Transcript Slide 1

Active Radars
R. A. Brown 2005 Miami
From Neil Tyson’s address/campaign
On the Future of NASA Univ. Wash. Jan 20, 2005
President’s commission --- “Vision”
Winners:
Space Exploration
(thing)
Planetary Science
Astrobiology
Astrophysics
Astronomy
Losers:
Beyond Einstein missions
Earth Science
“LEO (low earth orbits) are old hat and boring. NASA must do new stuff – space”
R. A. Brown 2004 EGU
Winds --- see elsewhere
R. A. Brown 2004 EGU
Toward a Surface Pressure Model Function
• A Scatterometer doesn’t measure Winds. It measures
Capillaries & Short Gravity Waves, related to zo., u*
• Fortunately, there exists a relation
U10/u* = F(z, zo, stratification…).
– Established over land, assumed over the ocean .
– Verified in the 26 years since Seasat.
There’s an easy extrapolation to:
• Fortunately, there exists a relation
UG/u* = F(z, zo, stratification, , ……).
– Established in UW PBL_LIB.  is a constant, a similarity
parameter.
– Verified in the last 10 years of satellite data
R. A. Brown 2004 EGU 2005 Miami
Try a direct correlation with pressure
Since VG = P / ( f )
Use ECMWF/NCEP surface pressures analyses get P and VG ;
substitute VG for U10 in the Model Function i.e. We’re using the
model surface pressures as truth rather than the PBL analysis of U10
Results: VG correlates with o as well as U10
Prospects:
* Better alias selection with scatterometer data alone
* High winds appear
* Low winds, directions appear
* Stratification, Thermal Wind Effects
R. A. Brown EGU 2004
Example of o
vs look angle
for U10 = 20m/s;
Incidence = 45
Example of o
vs look angle
for VG = 27m/s;
Incidence = 45
R. A. Brown 2004 EGU
R. A. Brown 2004 EGU
Producing smooth wind fields
Usually from rain
Direction information is poor
R. A. Brown 2003 U. ConcepciÓn
(JPL)
UW
Pressure field
smoothed
Raw scatterometer winds
JPL Project
Local GCM nudge
smoothed = Dirth
R. A. Brown 2003 PORSEC, U. ConcepciÓn
QuikSCAT
• QuikSCAT pressure gradients
are comparable to ECMWF in
general
– QuickSCAT implies stronger
gradients in frontal zones.
• Comparison with ECMWF
similar in both hemispheres
• Good correlation overall
• We’ve grown accustomed to
this level of quality in QS-SLP
Foster, 2005
Swath QuikScat
surface pressure
fields available at
http://PBL.atmos.
washington.edu
Dashed:
ECMWF
R. A. Brown 2005
Surface
Pressures
QuikScat analysis
ECMWF analysis
J. Patoux & R. A. Brown 2003 PORSEC, U. ConcepciÓn; 2005 Miami
Results from Satellite Scatterometer
surface pressure analyses:
 Agreement between satellite and ECMWF pressure fields
indicate that both Scat winds and the nonlinear PBL model
(VG/U10) are accurate within  2 m/s.
 3-month, zonally averaged offset angle <VG, U10> of 19°
suggests the mean PBL state is near neutral. This is the nonlinear
PBL predicted angle (18°).
• SLP gradients (e.g. from buoys) provide surface truth for VG,
hence U10.
 Swath deviation angle observations show thermal wind and
stratification effects, implying temperatures.
 Predicted higher winds from pressure gradients (than from
GCM or buoys) agree with OLE effect, observations.
 VG rather than U10 could be used to initialize GCMs
R. A. Brown 2004 EGU
R. A. Brown 2004 EGU
Storms & Fronts Analyses
Fronts: Location; Analysis; Frontogenesis; Prognosticators
In the second case, the system is decaying but a
secondary low is developing behind the remnants of
the cold front. Note also the correspondence between
convergence and clouds.
R. A. Brown, J. Patoux003
Analysis of QuikScat derived
fronts and pressure fields
suggests there are correlations
between frontal characteristics,
upper level conditions (PV) and
subsequent development of
explosive storms development
Patoux, J., PhD thesis Univ. of Washington, 2003. MWR in press
R. A. Brown 2004 EGU
Southern Hemisphere
Pressures
ECMWF & NSCAT
Comparison
• Surface Pressure Fields of 102
Storms surveyed for 1996:
* 25% good matches (-3 mb ave. diff.)
