=========================================================================
Date: 6 January 1995, 17:26:16 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: why study global warming
 
         Michael Tobis posted:
>Well, it would be stupid to perform those studies if we are going to ignore
>the results. (This seems to be James Shearer's position: why should I fund
>you if I'm not going to listen to your results? This seems silly to me but
>it has the virtue of consistency.)
 
         Actually my position is, why fund more studies of global warm-
ing if they are just going to conclude what the numerous previous
studies of global warming concluded (ie that adding CO2 to the
atmosphere will probably produce a warmer climate)?  Where is the value
added?
                    James B. Shearer
=========================================================================
Date: 6 January 1995, 17:33:25 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: scientific culture
 
         Michael Tobis posted:
>There is a difference between accusing a field of error and accusing it of
>deliberate distortion. ...
 
         I am not accusing climatology of either.  I am pointing out
that climatologists have an interest in promoting concern about and
study of global warming.  As is well known people's views are affected
by their interests.  For this reason it is reasonable and legitimate
to consider people's interests when evaluating their views.
         In this regard Rich Pulchasky posted:
>And what organization is that?  I post here as a private individual,
>not as a representative of some organization.
 
         If you are a employee of an environmental organization then
I believe it is reasonable to expect you to disclose this interest in
your posts to sci.environment.
                          James B. Shearer
=========================================================================
Date: 6 January 1995, 18:01:29 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: environmental decisions
 
         Rich Pulchasky posted (as quoted by Michael Tobis, I was off
the net for a few days and missed the original):
> To put up our positions in the same strawman fashion as Shearer prefers :-),
> Shearer wants the market to make environmental decisions, Tobis wants them
> to be made by scientific experts, and I want them to be made by elected
> legislatures.
 
         Actually my method of argument is reductio ad absurdum.
         The "market" can only make decisions in the context of the
structure of the society within which it exists.  Change the structure
(by passing a law for example) and the market may arrive at a different
decision.  So a position that the "market" should make decisions does
not really settle anything.  It is correct that I believe that society
should be structured so that markets have a large role.
         John McCarthy posted:
>Rich Puchalsky includes:
>
>      I've listened to what Tobis has said and I disagree; if he
>     wants to have his view prevail over mine he will have to
>     gather the political power to overcome those who beleive as
>     I do.
>
>I'm hoping we just did - at least partly.
 
         As I understand their positions, Puchalsky believes environ-
mental decisions should be subject to popular passions as expressed
in elections (among other things) whereas Tobis believes they should
be made by experts insulated from public opinion.
         I fail to see how the recent elections favor Tobis's position
over Puchalsky's.  Any change in environmental regulation will be a
result of the election (Puchalsky's preference) rather than a change
in expert opinion (Tobis's preference).  I don't recall the Republicans
pledging to insulate environmental regulation from political pressure.
One can favor democracy without being thrilled by the result of
every election.
                          James B. Shearer
=========================================================================
Date: 6 January 1995, 18:57:40 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: discounting the future
 
         Rich Pulchasky posted (as quoted by Carl Lydick, I was off
the net for a few days and missed the original):
>First of all, many discounters advocate an 8% rather than a 3% rate.
 
         Certainly it is possible to slant results by choosing an
inappropriate discount rate.
         Rich Pulchasky posted (as qouted by Carl Lydick):
>Keep in mind, also, that the example I gave above is mot a comparison
>of costs now vs costs later or deaths now vs deaths later.  It's costs
>now vs deaths later.  You're not saving any lives by permitting the
>pollution which will kill people later; you're just saving the company
>money.
 
         Whether it is legitimate to put dollar values on human lives
is a separate question.
                       James B. Shearer
=========================================================================
Date: 6 January 1995, 19:09:23 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: conflict resolution
 
         Dean Alaska posted:
>   ...  More scientific research will not resolve these conflicts
>because conflict resolution - short of the black&white victory/defeat
>paradigm - is a skill that our society is deeply lacking in.  ...
 
         Actually I think our society is pretty good at this,
certainly compromise outcomes are common contrary to what you say
above.  What societies do you think do a better job at conflict
resolution?  You don't have to look too hard to find societies that
do a worse job.
                          James B. Shearer
=========================================================================
Date: 9 January 1995, 18:19:44 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: learning curve effect
 
         Rich Puchalsky posted (in reply to someone else):
>: For example, let's say that a new process was discovered that saved
>: $1,000 off the cost of building a car.  Assuming that there is no
>: monopoly enforced by patents or other means, who benefits?  Well, the
>: car manufacturers can't keep the $1,000 because they are in a
>: perpetual price war, so the cost of cars goes down by a bit more than
>: $1,000.  Car buyers benefit.  Since more cars are sold, auto workers
>
>BTW, how does the cost of the car go down by _more_ than the amount
>saved, over the long run?
 
         I am not the original poster but I believe the argument is
as follows.
         The reduced cost of production of cars will be reflected in
the price causing increased sales and production.  The increased
production will cause the unit cost of production to decrease further
due to economy of scale and learning curve effects.
         This effect can be readily observed for products based on
new technology.  Once the price of a new product (such as CD players)
becomes low enough to attract a mass market, rapid further decreases
in price often occur because of the increased production.
         Advocates of alternative technology (such as solar photo-
voltaic) often claim such a process is on the verge of starting.
         For a mature technology such as cars one would expect that
the effect would be minor as the original poster indicated.
                          James B. Shearer
=========================================================================
Date: 9 January 1995, 18:57:47 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: disclosure
 
         I posted:
>         If you are a employee of an environmental organization then
>I believe it is reasonable to expect you to disclose this interest in
>your posts to sci.environment.
         Loren Muldowney responded:
>Certainly.  It is also reasonable to expect disclosure of employment
>in a for profit industry which depends upon environmental degradation
>for its bottom line, also stockholding in such industries, also one's
>entire asset portfolio and the location of one's actual personal
>residence, income tax bracket and deductions.  Certainly anyone
>employed by or profiting from financial markets and compound interest
>in general would be ethically required to disclose that fact, they
>may well shoot down any idea which points a finger at the fundamentally
>flawed concept of perpetual growth as a physical possibility and goal.
>Is that what you meant?
 
         Actually, no.
         I don't believe it is practical or desirable to enumerate
every minor interest you might have while participating in this group.
However I do believe major interests should be disclosed.  While
there may be borderline cases where reasonable people will disagree,
I believe the above case (employee of an enviromental orgranization
posting to sci.environment) is clearcut.  Note your employer has
an unique ability to retaliate if your employer finds your posts
offensive.
                    James B. Shearer
=========================================================================
Date: 9 January 1995, 20:04:33 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: Lucas vrs South Carolina Costal Council
 
         Jon Heaton posted:
>I am not familiar with the other 2 cases you cited, but I have read about
>poor Mr. Lucas.  It seems that Dave bought his beachfront property while
>fully aware of the laws against development on it. (It was in a hurricane
>zone and prone to periodic flooding and destruction, and was also in an
>unstable tidal area with a frequently shifting boundary) He then proceeded
>with plans for a large resort development on the property. When he was
>stopped due to the already existing laws, he sued. And not just for the
>price of the land which he had paid mind you, no, he sued for all of the
>future profits he was being "deprived" of obtaining.
         Kevin M. Okleberry posted:
>You obviously haven't read enough about Mr. Lucas.  True, he was prevented by
>an already existing law, but his property was bracketed by property that
>already had been developed into homes, so it wasn't as if he was developing
in a wilderness.  After the Supreme Court decision, the state of South Carolina
>was forced to buy his land from him.
 
         According to the New York Times (8/23/92) Lucas bought his 2
lots in 1986 (for $975000) and the law prohibiting all development on
them was passed in 1988.  (The law was later amended to allow owners
to apply for exemptions.)
         Dean Alaska posted:
>While I am not familiar with the first two cases, the third will require
>the state to spend millions to provide flood protection that will be lost
>due to this decision.  Thats real money they will have to spend.  And this
>developer was not planning on putting a retirement cottage on this
>property.  He wants to put a hotel there.  Maybe you thinkg the people
>(thats right, the people, not the govt) should have to pay this guy not
>to take natural protection from the rest of us, but I think he should
>have to pay for actions which will eventually cost taxpayers a lot of
>money.
 
         I believe this account is basically completely wrong.
         The Supreme Court did not rule Lucas was entitled to
compensation.  It ruled the South Carolina courts had used the wrong
standard in denying him compensation and sent his case back to South
Carolina for reconsideration.  I believe the decision specifically
noted that you do not have a right to put others at risk.
         Does anyone have a reference to an authoritative account of
this case and how it was finally resolved (if indeed it has been)?
                          James B. Shearer
=========================================================================
Date: 9 January 1995, 20:32:31 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: Tobis vrs Puchalsky
 
         Michael Tobis posted:
>It is not clear to me that Mr Puchalsky has listened to what I said, since
>he consistently summarizes my position in ways that are not only inaccurate
>but offensive. I am not opposed to democracy. I am in favor of competence, and
>I'm looking for ways to make democracy and competence coexist.
         Then in another post Michael Tobis said:
>If Mr Puchalsky proposes anything at all instead of rational risk assessment,
>it is direct democracy. Apparently on any given topic he prefers the opinion
of people who have not thought about it much to that of people who have thought
>about it. Thus regulations are fine, but regulatory bodies are suspect, and
>lobbying is to be preferred to study by professionals, but direct voting
>is best of all.
 
         This looks suspiciously like an inaccurate and offensive
summary of Puchalsky's position to me.
         I believe Puchalsky has indicated in previous posts that he is
reasonably satisfied with the present system (which is certainly not
direct democracy).
         If you are not satisfied with the current system perhaps you
should make some specific suggestions for improving it.  I would not
say the present system is perfect but it does have some virtues which
you do not seem to adequately appreciate.  I doubt your ability to
come up with a significantly superior system.
         Michael Tobis goes on:
>Probably California is the place which is furthest along in direct democracy,
>and one notes the precipitous decline in the relative economic position of
>California versus the other states as some evidence to bring to bear on this
>subject.
 
         Your advocacy of rational risk assessment would be more
convincing if you avoided irrational arguments like this.  The initi-
ative process has been in place in California for many years.  Many
Western states have similar provisions.  So perhaps you should look
elsewhere for an explanation of California's recent problems (which
I will predict will prove to be temporary even without elimination
of citizen initiatives).
                          James B. Shearer
=========================================================================
Date: 9 January 1995, 21:29:20 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: scientific culture
 
         Michael Tobis posted:
In general, practicing scientists are motivated not in whole but in significant
>part by truth. In disciplines with effective evaluation methods they rarely
>stray very far from monotonic improvement in understanding, at least in the
>aggregate.
>
>To suggest that the discipline is far from the truth because of bias is
>not as insulting as to claim that it is far from the truth because of
>dishonesty. It does, however, implicitly claim that the discipline is
>insufficiently mature as to have significant constraints on fashionable
>opinion. This is only insulting insofar as it is untrue.
>
>I would claim that in the case of climatology, the maturity exists in the
>scientific sense, that grossly unrealistic claims are excluded by the usual
>mechanisms of science.
 
         Well it appears to me that the evaluation methods that can
be applied to climatology are quite weak compared to other scientific
disciplines which are experimentally based.
         It is my understanding that to the extent current theories
can be tested they do not do very well.  For example they fail to
reproduce the present climate from first principles.
         Michael Tobis:
>If current understanding fails it will be because some real unidentified
process has not been accounted for. It will not be because the identified
>processes have been skewed to give a particular result, because the science
>has attained a state where there is very little leeway for bias to enter
>into the logic except through refutable error.
 
         It is my understanding that there are known identified
processes which are known to be important which the current models
do not even attempt to account for.
                    James B. Shearer
=========================================================================
Date: 10 January 1995, 19:06:35 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: discounting the future
 
         Mike Friedman posted:
> For example, let's say that a new process was discovered that saved
> $1,000 off the cost of building a car.  Assuming that there is no
> monopoly enforced by patents or other means, who benefits?  Well, the
> car manufacturers can't keep the $1,000 because they are in a
> perpetual price war, so the cost of cars goes down by a bit more than
> $1,000.  Car buyers benefit.  Since more cars are sold, auto workers
         Rich Puchalsky asked:
>BTW, how does the cost of the car go down by _more_ than the amount
>saved, over the long run?
         Mike Friedman replied:
>ARRGGH!!!  Please give me someone who knows economics!!!!!
>
>Because car manufacturers make a profit on their costs.  The more it
>costs to make a car, the higher the profit.  If you cut the cost of
>building a car you also cut the profit.  If your profit averages 25%
>then a car that costs $9,000 to make sells for $12,000.  When that
>drops to $8,000 then the price will drop to about $10,660.
 
