The Soil Moisture Active Passive (SMAP) radiometer is able to measure wind speeds up to 65 m s−1 without being affected by rain.
Photo credit: Seychelles Islands Development Company
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
Accurate measurements of ocean vector winds are necessary for improved ocean and weather forecasting for atmosphere and ocean numerical modeling (Atlas et al. 2001; Von Ahn et al. 2006; Chelton et al. 2006; Brennan et al. 2009; Candy et al. 2009; Chang et al. 2009). They are essential for understanding cyclogenesis and for forecasting tropical storms (Sampson and Schrader 2000; Katsaros et al. 2001, 2002; Sharp et al. 2002). In situ measurements from buoys, drifters, ships, or aircrafts are very valuable for this purpose, and they have been available for many decades. However, vast regions of the oceans are not covered by these spatially limited in situ wind observations. Weather forecasts for remote locations would not be accurate if relying only on in situ observations, particularly for extreme events.
The introduction of continuous satellite monitoring of ocean surface winds in the late 1980s has provided a much needed global view and a remedy for accurate global weather prediction (Bourassa et al. 2010). Two main types of satellite sensors are used to retrieve surface wind speed information over the oceans: microwave radiometers and scatterometers. Microwave radiometers are passive instruments that measure the emission of a rough ocean surface from which wind speeds can be derived (Wentz 1983, 1997; Wentz et al. 1986; Meissner and Wentz 2012; Zabolotskikh et al. 2015). Scatterometers are active sensors that measure the backscatter of a microwave radar signal that depends on wind speed and direction (Jones et al. 1977, 1982; Stoffelen and Anderson 1997; Wentz and Smith 1999; Ricciardulli and Wentz 2015). With their global daily coverage, these space-based wind measurements significantly improve the prediction skills when assimilated into numerical weather prediction (NWP) models (Isaksen and Stoffelen 2000; Von Ahn et al. 2006; Chelton et al. 2006). Multiple space-based microwave sensors were operating during 2015 and 2016. Among them were the U.S. Navy multichannel polarimetric radiometer WindSat (6.8–37 GHz, 4.4–0.8 cm; Gaiser et al. 2004), the Japanese Space Agency (JAXA) Advanced Microwave Scanning Radiometer (AMSR2; 7–89 GHz, 4.3–0.3 cm; Imaoka et al. 2010), the European Space Agency Advanced Scatterometer (ASCAT; C band, 5.2 GHz, 5.8 cm; Figa-Saldaña et al. 2002; Verspeek et al. 2010), and the National Aeronautics and Space Administration (NASA) Rapid Scatterometer (RapidScat) on the International Space Station (Ku band, 13.4 GHz, 2.2 cm; the mission ended in August 2016; http://winds.jpl.nasa.gov/missions/RapidScat/). All these sensors can measure wind speed at spatial resolutions of about 20–50 km. When verified versus buoys or aircraft measurements, the satellite measurements provide a wind speed accuracy of about 1.0 m s–1 (Verspeek et al. 2010; Meissner et al. 2011; Ricciardulli and Wentz 2015; Wentz 2015).
While these sensors give accurate and consistent wind measurements in the range of 0–30 m s–1, the instrument sensitivities and the reliability of the sensor winds at extreme wind speeds are still under investigation. One big challenge is that there are very few reliable in situ observations of extreme winds available for comparison. In addition, measurements of winds in storms are complicated by the presence of rain, which affects the radiometers (Wentz and Spencer 1998; Hilburn and Wentz 2008; Meissner and Wentz 2009) as well as the scatterometers (Stiles and Yueh 2002; Tournadre and Quilfen 2003; Draper and Long 2004; Hilburn et al. 2006; Weissman and Bourassa 2008; Portabella et al. 2012). These factors pose a challenge for the development and validation of high-wind retrieval algorithms. Currently, the most reliable measurements in tropical cyclones are airborne measurements from aircrafts flying into hurricanes, using dropsondes (Hock and Franklin 1999; Franklin et al. 2003), or the aircraft-mounted stepped-frequency microwave radiometer (SFMR; Jones et al. 1981; Uhlhorn and Black 2003; Uhlhorn et al. 2007; Klotz and Uhlhorn 2014).
A recent advance for wind measurements in storms is provided by passive sensors at a lower frequency (L band, 1.41 GHz, 21.4 cm), which are largely unaffected by rain and provide excellent sensitivity to wind speed even in very high winds. This capability was first explored by Reul et al. (2012) with observations from the Soil Moisture and Ocean Salinity (SMOS) European Space Agency mission (Kerr et al. 2010) using a radiometer at a resolution of 43 km. In their study, Reul et al. (2012) found that at L band the wind-induced brightness temperature increases approximately linearly with wind speed in high winds (greater than 25 m s–1). This first work has been recently extended to include an ensemble of storms intercepted by SMOS globally between 2010 and 2015 and using an improved wind speed retrieval algorithm (Reul et al. 2016). Validation of the SMOS surface wind speeds against SFMR and other sources shows a root-mean-square (RMS) error on the order of 5 m s–1. These results confirm that passive L-band sensors might be well suited for observations of marine storm-force winds (above 25 m s–1) from space and fill a gap in remote locations not covered by aircraft measurements.
The NASA Soil Moisture Active Passive (SMAP) mission (Entekhabi et al. 2010, 2014) was launched on 29 January 2015 and has been providing science data since April 2015. SMAP was designed as a combination of passive (radiometer) and active (scatterometer) L-band sensors to measure soil moisture and to provide information on the freeze–thaw state. Unfortunately, the radar transmitter failed in early July 2015, leaving only the radiometer operating since then. Though designed for hydrological applications over land surfaces, SMAP has excellent capabilities to measure ocean surface salinity (Meissner and Wentz 2016a,b) and ocean winds in storms.
The goal of this paper is to explain the strength of SMAP for measuring winds in tropical cyclones, even in the presence of heavy rain. We first show how to derive the signal of the wind-induced emission at L-band frequencies and summarize the main features of the wind retrieval algorithm. We then provide verification of the SMAP wind speeds and compare them with the results of several other spaceborne missions. It will become evident why SMAP has a unique advantage over many other sensors when it comes to measuring winds in intense storms.
THE SMAP L-BAND RADIOMETER.
The SMAP instrument flies in a polar 8-day repeat orbit at an average altitude of 685 km. It performs a conical 360° scan and maps out a contiguous swath of approximately 1,000-km width. The local times of the ascending and descending equatorial crossings are 1800 and 0600 local time (LT), respectively. The resolution of the SMAP radiometer is about 40 km.
The basic input to the retrievals of the ocean parameters salinity and wind speed are the SMAP level 1B antenna temperatures (Piepmeier et al. 2014, 2016). To convert the antenna temperatures into calibrated brightness temperatures at the ocean surface TB,surf, a number of spurious signals have to be removed. These include galactic radiation that is reflected from the ocean surface, Faraday rotation in Earth’s ionosphere, and atmospheric attenuation (Wentz and LeVine 2012; Meissner et al. 2014a, 2015; Meissner and Wentz 2016a,b; LeVine et al. 2015). Validation of the retrieved SMAP salinity against ground-truth observations shows that this calibration has a very high radiometric accuracy of about 0.2 K (Meissner and Wentz 2016b).
GEOPHYSICAL MODEL FUNCTION AND WIND SPEED RETRIEVALS.
Derivation of the geophysical model function.
To retrieve ocean surface wind speeds from SMAP, we need to take the difference between the SMAP-measured ocean surface brightness temperatures TB,surf and the brightness temperature of a flat ocean surface TB,surf,0 and match this difference to our radiative transfer model (RTM) of the wind-induced ocean surface emission. The component of the RTM that contains the wind-induced excess emissivity of the rough ocean surface is called the geophysical model function (GMF), and it is derived empirically. In doing so, we follow the same basic procedure that we have used for the Aquarius L-band radiometer (Meissner et al. 2014b). The SMAP radiometer TB measurements are collocated with wind speeds from WindSat (Wentz et al. 2013), which have been validated against ground-truth observations (Meissner et al. 2011). The equatorial crossing times of SMAP and WindSat are approximately the same, which makes WindSat an excellent candidate for creating matchups between the two instruments. We use a collocation time window of 1 h and observations that are contaminated by rain, land, or sea ice are removed. For the derivation of the GMF we use 6 months of SMAP–WindSat matchups containing a total number of more than 29 million collocations. We then subtract the computed brightness temperature of a flat ocean surface TB,surf,0 from the observed TB,surf and stratify this wind-induced excess emissivity (DTB,rough = TB,surf − TB,surf,0) with respect to the WindSat wind speed W. The computation of the flat surface TB,surf,0 is based on the Meissner–Wentz 2012 dielectric model of seawater (Meissner and Wentz 2004, 2012; Meissner et al. 2014b) and needs global ancillary input fields; ancillary sea surface salinity (SSS) is obtained from the top-layer salinity field of the Hybrid Coordinate Ocean Model (HYCOM) Global Analysis (GLBa0.08) Experiment 90.9 (Fox et al. 2002; Cummings 2005; Cummings and Smedstad 2013; available online at http://www.hycom.org). Ancillary sea surface temperature is available online (www.ncdc.noaa.gov/oisst; Reynolds et al. 2007; Reynolds 2009). Both ancillary fields have a spatial resolution of 0.25° and a temporal resolution of 1 day.
