"Our greatest responsibility is to be good ancestors."

-Jonas Salk

Wednesday, June 27, 2007

Weird weather

18 inches of rain in 24 hours, yesterday, just upriver from us at Marble Falls TX. The flood control people (the LCRA = "Lower Colorado River Authority") are going bonkers trying to smooth this event out and parcel out minimal flood damage. More rain is expected in the Colorado basin this week.

Such rainfalls are not unheard of in Texas, but usually they are associated with tropical storms, which this one isn't.

Here's hoping we don't get one of those soon.

No individual event is caused by climate change. By definition, "climate change" loads the dice in favor of certain types of events. Theory (Allen and Ingram, Nature 2002, v 419 pp 224 - 232) indicates an increasing likelihood of such severe local precipitation events in the foreseeable future.

Update: more here.


Heiko said...

There are two issues I have with this type of argument.

Firstly, it's true that it's like loading the dice and you can't attribute a single event. But, that doesn't answer the question whether for a particular event the dice have in fact been loaded in a direction favouring it or not.

It seems to me that this train of thought is regularly stated without asking that question, it's just assumed that climate change has made this event more likely.

It will, in many cases, but equally, for say certain storms, or say flooding in June in England, the dice may be loaded the other way by climate change, and the event so to speak happened in spite of the climate change induced trend, rather than because of it.

The second issue I have with it is perspective, and you know I've talked about that before. Namely, we know that more warmth means more evaporation, and as evaporation has to equal precipitation on a world scale, it'll mean more precipitation. But, to what degree does that mean more drought / more flooding,

and to what degree is this merely the equivalent of moving a little bit more towards the equator, as opposed to this notion of "a general increase in the freakishness/destructiveness of weather"?

I think the latter is quite poorly supported even on a world average scale. But at least, there you can argue it by looking at droughts or tropical storms.

On the "this increase in freakishness will occur everywhere" level, I think it's quite silly and not at all supported. Every place on Earth will see more droughts / more storms / more intense storms / more flooding?

Given enough warming, every place should see an increase in heat extremes and a decrease in cold extremes, but hail, storms, wind hoses, flooding, drought?

You may say, "ok, but what does this have to do with the dice loaded passage"?

It's that for any and every bad weather argument, it seems to be wheeled out in press reports, implying that the particular event has been made more likely by climate change,

and given the breadth of the coverage, that it is a worldwide thing and that every region will suffer more "weird" weather.

Michael Tobis said...

Hi, Heiko!

I make no claim that any individual experiencing a weird weather event can attribute the particular event to anthropogenic climate change. I explicitly stated the contrary.

Theory predicts that a larger proportion of precipitation will occur in severe events in a greenhouse world. I'll try to find the reference; I had little trouble last time I looked for it but I can't recall the author. It was a very readable review paper in Science or Nature.

Models concur.

I believe observations are consistent with this anticipated trend but I am not sure whether they are decisive as yet.

Mostly I was just reporting some weird weather, though. If you had experienced the like I think you'd mention it.

Heiko said...

Thanks, the idea of more heavy precipitation events seems very intuitive to me, so you don't really need to convince me on that one. Admittedly my intuition is based on weak evidence, namely I look at what rain's like in the summer/winter, and what it's like in England/Malaysia.

No weather event is directly attributable to climate change (a particular storm or heat wave anyway, and I am not sure this would hold when we get beyond a certain level of climate change, I mean a tropical cyclone hitting Ireland would be so unlikely in a pre-industrial climate that I'd be quite willing to say that it is attributable to climate change), but I digress, and certainly it's interesting to hear weird weather reported,

what troubles me when the point about "no particular event being attributable" gets made in press coverage (rather than your blog) is that I feel it is implied that while the event isn't attributable to climate change, it's been made more likely by climate change.

Taking the example of heavy rain in Texas, it may very well be that the change in weather patterns is such that in this particular locality the world trend is not replicated for heavy precipitation, and that for this particular location (and say season of the year) events like the one you report are now less likely to occur than in a pre-industrial climate.

There's two levels here, you can either blame all weird/extreme weather on climate change (which is a pretty poor position to take, and you've railed against yourself already), or you can say (which is much, much less strong, but still clearly not right) that for all locations in the world, climate change means, on average, more droughts, more floods, more storms, more intense storms.

I am probably overreacting anyway, after all, you can also argue that this "this particular event is not attributable" line isn't about implying very much at all, and in so far as it is, just that there'll be damages from climate change overall.

Michael Tobis said...

The claim is actually stronger than "more heavy precipitation events"; there is also the claim that there will be fewer light precipitation events.

Essentially, a hotter world is a riskier world for an individual specimen (of a species or an infrastructure).

Imagine a world where all the rainfall (balancing all the evaporation form all the oceans) occurred in one place, and that place changed from year to year. It would be extremely inhospitable in the dry places and the wet place.

The world I described is not possible, but the greenhouse world moves closer to it. There is a compelling dynamical argument for it.

John Fleck said...

My confusion was simpler than Heiko's. I didn't realize y'all had your own Colorado River. :-)

Heiko said...


Found a free pdf version, which you might like to link to in your post. It doesn't take that much searching, but people might give up when they see that Nature's version is behind a subsription screen.

Looking at the paper, the stuff about heavy precipitation events is a fairly short discussion. I am not sure I entirely understand it, but it sounds very much like they are saying there's a direct relationship between heavy precipitation events and temperature based on simple physics? I.e. it would likewise apply for the differences between summer and winter or between English rainfall patterns and tropical rainfall patterns?

I suppose that's what I thought before looking at the paper, and maybe they are trying to say something else?

You see I am not the least worried about that kind of change.