         This answer is completely wrong.  Car manufacturers make a
profit on their invested capital (the normal profit represents the
cost of capital).  If your new process requires the same amount of
capital the profit should remain (in the long run) $3000 per car.
                          James B. Shearer
=========================================================================
Date: 13 January 1995, 18:10:25 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: cost benefit analysis (was discounting the future)
 
         Rich Puchalsky posted:
>A polluting industry is located in the Third World.  Its ability to
>dump pollution indiscriminately represents a significant externality
>that it doesn't have to pay for; but one that causes statistical deaths
>to the people in its neighborhood.  For the sake of argument, let's say
>that almost all of the wealth generated by this facility goes into the
>pockets of the investors, who live halfway across the world.  The neighbors
>see no benefit from this facility that is anywhere commensurable to the
>hazards they undergo.  Meanwhile, the facility owners claim complacently
>that the economic benefits to them outweigh the cost that "society" is
>paying.
 
         I don't quite see the point of this example.  This situation
will only arise when the third world government is corrupt and/or
incompetent.  While it is unfortunate that some third world governments
are corrupt and/or incompetent I don't see how this can blamed on cost
benefit analysis.  What are you suggesting be done about this?  Do
you advocate a return to a colonial system for some countries?
         Rich Puchalsky also posted:
>Now, you will undoubtedly claim that figuring out which regulations to
>have is why we should do cost/benefit in the first place.  Unfortunately,
>cost/benefit would allow the unjust situation I describe above because
>it treats both costs and benefits as accruing to the same nebulous party.
         Expecting cost/benefit analysis to eliminate all injustice
in the world seems a bit unreasonable.  In this case however cost
benefit analysis would tell the third world country not to allow the
situation described to exist.
         I agree that it is unreasonable to expect people to support
actions that harm them materially just because cost benefit analysis
indicates a benefit to society as a whole.  However note this cuts
both ways (industries will oppose measures which harm them even if
cost benefit analysis shows they are in the public interest).
         It is sensible for society to be structured as much as
practical to prefer actions which do not harm anybody materially.
                    James B. Shearer
=========================================================================
Date: 13 January 1995, 20:56:56 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: scientific culture
 
         I have been expressing my view that there is a lot that
climatologists don't know and that this means that global warming
predictions are a bit shaky.  Before responding to the criticism of my
view, let me say that the main reason that I do not support drastic
measures to limit CO2 emissions is not that I question that these
emissions are likely to cause a warmer climate but that I am unconvinced
that costs of a warmer climate exceed the costs of avoiding it.
         I had posted:
>         Well it appears to me that the evaluation methods that can
>be applied to climatology are quite weak compared to other scientific
>disciplines which are experimentally based.
         Dave Halliwell responded:
>   Examples? This is a pretty sweeping statement. What evaluation methods
>are you thinking about, and what are their weaknesses?
 
         Science progresses by testing theories against experiments and
observations of the natural world.  Climatologists have essentially no
ability to experiment and a very limited set of observations.  While
climatologists may do the best they can considering these limitations
it is unreasonable to contend that their theories are as firmly based
as those in other scientific fields.
         I also posted:
>         It is my understanding that to the extent current theories
>can be tested they do not do very well.  For example they fail to
>reproduce the present climate from first principles.
         Dave Halliwell responded:
>  "Fail to reproduce"? What do you mean by "first principles"? Do you
>consider basic radiation transfer and geophysical fluid dynamics to be
>"first principles"? Do you consider models that simulate geographical
>distributions of temperature, precipitation, radiation, pressure, and
>atmospheric circulation patterns that show strong similarity to real
>patterns as "failing to reproduce"?
 
         See the article "Climate Modeling's Fudge Factor Comes Under
Fire" (Science, Sept. 9, 1994, p.1528) which I have cited before in
this group.  The second sentence in this article reads:  "Although
models of how the ocean and atmosphere interact are meant to forecast
the greenhouse warming of the next century, when left to their own
devices they can't even get today's climate right."
         Michael Tobis posted:
>I find James Shearer's offhand dismissals of climatology disconcerting.
>(I was going to say something stronger, but my mistakes here of late
>have taught me some lessons...) They are hard to reconcile with the
>astonishing seriousness and competence I have observed in its leading
>practitioners. To evaluate whether I have somehow been utterly
>bamboozled, I'd like to establish exactly how much Mr Shearer
>actually knows about the field, in the hopes of seeing some more
>specific criticisms.
 
         I don't mean to imply that I have believe climatology is
totally worthless.  I do believe the field's ability to make useful
predictions as a guide to public policy is and will remain limited.
I consider my knowledge of the field to be that of the intelligent
layman.
         Michael Tobis continued:
>Let me start by listing some points that I believe to be indisputable
>in the face of the evidence. Does Shearer dispute any of these?
>
>1) The surface temperature of a planet with an atmosphere exceeds that
>of an otherwise identical planet without an atmosphere in a way which,
>given albedo and atmospheric composition, can be predicted with precision.
 
         I am not sure exactly what you mean here but I think I disagree
with the "predicted with precision" part.  How do you account for the
different temperture during ice ages?  What about the differences in
projected response to CO2 forcing?
         Michael Tobis continued:
>6) There are no immediate constraints outside of policy decisions to continuing
>exponential growth in direct anthropogenic forcing of CO2 concentrations,
>and considerable economic incentives to continuing that growth.
>(Halocarbons appear to have be controlled, while the others are less well
>understood.)
>
>CO2 concentrations, essentially constant on historical time
>scales before the industrial revolution, will in the absence of concerted
>action exceed double preindustrial background values in the next fifty years,
>and will very likely exceed quadruple background values before the end
>of the next century.
 
         This is one scenario.  I don't think it is indisputable.  In
any case projecting CO2 emissions is outside the field of climatology.
I don't know about the carbon cycle, is studying that considered
climatology or something else.
         Michael Tobis again:
>9) The earth has warmed in the past century at a rate that is at the highest
>limits of natural variability, but is somewhat less than the best estimated
>sensitivity to CO2. Accounting for cooling by anthropogenic aerosols
>accounts for much of the difference. The agreement is not conclusive
>in a statistical sense, but the global average historical record is much more
>consistent with theoretical predictions than inconsistent with it.
 
         I don't accept that the limits of natural variability are
accurately known.  What do you believe to be the limits of natural
variability and on what basis?
         Michael Tobis:
>None of the above statements requires the fluid dynamical computational models
>that are so broadly criticized. All are based on observation or physical
>reasoning. As far as I know, all of them are considered settled.
>
>Does James Shearer concede these points as being based in good science?
 
         I accept that CO2 emissions will probably produce a warmer
climate.  I consider words like "indisputable" or "settled" to be
overstatements.
         Michael Tobis:
>If so, perhaps we can proceed from there to reasonable uses for the models,
>sensible approaches to resolving open questions, reasonable estimates of
>impacts, policy implications, and/or sensible responses.
 
         As I indicated above I am unconvinced that a warmer earth would
be extremely damaging.  I don't have the belief that human alteration
of the natural environment is evil that some posters to this group
appear to have.
                          James B. Shearer
=========================================================================
Date: 18 January 1995, 20:03:11 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: discounting the future
 
         Mike Friedman posted:
>Hmm.  Can someone from sci.econ please comment?  Don't costs have an
>impact on capital requirements?  Doesn't this tie in with inventory
>carrying costs?
 
         A reduction in capital requirements is a reduction in costs,
it is not fair to count it twice.
                           James B. Shearer
=========================================================================
Date: 18 January 1995, 21:21:27 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: scientific culture
 
         Here I reply to critics of my view that climatologists
predictions should not be considered totally reliable.
         I had posted:
>         See the article "Climate Modeling's Fudge Factor Comes Under
>Fire" (Science, Sept. 9, 1994, p.1528) which I have cited before in
>this group.  The second sentence in this article reads:  "Although
>models of how the ocean and atmosphere interact are meant to forecast
>the greenhouse warming of the next century, when left to their own
>devices they can't even get today's climate right."
         Len Evens responded:
>You should go back and read that article again.   What it referred to is
>the newer dynamic models which have a tendency to drift---I think for
>reasons having to do with numerical instabilities rather than underlying
>physics, but I am not sure.    The fudge factor is put in to balance
>that out and produce a stable model of current climate which is then
>perturbed.   The Science article quoted work done at MIT in which
>related models appeared to produce faulty predictions when this was
>done.   The article also quoted climate modelers like Manabe who
>questioned the relevance of the MIT work for actual models.   The article
>also remarked that
>the issue becomes less relevant as the models improve.  Indeed, this
>is a good example of just why additional funding for this work is
>justified.   It will support further critical work testing the validity
>of these coupled atmospheric ocean dynamic models as well as improved
>models.
>
>It should also be noted that the article mentioned that earlier less
>sophisticated models which did not involve the oceans as directly
>and only dealt with the equilibrium rather than the dynamics did not
>have this problem.   Finally, what Michael Tobis was talking about
>is the explanation for the basic difference---about 33 deg K---between
>the Earth with a greenhouse effect and the Earth without one.   That
>does not require complicated computer models.
 
         My interpretation of the article is that older models modeled
the atmosphere only with the oceans given.  These models could not
move too far from the present climate because the oceans did not change.
However these "models could not provide a clear window on the future".
(Since if the climate changed the oceans would change.)  However once
the oceans were included in the models the present climate was not in
equilibrium (indicating of course a problem in the models) and when the
models were run with the present climate as an initial condition they
would move away to a model equilibrium diferent from our climate.  Most
modelers perturb the models to force the present climate into equili-
brium a procedure of extremely dubious validity.  ("The implication
that flux adjustments disguise-but may not correct-a model's underlying
defects won't surprise other climate modelers.  Most agree with Warren
Washington of ... (NCAR) ... who says the practice "makes your model
look better than it really is; it could cover up deficiencies."").
The article does include near the end a quote from Manabe (probably
thown in for "balance") defending the practice as justified in "some
cases" as compensating for "computational bias".  (I am unconvinced
that two wrongs will make a right even so.)  However the article taken
as a whole clearly indicates that current climate models have serious
problems reproducing the current climate which is as I said originally
is one of the few ways in which they can be empirically tested.
         I posted:
>         Michael Tobis continued:
>>6) There are no immediate constraints outside of policy decisions to continuing
>>exponential growth in direct anthropogenic forcing of CO2 concentrations,
>>and considerable economic incentives to continuing that growth.
>>(Halocarbons appear to have be controlled, while the others are less well
>>understood.)
>>
>>CO2 concentrations, essentially constant on historical time
>>scales before the industrial revolution, will in the absence of concerted
>>action exceed double preindustrial background values in the next fifty years,
>>and will very likely exceed quadruple background values before the end
>>of the next century.
>
>         This is one scenario.  I don't think it is indisputable.  In
>any case projecting CO2 emissions is outside the field of climatology.
>I don't know about the carbon cycle, is studying that considered
>climatology or something else.
         Len Evens responded:
>Study of the Carbon cycle is certainly closely related to climatology.
>For example, two leading figures in the IPCC are Bert Bolin and
>J. T. Houghton.  The first is one of the world's leading experts on
>the Carbon cycle, and the second is a prominent climatologist.
 
>Since you really don't have any qualifications
>in the study of the Carbon cycle, it would seem of little relevance
>that you don't think the scenario is indisputable.  You've made
>a major criticism in the past about Carbon cylce modeling which I
>thought was based on a gross oversimplification and a basic lack of
>understanding of how those models work.   However, what you or I
>think is basically irrelevant here.   Unless you are willing
>to put your ideas to the test by submitting a paper to a peer
>reviewed scientific journal, you would be prudent to accept the
>conclusions of people like Bolin.   And they come to conclusions
>like that described by Michael above.
 
