Difference Between Collocated ASCAT And Quikscat Surface Wind Retrievals
Total Page：16
File Type：pdf, Size：1020Kb
1 1
2
3
4
5 Matching ASCAT and QuikSCAT Winds
6 7 Abderrahim Bentamy1, Semyon A. Grodsky2, James A. Carton2, Denis 8 Croizé-Fillon1, Bertrand Chapron1 9 10 11 12Submitted to : 13JGR-Oceans (rev. 07/29/2011) 14 15 16July 29, 2011 17
191 Institut Francais pour la Recherche et l’Exploitation de la Mer, Plouzane, France 20 212Department of Atmospheric and Oceanic Science, University of Maryland, College Park, 22MD 20742, USA 23 24Corresponding author: [email protected]
2 1Abstract
2Surface winds from two scatterometers, the Advanced scatterometer (ASCAT), available
3since 2007, and QuikSCAT, which was available through November 2009, show
4persistent differences during their period of overlap. This study examines a set of
5collocated observations during a 13-month period November 2008 through November
62009, to evaluate the causes of these differences. The difference in the operating
7frequency of the scatterometers leads to differences that depend on rain rate, wind
8velocity, and SST. The impact of rainfall on the higher frequency QuikSCAT is up to 1
9ms-1 in the tropical convergence zones and along the western boundary currents even after
10the rain flagging is applied. This difference from ASCAT is reduced by some 30% to
1140% when data for which the multidimensional rain probabilities >0.05 is also removed.
12An additional component of the difference in wind speed seems to be the result of biases
13in the geophysical transfer functions used in processing the two data sets and is
14parameterized as a function of ASCAT wind speed and direction relative to the mid beam
15azimuth. After applying the above two corrections, QuikSCAT wind speed remains
16systematically lower (by 0.5 ms-1) than ASCAT over cold SST<5oC. This difference
17appears to be the result of temperature-dependence in the viscous dumping which
18differentially impacts shorter waves and thus preferentially impacts QuikSCAT. The
19difference in wind retrievals also increases in the storm track corridors as well as in the
20coastal regions where the diurnal cycle of breeze projects on the time lag between
21satellites.
1 2 11. Introduction
2 Many meteorological and oceanic applications require the high spatial and temporal
3resolution of satellite winds [e.g. Grima et al, 1999; Grodsky et al, 2001, Blanke et al,
42005; Risien and Chelton, 2008]. Many of these studies also require consistent time
5series spanning the lifetime of multiple satellite missions, of which there have been seven
6since the launch of the first European Remote Sensing satellite (ERS-1) in August 1991.
7Creating consistent time series requires accounting for changes in individual mission
8biases, most strikingly when successive scatterometers operate in different spectral bands
9[e.g. Bentamy et al., 2002; Ebuchi et al., 2002]. This paper presents a comparative study
10of winds derived from the matching of two such scatterometers: the C-band Advanced
11SCATterometer (ASCAT) on board Metop-A and the higher frequency Ku-band
12SeaWinds scatterometer onboard QuikSCAT.
13 Scatterometers measure wind velocity indirectly through the impact of wind on the
14amplitude of capillary or near-capillary surface waves. This wave field is monitored by
15measuring the strength of Bragg scattering of an incident microwave pulse. If a pulse of
16wavenumber k impinges on the surface at incidence angle relative to the surface the
17maximum backscatter will occur for surface waves of wavenumber of k B 2k sin( )
18whose amplitude, in turn reflects local wind conditions [e.g. Wright and Keller, 1971).
19The fraction 0 of transmitted power that returns back to the satellite is thus a function
20of local wind speed and direction (relative to the antenna azimuth), and .
21 To date, the most successful conversions of scatterometer measurements into near-
22surface wind rely on empirically derived geophysical model functions (GMFs)
1 3 1augmented by procedures to resolve directional ambiguities resulting from uncertainties
2in the direction of wave propagation. Careful tuning of the GMFs is currently providing
3estimates of 10m neutral wind velocity with errors estimated to be around ±1 ms-1 and
4±20o, [e.g. Bentamy et al., 2002; Ebuchi et al., 2002]. However systematic errors are also
5present. For example, Bentamy et al. [2008] have shown that ASCAT has a systematic
6underestimation of wind speed that increases with wind speed, reaching 1 ms-1 for winds
7of 20 ms-1.
8 The GMFs and other parameters must change if the frequency band being used by
9the scatterometer changes and historically scatterometers have used two different
10frequency bands. US scatterometers as well as the Indian OCEANSAT2 use frequencies
11in the 10.95-14.5GHz Ku-band. In contrast the European Remote Sensing satellite
12scatterometers have all adopted the 4-8 GHz C-band to allow for reduced sensitivity to
13rain interference [Sobieski et al, 1999; Weissman et al., 2002]. To avoid spurious trends
14in a combined multi-scatterometer data set, the bias between different scatterometers
15needs to be removed.
16 There have been numerous efforts to construct a combined multi-satellite data set of
17winds [Milliff et al., 1999; Zhang et al, 2006; Bentamy et al, 2007, Atlas et al, 2011].
