TC-Winds are highly specialized for tropical cyclone (TC) scenes.

Find general purpose winds here: www.remss.com/measurements/wind


TC-Winds

The retrieval of satellite winds in storms is challenging and often requires dedicated algorithms. To help tackle this problem, Remote Sensing Systems has developed several new algorithms which allow for reliable wind measurements in Tropical Cyclones (TCs), including observations in rainy environments.


SMAP

The first such product is the SMAP global daily winds, released in 2017. The SMAP observations provide realistic wind retrievals in TCs and are not affected by rain. The SMAP wind dataset is extensively described and validated in Meissner et al. (2017). SMAP winds are processed in Near Real Time and are available at www.remss.com/missions/smap.


WindSat TC-Winds

Here we describe and provide data access to a new product: WindSat TC-winds. This dataset was developed in 2020 and has been processed for the entire duration of the WindSat mission (2003 to present). More information on the WindSat sensor can be found at www.remss.com/missions/windsat.

The WindSat TC-wind retrieval algorithm is a statistical linear regression and is based on the method by Meissner and Wentz (2009). It uses combinations of WindSat C-band and X-band channels to correct for the effects of rain and is trained using SMAP winds in tropical cyclone conditions, which include high winds and rain. This statistical training is possible as SMAP and WindSat have approximately the same ascending node (equatorial crossing) times. As such, it is easy to obtain match-ups between WindSat radiometer measurements and SMAP wind speeds that can be used in the algorithm training.

WindSat TC-winds are distributed as netCDF4 daily global 0.25° gridded maps, separated into ascending and descending passes. The TC-winds data are only valid over warm waters (SST > 20°C) and for winds greater than 10 m/s, as this algorithm has been specifically developed for TC conditions.

The netCDF files include the following variables:

  • TC-wind speed
  • wind direction
  • rain rate
  • ancillary sea surface temperature
  • time of observation

The WindSat wind directions become noisy and unreliable in very high rain.


AMSR-E and AMSR2 TC-Winds

We have extended the WindSat TC-wind algorithm to the AMSR-E and AMSR2 sensors, which also measure at C-band and X-band channels that can be used to correct for rain. Because the ascending node (equatorial crossing) times of the AMSR and SMAP are 5 – 6 hours apart, it is not feasible to create AMSR-SMAP match-up data and use the SMAP wind speeds directly to train the AMSR TC-wind algorithms. However, it is possible to adapt the WindSat TC-wind algorithm to AMSR. The AMSR brightness temperature (TB) measurements are shifted to the WindSat configuration, whose channel frequencies and Earth Incidence Angles differ slightly from both AMSRs. The shift is performed by using the RSS Radiative Transfer Model (Meissner and Wentz 2012; Wentz and Meissner 2016).

AMSR-E and AMSR2 TC-winds are distributed as netCDF4 daily global 0.25° gridded maps, separated into ascending and descending passes. The data are only valid over warm waters (SST > 20°C) and for winds greater than 10 m/s, as this algorithm has been specifically developed for TC conditions.

The netCDF files include the following variables:

  • TC-wind speed
  • rain rate
  • columnar water vapor
  • ancillary sea surface temperature
  • time of observation

TC fixes

Additionally, we provide Tropical Cyclone (TC) ASCII files with SMAP/WindSat/AMSR-E/AMSR2 10-min maximum-sustained winds (in kt) and wind radii (in nm) for the 34 kt (17 m/s), 50 kt (25 m/s), and 64 kt (33 m/s) winds for each pass over a TC in all tropical ocean basins. Images for these TC fixes are also provided as jpeg files. The TC fixes are written in a format easy to ingest for TC forecasts, as used in the Automated Tropical Cyclone Forecasting (ATCF) systems at the US Navy.