It is time to stop quivering in our boots in pointless fear of the future and just roll up our sleeves and build it.
- Ray Pierrehumbert

Thursday, May 5, 2011

Is there a trend in Atlantic Hurricanes?

Image via Wikipedia ; unattributed.

The denialists are mocking the idea of anthropogenically forced increase in Atlantic hurricanes, never mind tropical storms in general.

You could in fact plot "landfalling hurricanes in the US" and get a much noisier plot wherein there is a "robust decline", the idea that this counts as evidence against anthropogenic forcing being quite disingenuously implied.

But my point again is that (Update: perhaps) the system is getting wobblier. Suppose you take the event "over 15 Atlantic hurricanes named storms" as a matter of interest. These are the darker of the swans. The number of instances is too rare to extract a statistical trend, but that null result does NOT mean that such events aren't more likely than they used to be. And what of "over 25 Atlantic hurricanes named storms". We only have a single event. It's quite recent. But it's been followed by more or less ordinary seasons, and extraordinarily few US landfalling hurricanes (currently we are in a record three year gap since the last, and counting).

Now, just for the sake of argument, imagine that hurricanes pretty much go away for thirty years, and then suddenly come back in greater force than ever, with year after year competing with 2005 for a few decades. Then imagine they drop down to an unprecedented low, with no Atlantic tropical storms for a long while. Projecting this trend onto CO2 would yield nothing.

Yet it is not out of the range of the plausible, because several features of the atmosphere have to align for a strong tropical storm year, and each of them is subject not only to monotonic forcing but to wobbles as the system gets larger and larger transients. In other words, the entire complex behavior, by hypothesis, might be forced. The absence of a trend is not enough to say that it is not anthropogenic climate change.

Admittedly, this gets us in the quandary of what evidence to accept as supportive and what to reject. Certainly, explicitly predicted outcomes provide stronger support for the details of the science than complete surprises. Some people are still stuck in hypothesis testing mode, which I think is starting to get a little bit crazy. We can only test hypotheses that way. We cannot use that approach to establish that changing the radiative properties of the atmosphere is safe.

My claim is that we need to get our thinking out of nudge-world. This is not a "nudge":



(A bit out of date; please mentally extend the green line about to about 394 for today, and probably no less than 500 by the time we are done; perhaps more. I believe the chart is from Hansen's group but I found it here, courtesy of an oil company btw.)

I don't understand why people don't anticipate some ringing in a system that gets kicked this hard.

Update: A real meteorologist speaks on the subject. Less systems theory and philosophical handwaving, more details. Scary enough, but I still see signs of nudge-world thinking.

45 comments:

Layzej said...

You mention that the 0.7C rise in temperature is a kick rather than a nudge. Would it be possible to quantify the amount of energy required to affect that temperature change in standard American units so that an average Joe like me could understand? That is, area should be measured in football fields, data in libraries of congress, and energy (as in this case) should be measured in nuclear bombs.

How many nuclear bombs would it take to raise the temperature of the Earth by .7C? (Assuming that all of the energy was expressed as heat and disregarding the fact that nuclear bombs kick up dust or any other confounding factors)

This would really help someone like me get my head around what is meant by kick or nudge.

steven said...

agreed

Paul Daniel Ash said...

this gets us in the quandary of what evidence to accept as supportive and what to reject

I keep coming back to Hsu's "climate index" idea from the prediction market paper. Hsu suggests a broad basket of indicators:

(i) temperature extremes;
(ii) hurricane and tropical storm intensity;
(iii) hydrological cycle intensity (IOW more intense rainfall AND more droughts);
(iv) sea level rises; and
(v) ocean acidification

Like a stock market index, this seems like it could more accurately reflect the overall picture and dampen the thrash somewhat.

Michael Tobis said...

A couple of things wrong with the question.