18These efforts generally rely on utilizing a third reference product such as passive
19microwave winds or reanalysis winds to which the individual scatterometer data sets can
20be calibrated. In contrast to such previous efforts, this current study focuses on directly
21matching scatterometer data. We illustrate the processes by matching the current
22ASCAT and the recent QuikSCAT winds [following Bentamy et al., 2008]. This
23matching exploits the existence of a 13 month period November 2008 – November 2009
1 4 1when both missions were collecting data and when ASCAT processing used the current
2CMOD5N GMF [Verspeek et al., 2010] to derive 10m neutral wind.
32. Data
4 This study relies on data from two scatterometers, the SeaWinds Ku-band (13.4
5GHz, 2.2 cm) scatterometer onboard QuikSCAT (referred to as QuikSCAT or QS, Table
61) which was operational June, 1999 to November, 2009 and the C-band ASCAT
7onboard the European Meteorological Satellite Organization (EUMESAT) MetOp-A1
8(5.225GHz, 5.7 cm) launched October, 2006 (referred to as ASCAT or AS).
9 QuikSCAT uses a rotating antenna with two emitters: the H-pol inner beam with an
10incidence angle =46.25° and the V-pol outer beam with an incidence angle of =54°.
11Observations from these two emitters have swaths of 1400km and 1800km, respectively
12which cover around 90% of the global ocean daily. These observations are then binned
13into Wind Vector Cells of 25´25 km. QuikSCAT winds used in this study are Level 2b
14data estimated with the empirical QSCAT-1 GMF [Dunbar et al., 2006] and the
15Maximum Likelihood Estimator which selects the most probable wind direction. To
16improve wind direction estimates in the middle of the swath an additional Direction
17Interval Retrieval with Threshold Nudging algorithm is applied. The winds produced by
18this technique show RMS speed and direction differences from concurrent in-situ buoy
19and ship data of approximately 1ms-1 and 23°, and temporal correlations in excess of 0.92
20[e.g. Bentamy et al., 2002; Bourassa et al, 2003; Ebuchi et al, 2002]. The QuikSCAT
21wind product includes several rain flags determined directly from the scatterometer
11 2http://www.eumetsat.int/Home/Main/Publications/Technical_and_Scientific_Documentation/Technical_No 3tes/ 4http://www.knmi.nl/scatterometer/.
5 5 1observations and from collocated radiometer rain rate measurements from other satellites.
2The impact of rain on QuikSCAT data is indicated by two independent rain indices: rain
3flag and multidimensional rain probability (MRP).
4 ASCAT has an engineering design that is quite different from QuikSCAT. Rather
5than a rotating antenna it has a three beam antenna looking 45o (fore-beam), 90o (mid-
6beam), 135o (aft-beam) of the satellite track, which together sweep out two 550 km
7swaths on both sides of the track. The incidence angle varies in the range 34o-64o for the
8outermost beams and 25o-53o for the mid-beam, giving Bragg wavelengths of 3.2-5.1cm
9and 3.6-6.8 cm. Here we use level 2b ASCAT near real time data distributed by
10EUMETSAT and by KNMI at 25x25 km2 resolution. Comparisons to independent
11mooring and shipboard observations by Bentamy et al. [2008] and Verspeek et al. [2010]
12show that ASCAT wind speed and direction has accuracies similar to QuikSCAT.
13 To reevaluate wind accuracy we use a variety of moored buoy measurements
14including wind velocity, SST, air temperature, and significant wave height. These are
15obtained from Météo-France and U.K. MetOffice (European seas) [Rolland et al, 2002],
16the NOAA National Data Buoy Center (US coastal zone) [Meindl et al, 1992], and the
17tropical moorings of the Tropical Atmosphere Ocean Project (Pacific), the Pilot Research
18Moored Array (Atlantic), and Research Moored Array for African–Asian–Australian
19Monsoon Analysis and Prediction project (Indian Ocean) [McPhaden et al., 1998;
20Bourles et al., 2008; McPhaden et al., 2009]. Hourly averaged buoy measurements of wind
21velocity, sea surface and air temperatures, and humidity are converted to neutral wind at
2210m height using the COARE3.0 algorithm of Fairall et al. [2003].
1 6 1 Quarter-daily 10m winds, as well as SST and air temperatures, are also obtained
2from the European Centre for Medium Weather Forecasts (ECMWF) operational
3analysis. These are routinely provided by the Grid for Ocean Diagnostics, Interactive
4Visualization and Analysis (GODIVA) project2 on a regular grid of 0.5o in longitude and
5latitude.
63. Collocated observations
7 Both satellites are on quasi sun-synchronous orbits, but the QuikSCAT local
8equator crossing time at the ascending node (6:30 a.m.) leads the ASCAT local equator
9crossing time (9:30 a.m.) by 3 hours. This difference implies that data precisely
10collocated spatially are available only with a few hours time difference at low latitudes
11(Figure. 1) [Bentamy et al, 2008]. Here we accept data pairs collocated in space and time
12where QuikSCAT data is collected within 0< <4 hours and 50 km of each valid ASCAT
13Wind Vector Cell.
