Validation of Harmful Algal Bloom (Karenia brevis ) Induced Respiratory Irritation Forecast (RIF) Model In Southwest Florida From 2006 - 2023

Harmful algal blooms (HABs) occur when phytoplankton rapidly grow out of control, producing high biomass blooms and often toxins which are harmful to both people and animals. While most HABs produce toxins that must be ingested to be harmful, the toxins produced by the HAB species Karenia brevis (also known as red tide in the Gulf of Mexico) can become aerosolized, which lead to acute respiratory irritation. Exposure to the aerosolized toxins is a nuisance to most people (coughing, runny nose, watery eyes), and a health risk to those with asthma or other respiratory conditions. Risk of exposure depends on bloom intensity at the beach and the occurrence of onshore winds.

In response, NOAA’s National Centers for Coastal Ocean Science (NCCOS) developed a respiratory irritation forecast (RIF) model to provide beach-goers with a category-based forecast of respiratory irritation risk (‘very low’, ‘low’, ‘moderate’, or ‘high’ risk) at individual beaches in the Gulf Coast region. The inputs for the RIF model include: 1) K. brevis cell counts collected at individual beaches by state and county agencies in addition to local community scientists; 2) high resolution, 2.5km wind forecasts provided by NOAA’s National Weather Service (NWS), National Digital Forecast Database (NDFD); and 3) point-based beach shoreline orientation used to designate onshore and offshore winds. The forecast is publicly available at the NCCOS webpage (https://coastalscience.noaa.gov/science-areas/habs/hab-forecasts/gulf-of-mexico/) and updated every 3 hours over the following 36 hours during both bloom (~August-January) and non-bloom seasons. During the 2022-2023 bloom season, the forecast webpage received a total of ~135,000 unique page views with high respiratory irritation alerts reaching a total of ~155,000 people on social media through NCCOS and NWS Tampa Bay Weather Forecasting Office.

In order to test the RIF model logic, an analysis of modeled respiratory irritation hindcasts (2006-2023) were compared to same-day reports of respiratory irritation, based on frequency of coughs at individual beaches sourced from the Beach Conditions Reporting System (BCRS, Mote Marine Laboratory), which has been operating since 2006. To produce modeled hindcasts, historical cell counts were matched to same-day BCRS reports and then to local buoy wind data. The NDFD winds have shown to resolve sea breezes and can be evaluated against the meteorological measurements from local Florida buoys. After including wind data for Clearwater (CWBF1), Venice Beach (VENF1), Naples (NPSF1) from the National Data Buoy Center (NDBC), shoreline information was calculated for each historical match of BCRS and K. brevis cell count. Finally, the modeled hindcast was calculated and statistically compared to the BCRS category.

Validation efforts confirmed the overall assumptions of the model logic, including: 1) reports of higher respiratory irritation correlated with higher cell count concentrations; and 2) when cells were present, onshore winds led to higher respiratory irritation. However, we found that individual model categories, particularly hindcasts of ‘low’ and ‘moderate’ respiratory irritation, were less robust and will need to be reassessed. Furthermore, BCRS is not a direct measure of aerosolized brevetoxin presence, so some coughing (i.e., modeled false negatives) may result from other factors besides exposure to aerosolized toxins on the beach. Taken together, validation results suggest the RIF model logic accurately predicts ‘very low’ and ‘high’ risks of respiratory irritation at beaches across the southwest Florida coast, but that additional research is needed to better capture environmental conditions when respiratory irritation is ‘low’ or ‘moderate’.