Typhoon Roke bears down on Japan; 98L continues to grow more organized
Powerful Category 3 Typhoon Roke is bearing down on Japan, and is expected to hit the main island of Honshu on Wednesday morning, local time. Roke is on a dangerous track for Japan, one that would take the storm over some of the most heavily populated areas of the country. Heavy rains from Roke have already reached the coast of Japan, as seen on Japanese radar. However, Roke is starting to weaken, as seen in latest satellite imagery. The eye is no longer apparent, the cloud tops have warmed, and a slot of dry air has gotten wrapped into the storm's northwest side. Wind shear should continue to weaken Roke as it approaches landfall; shear is currently a high 20 knots, and will increase to 30 knots by Wednesday morning. Given the current weakening trend, I expect Roke is most likely to be a Category 2 typhoon at landfall.
Typhoon's Roke's storm surge, winds, and heavy rains will all be a concern. A damaging storm surge is likely to the right of where the center makes landfall, since Roke is a large storm whose winds are spread out over a wide area. If Roke tracks farther to the east than expected, a large storm surge may affect Tokyo Bay. Perhaps the biggest concern from the storm is heavy rain. The soils over much of Japan are saturated from the passage of Tropical Storm Talas during the first week of September. Talas was a very slow moving storm, and brought extreme rainfall amounts of over six feet to some portions of Japan. Roke is expected to bring up to 20 more inches of rain along its path. Roke could bring winds of 30 - 40 mph and heavy rains of 3 - 5 inches to the damaged Fukishima-Dai-Iche nuclear plant northeast of Tokyo.
Figure 1. MODIS image of Typhoon Roke taken at 1:45 pm local time (4:45 UTC) on Tuesday, September 20, 2011. At the time, Roke was a Category 2 storm with 110 mph winds. Image credit: NASA Earth Observatory.
Invest 98L continues to grow more organized
A tropical wave midway between Africa and the Lesser Antilles (Invest 98L) has increased in organization this afternoon, but still lacks a well-defined surface circulation. Satellite imagery shows a number of curved spiral bands have formed this afternoon, and the area covered by heavy thunderstorms has steadily increased. An ASCAT pass from 8:21 am EDT this morning did not capture the full circulation of the storm, but did show winds of 30 mph on the east side of the center. Wind shear as diagnosed by the SHIPS model has increased to a moderate 10 - 15 knots, and is predicted to stay moderate through Friday. Ocean temperatures are 28 - 28.5°C, well above the threshold typically needed for a tropical storm to spin up. Water vapor satellite images show 98L is embedded in a moist environment, but there is dry air to the system's northwest. Given that the shear has now increased to the moderate level, this dry air may begin to hinder development on Wednesday. An analysis of upper level winds from the University of Wisconsin CIMMS group shows a pattern favorable for development, with an outflow channel open to both the north and south available to ventilate the storm and allow 98L to efficiently lift plenty of moisture to high levels.
Figure 2. Afternoon satellite image of 98L.
The latest 8 am EDT (12Z) runs of the computer models show either no development of 98L, or development of 98L into a tropical depression or weak tropical storm by Saturday. 98L's westward motion of 5 - 10 mph should bring the storm into the Lesser Antilles Islands by Saturday, though the models are not in strong agreement about the forward speed of the storm. The GFDL model brings 98L into the islands on Friday, while the HWRF model keeps the storm east of the islands through Sunday. If 98L takes a more west-northwesterly path through northern Lesser Antilles, which has been the preferred track for tropical systems this year, the disturbance should encounter high wind shear in excess of 20 knots due to strong upper-level winds out of the west. This shear should make it difficult for 98L to intensify as it moves though the islands. However, if 98L takes a more southerly path across Barbados, as predicted by the GFDL model, the storm will miss seeing the high shear area that lies over the northern islands, and the storm would have more opportunity to strengthen. The most likely scenario I see at this point is for 98L to be a weak tropical storm on Saturday as it moves through the Lesser Antilles--but there is more than the usual amount of uncertainty in both the track and intensity forecast. NHC gave the disturbance a 70% chance of developing into a tropical depression by Thursday in their 2 pm Tropical Weather Outlook.
September temperatures return to normal over the U.S.
The summer of 2011 was the second hottest in U.S. history, but September of 2011 is so far shaping up to be an average one for temperature. A series of cold fronts and cold-cored low pressure systems have moved southwards out of Canada this month, bringing typical amounts of cool air to the country. If you want to select dates for the start and end of the U.S. heat wave of 2011, the dates to pick would be May 20 - September 4. During the period May 20 - September 4, 2011, the number of daily record high temperatures at the 515 major airports in the U.S. exceeded the number of daily low temperature records every day but one. That's an astonishing 107 out of 108 days! Only July 15 had more record daily lows than highs during that 108-day period. I doubt one could find a similar stretch of days anytime in U.S. weather history where such a lopsided ratio of high temperature to low temperature records existed. For the 3-month summer period of June, July, and August, 2703 daily high temperature records were set, compared to 300 daily low temperature records--a ratio of 9-to-1. Not surprisingly, the summer of 2011 wound up as the hottest summer in 75 years in the U.S., and was only 0.1°F cooler than the all-time record hottest summer, during the Dust Bowl year of 1936. But so far this September, the ratio of high temperature records to low temperature records has been close to 1-to-1. There were 283 daily high temperature records set during the first sixteen days of September, and 246 low temperature records. Eight of the first sixteen days of September have seen the lows outnumber the highs, and eight have seen the highs outnumber the lows. The latest 2-week forecast from the GFS model shows a continuation of pretty normal weather over the U.S. for the rest of the month, and September temperatures will end up close to average.