top border

Please note, not all links may be active. This site is a snapshot of an earlier time.

Model Muddle
Temperatures to drop sharply, then some uncertainty

Steve Scolnik @ 4:20 PM


Mostly sunny, mild. After a brief step backward, the seasonal calendar is about to take a jump forward to more wintry temperatures. What happens thereafter is subject to a bit of debate among the models, but the storm development noted in Dan's earlier post has disappeared from the run of the same model made 12 hours later.

By mid afternoon today, however, readings were well into the 60s throughout the area, with Dulles at 68° and National 66°. Salisbury, Culpeper, Leesburg, Manassas and Patuxent River all reached at least 70°. The air is also quite muggy with dewpoints generally in the mid to upper 50s.

Tonight and Tomorrow

Mostly cloudy, showers likely, then turning colder. Skies will become mostly cloudy overnight with showers likely and thunderstorms possible, mainly after midnight and through tomorrow morning. Lows will be 48-53°. Clouds will decrease tomorrow afternoon along with increasingly brisk and gusty northwesterly winds. After morning highs in the low 50s, temperatures will drop during the day to the low 40s by evening.

For the outlook through the rest of the week and into the weekend, scroll on down to Dan's post below.

Climate Corner: New MIT Model Results

Prof. Peter Stone of the MIT Center for Global Change Science gave a very interesting presentation last night on the science behind climate change modeling. With the current model inconsistencies over what may or may not develop as soon as early next week, this is an especially relevant time to review the difference between weather and climate modeling. PM Update asked Dr. Stone how he would explain this, and his answer was very similar to what has been said here in the past, but it bears repeating:

Even though climate is the long-term average of the weather, and even though the basic laws of physics apply equally to both problems, climate modeling is not simply taking a weather prediction model and running it for a much longer time. It is much easier, and more accurate, to predict the average conditions produced from the long-term equilibrium of a set of forces than to predict the specific details of the evolution from a beginning state to some point in the future. For the more technically inclined, this is the difference between an initial-value problem (hard) and a boundary-value problem (not so much). Weather prediction is an initial-value problem; it starts from a given set of conditions and predicts the details of their change over time. Climate prediction is a boundary-value problem; it applies a set of forces to a beginning state and predicts the final equilibrium result.

In the course of his lecture, Prof. Stone disclosed some tentative results, now in peer review for future publication, regarding likely future temperature increases. Based on a "business as usual" scenario (no reduction in CO2 emissions), his research indicates that the median prediction of the mean global temperature increase by the end of this century is 4.9°C, with a 90% confidence interval of 3.0-7.7°C. This is sharply higher than the results presented in the latest IPCC report of a range of 2.0°- 4.5°C for a doubling of CO2. To put this into perspective, the predicted increase produces a higher mean temperature than has occurred in the last 3 million years.

The slides from Prof. Stone's lecture are available on the MIT Club Seminar Series website.

Comments are closed for this archived entry | Link | email post Email this post