)

Shiqi Fang, North Carolina State University, [email protected]
Dario Del Giudice, NC State University, [email protected]
Daniel Obenour, NC State University, [email protected]

Abstract

The occurrence and related risks from cyanobacterial harmful algal blooms (HABs) have increased in western Lake Erie over the past decade. Information on the abundance and distribution of cyanobacteria toxins is still limited, yet plays a fundamental role in risk assessment and management. In this study, a space-time geostatistical modeling approach was used to estimate the spatiotemporal distribution of both chlorophyll a (Chl-a) and Microcystin (MC). The approach was developed and tested using ten years (2008-2017) of data from multiple in situ monitoring programs and remote sensing products. Compared with Chl-a, MC are more stochastic (noisy), but exhibit substantial spatio-temporal correlation. We explored the variability in bloom toxicity (i.e., the MC: Chl-a ratio) and analyzed trends with seasonality, climate and nutrient conditions. For example, our results show that the bloom toxicity is strongly related to in-lake nitrogen concentrations (e.g. TKN and nitrate, R2 = 0.47) in the western basin. The resulting Chl-a and MC estimates were then used to map cyanobacterial risk relative to established human health criteria over time. The risk maps can assist authorities to identify times and location with high toxin risks, potentially indicating the need for higher frequency monitoring and enhanced mitigation of nutrients.