cycloneye wrote:WHXX04 KWBC 051721
NCEP COUPLED GFDL HURRICANE MODEL FORECAST MADE FOR
TROPICAL DEPRESSION INVEST 99L
INITIAL TIME 12Z SEP 5
FORECAST STORM POSITION
HOUR LATITUDE LONGITUDE HEADING/SPEED(KT)
0 29.0 [b] 71.6 95./ 8.0
6 29.5 70.5 67./10.6
12 29.9 69.7 64./ 7.8
18 30.0 69.4 65./ 3.1
24 30.4 69.3 24./ 3.9
30 30.5 69.9 282./ 5.5
36 30.4 70.7 266./ 6.7
42 30.4 71.9 268./11.1
48 30.2 73.4 264./12.5
54 30.1 74.2 259./ 7.1
60 30.3 74.8 290./ 5.8
66 30.8 75.3 313./ 6.4
72 31.5 75.9 320./ 9.3
78 32.4 76.6 321./10.0
84 32.9 77.2 316./ 7.1
90 33.8 77.2 354./ 8.3
96 34.8 77.3 359./10.1
[/b]
http://my.sfwmd.gov/sfwmd/common/images ... r/gfdl.txt12z GFDL has landfall in NC.Initial position is almost where it is now.
Luis' last statement is a great observation imho and very pertinent! The initialized 12Z
GFDL longitude as of 12Z today is 71.6 W, which appears to be almost perfect. Therefore, this run has lots of credibility imho as of now. It now moves it to as far east as 69.3 W (12Z Thu). Due to this pretty far east position, which is realistic to me based on the
nice initialization, it gets the storm only back as far west as 77.3 W (NC).
Now, I'll compare this to some other 12Z models:
1)
GFS had good initialization and only gets it back to
NC.
2)
Canadian had a poor initialization of a too far west 73 W. Due to this, it only gets as far east as ~72.3 W (0Z Thu), which allows it to come back all the way to near the
GA/SC border.
3)
NOGAPS had a very poor initialization of a much too far west 74 W. Due to this, it only gets as far east as ~72 W (12Z Thu), which allows it to come back all the way to near
Charleston.
Summary for these 12Z models: GFS/GFDL had best initialization and, therefore, avoid SC/GA.
CDN/NOGAPS had too far west initialization, which allows GA/SC border and Charleston hits. I think the pattern is pretty clear and, therefore, continue to favor NC for a hit, assuming any of the SE is directly hit.