Dr. Jeff Masters' WunderBlog

Wunderground launches BestForecast; Giovanna kills ten in Madagascar
Posted by: Dr. Jeff Masters, 4:41 PM GMT on February 15, 2012 +36
Weather Underground's meteorologists have spent over five years developing and testing a powerful forecasting system called BestForecast, which has been used to provide all of our forecasts for non-U.S. locations for the past several years. After some final improvements made in the past year, the forecasts from BestForecast have become competitive with forecasts from the National Weather Service (NWS) over much of the U.S. As of today, BestForecast forecasts are the default on the site. What's cool about BestForecast is that we can make its forecasts specific for any site that collects weather data. We gather several years of weather data from a site and optimize the forecast to suit the unique microclimate of a particular station. Thus, "backyard meteorologists" that own and maintain one of the more than 22,000 personal weather stations that record and send live weather conditions to Weather Underground will now have a forecast specifically generated for their own backyard. BestForecast also gives the expected precipitation amounts (in inches), and provides ten days of forecast information, instead of the seven days provided by the National Weather Service.

Users can evaluate the reliability of these forecasts themselves and get a second opinion by switching back to the National Weather Service forecasts that were previously published. In some areas, the National Weather Service will out-perform BestForecast, so play around with using both, and see what works the best for your location. Web site visitors can switch between best forecast and NWS forecasts using the switch "BestForecast" ON|OFF at the top of the forecast page. To create transparency in our forecasts, wunderground.com will publish the recent accuracy of its temperature forecasts over the past 20 days for every location, alongside the accuracy of the NWS. The accuracy is given in terms of the Root Mean Square Error (RMSE), in degrees Centigrade. A lower RMSE is better. The "MaxT" number is the verification of the daily high temperature forecasts, while the "Average" number is for the hour-by-hour forecasts.

A video demonstration of BestForecast is available on the wunderground.com About Our Data page.


Figure 1. Visible image from NASA's Terra satellite of Tropical Cyclone Giovanna over Madagascar, taken at 10:45 UTC Wednesday February 15, 2012. At the time, Giovanna was a tropical storm with 45 mph winds. Note the extensive plume of runoff and sediment stirred up by the storm flowing southwards along the east coast of Madagascar. Image credit: NASA.

Tropical Cyclone Giovanna kills ten in Madagascar
At least ten people were killed by Tropical Cyclone Giovanna in Madagascar, which hit the island nation as a powerful Category 3 storm with 125 - 130 mph winds at 22 UTC Monday night. An estimated 600,000 people lived in areas that received hurricane-force winds, but the eyewall of the storm missed the capital of Antananarivo, which received peak winds of 38 mph, gusting to 55 mph. Many remote areas that were affected by the storm have not been heard from yet, so the full extent of Giovanna's damage is not yet known. Giovanna is currently in the Mozambique Channel between Madagascar and Mozambique as a tropical storm with 50 mph winds, and is slowly intensifying. Latest computer model forecasts from the GFS and ECMWF models conflict, with the ECMWF model predicting the storm will swing around and pass very close to the southern tip of Madagascar next week, and the GFS model predicting landfall in Mozambique this weekend. Meanwhile, Madagascar must also keep an eye on Tropical Cyclone Thirteen, which is gathering strength over the waters to the east of the island, and is on a course that will bring it close to Madagascar next week.