First of all, 0.7 C is neither a nudge nor not a nudge. This is the problem; a lot of people are just assuming that roughly speaking we increase everything everywhere by 0.7C and consider how the climate responds. Anything not explainable by that nudge is attributed to ordinary run of the mill bad luck.

No, it is better to look in terms of energy rather than temperature. The current net anthropogenic forcing is about equivalent to a bit less than 1% increase in solar energy hitting the surface, or about 2 watts per square meter; eventually going up to perhaps 8.

It is hard to express this in bombs, because the system is leaky. I suppose you could work it out in bombs per day, but I'm not sure that really helps; it confuses the issue of nuclear winter too.

But we can get an energy count. The numbers in fact work out quite neatly. 2 watts over the surface of 5.1*10^14 m^2 is 10^15 watts. Too big a number to think about, so let's work that out per capita: about 130 kW. I always think in terms of room space heaters, which are typically 1 kW. So, if there were no greenhouse effect, and we wanted to get equivalent warming to what we see now, every person now alive would have to purchase and run 150 space heaters 24 hours a day.

Actual world energy consumption is 1.5*10^13 W. So the current greenhouse effect is equivalent to about 60 times the current urban heat island effect. If everybody multiplied their energy usage 60-fold, it would equal the amount of energy being added to the system by anthropogenic radiative forcing. You can do that with 150 space heaters on the global average. But, if we proportion out this heating on a national basis, Americans would have to need 10.3 Kw * 60 or 600 space heaters apiece.

That's not all. The greenhouse effect is cumulative. Once we get to 4 x CO2, or 8 W/m^2, presuming that the balance of national usage stays constant, you will have to quadruple your efforts or be running something like 2400 space heaters all day every day to match your cumulative greenhouse has score.

Or, if you prefer a more entertaining route, about twelve thousand 52 inch TV screens, all day every day, powered by non-carbon sources.

steven said...

MT.
"Yet it is not out of the range of the plausible, because several features of the atmosphere have to align for a strong tropical storm year, and each of them is subject not only to monotonic forcing but to wobbles as the system gets larger and larger transients. In other words, the entire complex behavior, by hypothesis, might be forced. The absence of a trend is not enough to say that it is not anthropogenic climate change.

Admittedly, this gets us in the quandary of what evidence to accept as supportive and what to reject. Certainly, explicitly predicted outcomes provide stronger support for the details of the science than complete surprises. Some people are still stuck in hypothesis testing mode, which I think is starting to get a little bit crazy. We can only test hypotheses that way. We cannot use that approach to establish that changing the radiative properties of the atmosphere is safe"

I think this is an under appreciated point. As I said over on Lucia's; radiative physics is the best evidence of AGW. extreme events that may or may not occur with increasing frequency or amplitude as a result of us kicking the system are not really the "best" evidence. It's important for our understanding of potential impacts, its important for planning and preparation, but it's not really central to the core of what we know. when some folks treat it as core, failures in hypothesis testing are oddly taken as evidence that somehow the core science is wrong. GHGs warm the planet. It's entirely logically possible for increased forcing to cause fewer hurricanes. In no way could future of fewer hurricanes OVERTURN the basic physics. That's I guess my main point. if we never have another hurricane, that fact would say NOTHING about the core science. GHGs would still warm the planet. It would be an odd result, of course, but we would not throw out radiative physics if hurricanes disappear.

make sense?

Kooiti Masuda said...

Very regrettably, the time series of counts of hurricanes suffers from changes of observing methods. You would not miss hurricanes since 1976 when GOES is available. Even then, the algorithm to determine ranks of storms may have changed. To discuss whether the recent situation is extraordinary against the past records, we need some homogeneity, either by restricting the range of data or by making some adjustments.

Hisayuki Kubota, my collegue at JAMSTEC, has been working on centennial trends in typhoons in the western Pacific. A paper Kutoba and Chan 2009 GRL has been published. It says that there is no discernible trend in the count of tropical cyclone landfall in the Philippines though there are multi-decadal variations. According to Kubota's yet-to-be-published studies, the total count of northwestern Pacific typhoons has not changed so much, as far as we can deduce from records at limited locations. But possibly the tracks of storms have shifted a little.

