Improving Micro– and Macro–Algae Speciation Mapping for Management: A Field‐Validated Sentinel‐2 and Sentinel‐3 Intercomparison in the Indian River Lagoon

Abstract In recent years, the health of the Indian River Lagoon (IRL) on Florida's east coast has fluctuated due to recurring harmful algal blooms (HABs). These events typically involve cyanobacteria, dinoflagellates, or A. lagunensis (brown tide) causing water‐quality degradation and seagrass decline. The IRL is prone to HAB formation, and they have been observed more commonly in recent years. Although HABs can be non‐toxic, they can threaten critical primary producers like seagrass and destabilize sediments. Due to their impact, it is critical to improve methods of water quality and algal bloom monitoring. To do this, we apply a multi‐image varimax‐rotated principal component analysis (VPCA) to images from a harmonized Sentinel‐2 A/B multispectral imager (MSI) data set spanning three years (2020–2022). Field collected hyperspectral data and a coincident Sentinel‐3 A/B OLCI image is used to further validate the multi‐image VPCA. The VPCA method employs visible derivative water surface reflectance and unmixes the spectral signal to identify orthogonal components related to divisions of algae, suspended sediments, and seagrass. This advances water quality monitoring by improving unmixing and identification of optical components over time. This identification is supported by high correlation to a spectral library, field observation, published observations, and validation across sensors of varying resolutions. Through this research we find confident identification and mapping of in‐water components related to cyanobacterial (component 1 r 2 = 0.865) and dinoflagellate blooms (component 4 and 5 r 2 = 0.986 and r 2 = 0.978), and seagrass beds (component 3 r 2 = 0.916) with 10 m spatial resolution through time improving water quality monitoring and root cause identification.