Jeff Masters
Categories: Hurricane
  Permalink | A A A
Reader Comments
Display: 0, 50, 100, 200 Sort: Newest First - Order Posted
Viewing: 251 - 264

Page: 1 | 2 | 3 | 4 | 5 | 6 — Blog Index

251. uncwhurricane85 2:26 PM GMT on February 16, 2012    
Quoting RitaEvac:


i havent looked at normal SST for this time of year, but that seems abnormally warm, specially the gulf and just off the southeast coast near the GS!!??
Member Since: August 4, 2010 Posts: 0 Comments: 559
252. nfloridandr 2:29 PM GMT on February 16, 2012    
excerpt from nws tallahassee

On Saturday, however, the ingredients appear to be coming together
for possibly the most widespread and significant Severe Weather Event
of the season thus far for Southeast Alabama, Southwest and South
Central Georgia, and the Florida Big Bend and Panhandle! With 0-6km
Bulk Shear Values quite possibly exceeding 80 kts, 0-3km Bulk Shears
of 50 to 60 kts, along with an 850 mb jet in excess of 50 kts,
combined with vigorous 0-1 km Helicities between 300-500 m^2/s^2,
very strong diffluent flow aloft, a rapidly deepening Sfc Low which
the models keep trending further and further to our West, Surface
Dewpoints rising into the mid 60s which could result in ML Capes
nearing 500 J/KG. This is expected to set the stage for the
development of a Large MCS with the distinct possibility of
Vigorous forerunning Supercells capable of producing Significant
Dangerous Tornadoes, as well as possible widespread straight line
wind damage. In summary, the above mentioned severe weather
parameters are very high in this instance, making this a potential
event of quite rare proportions.

It should be noted, however, that this event is still a few days
away, so the timing intensity, and overall threat level could change
before Saturday, but ALL interests in our CWA should keep abreast of
the latest forecasts from the National Weather Service in
Tallahassee, FL and the Storm Prediction Center in Norman, OK, and
be prepared to take quick and decisive action if needed as this
could be a very rare event for our region.
Member Since: June 1, 2007 Posts: 0 Comments: 70
253. bluheelrtx 2:36 PM GMT on February 16, 2012    
Quoting Skyepony:


If you don't trust PWSs choose your closest airport or NWS...if you don't trust that data why bother?


I was just about to suggest that. There are a number of lousy weather stations in my area, so I set my preferred stations to airports only. That was before the Best Forecast thing came out, but when it did, the accuracy was much better. I toggled back and forth for a couple of weeks, then just left it on Best.
Member Since: November 8, 2006 Posts: 5 Comments: 220
254. Tropicsweatherpr 2:49 PM GMT on February 16, 2012    
The TCHP is warming in the Western Caribbean.

Member Since: April 29, 2009 Posts: 64 Comments: 8160
255. LargoFl 2:59 PM GMT on February 16, 2012    
Quoting nfloridandr:
excerpt from nws tallahassee

On Saturday, however, the ingredients appear to be coming together
for possibly the most widespread and significant Severe Weather Event
of the season thus far for Southeast Alabama, Southwest and South
Central Georgia, and the Florida Big Bend and Panhandle! With 0-6km
Bulk Shear Values quite possibly exceeding 80 kts, 0-3km Bulk Shears
of 50 to 60 kts, along with an 850 mb jet in excess of 50 kts,
combined with vigorous 0-1 km Helicities between 300-500 m^2/s^2,
very strong diffluent flow aloft, a rapidly deepening Sfc Low which
the models keep trending further and further to our West, Surface
Dewpoints rising into the mid 60s which could result in ML Capes
nearing 500 J/KG. This is expected to set the stage for the
development of a Large MCS with the distinct possibility of
Vigorous forerunning Supercells capable of producing Significant
Dangerous Tornadoes, as well as possible widespread straight line
wind damage. In summary, the above mentioned severe weather
parameters are very high in this instance, making this a potential
event of quite rare proportions.