The resulting wind-induced excess surface brightness temperature DTB,rough(W) for a rough surface is shown in Fig. 1 for two polarization states: V-pol and H-pol. The error bars in Fig. 1 are largely caused by sampling mismatches between SMAP TB and WindSat observations. Because of the large number of matchups that went into the derivation of the GMF, random noise is negligible. The SMAP GMF in Fig. 1 is consistent with the GMF for the Aquarius radiometer (Meissner et al. 2014b). Aquarius and SMAP share the same center frequency of 1.41 GHz.
SMAP wind-induced surface brightness temperature ∆TB,rough(W) at a reference surface temperature of 290 K for vertical (V-pol) and horizontal (H-pol) polarization as a function of surface wind speed W. The figure shows the difference between the TB of a rough ocean surface and the TB of a flat ocean surface. The symbols are the actual SMAP TB measurements from the SMAP TB–WindSat wind speed matchup data. The lines are the functional forms of the GMF that were fitted to the data. These GMFs for ∆TB,rough(W) are used in the SMAP wind speed retrievals. The error bars indicate the one-sigma standard deviation of the observed ∆TB,rough in each of the 1 m s–1 bins of the wind speed matchup data.
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
Wind speed retrievals.
Once the SMAP wind-induced GMF has been determined, we can retrieve ocean surface wind speeds from the observed TB,surf. The best wind speed estimate is obtained from a maximum likelihood estimator (MLE) that minimizes the difference between the measured TB,surf and the GMF for both V-pol and H-pol channels. To calculate the GMF for the MLE, we use the same ancillary sea surface salinity and sea surface temperature input fields as we did in the derivation of the GMF expressions. The SMAP wind speed retrieval algorithm does not rely on daily input from NWP models. The retrieved SMAP wind speeds are publicly available online (at www.remss.com/SMAP) with an average delay of about 7 h. They are resampled to a 0.25° grid and separated into ascending (PM) and descending (AM) swaths (Meissner et al. 2016c).
We have evaluated the performance of the retrieved wind speeds by comparing them with wind speeds from the collocated rain-free WindSat matchup set. The matchups that are used during testing are different from the ones that we have used for developing the GMF. The difference between SMAP and WindSat wind speeds yields a global RMS of about 1.5 m s–1 for rain-free ocean scenes. This means that the global accuracy of the SMAP wind speeds matches approximately that of most numerical weather prediction models, such as from the National Centers for Environmental Prediction (NCEP) or the European Centre for Medium-Range Weather Forecasts (ECMWF; Meissner et al. 2001). However, the global accuracy of the SMAP radiometer wind speeds is not as good as that of most other passive [WindSat, Special Sensor Microwave Imager (SSM/I), Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI), Global Precipitation Measurement (GPM) Microwave Imager (GMI), and AMSR] and active (ASCAT, QuikSCAT, and RapidScat) spaceborne sensors that measure ocean surface winds, whose accuracies are typically 1.0 m s–1 or better (Wentz 2015; Ricciardulli and Wentz 2015). The major strength of the SMAP winds lies in the retrieval of high wind speeds, which we will examine next.
HIGH WIND SPEEDS.
Sensitivity of the L-band radiometer at high wind speeds.
At wind speeds above 30 m s–1, the SMAP–WindSat matchup dataset becomes too sparse to derive a reliable GMF. However, the data in Fig. 1 clearly indicate that the GMF for both polarizations increases linearly with wind speeds above 20 m s–1. There is no sign that the L-band wind-induced emissivity signal saturates at high winds. This is likely because the emission from the wind-roughened ocean surface at high wind speeds is mainly caused by sea foam, which does not saturate and is expected to be linear (Nordberg et al. 1971; Monahan and O’Muircheartaigh 1980; Reul and Chapron 2003; Anguelova and Webster 2006). It has been demonstrated that the linear increase of ∆TB,rough at high winds holds true for C-band and higher frequencies (Uhlhorn et al. 2007; Meissner and Wentz 2012). The results from the European SMOS sensor (Reul et al. 2012, 2016) indicate that it also holds true at L band.
To obtain a first guess for the wind-induced GMF above 30 m s–1, we have linearly extrapolated the results from the SMAP–WindSat matchup sets in Fig. 1, which is indicated by the solid lines. We use this linearly extrapolated GMF for retrieving SMAP winds in storms. In the following sections, we will demonstrate how the results compare with other instruments and we will provide verification for our assumption of the linear increase of ∆TB,rough with wind speeds above 30 m s–1 up to at least 65 m s–1.
Major error sources.
Before addressing the details of the validation, we want to briefly discuss the major error sources that go into the retrievals of high wind speeds with SMAP. From the slope of the curves in Fig. 1, we estimate that above 15 m s–1 a radiometric error of 1 K translates roughly into an error of 2.5 m s–1 in the retrieved wind speed.
The dominant error source in the wind speed retrievals is the SMAP radiometer noise, which is approximately 0.7 K for observations over ocean scenes. This explains most of the observed global RMS difference of 1.5 m s–1 between SMAP and WindSat wind speeds. At high wind speeds, the wind directional signal becomes an additional source of error. Its size and impact can again be estimated from the Aquarius results (Yueh et al. 2013; Meissner et al. 2014b), which show a directional amplitude of about 0.6 K at high wind speeds. This translates roughly into a wind speed uncertainty of 1.5 m s–1, if no correction for the wind direction signal is performed.
We also need to assess the impact of systematic errors in the ancillary fields for SSS that are used in calculating the brightness temperature of a flat ocean surface TB. The ancillary SSS field from the HYCOM model, which is largely based on in situ salinity measurements by Argo drifters, has a high global accuracy of about 0.25 practical salinity units (psu), and its error impact on the wind speed retrievals can generally be neglected. However, there are few instances where the HYCOM SSS values do not accurately represent the SSS within the upper few centimeters of the ocean surface, which is sensed by the SMAP L-band radiometer. This is the case in areas with freshwater outflows from large rivers into the ocean, as, for example, the Amazon, Congo, or Ganges Rivers. Another important example is freshening in heavy rain, which leads to salinity stratification in the upper-ocean layer and therefore to significant inaccuracy in the ancillary HYCOM SSS input. This freshwater lensing effect is important at low wind speeds. However, with increasing wind speed the upper-ocean layer becomes increasingly mixed and therefore the SSS stratification becomes less important. At a wind speed of 12 m s–1, the rain freshening effect is estimated to be only to 0.07 psu (mm h–1)−1, leading to an error in wind speed of about 1 m s–1 for rain rates of 10 mm h−1. Above 20 m s–1 this effect is negligible (Meissner et al. 2014c; Boutin et al. 2016). Therefore, the error impact due to SSS stratification in the upper-ocean layer does not constitute a significant source of error for retrieving high wind speeds.
Other potential error sources are wave height or other parameters characterizing the sea state. Though they have been shown to be small below 20 m s–1 (Meissner et al. 2014b), it is unknown if they could play a role in hurricane-force winds. Very heavy precipitation results in a small wind speed bias, which we will discuss in more detail in the section about rain impact.
VERIFICATION: COMPARISON BETWEEN SMAP AND SFMR.
A challenging aspect of satellite retrievals at high winds is verifying them versus ground-truth observations. The most reliable in situ measurements in tropical cyclones uses dropsondes from aircraft to infer the wind speed based on the displacement of the sonde, which is tracked via GPS. Unfortunately, these measurements are sparse, and it is difficult to use them directly to validate our SMAP high wind speed retrievals. However, measurements taken by airborne SFMR provide a very good connection between the sparse dropsonde data and the observed SMAP wind fields. Because of this, we have chosen to use SFMR to validate the SMAP high wind speeds. As we have not used SFMR measurements in either developing the GMF or at any stage in the wind speed retrieval algorithm, they allow an independent verification of our results.
The SFMR instrument.