14 For the thirteen month period November 2008 through November 2009 this
15collocation selection procedure produces an average 2800 collocated pairs of data in each
161ox1o geographical bin. The data count is somewhat less in the regions of the tropical
17convergence zones due to the need to remove QuikSCAT rain-flagged data. The data
18count increases with increasing latitude as a result of the near-polar orbits, but decreases
19again at very high latitudes due to the presence of ice. At shorter lags of <3 hours
20collocated data coverage is still global, but the number of match-ups is reduced by a
21factor of two. At lags of <1 hours collocated data is only available in the extra tropics
22(Figure. 1).
12 behemoth.nerc-essc.ac.uk/ncWMS/godiva2.html
2 7 1 Next we explore possible impacts of nonzero time lag in the collocated data by
2examining the temporal lag decorrelation time of buoy winds. In midlatitudes where the
3characteristic time separation between collocated ASCAT and QuikSCAT data is less
4than 2 hours we examine wind measurements from NDBC buoys. The original buoy data
5is sub-sampled based on the timing of the nearest ASCAT (sub-sample 1) and QuikSCAT
6(sub-sample 2) overpasses. The mean bias between subsamples at these extra tropical
7NDBC mooring locations is close to zero with an RMS difference of less than 1ms-1, and
8a temporal correlation >0.95. Negligible time mean bias at the buoys suggests that time
9mean differences between collocated satellite wind fields (if any) is the result of
10differences in the satellite wind estimates rather than the time lag between samples.
11Similar results are found in an examination of data from the tropical moorings assuming
12 < 4 hr with no mean bias, RMS differences of 1.2 m/s, and temporal correlations >0.92.
13Finally we examine the impact of the collocation space and time lag using the ECMWF
14surface winds to simulate space and time sampling of ASCAT and QuikSCAT during the
15period November 20, 2008 through November 19, 2009. This comparison also shows no
16significant geographical patterns in the sampling bias.
17 Next we evaluate the accuracy of the satellite winds in the tropics by using this data
18window to find triply-collocated ASCAT, QuikSCAT, and tropical mooring winds (here
19we assume the mooring winds are unbiased). Both, QuikSCAT and ASCAT wind speeds
20(WQS and W AS ) agree well with collocated mooring wind speeds, Wbuoy , (Figure 2) with
21time mean differences less than 1 m/s at all locations. ASCAT has a slight -0.35 ms-1
22negative bias averaged over the moorings (Figure 3. In contrast, QuikSCAT mean wind
1 8 1speed errors have a mixture of negative and positive biases, roughly mirror the spatial
2distribution of rainfall and giving a mooring-average bias of -0.15 ms-1.
3 We evaluated the agreement of the time-dependent component of wind speed by
4examining RMS differences and correlations of the collocated satellite and buoy winds.
5This RMS difference averaged over all tropical buoys is 1.2 m/s and is well above the
6mean wind speed bias. The correlation of satellite and tropical buoy winds is 0.8 at = 4
7hr, but exceeds 0.9 at zero lag (Figure 4). At higher latitudes, where the winds have
8longer synoptic timescales, the collocated satellite-buoy wind time series have even
9higher correlations (0.89 at =4 hours, and 0.96 at less than half an hour, not shown).
104. Results
114.1 Global comparisons
12 Time mean difference of collocated QuikSCAT and ASCAT wind speed (
-1 13 W WQS WAS , Figure 5) is generally less than 1ms , but contains some planetary-
14scale patterns. In general, W >0 at tropical to midlatitudes where it exhibits a pattern of
15values in the range of 0.40-0.80 ms-1 resembling the pattern of tropical and midlatitude
16storm track precipitation. This pattern suggests that even after applying the rain flag there
17is a residual impact of rainfall on QuikSCAT wind retrievals (as previously examined by,
18e.g. Weissman et al., 2002]. Enhancing rain flagging by removing cases with MRP >
190.05 (5% of QuikSCAT data) decreases the frequency of positive differences in the
20tropical precipitation zones by some 30% to 40% and the number of large differences
21 W >1 ms-1 is reduced by 38% (Figure 5). However, the RMS variability of W
22doesn’t change with the enhanced rain flagging. Both with and without additional rain
1 9 1flagging RMS( W ) has a maximum of up to 3ms-1 in the zones of the midlatitude storm
2tracks of both hemispheres and approaches 1.8 ms-1 in the tropical convergence zones.
3These maxima in inter-instrument variability reflect a combination of still unaccounted
4for impacts of rain events, and impacts of strong transient winds in the midlatitude storm
5tracks, as well as deep convection events in the tropics.
6 In contrast to the tropics or midlatitudes, time mean W is negative at subpolar
7and polar latitudes. This is particularly noticeable at southern latitudes where the
8difference may exceed -0.5 ms-1. This striking high latitude inter-scatterometer difference
9remains even after applying enhanced rain detection to QuikSCAT (Figure 5).