         Consider what two recent surveys of the carbon cycle said.
         "The Global Carbon Dioxide Budget" by Eric T. Sundquist
Science, Vol 259, Feb 1993, p934-941.  On page 939 I read
         "Forecasts of future CO2 emissions vary widely, depending on
assumptions about such factors as population, economic growth, energy
supplies and technologies and land use.  A recent summary listed cul-
mulative emissions estimates for the years 1990 to 2100 ranging from
about 700 to 2100 Gt of C as CO2 (60).  Such diverse estimates lead to
comparably uncertain scenarios for future atmospheric CO2 concentra-
tions (38).  Estimates of future CO2 emissions are therefore viewed by
many as the major uncertainty in anticipating atmospheric trends.  On
the other hand much more confidence is generally placed in the ability
of the C cycle models to forecast atmospheric CO2 for a given emissions
scenario.  For example, the recent update report of the <IPCC> stated
"Future atmospheric CO2 concentrations resulting from given emissions
scenarios may be estimated by assuming that the same fraction remained
airborne as has been observed during the last decade, that is 46+-7%"
(38, p. 35).  Analysis of the deglacial and historical to modern CO2
budgets suggests that the uncertainties in net emissions forecasts
should be even greater and that confidence in modeled CO2 predictions
should be much less."
         "Atmospheric carbon dioxide and the ocean" by U. Siegenthaler
& J.L. Sarmiento, Nature, Vol 365, 9 September 1993, p119-125(p124).
         "The main uncertainty about the future growth of atmospheric
CO2 levels stems from the development of anthropogenic emissions, which
depend on political and socio-economic factors and cannot be predicted.
For instance there is a three-fold difference in the cumulative
emissions 1990-2100 between the highest (2,127 Gt C) and the lowest
(742 Gt C) of the siz 1992 IPCC scenarios, all based on plausible
assumptions about future developments.  ...  For a given emissions
scenario, the model predictions of future atmospheric CO2 depend on
assumptions about the "missing sink".  Calculations based on best
estimates for the sources and sinks (implying an unbalanced pertuba-
tions budget) yield higher future concentrations based on a balanced
budget achieved by assuming an extra (biospheric or ocean) sink; the
difference may be considerable, especially for low emission scenarios."
         I don't see any words like "indisputable" here.  (Nor did I
see Bert Bolin in the references (152 total) a little surprising if
he is really one of the leading experts.)
         Len Evens in another post:
>           ...   Shearer seems to suggest that the degree of confidence
>we can have in these predictions is low enough that it is rational
>to ignore them.   I don't think he has made a compelling case, but
>more important neither he nor I is really in a position to judge
>that.   That is why I suggest he enter arena with his ideas if he
>thinks they have merit.
 
         As should be obvious I do not believe that all but indis-
putable predictions should be ignored.  I do believe the future is
uncertain and that it is irrational to pretend otherwise.
         As for publishing, why should I try to publish what is already
well known to experts in the field.
         Rich Puchalsky asked:
>So what do critics like Shearer want as a response?  Can he suggest anything
>other than that the experts either direct him to publish in a journal, tell
>him to read the IPCC report, browbeat him, or ignore him?  Evens has simply
>chosen the first option.
 
         Evens is not an expert.  I believe my quarrel is not with the
experts but the misrepresentation of expert opinion in this group.
         May I suggest another option, careful consideration of my
points followed (of course) by acknowledgement of their validity.
         Dave Halliwell posted:
>>         I have been expressing my view that there is a lot that
>>climatologists don't know and that this means that global warming
>>predictions are a bit shaky.  Before responding to the criticism of my
>>view, let me say that the main reason that I do not support drastic
>>measures to limit CO2 emissions is not that I question that these
>>emissions are likely to cause a warmer climate but that I am unconvinced
>>that costs of a warmer climate exceed the costs of avoiding it.
>
>   This seems to be a rather odd interpretation of a series of posts that
>question the accuracy of climate models and the people that develop them.
>Which part of the climate models do you feel is responsible for the cost
>assessments that you disagree with?
 
         The only cost assessments that I am aware of were mentioned
some time ago in this group.  They were by economists and concluded
that the cost would be minor.
         Dave Halliwell posted:
>>         Science progresses by testing theories against experiments and
>>observations of the natural world.  Climatologists have essentially no
>>ability to experiment and a very limited set of observations.
>
>   I presume you exclude the "non"-experiments regarding thinks like gas
>laws, thermodynamics, radiation transfer, evaporation processes, cloud
>physics, and every other aspect of weather dynamics than _can_ be
>experimented with and is part of our understanding of the atmosphere? By
>claiming that the only way of testing climate models is to test them as a
>whole, then you are demonstrating a rather naive view of scientific
>evaluation.
>
>   Beginner-type question: how much is there in common between climate
>models and numerical weather prediction (NWP) models? (There is an
>intentionally-misleading aspect to this question.) Would you put NWP
>models in the same category regarding lack of suitable testing?
 
         No I would not put NWP models in the same category because
since they predict on a much shorter timescale (and in some cases for
smaller areas) the observational record is in effect much larger.
         Dave Halliwell:
>> While
>>climatologists may do the best they can considering these limitations
>>it is unreasonable to contend that their theories are as firmly based
>>as those in other scientific fields.
>
>   Again, this is a rather sweeping generalization that I disagree with.
>You could easily pick a specific field in physics for which your
>statement is true, but I can just as easily pick another field for which
>it is not. The sweeping generalization is basically without substance.
 
         There is a spectrum and climatology is on the shaky end.
         Dave Halliwell posted
>>         See the article "Climate Modeling's Fudge Factor Comes Under
>>Fire" (Science, Sept. 9, 1994, p.1528) which I have cited before in
>>this group.  The second sentence in this article reads:  "Although
>>models of how the ocean and atmosphere interact are meant to forecast
>>the greenhouse warming of the next century, when left to their own
>>devices they can't even get today's climate right."
>
>  This issue is not in our library at the moment; I think it is being
>bound. Len Evens has posted how the quote is taken out of context. All I
>can say is that basing your sweeping generalization on a _single_ out-
>of-context quote doesn't do much for your credibility.
>
>   If I were to give you a global map of mean annual temperature, what do
>you think the chances are that you would correctly identify it as being
>either model output or actual observations? Your claim implies that you
>would easily identify the model output as being unreasonable.
 
         The quote is not out of contex.  The entire first paragraph is:
"In climate modeling, nearly everyone cheats a little.  Although models
of how the ocean and atmosphere interact are meant to forecast the
greenhouse warming of the next century, when left to their own devices
they can't even get today's climate right.  So researchers have tidied
them up by "adjusting" the amount of heat and moisture flowing between
a model's atmosphere and ocean until it yields something like the
present climate.  But, as usual, cutting corners can eventually catch up
with you."
         My claim does not imply that I could identify the model output
as unreasonable but that an expert could (unless the model is fudged
to disguise its problems).
         Dave Halliwell posted:
>   I consider your knowledge of the field to be that of a rather naive
>and not-very-well-informed layman. I agree with Michael that it would be
>nice to see more specific criticisms from you, in the hope of
>establishing the limits of your knowledge. ...
 
         I feel confident that I know more about climatology than 95%
of the general public.  Do you disagree?
         I have repeatedly given specific criticisms.  It appears that
some members of this group have such an emotional stake in global
warming that they are unable to respond rationally to even mild critic-
ism.
         Dave Halliwell responding to me responding to Michael Tobis:
>>>9) The earth has warmed in the past century at a rate that is at the highest
>>>limits of natural variability, but is somewhat less than the best estimated
>>>sensitivity to CO2. Accounting for cooling by anthropogenic aerosols
>>>accounts for much of the difference. The agreement is not conclusive
>>in a statistical sense, but the global average historical record is much more
>>>consistent with theoretical predictions than inconsistent with it.
>
>>         I don't accept that the limits of natural variability are
>>accurately known.  What do you believe to be the limits of natural
>>variability and on what basis?
>
>   This question can be approach over an almost infinite range of
>temporal and spatial scales. Michael's statement specifies the last 100
>years as a special case over a longer period of time. Let's take the last
>100,000 years as an appropriate starting point. For that period, we can
>base our knowledge of climate on:
>
>   - instrumental records. (100-300 years)
>   - anecdotal descriptions (somewhat longer)
>   - dendrochronology (well over 1000 years)
>   - ice cores (entire period)
>   - palynology (entire period)
>   - geological evidence (entire period)
>
>   Each of these has its own character as far as spatial and temporal
>resolution is concerned, and tells us different things about climate
>(e.g. temperature, precipitation, etc.)
>
>   Would you care to make a more specific statement regarding what it is
>that causes you to reject any or all of these records as giving
>sufficient evidence of past variability?
 
         Sure, they do not have sufficent resolution for an accurate
determination of the limits of natural variability over a century.  I
note neither you nor Tobis seem willing to tell me what the limits of
natural variability are.
         Dave Halliwell replying to me:
>>         As I indicated above I am unconvinced that a warmer earth would
>>be extremely damaging.  I don't have the belief that human alteration
>>of the natural environment is evil that some posters to this group
>>appear to have.
>
>   Strawman alert again. The question of "damage" is a completely
>different issue from the assessment of climate model reliability. What
>you have just said is essentially the same as saying "I don't think that
>pulling the trigger will cause a bullet to fire because I don't think
>anyone will be hurt by the bullet".
 
         Actually it isn't.
         What I was trying to point out is that I am willing to grant
that there is say a 50% chance that the models are correct.  While the
difference between a 50% chance and a 100% chance may be important
psychologically and politically, objectively it probably does not make
too much difference in determing a rational policy response.
                       James B. Shearer
=========================================================================
Date: 23 January 1995, 21:41:54 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: scientific culture
 
         More arguing about climatology (my reply to Tobis).
         Michael Tobis posted:
>I enjoy arguing with Mr Shearer - he is the most interesting opponent on
>the greenhouse question I've encountered. (Dale Bass, to my disappointment,
>did a remarkably bad job, though he did obfuscate matters effectively.)
>However, I would recommend he find a consistent basis for his objections.
>
>So far, he's started by implying that climatologists were dishonest. Then
>he backed off to the position that they were unconsciously biased, and
>then he implied that, well, perhaps they weren't so biased but they weren't
>as competent as they claimed. At last count, he seemed to be claiming that
>climatology was reasonably sound, but that Halliwell and Tobis were, for
>reasons of their own (dishonesty? bias? incompetence?) misrepresenting
>the state of the art. At least, that's how I read it.
 
         Well of course I read it differently.
         I started by observing climatologists have an interest in
promoting concern about global warming.  I believe you eventually
conceded this but said it didn't matter because errors would be quickly
exposed by the mechanisms of science.  Ok so far?
         I then claimed it is harder (as compared to many other areas
of science) to test climatological theories against reality because
of an inability to experiment and a limited set of observations.  I
further claimed the global warming models fail one of the few ways in
which they can be tested in that they do not reproduce our present
climate.  Do you dispute this?
         You then diverted the discussion into how much I know about
climatology which I will concede is not everything.  However this is
irrelevant to the points above.  Nobody can have a detailed knowledge
of all scientific and technical fields.  Yet people must still make
decisions.  While one should be aware of relevant expert opinion it
is important to realize that experts are not infallible.
         My comment about "misrepresentation" was aimed more at Evens
and Halliwell, neither of whom has been willing to concede that the
Science article that I cited supports my position, than you.
         Some detailed responses:
         Michael Tobis posted:
>So your statement should not say that "current theories" don't represent
>climate well. Physical laws and mathematical theorems have as much
>validity in geophysics as anywhere else. You might fairly have said that
>"while atmospheric and oceanic models represent climate well, they have
>not yet been coupled with great success without the use of fudge factors".
 
         How about, "Confidence in the ability of the most general
climate models to accurately predict the effects of CO2 forcing is
decreased by the fact that these models incorrectly predict the effects
of zero forcing (ie they do not reproduce the current climate)"?
(Of course we also have "Confidence in the ability of less general
models to accurately predict the effects of CO2 forcing is decreased
by the fact that these models omit potentially important factors.")
         Michael Tobis posted:
>I claim that familiarity with the culture and the work shows that the
>constraints of physical science effectively are at work in climatology.
>Thus, climatology is susceptible to refutation, and thus the options
>for bias to enter are rather small.
 
