NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/sci/environm/12928

Internet Message Format | 1992-11-21 | 8.6 KB

Path: sparky!uunet!olivea!sgigate!sgi!cdp!alanm
From: [e-mail] (Alan McGowen)
Newsgroups: sci.environment
Subject: I=PAT, and a little sci-fi
Message-ID: <[e-mail]>
Date: 21 Nov 92 06:09:00 GMT
Sender: Notesfile to Usenet Gateway <[e-mail]>
Lines: 176
Nf-ID: #N:cdp:1466601925:000:8480
Nf-From: cdp.UUCP!alanm    Nov 20 22:09:00 1992


Some thoughts on ecosystem impacts and the I = PAT eqn. [I =
impact, P = population size, A = mean "affluence" or consumption,
T = impact of technology used to provide consumed goods or
services.]

The I = PAT idea is due to Ehrlich and Holdren [1].

Some posters have criticised the idea on the grounds that the
factors are not in general completely independent, i.e they say
that we might have A = A(P,T), T = T(P,A) or dependencies on
derivatives. It has also been suggested that only physical
scientists would ever notice such a complex situation ;-).
Actually, the problem may be that physical scientists have
inappropriate expectations about the generality which can be
expected from any detailed model in the life sciences.

I = PAT is *not* a detailed model -- it is a template for
building detailed models -- it is a *class* of models, if you
like. To build a detailed model, we need to specify the exact
impact and the factors of technology and affluence which are
relevant to it. It is also important to remember that the
response of ecosystems to impacts is not in general linear, even
though there are *qualitative* similarities of response to
impacts of different *kinds* (which is why the concept of impacts
is useful).

For example, consider the impact of logging on a forest
ecosystem. Very selective logging on a small scale may closely
approximate the normal effects of trees falling -- especially if
the wood is used within the ecosystem, so that its nutrients are
not permanently removed from the total capital. This is the
situation with traditional use of trees by indigenous forest
people. Here the impact may be smaller than what the ecosystem
can absorb without change of character ("loss of function"). At
the opposite extreme is clearcut logging on a scale affecting
entire watersheds.

The different *techniques* of logging give the different factors
of T. Note that T should be measured as some specific type of
important impact -- e.g. as a rate of habitat conversion per unit
of time per dollar in the case of nonselective logging.
Quantification is difficult but not impossible. For techniques
which produce extreme degredation remote sensing can be used to
measure extent of habitat destruction. For less destructive
techniques there are a panoply of diversity measures which could
be used to quantify impacts.

A is simply the average consumption of forest products of the
population responsible for the logging.  In an earlier post about
this subject, one objection raised was that the T for logging in
one area (say Indonesia) was multiplied by the A in another area
(say Japan) -- this violated the intuitions of one poster.
However, the As which are relevant are the ones which are
responsible for the logging operation -- i.e. the As at the
sources of consumption. A has to be modified if there are
nonmarket factors such as regulations restricting the use of a
given technology or its use in certain areas, since this reduces
consumption.

P is just the number of people with the A characteristic, i.e. PA
is the total consumption of the forest products produced by the
technology with impact T per unit of consumption. The result?

I(rate of habitat destruction due to clearcut logging)
= P(consumers of forest products from the logging)X
  A(mean consumption of forest products from the logging)X
  T(rate of habitat destruition for the given technology and the
    given ecosystem per unit of consumption)

Like more than a few basic ideas in science, this is scarcely
anything more than than bookeeping. Moreover, the factors are
independent enough to permit some prediction -- for example of
the likely effects on a specific ecosysem of a regulation
changing T (specifying logging methods) or A (restricting take).
Likewise, for cases where P is increasing rapidly and T is
changing slowly if at all (South American farmers with
chainsaws), again prediction is possible without considering
complex implicit dependencies among the factors.