10 The temporal correlation of QuikSCAT and ASCAT winds is high (>0.9) over
11much of the global ocean (Figure 5, CORR). It decreases to 0.7 (0.5 at some locations) in
12the tropical convergence zones. This decrease occurs because in the tropics the space
13scales of transient deep convective systems approache the scale of a single radar footprint
14while the temporal scales of transients compare with the time lag between the two
15satellites [Houze and Cheng, 1977]. In contrast, wind variability at mid and high latitudes
16has long multi-day synoptic timescales and large spatial scales which result in high
17correlations among collocated measurements.
184.2 Effect of geophysical model function
19 In this section we attempt to parameterize observed W (after applying enhanced
20rain flagging to the QuikSCAT data) as a function of geophysical and satellite-earth
21geometrical factors. The ASCAT GMF is parameterized by a truncated Fourier series of
22wind direction (DIR) relative to antenna azimuth (AZIM) for each of the antenna beams
23 =DIR-AZIM as
1 10 0 1 B0 B1 cos B2 cos 2 , (1)
2 where the coefficients B depend on wind speed and incident angle, . This
3suggests that a correction function dW (to be added to ASCAT winds in order to adjust
4them to QuikSCAT winds) should depend on similar parameters, dW (WAS , , ) . Due to
5the fixed, three beam observation geometry of ASCAT, the azimuthal dependence of
6 dW is reduced to dependence on ASCAT wind direction relative to the ASCAT mid-
7beam azimuth, DIRAS AZIM 1 . In initial experiments we find that dW does not
8depend on incidence angle (not shown).
9 Empirical estimates of this correction function dW are constructed by binning
10observed W as a function of wind speed and . Here we begin by considering
11estimates based on data from latitudes equatorward of 55o where the time mean W
12values are positive. Figure 6 reveals a clear dependence on . For ASCAT mid-beam
13looking in the wind vector direction ( =0o) W is small. Looking in the upwind
14direction ( 180o) QuikSCAT wind speed is stronger than ASCAT by 0.5 m/s.
15Looking in the cross-wind direction ( 90o) ASCAT wind speed is stronger than
16QuikSCAT wind speed by approximately 0.2 ms-1. Similar azimuthal dependencies show
17up when examining the difference between buoy wind speed and ASCAT wind speed
18(Figure 7) although data scatter is larger due to the smaller number of satellite-buoy
19collocations (compare right hand axes in Figures 6 and 7).
20 We next examine the dependence of W on both wind speed and azimuth (Figure
218). Independent of wind direction, QuikSCAT wind speed exceeds ASCAT at both low
-1 -1 22winds (WAS <3 ms ) and high winds (WAS >15 ms ), with a larger difference at high
1 11 1winds [see also Bentamy et al., 2008]. Independent of wind speed, QuikSCAT wind is
2stronger than ASCAT if the ASCAT mid-beam is oriented in the upwind direction (
3 180o, compare with Figure 6).
4 Due to sampling errors, binned W are noisy at high winds (for which less data is
5available) and are not symmetrical in azimuth (Figure 8a). To mitigate the impact of
6sampling errors we represent dW by its mean and first three symmetric harmonics
m3 m 7 dW P5 (WAS )cos(m) , (2) m0
m 8where the coefficients P5 (WAS ) are assumed to be fifth order polynomials of ASCAT
9wind speed. The polynomial coefficients in turn are estimated by least squares
10minimization (Table 2).
11 To evaluate the usefulness of this approximation we compare its time mean
12structure to that of the time mean of W in Figure 9. This comparison shows that most
13of the spatial structure in the time mean of W is retained in the time mean of dW .
14Applying the correction function (2) to instantaneous ASCAT winds leads to apparent
15decrease in the residual time mean wind speed difference (compare WQS WAS and
16WQS WAS dW panels in Figure 9). The unmodeled time mean difference includes
17positive values in the regions of the midlatitude storm tracks, which is to be expected
18because ASCAT underestimates strong winds (>20ms-1). Strong wind events are
19relatively rare, and the correction function dW is uncertain at those conditions (Figure
208a). Some unmodeled time mean dW is also present in coastal areas such as off Peru and
21Namibia, where we suspect there are sampling problems resulting from under-resolving
22of the prominent diurnal land-sea breezes.
1 12 1 After correction for our modeled dW the global histogram of W WQS WAS
2eliminates the slight 0.18 ms-1 positive bias in W (Figure 9). However large negative
3tails of the distribution remain, reflecting the presence of large negative values of W at
4high latitude mentioned previously. Here we explore a physically-based hypothesis for
5this high latitude negative W .
64.3 Impact of SST
7 As described in the Appendix radar backscatter 0 depends on the spectrum of the
8resonant Bragg waves. These near-capillary waves have amplitudes that depend on a
9balance of wave growth and viscous dissipation, and the latter is a function of SST
10because of the impact of SST on viscosity. We present a linearized form of the
11dependence of ΔW on SST through a Taylor series expansion of the growth-dissipation
o 12equation around the space-time mean value of SST (T0 =19 C)
2[ (T ) (T0 )](QS AS ) 13 W (3) ( a / w )C C DW (1 c )
14where notations and variable definitions are the same as in (A3). W is affected by
15deviation of local SST from T0 through kinematic viscosity (T ) (T0 ) , and
16inversely by increases in wind speed with some modifications due to changes in a .
