[p2p-research] List of articles on Indium/Gallium Supplies

[marc fawzi]

Clarification:

The kind of data I would bee looking for is the raw data given over a
much longer period

The reason is that if the volume or production is periodic at all then
the way the logistic regression is applied would have to be different
than how it's applied normally (assuming no periodicity in the
behavior of the dependent variable)

If no raw data exists or if it doesn't exist for say the last 50 years
then my own assumption would be that it's periodic at this time but I
would not know the period so logistic regression would be applied in
yet another way

I think what he did is he applied it without assuming any periodicity

and that's all besides the point that you cannot predict gallium
depletion based on gallium production volumes alone, although you can
predict the collapse of production (assuming you're applying the
regression correctly)

I definitely need to see the full analysis (or any other full
analysis) despite knowing that there is a definite flaw in this case
(which is the assumption that if production is about to collapse then
the resource is about to be depleted)

Marc

On Wed, Feb 25, 2009 at 12:49 PM, marc fawzi <marc.fawzi at gmail.com> wrote:
> minor typo in last para:
>
> "so production collapses but that doesn't mean resource IS depleted."
>
>
> On Wed, Feb 25, 2009 at 12:46 PM, marc fawzi <marc.fawzi at gmail.com> wrote:
>> If you look at the data, the author points out the fact that the URR
>> [ultimate recoverable resource] is much higher than his estimate
>> (which is based on the USGS production data alone) in case of Gallium
>> but he decides not to includes the reserves base because he mistrusts
>> the accuracy of the reserves data, which are also supplied by USGS,
>> citing "high uncertainty of the data for gallium."
>>
>> By deciding to use the production data alone all he is able to predict
>> is the depletion (or rather the "collapse") of production, not the
>> depletion of gallium itself. There is one humongous difference here. I
>> hope you understand.
>>
>> He says that "In principle, the peaks that we have reported could be
>> interpreted as due to factors other than depletion."
>>
>> (sorry but I personally find it funny)
>>
>> How can you predict depletion of a resource based on production data alone?
>>
>> Furthermore, the application of logistic regression to survey data
>> itself (to predict collapse of production) is not exposed in the
>> article and could contain some crucial errors, i.e. where is the SAS
>> model or where is the model design in general?
>>
>> For example, if the production volume shows periodicity where the
>> period is known then that takes a different kind of logistic
>> regression model than if the period is unknown, and yet another model
>> if it's non-periodic. So I am not sure if he's applying logistic
>> regression correctly (for collapse of production) given that I don't
>> see the calculation.
>>
>> This also implies that by not discussing the possibility that the
>> production volume (the detailed data not the cumulative data) may be
>> periodic his prediction of the collapse may not only be entirely
>> flawed (in the design of the logistic regression model) but also
>> obstructing a significant truth about production data in ecology and
>> mineralogy, which is that production peaks and collapses in periods
>> (up to the Feigenbaum constant, i.e. r < 3.6) and then it goes into
>> chaos.
>>
>> What causes the periodicity and eventually chaos is that the
>> production rate exceeds the sustainability limit for the given
>> production technology and/or the sustainability limit of the market
>> (he does not supply demand as another possibility rather than
>> depletion of the resource) so production collapses but that doesn't
>> mean resource is not depleted. The production grows again when more
>> efficient technology is developed for extraction or when the market
>> swings back into demand mode (from over supply.)  In fact that is what
>> may be happening with oil, too. We're reaching the first point of
>> bifurcation (the point at the which the quantitative behavior of the
>> system changes abruptly)  but it does not mean that more efficient
>> technology won't make deep oil reserves cheap again which will cause a
>> come back in oil production (after the collapse).
>>
>> The analysis given by that paper is silly at best.
>>
>> Marc
>>
>>
>>
>>
>> On Wed, Feb 25, 2009 at 2:58 AM, Tere Vadén <tere.vaden at uta.fi> wrote:
>>> There is some quantitative analysis here, as regards Gallium ("According
>>> to a logistic fit of the data, it peaked in the year 2000 (Fig. 5)")
>>> (and other minerals) though not explicitely related to solar:
>>>
>>> http://www.theoildrum.com/node/3086
>>>
>>> marc fawzi wrote:
>>>> I have not seen a numerical analysis of the "time and scale" problem
>>>> as it applies to solar energy.
>>>>
>>>
>>
>