The project team will establish a California regional hub for harmful algal bloom data, technology, and knowledge transfer, and then expand or export these technological tools to other regions on a national scale.

Why We Care
Harmful algal blooms (HABs) are persistent threats to coastal resources, local economies, and human and animal health throughout U.S. waters. HABs are expected to intensify and expand in range as oceans change in response to anthropogenic climate change, creating an immediate need for more effective strategies to monitor and communicate the risk of algal toxins to human and ecosystem health in U.S. waters.

In 2015–2016, a massive Pseudo-nitzschia bloom shut down the Dungeness crab fishery on the West Coast and impacted the entire marine food web from California to Alaska. Tightly associated with the anomalously warm ocean temperatures that have persisted in many parts of the California Current System since their onset in late 2013, that bloom and the ensuing toxic domoic acid (DA) event may be harbingers of things to come as scientists are better able to match the environmental conditions associated with marine heat waves to changes in the biogeography and intensity of some HABs.

In the last decade, the ocean science community has developed several novel sensors and methods for monitoring and predicting the diversity of HAB events, such as the Imagining FlowCytobot (IFCB) and the Environmental Sample Processor (ESP). Data from these sensors are assimilated into various biophysical modeling systems optimized for HAB prediction. Also, in recent years, efforts for monitoring HABs using microscopy have greatly increased within research laboratories and observing programs. The proliferation of technologies and HAB monitoring and prediction efforts has resulted in a surge of information made available to scientists, managers, and other user groups.

A negative consequence of this rapid progress is the lack of harmonization of protocols, technical frameworks, and approaches employed between groups working in the optical (microscopy), machine learning, and forecasting domains. A centralized data assembly center for HAB data and resources currently does not exist, and there is no directly funded group to coordinate across the growing HAB community. The central scientific doctrine of conducting analyses with reproducible methods is not achievable for HABs at a national level without improved coordination, implementation, and documentation of the growing set of complex analytical approaches, such as those of machine learning. It is clear that harmonization of approaches and convergence of technical resources will result in enhanced and standardized HAB data products that can be more easily and robustly integrated into management or monitoring frameworks.

What Are We Doing
The overarching goal of this project is to establish a California regional hub for HAB data, technology, and knowledge transfer, and then expand or export these technological tools to other regions on a national scale. This will promote harmonized data feeds and management-relevant products across the community-at-large. To accomplish this, this project set objectives that scale the effort from primarily the Imaging FlowCytobot and California, to a flexible integrated system capable of supporting regional and national efforts. The project team will:

• • • establish a Harmful Algal Bloom Data Assembly Center (HABDAC), and, using HABDAC cyberinfrastructure, support the California IFCB network with data capture, processing, visualization, and archive through operationalization of machine learning toolkits and development of improved data standards;
    • centralize and curate additional HAB data set types and various HAB forecast and nowcast bio-physical modeling systems;
    • centralize, assess, and optimize machine learning models and training data sets used by isolated groups across the community;
    • deploy a HAB Resource Catalog to enable users to discover HAB resources;
    • continue to centralize and curate new data set types and modeling systems;
    • promote the use of standardized data architecture at a national level through broad implementation of the HABDAC and HAB Resource Catalog; and
    • identify more permanent funding sources to operationalize this harmonized approach with IOOS and other interested partners.

Dr. Clarissa Anderson of SCCOOS-Scripps Institution of Oceanography leads this project. Co-investigators are Mr. Rob Bochenek (Axiom Data Science LLC), Dr. Heidi Sosik (Woods Hole Oceanographic Institution), Dr. Raphael Kudela (University of California – Santa Cruz), Dr. Henry Ruhl (CeNCOOS – Monterey Bay Aquarium Research Institute). The project is funded through the NCCOS Prevention, Control, and Mitigation of Harmful Algal Blooms Program.