17These effects are relatively small over much of the ocean but increase at high latitudes
18due to the increase in kinematic viscosity (Figure 10). The SST dependent bias varies
19inversely with wind speed and thus is further enhanced south of 60oS due to the
20combination of cold SSTs and relatively weak winds.
1 13 1 To summarize the impact of SST and winds on ΔW we present the QuikSCAT
2minus ASCAT collocated differences binned as a function of wind speed and SST
3(Figure 11). QuikSCAT wind speed is higher than ASCAT at low (<3 ms-1) and high
4(>12 ms-1) winds. In contrast, ASCAT wind exceeds QuikSCAT wind over cold
5SST<10°C under moderate wind 5 ms-1 6difference between the scatterometers is found at SST<5°C. In contrast to our 7expectations from (3) the comparison in Figure 11 indicates that the negative difference 8between ASCAT and QuikSCAT wind speeds rises in magnitude at moderate winds. The 9explanation for this behavior likely requires an improved parameterization of the wind- 10wave growth parameter and a more detailed radar imaging model (see Kudryavtsev et al., 112005 for further details). 125. Summary 13 The end of the QuikSCAT mission in 2009 and the lack of a follow-on Ku band 14scatterometer has made it a high priority to ensure the compatibility of QuikSCAT and 15the continuing C band ASCAT mission measurements. Examination of collocated 16observations during a 13 month period of mission overlap November 2008-November 172009 shows that there are currently systematic differences in the 10m neutral wind 18estimates from the two missions. This study is an attempt to identify and model these 19differences in order to produce a consistent scatterometer-based wind record spanning the 20period 1999 – present. 21 The basic data set we use to identify the differences and develop corrections is the 22set of space-time collocated satellite wind observations from the two missions and triply 23collocated satellite-satellite-buoy observations during the 13 month period of overlap. 1 14 1The first part of the study explores the acceptable range of spatial and temporal 2separations between observations for comparison. Satellite and buoy winds show high 3temporal correlation ( 0.8) for time lags of up to 4 hours and thus 4 hours was 4determined to be the maximum acceptable time separation for winds stronger than 3.5 5m/s. Requiring shorter time lags severely limits the number of collocated observations at 6lower latitudes. Similarly, spatial separations of up to 50 km were determined to be 7acceptable. 8 An examination of the collocated winds shows a high degree of agreement in 9direction, but reveals systematic differences in speed that depend on rain rate, the 10strength of the wind, and SST. The impact in wind speed of rainfall on the higher 11frequency of the two scatterometers, QuikSCAT, is up to 1ms-1 in the tropical 12convergence zones and the high precipitation zones of the midlatitude storm tracks even 13after rain flagged data is removed. This 1 ms-1 bias is cut by half by removing QuikSCAT 14wind retrievals with multidimensional rain probabilities >0.05 even though this additional 15rain flagging reduces data amount by only 5%. For this reason we recommend the 16additional rain flagging. 17 We next examine the causes of the differences in collocated wind speed remaining 18after removing QuikSCAT winds with MRP>0.05. Examination of histograms shows 19that the differences are partly a function of ASCAT wind speed and direction relative to 20the mid beam azimuth, a dependence strongly suggesting bias in the ASCAT geophysical 21model function. This dependence is empirically modeled based on the collocation data, 22and is then removed. The bias model reduces collocated differences by 0.5 ms-1 in the 23tropics where it has large scale patterns related to variability of wind speed and direction. 1 15 1 After applying the above two corrections, a third source of difference is evident in 2that QuikSCAT wind speed remains systematically lower (by 0.5 ms-1) than ASCAT 3wind speed at high latitudes where SST<5oC. An examination of the physics of wind 4ripples implicates the SST-dependence of wave dissipation, an effect which is enhanced 5for shorter waves. Again, the higher frequency scatterometer, QuikSCAT, will be more 6sensitive to this effect. 