* 70% misplaced average an 280 km
* 5% missed entirely (vs 20% in 1990)
R. A. Brown 2003 PORSEC, U. ConcepciÓn; 2005 Miami
Revelations from scatterometers
• Great global surface pressure fields are available daily.
• The winds are higher; the low pressures are lower &
more frequent; heat fluxes are greater; and surface
stress is much greater than climatology states.
• Data on storms and fronts is exceptional. (Patoux, J., G.J.
Hakim and R.A. Brown, 2004: Diagnosis of frontal instabilities over the
Southern Ocean, Monthly Weather Review, in press)
• Storms frequency, strength and statistics are different
• Fronts (defined as wind convergence zones) are ubiquitous,
persistent and provide new data
The global marine wind and pressure data from scatterometers
(the 19 deg turning) and SARs (OLE surface imprints) indicate
that the nonlinear OLE (Rolls) are present 50 – 70% of the time.
Hence the nonlinear PLB model prevails (the Ekman solution
does not exist). While this is a nonlinear finite perturbation, it
can have large effects on measurements 10-km or less and in
the mean.
Air-Sea fluxes are non-homogeneous, take place in advective
plumes, and interact with the inversion. New PBL models are
needed to get good heat and momentum fluxes for ocean and
climate modeling.
K-theory (diffusion modelling) is physically incorrect for
modeling these fluxes.
Winds are higher than climate and ocean modelers thought.
• In addition to better initialization of GCMs;
• Global marine surface winds and pressures are
best available
• Storms and fronts analyses are revolutionary.
– Provides surface truth for storms
– Provides statistics for storms
– Possibility of predictors of storms genesis.
• Winds and fluxes are different than climatology
records. Climate and ocean dynamics modelers
take note.
The Contributions of Scatterometry
• The microwave scatterometers have now provided over two decades of
inferred surface winds over the oceans. These data have been extensively
studied and compared to in situ measurements so that they comprise a
‘surface truth’ base comparable to other sources of winds. In fact, in many
cases these products are revolutionary, changing the way we view the
world. Examples are:
Mainline products
•
•
The surface winds are exceptional in resolution and coverage.
The nonlinear solution applied to satellite surface winds provides sufficient
accuracy to determine surface pressure fields from satellite data alone. We
can uniquely offer a continuous record of QS-derived surface pressure
fields: these pressure fields extend through the Tropics - a region that is
poorly characterized by numerical weather forecast models - and contain
fine details that are absent from numerical model analyses
• The pressure fields can be used as a low pass filter to aid ambiguity
selection and provide smooth wind fields from scatterometer data alone.
From the Scatterometers (2)
Storms and Weather Revelations
• The scatterometer data allow study of the development of fronts in general
and frontal waves in particular: QS reveals mesoscale features that are not
captured by numerical models or other satellite-borne instruments, in
particular the surface signature of frontal instabilities that sometimes
develop into secondary cyclones. (Patoux, J., G.J. Hakim and R.A. Brown,
Diagnosis of frontal instabilities over the Southern Ocean, Monthly Weather
Review, in press).
•
These data allow us to build a climatology of primary and secondary
cyclones (in particular their kinematics as revealed by scatterometer winds),
to test the hypothesis that explosive frontal storm development may have
surface predictors (e.g. surface PV anomaly coincident with upper-level
vorticity) and to investigate the possibility that the strength of storms and
fronts is increasing due to global warming
• Capturing storm and frontal dynamics require at least 25-km resolution.
New revelations are stimulating and surely forthcoming from QuikScat data.
(Brown, R.A., Comments on the synergism between the analytic PBL model
and remote sensing data, Bound.-Layer Meteor., in press)
From the scatterometers (3)
Basic Revelations
• The numerical global models of the 90s were inadequate in representing
Southern Hemisphere and tropical weather systems. In 1991, they missed
20% of the So. Hemisphere storms. After the scatterometer revelations, the
numerical models improved resolution and incorporated satellite data so
they now (2004) miss only 5% of the storms (tho miss-locating 70% by an
average 250km). QuikScat data forms the basis of this evaluation.
• The data indicate that the global climatology surface wind record is too low
by 10 – 20%.
Basic Science
•
There is evidence from these data that the secondary flow characteristics
of the nonlinear PBL solution (Rolls or Coherent Structures) are present
more often than not over the world’s oceans. This contributes to basic
understanding of air-sea fluxes.
