Artificial Neural Networks to reconstruct incomplete satellite data: application to the Mediterranean Sea Surface Temperature

Satellite data can be very useful in applications where extensive spatial information is needed, but sometimes missing data due to presence of clouds can affect data quality. In this study a methodology for pre-processing sea surface temperature (SST) data is proposed. The methodology, that processe...

Full description

Saved in:
Bibliographic Details
Published in:Nonlinear processes in geophysics 2008-01, Vol.15 (1), p.61-70
Main Authors: Pisoni, E, Pastor, F, Volta, M
Format: Article
Language:English
Subjects:
Astrophysics
Cosmology and Extra-Galactic Astrophysics
Earth Sciences
Meteorology
Neural networks
Ocean temperature
Physics
Research methodology
Sciences of the Universe
Validation studies
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Satellite data can be very useful in applications where extensive spatial information is needed, but sometimes missing data due to presence of clouds can affect data quality. In this study a methodology for pre-processing sea surface temperature (SST) data is proposed. The methodology, that processes measures in the visible wavelength, is based on an Artificial Neural Network (ANN) system. The effectiveness of the procedure has been also evaluated comparing results obtained using an interpolation method. After the methodology has been identified, a validation is performed on 3 different episodes representative of SST variability in the Mediterranean sea. The proposed technique can process SST NOAA/AVHRR data to simulate severe storm episodes by means of prognostic meteorological models.
ISSN:1607-7946
1023-5809
1607-7946
DOI:10.5194/npg-15-61-2008