I was discussing the idea of an interactive pie chart showing ghgs on the discussion about trace gases with Mark R:

http://www.skepticalscience.com/thread.php?t=2567&r=48

I think I should have the web site I was developing for a friend finished this week (hopefully it will be uploaded to the server tomorrow) so am up for another challenge (at a more sedate pace this time) and have a go.

Mark R said:

I think the physics is doable.

We need an estimate of climate response to radiative forcing and can prescribe temperature changes based on the sensitivity or, if we want to see it through time, by typical feedback equations.

Then to calculate the change in water vapour it's just the Clausius-Clapeyron equation (RH has stayed pretty constant so far and models tend to expect it to remain like that).

The radiative transfer is the difficult part: we need access to atmospheric codes but I think we can do that. I suspect someone in my department has them if no-one else here can find them... then we can just create a multi-dimensional lookup table (or approximate equations) for different temperatures, WV, CO2, other greenhouse gases. Once you've defined your 'fraction' properly it should be doable from this.

So if someone can work out this stuff??
Meanwhile I'll re-read the discussion and have a think.

I have stuck it in the general chat area at least to start it off.

Just working on a mock up of the graphics for this idea based on Mark Rs comments.

Looking at using a word tag cloud instead of pie charts, it's looking good so far.

There's an interactive version of ModTran online.

Ideally we'd have something offline that we could run through shellscripts to produce a lot of data.

Quick mock up I produced this afternoon:

The processing would start when the slider is dropped, after being dragged. The X would show the number in parts per million. The background colour will be the same colour as the indicator on the right. So as the back radiation is increased then the colour would get redder. The size of the gas words would change depending on the gases proportion of the back radiation.

(I assume watts/m² is the correct units??)

Is there any interest in this??

And how hard or easy would it be to use Modtran to create the data??

Does my mock up generate any thoughts?

Hi Paul,

I've been eager to see where you are going with this project. I would like an interactive illustration that helps me get a handle on the nature of the calculations used in quantifying the greenhouse effect and a visual to explain some of the concepts. I think that's what you are doing, and if so, I need more explanation of what I'm supposed to understand from the tag cloud interface.

I'll give an example. In this illustration I tried to show the various greenhouse gases with their relative contribution to the greenhouse effect while also implying the major wavelengths each is active in. Something like this could have an interface where you can remove molecules via checkboxes next to each or even add other known molecules for other hypothetical atmospheres. E.g., I'd like to see it work for Titan's methane+ atmosphere. And of course, concentration of each could be changed. You could also add an adjustable sun and change latitude to see what different temperatures you'd get.

I was playing what another way to collect information about wavelengths received, absorbed and reradiated in this very unfinished diagram:

So, are we thinking of the same type of thing?

jg

jg,

In diagram 1:

Is the thickness the absorptive coefficient, and the angular variable the frequency? Is the radial coordinate the radial distance, or something else?

Hi JG I like your first diagram it makes a planet into a 'pie chart' and everyone can see what it is supposed to represent. I think it does need some refining though.

What are the red bars??

I think having two pictures of each molecule is a bit confusing. I think you need to have one for each gas, either within their 'slice' or somewhere else. Maybe have a key of some sort.

I think whatever approach is taken, assuming the main point is to show earths atmosphere, then the CO2 ppm should be indicated in some way to point out clearly that small amounts of CO2 can cause big changes. Effectively visually and interactively rebutting the trace gas meme.

BTW MarkR has implied that the changes in 'slice' sizes would be each gases back radiation contribution rather than the gases physical quantity. Do we have two conflicting views about what should be shown here? Or is it just that the output of the physics calculations gives the back radiation rather than quantities??

What do people think about making sure the maths and data behind the interactive animation is clearly published so that we don't get cries of 'it's a fix' or similar.

Forgive me for being slow to reply. My main intention was to keep Paul's topic alive. I'm very eager to see the work turn into something we can use. My own idea is half baked, and I hoped I'd find time to elaborate via illustration, but I won't be able to do so till tomorrow or Monday.

My original diagram, which I grabbed one frame of and shared above, showed the atmosphere's constituents as a pie chart (whose middle is blocked by the earth), so you have and N and O then Ar, and last a tiny sliver with all the GHGs. The sliver then becomes the whole diagram with the arc loosely representing the wavelengths at which the various gases are most active as absorbers. I imaging the image toggling between these views.

The thickness shows the percent aborbed at the implied wavelength. I figured that I could program a radial graph that would draw lines at each wavelength (rounded to a manageable unit) and extend their length as need to show percent of absorbtion. 

In describing this, I see that this is not showing contribution to back radiation that Paul originally proposed. Perhaps a third state would turn the graph into a total watts of back radiation where the arc represents back radiation and can grow a little if increasing GHGs require larger values. Each gas's contribution to back radiation would be shown by thickness an arc originating from the gas symbol on the left.

thanks,

jg

JG

I proposed what your diagram shows, eg. the pie slices determined by quantity of gases (%).

It was MarkR that suggested the Back Radiation version.

Maybe it would be possible to switch between the two modes??

Sounds like we have a common objective that we can work on. I'll be gone for a day, but I'll check in tomorrow to see anything you add. Shall I start drawing a hypothetical user interface?

Sure have a go.

I'm thinking Flash might be the best way to programme this up. It depends how the graphics turn out and how it can be broken down into usable parts. Or alternatively it is probably time to consider html5 and that route.

We need someone to volunteer to help with the science and data.

Hi JG

Have you dropped this idea?

Paul,

Here is a very unfinished diagram. I'm sharing it now just to show that I'm really working on the idea. It's been awhile so you would have reason to believe I've dropped it. I haven't, rather, I've had a lot conflicts lately.

