[ExI] Hurricane Sandy

[Anders Sandberg]

On 31/10/2012 10:20, Charlie Stross wrote:
> Per news reports I've seen in the UK, MTA commissioned a report on mega-hurricane disaster of Sandy proportions a couple of years ago, but hadn't upped their maximum anticipated disaster to match -- Sandy exceeded their "worst case" assumptions, and even after they pump everything dry and repair the damage, Sandy-proofing the subway against a future hurricane of identical proportions will probably cost billions.

One of the (mathematically) cool and (practically) horrifying aspects of 
long tail disasters is that "if it is bad, it is likely far worse than 
you thought". One can never make a bad enough worst case scenario.

[ If you have a damage threshold T, the exceedance X-T of how much worse 
the actual disaster X is, will have an expectation that is on the order 
of T or larger - when bad things happen, they tend to be really bad, not 
just a millimeter above the levee. ]

You can reduce many risks by adding big safeguards, but that is just the 
first step. Security in depth, resiliency, ability to quickly recover 
etc. matter a lot more than a high dam. New York will recover because it 
is a rich city full of people willing and able to work and pay for 
fixing it, but no doubt there are ways of making it even more resilient.


>> I was really looking for preemptive strikes against the hurricane
>> itself.  If cloud-seeding can make rain, is there a similar something
>> that can artificially/manually downgrade a hurricane?
> The energy density of a hurricane is staggering, and it's driven by humidity and temperature gradients over an area the size of a sub-continent. It's not immediately obvious that we can do anything about this short of large-scale geoengineering to reduce global energy inputs. (It's not "global warming" so much as "global absorbing-more-energy-in-a-dynamic-system", resulting in wilder fluctuations and more extreme weather events. Such as Sandy.)

Local OTEC cooling has been mentioned. One could go to the megascale 
"what could possibly go wrong" approach with orbital solar shades too. 
An intermediary approach would be JoSHs nanotech based climate control 
clouds (floating minimachines that regulate albedo and reflectance).

None of these are enough to stop an ongoing hurricane (I think), but 
they might be enough to do chaos control. The neat thing about chaotic 
dynamical systems is that they are unstable and happy to weer into other 
directions if given just the right push. You need to measure the 
situation very well, simulate it very fast, and keep on pushing in order 
to get the effect. You will not be able to achieve perfect control - 
this is a probabilistic method - but it would likely reduce risks a lot. 
At the price of having the Atlantic Weather Control Board responsible 
for the inevitable slipups *even if they do a perfect job*.





-- 
Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University