At 12:29 AM 12/17/2004, Mark Wallace wrote:
>They still can't tell me if it's going to rain tomorrow. The charts cited
>above rely on something called a ""Heidke Skill Score,"" which is a
>controversial verification technique, even within NOAA:
>
http://ams.confex.com/ams/84Annual/tech ... _68014.htm
This is an interesting point, though I've seen it claimed elsewhere that
forecasters don't understand Heidke so it seems unlikely they'd
deliberately manipulate it. It would be interesting to know whether Garrity
scores are going up or not.
I think I dragged a red herring through the debate though. The improvement
in short term forecasts could be entirely due to improved data, for
example. However, this is the wrong question to ask of GCMs. Inability to
perform point predictions is not, by itself, a reason to reject a model.
There are lots of useful models of things like turbulent flow that
adequately reproduce the behavior of a phenomenon without actually
replicating any particular instance.
> > A corrolary to ""the map is not the territory"" might be ""you can't plan a
> > trip with the territory.""
>If the map isn't that good, it might be optimal not to take the trip in
>the first place.
> > It's not irresponsible to use primitive results
> > for purposes of public policy; using nothing at all is worse.
>I respectfully disagree. While results are still primitive, the only
>sensible public policy is ""hands- off."" The compulsion to DO SOMETHING
>(i.e., act for the sake of acting, before carefully considering the
>consequences) should be resisted. [...]
This is not an application for JWF's (possibly apocryphal) admonition to
""don't just do something, stand there."" We can't choose not to take the
trip (unless we all join the Voluntary Human Extinction movement) - we have
to choose which trip to take. Currently, we are taking a trip that
subsidizes resource extraction and consumption, taxes goods (wages) instead
of bads (emissions), ignores externalities associated with energy use
(especially in the transport sector), regulates the pollutants that are
controlled by inefficient means, prices electric power at average instead
of marginal cost, and values carbon emissions at 0 $/ton. That's hardly
""hands-off."" It's a ""hands-on"" policy based on assumptions that technical
progress will continue to outstrip declining resource quality, sensitivity
of climate to 2x CO2 is 0, etc.
>[...] In particular, it is irresponsible to propose (in effect) limits on
>energy consumption without making the slightest attempt to evaluate the
>costs of imposing those limits. Doctors understand that ""first of all, do
>no harm"" is an important principle when dealing with the unknown. I see
>no such humility among Global Warmers.
In fact there have been many attempts to evaluate the cost of limiting
emissions - there is an entire field of Integrated Assessment Modeling of
global change. There must be at least 20 significant IAMs out there, not
counting dozens of variants and small models. All of these explicitly trade
the benefits of avoided climate change against the economic costs of
current emissions reductions. Even the most conservative models - for
example, DICE (by economist William Nordhaus, one of the earliest and most
vehement critics of World Dynamics/Limits to Growth, hardly a ""Global
Warmer"") suggest an initial carbon tax in the ballpark of $10/ton, rising
over time. By ""conservative"" I mean that these models arrive at their
conclusions using assumptions that generally disfavor GHG emissions
reductions - infinite sinks for carbon, infinite technology, full
substitutability of environmental and market goods, discounting the welfare
of future generations for pure impatience, etc.
> > It's easy
> > enough to build policy models with an appreciation of the uncertainty.
> Even if
> > our subjective distribution of the climate sensitivity to 2x CO2 has
> zero mean,
> > it's likely to be optimal to take some action in order to prevent
> scenarios in
> > the tails.
>That depends on whether or not ""some action"" will keep billions of people
>in a poverty they might otherwise have escaped had more energy been
>available. And by the way, poverty, anywhere in the World, is a leading
>cause of death.
It's a gross overstatement to claim that GHG emissions reductions will keep
billions of people in poverty. The energy sector is not that big in the
economy - even a ""ruinous"" total elimination of emissions in the DICE model
doesn't cause a decline in global GDP/capita; it just requires foregoing a
couple of years of growth.
The absolute global growth impacts of climate damage and emissions
reductions are insignificant compared to distributional issues. From a
welfare standpoint, the victims of climate change will be
disproportionately in places like Bangladesh, and disproportionately unable
to use the products of growth to invest in things like irrigation and
coastal defence. Again, some existing IAMs make this tradeoff explicitly.