         Would you accept the same claim with "climatology" replaced
with "studies of racial differences in intelligence"?  If not, what
is the distinction?
         Michael Tobis posted:
>But *given* albedo and atmospheric composition and solar forcing, there
>is no difficulty in calculating equilibrium surface temperature, which is
>all I asserted.
 
         If all you are trying to claim here, is that if everything
else stayed the same adding CO2 would increase the temperature then I
accept this.  The difficulty is that everything will not stay the
same because of a complex and poorly understand set of interactions.
In particular your "given" albedo, atmospheric composition and solar
forcing may not be in equilibrium which certainly does present a
difficulty in calculating an equilibrium surface tempature.
 
         Michael Tobis:
>This is debatable, but hair splitting. However, if you don't choose to
>put it under the rubric of climatology, it is irrelevant to your claim
>that climatology is fatally tinged by bias.
 
         I of course have never made such a claim.  Try replacing
"fatally" with "possibly".
         Michael Tobis first posted:
>9) The earth has warmed in the past century at a rate that is at the highest
>limits of natural variability, but is somewhat less than the best estimated
>sensitivity to CO2. Accounting for cooling by anthropogenic aerosols
>accounts for much of the difference. The agreement is not conclusive
>in a statistical sense, but the global average historical record is much more
>consistent with theoretical predictions than inconsistent with it.
         My response was:
>          I don't accept that the limits of natural variability are
> accurately known.  What do you believe to be the limits of natural
> variability and on what basis?
         Michael Tobis then posted:
>"Limits" was a poor choice of word. The observed change has been rapid
>compared to the paleoclimate record.
 
         Ok how about quantifying "rapid".  Based on the record, what
proportion of the time could one expect the observed change or larger?
What are the error bars on your number?  For that matter what are the
error bars on the amount of warming in the last century?
                    James B. Shearer
=========================================================================
Date: 26 January 1995, 22:13:16 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: scientific culture
 
         Still more on climate models:
         Michael Tobis posted:
>I think you don't understand the legitimate purposes of computer modelling,
>and how they fit in with the public policy questions. This is hardly
>surprising, because there are many people who do the modelling who
>can't formulate it very well for public consumption.
 
         It is true I am extremely skeptical of large complicated
computer models of poorly understood systems.  I could elaborate on
why, but I will refrain as Michael Tobis continued:
>Because of this, and despite the fact that I think there are several good
>reasons for continuing the modelling efforts, I prefer to leave the models
>entirely out of the discussion for the time being.
 
         If we leave the models out of the discussion, I am not sure
what there is to discuss.  I accept that simple unrealistic models
show warming.
         Michael Tobis posted (replying to me):
>:          My comment about "misrepresentation" was aimed more at Evens
>: and Halliwell, neither of whom has been willing to concede that the
>: Science article that I cited supports my position, than you.
>
>I agree with something David said in his heated response - the article
>refers to a single class of model among a whole spectrum. Further, it
>does not refer to a complete failure of the models - they do produce
>something resembling the real climate without the "flux corrections" -
>and this is how one of the critics quoted in the article does run them.
>Finally, the flux corrections themselves provide some information on how
>to better parameterize precipitation and evaporation, which after all do
>happen on a scale too small to be resolved by global 3-D coupled dynamical
>models.
>
>So I don't concede that the article offers as much support to your position
>as you seem to think. However, again I'd like to see how far we can get
>without reference to fluid dynamical computer models. So I'm not conceding
>the point. I'm trying to establish what a rational position would be, for
>purposes of argument, if the models were indeed worthless.
 
         The class of models is those which try to model the entire
system.  This is the most stringent test of how well we understand the
the system.  A purely empirical model could reproduce the present
climate without understanding it.  However it would offer no guide for
predicting the effects of CO2 forcing.  Other models on the spectrum
are semi-empirical and for this reason could err in predictions if the
empirical part of the system behaved in an unexpected way.
         The article describes the model climates as "quite unreal".
Btw if the climate after global warming "resembles" the current
climate does that mean the difference is unimportant?
         Do you concede that the article offers some support for my
position?
         Michael Tobis posted:
>What baffles me is the next step that you seem about to take - which is
>to presume that, in the light of these uncertainties, the rational course
>is to act as if the global response were *the smallest non-refutable
>response*. Why should we presume on our luck that much? Do you always
>put your whole paycheck on a 35:1 payoff on a roulette wheel before you
>take it home?
 
         It would simplify these discussions if people would argue
with what I actually say not with what they think I am about to say.
People may use scientific uncertainty to justify irrational acts,
however this is no excuse for pretending scientific uncertainty is
not present when in fact it is.
         Michael Tobis earlier posted:
>"Limits" was a poor choice of word. The observed change has been rapid
>compared to the paleoclimate record.
         I asked:
>          Ok how about quantifying "rapid".  Based on the record, what
> proportion of the time could one expect the observed change or larger?
> What are the error bars on your number?
         Michael Tobis responded:
>Good questions, which I can't answer, and for which I'm not sure there is
>a good answer. My assertion is based largely on Wigley & Raper, 1990,
>"Natural Variability of the Climate System and Detection of the Greenhouse
>Effect", _Nature_ vol 344 pp 324-327. They claim that the observed
>variation is outside the limits of unforced autovariation in the ocean
>system (the only component of the climate with internal variability on
>the time scale in question), but claim that it's premature to identify
>the forcing with the greenhouse forcing. I suppose I may as well admit that
>it was a modelling study, though.a
 
         This claim is not based on the paleoclimate record and so does
not support your claim.  Btw is it known that the solar "constant" does
not vary on hundred year time scales?  You appear to assuming that it
doesn't.
         I also asked:
> For that matter what are the
> error bars on the amount of warming in the last century?
         Michael Tobis responded:
>It's about 0.5 K, +/- 0.25.
 
         Well if there is that much uncertainty in the amount of
warming in the last century (for which the data must be far superior to
that for earlier periods) it would appear that the uncertainty in the
average variability based on say the last million years must be so large
that the only thing you can safely say is that the temperature change
in the last century is in the same direction as predicted by global
warming models.  Do you disagree?
                          James B. Shearer
=========================================================================
Date: 30 January 1995, 21:31:23 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: scientific culture
 
         More climatology (reply to Halliwell):
         Dave Halliwell asked:
>   Beginner-type question: how much is there in common between climate
>models and numerical weather prediction (NWP) models? (There is an
>intentionally-misleading aspect to this question.) Would you put NWP
>models in the same category regarding lack of suitable testing?
         I replied
>         No I would not put NWP models in the same category because
>since they predict on a much shorter timescale (and in some cases for
>smaller areas) the observational record is in effect much larger.
         Dave Halliwell replied:
>    And you just stepped beautifully into the misleading part of the
>question. I note that you didn't even *try* to answer the first question.
>As for the second question, since global atmsopheric GCMs are strongly
>related to NWP models, they benefit from the validation available for NWP
>models. A great deal of the basic physics and numerical techniques are
>similar. GCMs are, in a crude sense, weather models that are used to
>accumulate climate statistics, just as observations of climate are an
>accumulation of observations of weather. If you can reliably model
>weather, why can't the statistics give a good indication of climate? There
>*are* differences between GCMs and NWP models, but I'll bet that you can't
>describe even one in much detail.
 
         This is a bunch of nonsense.  Since NWP models only attempt
to predict the weather for a few days they do not have to accurately
model parts of the system that change slowly .  GCMs must include
many interactions that NWP models can neglect.
         Michael Tobis first posted:
>9) The earth has warmed in the past century at a rate that is at the highest
>limits of natural variability, but is somewhat less than the best estimated
>sensitivity to CO2. Accounting for cooling by anthropogenic aerosols
>accounts for much of the difference. The agreement is not conclusive
>in a statistical sense, but the global average historical record is much more
>consistent with theoretical predictions than inconsistent with it.
         My response was:
>          I don't accept that the limits of natural variability are
> accurately known.  What do you believe to be the limits of natural
> variability and on what basis?
         Michael Tobis then posted:
>"Limits" was a poor choice of word. The observed change has been rapid
>compared to the paleoclimate record.
         I asked:
>         Ok how about quantifying "rapid".
         Dave Halliwell answered:
>  Changes of several degrees over a period of roughly a century.
 
         Fine, then the change in the last century was not "rapid'
as Tobis claimed.  You should continue this argument with Tobis.
         I asked:
>  Based on the record, what
>proportion of the time could one expect the observed change or larger?
         Dave Halliwell answered:
>   Based on instrumental and paleoclimate records, never in the last
>100,000 years at least. The Younger-Dryas is probably the closest event
>in the record to the rapidity of change we might expect.
 
         Please pay attention.  The question was about the "observed
change" in the last century, not the "change we might expect".
                    James B. Shearer
=========================================================================
Date: 31 January 1995, 22:01:34 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
 
         Michael Tobis posted:
: >I think you don't understand the legitimate purposes of computer
: >modelling,and how they fit in with the public policy questions.
         I responded:
:          It is true I am extremely skeptical of large complicated
: computer models of poorly understood systems. ...
         CDChase then posted:
>++ Has anyone else noticed that "large, complicated, computer models" are
>used everywhere including for environmental work? Why is this line of
>argument dot used to debunk forecasts in all fields?
 
         It is, however this is sci.environment.
         CDChase continued:
>"large, complicated, computer models"are used to predict most aspects of
>the future for planning purposes. I'm sure the federal budget forecasts
>and all the financial models people are using around Wall Street and to
>try and manipulate financial markets are all based on "large, complicated
>computer models."
 
         Using large complicated computer models to predict the future
is in many cases the modern equivalent of consulting an oracle, an
irrational but psychologically understandable response when people are
required to make important decisions in the presence of uncertainty.
         Federal budget forecasts are notoriously inaccurate (and often
politically biased as well).  I am sure you can find models that will
happily calculate the effect of changing the capital gains tax or the
minimum wage on the US economy in five years.  You would be unwise to
give great weight to the answer.
         Wall Street models may be better because of the continous
feedback and the painful consequences of errors.  Nevertheless Wall
Streeter's are continually coming up with models that work well most
of the time but then fail castrophically when something unexpected
happens.  Remember portfolio insurance or the Granite funds.
         When confronted with the output of a large computer model one
should ask how the model was validated.  If there is no convincing way
to validate the model the results are probably garbage (by which I
mean are no more reliable than the results of simple models).
         CDChase added:
>=) better to use small, simple computer models which have no relationship
>at all to the real world=)
 
         Simple models have the virtue that their flaws tend to be more
obvious which discourages placing excessive weight on their results.
For some reason people will often give greater weight to models that
they do not understand.
                          James B. Shearer
=========================================================================
Date: 7 February 1995, 20:56:40 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
         Michael Tobis asked:
>Finally, I would like to know where you are getting your information. It does
>not appear to arise from the scientific press, much less referred journals.
 
         You don't consider the front pages of Science the "scientific
press"?  Let me point out that the current (1/27/95, p454) Science
contains an article "Darker Clouds Promise Brighter Future for
Climate Models" suggesting current climate models substantially err in
the way they handle clouds (specifically that they underestimate the
amount of incoming solar radiation absorbed by clouds).  Some quotes:
         "As mirrors of the real world, climate models are far from
perfect.  These computer simulations of how solar energy and Earth's
ocean and atmosphere can't even get today's climate entirely right.
And when they're asked to prognosticate the results are even worse:
... These shortcomings are no great surprise, given the number of
climate processes that are poorly understood or totally unknown.
... "This is such a basic thing; it throws a big monkey wrench into
the modeling works".  But unlike most monkey wrenches thrown into
machinery, this one may effect some much-needed repairs. ... "It's
Mother Nature doing something, something we don't understand." ...
"fairly dramatic" effects on climate models ... Modelers are eager to
see what surprises come out of greenhouse simulations with more
realistically absorbant clouds.  "There's no way to guess what this
would do ...""
                          James B. Shearer
Btw:  I have not been challenging your claim that given atmospheric
composition, solar radiation and albedo the surface temperture can
be calculated from first principles.  (Is this in fact your claim?)
I now suspect you have been blowing smoke.  How do you calculate the
effects of clouds from first principles?
=========================================================================
Date: 8 February 1995, 20:03:03 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
         More on computer modeling particulary climate models (replies
to Evens, Tobis and Halliwell):
         Leonard Evens posted:
>                ...   There are circumstances in which a simple model
>is just not accurate enough to tell us what is going to happen, so
>we have no choice but to use more elaborate models.   ...
 
