Neural Networks and Statistical Methods for Climate Studies
Advanced statistical techniques can be used for a variety of applications
in Meteorology and Climatology. Linear or nonlinear regression
can be used for the approximation of direct or inverse radiative
transfer models for remote sensing purposes. Classification algorithms
are usefull to define prototypes of airmass, types of surface,
describe indices of vegetation, desert, ice or wetland. Component
extractions techniques help in analyzing huge datasets of observations
or model outputs, they can also improve our understanding of the
processes involved in the atmosphere or the ocean.
Research Themes:
Classification of Surface Properties
(See dedicated WEB page)