To continue with the same example, there is a regional version of DICE,
called RICE, that includes seven regions at varying levels of development.
It still concludes that modest carbon taxes are beneficial. It does this
using an optimization framework that assigns greater weight to the welfare
of individuals in wealthier regions (in order to prevent large capital
flows to the developing world). If you relax that morally bankrupt
assumption, and set climate policy with fair welfare weights, I expect that
the outcome would show that emissions reductions in the developed world are
the ultimate form of development aid, because unlike monetary payments they
can't be squandered by institutions.
In order to favor doing absolutely nothing, you have to believe that
climate change won't happen, or won't be damaging, and that there's zero
chance of being wrong. For the rich to be cheerleaders for growth in the
name of the poor, while subjecting the poor to greater risks than the rich,
seems at best self-serving.
>I haven't seen any models of the effects on human well-being of reducing
>energy consumption. This is an area where the SD community could make a
>major contribution. SD is supposed to excel at pointing out unintended or
>unobvious consequences. Here is an opportunity to do so. It is just
>astonishing to me that SD has not been used to look at the cost of
>reducing energy consumption. Or have I missed something?
You have definitely missed something. There has been a great deal of energy
work in SD. Check the SD bibliography for Andy Ford's papers or MIT
Archives online for dissertations (John Sterman's and mine), to name just a
few. The energy-economy models in SD, like climate IAMs, frequently include
an explicit utility function or other output metrics representing the
benefits of goods and services consumption. Energy is seldom a direct
component of utility because it's not energy per se that is in demand.
People want warmth, light and mobility, not gigajoules. The real question
is how to deliver energy services efficiently. In SD and other models, the
direct cost of reducing energy consumption is generally the substitution of
capital for energy through installation of more efficient, but more
expensive, devices.
To the extent that there are costs associated with energy use that are not
internalized in current markets there can be negative net costs for many
energy/emissions reducing policies. Obvious examples include surface air
pollutants, which go nearly untaxed in the transport market. Not-so-obvious
examples include things like insurance externalities (i.e. making some of
the fixed cost of vehicle insurance into a variable cost through a gas tax
would have net benefits).
In addition, bottom-up engineering-economic studies generally indicate
potential for negative costs of energy conservation, even when
externalities are neglected. This might be due, for example, to biases in
perception of costs and benefits. Economists have more or less unilaterally
declared that this ain't so, e.g. due to hidden costs of implementation,
but on the basis of fairly flimsy evidence and circular logic in my opinion.
> > The >10 degree jump at the end of the Younger Dryas and the <1 degree
> > increase in the 20th Century should not be compared because the first is
> > regional and the second is global.
>I am curious if the poster knows this for a fact and, if so, how. The
>article makes no mention of such a limitation (that the increase was
>regional, only). Just because the data was collected at only one location
>doesn't mean that the increase couldn't also have been occurring elsewhere.
There are several reasons one should use caution when comparing regional
and global temperatures. First, changes at high latitudes tend to be bigger
than changes in the tropics. Second, if local climates were just random one
would expect regional variations to exceed global variation because the
mean of a bunch of measurements is more stable than the individuals.
Third, climate isn't random, but one would still expect the same kind of
behavior for physical reasons: for example, a warm current could shift from
the east of Greenland to the west, dramatically changing local climate
without having much effect on the global energy balance.
The increase at the end of the Younger Dryas is visible elsewhere in the
global record, though not all abrupt changes are. (""This abrupt event can
be found in paleo records from many parts of the world, although not
necessarily to such an extreme degree.""
http://www.ngdc.noaa.gov/paleo/ctl/abrupt.html ; See also
http://www.ncdc.noaa.gov/paleo/abrupt/data_glacial.html .) However, the >10
and <1 degrees are not on the same yardstick.
Tom
****************************************************
Tom Fiddaman
Ventana Systems, Inc.
http://www.vensim.com
PO Box 153 Tel (406) 578 2168
Wilsall MT 59086 Fax (406) 578 2254
Tom@Vensim.com http://www.sd3.info
****************************************************