I do expect (with sadness) that anthropogenic climate change will cause remarkable changes in activities of tropical cyclones. But apparently it has not revealed itself yet.

Kooiti Masuda said...

I agree with MT that the symptoms of anthropogenic climate change should not be framed as effects of increase of global mean temperature, but as effects of radiative forcing.

Kooiti Masuda said...

Regrettably, the time series of counts of hurricanes suffers from the changes of observing methods. You would not miss a hurricane since 1976 when GOES is available, but even then, there may be changes of alrogithms to determine ranks of storms. When we want to discuss whether the recent situation is extraordinary against the history, we need some homogeneity, either by restricting the range of analysis to well-observed part or by making adjustment.

Kooiti Masuda said...

As for the centennial trends of northwest Pacific typhoons, Hisayuki Kubota, my colleague at JAMSTEC, has been working. A paper Kubota and Chan 2009 GRL has been published. It says that there is no discernible trend in the count of tropical cyclone landing on the Philippines, though there are multi-decadal variations. His subsequent studies (yet to be published) shows that there is no discernible trend in the count of tropical cyclones in the whole northwest Pacific either, as far as we can deduce from locations where long records are available. But it is possible that the tracks has shifted somewhat.

I do expect (sadly) that anthropogenic climatic change will cause remarkable changes in the activities of tropical storms. But apparently they have not revealed themselves yet.

Michael Tobis said...

steven's point is sort of the flip side of mine. I agree with what he says, though it differs in emphasis.

He is arguing that severe events not be used in hypothesis testing. I agree.

I am arguing that even though statistical attribution of extreme events is not possible, that does not mean that the causality is absent. These points are compatible but not identical.

Michael Tobis said...

Paul, though I can see merit in the general idea, Hsu's particular basket makes no sense to me. What is that supposed to measure, and why?

Michael Tobis said...

Masuda-san, thank you for your input, and particularly on your support for the key point about focus on radiative forcing rather than temperature.

I believe the habitual attention to global mean temperature as a measure of the degree of impact is a misguided simplification.

The physical sensitivity of global temperature is of great scientific interest given the paleoclimate evidence. But I don't think it provides a good measure of impacts and emphasis on temperature is thus misleading in public communication.

We do not know the transfer function from temperature to damage. Also there is good reason to believe that impact depends greatly on rate of change as well as amplitude.

Also, it is easy enough to talk in watts rather than in degrees, and it leaves far less room for misdirection or obfuscation or innocent confusion.

Layzej said...

A lot of people doing a lot of little things is still difficult for me to get my head around. A lot of something unimaginably big somehow clarifies this. So at the risk of confounding the issue of nuclear winters...

I'm not so good at the maths so please feel free to correct me:

1 nuke = 15 kilotons of TNT = 62 gigajoules = 17430 kilowatt hours

1,000,000,000,000 kw/17430 kwh = 57,372,346 nukes/hour to sustain a 0.7C rise in temperature?

It seems like my math must be wrong. If not, that is a heck of a kick.

Michael Tobis said...

I think you are off by 10^3 according to this widget.

Only 57,000 nukes per hour.

Paul Daniel Ash said...

I'm not sure if those particular measures make sense, but Hsu raises them in the context of likely impacts of global climate change.

Is it meaningful at all to speak of a hurricane anomaly, rainfall anomaly, etc. in the sense of a departure from mean? If so, maybe an index of those anomalies could give an informative "weirdness" trend over time.

Word Verification: hobbe, nasty, brutish, and truncated

Michael Tobis said...

No, anomaly doesn't capture it. Maybe fourth power of anomaly would, to some extent. The idea is to define the weirdness that many of us think we perceive. So it's the extreme outliers that matter.