SFMRs are flown regularly on hurricane-penetrating aircrafts by the National Oceanic and Atmospheric Administration (NOAA) and the U.S. Air Force Reserve Command (AFRC) through tropical cyclones in the Atlantic and east Pacific Oceans. One aircraft flight goes through the storm typically two to four times and records the winds near the eyewall and the eye of the cyclone and along radial legs extending outward from the storm center. The SFMR data are publicly distributed by NOAA’s Hurricane Research Division (HRD) of the Atlantic Oceanographic and Meteorological Laboratory (AOML; available online at www.aoml.noaa.gov/hrd/data_sub/).
The SFMR measures brightness temperatures at six closely spaced C-band channels (4–8 GHz, 7.5–3.8 cm). The presence of multiple C-band frequencies enables it to discriminate between wind and rain signals and thus retrieve both ocean surface wind speeds and surface rain rates simultaneously (Uhlhorn et al. 2007). Klotz and Uhlhorn (2014) present an improved retrieval algorithm and thorough validation of the SFMR wind speeds by comparing observations between 1999 and 2012 with wind speed measurements from dropsondes. They find a strong correlation between SFMR and dropsonde winds over a wide wind speed range up to 70 m s–1. They report a linear correlation coefficient of 0.92, a total RMS of 3.9 m s–1, and a linear fit between the two given by the equation WSFMR = 0.98Wdropsonde + 0.73 m s–1. The results of Klotz and Uhlhorn (2014) establish that the SFMR can be used as a connection between the dropsonde in situ observations and the retrieved SMAP wind fields. For our verification, we do not use SFMR observations whose wind speed is below 15 m s–1, as the signal-to-noise ratio in the SFMR measurement becomes unfavorable at lower wind speeds (J. Carswell 2015, personal communication).
Creating SMAP–SFMR matchups: Hurricane Patricia.
The SFMR measurements are processed with a 10-s running mean and at a spatial sampling of approximately 3 km. To make a meaningful comparison with spaceborne observations, it is necessary to resample the SFMR data to the approximate resolution of the satellite measurements. Moreover, the time difference between SFMR and the satellite measurements also needs to be considered. We want to demonstrate the SFMR–SMAP matchup procedure for the case of Hurricane Patricia shortly before its landfall at the Pacific coast of Mexico on 23 October 2015. Figure 2a depicts the wind speeds of the two SFMR flight segments through the tropical cyclone. The time of the first (lower) segment was about 1730 UTC, and the time of the second (upper) segment was about 2030 UTC. The SMAP pass over Patricia occurred at about 1330 UTC. The storm moved about 100 km between the time of the SMAP observation and the first SFMR segment at 1730 UTC. Though 4 h lie between the SMAP and the SFMR passes, the intensity of the storm changed very little. The maximum sustained winds from the best track data of the NOAA National Hurricane Center (NHC; Kimberlain et al. 2016) show a decrease of less than 3% during these 4 h. Figure 2b zooms into the storm center. To create a matchup between SMAP and SFMR, we take the 1730 UTC SFMR segment and shift it onto the SMAP wind field so that storm centers of SFMR and SMAP visually coincide. We then average the SFMR observations along the flight track straight into the same 0.25° cells of the SMAP wind speed observations. The spatial–temporal along-track averages of the SFMR wind speed measurements aim to approximately represent the wind fields that are seen by typical spaceborne missions, including SMAP. Therefore, it is reasonable to use them for creating a SFMR–satellite matchup observation. The time series for the original SFMR, the resampled SFMR, and the SMAP wind speeds along the SFMR track are shown in Fig. 3. They suggest a strong correlation between SMAP and SFMR along the whole SFMR track and over a wide wind speed range (between 15 and 70 m s–1) for that particular storm. The figure highlights that during Patricia the SFMR did record wind speeds above 60 m s–1 over spatial scales of about 25 km, and the maximum winds exceeded 90 m s–1 before resampling. The resulting high wind speeds of the resampled SFMR data should therefore be observable by active and passive spaceborne sensors such as SMAP, whose typical resolutions are between 20 and 50 km.
(a) NOAA SFMR flight segments through Hurricane Patricia on 23 Oct 2015. (b) SMAP wind field and SFMR matchup (black line with circles). To achieve a matchup between SMAP and SFMR, the lower SFMR segment, which is closest in time to the SMAP overpass, is shifted so that the SFMR and SMAP storm centers coincide (black line). The SFMR wind speed observations are then averaged along track into 0.25° boxes and compared with the SMAP wind speed cells (circles).
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
Time series of the SMAP–SFMR matchups from Fig. 2 along the SFMR track. The red dotted lines are the actual SFMR measurements. The red circles on the solid line show the SFMR measurements after averaging them along track into 0.25° boxes. The time values of the x axis refer to the SFMR flight. The SMAP overpass occurred about 4.5 h earlier and has been shifted to the SFMR time.
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
SMAP–SFMR matchup data for tropical cyclones during 2015.
In this analysis, we have created SMAP–SFMR matchup wind speeds for eight tropical cyclones during 2015. We have allowed a maximum time collocation window of up to 4.5 h between the SMAP and the SFMR observation. Because there is only one year of SMAP data available for our analysis, choosing a tighter time window would create an insufficient number of matchup data at high wind speeds. To avoid cases where the storm has changed too much, we estimate the change in intensity between the SMAP and the SFMR overpasses from the maximum sustained winds in the NHC best track data and require that this intensity change does not exceed 7%. In addition, we have excluded data if the storm centers of SMAP and SFMR are more than 100 km apart. We should also note that this matchup procedure between SFMR and SMAP observations becomes inaccurate in areas of very large wind speed gradients as they occur near the eye of the cyclone. Performing a one-dimensional average along the SFMR track puts a relative larger weight to the low wind speeds in the eye than the two-dimensional average over the SMAP antenna gain does. Therefore, close to the eye, the resampled SFMR data are systematically lower than the satellite measurements. To avoid a corresponding sampling error in the matchup data, we have excluded observations for the matchups that fall within the eye of the cyclone or within a corresponding area of the adjacent eyewall.
Statistics for the full matchup dataset are shown in Fig. 4 and depict a strong correlation between SMAP and resampled SFMR wind speeds for these eight storms. Table 1 breaks these matchups into different wind speed regimes and displays statistics for each. The results of Fig. 4 and Table 1 demonstrate that our assumption about the linear increase of the wind-roughened surface emission ∆TB,rough with wind speed holds to a very good approximation. It is worth noting that the uncertainties in Fig. 4 and Table 1 are not solely due to SMAP, but they also include the aforementioned uncertainties of the SFMR as well as sampling mismatch between the two observations. It is reasonable to allocate about half of the errors in Fig. 4 and Table 2 to the SMAP wind speeds. This means that the SMAP L-band radiometer is indeed capable of measuring wind speeds above 15 m s–1 with an accuracy of 11% or better.
Statistics of SMAP–resampled SFMR matchups for eight tropical cyclone overpasses during 2015. The linear regression fit is shown as the solid black line. For reference, perfect correlation is shown as the dashed black line. The total number of matchup scenes is 198. The legend lists the times of the SMAP overpasses. The times in the parentheses indicate the average time difference (h) between the SFMR and the SMAP passes. If this time difference exceeds 1 h, we have also listed the relative change (%) in the maximum storm intensity at the SFMR overpass compared to the SMAP overpass as estimated from the NHC best track data. For a valid matchup, we require that this relative intensity change between the SFMR and the SMAP pass does not exceed 7%.
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
Statistics of SMAP–resampled SFMR matchups for different wind speed regimes. The statistics include only matchups for which both the SMAP and the resampled SFMR wind speeds exceed 15 m s–1. The wind speed range in the first column refers to the average between SMAP and resampled SFMR wind speed.
Statistics of SMAP–resampled SFMR matchups for different rain rates. The SFMR rain rate has been averaged to 0.25°. The statistics include only matchups for which both the SMAP and the resampled SFMR wind speeds exceed 15 m s–1.
SMAP WIND SPEED IN SELECTED INTENSE STORMS.
In the following section, we show three examples of SMAP wind speed retrievals for notably intense tropical cyclones that occurred during the 2015 and early 2016 seasons and compare with results from other sources.
Hurricane Patricia.
Patricia was one of the strongest cyclones that have ever been recorded in the tropical eastern Pacific, reaching its peak intensity at around 1200 UTC 25 October 2015 (Kimberlain et al. 2016), shortly before its landfall off the Mexican coast (Fig. 3). We have used it to demonstrate the SMAP–SFMR validation.
