全球变化知识资源中心

In this study, we simulate three-dimensional transport of algal blooms in Lake Erie using a combination of remote sensing and hydrodynamic modelling. The remote sensing algorithms use data from the Sentinel-3 OLCI satellite sensor to derive chlorophyll-a concentration from cyanobacteria blooms in Lake Erie. The derived chlorophyll-a concentration initializes an algal bloom transport model driven by the lake component of the Water Cycle Prediction System for the Great Lakes, a system of coupled atmosphere-lake-hydrological models operated out of Environment and Climate Change Canada. The bloom is modelled as Microcystis aeruginosa, a buoyant species that is often dominant in harmful algal blooms in western Lake Erie. Short-term (a few days) predictions of algal bloom transport from July 27 to October 8, 2017 are modelled in both Eulerian and Lagrangian frameworks. The Eulerian framework is used to evaluate the sensitivity of model results to the initial vertical distribution of the bloom. In this work, the Lagrangian framework is limited to two-dimensional surface confined particles. We use several error metrics to evaluate model predictions. We find that results are sensitive to the buoyancy velocity for cases where the bloom was initially distributed over a large portion of the water column. An initial vertical distribution selected from modelled chlorophyll-a half depth shows the highest accuracy for the entire range of buoyancy velocities tested. We also find that the Pierce skill score is difficult to interpret, particularly in cases where bloom intensity is greatly overpredicted by the model. Crown Copyright (C) 2018 Published by Elsevier B.V.