A probabilistic modelling framework for generalising the cause-and-effect relationships of the toxic cyanobacterium Lyngbya majuscula's harmful algal blooms in Moreton Bay, Queensland, Australia

Abstract

The increase of harmful algal blooms (HABs) in estuaries and coastal waters is a global concern for ecological, economic, and health reasons. As such, there is a need to develop suitable management strategies based on an understanding of algal blooms’ cause-and-effect relationships. Extant studies have significantly improved our understanding of the eutrophication processes that trigger HABs in various environments. However, Moreton Bay in Australia is a complex estuarine system, with inflows from major rivers, tidal flows, and an ocean barrier of islands and sand bars. Such biophysical and spatial features lead to significant variability in the ecosystem which makes it difficult to generalise such cause-effect relationships. This has raised debate among scientific communities about bloom initiation and development which, in turn, has obstructed appropriate management planning. Moreton Bay has been subject to HABs of the cyanobacterium Lyngbya majuscula with confounding causes since the late 1990s. There is still no certainty about whether naturally driven factors, human-related changes, or a combination of both are underpinning the frequency and expansion of L. majuscula blooms and distribution in Moreton Bay. Despite efforts to model and predict the initiation of L. majuscula, scientific explanations on the cause-and-effect relationships have had limited success. While extant models have considered significant environmental and socioeconomic factors, variables influencing the generic occurrence of hypoxia (depleted dissolved oxygen (DO) concentration below 3 mg/L) have yet to be thoroughly incorporated due, in part, to insufficient monitoring data collected on DO and chlorophyll-a (chl-a). Moreover, the relationship between L. majuscula’s preferred habitat and the ability of benthic sediments to induce hypoxia and provide bioavailable nutrients for initiation and growth has not been investigated. This thesis adopts the hypothesis that hypoxic water conditions at the benthos are an essential cause of L. majuscula initiation, specifically when bioavailable nutrients are concomitantly released from the sediments to the water column. Such processes may be described in terms of spatial-temporal patterns and may be statistically generalised by probabilistic models to distinguish influential factors in the initiation and occurrence of L. majuscula in Moreton Bay. These models may also provide the ability to identify locations under threat of experiencing HABs.