SMAP, ASCAT, RapidScat, and WindSat were all able to observe this storm within 4 h of one another on this day. Figure 5 displays the wind fields for these sensors, for the NCEP Global Data Assimilation System (GDAS) model 10-m winds (at 0.25° resolution or about 25 km; available online at www.nco.ncep.noaa.gov/pmb/products/gfs/), and for WindSat rain rates. The size of the storm can be characterized by the average hurricane-force wind radius (33 m s–1). For both SMAP (Fig. 5a) and ASCAT (Fig. 5b) this was about 50 km, consistent with the NHC best track value of 46 km. However, ASCAT observed maximum winds of 37 versus 65 m s–1 measured by SMAP. The SFMR results in Fig. 3 leave little doubt that in Patricia wind speeds reached values of about 70 m s–1 on spatial scales of 20–50 km. The significantly lower ASCAT maximum winds are a sign of degradation due to its decreased sensitivity at extreme winds. The maximum winds of WindSat (Fig. 5c) and RapidScat (Fig. 5d) both stayed below 33 m s–1. Intense rain (Fig. 5f) affects these two sensors by attenuating the backscatter signal in RapidScat and causing spurious “rainbands” in the WindSat wind field (see sidebars on scatterometers and WindSat). The 1200 UTC NCEP GDAS 0.25° 10-m winds (Fig. 5e) show a maximum of 70 m s–1 but also a very large hurricane-force wind radius of about 100 km. Figure 6 shows histograms of the collocated wind speed distributions of all four sensors. Below 30 m s–1, the four distributions are in reasonably good agreement. Only SMAP was able to see wind speeds much higher than 37 m s–1. SMAP is therefore the only observed wind field that accurately captures both storm peak intensity and hurricane-force wind radius.
SCATTEROMETERS
Scatterometers are spaceborne radars used to measure the ocean surface wind speed and direction. They operate at microwave frequencies with wavelengths of about 2 (Ku band), 5 (C band), and 20 cm (L band). The incident wave resonates with ocean capillary waves, whose wavelengths range from a few millimeters to a few centimeters, via Bragg scattering. A signal is backscattered to the scatterometer. The intensity of the backscatter relative to the transmitted intensity depends on the wind speed and direction and on frequency, polarization, and Earth incidence angle. The longest scatterometer mission to date has been the NASA QuikSCAT (Ku band, VV-pol, and HH-pol) operating from 1999 to 2009, which has been very useful for wind data assimilation in NWP models. Scatterometer missions considered in this paper are the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT) ASCAT (C band, VV-pol, on MetOp-A) and the NASA RapidScat (Ku band, VV-pol, and HH-pol) mounted on the International Space Station.
The scatterometer wind retrieval algorithms are based on empirical models (GMFs), which are developed by matching the observed backscatter signal as a function of frequency, polarization, and Earth incidence angle to collocated ground-truth wind measurements. At wind speeds below 30 m s–1, the ground truth is represented by buoys, by other well-calibrated satellite measurements, or by NWP models, for example, NCEP GDAS or ECMWF. For very high winds (above 30 m s–1) the GMF is typically extrapolated and tuned to make the retrieved winds match winds in storms observed by aircraft-mounted instruments or to the H*Wind field analyses from the National Hurricane Center (Powell et al. 1998; DiNapoli et al. 2012).
Scatterometer measurements are impacted by rain. At winds greater than 20 m s–1 rain results in a negative bias in the wind measurements due to the attenuation of the direct and backscattered signal. All rain effects are proportional to the rain intensity, and they affect high frequencies (Ku band) much more than lower ones (C band). The effect at L band seems to be negligible. The rain bias at a wind speed of 30 m s–1 and at a rain rate of 8 mm h–1 is approximately 5 m s–1 for Ku band and about 1 m s–1 for C band (Ricciardulli and Wentz 2015; Ricciardulli et al. 2015).
Often, observations from each scatterometer are processed by more than one institution, using slightly different GMFs and calibration targets as well as different wind algorithms. In this article, we use the Remote Sensing Systems (RSS) ASCAT version 2 data that are output on a 12.5-km grid with a 25-km resolution (Ricciardulli 2016; Ricciardulli and Wentz 2016). The RapidScat data we use are the level 2B ocean wind vectors in 12.5-km slice composites version 1.1 (available online from the NASA Physical Oceanography Distributed Active Archive at http://dx.doi.org/10.5067/RSX12-L2B11). The actual resolution of these data is approximately 20 km.
WINDSAT
The spaceborne WindSat polarimetric radiometer (Gaiser et al. 2004) was developed by the Naval Research Laboratory (NRL) and operates at five frequencies between C band (6.8 GHz, 4.4 cm) and Ka band (37.0 GHz, 0.8 cm). It has been operating since 2003 with just a few brief interruptions. WindSat has been providing the meteorological and oceanographic communities with high-quality observations of ocean surface wind speeds and direction since 2003.
Our study uses data from the RSS version 7.0.1 WindSat Ocean Suite (Wentz et al. 2013). For developing the GMF of the SMAP rough surface emission (Fig. 1), we have used rain-free WindSat wind speeds at a resolution of 32 km. The simultaneous presence of C-band (6.8 GHz, 4.4 cm) and X-band (10.7 GHz, 28 cm) channels provides the WindSat sensor with limited capability to measure wind speeds through rain (Meissner and Wentz 2009; Meissner et al. 2011) using a wind retrieval algorithm that has been trained with H*Winds (Powell et al. 1998). The WindSat all-weather wind speeds have been used for the storm comparisons in Fig. 5 (Hurricane Patricia) and the time series of Tropical Cyclone Winston (Fig. 7). Extreme values of wind speed (>40 m s–1) and rain rate (>10 mm h–1) have not been included in the dataset with which the statistical WindSat all-weather wind algorithm has been trained. As a consequence, the WindSat all-weather algorithm is less accurate in very intense tropical cyclones. An all-weather algorithm has also been developed for the JAXA AMSR2 radiometer (Zabolotskikh et al. 2015), which has channels at two different C-band frequencies in addition to the X band but has not been used in this analysis.
Hurricane Patricia on 23 Oct 2015 from various sources. Wind speeds from (a) SMAP, (b) RSS ASCAT, (c) RapidScat, (d) WindSat all-weather, (e) NCEP GDAS 0.25°, and (f) rain rate from WindSat.
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
Histograms of collocated wind speeds for SMAP, RSS ASCAT, RapidScat, and WindSat for Hurricane Patricia on 23 Oct 2015. The bin size is 2 m s–1. The figure counts only wind cells in a 0.25° grid for which all four sensors have valid observations. The times indicate the average time of the satellite passes. Among all four sensors, only SMAP exceeds 36 m s–1 reaching a maximum of 64 m s–1.
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
Time series of various assessments of the maximum wind speed in Tropical Cyclone Winston during the 15-day period from 10 to 25 Feb 2016. Black line and triangles are NCEP GDAS 0.25° model output. Red circles indicate SMAP peak winds. Blue squares are RSS ASCAT peak winds. Green diamonds indicate RSS WindSat all-weather peak winds. Purple dashed lines and triangles are the maximum 10-min sustained wind speed from the JTWC storm warning.
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
Tropical Cyclone Winston.
On 20 February 2016, the Fiji Islands were hit by Winston, one of the most severe tropical cyclones in history. With 1-min maximum sustained winds reaching 80 m s–1, Winston was second in strength only to Super Typhoon Haiyan. Winston persisted for more than 2 weeks, taking an unusual track that included a sudden reversal of direction and two different periods of rapid intensification.
As is often the case for Southern Hemisphere storms, due to the remote storm locations, there were no in situ wind observations available from hurricane hunters’ aircrafts or dropsondes. In these cases, forecasters rely on visible and infrared satellite imagery of the storm and its evolution using the Dvorak technique (Velden et al. 2006). SMAP, ASCAT, RapidScat, and WindSat gave important information about the storm intensity and its track. We have compared the intensities that were measured by the four satellites with maximum sustained winds from the storm warnings and preliminary best track data of the Joint Typhoon Warning Center (JTWC) (available online at https://metoc.ndbc.noaa.gov/jtwc or the links at https://en.wikipedia.org/wiki/Cyclone_Winston). Without SFMR flights, there is no in situ spatial information available in this case, and a direct comparison for the winds that are seen by the spaceborne sensors cannot be made. To be comparable to the surface roughness of the satellite observation, the surface winds need to be sustained for at least 10 min. To make a more meaningful comparison, we have therefore scaled the 1-min sustained winds from the JTWC storm warnings to 10-min sustained winds using the scale factor of 0.93, as recommended by the World Meteorological Organization (Harper et al. 2010). In the cases when SFMR measurements are available, this scaling is not done, as its effect is already included in the spatial averaging.
