Analysis of extreme low pressure events like hurricanes and extra-tropical storms thanks to altimetry
8-10 rue Hermès,
The relationship between atmospheric Sea Level Pressure (SLP) and the Sea Level Anomalie (SLA) measurements is investigated during storms and very low pressure conditions. Indeed the Inverse Barometer (IB) response has been extensively studied for normal meteorological conditions (Wunsh and Stammer 1997; Mathers 2000; Carrère 2003); but it remained uncertain that there exists a well SLP/SLA correlated signal during storms and hurricanes which are generally characterized by heavy rains, high sea states and strong winds.
The 2003-2004 period has been studied to benefit from the availability of 3 altimeter missions (ENVISAT, TP and Jason-1), an extensive observing network deployed by NOAA in the Atlantic ocean, high resolution ECMWF products (pressure and wind at 0.5°) and from the QuikScat scatterometer measurements.
Specific altimeter treatments have been applied because of too many missing measurements due to heavy rains and strong winds. The C- or S-band measurements are used because less impacted by rain, an expected Ku-band backscatter coefficient is defined for data impacted by rain, and continuous along-track wind speed profiles are computed thanks to the Young algorithm (Young 1993). Due to a non negligible error bar of 11 cm rms, these treatments were only applied on TCs observations, for which classical validated data are the less numerous. Along-track altimeter data were also filtered to remove ocean signal not related to atmospheric pressure : 600 km wavelength for TCs and 1500 km wavelength for extra-tropical storms which have significantly larger spatial scales.
The ability of altimetry to detect extreme low pressure events has been demonstrated during extra-tropical storms, with good correlation between SLP and SLA. It came out that one needs to use all storms cases at once to get a stable regression coefficient. The validation of the regression model showed interesting results (no bias and 6 cm standard deviation error if compared to ECMWF pressure, and 5 cm rms error if compared to colocalised altimeter-buoys database); but a small residual error-SLA dependency still remains (error = SLPrestored-SLPecmwf).
For tropical cyclones, selected cases showed that altimetry could provide precise information about wind speed (Quilfen et al., submitted to JGR) compared to scatterometers for instance. However, the retrieval of surface pressure from altimetry during TCs was not possible due to the poor quality of the pressure datasets available during these events, to the more complex response of the ocean medium to such extreme forcing (maximum storm surge is due to wind during TCs), and to the heavy rains characterizing such extreme weather conditions.