Over the past several weeks, I have gotten many emails, tweets, comments, etc. that have asked me for my thoughts on what the upcoming summer is going to be like. I usually try to avoid such prognostication, but since this blog affords me the ability to talk for a while, say it once, and just keep on refering to it later... I might as well give this a whirl :-)
Why I don't like seasonal prediction... let me count the ways... :-)
1. It's difficult (just slightly better than random) and can make someone seem like they don't know what they're talking about really quickly.
I have a tough enough time with the next seven days, let alone trying to predict something 100 days into the future. Ironically enough, climate prediction (trying to ascertain the potential shift in globally averaged temperature and precipitation anomalies) is easier than seasonal prediction-- and that 20 years and more into the future. The reason is that the looong range stuff is averaged out over the whole Earth (ie., you don't actually have to be accurate anywhere specific, so long as they all balance out ;-)...The tighter the time frame, the more location-specific you have to be, and the closer that gets to simply predicting the weather (which is hard).
Here's a good (and recent) case in point: The UK Met Office, one of the most renowned weather institutions in the world, and owners of the most accurate weather prediction model in existence (the 'Euro' or 'ECMWF'), recently said this in a March 23rd forecast of April weather:
"The forecast for average UK rainfall slightly favours drier-than-average
conditions for April-May-June as a whole, and also slightly favours April being the driest of the 3 months.
With this forecast, the water resources situation in southern, eastern and central England is likely to deteriorate further during the April-May-June
Well, because I'm including this here, you can imagine what all happened in Britain during the month of April. How about the wettest April in history...
Interestingly enough, sometimes these sorts of predictions never get called on by the general viewer...because their attention is distracted again by another future projection (like the coming hurricane season, or the winter snowfall forecast). I guess I could just hope y'all do the same thing, but then that would mean I'd have to provide those forecasts too ;-) Anyway, it's with this kind of experience that I tepidly approach this kind of forecast.
2. A seasonal forecast may not be all that useful when you get down to it.
Suppose I gave you a 100% accurate forecast for a "colder" and "drier" summer. What does that mean, actually? Well, that means that at the end of the summer we'd look back and see if the sum total of moisture and the average of temperature is where they were predicted to be compared to normal. What it doesn't do, is pick out individual events to highlight, or identify any potential extreme events that may lie down the road. It's possible to hit record lows in an overall warm summer. It is also 'possible' to get an entire summer's worth of rain in a single week, cause massive flooding, and still register below average for the whole 3-month period. In truth, seasonal forecasts are only really useful for seasonal purposes-- people who manage energy networks or agricultural interests; that sort of thing. Those folks are trained in appropriately managing the information issued (ie, not the type of person who takes a single heavy rain event and says "but you said it was going to be dry this summer" :-) The UK forecast above was technically for a three month period-- they just happened to unfortunately highlight April. Perhaps they may still yet be right about the three months as a whole (but I doubt it).
To further illustrate this point, here's the National Weather Service's forecast for "June/July/August" for our area:
...now if you're wondering if a forecast like this is of any real use-- you're not alone! :-)
The making of a seasonal forecast
The elements to a forecast out the next 7-10 days are completely different than that of a seasonal forecast. If you think about it, meteorologists would be looking at the weather upstream-- what's going on in the Central Plains, West Coast, etc. to get an idea of our weather. There are known values on the map that contribute the end result. When forecasting for a season itself (a specific day months away is impossible in my opinion), you'll be looking at larger scale contexts and teleconnections, rather than the weather 'upstream'. All of the individual weather events that will happen in the summer aren't even an twinkle in the eye of the Earth-- meaning you can't just stare at someplace on the other side of the planet and expect that weather to hit us in July.
There are two of the things I look at when thinking 'season':
1. Current trends
2. Air/Ocean circulations (like El Nino)
Here's how we've done in Huntington in terms of temperature since last year:
Now, at first glance, it's instantly noticeable how crazy March was in the context of 'normal'. April may have provided a little balance, but on the whole, the numbers are moving back toward the same numbers we saw last May. Last summer we saw the most 90-degree days in years, despite it having the appearance of being a cooler than average summer. Since we're already heading back up to where we left off compared to last year, if the pace continues as it currently is we'll be warmer than normal this summer.
For precipitation, it looks like this:
|NRCC - April Precip Percentage||NRCC - 3-Month Precip Percentage||NRCC - 6-month Precip Percentage|
As you can see, precipitation is a difficult animal. Some places have been flooded with rain recently (and you can imagine what May will look like in southern Ohio after the several inches of rain they've received this past week alone). Indeed, the trend had been pointing toward a lower-than-average rainfall, but recent storms are turning that tide around. Since January 1st, the Huntington area has received about 12" of liquid equivalent precipitation (meaning melting the snow down too to add it to the rain totals)... This is about 2" below the average, and much less than the 22" we had received to date last year.
