Get documents about "Wind Direction Editors for Use in the Generation of
Document Sample
Wind Direction Editors for Use in the Generation of Synthetic Aperture Radar Wind Speed Imagery
Todd D. Sikora, Millersville University
George S. Young, Pennsylvania State University
Nathaniel S. Winstead, Johns Hopkins University Applied Physics Laboratory
Background Method Results
• SAR provides a unique capability for high • Manual editing of gridded wind direction data
resolution remote sensing of wind speed over – Allows for the broadest combination of data types and the
oceanic regions maximum input of human insight
– Disadvantage is amount of analyst’s time and
• Requires accurate a priori wind direction fields computationally demanding re-derivations of the wind
speed
•1200 m pixel SAR-derived wind speed field for a barrier jet near 60 degrees north
• Manual morphing of gridded wind direction and 143 degrees west in the Gulf of Alaska derived from SAR backscatter data
collected by RADARSAT-1 at 0310 UTC on 18 February 2000
data – Left panel shows SAR-derived wind speed field with the barrier jet too enhanced because it was not resolved in the
first-guess wind direction field
– Reposition and reshape features of the existing wind
– Right panel shows the wind speed field re-derived after manually editing the wind direction field to correct this wind
direction field direction error
• No one source of wind direction fields is • Identify error signatures
– Fronts: Fast and slow bands
entirely adequate for this application – Singularities: Hourglass signature
• Specify paired (before and after transformation) control points
–NWP model wind direction fields can have position errors – The wind direction data from the before point will be relocated to the after point
and either missing or extraneous synoptic and mesoscale • Displaced information of the control point pairs is used to interpolated to
wind shift features the location of each first-guess wind direction point using objective
analysis
–Scatterometer observations can be contaminated by the – Cubic interpolation inside the region of the control points and nearest neighbor interpolation outside the
region of the control points
mishandling of the ambiguity resolution near frontal and
– Disadvantage • 1200 m pixel SAR-derived wind speed field for a cyclone near 55 degrees north
vortex wind shifts
• Does not allow the analyst to make full use of all available data sources and 180 degrees west in the Bering Sea derived from SAR backscatter data
–SAR can provide wind direction data only in those collected by RADARSAT-1 at 1818 UTC on 6 December 2000
regions where meteorological phenomenon yield along- • Semi-automated morphing – Left panel shows SAR-derived wind speed field with HG marking the location of the hourglass signature of a
misplaced cyclone in the first-guess wind direction field
wind streaks in the surface wind speed field – For the case of a misplaced singularity – Right panel shows the same wind speed field re-derived after semi-automated morphing of the wind direction field
• The problem has many fewer degrees of freedom than correcting the reduces, but does not eliminate, the position error
• This paper presents methods for editing gridded position of a front
(e.g., scatterometer or NWP model) wind – The analyst provides the Nelder-Mead simplex algorithm
direction data to mitigate wind direction errors with the singularity’s position in the first-guess wind
direction field
• The algorithm then iterates to a solution that minimizes the hourglass
(cos2φ) signature
– Disadvantage
• Typically requires many iterations to arrive at a solution and thus requires
more wall clock time • 1200 m SAR-derived wind speed field for a cold front near 58
• All examples presented are for correcting NWP degrees north and 154 degrees west in the Bering Strait derived
from SAR backscatter data collected by RADARSAT-1 at 1957
model wind directions UTC on 08 January 2004
– Left panel shows SAR-derived wind speed field with the fast-band signature of a misplaced front
This work was supported in part by grant ATM-0240869 from the National Science Foundation and grants N00014-04-10539, in the first-guess wind direction field (The true frontal position is marked TF)
N00014-04-WR-20365, N00014-05-WR-20319, and N00014-06-10046 from the Office of Naval Research. The SAR imagery used in – Right panel shows the same wind speed field re-derived after manually morphing the wind
this paper were provided by the National Oceanic and Atmospheric Administration’s Ocean Remote Sensing Program via the Alaska direction field to correct this position error and sharpen the frontal gradient
SAR Demonstration Project.
Related docs