•
The revealed dynamics of the typical PBL indicate that K-theory models
are physically incorrect. This will mean revision of all GCM PBL models.
QuikScat data will help convince them.
• The conclusions from these observations are important yet often ignored by
the modeling community. The continued accumulation of data from
QuikScat is essential to wake up the community.
SLP from Surface Winds
• UW PBL similarity model
N
10
U
u
10
 log
k
zo  u 
G
 f (P, T10 , SST , q10 , CS , )
u
N)
(U
• Use “inverse” PBL model to estimate
G
N
from satellite U10
• Use Least-Square optimization to find best fit
P(UGN )
SLP to swaths
• Extensive verification from ERS-1/2, NSCAT,
QuikSCAT
R. A. Brown 2005
Using SLP to Assess Direction
• Winds derived from SLP are optimal
smooth winds
• Arbitrary threshold of 35o from Model U10
used to distinguish potentially wrong
ambiguity choice
• Look for an ambiguity with closer direction
to Model winds in these cases
R. A. Brown 2005
Taking measurements in the Rolls
Hodograph
Hodograph
from convergent zone from center zone
1-km
The OLE
winds
Station A
3
U
2
Z/
2 - 5 km
The
Mean
Wind
1
Station B
V Mean Flow Hodograph
RABrown 2004
Taking measurements in the Rolls
Hodograph
Hodograph
from convergent zone from center zone
1-km
The OLE
winds
Station A
3
U
2
Z/
2 - 5 km
The
Mean
Wind
1
Station B
V Mean Flow Hodograph
RABrown 2004
The Gradient Wind Correction
The gradient wind correction is described in Patoux and Brown
(2002) and uses the simple balance of forces in natural
coordinates shown in the figure
R. A. Brown 2003 U. ConcepciÓn
Geostrophic balance
Gradient balance
R. A. Brown 2004 EGU
On the right, the gradient wind correction has been included. The
obtained pressure field is very similar to the uncorrected one, except
for the center of the anticyclone, where the radius of curvature is
smaller, and the effect of the correction bigger. The pressure gradients
are weaker and the central area of the high is flatter, which seems in
better agreement with ECMWF.
R. A. Brown 2003 U. ConcepciÓn
Best surface winds, pressures available
Much Better Storms Depiction
Shows Evolution of Fronts & Cyclones
Proof of Rolls (OLE) Ubiquity (PBL model)
Higher Winds (heat fluxes)
for Climate models
Better Hurricane PBLs, GCM
Initializations, forecasts
R. A. Brown 2005
Revelations from scatterometers
* Great global surface marine winds are available daily.
* Great global surface pressure fields are available daily.
* Ship or Buoy winds are not good surface truth in general.
* GCM PBL models have the wrong physics.
* The oV saturates (due to white water) @ U10 ~ 35 m/s,
but the oH does not saturate even at U10 ~ 65 m/s.
* The winds are higher, the low pressures are lower &
more frequent, heat fluxes are greater and stress much
greater than climatology states. Climate modelers take note.
* Scatterometer derived pressure fields can be used to
de-alias winds and correct (smooth) o single or small area
anomalies (rain or nadir/edge ambiguities).
* Data on storms and fronts is revolutionary. (Patoux,
J., G.J. Hakim and R.A. Brown, 2004: Diagnosis of frontal instabilities over
the Southern Ocean, Monthly Weather Review, in press)
1. The winds are non-homogeneous at the surface over a 0.1-5 km
horizontal distance. Upper high velocity wind is advected to the
surface in lines. OLE must be taken into account in surface truth
measurements (In the average and point values).
(Brown, Canadian Jn. Remote Sensing, 28, 340-345, 2002)
2. The average wind profile is different from the Ekman solution –
and from a profile 100m away – the nonlinear wind solution (and
hence fluxes of momentum, heat, CO2 etc.) is 10-50% different,
depending on stratification and thermal wind. In a satellite’s 25-km
footprint there will be 10-OLE so the periodic effect will be the
average. Not true at 6-km or less --- the SAR resolution. (Brown &
Foster, The Global Atmos.-Ocean System, 2, 163-183, 1994; 185-198, 1994; 199-219, 1994.)
3. The PBL contains advecting flow not amenable to diffusion
modeling. Numerical models cannot portray correct physics of
mean flow without extreme increase in resolution.
R. A. Brown 2005
Dashed:
ECMWF