Archive for July, 2008|Monthly archive page

Amazon.com outage left our map void

For an extended period between Saturday and Sunday, some of you may have noticed that the landmass images for our tracking map were failing to appear.

We traced the problem back to Amazon.com, whose Simple Storage Service (S3) experienced an outage during the same period. This failure on Amazon’s part meant that we could not deliver map images to anyone visiting our site or viewing the map on PalmBeachPost.com. This was particularly ill-timed given the need to track Tropical Storm Dolly and Tropical Storm Cristobal.

The problem was resolved by Amazon engineers at 8:12pm (Eastern) Sunday evening. We at Stormpulse.com have also taken steps so that in the event of another Amazon failure, we will still be able to deliver these images.

Sorry for the inconvenience, and thanks for using Stormpulse.com.

UPDATE: Since writing this blog post, we have implemented a failover system such that future S3 outages will not impact our ability to show our basic map tiles.  Thanks again for your patience.

Advertisements

PalmBeachPost.com embeds our hurricane-tracking map

The Palm Beach Post has begun embedding Stormpulse’s tracking map on their own hurricane site, Storm2008.

We are very happy to be able to take a step forward in getting our tracking map in front of a larger audience, and the Post is happy to receive a major upgrade to their storm mapping displays. This enhancement also coincides with what is turning out to be an active July, with Tropical Depression Three forming as I write this.

The Palm Beach Post is the first, but hopefully not the last, place where our embeddable map should surface. Anyone interested in embedding the map on their own site, using a simple JavaScript tag, is welcome to reach out to us via this blog or our email address (stormpulse at gmail dot com). We look forward to hearing from you.

Hurricane Bertha highlights challenge in forecasting

In the National Hurricane Center’s Discussion #19 on now-Hurricane Bertha the Forecaster Rhome states:

“BERTHA’S INTENSIFICATION TO MAJOR HURRICANE TODAY HIGHLIGHTS THE DIFFICULTIES OF FORECASTING RAPID INTENSITY CHANGES.”

I wonder if this difficulty will ever go away through sheer computing power, or if something a bit more human is required. Could human-consensus forecasting address this weakness? Still unproven, but it seems worth a shot.

Tropical Storm Bertha stubbornly heads west

Tropical Storm Bertha, considered yesterday an almost certain fish-spinner, continues to head west, and her turn to the open Atlantic appears a bit delayed and a bit less severe than originally forecast.

00pm EST

Latest NHC forecast path for TS Bertha, 7.4.08, 11:00pm EST

Still, sea surface temperatures remain too cool for rapid intensification (mid-Atlantic buoy 41041 reports 79.52 degrees F), but she should hit warmer waters by Wednesday (80-82F). By that time, if she follows the official NHC forecast path, she will be approximately 1300 miles east-southeast of Miami. She is presently 2395 miles east-southeast of San Juan, Puerto Rico.

45pm EST

Water vapor image for TS Bertha, 7.4.08, 11:45pm EST