The Scatterometer

1. The Scatterometer R. A. Brown 2003 U. ConcepciÓn

2. Definitions: Backscatter: microwave radar returned signal from cm-scale (capillary or short gravity) waves Model Functions: Satellite sensor surface Model Functions are the empirical correlations between observed satellite backscatter and geophysical parameters. Look Angle: The angle between radar look direction and surface wind forcing Offset angle: The angle between surface wind and surface geostrophic wind ( U10 and VG) PBL (Planetary Boundary Layer): A filter through which all data must pass to use scatterometer data for atmospheric study R. A. Brown 2003 U. ConcepciÓn

3. Scatterometry Basics appraisal  Data: cm-scale, average density in a footprint. 50km  25km  7km  100m (SAR) Theory:State: 1-10, poor to excellent  Wind generation of water waves 1  % energy into short/long waves 2  Wave-wave interaction 3  PBL wind (log layer, Ekman layer) 8 Parameterizations  U10 (u*) land 7  U10 (u*) ocean 5  PBL (similarity) 7  Scatterometer Model Function  u* (o) 4  U10 (o) 8 P (o) 7  Note: There’s room for new phenomena R. A. Brown 2003 U. ConcepciÓn R.A. Brown 2000

4. The microwave backscatter can be expected to vary with: • Wind stress, hence • Surface wind speed • Incidence angle • Angle between radar look • and wind forcing  A Measure of the Success of a wind vector Model Function is the recognizable periodic response of backscatter in a global data set.  This will appear if data are binned in small increments of wind speed and incidence angle, and backscatter vs look angle are plotted.  The backscattershould rise and fall sinusoidally as look direction is up, down, or across the wind. R. A. Brown 2003 U. ConcepciÓn R.A. Brown, 11/99

5. The Struggle to get Satellite scatterometers R. A. Brown 2003 U. ConcepciÓn

6. A Brief History of Scatterometers 1970 Conception SeaSat Built --- with Scat, SAR, SMMR, Alt SeaSat Launch --- Lasts 99 days 1980 NSCAT conceived and built Dark Ages: Launch $ to Gulf & Carribean wars, to Refurbish battleships, to build 200 ships, Star Wars 1990 ERS-1 Launch NSCAT launched on ADEOS --- 9 mos. ERS-2 Launch Quikscat Launch 2000 Dark Ages – II USA? Star Wars – II SeaWinds on ADEOS - II ESA A-SCAT 2010 R. A. Brown 2003 U. ConcepciÓn

7. Satellite Measurements • Scatterometer Algorithms Correlated to • global buoy data • global NMC surface winds • Radiometer Algorithms (SSM/I) Altimeter, etc • Correlated to buoys • Lidar Measurements (Satellite) • Doppler return • Point measurement: Averaging problem R. A. Brown 2003 U. ConcepciÓn

8. Revelation # 1 • Storms are: • Often misplaced by numerical models; • Stronger (deeper Pressures), • More frequent than found in GCMs and climatology records R. A. Brown 2003 U. ConcepciÓn

9. Storms Conclusions  Resolution is better  Location, Fronts, Storm Strength are better  New dynamics revealed with respect to climatology,Storms contribute : 10% total latent heat flux 20% sensible heat flux 30% total ocean stress Storms analysis with Satellite scatterometers data compared to conventional (GCM) analyses: R. A. Brown 2003 U. ConcepciÓn

10. Revelation # 2 There exist large regions of High Winds (1000km2/storm) that nobody knows of…… These do not appear in: • GCM analyses • Buoy data • Climate data • Satellite data R. A. Brown 2003 U. ConcepciÓn

11. Revelation # 3 Fronts: Defined as lines of different sea state --- roughness variation Are: • Ubiquitous • Persistent • Mysterious R. A. Brown 2003 U. ConcepciÓn

12. Applications Better GCM Progs (initialization) Better Storms Definition Evolution of Fronts & Cyclones Higher Winds (heat fluxes) in Climate models Proof of Rolls (OLE) Ubiquity Better Hurricane PBLs, Initialization, forecasts R. A. Brown 2003 U. ConcepciÓn

13. Storms analysis with Quikscat R. A. Brown 2003 U. ConcepciÓn

14. In this first example, the black dot shows where the frontal wave will later develop (note that it is an ideal col point). The main feature of interest is highlighted by a small black arrow in the bottom panels. As the surface front develops and strengthens, a short wave on the upper-level jet moves in. The subsequent growth of that upper-level feature is coincident with the deepening of the frontalwave beneath it, which suggests a baroclinic growing mode in the later stages. R. A. Brown 2003 U. ConcepciÓn