         This is the thinking that leads to trouble.  The alternative
is to accept that we can not tell with certainty what is going to
happen.  An elaborate model should be preferred to a simple model only
when it can be convincingly demonstrated that the elaborate model can
be expected to give better predictions in practice.
         Leonard Evens posted:
>In this respect, climate models are somewhere `in between' .   The
>underlying physical and mathematical principles are pretty well
>understood.  ...
 
         I disagree, the underlying physical principles are not well
understood in any effective sense (saying everything follows from
Schrodinger's equation is not helpful).
         Leonard Evens also posted:
>                                                 ...  Those who
>take the conclusions of these models seriously think that it is
>likely that what they tell us is at least as accurate as what
>simple models tell us, particularly when many different models
>using somewhat different approaches to the areas of uncertainty
>result in similar predictions.
 
         The predictions are not similar.  For CO2 doubling complicated
models predict 1.5-4.5 C, a simple model 1 C.  If the complicated model
predictions are "similar", then they are all similar to the simple
model so where is the value added?
         Michael Tobis asked (concerning econometric models):
>What comparative weight would you give to such models compared to climate
>models?
 
         In general I give low weight to both.  A more specific answer
would depend on exactly what you are trying to predict with the models.
         I posted:
>          Wall Street models may be better because of the continous
> feedback and the painful consequences of errors.  Nevertheless Wall
> Streeter's are continually coming up with models that work well most
> of the time but then fail castrophically when something unexpected
> happens.  Remember portfolio insurance or the Granite funds.
         Michael Tobis responded:
>In the case of climate models, this clearly indicates that the models are
>intrinsically optimistic, rather than alarmist as is often alleged.
 
         An absurd conclusion.  Wall Street models attempt to identify
profitable trades.  A pessimistic model misses some profitable trades.
This is unlikely to cause serious problems.  An optimistic model
identifies some unprofitable trades as profitable a recipe for disaster.
         Michael Tobis posted:
>Agreed. Things like the Club of Rome models have value only for didactic
>purposes, and have essentially no predictive value.
>
>That you believe the same holds for climate models shows yet again that you
>do not know what a climate model is.
 
         What is being shown here is the truth of the proverb about the
mote in your neighbor's eye.
         I posted:
>          Simple models have the virtue that their flaws tend to be more
> obvious which discourages placing excessive weight on their results.
> For some reason people will often give greater weight to models that
> they do not understand.
         Michael Tobis responded:
>In your own case, this is obviously untrue. However, it is a bit irritating
>that you consistently fail to make an effort to improve your understanding
>of the models, while continuing to criticize them.
 
         A belief that one is obligated to understand a model before
criticizing it is dangerously wrong, making incomprehensible models
above reproach.  Btw how much do you actually know about the
macroeconomics models you denigrate.  You certainly haven't shown much
understanding of economics (along with practically everyone else in
this group).
         Michael Tobis posted:
>Climate models have fewer degrees of freedom than the system they simulate,
>and they can be validated in several ways:
>
>1) reproducing the spatial structure of contemporary climate
>
>2) reproducing the seasonal cycle of contemporary climate
>
>3) reproducing paleoclimate conditions
>
>4) predicting the effects of perturbations like Mt Pinatubo
>
>5) reproducing physical conservation principles like mass conservation, energy
>conservation and conservation of angular momentum
>
>and believe it or not
>
>6) reproducing the course of greenhouse gas response to date
>
>In all of these matters, any sensible measurement shows skill above zero,
>though below perfection.
 
         These are all very weak forms of validation.  Strong valida-
tion would be comparing the model answers to the correct answers over
the entire range of conditions the model is suppposed to handle.
         I will also mention that it is very dangerous to use any data
used to develop the model to validate it.  This is shown over and over
by the failure of stock trading (or sports betting) schemes which work
great retrospectively when applied to future events.
         Do you doubt that it is possible to come up with obviously
wacko models which pass all or most of the above tests.  Consider for
example the following model for predicting climate.  Find the year x
in the last 100000 with climate most similar to this year's climate.
Then predict next year's climate will be that of year x+1.  This model
passes the above tests with flying colors.  Are you impressed?  (One
might quibble that the climate records for the last 100000 years are not
good enough to actually implement this model, however since you have
making been claims about how well paleoclimatic conditions are known
this shouldn't bother you.)
         Michael Tobis asked:
>Also, I remain curious how you would design a greenhouse gas policy in the
>absence of any information at all from the models you like to criticize.
 
         I would use simple models.
         Dave Halliwell:
>   Every model must be assessed according to its goals. Models _will_
>leave out things that are not relevant to the problem at hand. You will
>never get a "model of everything", which seems to be what Shearer wants
>to see. My guess is that by the time the newest coupled atmosphere-ocean
>GCMs sort out their drift tendencies and give "accurate" climates, the
>modelling community will have moved on to add yet more factors to the
>climate models, and these current "newest" models will be shifted into the
>"simpler" class. At that stage, Shearer will _still_ be saying "I accept
>that simpler models show warming, but...", and he'll still be arguing
>that something has been left out that should be in there (and that this
>is why the models still show warming when he is sure that they
>shouldn't). [Note: we of course must wait to see whether more detailed
>moselds will or will not show the same warming that current models do.]
 
         What I want is for climate modelers and others to quit over-
selling the reliability of model predictions.
         When the most realistic models give the same results (and when
making them more realistic still does not significantly change the
results) I will give greater weight to their results.  I don't expect
this to happen anytime soon.  (Btw the results could still be wrong
because of some factor no one considered.  There is no getting around
the fact that climatology suffers from a distinct lack of real data.)
         The statement that I am sure the models shouldn't show
warming is the usual Halliwell misrepresentation.  Unlike some
people I am not sure of what the models should show.
                     James B. Shearer
=========================================================================
Date: 9 February 1995, 13:20:30 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: excess profits (was discounting the future)
 
         Mike Friedman posted:
>If the car industry was an oligopoly then there would be excess
>profits.  The whole point of collusion is to generate excess profits.
 
         The excess profits have been captured by the UAW in the form
of excess wages.  Do you deny that the US car industry pays above
market wages or that splitting up the companies would put downward
pressure on these wages?
                          James B. Shearer
=========================================================================
Date: 10 February 1995, 21:40:15 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: climatology
 
         The February 3, 1995 Science also has a global warming article
"Is the World Warming or Not?" (p. 612).  It concludes that it is too
soon to tell.  It gives the warming in this century as .4 C and
suggests this is on the low side of greenhouse model predictions.
         There are also a couple of climate papers in the back
section.
                          James B. Shearer
=========================================================================
Date: 10 February 1995, 21:53:45 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
         Snark posted:
>There is an interesting letter (with reply) in this month's "Physics
>Today," questioning some details of some global warming models.  I
>don't understand the arguments well enough to comment knowledgeably.
         Robert Parson commented:
> The letter is by Richard Lindzen, the most eminent global warming
>critic, and the replies are by Henry Charnock, Keith Shine, and Rober Kandel.
>It's actually the latest iteration in an exchange that has been going on
>for a couple of year (_Physics Today_ seems to have an appallingly long
>publication lead time - Lindzen is responding to letters published in
>December 1993.) Lindzen argues, on the basis of to-be-published
>calculations, that the equilibrium response to doubling CO2 is 0.5
>degrees C for clear-sky conditions and only 0.22 C for 40% cloud cover.
>This is much smaller than the IPCC estimate of 3 degrees C. The
>difference arise from the treatment of water-vapor feedback, but I
>should let someone closer to the field summarize it.
 
         This is not quite right.  The calculation is for the no
feedback case.  I believe the IPCC estimate is with all feedbacks.
The conventional wisdom no feedback estimate would be less (Tobis said
1 C).  It appears Lindzen's view of what no feedback means differs from
other modelers.  It also appears that calculating the no feedback
effect is not as straightforward as Tobis has been claiming.
                          James B. Shearer
PS:  The letters are in the February 1995 Physics Today starting at
page 78.
=========================================================================
Date: 13 February 1995, 19:30:30 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: funding climatology (was computer modeling)
 
         Len Evens posted:
>I again hope that James Shearer will reverse his previously stated
>opinions on the need for funding research in this area.
 
         My previously stated opinion is that climatologists should
not be given a blank check which I see no reason to reverse.  More
money for climatology means less money for other federal programs which
also have supporters.  Are you willing to cut medicare to fund
climatology research?
         I would fund climatology at some level.  Since I am unfamilar
with the details of current funding I don't really have a strong
opinion about whether it is too high, too low or just right.  Can you
summarize current federal support of climatology research (or do you
support funding on blind faith)?
         I might also change funding priorities within climatology.
For example I might give more money to observation based research and
less to running intricate computer models.  In this regard I have seen
a computer time cost figure of $100000 for a single run of a big model.
This is lot of money given my opinion of the reliability of these
models.  In this regard does anyone have figures for the computational
costs of running the big models and how this compares to the total
climatology research budget?
                          James B. Shearer
=========================================================================
Date: 13 February 1995, 20:08:19 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
         Michael Tobis asked:
>I would like Mr Shearer to provide a *single* verifiable in-context quotation
>of a climate modeller overselling the reliability of GCMs, which seems to be
>the class of model Mr Shearer is criticizing, or of any other approach.
 
         I was thinking of you and Halliwell.  Halliwell for example has
contended that climate models are as reliable as weather prediction
models (or at least he jumped all over me when I stated they aren't).
         For another example consider the testimony of James E. Hansen
before the US Senate Subcommitte on Energy and Natural Resources on
June 23, 1988 as quoted in "The Challenge of Global Warming" edited by
D.E. Abrahamson, Island Press, 1989 (p 36-38).
         "The present observed global warming is close to 0.4 <degree>
C relative to "climatology," which is defined as as the 30-year (1951-
1980) mean.  A warming of .4 <degree> C is three times larger than the
standard deviation of annual mean temperatures in the 30-year
climatology.  The standard deviation of .13 <degree> C is a typical
amount by which the global temperature flucuates annually about its
30-year mean; the probability of a chance warming of three standard
deviations is about 1%.  Thus we can state with about 99% confidence
that current temperatures represent a real warming trend rather than
a chance fluctuation over the 30-year period."
         Do you wish to defend this calculation?
                          James B. Shearer
=========================================================================
Date: 13 February 1995, 20:52:16 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: free-market environmentalism
 
         Kit Rawson asked:
>I am interested in a discussion of free-market environmentalism.
>Many people criticise the current environmental regulatory
>structure (as ineffective as it is).  Beyond criticism, however,
>I have not heard much detailed discussion of free-market
>alternatives to governmental regulation.  I can understand the
>arguments against regulation.  However, I cannot support moving
>away from regulation without a better understanding of how we
>can preserve some of the Earth's natural functions without it.
>I'd also be very interested in books and other sources with more
>information on this subject.
 
         Actually free-market environmentalism does not really mean
eliminating regulation.  Instead it proposes that regulation be done
in more economically efficient ways than at present.  For example
by taxing pollution rather than setting permitted levels.  It should
be easy to find references.  This is related to cost/benefit analysis
which also has a vast literature.
                          James B. Shearer
=========================================================================
Date: 16 February 1995, 19:05:50 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
         Michael Tobis asked:
>I would like Mr Shearer to provide a *single* verifiable in-context quotation
>of a climate modeller overselling the reliability of GCMs, which seems to be
>the class of model Mr Shearer is criticizing, or of any other approach.
         I replied in part:
>          For another example consider the testimony of James E. Hansen
> before the US Senate Subcommitte on Energy and Natural Resources on
> June 23, 1988 as quoted in "The Challenge of Global Warming" edited by
> D.E. Abrahamson, Island Press, 1989 (p 36-38).
>          "The present observed global warming is close to 0.4 <degree>
> C relative to "climatology," which is defined as as the 30-year (1951-
> 1980) mean.  A warming of .4 <degree> C is three times larger than the
> standard deviation of annual mean temperatures in the 30-year
> climatology.  The standard deviation of .13 <degree> C is a typical
> amount by which the global temperature flucuates annually about its
> 30-year mean; the probability of a chance warming of three standard
> deviations is about 1%.  Thus we can state with about 99% confidence
> that current temperatures represent a real warming trend rather than
> a chance fluctuation over the 30-year period."
         Michael Tobis responded:
>I don't like the presentation, since there seems to be an unstated and
>false assumption that interannual temperature variation is uncorrelated.
>However, I did specify *in-context* and a lot of context is missing from the
>above, which may mitigate the statement as made by Hansen, though not as
>quoted by Abrahamson. (fwiw, I don't like Abrahamson's book very much.)
 