It should be noted, however, that this event is still a few days
away, so the timing intensity, and overall threat level could change
before Saturday, but ALL interests in our CWA should keep abreast of
the latest forecasts from the National Weather Service in
Tallahassee, FL and the Storm Prediction Center in Norman, OK, and
be prepared to take quick and decisive action if needed as this
could be a very rare event for our region.
ty for posting this, a very real dangerous event is beginning to form, as the models are slowing coming together, people need to start paying attention to whats forming in the days to come,even here in the tampa bay area the local weather guy said to watch for potential severe storms sat night and sunday morning, i hope That changes for us, but we can use any rain it can give us. had I-75 closed yesterday because of brush fires, we are very dry
Member Since: August 6, 2011 Posts: 0 Comments: 22426
256. GeorgiaStormz 3:06 PM GMT on February 16, 2012    
With my usual bad luck, this is all south of my area in GA but from the SPC Day 3 Convective Outlook:

People really need to watch this and see what happens, it could be interesting. If discrete cells form ahead of the main line, we could have some storms similar to the April 14-16,2011 Tornado Outbreak.
Member Since: February 11, 2012 Posts: 0 Comments: 7162
257. RitaEvac 3:10 PM GMT on February 16, 2012    
Heavy rain event to take place near TX and LA coasts and offshore. Models pegging at some 4 inches somewhere. Pacific flow is in full throttle

Member Since: July 14, 2008 Posts: 1 Comments: 8899
258. nfloridandr 3:14 PM GMT on February 16, 2012    
last time there were discussions about a "rare" severe weather event we had tornado pass very near our home. it was so unusual you could hear the jet (roar) overhead for over an hour before the storms finally rolled in.
Member Since: June 1, 2007 Posts: 0 Comments: 70
259. islander101010 3:17 PM GMT on February 16, 2012    
Quoting HadesGodWyvern:


invest west of Peru... hmm

LOW PRESSURE CENTER 1012 MB AT 09S 87W.

SCATTERED MODERATE CONVECTION WITHIN 180 NM E QUADRANT OF CENTER
good surf for costa rica coming
Member Since: September 11, 2010 Posts: 0 Comments: 2967
260. sunlinepr 3:18 PM GMT on February 16, 2012    
Member Since: August 2, 2010 Posts: 19 Comments: 8445
261. sunlinepr 3:27 PM GMT on February 16, 2012    
Areas with 10,000 Bq/m² of radioactivity considered "highly contaminated" by study — Over 30,000 square kilometers in Japan exceed level, 8% of nation

Published: February 16th, 2012 at 8:34 am ET By ENENews

Japan’s science ministry says 8 per cent of the country’s surface area has been contaminated by radiation from the crippled Fukushima nuclear plant.
It says more than 30,000 square kilometres of the country has been blanketed by radioactive caesium.
The ministry says most of the contamination was caused by four large plumes of radiation spewed out by the Fukushima nuclear plant in the first two weeks after meltdowns.
The government says some of the radioactive material fell with rain and snow, leaving the affected areas with accumulations of more than 10,000 becquerels of caesium per square metre. [...]
Member Since: August 2, 2010 Posts: 19 Comments: 8445
262. sunlinepr 3:28 PM GMT on February 16, 2012    
Member Since: August 2, 2010 Posts: 19 Comments: 8445
263. SPLbeater 4:43 PM GMT on February 16, 2012    
Quoting GeorgiaStormz:
With my usual bad luck, this is all south of my area in GA but from the SPC Day 3 Convective Outlook:

People really need to watch this and see what happens, it could be interesting. If discrete cells form ahead of the main line, we could have some storms similar to the April 14-16,2011 Tornado Outbreak.


i am just inside the 15% line up in the Piedmont of NC
Member Since: August 4, 2011 Posts: 43 Comments: 3953

Viewing: 251 - 264

Page: 1 | 2 | 3 | 4 | 5 | 6 — Blog Index

New Comment
Community Standards Policy Comments will take a few seconds to appear.
Post Your Comments
Please sign in to post comments.
Not only will you be able to leave comments on this blog, but you'll also have the ability to upload and share your photos in our Wunder Photos section.
About JeffMasters
Jeff co-founded the Weather Underground in 1995 while working on his Ph.D. He flew with the NOAA Hurricane Hunters from 1986-1990.

Local Weather
Partly Cloudy
66 °F
Partly Cloudy
Community Activity