15. In this second example, the environment in which the surface frontal wave rows is very different. A large meridionally elongated surface trough extends North to Brazil and has a very similar counterpart at upper-levels, in the form of a meridional elongated trough with a strong vorticity signature. The upper-level front is lagging slightly westward compared to the surface front. The upper-level trough and high-vorticity are already present above the surface front 48 hours before the occurrence of the frontal wave. R. A. Brown 2003 U. ConcepciÓn

16. In the third example, a similar development is observed, with a front oriented in a southeast-northwest direction. R. A. Brown 2003 U. ConcepciÓn

17. 1. In the first case, the front strengthens as a result of frontogenesis due mainly to the divergent ageostrophic cross-frontal circulation. Then three processes become important: (i) The along-front stretching associated with the environmental flow decreases with time and reaches a minimum of 0.2e-5 s-1. (ii) The environmental flow becomes frontolytic. (iii) A short upper-level wave moves in above the surface front. The front then becomes unstable and a secondary cyclone develops, with a clear signature in both the surface and upper-level fields. 2. In the second and third case, the large-scale trough in which the front is embedded also exists at upper-levels, with high values of vorticity. The frontal surface has a slight westward tilt with height. 3 Frontogenesis is relatively high but not due for the most part to the divergent ageostrophic wind. The environmental flow does not become frontolytic although the along-front stretching associated with it decreases. When it reaches a minimum of 0.2e-5 s-1 in the second case and 0.6e-5 s-1 in the third case, the secondary cyclone develops. The first scenario resembles the typical frontal waves observed in the North Atlantic (e.g. Rivals et al., 1998). In all cases, there is a clear connection between the surface and the upper-levels, and the subsequent cyclone growth seems to have a baroclinic component. However, the ``timing'' of the connection with the upper-level feature in the first case (i.e. with the fast-moving short wave in the upper-level jet) is more critical than in the second and third case. R. A. Brown 2003 U. ConcepciÓn

18. The fourth example is a front with a south-north orientation and a tail reaching southwest Australia. There is a well-formed upper-level trough corresponding to the surface trough, existing in the early stages of the front life cycle. However, over its 8-day life cycle, there is no sign of instability or wave growth in the front. Since high values of vorticity do exist at upper-levels, it suggests that the surface conditions might not be optimal in this case for the growth of a vortex on the tail of the front. The environmental stretching deformation might not relax to sufficiently small values. R. A. Brown 2003 U. ConcepciÓn

19. In this last example, a long-lived front is shown over the Indian Ocean. An instability can be detected (black dot) but no vortex deepens significantly. An incipient frontal wave, as one might call it, appears at one synoptic time, but is damped out in the following hours by a strengthening of the front. The front maintains its shape (a straight line) and orientation (southeast-northwest) for 4 more days as it moves eastward. Then only, as the tail of the front approaches Australia, a vortex appears and grows into a small cyclone. R. A. Brown 2003 U. ConcepciÓn

20. We have shown, Scatterometer versus GCM: * Storm Fronts are located more accurately, (e.g. Brown, 1980, Patoux & Brown, 2001) * Fronts have more detail, & more extent (Dickinson & Brown, 1996, QuikScat ‘99) * Pressure fields are more accurate (Brown & Levy, 1986; Brown & Zeng, 1994; Zeng & Brown, 1997, Brown, 2000, Patoux & Brown, 2001) * There are 5% more storms in Tropics & Southern Hemisphere than seen by GCMs, 10% more than appear in climatology (Brown & Zeng, 1994) * GCM and buoy winds, hence many (not all) scatterometer model functions, have systematic low winds (10%). (Foster & Brown, 1994; Zeng, 1996, Brown & Zeng, 1996, 2000). * Scatterometer data and UW PBL model suggest very high wind regions (>40m/s) are missing from the GCM analyses. R. A. Brown 2003 U. ConcepciÓn

21. Programs and Fields available http://pbl.atmos.washington.eduQuestionsto rabrown, neal or jerome@atmos.washington.edu • Direct PBL model: PBL_LIB. (1975) An analytic solution for the PBL flow with rolls, U(z) = f( P, To , Ta , ) • The Inverse PBL model: Takes U10 and calculates surface pressure fields P (U10 , To , Ta , ) (1986) • Pressure fields directly from the PMF: P (o) along all swaths (2003) • Surface stress fields from PBL_LIB corrected for stratification effects along all swaths (2004) R. A. Brown 2003 U. ConcepciÓn