Monday, February 16, 2015

Weather Monday 2/16/15 Weather Forecasting Part 3: Other methods of forecasting


Wintertime Blues is the name of this quilt. That is what we have in the eastern USA. We are ready for summer, but the forecast looks like more bitter cold and snow. Yikes!

This time I want to talk about other methods of forecasting. Some are easier to do than others. Some of more accurate than others.




Persistence Forecast

This forecast says what it is doing now it will be doing in the future. This would be like what you see in the table below. Boy I wish that were our forecast

Today Tomorrow The day after tomorrow
image image image

This is certainly and easy method, but is only good for very short range forecasts. Weather can change very quickly so I would not put any money bets on a persistence forecast. But it is so easy even my cat could make this kind of forecast!

Guarding the Catnip (or forecasting the weather) Maybe he can do better than the groundhog!








Steady State or Trend

This method is a bit more sophisticated than using persistence and does require knowledge of meteorology.This method assumes that weather systems such as a cold front will continue to move at the speed that have been moving at and in the same direction. This map from ww2010 shows a cold front moving a distance of 800 miles in 24 hours. We would assume that it will continue to move at the speed and in that direction so we know where it will be 24 hours later. Again this is only good for short range forecasts since systems can change speed and direction.



Analogue Forecast

This method relies on a forecaster recognizing the current weather situation as something he/she has seen before. So a weather person would have to have a good knowledge of past weather scenarios. So you would think that a forecaster with lots of years of experience would do better with this method. The only trouble is seeing a similar situation in the past does not mean it will turn out exactly like that this time. In fact you can pretty much guarantee that it will not turn out the same. I guess this method is like a quilter working in a series. Each quilt will not be exactly like the one before. This method is actually better for longer range forecasts where you are just looking for trends.


The weather forecast for this winter made back in the early fall. Looks like it is shaping up to be just like they said it would be.






Ensemble Forecast

There are several different computer models that are used to forecast the weather. All of these models are run and then compared. If they forecast a similar outcome then there is pretty good confidence in the forecast. If the models all predict something different then there is a high degree of uncertainly. So it is less clear if you will get snow or freezing rain. Or if you will get 6 inches or 12 inches of snow.

Climatological Forecast

This type of forecast is based on the average weather (climate) of a particular location. For example, July and August are nearly rain free in Los Angeles. It rains only 1 day in 90 in this area in the summer. That is only a 1% chance of rain. So if I made a forecast that it was going to be sunny this year on the 4th of July I would have a chance of 99% of being right. This would not be that easy to make in other locations.

Probability Forecast

What is the chance that a particular event will happen. This is given a percentage. The higher the % the more likely it will happen. A 90% chance of snow is certainly more likely than a 20% chance.

How do forecasters stack up against persistence and climatological forecasts?



The forecasters using their models and skills do much better early on.  After 9 days you may as well just go with the what you would expect on that particular day (the climate). Looks like persistence is the worst method for forecasting.




Thanks for reading. Not sure what the topic will be next Monday.



Susan said...

This is very timely info for me! I live in NC so I've been watching a lot of weather forecasts and it looks like I will be for another day or 2. :)

The Idaho Beauty said...

Once again, I am enjoying your lovely quilts - you really are a talent. As for the weather info, I'm not sure I've lived anywhere that I was so aware of the consistent speed with which storms moved through as when I lived in Wisconsin. At one point I lived about an hour and a half east of the twin cities and listened to a radio station in Minneapolis. So I'd get advance warning of a thunderstorm or high winds or tornadoes or whatever as the station reported it happening there, soon learning how long it almost always would take show up on my doorstep - pretty much like clockwork!.

As for the percentage forecast - that brings back grade school memories. Forecasting was always this mystical magical thing before we studied it as part of our science block. I remember how fun it was to go outside, take all the readings, work out the percentages to come up with our forecasts which then usually came true. That whole odd thing about "chance of rain" suddenly made sense. Didn't think any less of the weathermen, just had seen behind the curtain and what was hiding there was pretty cool.