Figure 7 shows the time series of maximum wind speeds of three satellites: the JTWC 10-min maximum sustained winds together with the maximum 10-m winds from the 0.25°NCEP GDAS. These data are displayed for the whole duration of the storm (10–26 February 2016). Overall, the temporal evolution of the maximum wind speed observed by SMAP appears consistent with the preliminary best track data over the whole duration of the storm, with the exception of a couple of outliers. The SMAP intensity is also consistent with the NCEP GDAS model output. Above 55 m s–1, both SMAP and NCEP GDAS show lower maximum wind speeds compared to the preliminary JTWC best track data. Despite these differences, SMAP, preliminary JTWC best track data, and NCEP GDAS exhibit the two major intensification stages of the storm. SMAP observations reached winds well above 50 m s–1. ASCAT, WindSat, and RapidScat (not shown in Fig. 7) seem to consistently reach a maximum of only about 35–38 m s–1, and they do not clearly exhibit the two major intensification stages.
Tropical Cyclone Fantala.
The category 5 Tropical Cyclone Fantala recorded the strongest sustained winds in the southwestern Indian Ocean. The storm warning by the Tropical Cyclone Center in La Reunion at 1800 UTC 17 April 2016 (available online at www.meteo.fr/temps/domtom/La_Reunion/webcmrs9.0/anglais/) indicates the maximum 10-min sustained winds of 69 m s–1 and a hurricane wind radius of approximately 63 km. The value for the intensity was again based on estimates from the Dvorak technique. From the devastation that the storm caused when passing over the atoll of Farquhar (Seychelles; see picture on opening page) at around 1500 UTC, the Tropical Cyclone Center estimates that there must have been gusts in excess of 77 m s–1 (P. Caroff 2016, personal communication). SMAP (Fig. 8a) and ASCAT (Fig. 8b) were able to observe the storm close to the time when it reached its maximum intensity. The wind fields of both satellites are shown in Fig. 8. SMAP has a maximum wind speed of 70 m s–1 (category 5, according to the Saffir–Simpson hurricane wind scale) and a hurricane-force wind radius of about 55 km. ASCAT shows a much weaker and smaller storm with a maximum wind speed of only 35 m s–1 (category 1, according to the Saffir–Simpson hurricane wind scale) and a hurricane wind radius of only about 10 km. For comparison, the values of the NCEP GDAS 0.25° analysis (not shown) were 66 m s–1 and 110 km, respectively. As we have observed in several other cyclones, the 0.25° NCEP GDAS gives reasonable estimates for the maximum winds but the hurricane-force wind radii tend to be too large.
Wind speed field of Tropical Cyclone Fantala on 17 Apr 2016 as seen by (a) SMAP and (b) RSS ASCAT.
Citation: Bulletin of the American Meteorological Society 98, 8; 10.1175/BAMS-D-16-0052.1
RAIN IMPACT.
Next to its sensitivity over a very wide wind speed range, another strength that L-band radiometers have over sensors at higher frequencies in regard to measuring winds in tropical cyclones is the fact that at L-band frequencies high winds are minimally impacted by rain. We have computed the statistics of the SMAP–resampled SFMR matchup data in different rain-rate regimes using the rain rate measured by SFMR. As we had done for the SFMR wind speeds, we have also resampled the SFMR rain-rate values by averaging the data that were recorded at high spatial and temporal resolution into 0.25° grid cells in order to approximately represent the rain rate that is seen by SMAP. These results are shown in Table 2. The bias and standard deviation do not indicate any systematic degradation of the wind speed statistics for SMAP versus the resampled SFMR wind speeds with increasing rain rate. Even within the bin with very high rain rates (>15 mm h–1) we do not observe a significant variability of the wind speed difference between SMAP and SFMR.
The reason for the negligible effect of rain onto the SMAP wind speed is rooted in the radiative transfer properties of L-band radiation. The atmospheric attenuation by water drops is very small at L band, even in heavy rain, when compared to the higher frequencies typical of other satellite sensors (Wentz 2005; Reul et al. 2012). Wentz (2005) has shown that at the extreme rain rate of 30 mm h–1 the L-band brightness temperature change due to cloud liquid water absorption is only 0.3–0.4 K, resulting in a systematic bias of about +1.0 m s–1 in the retrieved wind speeds. This is a relatively small contribution to the total error statistics at higher winds, about an order of magnitude smaller than at Ku band (see sidebar on “Scatterometers”). For frozen precipitation, such as cloud ice, snow, hail, or graupel, the absorption at L band is at least an order of magnitude smaller than for cloud liquid water (Ulaby and Long 2013), and the impact on wind speed is negligible. Rain roughening of the ocean surface through splashing effects is a potential error source at lower wind speeds but plays little or no role at high winds (Tang et al. 2013; Boutin et al. 2014, 2016).
DISCUSSION AND CONCLUSIONS.
In this paper, we have demonstrated the capability of the spaceborne SMAP radiometer at the L-band frequency to give accurate estimates of the intensity and radii of hurricane-force winds. We have shown that the SMAP wind speeds are impacted very little by precipitation, even at high rain rates.
Current wind measurements from space are taken from microwave radiometers and scatterometers at higher frequencies (C and Ku band). These measurements are often used in NWP models and for predicting the evolution of extreme events. However, there are limitations for these scatterometers when it comes to tropical cyclones. Copolarized (VV-pol and HH-pol) radar backscatter loses sensitivity at high wind speeds and starts to saturate at hurricane-force winds (Donelan et al. 2004; Hwang et al. 2013; Hwang and Fois 2015; Sapp et al. 2016), which makes accurate estimates of the intensity difficult. Another limitation is the radar signal attenuation in intense rain at hurricane-force winds, which is severe at Ku band and, despite being lower, still relevant at C band (see sidebar on “Scatterometers”).
The SMAP wind speed retrievals are based on an RTM for the microwave emission of a wind-roughened ocean surface. At wind speeds above 15 m s–1, the microwave emission is mainly caused by whitecaps. The sea surface emissivity signal appears to increase linearly above 20 m s–1 and shows no sign of saturation at any wind speed. The SMAP radiometer winds show very good correlation with the airborne SFMR observations over a wide wind speed range between 15 and 65 m s–1, even in intense rain. We have compared SMAP wind fields for some recent storms (Patricia, Winston, and Fantala) with the wind fields observed by three other space-based missions (ASCAT, WindSat, and RapidScat). In these storms and at the satellite footprint scales of 20–50 km, none of the other instruments is able to observe wind speeds greater than 37 m s–1, while SMAP winds reach 55–70 m s–1 for all of them. This difference is much higher than the 10% uncertainty associated with the satellite measurements at this wind speed range.
Though not the focus of this article, we note that SMAP has a limited capability to measure wind direction at wind speeds above 15 m s–1 aided by its polarimetric channels. To reduce the noise, it is necessary to decrease the spatial resolution to about 100 km and thus is mainly useful in large extratropical cyclones.
The capabilities of SMOS and SMAP to measure extreme winds are very promising for predicting the evolution of extreme storms. Being spaceborne, SMOS and SMAP also have the advantage of observing remote locations on a large scale. This could be of critical importance for forecasting storms in remote locations, where direct measurements of wind speeds are scarce, as was the case for the storms Winston and Fantala presented here. When both airborne and satellite measurements are available, airborne measurements and dropsondes will always be considered as the ultimate ground truth for verification of the space-based observations of extreme winds.
We regard the L-band radiometers SMOS and SMAP as complementary to the current spaceborne scatterometer missions. Scatterometers provide better accuracy for both wind speed and direction at winds below 30 m s–1. The scatterometers can also achieve better spatial resolution than the passive sensors and thus measure winds closer to the coast. However, scatterometer-retrieved wind speeds have higher uncertainties and biases in intense storms.
It is desirable to increase the validation database of SMAP high-wind retrievals with additional SFMR matchups, as they become available. This could lead to some fine-tuning of the SMAP wind-induced emissivity signal. Once completed, SMAP wind speeds can become a “space-based ground truth” for winds in tropical and extratropical storms. We expect this dataset to be useful for calibrating the Cyclone Global Navigation Satellite System (CYGNSS), a NASA constellation of small GPS satellites that are specifically designed to measure tropical cyclone winds in rain, with the purpose of improving prediction of extreme events (Ruf et al. 2013, 2016). CYGNSS was successfully launched in December 2016. Similarly, we expect SMAP winds to be useful for calibration/validation of the European Space Agency (ESA) cross-polarized (VH-pol) synthetic aperture radar (SAR) Sentinel-1, which is also capable of measuring hurricane-force wind speeds (Mouche and Chapron 2015).
