150 replies to this topic

[Boar Wrinklestorm]

Leaky integrator has been used to model land/climate coupling before, here (PDF) It seems like Carson (1982) was the very first ... (but can't seem to get hold of that paper - any helpers?)

In Cox et al (1999) reference is made, as well as weaknesses in the generality (it is referred to as 'obvious limitations') to model criteria that use the leaky integrator (here called a leaky bucket - but it's the same thing) to model soil moisture.

Interestingly, it also models the 'overflowing' of the bucket - something I have not considered, since the LI model assumes an infinitely large bucket. This seems to open up the notion of a limiting maximum with respect to quantity in the bucket. This is an observable phenomena in the Vostok ice core record (Petite et al) where we can observe that there is an apparent maximum temperature of the climate.

This is something that the VP/CB LI model didn't really do. More work to do!!

Edited by VillagePlank, 13 March 2010 - 08:00 .

[Hiya]

VillagePlank, on 13 March 2010 - 07:38 , said:

Leaky integrator has been used to model land/climate coupling before, here (PDF) It seems like Carson (1982) was the very first ... (but can't seem to get hold of that paper - any helpers?)

In Cox et al (1999) reference is made, as well as weaknesses in the generality (it is referred to as 'obvious limitations') to model criteria that use the leaky integrator (here called a leaky bucket - but it's the same thing) to model soil moisture.

Interestingly, it also models the 'overflowing' of the bucket - something I have not considered, since the LI model assumes an infinitely large bucket. This seems to open up the notion of a limiting maximum with respect to quantity in the bucket. This is an observable phenomena in the Vostok ice core record (Petite et al) where we can observe that there is an apparent maximum temperature of the climate.

This is something that the VP/CB LI model didn't really do. More work to do!!

That Carson (1982) reference is probably a book not a paper.

[Boar Wrinklestorm]

Hiya, on 13 March 2010 - 08:55 , said:

That Carson (1982) reference is probably a book not a paper.

Ah - didn't realise that. Will search on Amazon with me credit card

[Captain_Bobski]

VillagePlank, on 13 March 2010 - 07:38 , said:

Leaky integrator has been used to model land/climate coupling before, here (PDF) It seems like Carson (1982) was the very first ... (but can't seem to get hold of that paper - any helpers?)

In Cox et al (1999) reference is made, as well as weaknesses in the generality (it is referred to as 'obvious limitations') to model criteria that use the leaky integrator (here called a leaky bucket - but it's the same thing) to model soil moisture.

Interestingly, it also models the 'overflowing' of the bucket - something I have not considered, since the LI model assumes an infinitely large bucket. This seems to open up the notion of a limiting maximum with respect to quantity in the bucket. This is an observable phenomena in the Vostok ice core record (Petite et al) where we can observe that there is an apparent maximum temperature of the climate.

This is something that the VP/CB LI model didn't really do. More work to do!!

Important to note (or rather to emphasise) that this previous use of the leaky bucket, Cox et al 1999, does not use the leaky bucket the same way that we are using it. We, of course, are using the leaky bucket to model the total heat capacity of the entire planet, whereas Cox taks about its use within the Earth climate system.

(Just thought I'd underline that before anyone jumps up and says that the LI has already been discredited! )

I had assumed that an overflowing would occur at some point, since the peak temperature of the Earth appears to be around 22C (historically, the Earth has spent far longer at 22C than it has at its lowest point of about 12C - we're now at somewhere around 14C, I believe). It hadn't occurred to me that this would have any relevance to the LI in the temperature region that we're looking at, since it's well below the 22C "cutoff" point. But would a limiting factor at the top end cause other limitations, or different behaviour, further down the scale?


CB

That Carson 1982 thing is, indeed, a book - or rather it is a section within a book entitled "Land surface processes in atmospheric general circulation models" by PS Eagleson. The link below links to places (in the US) where you can buy it. Pre-owned off Amazon.com for around $80 - you can get a preowned copy from Amazon UK for about £80. Yikes! It's a better deal off Amazon US, but I don't know if it's worth chasing up.

http://openlibrary.o...culation_models

EDIT - Not sure if this page might be of some help in the LI:

http://www.unu.edu/u...5e/80635E0j.htm

It references Carson 1982, and it has a table of average monthly values for albedo at various latitudes (which is the point where Carson comes in).

