Winds (Continuation)

Following the development of the scatterometer community, SAR-based wind retrieval has been originally based on a single observed quantity; the co-polarized Normalized Radar Cross Section (NRCS) in the VV channel. The use of HH polarization has been integrated later using the co-pol ratio at first. Some developments on the use of dual-pol (co- and cross-pol) wind field retrieval applied to high wind situation have been carried out lately using Sentinel-1 (see Mouche et al 2017). With S-1, the SAR community has now access to a overwhelmed amount of dual-pol SAR data over the ocean sea surface. However, the capacity for TOPS EW/IW S-1 mode to preserve polarization phase is often neglected. This talk/paper aims at studying the benefit of the dual-polarimetric information provided by SLC TOPS-mode S-1 products (including phase).

 Ocean sea surface is often used to calibrate the polarimetric information via the reflection symmetry approach. The real and the imaginary parts of the co-cross polarization coherence should be equal to 0 from sea surface. However, some studies (e.g. Tsai et al 2000) have shown this condition is theoretically met at specific conditions such up or down wind situation. (Zhang et al. 2012) showed first analysis with Radarsat-2 full polarimetric data from a limited sea of wind condition and incidence angle range.

Here, we have adopted a massive approach with more than 3800 SLC IW products processed (about 27 TB). The co-cross polarization coherence (magnitude and phase) has been computed together with collocated ECMWF wind speed/direction. 

In this talk/paper, we study this massive processing and provide some analysis beyond the-state-of-the-art:

even in the case of reflection symmetry condition (low wind situation, up/down wind), the coherence magnitude does not reach exactly zero. The Maximum Likelihood Estimator (MLE) coherence is study together with the issue of non-stationarity.

a dependence on incidence angle and wind condition is found for the co-cross polarization magnitude and phase, opening the perspective to fully define a Polarimetric Geophysical Model Function (PGMF)

a connection between the polarimetric scattering world and the cross-spectral analysis of the co- and cross-pol channels. The latter can help understanding the displacements of scatterers along the wind waves depending on the relative wind direction and the polarization.

some striking features are shown, and some attempts to link them to the occurrence of breaking waves are provided.

the possibility to add this analysis for wind field retrieval, and lever the 180° ambiguities regarding the wind direction. Its applicability with respect to S-1 quality will be discussed (noise level, azimuth variation within TOPS burst...)

This study has been partly funded by the ESA SEOM S-1 Ocean studies, and the Mission Performance Center (MPC) S-1 activities.

The 2017 Atlantic hurricane season was with three major hurricanes a particular active one. The Category 4 hurricane Irma made landfall on the Florida Keys on September 10^th 2017 and was imaged several times by the ESAs Sentinel-1A and B  satellites as well as RADARSAT-2 in C-band, and the TerraSAR-X satellite in X-band.In the following as well images of hurricanes Jose and Maria were acquired.

Imaging the hurricanes by space radar gives the opportunity to observe the sea surface and thus measure the wind field and the sea state and sea surface motion under hurricane conditions through the clouds even in this severe weather, although rain features, which are usually not observed in SAR become visible due to damping effects.

The high resolution TerraSAR-X imagery showed sea state under the hurricane as well as the footprint of individual tornadoes on the sea surface together with their turbulent wake imaged as a dark line due to increased turbulence. The water-cloud structures of the tornadoes are analyzed and their sea surface structure is compared to optical and IR cloud imagery. An estimate of the wind field using standard XMOD algorithms is provided, although saturating under the strong rain and high wind speed conditions.

The Copernicus Sentinel-1 A and B satellites, which are operating in C-band provided several images of the sea surface under hurricane Irma, Jose and Maria. The data were acquired daily and converted into measurements of sea surface wind field u₁₀ and significant wave height H_s over a swath width of 280km about 1000 km along the orbit using as well ESA´s SNAP toolbox..

The wind field of the 2017 hurricanes as derived from Sentinel VV or HH imagery by CMOD  is as well provided by NOAA on their web server. In the hurricane cases though the wind speed saturates at 20 m/sec and is thus too low in the area of hurricane wind speed. The HV imagery now permits to retrieve hurricane wind speed above 60m/sec by simple linear algorithms.HV hurricane imagery as well shows a different eye size and is compared to cloud imagery.

As Sentinel imagery is as well provided as complex imagery, sea surface movement can be analyzed. Sea surface motion is derived from Doppler centroid measurements showing a strong westward motion in the northern part of the hurricane and an eastward one in the southern part.

A new theory on hurricane intensification based on Kelvin-Helmholtz instability is discussed and a first comparison to the SAR data is given.

