MABI: A novel Mixed Algal Blooms Index based on color space transformation

Abstract

Harmful algal blooms (HABs) pose serious threats to coastal economies and ecosystems, yet effective monitoring remains challenging due to diverse bloom types and complex environmental conditions. This paper proposes a Mixed Algal Blooms Index (MABI) that uses a new color space to improve HABs detection. By employing Sentinel-2's near-infrared, short-wave infrared, and green bands to calculate tristimulus values—replacing traditional RGB bands—MABI significantly enhances the distinction between algal blooms and water. And an improved grid-based Otsu automatic threshold segmentation algorithm is utilized to extract algal blooms. The inter-class distance is employed as an indicator to compare 14 commonly used algal blooms indices. Validation across nine global sites, covering coastal and inland areas, shows MABI's robustness, with an overall accuracy of 0.98 and a Kappa coefficient of 0.95. Compared to traditional algal bloom indices, the proposed MABI shows notable advantages in detecting blooms, effectively identifying both mixed blooms from multiple algae species and single-species blooms. We also verified the effectiveness of MABI with Landsat-8, and the combination of Landsat and Sentinel-2 imagery is expected to enhance its capability to monitor the full lifecycle of algal blooms. While MABI shows some resistance to thin clouds and shadows, its detection accuracy can still be affected in optically complex waters. Therefore, careful threshold selection or combining with other indices is recommended for comprehensive assessment. This study utilized Google Earth Engine (GEE) for data acquisition, processing, algorithm development, and validation, offering an efficient and reliable tool for accurately monitoring HABs with wide-ranging applications.