Like in baseball, it is possible to invent a whole lot of statistical measures after the fact, though.

Richard said...

Michael:

Are you not really saying that you suspect in your gut that there is attribution - but their is no evidence to support your gut yet?

You expect that later such evidence will turn up - but it is not there as of yet?

That is just an opinion - not science, right?

manuel "moe" g said...

Richard,

I believe Michael is saying the opposite:

If you roll two dice, you expect a certain rate of 12s to be rolled. "Rolling a 13" means a number outside of expectation to a degree that it suggests the game has been changed.

Many climate bloggers and reporters say:

[1] that these "rolling a 13" events are not caused by an unstable state enabled because more energy is available because of radiative forcing.

[2] and that these "rolling a 13" events could never be considered evidence of an unstable state enabled because more energy is available because of radiative forcing.

Michael is saying that statement [2] goes too far. I am inclined to agree with Michael. You can model an unstable state enabled because more energy is available because of radiative forcing. You can notice extreme weather events in your model. You can compare to measurement and reporting of extreme weather events now. You can then make a causality judgement.

In less politically contentious issues, precisely this thing is done, and sound decisions are thus made about helpful interventions, taking into account economic constraints and comparisons of outcomes based on values.

Richard, Michael, I fear I may have made a mistake in trying to distill this point. I appreciate any help to fix any mistakes and any unclear phrasing.

Michael Tobis said...

Richard, the definition of "science" is extremely context dependent and I smell a trap.

Something can be true and yet not be provable. Statistical attribution is a third case. Something can even be true and sceintifically provable but not statistically attributable. e.g., "There was a sudden and deep global cooling lasting a bit over a millennium that occurred around 12,000 years ago." That's a fact, in that there is nobody actively contesting it and plenty of work based on it. But as a single event, it has no associated statistics.

The tiresome attention to statistical proof is not always the best way to approach earth science.

Do I still suspect that the 2005 hurricane season was anthropogenic? Yes, I consider it more likely than not. Can I make a publishable case? No, obviously not. Does that make it speculation? Sure.

Is it true? Well I already said I thought so.

Is it sufficiently compelling to motivate some change in behavior? Well, the insurance companies running away from Florida in droves sure think so. Do you?

Is it "science" that such a link is established? No, of course not. But neither is the claim that the link is refuted. It's just as yet indeterminate, that's all. The science as it stands says that one should not behave as if the case were certain either way.

steven said...

"
I am arguing that even though statistical attribution of extreme events is not possible, that does not mean that the causality is absent. These points are compatible but not identical."

precisely.

Andy S said...

I'm not sure that "nudge" or kick" is a good metaphor. What we are doing to the atmosphere is cumulative, not transient. What's more, most of the cumulative effect will endure for the next few centuries and a good fraction of it will last essentially forever. We are not kicking or nudging the climatic camel but instead piling straw after straw on its back.

I'll attempt to answer Layzej's question in a different way, ie, in terms of stock instead of flow. At the end of David Archer's The Long Thaw he calculates that the "good" energy (the energy we can use) that we get from a gallon of gas will produce 40 million times the "bad" energy resulting from the lifetime effects of the CO2 emitted from burning it.

Michael Tobis said...

Yes, on human time scales it is more like a ramp. It will look sort of like a kick a long time after we stop.

But ramp responses also allow for ringing.

A central point here is that I am really exasperated by people who, when they don't find unambiguous trends, conclude that nothing is happening. And that is nudge thinking.

Grypo Saurus said...

The reason for using for temperature in place of radiative forcing was mainly to point to the heat coming from the rise in temperatures in the Gulf that are one aspect of a possible tornado enhancer. Otherwise, I agree with the idea of using "energy" and "forcing" as more descriptive words of the disease, so to speak. Temperature is just a measurable symptom. I've been saying this for a while.

Belette said...

