Harmful algal blooms or HABs pose a grave threat to humans and wildlife in the San Francisco Bay, though they remain somewhat rare and small in scale. Despite the presence of many algal species, the San Francisco Bay generally does not experience severe blooms, and seasonal chlorophyll-a concentrations generally peak around 10-20 µg/L. HABs, while commonly fueled by a relatively high nutrient load, are thought to be subdued by a combination of lower light condition, high turbidity, and mixing. However, any disturbance in one or more of these limiting conditions can result in a major algal bloom due to this availability of nutrients. This was demonstrated in August of 2022, when a sudden bloom of Heterosigma akashiwo resulted in extreme chlorophyll-a concentrations, with peaks measured in the ranges of 200-400 µg/L or more, followed by a drop in dissolved oxygen levels in the SF Bay and subsequent fish die-offs.

While many efforts are underway to understand the exact mechanisms that created conditions for such a bloom, this event highlighted the need for a centralized location to monitor various environmental conditions that may signal the next such incident. Considering the San Francisco Bay’s expansive geography and its hydrologic connections to the massively complex Sacramento-San Joaquin Delta, a tool must capture the big picture while also reporting on local, small-scale events. Various tools exist, such as products developed by SFEI (San Francisco Estuary Institute) for remotely sensed data or continuous sensor data such as the CA State Water Boards FHAB California tool, but they do not provide a way to easily compare data visualizations from both monitoring and satellite remote sensing data.

The dashboard presented here combines continuous sensor data (SFEI and United States Geological Survey), discrete sampling data (USGS, Baykeeper), and remotely sensed imagery (Sentinel-3 OLCI) both in the San Francisco Bay and Sacramento-San Joaquin Delta into a single, online user-interface so as to bring these disparate datasets together, streamlining the comparison of trends across different data, and highlighting the interplay of changing environment conditions. The web-application is built on a stack of Postgres, Node, Vue, MapServer, OpenLayers, and D3 to create a user interface where data visualizations can be isolated into components, combined, synchronized, and rearranged as needed.

Remote-sensing imagery is processed more-or-less from scratch, beginning with the atmospheric correction and cloud detection and masking of Level-1 radiance files. From there, chlorophyll products are derived using a red-edge algorithm developed by Raphe Kudela et al. (2024), modifying the original algorithm given in Wynne et al. (2022) specifically for use in the San Francisco Bay. Turbidity products are derived using the algorithms developed by Nechad et al. (2010). All this is done via Python scripts that run autonomously on SFEI servers each day. These scripts use the Sentinel Application Platform (SNAP) provided by the European Space Agency (ESA) and incorporate some GDAL, providing the greatest flexibility and control in adjusting outputs and algorithms, as well as the ability to quickly update any piece of the pipeline.

Continuous sensor data are pulled through various web API services via automated Python scripts, formatted, and ingested into PostgreSQL databases. Because of differences in the design of each API and returned data, each script and ingestion process is bespoke for the dataset it connects to, but in the end, the result is an enforced consistency in data structure for display in the web application. Each data source can be queried, processed, and displayed in a consistent manner.

As part of the larger Monitoring and Event Response Research Program (MERHAB) to develop an integrated network for HAB detection and monitoring across the region, this dashboard serves as a hub for presenting the various efforts across the program and represents a key piece of communicating the data in one, centralized location. The plan is to accommodate more data sources as the system grows. In doing so, it will hopefully become an important tool for monitoring HAB events, understanding the drivers of such events, and supporting management decisions for ensuring the continued health of the San Francisco Bay and Sacrament-San Joaquin Delta.