7 Two additional sources of differences in the collocated observations are also 8identified. Differences in wind retrievals occur in storm track zones due to 9underestimation of infrequent strong wind events (>20 ms-1) by ASCAT. Differences in 10wind speed estimates are also evident in the Peruvian and Namibian coastal zones where 11diurnal land-sea breezes are poorly resolved by the time delays between passes of the two 12scatterometers (up to 4 hours in this study). Here the differences seem to be a result of 13the limitations of the collocated data rather than indicating error in the scatterometer 14winds. 15 Acknowledgements. This research was supported by the NASA International 16Ocean Vector Wind Science Team. We thank J. F. Piollé and IFREMER/CERSAT for 17data processing support. The authors are grateful to ECMWF, EUMETSAT, CERSAT, 18GODIVA, JPL, Météo-France, NDBC, O&SI SAF, PMEL, and UK MetOffice for 19providing buoy, numerical, and satellite data used in this study. 1 16 1 Appendix. Impact of SST on the Bragg scattering 2 The purpose of this Appendix is to evaluate deviation in retrieved scatterometer 3wind speed due to changes in SST. This evaluation is done for the pure Bragg scattering 4and using simplified model of the Bragg wave spectrum. 5 Radar backscatter, 0 , is mostly affected by energy of the resonant Bragg wave 6component. According to Donelan and Pierson [1997] for the pure Bragg scattering 7regime radar backscatter 0 is a positive definite function of the spectrum of the resonant 8Bragg wave component whose energy depends on the balance of wind wave growth 0 9parameter w and viscous dissipation that is a function of SST, F( w ) . 10For simplicity, we next consider waves aligned with wind and use the Plant [1982] 2 11parameterization w a / wC (u* / c) where a / w is the air to water density ratio, 12C =32 is an empirical constant, u* CD W is the friction velocity in the air, CD is the 13neutral drag coefficient that is parameterized as in Large and Yeager [2009]. This w 14works over wide range of u* / c but fails at small values corresponding to threshold 15winds [Donelan and Pierson, 1987]. It should be also corrected by a factor (1 c ), 16where c is the wind-wave coupling parameter to account for the splitting of total wind 17stress in the wave boundary layer into wave-induced and turbulent components [Makin 18and Kudryavtsev, 1999; Kudryavtsev et al., 1999) 2 19 w a / wC (u* / c) (1 c ) , (A1) 1 17 1Viscous dissipation depends on the wavenumber, k , and frequency, , of the Bragg 2 1 2component, 4k . It also depends on the water temperature, T , through the 3kinematic viscosity, (T ) , that explains temperature behavior of the wind-wave growth 4threshold wind speed [Donelan and Pierson, 1987, Donelan and Plant, 2009]. We also 5assume that scatterometer calibration, 0 (W ) , corresponds to the globally and time mean o ~ 6sea surface temperature T0 =19 C. Then an expected error in retrieved wind speed W due 7to deviation of local temperature from T0 , T T T0 , is calculated by differentiating 8along (T,W ) const : ~ / T / 9 W (k) T a a w (A2) / W w / W 10The first term in (A2) accounts for changes in retrieved wind speed due to changes in the 11air density [Bourassa et al., 2010]. For chosen w this term doesn’t depend on k , thus 12doesn’t contribute to the wind difference between Ku- and C-band. Temperature 13dependent difference between QS and AS winds is explained by the viscous term in (A2) ~ ~ 2 (QS AS ) 14 W W (kQS ) W (k AS ) , (A3) ( a / w )C C DW (1 c ) 15where (T) (T0 ) . 1 18 1References 2Atlas, R., R. N. Hoffman, J. Ardizzone, S. M. Leidner, J. C. Jusem, D. K. Smith, and D. 3 Gombos (2011), A Cross-calibrated, Multiplatform Ocean Surface Wind Velocity 4 Product for Meteorological and Oceanographic Applications. Bull. Amer. Meteor. 5 Soc., 92, 157–174. doi: 10.1175/2010BAMS2946.1 6Bentamy, A., K B. Katsaros, M. Alberto, W. M. Drennan, and E. B. Forde (2002), Daily 7 surface wind fields produced by merged satellite data. American Geophys. Union, 8 Geophysical Monograph Series, 127, 343-349. 9Bentamy, A., H-L Ayina, P. Queffeulou, and D. Croize-Fillon (2007), Improved Near 10 Real Time Surface Wind Resolution over The Mediterranean Sea. Ocean Sci., 3, 259- 11 271. 12Bentamy, A., D. Croize-Fillon, and C. Perigaud (2008), Characterization of ASCAT 13 measurements based on buoy and QuikSCAT wind vector observations, Ocean Sci., 14 4, 265–274. 15Blanke, B., S. Speich, A. Bentamy, C. Roy, and B. Sow (2005), Modeling the structure 16 and variability of the southern Benguela upwelling using QuikSCAT wind forcing, J. 17 Geophys. Res., 110 ,C07018, doi:10.1029/2004JC002529. 