Assessing the effectiveness of Landsat 8 chlorophyll a retrieval algorithms for regional freshwater monitoring

• DOI: 10.1002/eap.1708
• Publication Date: 2018-05-30T18:44:16Z
• Authors: Jonah Boucher; Kathleen C. Weathers; Hamid Norouzi; Bethel Steele
• Keywords: Proxy (statistics)

Abstract Predicting algal blooms has become a priority for scientists, municipalities, businesses, and citizens. Remote sensing offers solutions to the spatial temporal challenges facing existing lake research monitoring programs that rely primarily on high‐investment, in situ measurements. Techniques remotely measure chlorophyll (chl ) as proxy biomass have been limited specific large water bodies particular seasons narrow chl ranges. Thus, first step toward prediction of is generating regionally robust algorithms using remote data. This study explores relationship between in‐lake measured data from Maine New Hampshire, USA lakes sensed retrieval algorithm outputs. Landsat 8 images were obtained then processed after required atmospheric radiometric corrections. Six previously developed tested regional scale 11 scenes 2013 2015 covering 192 lakes. The best performing across both states had 0.16 correlation coefficient ( R 2 P ≤ 0.05 when within 5 d, improved 0.25 only used. strength varied with specificity time window relation in‐situ sampling date, explaining up 27% variation several scenes. Two published 8's Bands 1–4 correlated , late‐summer scenes, they accounted 69% A sensitivity analysis revealed longer difference measurements satellite image increased uncertainty models, an effect year indices was demonstrated. model based validated independent images. These results suggest that, despite including seasonal effects low thresholds, could be effective accessible regional‐scale tool