In hurricane studies, the secondary eyewall is important in tropical cyclone (TC) evolution and intensification and is routinely assumed to be axisymmetric. A unique opportunity to investigate the secondary eyewall in two-dimensions is provided by the high spatial resolution (about 1 km) sea surface winds observed by spaceborne synthetic aperture radar (SAR) over hurricane Ike (2008). Hurricane Ike contains an asymmetrical secondary eyewall which is different from classical Eyewall Replacement Cycle (ERC) theory. Here, we extract the asymmetric characteristics using our SHEW (Symmetric Hurricane Estimates for Winds) model, which is based SAR-derived wind fields capturing the eyes of hurricanes. Thus, we analyze the related hurricane evolution by comparisons of SAR imagery with SFMR (stepped frequency microwave radar) aircraft measurements. We find the following characteristics for the asymmetric secondary eyewall: the primary eyewall does not weaken as long as the secondary eyewall exists, the asymmetric secondary eyewall does not contract, and the concentric eyewalls endure for a long period, more than 30 hours. We suggest that, as observed by SAR images, this persistence results from the low wind area in the secondary eyewall supplying a pathway for the boundary layer inflow to maintain the primary eyewall intensity, and reduce the energy needed for secondary eyewall contraction.

TC intensity may change due to the oceanic, environmental and internal dynamic processes. In the axisymmetric framework typical of hurricane processes, earlier studies demonstrated that heating in the secondary eyewall induces a negative tendency for the tangential winds inside the heating annulus. This can cut off the boundary layer inflow to the primary eyewall resulting in weakening of hurricane intensity. However, as an internal process, ERC theory does not eliminate other contributions such as those due to warm oceanic water. In fact, the convective rings defined as annular regions of active convective heat release result from evaporation latent heat from the warm ocean surface. Large-scale vertical wind shear has been proposed as a driver for hurricane asymmetry and is supposed to be the most important environmental contributor for TC intensity evolution. Therefore, if the asymmetric secondary eyewall of hurricane Ike (2008) is driven by vertical wind shear, our suggestion that the boundary layer inflow pathway may act as a supplementary internal process to the ERC theory provides the connection between the large-scale atmospheric environment and TC intensity changes. Additional studies should evaluate mechanisms related to the boundary layer inflow pathways in the hurricane asymmetric secondary eyewall, which can potentially improve intensity forecasts.

References:

Zhang and Perrie, (2018): Effects of Asymmetric Secondary Eyewall on Tropical Cyclone Evolution in Hurricane Ike (2008). In press GRL.

Zhang, G., B. Zhang, Perrie….. (2014), A hurricane tangential wind profile estimation method for C-band cross-polarization SAR, IEEE TGRS, doi:10.1109/TGRS.2014.2308839.

Zhang, Perrie,…. (2017a). A Hurricane Morphology and Sea Surface Wind Vector Estimation Model Based on C-Band Cross-Polarization SAR Imagery. IEEE TGRS, 55(3), 1743-1751.

Zhang, G., Li, X., Perrie,….. (2017b). A Hurricane Wind Speed Retrieval Model for C-Band RADARSAT-2 Cross-Polarization ScanSAR Images. IEEE TGRS.

Systematic Envisat Advanced Synthetic Aperture Radar (ASAR) acquisitions over the Agulhas Current between 2007 and 2012 are used to investigate the signature on high resolution (1km) satellite-derived winds of this intense and warm current, which velocity field is provided by Globcurrent products. SAR winds show an increase in magnitude over the Agulhas Current up-, down- and cross-current wind conditions with contributions from both the ocean surface current and Sea Surface Temperature (SST). In down-current conditions, when relative wind is expected to be reduced, SAR wind speeds typically increase by 1 to 2 m/s over the Agulhas Current consistent with the effect of SST perturbation on surface wind speed. In up-current wind regimes, intense spurious increases of up to 20 m/s in the retrieved SAR wind speeds are observed at the inshore edge of the Agulhas Current. On average, SAR winds speeds at the northern Agulhas Current wall increase by 4 to 6 m/s when the winds are against the current and 1 to 2 m/s when the winds are with the current. Similar but less intense wind accelerations are observed in the Jason-2 dataset in up-current wind conditions. Using co-located ocean current, SST and significant wave height information, we argue that the strong acceleration in SAR ocean winds observed at the inshore Agulhas Current front cannot be explained only by the presence of SST gradients or localised current convergence / divergence at the front. Our analysis rather suggests that wave-current interactions in regions of strong current shear produce enhanced sea surface roughness signatures and lead to artificially high estimates of SAR-derived wind speeds. There is future potential in using high resolution SAR imagery to map strong ocean current fronts and thus improve our ability to monitor ocean surface currents from space.