ACKNOWLEDGMENTS
This research was supported by the NASA SMAP Science Utilization Program SUSMAP (Contact NNH17CA04C) and the NASA Ocean Vector Winds Science Team OVWST (Contract NNH14CM09C). We thank the reviewers for their comments and suggestions, which helped to improve the manuscript. We are thankful to Philippe Caroff (Tropical Cyclone Center La Reunion) for information about Tropical Cyclone Fantala and to Brad Klotz (University of Miami/CIMAS and NOAA/HRD), Jim Carswell (Remote Sensing Solutions), Paul Chang, and Zorana Jelenak (NOAA/NESDIS/STAR) for useful discussions about the SFMR measurements. The SFMR data used in this validation study are available at the NOAA Hurricane Research Division of the Atlantic Oceanographic and Meteorological Laboratory (AOML).
REFERENCES
Anguelova, M., and F. Webster, 2006: Whitecap coverage from satellite measurements: A first step toward modeling the variability of oceanic whitecaps. J. Geophys. Res., 111, C03017, doi:10.1029/2005JC003158.
Atlas, R., R. Hoffman, S. Leidner, and J. Sienkiewicz, 2001: The effects of marine winds from scatterometer data on weather analysis and forecasting. Bull. Amer. Meteor. Soc., 82, 1965–1990, doi:10.1175/1520-0477(2001)082<1965:TEOMWF>2.3.CO;2.
Bourassa, M., and Coauthors, 2010: Remotely sensed winds and wind stresses for marine forecasting and ocean modeling. Proc. OceanObs’09, Venice, Italy, ESA. [Available online atwww.oceanobs09.net/proceedings/cwp/Bourassa-OceanObs09.cwp.08.pdf.]
Boutin, J., N. Martin, G. Reverdin, S. Morisset, X. Yin, L. Centurioni, and N. Reul, 2014: Sea surface salinity under rain cells: SMOS satellite and in situ drifters observations. J. Geophys. Res. Oceans, 119, 5533–5545, doi:10.1002/2014JC010070.
Boutin, J., and Coauthors, 2016: Satellite and in situ salinity: Understanding near-surface stratification and subfootprint variability. Bull. Amer. Meteor. Soc., 97, 1391–1407, doi:10.1175/BAMS-D-15-00032.1.
Brennan, M., C. Hennon, and R. Knabb, 2009: The operational use of QuikSCAT ocean surface vector winds at the National Hurricane Center. Wea. Forecasting, 24, 621–644, doi:10.1175/2008WAF2222188.1.
Candy, B., S. English, and S. Keogh, 2009: A comparison of the impact of QuikScat and WindSat wind vector products on Met Office analyses and forecasts. IEEE Trans. Geosci. Remote Sens., 47, 1632–1640, doi:10.1109/TGRS.2008.2009993.
Chang, P., Z. Jelenak, J. Sienkiewicz, R. Knabb, M. Brennan, D. Long, and M. Freeberg, 2009: Operational use and impact of satellite remotely sensed ocean surface vector winds in the marine warning and forecasting environment. Oceanography, 22, 194–207, doi:10.5670/oceanog.2009.49.
Chelton, D., M. Freilich, J. Sienkiewicz, and J. von Ahn, 2006: On the use of QuikSCAT scatterometer measurements of surface winds for marine weather prediction. Mon. Wea. Rev., 134, 2055–2071, doi:10.1175/MWR3179.1.
Cummings, J., 2005: Operational multivariate ocean data assimilation. Quart. J. Roy. Meteor. Soc., 131, 3583–3604, doi:10.1256/qj.05.105.
Cummings, J., and O. Smedstad, 2013: Variational data assimilation for the global ocean. Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications, Vol. II, Springer-Verlag, 303–343, doi:10.1007/978-3-642-35088-7_13.
DiNapoli, S., M. Bourassa, and M. Powell, 2012: Uncertainty and intercalibration analysis of H*Wind. J. Atmos. Oceanic Technol., 29, 822–833, doi:10.1175/JTECH-D-11-00165.1.
Donelan, M., B. Haus, N. Reul, W. Plant, M. Stiassnie, H. Graber, O. Brown, and E. Saltzman, 2004: On the limiting aerodynamic roughness of the ocean in very strong winds. Geophys. Res. Lett., 31, L18306, doi:10.1029/2004GL019460.
Draper, D., and D. Long, 2004: Evaluating the effect of rain on SeaWinds scatterometer measurements. J. Geophys. Res., 109, C02005, doi:10.1029/2002JC001741.
Entekhabi, D., and Coauthors, 2010: The Soil Moisture Active Passive (SMAP) mission. Proc. IEEE, 98, 704–716, doi:10.1109/JPROC.2010.2043918.
Entekhabi, D., and Coauthors, 2014: SMAP Handbook. National Aeronautics and Space Administration, 192 pp. [Available online at https://smap.jpl.nasa.gov/.]
Figa-Saldaña, J., J. Wilson, E. Attema, R. V. Gelsthorpe, M. Drinkwater, and A. Stoffelen, 2002: The Advanced Scatterometer (ASCAT) on the Meteorological Operational (MetOp) platform: A follow on for European wind scatterometers. Can. J. Remote Sens., 28, 404–412, doi:10.5589/m02-035.
Fox, D., W. Teague, C. Barron, M. Carnes, and C. Lee, 2002: The Modular Ocean Data Assimilation System (MODAS). J. Atmos. Oceanic Technol., 19, 240–252, doi:10.1175/1520-0426(2002)019<0240:TMODAS>2.0.CO;2.
Franklin, J., M. Black, and K. Valde, 2003: GPS dropwindsonde wind profiles in hurricanes and their operational implications. Wea. Forecasting, 18, 32–44, doi:10.1175/1520-0434(2003)018<0032:GDWPIH>2.0.CO;2.
Gaiser, P., and Coauthors, 2004: The WindSat spaceborne polarimetric microwave radiometer: Sensor description and early orbit performance. IEEE Trans. Geosci. Remote Sens., 42, 2347–2361, doi:10.1109/TGRS.2004.836867.
Harper, B. A., J. D. Kepert, and J. D. Ginger, 2010: Guidelines for converting between various wind averaging periods in tropical cyclone conditions. World Metrological Organization WMO/TD 1555, 64 pp. [Available online at www.wmo.int/pages/prog/www/tcp/documents/WMO_TD_1555_en.pdf.]
Hilburn, K., and F. Wentz, 2008: Intercalibrated passive microwave rain products from the Unified Microwave Ocean Retrieval Algorithm (UMORA). J. Appl. Meteor. Climatol., 47, 778–794, doi:10.1175/2007JAMC1635.1.
Hilburn, K., F. Wentz, D. Smith, and P. Ashcroft, 2006: Correcting active scatterometer data for the effects of rain using passive microwave data. J. Appl. Meteor. Climatol., 45, 382–398, doi:10.1175/JAM2357.1.
Hock, T., and J. Franklin, 1999: The NCAR GPS dropwindsonde. Bull. Amer. Meteor. Soc., 80, 407–420, doi:10.1175/1520-0477(1999)080<0407:TNGD>2.0.CO;2.
Hwang, P., and F. Fois, 2015: Surface roughness and breaking wave properties from polarimetric microwave radar backscattering. J. Geophys. Res. Oceans, 120, 3640–3657, doi:10.1002/2015JC010782.
Hwang, P., D. Burrage, D. Wang, and J. Wesson, 2013: Ocean surface roughness spectrum in high wind condition for microwave backscatter and emission computations. J. Atmos. Oceanic Technol., 30, 2168–2188, doi:10.1175/JTECH-D-12-00239.1.
Imaoka, K., M. Kachi, M. Kasahara, N. Ito, K. Nakagawa, and T. Oki, 2010: Instrument performance and calibration of AMSR-E and AMSR2. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci., 38, 13–18. [Available online at www.isprs.org/proceedings/XXXVIII/part8/pdf/JTS13_20100322190615.pdf.]
Isaksen, L., and A. Stoffelen, 2000: ERS-Scatterometer wind data impact on ECMWF’s tropical cyclone forecasts. IEEE Trans. Geosci. Remote Sens., 38, 1885–1892, doi:10.1109/36.851771.
Jones, W., L. Schroeder, and J. Mitchell, 1977: Aircraft measurements of the microwave scattering signature of the ocean. IEEE Trans. Antennas Propag., 25, 52–61, doi:10.1109/TAP.1977.1141533.
Jones, W., P. Black, V. Delnore, and C. Swift, 1981: Airborne microwave remote-sensing measurements of Hurricane Allen. Science, 214, 279–280.
Jones, W., L. Schroeder, D. Boggs, E. Bracalente, R. Brown, G. Dome, W. Pierson, and F. Wentz, 1982: The SEASAT-A satellite scatterometer: The geophysical evaluation of remotely sensed wind vectors over the ocean. J. Geophys. Res., 87, 3297–3317, doi:10.1029/JC087iC05p03297.