Air / Ocean Circulations
This is more into the 'tea-leaves' kind of thing-- the deep end of the pool that I'd prefer floaties for :-) In the last 30 years, climate scientists and meteorologists have identified many circulations in both air and ocean that are long-term in nature, and contribute to the overall weather pattern. The famed "El Nino / La Nina" pattern is perhaps the most understood (though still not easily predictable). Others include the Atlantic Multidecadal Oscillation, the Madden-Julian Oscillation, the Pacific Decadal Oscillation, etc. The two that I concentrate on are the NAO and the ENSO (which stand for the "North Atlantic Oscillation" and the "El Nino Southern Oscillation")
North Atlantic Oscillation
This invovles an atmospheric cycle between a prevailing area of low pressure near Iceland and a prevailing area of high pressure near the Azores. The neat thing about it is that it's not driven by the ocean, but the bad thing is that it's not driven by the ocean (hence a more volatile changer and more of a 'manifestation' of current climate than an easy predictor of a future one). Anyway, here's how this looks. The NAO has a 'positive phase' and a 'negative phase' (climatologists like to boil everything down to a simple 'index' number. So by assessing the relative strengths and positions of these two semi-permanent atmospheric features, climatologists can get data that is useful.
So you can see the different phases of the oscillation and what it has come to look like on the ground. Here's another way of looking at how the NAO index is correlated to the weather we experience locally:
|NAO Index - Temperature Correlation||NAO Index - Precipitation Correlation|
Here's how the above maps work (click for a larger image)... If the correlation is positive, then whenever the index is up, the temperatures (or precipitation) will be up. If the correlation is negative, than when the index is up, the temperatures (or precipitation) will be down. For our purposes, we can tell a lot more from the NAO in terms of temperature, but precipitation is okay as well (if you're looking at a multi-month generality).
This is what the NAO index is currently at, and also where it is forecast to be:
Fluctuations inside 1 / -1 are not considered that dominant of an impression, but it is interesting to note that though negative, the NAO is forecast (in ensemble averages) to get back to zero or even slightly positive). This would imply a wettening of the climate picture, though given the breakdowns as it particularly relates to summer there does not appear to be much correlatable skill here-- think about it: Most of the heavy moisture makers that great the Atlantic states are either the Nor'easters (winter) or hurricanes (fall). So even though on the surface it doesn't look like we'd learn much locally from the NAO this go around, it can tell us things anyway because it can influence the weather that eventually gets here.
El Nino - Southern Oscillation
This is the more famous one, although even farther away from the tri-state area. This does, however, concern the largest ocean on the planet and strongly influences the 'average weather experience' for much of the globe. As you may know, the Southern Oscillation is the cyclical sea surface temperature anomalies in the southern Pacific ocean. Whenever the temperatures are colder than normal, we call it a "La Nina" time, while the warmer than normal is the "El Nino". Now, of course, that's averaged out over the whole planet. Here's what it looks like for the US:
|El Nino - Temperature Impacts||La Nina - Temperature Impacts|
|El Nino - Precipitation Impacts||La Nina - Precipitation Impacts|
So, of note: (1) La Nina events bring warmer-than-average temperatures typically to the area, even though for the world it's a cooler-than-average picture; (2) El Nino events are typically drier for our area, while La Nina are wetter.
This is the current state of affairs as it relates to ENSO:
|Southern Pacific Sea Surface Temperatures||Southern Pacific Sea Surface Temperatures|
So we're currently in the last stages (as far as we know) of a La Nina period (one where we saw warmer than normal temperatures and near-normal precipitation thanks to the recent storms). The forecast is for us to continue to progress out of La Nina and onward to an El Nino summer. Now, there are two ensemble members that want us to tumble back into a La Nina, but most indidcations are for El Nino.
---Tacking A Breath---
Okay, back. :-)
What is "Normal"
I thought it was important to add this bit-- and a hat tip to the commenters for making sure I got this in here. A caveat first: "Normal" means precious little as far as determining the weather experience day-to-day. If we have 3 days that are 10-degrees above normal, followed by 3 days that are 10-degrees below normal, the 'average' would come out to 'normal', but yet not really reflect the experience. However, when talking about a 'seasonal' forecast, we're not after the weather experience per se, but rather the longer-term signal that rises out of that noise.
||Record Low||Normal Precip|
Putting it all together
Typically there's one more thing to look at, which is called "analogous years". This would be where we identify years past that has progressed similarly to the present one, and we examine how those years did in reference to the forecast period. It's another ingredient to add to the pot, but it's not as necessary here. Typically it's more useful for winter weather forecasting, because just saying that a season is going to be "wetter" than normal means nothing as far as how much of that water is actually going to be snow flakes.
So going with what we have, I'm going to say that this summer is going to be Warmer than normal, and Drier as well. It's not too difficult to be drier than last year though, but I'm talking drier than normal ;-). Predicting seasonal-scale weather does not afford us the ability to declare something like "We'll have 4 events with greater than 3" of rain", or "Our highest temperature for the summer will be 102°", but those who are interested in seasonal weather prediction for their business would already know about that.
I'm going to refer to this post in future climate / seasonal forecasts, should I decide to do this ever again ;-) ... This will keep me from having to type certain things out over and over again. Likewise, I'll try to keep coming back to this post when I can, so if you have any questions on this subject, I'll be happy to answer them here as well.
...So there you have it :-)
Thank you for bearing with me...