> But my point again is that the system is getting wobblier

I don't think you have made your point convincingly. If you think you can, I think you should, because it is an interesting assertion, that I've seen other people make. But anecdotal lists of stuff isn't really support, or at least not good enough.

Michael Tobis said...

Belette, agreed, not good enough.

I should have said instead that I both 1) expect that the system will get wobblier and 2) suspect that it already has. But I not only don't have proof, I am not sure I know how such proof would be constructed.

Grypo Saurus said...

Micheal,

Random thoughts.

There must be some way to put the black swan data over the last 2-3 years in some perspective relative to, at least, the last 100 years. Small tornadoes aside, it really is about the chance of occurring of all the global flooding, the F4/F5 tornadoes, the heat, hurricanes (many although no landfall), monsoons in Asia, ice melt, etc. Then these should correlate with El Nina/La Nina far enough back to how this latest heat exchange "kicked" the system enough to cause this "wobble", if it exists.

Michael Tobis said...

Right, but how do we avoid baseball-ism?

"That's the greatest differential in extra-base hit percentages between home and away games by a rookie third-baseman in major league history!"

It's easy to say event X or event Y seems weird.

It is not easy to define weird in a pariori sense, or count weird events from a time when people were less attuned to them.

Suppose I postulate four more large scale weird events in the next 12 months. Now come back in a year and test my hypothesis. How?

Should we count the giant tsunami? That was the weirdest thing of all, after all. OK, no, because we really don't think that was forced. But now, how do we draw the line? Aren't we begging the question?

I am open to any definition. I think persistent jet stream anomalies may be a good place to look, for instance.

Andy F said...

Grypo Saurus,

Just how many regional 100/500/1000 year events would you expect to see in, say, the past 12 months?

I tried make the analogous case in space here wrt to the Memphis flooding: you might expect to see 500 year rains in O(20) stations out of 10000, but you don't expect all those stations to happen to be in western Tennessee.

Grypo Saurus said...

(I'm a big baseball fan, so that's very funny to me;)

Yes, we would have to neglect any event that is much more likely to be seen now, as opposed to, say 50 years ago, which eliminates small tornado frequency, but the point would be to look at major events, one's people don't miss, or ones we have good proxy-data for. Also, we would be looking at them tied to the natural oscillations of the ocean. We also (seem to, I'm not exactly sure) had some idea how often these events are supposed to happen, but unfortunately we do not have 1000 years of data to help that problem and will just be redefining those scales. So priori "weird" is going to be subjective unless those once in a 1000 year events are really only supposed to happen once in a thousand years. How likely is it that a lot of, seemingly unrelated, once in a hundred year events are all happening in the last two years, last decade, etc. And how accurate are those assessments?

Jim Bouldin said...

I'm still having a hard time with the way you discuss this. I could be wrong but it seems to me that you think of attribution as the detection of a statistically significant difference in something. That's not attribution, that's statistics. Attribution involves assessing the relative efficacies of various cause and effect possibilities in explaining a set of observations.

My point is that not only are you correct that getting no trend does not mean there is nothing happening, but that you are selling short the power of attribution. I don't care how rare these black swan events are, if you have reasonably well developed physical models that you can evaluate against each other, relative to the observed data, then you *will* get a probabilistic evaluation of what is most likely happening, within the constraints of the knowledge embedded in those models.

David B. Benson said...

Atlantic hurricanes are promoted by high tropical SSTs. On the other hand, sand blowing off the Sahara tends to kkep tropical storms from developing eyewalls. So does El Nino conditions which promote winds aloft, knocking off the tops of nascent hurricanes (hurricanelets?) and preventing eyewall formation.

So it is conceivable that as the climate moves towards permanent El Nino conditions there will be fewer hurricanes; instead just more very large (and wet) tropical storms.

Steve Scolnik said...

"I think persistent jet stream anomalies may be a good place to look, for instance." (MT @3:52 PM)

Stu Ostro of The Weather Channel has been doing just than in terms of enhanced frequencies of cut-off low circulations.

steven said...

