Premium
Prediction of sea surface temperatures in the western Mediterranean Sea by neural networks using satellite observations
Author(s) -
GarciaGorriz Elisa,
GarciaSanchez Joan
Publication year - 2007
Publication title -
geophysical research letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.007
H-Index - 273
eISSN - 1944-8007
pISSN - 0094-8276
DOI - 10.1029/2007gl029888
Subject(s) - ocean gyre , mediterranean sea , climatology , sea surface temperature , satellite , mediterranean climate , geology , artificial neural network , environmental science , oceanography , meteorology , geography , computer science , subtropics , artificial intelligence , archaeology , aerospace engineering , fishery , engineering , biology
We use artificial neural networks (ANNs) to predict sea surface temperatures (SSTs) in the western Mediterranean Sea. The ANNs are trained with meteorological variables as input and concurrent satellite‐derived SSTs as target. The trained ANNs predict well both the seasonal and the interannual variability of SST in that region. We also reproduce the impact of the heat wave that occurred during the summer of 2003 on the SSTs of the western Mediterranean Sea. The ANN technique allows us to predict SST maps in the western Alboran Sea for time coordinates before SST satellite availability. The presence and later partial collapse of the western Alboran gyre throughout 1980 is detected with good agreement by both the ANN predictions and the concurrent results from a 3‐D circulation model. The same methodology is used to reconstruct incomplete SST satellite images.