         To clarify, the book contains what appears to be Hansen's
entire statement.  I extracted the above paragraph.
         Michael Tobis continued:
>I suspect all Hansen was trying to say was that there is a verifiable
>warming, not that it is verifiably outside the range of natural variability.
>If so, I would support the statement in such a context. If the statement
>was meant to imply that there is a real warming *outside of natural
>variability*, the reasoning is inadequate and the presentation misleading.
 
         The second bullet at the start of Hansen's statement (p. 35 in
the book) is "The global warming is now sufficiently large that we can
ascribe with a high degree of confidence a cause and effect relationship
to the greenhouse effect."  The quoted paragraph came from a section
titled "Relationship of global warming and greenhouse effect".  The
conclusion of this section (p. 40 in the book) reads:
         "Global warming has reached a level such that we can state
with a high degree of confidence a cause and effect relationship between
the greenhouse effect and the observed warming.  Certainly further study
of this issue must be made.  The detection of a global greenhouse signal
signal represents only a first step in analysis of the phenomenon."
         If this doesn't convince you of the context, I suggest you
look up the staement yourself.
         Michael Tobis continued:
>However, it has nothing to do with dynamic climate models, so it doesn't
>provide you with any sort of example whatsoever.
 
         You said "or of any other approach".  Here we have a prominent
modeler making a "misleading" presentation to Congress based on
"inadequate" reasoning.  This indicates to me that at least some
modelers can not be trusted to give an objective account of the
reliability of their work.
         I stated (in reply to Evens):
>          I disagree, the underlying physical principles are not well
> understood in any effective sense (saying everything follows from
> Schrodinger's equation is not helpful).
         Tobis responded:
>It would if the models were an implementation of Schrodinger's equation.
>In fact, they implement the Navier-Stokes equations, and the result is
>pretty good.
 
         The models contain much more than the Navier-Stokes equations.
In any case it is my understanding that it is computationally infeasible
to solve the Navier-Stokes equations directly because of turbulence.
         I said:
>          The predictions are not similar.  For CO2 doubling complicated
> models predict 1.5-4.5 C, a simple model 1 C.  If the complicated model
> predictions are "similar", then they are all similar to the simple
> model so where is the value added?
         Michael Tobis responded:
>The purpose of the models is not to yield a sensitivity, though even IPCC
>is guilty of allowing this impression. The purpose of the models is
>to give some insight into the spatial structure and the transient response,
>which the simpler models can't. For policy purposes, they present plausible
>answers ot the question "1.5-4.5 C, so what?"
 
         If the purpose is insight into the transient response, why are
results usually given for the equilibrium response?  If the models
don't agree on the sensitivity, why should I expect them to be reliable
for the transient response and the spatial structure which are
generally considered to be harder to predict?
         I posted:
>          Wall Street models may be better because of the continous
> feedback and the painful consequences of errors.  Nevertheless Wall
> Streeter's are continually coming up with models that work well most
> of the time but then fail castrophically when something unexpected
> happens.  Remember portfolio insurance or the Granite funds.
         Michael Tobis responded:
>In the case of climate models, this clearly indicates that the models are
>intrinsically optimistic, rather than alarmist as is often alleged.
         I countered:
>          An absurd conclusion.  Wall Street models attempt to identify
> profitable trades.  A pessimistic model misses some profitable trades.
> This is unlikely to cause serious problems.  An optimistic model
> identifies some unprofitable trades as profitable a recipe for disaster.
         Michael Tobis replied:
>Nothing absurd about it. The models tell us what will happen in the absence
>of unknown phenomena. If some unknown phenomenon kicks in, it will break
>the model, but unfortunately, it may also break the real climate system.
>
>The concession that some phenomenon is equally likely to arise that will make
>the adjustment to a new radiative equilibrium essentially painless as one
>that will make the situation worse than we understand seems to me yielding
>far too much. The claim that we should base our behavior on the expectation of
>such a deus ex machina is hard for me to characterize without being rude.
 
         Your original argument was that the catastrophic failure of
some Wall Street models indicates climate models are intrinsically
optimistic.  This argument is obviously absurd.  You may have some other
support for your belief that climate models are intrinsically
optimistic, but you have not presented it.  I will note that most
predictions of disaster prove pessimistic.
         I posted:
>          A belief that one is obligated to understand a model before
> criticizing it is dangerously wrong, making incomprehensible models
> above reproach.
         Michael Tobis replied:
>No, I express a belief that one should TRY to understand a model before
>criticizing it. You haven't made the effort.
 
         This is unrealistic.  People must decide how much weight to
give computer models without examining each one in detail.  As I said
before I discount the results of any model which cannot be convincingly
validated.  The climate models have not been convincingly validated.
Your arguments to the contrary are basically wishful thinking.  I also
find it amusing that you agree complicated computer models in other
fields are of doubtful validity but you expect me to believe
climatology is different.
         I am willing to expend some effort to learn more about the
models.  Do you have some accessible references?  For that matter are
there any climate models, which you are willing to defend, available
on the net?
         I said:
>          These are all very weak forms of validation.  Strong valida-
> tion would be comparing the model answers to the correct answers over
> the entire range of conditions the model is suppposed to handle.
         Michael Tobis replied:
>Yes, this is a common problem in software engineering, not just in
>climate modelling. It woould take an essentially infinite amount of
>time to verify a model in this way. It would also take a few billion years
>to verify almost any significant software product. We have to apply
>intelligence rather than brute force testing.
 
         This is not the problem.  If the model computed the correct
answer for 100 random inputs chosen from the entire input space this
would be strong evidence that the model computes the correct answer at
least 90% (for example) of the time.  This would not follow if the 100
inputs were chosen from a small subset of the input space.  This is
why I said the "entire range".
         The real problem is that it is impossible to test any computer
program if you don't know what the right answer is.
         I said:
>          I will also mention that it is very dangerous to use any data
> used to develop the model to validate it.  This is shown over and over
> by the failure of stock trading (or sports betting) schemes which work
> great retrospectively when applied to future events.
>
>          Do you doubt that it is possible to come up with obviously
> wacko models which pass all or most of the above tests.  Consider for
> example the following model for predicting climate.  Find the year x
> in the last 100000 with climate most similar to this year's climate.
> Then predict next year's climate will be that of year x+1.  This model
> passes the above tests with flying colors.  Are you impressed?  (One
> might quibble that the climate records for the last 100000 years are not
> good enough to actually implement this model, however since you have
> making been claims about how well paleoclimatic conditions are known
> this shouldn't bother you.)
         Michael Tobis responded:
>This is the part I'd like to respond to at some length. It shows a complete
>misunderstanding of what a climate model is. I wonder if anyone else reading
>is familiar with the type of model Mr Shearer is confusing with GCMs, as
>well as with GCMs themselves, and can elucidate the difference. If not, I'll
>try in a few weeks when I find the time.
 
         I was not claiming the purely empirical model presented above
resembles the GCM computer models.  I was giving it as an example to
show your validation tests are not very strong.
         In general models may be empirical looking for patterns in
historical data and predicting the future based on these patterns or
they may be intelligent attempting to explain the past data as the
consequence of some general laws and then using these laws to predict
the future.  Hybrids are also possible.  The reliability of empirical
models cannot be easily estimated from the same data used to derive
them.  It is too easy to find patterns in random variations.  Empirical
models also cannot be expected to do well if applied to conditions
outside the range of the historical data used to derive them.
         Any proposed model should be tested against a few simple
purely empirical models.  Early attempts at numerical weather
prediction did not do as well as the empirical methods then in use.
         The current general climate models are hybrids containing
large empirical components.  The empirical portions decrease one's
confidence in their ability to predict climate conditions outside
the range of historical data.
         Michael Tobis asked:
>Also, I remain curious how you would design a greenhouse gas policy in the
>absence of any information at all from the models you like to criticize.
         I replied:
>          I would use simple models.
         Michael Tobis responded:
>But the simple models just yield a global average temperature change, and
>do not yield a transient response or a spatial response. Thus, they yield
>much less information for sensible weighing of costs and benefits.
 
         What you believe to be additional information, I believe to be
noise.  Any cost/benefit analysis of global warming (and mitigation
measures) is subject to substantial uncertainty and it is unreasonable
to pretend otherwise.
                          James B. Shearer
=========================================================================
Date: 17 February 1995, 20:57:22 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
         I posted (replying to Tobis):
>Btw:  I have not been challenging your claim that given atmospheric
>composition, solar radiation and albedo the surface temperture can
>be calculated from first principles.  (Is this in fact your claim?)
>I now suspect you have been blowing smoke.  How do you calculate the
>effects of clouds from first principles?
         Dave Halliwell responded:
>    What do you mean by "first principles"? Would you consider Beer's
>Law to be "first principles"? Are you familiar with such things as "the
>two-stream approximation" for radiative transfer?
 
         From the Science article I mentioned earlier (1/27/95, p454)
"they found that, on a global average, clouds absorb more than 25 watts
of solar radiation per square meter ... rather than the 6 ... predicted
by theory".  So if modelers are calculating the effects of clouds from
first principles they aren't doing it correctly.
         Leonard Evens posted:
>                ...   There are circumstances in which a simple model
>is just not accurate enough to tell us what is going to happen, so
>we have no choice but to use more elaborate models.   ...
         I replied:
>         This is the thinking that leads to trouble.  The alternative
>is to accept that we can not tell with certainty what is going to
>happen.
         Dave Halliwell asks:
>   Define "with certainty", and give _one_ example where science can
>make a prediction "with certainty".
 
         Ok, I should have said "with the desired degree of certainty"
instead "with certainty".  This sort of quibble does not address my
point which is elaborate models are not necessarily better than simple
models.
         I posted:
>  An elaborate model should be preferred to a simple model only
>when it can be convincingly demonstrated that the elaborate model can
>be expected to give better predictions in practice.
         Dave Halliwell responded:
>   In Len's post, he has said that he is considering a situation
>where simple models are not "accurate enough". An example would be
>where the simple models leave out something that we know can affect
>the result. You have made the argument for _ignoring_ simple models
>because they do not include things like ocean circulation. Thus YOU
>favour more complicated models, using as your sole justification the
>fact that simpler models don't include "everything".
>
>   Now, in response to your claim above, why is it that _you_ prefer
>more elaborate models, since you also claim that you are not convinced
>that they are demonstrably better than the simpler ones?
 
         I have never argued for ignoring simple models.  I have argued
that their estimate of the effects of CO2 forcing is imprecise because
they ignore important aspects of the climate system.  More complicated
models could in principle give better estimates.  However I am not
convinced with the current state of the art that they actually do, and
for this reason I prefer simpler models.
         I said:
>         A belief that one is obligated to understand a model before
>criticizing it is dangerously wrong, making incomprehensible models
>above reproach.
         Dave Halliwell commented:
>   A belief that one's criticisms, made from a position of little or
>no understanding, can't possibly be wrong is even more dangerous.
>
>   A model incomprehensible by _anyone_ would be suspect. Models that
>_are_ understood by a large number of people are not incomprehensible.
>Thus climate models are not in that class, and you are constructing a
>strawman.
 
         I have no such belief that I "can't possibly be wrong".
Anybody should be open to the possibility that they are mistaken, even
"experts".
         As for climate models, how many lines of computer code are in
the big models?  How many people have read and understood every line?
I doubt very much it is a "large number".
         I said:
>         I will also mention that it is very dangerous to use any data
>used to develop the model to validate it.  This is shown over and over
>by the failure of stock trading (or sports betting) schemes which work
>great retrospectively when applied to future events.
         Dave Halliwell commented:
>    Now, if you could actually demonstrate that this is what has been
>done with climate models, then you'd have a point. For starters, why
>don't you take Michael's list and tell us which ones suffer from this
>fault? Describe, in detail, just what data is used to develop the model,
>and what data is used to validate it.
 
