satellite image of study area with ocean very orange before hurricane, and orangeness much reduced after hurricane and replaced with green blue

Observing Red Tides Comparative study of red algae blooms in Western Florida in 2022 - 2025

Mariam Gedenidze, Stefanía Sibille, Grández Yuyang Sun, Drake Christianson

Abstract 

This study compares two harmful red tide (Karenia brevis) events along Florida’s west coast—October 2022–February 2023 (no major hurricane) and October 2024–February 2025 following Hurricane Milton—using Sentinel-2 multispectral imagery processed in Google Earth Engine and ENVI. Monthly median composites (October, December, February) over an ~835 km² region near Sanibel Island were analyzed with four spectral indices (NDCI, Red Tide Index, Chlorophyll Index, and Floating Algae Index), followed by a supervised Maximum Likelihood classification to map three classes: red algae, water without algae, and land. The Floating Algae Index provided the clearest separation, while NDCI was confounded by terrestrial vegetation. Contrary to the initial hypothesis that hurricane-driven nutrient pulses would amplify blooms, the post-hurricane period showed consistently smaller algae extent than the 2022–2023 case (reductions of ~2,800 ha in October, ~20,600 ha in December, and ~14,800 ha in February). Potential explanations include classification and mosaic artifacts, flood-related land/water confusion, and nutrient flushing. The study demonstrates the utility of Sentinel-2 for HAB monitoring and recommends broader spatial/temporal coverage and integration with in-situ measurements for robust attribution.

satellite image of study area - southwest florida coast

Figure 1. The region of interest, with scale bar and north arrow.

workflow classification

Figure 2. Classification Workflow.

images of red algae distribution before and after hurricane

Figure 3. Supervised classification results based on FAI index.

Summary characterisitics of red algae, water without algae, and land in october, december, and february

Table 1. Summary statistics (hectares of coverage) of classification results, with computed differences between the two cases.

a series of six images of study area, with ocean area in black and land in yellow

Appendix I. NDCI Index of the study area. A BLUE/GREEN/RED/YELLOW color table has been applied with linear stretch (thresholds 0 and 0.25).

satellite image of study area with ocean very orange before hurricane, and orangeness much reduced after hurricane

Appendix II. Redtide Index of the study area. A BLUE/GREEN/RED/YELLOW color table has been applied with linear stretch (thresholds 0.95 and 0.98).

satellite image of study area with ocean very orange before hurricane, and orangeness much reduced after hurricane and replaced with green blue

Appendix III. Chlorophyll Index of the study area. A BLUE/GREEN/RED/YELLOW color table has been applied with linear stretch (thresholds 0.5 and 2.0).

satellite image of study area with ocean very orange and land very yellow before hurricane and ocean much more black and green with a smaller outburst of orange after hurricane

Appendix IV. FAI of the study area. A BLUE/GREEN/RED/YELLOW color table has been applied with linear stretch (thresholds -0.01 and 0).