Jim

"but that you are selling short the power of attribution."

what is the power of attribution.

are you talking about making the case for AGW?

That case is closed.

Jim Bouldin said...

Steven,

I'm talking generally. Michael is (I think) claiming that that the occurrence of rare events cannot provide evidence one way or the other about what is going on, physically. I am claiming that it can--not in a "hmmm all these extreme events seem weird", type of way, but rather in a definite quantifiable way. And that way involves the comparison of different possible models, i.e., attribution studies.

steven said...

Jim,

I suppose if one had a regional model with enough granularity that could predict tornado genesis, then that would be down the right path for attribution. absent that we've got vague descriptions like more X should lead to more Y.. ya but

Michael Tobis said...

Jim, you miss my point. My point is that while it may be difficult to attribute any of these single events there is an informal perception of general global weirdness that ought to be reducible to a formal, testable claim.

So I'm trying to encourage exploration of what such a formal claim would look like, in order to both be a reasonably sound representation and test of our informal perceptions.

Of course, a thorough understanding of phenomenology is better than a statistical test. This is another point to be made to those obsessed with statistical attribution.

On the other hand, post hoc explanations aren't all that satisfying compared to predictions.

Nobody saw the Russia/Pakistan/China thing coming before last summer. If it returns, we'll say "oh that again" and perhaps respond better. And we'll say, "maybe this is a recurring pattern in the changed climate".

It would be better if someone had noted this pattern in model projections and written a paper about it. As far as I know, even that didn't happen.

But "explaining" it in terms of climate change is, well, a strange and daunting challenge when you get down to the level of detail.

OK, now we have the most tornadic month in history to add to the burgeoning pile of astonishing outliers. Same problem. What sort of post hoc explanation is needed to link it to climate? And what purpose would be served?

I'm not saying it shouldn't be done, but I'm suggesting it's not important outside of pure science.

On the other hand, if the overall trend of rare and unprecedented events is up already, if both of these events and a host of others can be connected through some sort of dynamic understanding of how the perturbed climate differs from a quasi-equilibrium climate, now that's a result to take notice of.

But the thing is, it's an extra-scientific perception now. We don't even know how to formulate it as a hypothesis!

Jim Bouldin said...

Steven, I agree, and your latter approach seems to be the one that people have taken (and which I'm advocating). I'm assuming there are fairly solid statistical relationships between, e.g. the frequencies of frontal collisions, thunderstorm formation, and tornadoes. Of course, those could change, I realize that too. And the whole thing could just move spatially, as someone mentioned. No real change in frequency or intensity, just different geographically.

Jim Bouldin said...

Michael, you have the habit of saying a whole bunch of sensible things...and then throwing me completely. I was doing OK until you said "I'm not saying it shouldn't be done, but I'm suggesting it's not important outside of pure science." "Pure science" is in my mind, the whole point of why it's important.

Anyway this is a great discussion that gets to the heart of how science is, or should be, done. More later.

Paul Daniel Ash said...

Blogospheric harrumphing at the question "What sort of post hoc explanation is needed to link it to climate?" in 3... 2...

steven said...

OK, now we have the most tornadic month in history to add to the burgeoning pile of astonishing outliers. Same problem. What sort of post hoc explanation is needed to link it to climate? And what purpose would be served?

#######
at the risk of belaboring a point I made else. In normal science no one would think twice about making ad hoc explnations and tenuous hypothesis. in normal science if somebody predicted more hurricanes and we saw fewer it would be no big deal. But in THIS environment these kind of normal theory extensions get waaaay more attention then they deserve, especially when they go wrong. we dont allow "honest error"

In normal science these hypothesis at the boundary of understanding serve to extend the reach of the theory. But in our current circumstance they get used two ways.