         The article "Climate Modeling's Fudge Factor Comes Under Fire"
(Science 9/9/94, p. 1528) points out that many models have been forced
into agreement with today's climate.  Hence they cannot be validated by
their ability to reproduce contemporary climate.
         More generally I consider large computer models "guilty until
proven innocent".  Hence I believe the burden of proof is on you to
show this has not occurred, not on me to show that it has occurred.
As I said there are numerous examples from other fields where people
have placed excessive faith in models for this reason.  I will also
mention that there have been studies of successive determinations of
physical constants which show that improvements in precision often
cause the accepted values to jump outside the claimed error bars of
the previous determinations.  It has been suggested that this occurs
because experimenters stop debugging their experiments as soon as the
results appear reasonable.  Is it possible that climate modelers stop
debugging their codes as soon the results seem reasonable?  I find it
hard to believe climatology is immune to pitfalls which have repeatedly
tripped up researchers in other fields.
         I said (in reply to Tobis):
>         Do you doubt that it is possible to come up with obviously
>wacko models which pass all or most of the above tests.  Consider for
>example the following model for predicting climate.  Find the year x
>in the last 100000 with climate most similar to this year's climate.
>Then predict next year's climate will be that of year x+1.  This model
>passes the above tests with flying colors.  Are you impressed?  (One
>might quibble that the climate records for the last 100000 years are not
>good enough to actually implement this model, however since you have
>making been claims about how well paleoclimatic conditions are known
>this shouldn't bother you.)
         Dave Halliwell commented:
>    To suggest that your "climate model" bears any resemblance at all
>to the types of climate models that are actually used only demonstrates
>your ignorance again. To begin with, your model can't even begin to
>*try* to model three of the six items on Michael's list, so the chances
>of it "passing with flying colors" are rather remote.
>
>   Perhaps you would like to explain to us how _your_ "climate model"
>accepts input that tells it to model the changes in Michael's list?
 
         I made no suggestion that this ("obviously wacko") model
resembles those in actual use.  I was just pointing out it would pass
Tobis's tests.  This model predicts the past perfectly and for that
reason would pass every one of Tobis's tests.
         Michael Tobis asked:
>Also, I remain curious how you would design a greenhouse gas policy in the
>absence of any information at all from the models you like to criticize.
         I replied:
>         I would use simple models.
         Dave Halliwell responded:
>   Yet you have clearly argued for rejection of the results of the
>models you now claim to want to use. You haven't even shown any clear
>understanding of just what a "simple" climate model _is_, or what it
>can do.
 
         I have argued that simple models cannot be exepected to give
a precise estimate of the effects of CO2 forcing.  I have never argued
that they are completely worthless.  I posted the following to this
group over a year ago (on 12/7/93, in reply to Tobis):
>         This is the sort of reasoning used in arriving at figures for
>what regulation costs the economy.  Of course regulations may have
>benefits in which case they should be estimated as well.  You may
>object that such reasoning ignores many things (which I have dismissed
>as quibbles) and hence is imprecise.  However such objections apply
>also to the climate models which you have been defending.  In both
>cases one needs to be aware that there is some uncertainty in the
>results obtained however I do not believe this justifys totally
>ignoring them (as has been advocated in this group).
         This remains my position.
         Dave Halliwell posted:
>   There have already been *several* levels of "making them more realistic
>still does not significantly change the results", which is the criterion
>you stated for you to "give greater weight to their results".
 
         If the models all give the same results where did the 1.5-4.5
C range come from?  Or are you contending 1.5 C and 4.5 C are not
significantly different?
         Michael Tobis asked:
>I would like Mr Shearer to provide a *single* verifiable in-context quotation
>of a climate modeller overselling the reliability of GCMs, which seems to be
>the class of model Mr Shearer is criticizing, or of any other approach.
         I replied (in part):
>         I was thinking of you and Halliwell.  Halliwell for example has
>contended that climate models are as reliable as weather prediction
>models (or at least he jumped all over me when I stated they aren't).
         Dave Halliwell comments:
>   Glad to see that Shearer has selected something that fulfills Michael's
>request for a *verifiable*, in-context quotation. His response to Michael
>also demonstrates why he seems to hold the position that he does: either
>he can't read, or he doesn't understand english.
>
>   For the record, here are portions of four posts that I have made
>in the last month or so regarding climate models and weather models.
>NOTHING that I have said can be reasonably interpreted as a claim "that
>climate models are as reliable as weather prediction". In fact, I have
>acknowledged that there are differences bwetween the two. The position of
>Shearer's that I have challenged was his claim that weather models
>contributed _nothing_ to climate models. I challenge him to provide a
>_single_ quote from my posts that makes _any_ direct comparison of the
>overall reliability of climate and weather models.
 
         I have made no claim that weather models contributed _nothing_
to climate models.  The start of our exchange on this topic was as
follows.
         I stated:
>         Science progresses by testing theories against experiments and
>observations of the natural world.  Climatologists have essentially no
>ability to experiment and a very limited set of observations.
         You asked:
>   Beginner-type question: how much is there in common between climate
>models and numerical weather prediction (NWP) models? (There is an
>intentionally-misleading aspect to this question.) Would you put NWP
>models in the same category regarding lack of suitable testing?
         I replied:
>         No I would not put NWP models in the same category because
>since they predict on a much shorter timescale (and in some cases for
>smaller areas) the observational record is in effect much larger.
 
         You then proceeded to flame me for the above remark.  If you
in fact agree that weather prediction models are more reliable than
climate models I don't see the point of criticizing me for saying it.
Similarly I don't see the point of asking whether I can describe the
differences between climate models and weather models if you agree
that they are different.
                          James B. Shearer
PS:  My access to sci.environment seems to have gone away so this may
be my last post for a while.
=========================================================================
Date: 24 February 1995, 20:15:21 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
         I have found an alternative access to sci.environment so (for
better or worse) I will continue to post as I have time.
         Tobis and Halliwell seem to have completely missed the point
I was trying to make with the model I posted.  I originally said:
>         I will also mention that it is very dangerous to use any data
>used to develop the model to validate it.  This is shown over and over
>by the failure of stock trading (or sports betting) schemes which work
>great retrospectively when applied to future events.
>         Do you doubt that it is possible to come up with obviously
>wacko models which pass all or most of the above tests.  Consider for
>example the following model for predicting climate.  Find the year x
>in the last 100000 with climate most similar to this year's climate.
>Then predict next year's climate will be that of year x+1.  This model
>passes the above tests with flying colors.  Are you impressed?  (One
>might quibble that the climate records for the last 100000 years are not
>good enough to actually implement this model, however since you have
>making been claims about how well paleoclimatic conditions are known
>this shouldn't bother you.)
          Michael Tobis replied:
>The more I look at this, the less sense it makes. What is your metric of
>"similar"? In practice, this method is a clumsy way of identifying tha
>El Nino phase, and I think its skill would be quite small on the tests I
>mentioned for a 1-year prediction, and negligible for a 5-year prediction.
>
>What are you trying to prove here? That you think climate is a scalar
>variable?
          Dave Halliwell posted the following concerning his version
of the model:
>   As an example, the model prediction for 1864 is based on the observed
>T of -0.23 in 1863. The closest value in the existing record was -0.18 in
>1859. This meant that the 1860 T of -0.38 was used for the prediction.
>The error was thus 0.08 (warm).
 
          Actually the closest year to 1863 in the record is obviously
1863.  This will be true for any reasonable metric for "similar" which
is why I did not bother to specify one.  Thus the model predicts the
climate of 1864 will be that of 1864 with an error of 0 and in general
predicts the past perfectly.  Of course you may object that this happy
state will be unlikely to continue if we attempt to use the model to
predict the future.  However that is exactly the point I was trying to
make.
          The past record is embedded in the above model in an obvious
way, however there are much more subtle ways of doing the same thing.
The effect will still be a model that does not predict the future as
well as it predicts the past.  Tobis and Halliwell may deny that this
could be happening with climate models, however I think it is almost
certainly happening to some extent.
          On another point I posted:
>          I am willing to expend some effort to learn more about the
> models.  Do you have some accessible references?  For that matter are
> there any climate models, which you are willing to defend, available
> on the net?
          Michael Tobis replied:
>You can't run a GCM. Alas, the user interface isn't the string point of
>these models, but more to the point they take enormous computational
>resources. However, you can learn about them. The best introductions are:
 
          You might be surprised at what computational resources I have
access to.  In any case I didn't plan to attempt to do more than few
timesteps.  Just reading the source might be illuminating.  Is source
available?  Is it a requirement for publication of model results that
the source be freely available?  (I certainly think it should be a
requirement.)
          Also as Dave Halliwell keeps repeating there are smaller
models.  Is the source to any of them available?
          Halliwell has contended that source is not needed to under-
stand a model.  I may dispute this at length if I find time, for now
I will just note the idea that a computer program is necessarily a
faithful implementation of a short description of what the model is
supposed to be is extremely naive.
                          James B. Shearer
=========================================================================
Date: 28 February 1995, 19:33:37 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: natural dry ice?
 
         According to my Handbook of Chemistry and Physics, CO2 freezes
at -78.5c (or about -110f).  It is sometimes colder than that in
Antarctica.  Does this cause CO2 to freeze out of the atmosphere?  If
not, how much colder would it have to get for this to happen?  How
much colder still for significant amounts?  Is there any chance such
conditions have occurred in the past?
                         James B. Shearer
=========================================================================
Date: 3 March 1995, 17:28:58 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: computer modeling (was scientific culture)
 
         I have been arguing with Michael Tobis and others about how
much weight should be given to large computer models of the effects of
CO2 forcing.  I do not seem to be convincing anyone so I will try
another approach.
         Suppose studies are made of the effects of injecting dust into
the stratosphere.  Suppose the same sorts of computer models in which
Tobis has so much faith indicate that this would provide a cheap way
of reversing any ill effects from adding CO2 to the atmosphere.  Would
Tobis then advocate basing public policy on these models?
                          James B. Shearer
=========================================================================
Date: 10 March 1995, 19:03:59 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: wind energy economics
 
         Paul Dietz posted:
>That current wind generation is a small fraction of the US
>electricity production is *good* news for wind advocates.
>It means that wind is still near the beginning of the learning
>curve.
>
>A learning curve is a plot of the cost of a unit of production of some
>object vs. cumulative production.  Empirical models of the learning
>curve use a power law approach; a typical curve in aerospace, I
>understand, has unit cost going as (production)^-0.2.  I don't
>have any information on the learning curve in the wind power
>industry.
>
>Under this curve, the cost of wind energy would drop to 63%
>of its current cost after production increased by a factor
>of 10, to 40% after a factor of 100, and to 25% after a factor
>of 1000.
>
>A factor of 1000 increase in power output would be an average
>production of 400 GW.  If current cost is $.04-.06/kWh, this would
>give electricity at $.01-.015/kWh.  This is quite cheap.
 
         This argument is misleading.  The economics of wind energy
are quite site dependent.  Greatly expanding production would involve
use of inferior sites.  This effect could easily dominate.
         Furthermore you are assuming the regulatory environment would
remain the same.  The example of nuclear power shows this can not be
assumed.
         I also doubt current costs are $.04-.06/kwh.  Where did this
number come from?
         Finally, technically, I believe you should be looking at
cumulative production not annual production.
                          James B. Shearer
=========================================================================
Date: 10 March 1995, 19:38:22 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: intellectual honesty
 
         Len Evens posted:
>Also, I think you should give Michael and the rest of us credit
>for intellectual honesty.   I don't know if you meant it that way,
>but there is a hint of an implication in what you wrote above
>that we are comfortable with the predictions of models because
>we like the course of action they can be used to justify.
 