This shows what one might get choosing the options "greenhouse gasss only" and "IR absorptions". The angular span of the arc represents a percentage of each gas's contribution. (I wasn't trying to be accurate, just quick to show the visual idea). Other options would change the graph, but I need to sketch those out.

jg

Looking good. Agree that it needs refinement.
The code could decide whether to show the molecule based on the room available.

Question is what would happen if the segment width is to small to show??
Maybe just show a single pixel wide line.

I think numbers for the amounts would have to be shown??

I'm a fan of minimum colour sets/ranges for graphics.
(added: I meant a limited/appropriate colour pallete)

Is Mark R going to help with the maths/science.

I'm up for the developer/coding role but I would need help with the maths/science.

Hi Paul,

If I had my fantasy program, it would start with these features:

Here I'm showing features that allow you to adjust both the planetary surface, the atmosphere, and the sun not only to imagine Earth at different times but to speculate about different planets. That's asking a lot for someone to program, but I figure there's no harm in asking if what you have in mind will scale up to this.

This sketch shows how an option to view atmspheric concentration would appear. I'll sketch how greenhouse effect may appear.

Question: what size do you picture this? I've constrained this sketch to the SKS 570-pixel width limit, but we don't have to stick to this. We could do a lot in 800 or more.

jg

I wouldn't worry about size. The bars and pies would be generated by the code as vector graphics, probably the only thing needed from you would be the planet graphic as a jpg or png file.
I like the redesign into a bar chart and the colours.

You have a yellow line that drops down to the earth surface that represents the bottom of the bar chart, when it passes the pie chart, can it represent all the ghgs and be a pie chart section?? The bar chart would of course remain as you have it. If you don't understand what I mean I can modify it to demonstrate what I mean.

Not sure about your additions, although I know what you are getting at.

We still need some science input either to work out some equations that can be used in the code, or to generate data tables to do a similar job. Mark R may be to busy, is there anyone else??

What does John Cook think??

OK, I wasn't sure if I was interpreting your design correctly.

I've been out of action for the past week, my girlfriend was visiting from Italy and all my science time was spent on work...

I'll play with the online Modtran a bit later. My concern is that getting it to match IPCC values is very difficult, because the IPCC averages globally but for ModTran each run can only pick a certain atmosphere type.

I think we might have to do it as an example in a 'US standard atmosphere'. What I need is an offline Modtran I can run shellscripts on to speed this stuff up :/

Ok Mark, what exactly will you be producing from Modtran?
I'm assuming some tables?

Also whatever it is, you'll have to give me some instructions or/and an example of the use of them/it.

I have to say btw that this idea is coming along and is begining to look impressive, I think the new graphics concept (the mix of pie and bar charts) get subject across really well.

I'll wait a little longer before I start thinking about code development.

There's a question of 'cause' as well that needs to be sorted and well explained.

Because of absorption line overlap, the heating of water alone + CO2 alone would be greater than water + CO2 heating (they 'share' some of the heating burden if you will).

You can define CO2's concentration in lots of different ways.

Total current greenhouse effect with 0 CO2 / total current GHE.

Or is it: Effect of CO2 ON ITS OWN / total current GHE.

Skeptics tend to use the first. Logically I think the second makes more sense since CO2 is long lived. Start with a very cold atmosphere and little or no CO2. At CO2 and you get warming/more vapour. Without CO2 the vapour wouldn't be there anyway...

Example way of changing the relative values:

Looking up @ 0.0001 km, US standard atmosphere, constant current day temperature:

Radiation coming down:

No greenhouse gases = 39.8 W m-2

375 ppm CO2, no WV = 99.6

375 ppm CO2, WV (current RH) = 256.9

0 ppm CO2, current WV = 228.247

So, you can take the total greenhouse effect as 256.9 - 39.8 = 227.1 W m-2

And if you take water vapour first and say CO2 is the leftover, then you get CO2 as 13% of the surface greenhouse effect.

If you take CO2 first and say water vapour is the leftover then you get CO2 as 26% of the GHE.

Physically it's more accurate to take water vapour as the leftover, or to account for the temperature (no CO2, temperature falls, atmosphere cools, less water avpour, cooling... ice age, near zero atmospheric CO2). Deciding how to define and communicate these values could be difficult.

There will need to be some supporting 'technical' notes that explains some (or a lot) of the methods/science.

I think one thing we need to clarify to start with is what can be 'set' by a control on the interface and what can be 'calculated' or 'looked up from data'.

Some issues that need answers...

1. Does CH4 and NO2 have to be set and fixed by a control?? I'm assuming they would, I find it difficult to see how they would be calculated at any point. Correct me if i'm wrong.

2. I assume H2O vapour quantities can be calculated based on warming or cooling. Hence will vary, but will need a control to set the initial value.

3. I'm inclined to only show the GHE/forcings on the bar chart and only show proportions of gases in ppm on the pie chart. Maybe there could be a supplementary pie chart in a window that shows only the GHG gases ppm? What do you think JG??

I think we need to nail down what will be displayed, what needs controls and what can be calculated/looked up.
Once we are happy with some of these issues, JG can re-work the graphics.
There is no doubt that compromises will need to be made.

You define the CH4/NO2 values at the beginning. It's purely radiative transfer, it doesn't calculate anything else! You just say 'this is what the atmosphere has got in it' and then it works.

You can set the surface temperature, water vapour (although the height distribution is calculated in the model) and other gases (ozone, CH4, CO2)

The free online one doesn't include any other gases. And it costs over $1,000 to get the proper program! I haven't found anyone with it yet, but I might bump into someone in my department who'd do it for me.

Looks like we are waiting to see what or if you can do Mark.