18Bourassa, M. A., D. M. Legler, J. J. O’Brien, and S. R. Smith (2003), SeaWinds 19 validation with research vessels, J. Geophys Res., 108(C2), 3019, 20 doi:10.1029/2001JC001028. 21Bourassa, M. A., E. Rodriguez, and R. Gaston (2010), NASA's Ocean Vector Winds 22 Science Team Workshops. Bull. Amer. Meteor. Soc., 91, 925–928. doi: 23 10.1175/2010BAMS2880.1 1 19 1Bourle`s, B., and Coauthors (2008), The Pirata Program: History, accomplishments, and 2 future directions, Bull. Am. Meteorol. Soc., 89, 1111– 1125, 3 DOI:10.1175/2008BAMS2462.1. 4Donelan, M. A., and W. J. Pierson Jr. (1987), Radar scattering and equilibrium ranges in 5 wind-generated waves with application to scatterometry, J. Geophys. Res., 92(C5), 6 4971– 5029, doi:10.1029/JC092iC05p04971. 7Donelan, M. A., and W. J. Plant (2009), A threshold for wind-wave growth, J. Geophys. 8 Res., 114, C07012, doi:10.1029/2008JC005238. 9Dunbar, S., and Coauthors (2006), QuikSCAT Science Data Product User's Manual 10 Overview & Geophysical Data Products, Jet Propulsion Laboratory, ftp://podaac- 11 ftp.jpl.nasa.gov/allData/quikscat/L2B/docs/QSUG_v3.pdf 12Ebuchi, N., H. C. Graber, and M. J. Caruso (2002), Evaluation of wind vectors observed 13 by QuikSCAT/SeaWinds using ocean buoy data, J. Atmos. Ocean. Technol., 19, 14 2049-2069. 15Fairall, C.W., E.F. Bradley, J.E. Hare, A.A. Grachev, and J.B. Edson (2003), Bulk 16 Parameterization of Air–Sea Fluxes: Updates and Verification for the COARE 17 Algorithm, J. Clim., 16, 571–591. 18Grima, N., A. Bentamy, K. Katsaros, and Y. Quilfen (1999), Sensitivity of an oceanic 19 general circulation model forced by satellite wind stress fields, J. Geophys. Res., 20 104(C4), 7967-7989, doi:10.1029/1999JC900007. 21Grodsky, S. A., and J. A. Carton (2001), Coupled land/atmosphere interactions in the 22 West African Monsoon, Geophys. Res. Lett., 28(8), 1503–1506, 23 doi:10.1029/2000GL012601. 1 20 1Houze, R. A., and C.-P. Cheng (1977), Radar Characteristics of Tropical Convection 2 Observed During GATE: Mean Properties and Trends Over the Summer Season. 3 Mon. Wea. Rev., 105, 964–980. doi: 10.1175/1520- 4 0493(1977)105<0964:RCOTCO>2.0.CO;2 5Kudryavtsev, V., V. Makin, and B. Chapron (1999), Coupled sea surface‐atmosphere 6 model 2. Spectrum of short wind waves, J. Geophys. Res., 104(C4), 7625-7639, 7 doi:10.1029/1999JC900005. 8Kudryavtsev, V., D. Akimov, J. Johannessen, and B. Chapron (2005), On radar imaging 9 of current features: 1. Model and comparison with observations, J. Geophys. Res., 10 110, C07016, doi:10.1029/2004JC002505. 11Large, W. G., and S. G. Yeager (2009), The Global Climatology of an Interannually 12 Varying Air-Sea Flux Data Set, Clim. Dyn., 33, 341-364. 13Makin, V., and V. Kudryavtsev (1999), Coupled sea surface‐atmosphere model 1. Wind 14 over waves coupling, J. Geophys. Res., 104(C4), 7613-7623, 15 doi:10.1029/1999JC900006 16McPhaden, M. J., and Coauthors (1998), The Tropical Ocean-Global Atmosphere 17 observing system: A decade of progress, J. Geophys. Res., 103(C7), 14,169– 14,240, 18 doi:10.1029/97JC02906 19McPhaden, M. J., and Coauthors (2009), RAMA: The Research Moored Array for 20 African–Asian–Australian Monsoon Analysis and Prediction. Bull. Amer. Meteor. 21 Soc., 90, 459–480, doi: 10.1175/2008BAMS2608.1 22Meindl, E.A., and G.D. Hamilton( 1992), Programs of the National Data Buoy Center, 23 Bull. Amer. Meteor. Soc., 4, 984-993. 1 21 1Milliff, R. F., W. G. Large, J. Morzel, G. Danabasoglu, and T. M. Chin (1999), Ocean 2 general circulation model sensitivity to forcing from scatterometer winds. J. 3 Geophys. Res., 104(C5), 11 337–11 358, doi:10.1029/1998JC900045 4Risien, C.M., and D.B. Chelton (2008), A Global Climatology of Surface Wind and Wind 5 Stress Fields from Eight Years of QuikSCAT Scatterometer Data, J. Phys. Oceanogr., 6 38, 2379-2413. 7Rolland, J., and P. Blouch (2002), Les bouées météorologiques, Météo France 8 Publication, La météorologie 39, available online at 9 http://documents.irevues.inist.fr/bitstream/handle/2042/36252/meteo_2002_39_83.pd 10 f 11Sobieski, P. W., C. Craeye, and L. F. Bliven (1999), Scatterometric signatures of 12 multivariate drop impacts on fresh and salt water surfaces, Int. J. Remote Sens., 20, 13 2149-2166. 