Also, the map on average precipitation changes is very interesting. It seems that the Mediterranean would be the major area with decreased precipitation, with no other important land areas seeing major decreases.

On the other hand, most of Asia will see an increase, particularly India stands out on that front.

Michael Tobis said...

No. You aren't thinking of this correctly at all.

Everything is very tightly coupled. Climate change is not like the difference between north and south nor is it like the difference between summer and winter.

Nor are the seasons especially like latitude shifts. Your suggesting both gives me the idea that you aren't especially interested in the atmosphere. That is fine, not everyone needs to share my obsessions, but then you should limit expressing your opinions to matters you actually find interesting.

The meridional temperature gradients at all altitudes and all ocean depths are the primary control of the physics of the weather.

The idea that a 3 C global temperature change will amount to a 3C local temperature change with modest shifts in rainfall patterns is not even possible. Many previously consistent patterns will change.

If the continental ice sheets remain relatively stable, 3 C will be tolerable, but it will be quite stressful; many places will experience shifts upward of 10 C as a consequence.

Ecological disruption, at the least, will be very large. The anthropogenic extinction event will be greatly exacerbated.

So we had better stop there.

Heiko said...

It's certainly not like the seasons or latitude in all respects, but temperature does vary with both, as it would with climate change.

The explanation for more heavy precipitation events in that paper sounds awfully like at least this particular change is quite closely aligned with temperature, ie there are basic physical reasons why when it's hotter you are less likely to get drizzle and more likely to get a big downpour. Is that right, or is there something more behind the prediction for more heavy precipitation events that's escaped my attention?

(In many respects, I am probably far too interested in the atmosphere, at least, I often think to myself I spend too much time looking at what would to most people be arcane details ...)


Incidentally, a quick google gave me this temperature map for a mean temperature rise of 3.2C. The maximum local rise in the map is 7.8C and the minimum is 0.5C.

What strikes me there is that the populated areas that get hit with the biggest increases are tropical/subtropical (India, tropical South America, Africa), ie already too hot.

Michael Tobis said...

Heiko, I'm afraid you made my point. If you were really following closely you would not have considered that map representative of current modeling.

(Here's a link for those having trouble with Heiko's text.)

Did you look up the vintage of the model, HadCM2_IS92A?

Do you think scientific computation has made no progress since 1992?

Do you suppose the climate system is somehow, uniquely among complex physical systems, unsuitable for simulation? Or do you think that scientists interested in climate are uniquely incompetent?

I doubt it, so I wonder if someone lead you to believe that this is a serious prediction?

Here's the 2xCO2 surface temperature plot from CCSM3 which is broadly similar to all the contmporary AR4 CGCMs. Plots that are not broadly similar should be treated with suspicion. Top map = 2 x CO2 equilibration, middle map = unforced control (equilibration to fixed 1990 forcing), bottom is the difference.

Your first recourse should be IPCC AR4 WGI, not Google.

Heiko said...

I am not sure I get your point here. Yes, I used google, but that gave me the Hadley Centre of the Met Office, they've got two versions of their model there, which differ slightly from each other. I didn't expect major changes between them, and it seems the bit that surprised me (India getting particularly hot) has been revised a little.

And the map you've presented is for a higher mean temperature change and uses a different colour coding, so it's not that easy to compare the two.

What I did the google for was to (quickly) check whether climate models really showed that 3C average meant "many places will experience shifts upward of 10 C as a consequence.".

You are right I could have looked at AR4 instead, but it's not very search friendly at all. I know that the graphics I am interested in are buried in a 100 page pdf (which I have looked at a few weeks ago, skim reading all of it, but properly digesting maybe a fifth).


Not only are the graphs buried on page 766 (20 out of 100 for the pdf), you then have to go back a few pages to check the related mean warming (3.13 C for scenario A2).

And I still don't find many places experiencing shifts upwards of 10C for a mean warming of 3C.

Incidentally, if AR4 is the source to use, why didn't you link to that yourself?

And you haven't said anything yet about heavy precipitation events. My question there is, are we just expecting more of those based on basic physics and temperature (ie hotter air can hold more moisture, and therefore heavy precipitation requires hot air, in short), or is there something more behind this claim that I haven't picked up on?

Heiko said...

You may think that you've answered the question about heavy precipitation events by pointing to fewer light precipitation events. I am not sure why that's going beyond the simple physics type / temperature explanation, because cold air holds less water vapour and it would therefore seem easier to have light precipitation with cold air.

What I am wondering is whether the more heavy precipitation / less light precipitation story goes beyond that direct temperature link? Is there something in the modelling of global circulation patterns that suggests that on average we'll see a more pronounced trend in concentrating precipitation than would be expected from the fairly simple fact that hot air can hold more moisture?

I hope I am not being too much of a nuisance here, and asking too silly questions.

Michael Tobis said...

I should have said 10 C winter mean temperature changes, not annual mean. My apologies.

These are important as direct impacts, as winter temperature sets the poleward range for many potentially invasive species. Here is an example from Alaska that should be noted.

Regarding your other questions, the trouble is that you are asking for an introduction to meteorological dynamics without being aware that you are doing so, and perhaps only dimply being aware that there is a rich and mathematically rigorous science there.

The way you are thinking about it is badly off base, but I'd rather not be too hasty in answering.

Let me ponder the best way to address this without making you read Wallace and Hobbs. If I don't get back to it feel free to remind me.

If you do want to get serious, "Atmospheric Science" by Wallace and Hobbs is a good place to start. I also think "Physics of Climate" by Peixoto and Oort is a very nice introduction. Both assume only undergraduate physics and calculus.

Michael Tobis said...

"dimply" should read "dimly" above