January 20, 2011

A Great Irony, A Little Rain, And Then Dry

I believe both of these statements are true:

(1) Weather forecasting is now much more skillful than ten years ago.
(2) There has never been as big a gap between what meteorologists know and the information provided to the public.

Kind of ironic, right? Today, with huge amounts of data coming in from weather satellites and new observing systems, massive computer modeling systems that provide high-resolution forecasts, improved analysis and display systems, and other technical advances, weather forecasts has become hugely more skillful. I can easily prove this with a wide variety of verification statistics (here is an example from the NWS).
Technical improvements are giving my profession increasing capability to predict the spatial variation of weather and its timing. The amount of information is daunting, butunfortunately the means of communicating the information has not kept up.

How certain are we of the forecast? How will weather vary locally? What is happening right now and during the next few hours? None of this is being effectively provided to the public. Forecasts are often excessively vague or "broad brushed" as it is known in the field. Clearly failing forecasts are not updated in a timely way.

Want an example? How about today? Here is the forecast provided when you woke up this morning:
EVERETT AND VICINITY-SEATTLE/BREMERTON AREA-TACOMA AREA-
INCLUDING THE CITIES OF...EVERETT...EDMONDS...LYNNWOOD...
MARYSVILLE...ARLINGTON...SEATTLE...BREMERTON...SILVERDALE...
FEDERAL WAY...KENT...TACOMA...LAKEWOOD...PUYALLUP...SUMNER
400 AM PST THU JAN 20 2011

.TODAY...CLOUDY WITH A CHANCE OF RAIN THIS MORNING...THEN RAIN
LIKELY THIS AFTERNOON. HIGHS IN THE LOWER TO MID 40S. SOUTH WIND
10 TO 20 MPH.
This is the official National Weather Service zone forecast for the Puget Sound region. Now what would you conclude from this regarding the chances of rainfall vary around the region? Well, there is little information about it. But what did my profession know?

Here are rainfall totals from a computer model run the night before for three 3-h periods (7 AM-10 PM, 10-1 PM, 1 PM-4 PM)





What would you conclude from this model guidance? Perhaps a few light showers with Puget Sound generally dry (due to rainshadowing in the lee of the Olympics).

What happened? As seen by the radar (see examples below) there was only some light sprinkles and the rainshadowing was very evident.

There was good information that the rain would be greater over the eastern Sound than the western Sound, but that message was not being delivered.

With a 6-h forecast cycle the weather predictions are often not updated when the situation changes. However, when something critical is occurring the NWS can and often does release special statements.

Sometimes forecasters are sure about a forecast and sometimes they realize there is a lot of uncertainty. Most of the time this information is not readily available. For the public the only was to get an inkling of the forecaster confidence is in the Forecast Discussion, something that is available on the NWS web site (and on my department weather pages). A portion of the discussion from this morning is:
This issue of communication of forecasts is something I will come back to.

Tomorrow we will see some rain over the region, enough to
bring some of the rivers up again, but then
something unexpected will occur: an extended DRY spell
for most (the southern two-thirds) of the state for several days.
Above average temperatures. That should end the threat of
flooding for a while. More on that in my next blog.

Keep in mind we have only one more month of winter left.
Spring starts on March 21st?
Not here west of the Cascades!
By the time you get past Feb 25th the threat of anything
serious is gone--rare to get a major windstorm, snowstorm,
or flood after that date. And by that time the bulbs are
pushing up, the cherry trees are in blossom, and
the garden centers are full of fertilizer and seeds.

15 comments:

  1. I am planning and outdoors trip for this weekend and I am trying to understand cloud coverage better. Skiing and I am not a big fan of white out skiing. How about a quick blog post on reading cloud cover data for the outdoor enthusiast. Thanks. Excellent blog.

    ReplyDelete
  2. On another facet of the last post, can you give us some indication of how much better your 1.3km runs do for rainfall and winds than the 4km runs?

    I often find very different wind directions and somewhat different velocities on the small models that I run. I don't do much with precip, relying on NWS to do that part. If the 1.3km output data could be made available somehow, I'm sure someone else could run plotting routines on a free amazon cloud account and publish the plots. The plots don't take nearly the resources the model does.

    TJ.Olney

    ReplyDelete
  3. Cliff,

    I definitely agree with your point, especially with regards to "nowcasts", and look forward to advances in this area.

    A particular pet peeve of mine is moderate wind storms that never make it into a forecast. It seems that a few times a year we get hit with 30 or 40 knot winds when the forecast is much lower, 15 knots or less... all while I am out in a very wind averse boat that could make it back to the dock if we had an hour of warning.

    If I am paying careful attention and am worried about the possibility, sometimes I can suspect that it may happen based on the weather information available combined with my fairly weak meteorological skills, other times I am completely surprised.

    Even having a 1 or 2 hour warning that there is a 50% chance of such an event would be invaluable.