Edited by Captain_Bobski, 15 March 2010 - 12:46 .

[7andY]

Captain_Bobski, on 15 March 2010 - 12:24 , said:

Important to note (or rather to emphasise) that this previous use of the leaky bucket, Cox et al 1999, does not use the leaky bucket the same way that we are using it. We, of course, are using the leaky bucket to model the total heat capacity of the entire planet, whereas Cox taks about its use within the Earth climate system.

(Just thought I'd underline that before anyone jumps up and says that the LI has already been discredited! )

I had assumed that an overflowing would occur at some point, since the peak temperature of the Earth appears to be around 22C (historically, the Earth has spent far longer at 22C than it has at its lowest point of about 12C - we're now at somewhere around 14C, I believe). It hadn't occurred to me that this would have any relevance to the LI in the temperature region that we're looking at, since it's well below the 22C "cutoff" point. But would a limiting factor at the top end cause other limitations, or different behaviour, further down the scale?


CB

That Carson 1982 thing is, indeed, a book - or rather it is a section within a book entitled "Land surface processes in atmospheric general circulation models" by PS Eagleson. The link below links to places (in the US) where you can buy it. Pre-owned off Amazon.com for around $80 - you can get a preowned copy from Amazon UK for about £80. Yikes! It's a better deal off Amazon US, but I don't know if it's worth chasing up.

http://openlibrary.o...culation_models

Bit of a long shot, perhaps, but you could try your local Lending Libray. If it's in their system in the UK, you can order it through them.
7&Y

[Boar Wrinklestorm]

Yup - time for another battering on my credit card

[Boar Wrinklestorm]

Captain_Bobski, on 15 March 2010 - 12:24 , said:

Important to note (or rather to emphasise) that this previous use of the leaky bucket, Cox et al 1999, does not use the leaky bucket the same way that we are using it. We, of course, are using the leaky bucket to model the total heat capacity of the entire planet, whereas Cox taks about its use within the Earth climate system.

Indeed - it might even be construed that it's even more general than that since we are using the arbitrary count of sunspots of which might have any number of hitherto unknown effects which, of course, might indeed be the reason for the latency.

An example of why this might be the case is the recent tree-proxy data that links to solar activity, but a year or so behind. Certainly a find, but it doesn't seem to exhibit the relationship that the LI does. That is to say, it seems almost discrete and boolean such that the latency is not variable.

(And a general theory of latency that does not rely on instant responses tuned to the speed of light; I mean who needs discrete packets when one can just hold a little bit back depending on how much you're already holding ..... Here comes the wave function ... )

Edited by VillagePlank, 15 March 2010 - 14:53 .

[Boar Wrinklestorm]

Those darn scientists - they never stop, do they?

A crucial groundbreaking paper is published in Nature, here, that describes research and the physical creation of a memristor. Essentially, as I understand it without actually reading the paper, this is inspired by Hodgin-Huxley model. As a consequence of this paper (PDF), an abstraction - the leaky integrator - can be used to model the action potential on a soma membrane.

Like the memristor, this provides a sort of capacitance, giving rise to a hysteresis type effect, and is the physical basis for biological neural processing, and neural memory.

More later when I get time to read the paper in Nature, but well done to those scientists.

Edited by VillagePlank, 09 April 2010 - 07:34 .

[Boar Wrinklestorm]

[jethro]

Any chance of an explanation of that graph for thicko here?

[Boar Wrinklestorm]

jethro, on 02 June 2010 - 15:09 , said:

Any chance of an explanation of that graph for thicko here?

Sorry about that - I've used this a place-holder to discuss a new angle of the LI with CBob. Please feel free to remove it

(I did mean to edit it out, myself....)