Satellite chlorophyll retrievals with a bipartite artificial neural network model
Abstract:An artificial neural network (ANN) model with a bipartite classification scheme is developed to retrieve the chlorophyll‐a concentration (Chl) from sea‐viewing wide field‐of‐view sensor (SeaWiFS) data. Bio‐optical data derived from the SeaWiFS bio‐optical algorithm mini‐workshop (SeaBAM) are used to verify this bipartite artificial neural network (BANN) model. In comparison with SeaWiFS operational algorithms and a general ANN model, the BANN model significantly increases the accuracy of Chl retrieval not only on a log scale but also on a normal scale. The BANN model can significantly improve the accuracy of Chl especially in the high Chl region. The model also performs well in a test with in situ measurements from Taiwan coastal waters. The biases induced by errors in atmospheric correction are also reduced in the coastal water case.
Document Type: Research Article
Affiliations: 1: Department of Marine Environmental Informatics, National Taiwan Ocean University, Keelung, Taiwan 2: Department of Meteorology, University of Maryland, College Park, Maryland, 20742, USA 3: Department of Environmental Biology and Fisheries Science, National Taiwan Ocean University, Keelung, Taiwan
Publication date: April 20, 2006