    ReplyDelete
  4. Interesting post! Looking forward to more on the topic of communicating forecasts. A few years ago I discovered the Forecast Discussion and even though there is much I don't understand, I did realize it provided much more detailed information than the local media. Now I read it almost every morning! Sometimes the forecasters comments/speculations are a bit of poetry.

    ReplyDelete
  5. Cliff -- where is this la nina everyone talked about? I remember a lot more snow days in years past, and this year seems to be a bust for us, but is hammering the rest of the country.

    ReplyDelete
  6. The big question I see is "how do forecasters better communicate this knowledge?"

    The part that I find most frustrating is forecasters don't have an effective method of communicating uncertainty (I usually get a feel for this by reading the discussion). The forecast most of the public sees is "52 degrees and cloudy" even when it's 14 days out where we have little to no idea what the weather will be.

    Ultimately, this leads people to think forecasting skill is lower than it is because the long range forecasts tend to be broad brush and often are wrong.

    I'd love to hear some of your thoughts on how forecasters could better communicate uncertainty (or certainty) in their forecasts.

    ReplyDelete
  7. I have a couple of thoughts on this: National Weather forecasters are frequently moved around the country and tend not to be very familiar with local conditions such as the rain shadow affect and the convergence zone. There an abundance of information but not the depth of experience you get from living in a location for many years.

    The best climatologists go into research and teaching. The late Dr. Reed is a great example of this. We used to have a forecasting contest every spring at the UW Dept. of Atmospheric Sciences. Invariably Dr. Reed won this contest, handily beating the National Weather Service forecasts on an almost daily basis.

    ReplyDelete
  8. People have unrealistic expectations regarding La Niña. A few points (I used to work in a climate research group, abeit studying past climate records not forecasting)...

    1) As Cliff has said before, La Niña is correlated with somewhat cooler temperatures and higher rainfall. That doesn't mean snowstorms, usually it just means more rain and slightly colder.

    2) Also, the ENSO (El Niño Southern Oscillation - meaning the El Niño and La Niña cycle) correlates much more strongly with SPRING weather up here, not winter weather. To some degree it depends on the when the El Niño or La Niña starts - but look for a more significant impact on spring (unfortunately).

    3) It's only probability. It's more likely that the winter and/or spring will be cooler and wetter during a La Niña, but it's never a 100% sure thing. One of our drier winters in recent memory occurred during a La Niña this past decade (I forget the year, and am too lazy to look it up).

    ReplyDelete
  9. I know you addressed this earlier, but in my 25 years of coaching track and field and standing outdoors from March through May, I have come to the conclusion that "winter" lasts until mid April, sometimes later. The best skiing, with the deepest, most consistent snow, and the consistent gusty winds and rain down low, seem to always happen in that time. In fact, I've always felt like we get a period of fairly clear stable weather with comfortable (for the month) daytime highs in February. It's better weather for working outside than you'll find on many (most?) "spring" days.

    ReplyDelete
  10. Point taken, and more granularity would be welcome, but the "click on map" feature on http://www.wrh.noaa.gov/sew/ is going (I think) in the right direction. Thanks again for this interesting blog.

    ReplyDelete
  11. Hi Cliff,
    Just got a call from a staff member of my state rep regarding the implementatio of the common core math standards. I wrote a letter after reading (and then doing additional research). It has gotten their attention! Keep up all the great work.
    Tom
    PS: Wouldn't it be great if the weather forecasters on TV replaced some of their fancy graphics that don't visualize much with some of the graphics shown in the models. This and the discussion by you and the National weather service is what I go on now.

    ReplyDelete
  12. We here in New England are having a good old fashion winter , cold and snowy. Looking at below 0 and more snow next week.



    http://vacationideas-nickmeg.blogspot.com/2011/01/this-years-new-england-winter-weather.html

    ReplyDelete
  13. I'm curious if certain weather patterns or situations lend themselves to greater variance when comparing forecasts to actual outcome. This year it seems to me many of the forecasts are way off base, missing on occurrence of an event as well as it's intensity. i like the probability forecast which is linked to this site, but it seems to be missing the mark as well. Is this just business as usual, or is there more going on--the type of La Nina, for example?

    ReplyDelete
  14. What a beautiful rain shadow today! For once we were in the rain shadow. Bremerton was high and dry!

    What a bust so far for winter snowfall. A La Nina year, yet BELOW average throughout sections of the Cascades. Skiing at Stevens has not been stellar. I know it's not over, but even the long range forecasts look pretty run-of-the-mill. Nothing exciting.

    Will this La Nina winter end with a whimper? Another great post!

    ReplyDelete
  15. And what is with the lack of storms out of the Northwest that bring 2-3 feet of snow to the mountains? Seems like we've received tons of southwest flow and numerous atmospheric rivers. Flooding has been more pronounced than I would've imagined for La Nina. Seems odd for a La Nina year.

    ReplyDelete

Please make sure your comments are civil. Name calling and personal attacks are not appropriate.

Should I Let AI Write My Blogs?

I spend a lot of time thinking about and writing my blogs.   Could AI apps do the work for me?   Could any of you tell the difference? Well,...