1. to sell concern ( when they are correct... fine)

2. to spread doubt (when they are wrong.. not so fine )

in normal science an honest error at the edge of understanding just gets corrected. no big deal. In today's situation these honest errors get used for other purposes.

Not judging anything here, just describing what actually happens

Jim Bouldin said...

Michael, I think Steven just spoke for me. A *huge, important* part of science is what you call "post-hoc" analysis--looking at data and trying to model explanations for it. Then observing new things and seeing how well they do, or do not, line up with your previously constructed models. I don't see why you have a problem with that. You don't get good future predictions without a lot of post-hoc analysis.

You ask what "such a formal claim would look like". I think there's enough of a basis in how that basically looks already, and I really don't see that it looks any different from how most of environmental science is done. The only difference is the statistical confidence you're going to have that some hypothesized effect, e.g. increased tornado frequency, is already evident. The power to do that would be increased by making the analysis multivariate instead of univariate. Since you're then looking at several effects at once, you seemingly have to approach this through GCMs, which capture the global forcings, and which are then linked to a suite of more spatially resolute models that can downscale and translate GCM output to the various effects of interest.

Is that what you're looking for?

I'll admit that I need to read more of the literature on this though. I've started with the tornado stuff.

Michael Tobis said...

Jim, I think it;s always a welcome turn of events when someone with expertise in another field starts looking at climate in detail. It tends to be both refreshing and frustrating for us, as the attentions appear somehow selective.

I'm a little baffled that you don't see what I'm saying. The number of scales and dimensions of the system is very large. The question is what sort of event to take account of, and presuming that we are looking at events whose repeat interval is low, how to cope with the intuitive sense that they are coming on far too often to consider ourselves in "nudge" territory anymore.

I already stipulate Roger Pielke Jr.'s point that no trend in tornados or hurricanes is visible. Some people argue that the trend in severe hurricanes is to increase, which would support our intuition.

Calling this a "multivariate" hypothesis is at best premature. I'm trying to make a link between the informal impression and a real hypothesis, so we can establish whether our intuitions are misleading us. I think they aren't, but I don't know how to test that, and that's partly because I don't know how top assert it as a hypothesis.

I know people give Popper far too much credit but in the present case his point of view is very applicable. We are asserting a vague hypothesis. How do we go about turning it into something we can test?

You can say "mutlivariate" which is fine, but you aren't helping me assert what those variables are.

Persistent jet stream anomalies is a good place to start, and cutoff lows has the additional advantage that we can count them manually. But not all jet stream excursions that are extreme show cutoff lows. And at best this will only capture extratropical anomalies, but it will probably capture most of them.

So, yes, I think we should look at reanalysis data, and we should look at things like CAPE and jet stream position in it, to look for the sorts of things that people on the ground actually notice. A good start. But how to codify and cluster the "normal" climate and how to identify the outliers remains less than obvious.

susan said...

This may be off the wall, but would variation in the AO be useful, or related arctic changes?
--
Recently good booted again (once again I was doing something complicated in another window, think that's part of the problem) but will try again. This means I already quoted you on comments that will hopefully go through on RC. I will desist in future if you ask me not to do it again.
--
I've been excited to be learning something new - difficult but exhilarating. Having trouble understanding how scientists can say what they say about what seems to my ordinary observant eye (age 63) to be radical changes in weather.

Something I saw yesterday felt like a bullseye, but when I came back to find it I couldn't. Decided to put all mt comments in recent articles on weather here in one text document. That's some body of work you're putting together - great stuff!

Michael Tobis said...

Please, quote me far and wide, and let me know if you can come up with any way for me to get paid to do this sort of thing!

If your browser crashes, you need to make sure your browser is up to date and get more memory for your computer. If Google sends you an error message, though, it has nothing to do with the number of windows you have open; that cannot affect them on their end. Either way, it's great that this stuff almost works. We should probably be grateful more than irritated.

Jim Bouldin said...
This comment has been removed by the author.