The paper is organised as follows: after the introduction, the second part describes the database and the methodology; the third one is dedicated to the detection of tropical cyclones and describes the specific treatments developed for altimeter data; the fourth part gives results about the analysis of ETDs cases and more precisely on the possibility of retrieving a SLP signal from altimeter measurements during ETDs. The last part gives a complementary analysis of the SLP-SLA relationship in the case of ETDs from barotropic model outputs (MOG2D, [1]).2.?Database and Methodology2.1. DatabaseThe 2003/2004 time period has been chosen for the analysis because it is covered by several independent databases:- the ENVISAT, Topex/Poseidon and Jason-1 altimeter missions;- an extensive observing network deployed in the Atlantic ocean by the National Oceanic and Atmospheric Administration (NOAA).

The NOAA hosts the National Hurricane Center (NHC) and the Hurricane Research Division (HRD), which has defined an experimental wind analysis tool to provide regular high-resolution wind fields for tropical cyclones ([13]; http://www.solar.ifa.hawaii.edu/Tropical/tropical.html). This database gives an extensive list of tropical storms which have occurred on all ocean basins, with information on the track of the storm and estimates of the maximum sustain winds, wind gusts and the minimum central pressure. However these estimates give a measure of the storm��s intensity but not of the wind or SLP field which can be easily compared with the altimeter ground track measurements;- a collocated JASON/buoy database: buoy data include the NDBC network, data available via M��t��o-France, and the TAO array;- the ECMWF pressure analyses at 0.

5 degree-6 hour resolution;- the QuikSCAT scatterometer wind measurements; QuikSCAT winds have been assimilated into the ECMWF Numerical Weather Prediction (NWP) model since 2002.The ECWMF global pressure Drug_discovery fields are used to provide long time series of surface pressure with global space/time coverage. However, in this study, we are mostly interested in low and very low pressure systems. In such conditions, NWP models such as ECMWF suffer from limitations related to their coarse space and time resolution, to very few assimilated SLP measurements (aside from those of ships of opportunity limited to the ships��main tracks), and the fact that the dense scatterometer winds are severely under-sampled when assimilated in the NWP. However, we can derive SLP fields from scatterometer wind measurements using an atmospheric planetary boundary layer (PBL) model [14,15]. These QuikSCAT-derived SLP fields have the advantage of retaining the fine scale structures present in the QuikSCAT wind fields.