Abstract: The Harmful Algal Blooms (HABs) are one of the main threats in the freshwater aquatic ecosystem. In particular, cyanobacterial bloom often leads to the death of aquatic organisms due to toxin production and dissolved oxygen depletion. It is well-known fact that traditional field monitoring is not able to cover larger area and has low repeatability. And remote sensing can be applied to cyanobacteria monitoring by observing specific pigment, phycocyanin (PC), with high spatiotemporal resolution. However, remote sensing is able to observe near surface of waterbody. Vertical distribution pattern of cyanobacteria operates as a hindrance in accurate remote sensing. Because of vertical migration of cyanobacteria which is varying pigment concentration near water surface, the spectral signal from waterbody may be affected according to observation time of day. In this regard, this study aims to propose and evaluate new bio-optical algorithm using cumulative pigment concentration by depth to mitigate vertical migration effect. This study was conducted on a lake water and high spectral and spatial resolution reflectance data were gathered using UAV mounted hyperspectral sensor. In-situ remote sensing reflectance (Rrs) was measured by portable spectroradiometer to evaluate performance of UAV-based remote sensing. Surface and vertical profile of PC were measured by YSI EXO-2 water-quality multi-probe. The cumulative concentration was calculated by 0.5m depth interval. To estimate PC concentration using UAV and in-situ reflectance, two-band ratio algorithm, was applied and 709 and 665nm bands were used. The performance of bio-optical algorithms was evaluated using r-squared values. As a result of the bio-optical algorithm, using in-situ Rrs and surface PC concentration, r-squared was 0.54, while UAV-based reflectance showed 0.34, which was a rather low value. In case of results using in-situ Rrs and cumulative PC, improved r-squared performance were observed for all cumulative depth. The range was 0.76 to 0.85 depending on the cumulative water depth of 0.5 to 3.0m. The optimal cumulative depth was 0.5 m, and r-squared gradually decreased as the cumulative depth increased showing minimum at 3.0m. The regressions between UAV-based surface reflectance and cumulative concentration also indicated better r-squared values for all cumulative depth with range of 0.82 to 0.78. The r-squared increased from 0.5 to 1.0m with a maximum value, then gradually decreased. This study demonstrated that effectiveness of cumulative concentration on remote sensing. In future, considering light attenuation in freshwater would helpful for revealing relationship between remote sensing signal and vertical distribution pattern of algal community.