Michael Tobis wonders whether my position is that I is an order of
magnitude too great, and that P has to come down by this much.
Actually, I from all sources is probably 2-3 orders of magnitude
above what would be stable in the sense of a coevolutionary
stable state. I don't believe that any one of the three factors P,
A, or T could be expected to absorb the entire reduction, and
would advocate an order of magnitude reduction in P, and another
roughly two orders in AT. The technologial optimists of course
think that T can be brought as low as you please, but history
points the other way: T has been increasing, not decreasing. A
Brazilian farmer with a chain saw and a little gasoline has a
vastly greater T at his disposal than Spanish Conquistador or a
native of the rainforest. Technology has tended to evolve to make
it easier for humans to rearrange things, and that is the whole
problem. If we are to get huge reductions in T we are going to
need some stong nonmarket incentives (i.e. regulations) to get
the same level of service at much lower impacts.

To address the worries that we can't think about the future with
I = PAT because the "real" model is something like

I(t) = P(t, A, T, dP/dt, dA/dt, dT/dt, I, dI/dt) x
       A(t, P, T, dP/dt, dA/dt, dT/dt, I, dI/dt) x
       T(t, P, A, dP/dt, dA/dt, dT/dt, I, dI/dt)

which roughly means "its too complicated to think about", I'd
like to conclude with a philosphical tale which I will present as
a science fiction story.

     A Dialogue of the Dead and the Not-Yet Dead

The Space Empire Empidonax, deciding that they prefer Earth's
rainforests to Earth's humans, determines to end human
technology. They send their latest model space-time cruiser to
Earth, and after stopping a few cars on rural roads, cutting six
or seven crop circles, and leaving kindred signs of Higher
Intelligence, they kidnap four people who post in this group:
an ECONOMIST, a PHYSICIST, a COMPUTER SCIENTIST and a
METEOROLOGIST. The four are transported back in time to the late
17th century, where they are brought face to face with the
English scientist Robert Hooke, president of the Royal Society of
London, who has been busy measuring springs and is about to give
the world Hooke's Law -- the force of a spring is directly
proportional to its displacement from equilibrium. The four time
travelers are dismayed at his simplemindedness.

PHYSICIST: "Look, Hooke, you've fastened onto the first silly
model you've come across. The real situation is far more
complicated. Nonlinear oscillators, chaos... you have no concept.
Just look at what happens when you stretch your spring too far --
it breaks. Start by modelling that. Have you thought about what
happens if the oscillation occurs in a non-Euclidean manifold,
or if energy is quantized? I didn't think so. It's utterly
useless."

ECONOMIST: "Not only that, Hooke, but you have committed a
logical error: you say that EVERY spring obeys your law, but you
haven't examined EVERY spring. You've only examined a FEW
SPRINGS. Your conclusion doesn't follow from your premises."

COMPUTER SCIENTIST: "You don't seem to know how to use
statistics. I think your relation is emotive -- you discard the
cases you don't like. You should study some arithmetic. He who
refuses to --"

METEOROLGIST (cutting in): "Yes, I'm convinced that scientific
objectivity (which has nothing whatsoever to do with history or
cultural background) forces me to conclude that the situation
is probably more complex than Mr. Hooke thinks it. Real springs
are not linear, except in some very local applications."

HOOKE: "Gentlemen, I'm amazed at how simple I have been. You have
quite convinced me."

[Vast upheaval, in which the four time travelers cease to exist.
Hooke never publishes his law, no theory of harmonic oscillators
is ever developed, no electrodynamics, quantum mechanics,
electronic devices or computers are ever developed. The Earth in
the 20th century is quite safe for rainforests.

The Space Empire of the Empidonax is greatly pleased.]


The moral: thou shalt not multiply implicit variables beyond
empirical necessity without well-supported and *applicable*
theoretical cause, or in other words Keep It Simple, Stupid.


[1] P.R. Ehrlich and J.P. Holdren, "Impact of Population Growth,"
Science, vol. 171, pp. 1212-17 (1974)

------------
Alan McGowen