         People are more comfortable with data that confirms their
existing beliefs, particularly strongly held ones, than with data
which refutes them.  I see no reason to believe Michael and the rest
of you are exceptions.
         It is possible with effort to compensate for this human trait
to some extent.  I try to do this motivated in part by the memory of
past misjudgements this trait has caused me to make.  However I do not
believe I or anyone else can be completely objective.  Some of Michael
Tobis's past posts lead me to doubt he adequately recognizes the
dangers.
                          James B. Shearer
=========================================================================
Date: 10 March 1995, 20:58:09 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: global warming - active countermeasures
 
         I posted:
>          Suppose studies are made of the effects of injecting dust into
> the stratosphere.  Suppose the same sorts of computer models in which
> Tobis has so much faith indicate that this would provide a cheap way
> of reversing any ill effects from adding CO2 to the atmosphere.  Would
> Tobis then advocate basing public policy on these models?
         Michael Tobis replied in part:
>However, first order cancellation is better than no cancellation at all. I
>think that deliberate releases of aerosol to counter the greenhouse effect
>will likely occur, and I will probably support them if I am still around
>when matters get severe enough to warrant this.
>
>I would not advocate such a policy before matters become severe, and would
>prefer a policy that avoided such a scenario as much as possible.
 
         If active countermeasures may be feasible, why are they almost
never presented as options?  How much space in the IPCC report is
devoted to analysis of such countermeasures (as opposed to restricting
CO2 emissions)?  How much funding is study of such options currently
receiving?
         Why do you prefer options which restrict economic growth?  What
cost/benefit analysis justifys this preference?
                          James B. Shearer
=========================================================================
Date: 14 March 1995, 18:35:13 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: intellectual honesty
 
         Michael Tobis posted:
>Through all our discussions, though you raise numerous interesting
>points, you seem to be constantly disregarding that the anticipated
>perturbation to the radiative balance, given no restraints on carbon
>emission and no sudden arrival of cheap non-carbon energy, is very large
>compared to natural perturbations on similar time scales. I wonder
>if you have been objective about this, the basic substantive question
>at the root of this issue.
 
         I don't agree this is the basic substantive question.  I
believe the basic substantive question is whether CO2 additions to the
atmosphere are likely to be harmful and if so what is the most cost
effective means of reducing the expected damage.
         As I have said before, I accept that it is reasonable to
expect CO2 forcing to produce some global warming.  This does not
by itself justify restrictions on CO2 production.  This requires a
cost/benefit analysis.  I have yet to see anything resembling such
an analysis.
                           James B. Shearer
=========================================================================
Date: 14 March 1995, 19:08:31 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: global warming - active countermeasures
 
         I asked why little consideration is given active measures
to counteract global warming caused by CO2 forcing.
         Len Evens replied in part:
>                    ...   In any case, even if it were deemed appropriate
>to try to cool the planet by global engineering involving aerosols,
>we would have to know quite a lot about greenhouse gases and their
>effects to know how to do that.   For example, suppose it did turn
>out that the skeptics were perfectly correct and nothing much will
>happen to climate some 50 years down the line.   Then it would be
>a bit silly to spend a lot of time contemplating how we would inject
>aerosols in the stratosphere to cool the planet without at the same
>time producing other unanticipated undesirable side effects.
>Thus understanding the consequences of enhanced greenhouse warming
>is necessary as a first step in any case.
 
         I don't understand this objection.  By this logic it would
be even sillier to spend a lot of money restricting CO2 emissions
(which I believe you support) to combat a possibly illusory problem.
Studying active countermeasures could be a comparatively cheap form of
insurance.
         Len Evens also posted:
>I for one admit that, without knowing exactly how, I think humanity
>could somehow muddle through dealing with limitations on fossil
>fuel emissions.    After all, it would seem that we can exercise
>some degree of control over how our economic and social systems
>work, and I don't think economic growth is completely dependent
>on current use of energy sources.   Also, radical changes in social
>and economic behavior have occurred in this century and human
>beings have mostly coped with them.   Although dealing with
>limitations on fossil fuel emissions should not be easy, I
>don't see why it is different in magnitude or nature from other
>challenges our societies have faced.   ...
 
         Why don't you believe we could "muddle through" global
warming as well?  The question of course is which would be easier.
         Michael Tobis replied in part:
>Cancelling the first order surface warming should be considered symptomatic
>relief mostly. (It would prevent thermal expansion of the oceans, though,
>which may turn out to be a large part of the cost). ...
 
         What do you have against symptomatic relief?  I believe you
have stated the climate change to date is not worth worrying about.
Could active countermeasures reduce the expected climate change into
the not worth worrying about area?
         Michael Tobis continued:
>                                                ... Exponentially growing
>second order perturbations of the atmosphere would still eventually be
>expected to cause significant climate change. ...
 
         Before or after the supply of fossil fuels is exhausted?
Btw could the aerosol cooling be made latitude dependent?  If so this
would allow cancellation of second order effects as well, would it not?
         Michael Tobis added:
>                   ... So I expect that regardless of if and when this
>endeavor takes place that growth in carbon emissions eventually will need
>to be curtailed. And since the forcing is roughly an integral of the
>carbon emissions, the sooner we stop increasing the annual emissions,
>the better in the longest run.
 
         On the other hand the longer you wait the less likely you
are to impose costly unneeded restrictions.
         Michael Tobis also posted:
>I also think that the "restriction on economic growth" posed by carbon
>taxes is not especially enormous. I don't understand why this should
>terrify people (except coal miners) so. There are other ways of getting
>energy.
 
         The problem is they are all much more expensive than fossil
fuels.  Of course if you are rich this may not bother you much.
                          James B. Shearer
=========================================================================
Date: 14 March 1995, 20:24:28 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: why growth?
 
         Len Evens asked:
>I do have one question though.   Why is economic growth in itself
>a positive good?   It would seem necessary in the face of growing
>population, but one could certainly envision a stable society
>in which growth was not necessary.   Maybe that can be proved
>beyond a reasonable doubt to be impossible?
 
         Most people believe it is better to be rich than poor and
for this reason favor economic growth.  Of course you are free to
disagree, however in a democratic society it would be unreasonable
to expect your eccentric view to be the basis of public policy.
         The question is not whether economic growth is necessary
but rather whether it is desirable.  All sorts of strange societies
are probably feasible, this does not mean most people would find them
attractive.
         Michael Tobis posted:
>Finally, I confess that I still don't understand what "economic growth"
>means, at least in the modern countries. Soemtimes it reminds me of one
>of those circular Escher stairways, where locally there's a difference
>between ascending and descending, but in the long run you stay in
>exactly the same place. Assuming food and shelter is under control,
>it isn't at all clear to me that over the long time scale of interest
>here, whether aggregate wealth is a well-defined quantity.
 
         I find it hard to take this comment seriously.  First it is
my understanding that much of the projected increase in CO2 emissions
will occur in countries which are not presently wealthy (which is what
I presume you mean by "modern").  Second is it your contention that
the United States should not worry about climate change unless it is
likely to prevent us from feeding and sheltering our population since
any other measures of wealth are illusory?  Do you believe the belief
of most Americans that their lives would be better if they had more
money is irrational and should be ignored in deciding public policy?
What exactly is your point?
                          James B. Shearer
=========================================================================
Date: 14 March 1995, 20:48:54 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: wind energy economics
 
         I posted (in response to Paul Dietz):
>          I also doubt current costs are $.04-.06/kwh.  Where did this
> number come from?
         Tom Gray replied:
>No, this is correct.  I'd broaden the range to 4-7 cents/kWh.  Several
>large (relatively speaking) wind projects have been bid within this
>range within the past two years.  There is even a contract in Texas for
>a 40-MW project that is priced at 3.9 cents/kWh.  Source for this
>information is the 1995 Industry Status Report from the American Wind
>Energy Association.
 
         Interesting, what does this mean exactly?  Does a private
company assume the entire risk of the project with the utility only
committing to buy any power produced (up to a limit) at the stated
price or what?  Are these projects subsidized in any way (free land,
tax breaks or credits etc.)?  Is the price subject to an escalation
clause?  What would a comparable cost for a natural gas based project
be?
         Also how does the total cost divide between capital cost
and operating expenses?
                          James B. Shearer
=========================================================================
Date: 17 March 1995, 19:44:03 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: wind energy economics
 
         I asked:
> What would a comparable cost for a natural gas based project be?
         Tom Gray replied:
>_Powering the Midwest_, a 1993 study by the Union of Concerned
>Scientists, places the cost of power from a combined-cycle gas plant at
>4.1-8.6 cents/kWh, and the cost of power from wind at 2.7-7.3
>cents/kWh, including 0.6 cents/kWh for the value of the tax credit.
>This is levelized busbar cost.
>
>Gas prices are assumed to escalate at 3% above inflation.  Transmission
>costs for gas are $100-$250/kW, for wind $0-$250/kW.
 
         The UCS doesn't have much credibility with me.  The gas price
escalation assumption seems biased.  I think assuming constant real
gas prices would be more neutral.  Actually in practice don't utilities
often assure a supply (through long term contracts or buying a gas
field)?  The 4.1-8.6 range also seems large.  Where is the uncertainty
coming from?  Does the study break the cost down into capital and
operating costs?  Does combined cycle mean cogeneration plants?  Is
transmission cost the capital cost of added transmission lines to
connect to the grid?  Why would this be higher for gas?
         I said (in response to a comment of Michael Tobis about
there being alternatives to fossil fuels)
>          The problem is they are all much more expensive than fossil
> fuels.  Of course if you are rich this may not bother you much.
         Tom Gray replied:
>Prove it.
 
         I should have said much more expensive if used to replace a
substantial fraction of current fossil fuel use.
         I will confine my reply to Tom Gray to wind as an alternative.
         First wind may be becoming competitive as a means of generating
electricity however I believe generating electricity accounts for well
under 50% of current fossil fuel consumption.
         Second the cost figures for wind as a means to generate
electricity ignore costs imposed by the variable nature of wind
generator output.  These costs would become significant if wind power
began to account for a significant fraction of total power generated.
(I have seen 20% of total power generation as the point at which
significant problems would begin).
         Third wind does not appear to be as competitive in the third
world countries projected to account for an increasing fraction of CO2
emissions.
         For these reasons I don't believe wind can reduce future CO2
emissions by more than a small fraction without significantly increased
costs.
         Another way of looking at this is to ask what level of carbon
taxes would be required to reduce CO2 emissions to the desired levels.
If there were reasonably priced alternatives only modest taxes would be
required.  However I don't think that is what is being proposed.
         Btw if wind is really cheapest now in the US there is no need
for the tax credit is there?
                          James B. Shearer
=========================================================================
Date: 24 March 1995, 18:21:14 EST
From: JBS      at YKTVMV
To:   usenet-poster at polecat.newsgate.ibm.com
Newsgroups: sci.environment
X-Post-Me: Yes
Subject: wind energy economics
 
         Tom Gray challenged (regarding cost of natural gas fueled
power stations):
>In the immortal words of Tom Moore, if you don't like their numbers,
>how about coming up with some yourself?   Yes, costs are broken into
>capital and operating costs.  I imagine that transmission costs include
>pipelines, no?
 
         Ok, capital cost 300 $/kw (source Amoco publication Span
1989 issue 3, references EPRI).  Assuming 30 year amortization period,
3% real interest rate and 20% down time capital cost per kwh is about
($300*.051)/(8760*.8) = .22 cents.
         Fuel cost, assume natural gas price 2 $/1000cf, energy content
300 kwh/1000cf and generator efficiency of 40%.  Then fuel cost per kwh
is 2$/(300*.40) = 1.67 cents.
         Neglecting all of expenses (such as non-fuel operating costs)
this is a total cost of 1.89 cents/kwh substantially below the UCS
estimates.
         Perhaps the above figures are somewhat optimistic, however I
doubt wind is competitive with natural gas generators on purely
economic grounds.  Perhaps you could give your figures computed in the
same way.
         I believe in some third world countries gas is very cheap as
it is a byproduct of oil production with no ready market.  In these
countries power from gas plants would be even cheaper than indicated
above.
         Tom Gray posted:
>I see that confining your answer to wind helps you escape a number of
>difficulties, such as geothermal, solar thermal, and biomass-fired
>power, not to mention energy efficiency, which a number of analysts
>have concluded is substantially *cheaper* than adding new fossil to supply.
 
         I consider wind the most competitive of the "green" technol-
ogies.  Do you really