Katsaros, K., E. Forde, P. Chang, and W. Liu, 2001: QuikSCAT’s SeaWinds facilitates early identification of tropical depressions in 1999 hurricane season. Geophys. Res. Lett., 28, 1043–1046, doi:10.1029/2000GL011646.
Katsaros, K., P. Vachon, W. Liu, and P. Black, 2002: Microwave remote sensing of tropical cyclones from space. J. Oceanogr., 58, 137–151, doi:10.1023/A:1015884903180.
Kerr, Y., and Coauthors, 2010: The SMOS mission: New tool for monitoring key elements of the global water cycle. Proc. IEEE, 98, 666–687, doi:10.1109/JPROC.2010.2043032.
Kimberlain, T., E. Blake, and J. Cangialosi, 2016: National Hurricane Center tropical cyclone report (TCR): Hurricane Patricia. NOAA/NWS Rep. EP 202015, 32 pp. [Available online at www.nhc.noaa.gov/data/tcr/EP202015_Patricia.pdf.]
Klotz, B., and E. Uhlhorn, 2014: Improved stepped frequency microwave radiometer tropical cyclone surface winds in heavy precipitation. J. Atmos. Oceanic Technol., 31, 2392–2408, doi:10.1175/JTECH-D-14-00028.1.
LeVine, D., E. Dinnat, T. Meissner, S. Yueh, F. Wentz, S. Torrusio, and G. Lagerloef, 2015: Status of Aquarius/SAC-D and Aquarius salinity retrievals. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 8, 5401–5415, doi:10.1109/JSTARS.2015.2427159.
Meissner, T., and F. Wentz, 2004: The complex dielectric constant of pure and sea water from microwave satellite observations. IEEE Trans. Geosci. Remote Sens., 42, 1836–1849, doi:10.1109/TGRS.2004.831888.
Meissner, T., and F. Wentz, 2009: Wind vector retrievals under rain with passive satellite microwave radiometers. IEEE Trans. Geosci. Remote Sens., 47, 3065–3083, doi:10.1109/TGRS.2009.2027012.
Meissner, T., and F. Wentz, 2012: The emissivity of the ocean surface between 6 and 90 GHz over a large range of wind speeds and Earth incidence angles. IEEE Trans. Geosci. Remote Sens., 50, 3004–3026, doi:10.1109/TGRS.2011.2179662.
Meissner, T., and F. Wentz, 2016a: RSS SMAP level 2C sea surface salinity V2.0 validated dataset. Remote Sensing Systems, accessed July 2016, doi:10.5067/SMP20-2SOCS.
Meissner, T., and F. Wentz, 2016b: RSS SMAP salinity: Version 2 validated release. Release Notes Algorithm Theoretical Basis Document (ATBD) Validation Data Format Specification, RSS Tech. Rep. 091316, 23 pp. [Available online at ftp://podaac-ftp.jpl.nasa.gov/allData/smap/docs/V2/RSS_SMAP-SSS_V2.0_TechnicalDocumentation.pdf.]
Meissner, T., D. Smith, and F. Wentz, 2001: A 10 year intercomparison between collocated Special Sensor Microwave Imager oceanic surface wind speed retrievals and global analyses. J. Geophys. Res., 106, 11 731–11 742, doi:10.1029/1999JC000098.
Meissner, T., L. Ricciardulli, and F. Wentz, 2011: All-weather wind vector measurements from intercalibrated active and passive microwave satellite sensors. Proc. 2011 IEEE Int. Geoscience and Remote Sensing Symp. (IGARSS), Vancouver, BC, Canada, IEEE, doi:10.1109/IGARSS.2011.6049354.
Meissner, T., F. Wentz, D. LeVine, and J. Scott, 2014a: Aquarius salinity retrieval. Algorithm Theoretical Basis Document, Addendum 3, Remote Sensing Systems, 22 pp. [Available online at ftp://podaac-ftp.jpl.nasa.gov/allData/aquarius/docs/v3/AQ-014-PS-0017_AquariusATBD_Level2_Addendum3_DatasetVersion3.0.pdf.]
Meissner, T., F. Wentz, and L. Ricciardulli, 2014b: The emission and scattering of L-band microwave radiation from rough ocean surfaces and wind speed measurements from the Aquarius sensor. J. Geophys. Res. Oceans, 119, 6499–6522, doi:10.1002/2014JC009837.
Meissner, T., F. Wentz, J. Scott, and K. Hilburn, 2014c: Assessment of rain impact on the Aquarius salinity retrievals. Proc. Ocean Salinity Science Workshop, Exeter, United Kingdom, ESA. [Available online at www.smos-sos.org/presentations-ocean-salinity-science-workshop.]
Meissner, T., F. Wentz, D. LeVine, and P. De Mattheis, 2015: Aquarius salinity retrieval. Algorithm Theoretical Basis Document, Addendum 4, Remote Sensing Systems Rep. 071515, 16 pp. [Available online at ftp://podaac-ftp.jpl.nasa.gov/allData/aquarius/docs/v4/AQ-014-PS-0017_AquariusATBD_Level2_Addendum4_DatasetVersion4.0.pdf.]
Meissner, T., L. Ricciardulli, and F. Wentz, 2016: Remote sensing systems SMAP daily sea surface winds speeds on 0.25 deg grid, version 0.1. (BETA). Remote Sensing Systems, accessed July 2016. [Available online at www.remss.com/missions/smap/winds.]
Monahan, E., and I. O’Muircheartaigh, 1980: Optimal power-law description of oceanic whitecap coverage dependence on wind speed. J. Phys. Oceanogr., 10, 2094–2099, doi:10.1175/1520-0485(1980)010<2094:OPLDOO>2.0.CO;2.
Mouche, A., and B. Chapron, 2015: Global C-band Envisat, RADARSAT-2 and Sentinel-1 SAR measurements in copolarization and cross-polarization. J. Geophys. Res. Oceans, 120, 7195–7207, doi:10.1002/2015JC011149.
Nordberg, W., J. Conaway, D. Ross, and T. Wilheit, 1971: Measurement of microwave emission from a foam-covered, wind-driven sea. J. Atmos. Sci., 28, 429–433, doi:10.1175/1520-0469(1971)028<0429:MOMEFA>2.0.CO;2.
Piepmeier, J., P. Mohammed, G. De Amici, E. Kim, J. Peng, and C. Ruf, 2014: SMAP calibrated, time-ordered brightness temperatures L1B_TB data product. Algorithm Theoretical Basis Document Revision A, 83 pp. [Available online at http://nsidc.org/data/docs/daac/smap/sp_l1b_tb/pdfs/278_L1B_TB_RevA_web.pdf.]
Piepmeier, J., P. Mohammed, J. Peng, E. J. Kim, G. De Amici, and C. Ruf, 2016: SMAP L1B radiometer half-orbit time-ordered brightness temperatures, version 3 [RFI-filtered antenna temperatures]. National Snow and Ice Data Center, accessed July 2016, doi:10.5067/YV5VOWY5V446.
Portabella, M., A. Stoffelen, W. Lin, A. Turiel, A. Verhoef, J. Verspeek, and J. Ballabera-Poy, 2012: Rain effects on ASCAT-retrieved winds: Toward an improved quality control. IEEE Trans. Geosci. Remote Sens., 50, 2495–2506, doi:10.1109/TGRS.2012.2185933.
Powell, M., S. H. Houston, L. R. Amat, and N. Morisseau-Leroy, 1998: The HRD real-time hurricane wind analysis system. J. Wind Eng. Ind. Aerodyn., 77–78, 53–64, doi:10.1016/S0167-6105(98)00131-7.
Reul, N., and B. Chapron, 2003: A model of sea-foam thickness distribution for passive microwave remote sensing applications. J. Geophys. Res., 108, 3321, doi:10.1029/2003JC001887.
Reul, N., J. Tenerelli, B. Chapron, D. Vandemark, Y. Quilfen, and Y. Kerr, 2012: SMOS satellite L-band radiometer: A new capability for ocean surface remote sensing in hurricanes. J. Geophys. Res., 117, C02006, doi:10.1029/2011JC007474.
Reul, N., B. Chapron, E. Zabolotskikh, C. Donlon, Y. Quilfen, S. Guimbard, and J. F. Piolle, 2016: A revised L-band radio-brightness sensitivity to extreme winds under tropical cyclones: The five year SMOS-storm database. Remote Sens. Environ., 180, 274–291, doi:10.1016/j.rse.2016.03.011.
Reynolds, R., 2009: What’s new in version 2. NOAA/NCEP Rep., 10 pp. [Available online atwww.ncdc.noaa.gov/sites/default/files/attachments/Reynolds2009_oisst_daily_v02r00_version2-features.pdf.]
