The science of weather forecasting
Today蘑菇视频 four-day rain outlook is as accurate as the one day forecast 30 years ago. The remnants of tropical storm Gordon were forecast to bring heavy rain to Southwestern Pennsylvania and the forecast was spot on with 4 to 7 inches of rain falling across a wide area and causing widespread flooding.
Rains of this magnitude are very rare across our area and the ample lead time gave folks a chance to prepare. We were a bit lucky as the rain was spread out over thre full days, giving streams a chance to carry off some of the water. Had this much rain fallen in 24 to 36 hours, the results would have been much worse.
As I write this article, forecasters are busy making their predictions as to where Hurricane Florence will strike. At this point, it looks like North Carolina is in the direct path. Over the past several years, just about every hurricane forecast has been accurate to the point of predicting 2 or 3 days in advance the actual land fall, wind speeds, storm surge and rainfall amounts. What hasn鈥檛 been accurate is our governments response to these storms. One only has to think back to Hurricane Katrina in Louisiana and Maria last year in Puerto Rico. In both cases, many thousands died and the response, particularly in Puerto Rico, was shameful.
So much has changed in the forecasting of these storms. In 1900, The Great Hurricane of Galveston, Texas, came ashore almost unannounced and killed thousands and completely destroyed the island. Even with our super computers and the forecast models they are able to produce, there is still the human element needed. National Weather statistics show that the human factor improves today蘑菇视频 forecasts by 25 percent over just computer forecasting.
As the computers get better, I am sure the computer models will continue to do better. Will we be able to give you a 100 percent accurate forecast for your daughter蘑菇视频 wedding on the 19th of November? Forecasts are getting better, but we are not there yet.
Twenty years ago, National Weather service radar failed to detect a freakish rainfall storm that developed over Fort Collins, Colorado. On the night of July 28, 1996, residents were lulled to sleep by the sound or light rain drumming on their rooftops with no worry of what was to come. By midnight, two inches of rain had fallen over the region, but an area just Southwest of Fort Collins received 14 inches of rain, causing a flash flood on normally docile Spring Creek that, killing five people. The rains swamped homes and sent families scrambling up trees and rooftops in their pajamas.
Looking at the storm afterward, the national Weather Service said the radar simply did not pick up the isolated fierce rains. Nobody even reported the heavy rain from backyard rain gauges. This incident led to the establishment of some 30,000 volunteers who have built a database of weather watchers throughout the country. I have been part of this network for over 40 years, and this past week made morning calls to the weather service on the 7.02 inches of rain that fell in Chalk Hill.
A Texas volunteer not long ago reported over 7 inches of rain in his gauge while nearby gauges had no reports or just a small amount. The weather service people thought the 7 inches was an error until they looped the radar to find a tiny convective storm that formed and died in the same spot producing the rain.
Southwestern Pennsylvania had a few of these types of isolated convective storms earlier this summer that brought some local flash flooding.
Computers are a big help but the network of volunteer observers are still vital to a good forecast.