14Verspeek, J.; A. Stoffelen, M, Portabella, H. Bonekamp, C. Anderson, and J.F. Saldana 15 (2010), Validation and Calibration of ASCAT Using CMOD5.n, IEEE Transactions 16 on Geoscience and Remote Sensing, 48, 386-395, doi: 10.1109/TGRS.2009.2027896 17Weissman, D. E., M. A. Bourassa, and J. Tongue (2002), Effects of Rain Rate and Wind 18 Magnitude on SeaWinds Scatterometer Wind Speed Errors, J. Atmos. Oceanic 19 Technol., 19, 738–746. doi: 10.1175/1520- 20 0426(2002)019<0738:EORRAW>2.0.CO;2 21Wright, J. W., and W. C. Keller (1971), Doppler spectra in microwave scattering from 22 wind waves, Phys. Fluids, 14, 466-474. 1 22 1Zhang, H.-M., J. J. Bates, and R. W. Reynolds (2006), Assessment of composite global 2 sampling: Sea surface wind speed, Geophys. Res. Lett., 33, L17714, 3 doi:10.1029/2006GL027086. 1 23 1Table 1 Orbital parameters Satellite QuikSCAT METOP-A Recurrent period 4 days 29 days Orbital Period 101 minutes 101 minutes Equator crossing Local Sun 6:30 a.m 9:30 a.m. time at Ascending node Altitude above Equator 803 km 837 km Inclination 98.62o 98.59o 2 5 m m m i 3Table 2 Coefficients ai of P5 (WAS ) ai (WAS ) in (2) which parameterizes the i0 4difference between QuikSCAT and ASCAT wind speeds in terms of ASCAT wind speed 5(WAS ) and wind direction relative to ASCAT mid-beam azimuth ( ). Columns represent m 6coefficients ai for angular harmonics cos(m) in (2), m=0,1,2,3. 7 1 cos cos 2 cos3 Factor 1 1.6774 0.1427 0.0894 0.2229 x 1 WAS -0.7974 0.2091 0.2806 -0.2903 x 1 2 WAS 1.3854 -0.7235 -0.6268 0.8371 x 0.1 3 WAS -1.1416 0.7916 0.5320 -0.9592 x 0.01 4 WAS 0.4476 -0.3579 -0.1906 0.4681 x 1e-3 5 WAS -0.6307 0.5709 0.2322 -0.8181 x 1e-5 1 24 1 2 3 4 5 6Figure 1. Number of collocated QuikSCAT and ASCAT data on a 1ox1o grid during 7November 2008-2009 for time separations less than that shown in the left-upper corner. 1 25 1 2 3Figure 2. Difference between time mean satellite and buoy 10m neutral wind speed (ms-1) 4for QuikSCAT (QS,top) and ASCAT (AS,bottom). 1 26 1 2 3Figure 3.Scatter diagram of satellite and buoy 10m neutral wind speed for ASCAT (AS) 4and QuikSCAT (QS).Symbols represent time mean collocated winds at buoy locations. 1 27 1 2 3Figure 4. Lagged correlation of collocated satellite and buoy winds as a function of time 4separation between satellite and buoy data. Solid line shows the median value and 5vertical bars are bounded by 25 and 75 percentiles of correlation for different buoys. 1 28 1 2 3 4Figure 5. Mean difference between collocated QuikSCAT and ASCAT wind speed ( W 5), their standard deviation (STD), and temporal correlation (CORR). (Left) only rain flag 6is applied to QuikSCAT, (right) rain flag and multidimensional rain probability 7(MPR<0.05) are both applied. 1 29 1 2 3Figure 6. Difference between collocated QuikSCAT (QS) and ASCAT (AS) wind speed 4as function of ASCAT wind direction relative to the mid beam azimuth (DIRASC-AZIM1). 5Data is averaged for all wind speeds and binned into 10o in the relative wind direction. 6Downwind direction corresponds to 0o. Wind direction histogram is shown in gray 7against the right axis. 1 30 1 2 3 4Figure 7. The same as in Figure 6, but for buoy minus ASCAT collocated wind speed 5difference. 1 31 1 2Figure 8. Collocated QuikSCAT-ASCAT neutral wind speed difference (m/s) binned into 31m/s by 10o bins in ASCAT wind speed and wind direction relative to the ASCAT mid- 4beam azimuth DIRAS AZIM1. (a) Binned data, evaluated with data from 55S -55N 5to exclude high latitude areas of negative W (see Figure 5). (b) data fit by equation (2). 1 32 1 2Figure 9. (top panel) Wind speed difference (WQS WAS ). QuikSCAT rain flag and 3MRP<0.05 are applied (the same as in Figure5). (2nd panel) Wind speed difference after 4applying the correction function dW (WAS ,) . (3rd panel) Spatial distribution of time 5mean dW . (lower panel) Histogram of wind speed difference before (gray bar) and after 6(empty bar) applying of dW. 1 33 1 1 nd 2Figure 10. (top panel) Time mean inverse wind speed from ASCAT, WAS . (2 panel) 3Difference in kinematic viscosity (m2s-1) of the sea water relative to the viscosity 4evaluated at global mean SST (19oC). (third panel) QuikSCAT-ASCAT wind speed 5difference after applying the correction function dW (the same as in Figure 9). (bottom 6panel) SST dependent difference between QuikSCAT and ASCAT wind speed W (T) 7based on (Eq. 3) at =45o. 1 34 1 2 3 4 5Figure 11. Collocated QuikSCAT-ASCAT neutral wind speed difference (ms-1) binned 1o 6in SST and 1m/s in wind speed from the ECMWF analysis collocated data. There are on 7average 128,000 collocations in each bin. Only wind speed difference based on more than 82,000 collocations is shown. 1 35