Diapositiva 1

1
GOME and SCIAMACHY Ozone Processors
Alessandro Burini Fabio Del Frate
2
Neural Network and Ozone Retrieval
The Atmosphere monitoring plays a key role in
understanding our planet and its changing. The
derivation of atmospheric parameters from
satellite remote sensing instruments plays a
key-role in monitoring the Earth's atmosphere
and understanding the chemical and physical
processes therein. However, the atmospheric
structure and trace gases concentrations result
from a complex interaction between radiative,
physical and chemical processes.
3
Neural Network and Ozone Retrieval
The most widely used methods to obtain ozone
concentration profiles from nadir satellite data
are based on Optimal Estimation schemes, that
involve the use of a direct radiative transfer
code and some a priori information. Neural
Networks (NNs) emerged in the last years as an
alternative useful tool to face with such
complexity NNs allow ozone profiles retrievals
in a very short computational time and without
the need of a priori and/or climatological data.
The Neural Network has been trained by using
model simulations and real ozone profiles,
measured by ozone sonde.
4
Neural Network Training
A fast processing provides the retrieved data to
users, such as research agencies, universities or
companies involved in Atmospheric System
Management. The output is provided by Geo_Tiff
files, one per each pressure level. Gome Ozone
Processor retrieves ozone profile at 20 pressure
levels. SCIAMACHY Ozone Processor covers the
first 9 pressure levels. Elaboration time is of
about 5 minutes for Gome data and 30 minutes for
Sciamachy data.
5
GOME
The Global Ozone Monitoring Experiment (GOME) on
ERS is a nadir-viewing spectrometer observing
solar radiation transmitted through or scattered
from Earth's atmosphere or from its surface. The
recorded spectra is used to derive a detailed
picture of the atmosphere's content of ozone,
water vapour, and other trace gases.  3 days for
full world coverage
6
Pressure Level 1 (1000 hPa)
7
Pressure Level 5 (31.62 hPa)
8
Pressure Level 8 (10 hPa)
9
Pressure Level 13 (0.56 hPa)
10
Pressure Level 17 (0.03 hPa)
11
SCIAMACHY
SCIAMACHY Acquisition Geometry
Pressure level 5 Ozone Concentration
12
SCIAMACHY Output
Ozone concentration map for level 0 retrieved by
the Sciamachy Ozone Processor. Sciamachy has got
a better spatial resolution than Gome, but a
worse coverage
13
Gome and Sciamachy Validation
The first tests about the precision of the
retrieved ozone concentration profiles give good
results, with an enhancement of the accuracy
using Sciamachy data (due of the better spectral
resolution). The processing time is shorter than
the standard processor.