first results from a high altitude sfmr
DESCRIPTION
First Results from a High Altitude SFMR. Alan S. Goldstein, NOAA/Aircraft Operations Center Dr Eric Uhlhorn, NOAA/Hurricane Research Division. Basic SFMR Theory. Data Handling. Raw counts converted to Tb’s using AOC-derived calibration Linear Least-Squares fit outlier filter – same as P-3s - PowerPoint PPT PresentationTRANSCRIPT
First Results from a High Altitude SFMRFirst Results from a High Altitude SFMR
Alan S. Goldstein, NOAA/Aircraft Operations Center
Dr Eric Uhlhorn, NOAA/Hurricane Research Division
Alan S. Goldstein, NOAA/Aircraft Operations Center
Dr Eric Uhlhorn, NOAA/Hurricane Research Division
Basic SFMR Theory
Brightness Temperatures (theoretical)
100
120
140
160
180
200
220
4 5 6 7 8GHz
Ke
lvin 0 WS 0 RR
30 WS 0 RR
0 WS 30 RR
30 WS 30 RR
Data Handling
• Raw counts converted to Tb’s using AOC-derived calibration• Linear Least-Squares fit outlier filter – same as P-3s• Standard retrieval algorithm and emissivity curve• SSTs from MW-IR data• Data comparison Wind Speeds and Rain Rates are 30 sec averages• Winds in HDOB example are highest 10 sec avg over 30 sec span• Dropsonde reference winds are .83 * L150 from TempDrop reports
Typical Flight Track
080911n Flight Track - Hurr Ike
21
23
25
27
29
31
-98 -96 -94 -92 -90 -88 -86 -84 -82
Triangles are dropsonde splash points
30 sec avg vs Dropsonde
080911n - Hurr Ike
0
5
10
15
20
25
30
35
40
17
:32
:00
18
:02
:00
18
:32
:00
19
:02
:00
19
:32
:00
20
:02
:00
20
:32
:00
21
:02
:00
21
:32
:00
22
:02
:00
22
:32
:00
23
:02
:00
23
:32
:00
0:0
2:0
0
0:3
2:0
4
SFWS (m/s)
SFRR (mm/hr)
Sonde WS (m/s)
Error Scatter Plot – 5 Flights, 128 points
2.5 m/s RMS Error0
5
10
15
20
25
30
35
0 5 10 15 20 25 30 35
Sonde m/s
SF
MR
m/s
Potential Real-Time ProductPeak 10 Sec Wind Average in 30 Sec Window
25
27
29
31
-94 -92 -90 -88 -86 -84
>64 kts
>50 kts
>34 kts
<34 kts
Differences from Low Altitude and TurboProp Ops
• Bigger footprint (~5x) means some data smoothing• Faster airspeed may produce noisier data (may not be noticed at 10 second average; may be reduced by footprint data smoothing)• Roll angle issues
• Sometimes G-IV makes gentle turns that do not exceed the roll limit threshold – emissivity equation should compensate• Ocean patch being sampled can be further off-track• Don’t want to tighten up on roll threshold because of aircraft behavior in turbulence
Some Remaining Issues
• Intervening Precipitation• Performance seems reasonable, but more data needed• Model makes assumption that all precip is below 4 km, may not be accurate for tropical convection• May need to wait for Tail Doppler Radar (TDR) to characterize vertical precip profile
• Use aircraft pitch angle in retrieval algorithm• RF interference from TDR (probably a non-issue)• Transition to Operations
• Decision – are we ready to share?• Integrate AOC algorithm into G-IV data system – 3/09• How to get data to the ground - HDOB? But some fields (e.g. humidity) will need to be left blank