Reynolds, R., T. Smith, C. Liu, D. Chelton, K. Casey, and M. Schlax, 2007: Daily high-resolution blended analyses for sea surface temperature. J. Climate, 20, 5473–5496, doi:10.1175/2007JCLI1824.1.
Ricciardulli, L., 2016: ASCAT on MetOp-A data product update notes. Remote Sensing Systems Tech. Rep. 040416, 5 pp. [Available online atwww.remss.com/papers/rsstech/2016_040416_RSS_ASCAT_V2_update.pdf.]
Ricciardulli, L., and F. Wentz, 2015: A scatterometer geophysical model function for climate-quality winds: QuikSCAT Ku-2011. J. Atmos. Oceanic Technol., 32, 1829–1846, doi:10.1175/JTECH-D-15-0008.1.
Ricciardulli, L., and F. Wentz, 2016: Remote Sensing Systems ASCAT C-2015 daily ocean vector winds on 0.25 deg grid, version 02.1. Remote Sensing Systems, accessed July 2016. [Available online atwww.remss.com/missions/ascat.]
Ricciardulli, L., F. Wentz, and T. Meissner, 2015: Bringing consistency among scatterometer winds using radiometer observations. Proc. IOWST Meeting, Portland, OR, NASA, 17 pp. [Available online at https://mdc.coaps.fsu.edu/scatterometry/meeting/docs/2015/ClimateDataRecordDevelopmentAndAnalysis/Ricciardulli_ovwst_2015.pdf.]
Ruf, C., A. Lyons, M. Unwin, J. Dickinson, R. Rose, D. Rose, and M. Vincent, 2013: CYGNSS: Enabling the future of hurricane prediction [Remote Sensing Satellites]. IEEE Geosci. Remote Sens. Mag., 1, 52–67, doi:10.1109/MGRS.2013.2260911.
Ruf, C., and Coauthors, 2016: New ocean winds satellite mission to probe hurricanes and tropical convection. Bull. Amer. Meteor. Soc., 97, 385–395, doi:10.1175/BAMS-D-14-00218.1.
Sampson, C., and A. Schrader, 2000: The Automated Tropical Cyclone Forecasting System (version 3.2). Bull. Amer. Meteor. Soc., 81, 1231–1240, doi:10.1175/1520-0477(2000)081<1231:TATCFS>2.3.CO;2.
Sapp, J., S. Alsweiss, Z. Jelenak, P. Chang, S. Frasier, and J. Carswell, 2016: Airborne co-polarization observations of the ocean-surface NRCS at C-band. IEEE Trans. Geosci. Remote Sens., 54, 5975–5992, doi:10.1109/TGRS.2016.2578048.
Sharp, R., M. Bourassa, and J. O’Brien, 2002: Early detection of tropical cyclones using SeaWinds-derived vorticity. Bull. Amer. Meteor. Soc., 83, 879–889, doi:10.1175/1520-0477(2002)083<0879:EDOTCU>2.3.CO;2.
Stiles, B., and S. Yueh, 2002: Impact of rain on spaceborne Ku-band wind scatterometer data. IEEE Trans. Geosci. Remote Sens., 40, 1973–1983, doi:10.1109/TGRS.2002.803846.
Stoffelen, A., and D. Anderson, 1997: Scatterometer data interpretation data: Measurement space and inversion. J. Atmos. Oceanic Technol., 14, 1298–1313, doi:10.1175/1520-0426(1997)014<1298:SDIMSA>2.0.CO;2.
Tang, W., S. Yueh, A. Fore, G. Neumann, A. Hayashi, and G. Lagerloef, 2013: The rain effect on Aquarius’ L-band sea surface brightness temperature and radar backscatter. Remote Sens. Environ., 137, 147–157, doi:10.1016/j.rse.2013.06.016.
Tournadre, J., and Y. Quilfen, 2003: Impact of rain cell on scatterometer data: 1. Theory and modeling. J. Geophys. Res., 108, 3225, doi:10.1029/2002JC001428.
Uhlhorn, E., and P. Black, 2003: Verification of remotely sensed sea surface winds in hurricanes. J. Atmos. Oceanic Technol., 20, 99–116, doi:10.1175/1520-0426(2003)020<0099:VORSSS>2.0.CO;2.
Uhlhorn, E., J. Franklin, M. Goodberlet, J. Carswell, and A. Goldstein, 2007: Hurricane surface wind measurements from an operational stepped frequency microwave radiometer. Mon. Wea. Rev., 135, 3070–3085, doi:10.1175/MWR3454.1.
Ulaby, F., and D. Long, 2013: Microwave Radar and Radiometric Remote Sensing. University of Michigan Press, 1116 pp.
Velden, C., and Coauthors, 2006: The Dvorak tropical cyclone intensity estimation technique. Bull. Amer. Meteor. Soc., 87, 1195–1210, doi:10.1175/BAMS-87-9-1195.
Verspeek, J., A. Stoffelen, M. Portabella, H. Bonekamp, C. Anderson, and J. Figa-Saldaña, 2010: Validation and calibration of ASCAT using CMOD5.n. IEEE Trans. Geosci. Remote Sens., 48, 386–395, doi:10.1109/TGRS.2009.2027896.
Von Ahn, J., J. Sienkiewicz, and P. Chang, 2006: Operational impact of QuikSCAT winds at the NOAA Ocean Prediction Center. Wea. Forecasting, 21, 523–539, doi:10.1175/WAF934.1.
Weissman, D., and M. Bourassa, 2008: Measurements of the effect of rain-induced sea surface roughness on the QuikSCAT scatterometer radar cross section. IEEE Trans. Geosci. Remote Sens., 46, 2882–2894, doi:10.1109/TGRS.2008.2001032.
Wentz, F., 1983: A model function for ocean microwave brightness temperatures. J. Geophys. Res., 88, 1892–1908, doi:10.1029/JC088iC03p01892.
Wentz, F., 1997: A well-calibrated ocean algorithm for Special Sensor Microwave/Imager. J. Geophys. Res., 102, 8703–8718, doi:10.1029/96JC01751.
Wentz, F., 2005: The effect of clouds and rain on the Aquarius salinity retrieval. Remote Sensing Systems Tech. Memo. 3031805, 14 pp. [Available online at http://images.remss.com/papers/aquarius/rain_effect_on_salinity.pdf.]
Wentz, F., 2015: A 17-year climate record of environmental parameters derived from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager. J. Climate, 28, 6882–6902, doi:10.1175/JCLI-D-15-0155.1.
Wentz, F., and R. Spencer, 1998: SSM/I rain retrievals within a unified all-weather ocean algorithm. J. Atmos. Sci., 55, 1613–1627, doi:10.1175/1520-0469(1998)055<1613:SIRRWA>2.0.CO;2.
Wentz, F., and D. Smith, 1999: A model function for the ocean-normalized radar cross section at 14 GHz derived from NSCAT observations. J. Geophys. Res., 104, 11 499–11 514, doi:10.1029/98JC02148.
Wentz, F., and D. LeVine, 2012: Aquarius salinity retrieval. Algorithm Theoretical Basis Document, Remote Sensing Systems Tech. Rep. 082912, 45 pp. [Available online at ftp://podaac-ftp.jpl.nasa.gov/allData/aquarius/docs/v2/AQ-014-PS-0017_AquariusATBD_Level2.pdf.]
Wentz, F., L. Mattox, and S. Peteherych, 1986: New algorithms for microwave measurements of ocean winds: Applications to SeaSat and the Special Sensor Microwave Imager. J. Geophys. Res., 91, 2289–2307, doi:10.1029/JC091iC02p02289.
Wentz, F., L. Ricciardulli, C. Gentemann, T. Meissner, K. Hilburn, and J. Scott, 2013: Remote Sensing Systems Coriolis WindSat daily environmental suite on 0.25deg grid, version 7.0.1 LF wind speeds. Remote Sensing Systems. [Available online at www.remss.com/missions/windsat.]
Yueh, S., W. Tang, A. Fore, G. Neumann, A. Hayashi, A. Freedman, J. Chaubell, and G. Lagerloef, 2013: L-band passive and active microwave geophysical model functions of ocean surface winds and applications to Aquarius retrieval. IEEE Trans. Geosci. Remote Sens., 51, 4619–4662, doi:10.1109/TGRS.2013.2266915.
Zabolotskikh, E., L. Mitnik, N. Reul, and B. Chapron, 2015: New possibilities for geophysical parameter retrievals opened by GCOM-W1 AMSR2. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 8, 4248–4261, doi:10.1109/JSTARS.2015.2416514.
