Retrieval of Atmospheric Temperature Profiles from AMSU-A Measurement Using Artificial Neural Network and Its Applications for Estimating Tropical Cyclone Intensity for ‘Gonu’ and ‘Nargis’ - Monitoring and Prediction of Tropical Cyclones in the Indian Ocean and Climate Change - page 398

IAPP software package is due to imperfection of forward model calculations,
which may not be fully capable of extracting the low-level atmospheric
information and uncertainties in surface emissivity and cloud parameters. The
accuracy of retrievals depends on surface type (sea and land). From both the
figures it can also be noticed that the RMS errors are less for Port Blair as
compared to Chennai station. This may be due to the constant emissivity over
the sea compared to highly variable emissivity over land. The Port Blair station
is situated over the ocean and Chennai is mainly over the land. Although some
more cases need to be evaluated in order to obtain a statistical significant result.
4. Retrieval of Temperature Anomaly RT a for the
Cyclone 'Gonu' and 'Nargis'
One of the most exciting capabilities of the AMSU is the observation of tropical
cyclones and it has been observed that the AMSU-A instrument, which was
designed for temperature sounding can also observe the upper-level warm core
of a tropical cyclone (Kidder et al., 2000). A Tropical Cyclone (TC) is a warm
core low-pressure system. The strength of the warm core is directly related to
the storm's intensity. Over the past two decades, several studies (Velden et al.,
1991; Yao et al., 2005) have documented the TC intensities using satellite-
borne passive microwave radiance data. Brueske and Veldon (2003) developed
an MSLP estimation technique at CIMSS using AMSU-A 54.96 GHz (~200
mb) radiance anomalies.
The section 3.1 shows that the best results can be obtained by the neural
network at the lowest levels and near the 300-100 hPa level. Therefore to study
warm core anomaly of a cyclone the network-based retrieval algorithm was
applied for computing the anomaly at 200 hPa level near the centre of the
Super Cyclone 'Gonu' over the Arabian Sea and Very Severe Cyclone 'Nargis'
over Bay of Bengal formed on 31 st May-6 th June 2007 and 26 th April-2 nd May
2008 respectively. Then the temperature anomalies were correlated to MSLP
and wind speed obtained from Ocean Vector Winds Science, USA (http://
podaac.jpl.nasa.gov) and finally related to the intensification of the cyclone.
The temperature profiles are retrieved using the network developed in section
2.2. In this section, anomaly at 200 hPa from retrieved temperature profile will
be referred to as RT a .
4.1 Super Cyclone 'Gonu' and Very Severe Cyclone 'Nargis'
For cyclone 'Gonu', the retrieval data was selected to begin on 31 st May 2007.
The system 'Gonu' initially appeared as a low level circulation developed over
the Southeast Arabian Sea and intensified into a vortex with intensity T 1.0 on
Dvorak's scale (The Dvorak technique is a method using Infrared and/or visible
satellite imagery of geostationary satellite to quantitatively estimate the intensity
of a tropical system), centered at 12.5N/75.5E. Initially it moved in the northerly