Ultimate space simulation software

  • 1
  • 5
  • 6
  • 7
  • 8
  • 9
 
User avatar
Watsisname
Science Officer
Science Officer
Posts: 973
Joined: 06 Sep 2016
Location: Bellingham, WA

General global warming / climate chage discussion

20 Jun 2017 18:23

Oftentimes the justification in science is that "it happens to work". :)  In my university physics program I worked with a project for imaging the atomic structure of metal surfaces using a scanning tunneling electron microscope, and I saw how they make the probe tips.  It felt like the most redneck, questionably unscientific method for making such a high precision instrument that I had ever seen.  It happens to work.  And you test that it works by imaging with it.

So we see that reconstructions of temperature are not just mappings of temperature change at the location of proxy.  They are determinations of temperature by how the proxy behavior relates to the surrounding climate dynamics.  Oftentimes there are relationships that are widely separated in space.  But these aren't merely assumptions -- they are testable relationships.  Usually there is a good physical understanding for why they work the way they do, but even if there isn't, they make predictions that you can check.  They're also not perfect, and sometimes the nature of the relationship is "magic", but they happen to work for producing a sensible, coherent record of planetary temperature.  And you can test that this works in a variety of ways.

Uncertainty in this record is based on statistical analysis of the coverage and resolutions of the different proxies, and variations in the results between different reconstructions and methods (the "spaghetti graphs").  Confidence in the record comes from broad agreement between these reconstructions, with known sources of change in radiative forcing and the fundamental physics, and our knowledge of drivers and effects of climate change throughout Earth history.

The resulting picture is quite clear and robust.  Global temperature through the Holocene slowly declined, and now it is rising rapidly with the enhanced greenhouse effect.  The magnitude of this forcing now dominates all other natural factors, and changes in the greenhouse forcing are key to understanding not only this period, but much of geologic history.  Geology even provides us with a really good example of what happens as a result of this experiment we're now performing.
 
User avatar
midtskogen
Pioneer
Pioneer
Topic Author
Posts: 423
Joined: 11 Dec 2016
Location: Oslo, Norway
Contact:

General global warming / climate chage discussion

20 Jun 2017 23:55

Watsisname wrote:
Source of the post Oftentimes the justification in science is that "it happens to work"

I'm sure that has been said about augury as well.  When "it happens to work", science must require that you have such a large observational basis for that without a single exception so you can generalise by induction, or that you can describe a physical mechanism explaining why which has been independently verified.  For instance, as an example of modern augury, "when the swallows fly low, the weather will turn wetter".  Obviously, that fails generalisation, but we can propose a physical explanation.  The swallows hunt insects and insects will fly lower closer to safety when the humidity and the risk for rain increase.  We can test this and get reasonably good results, but it doesn't always mean that it's the best science.  Humans are masters in finding patterns, but they do not always exist.  That's how our minds are wired.  So we must be extra careful.

I don't know how the NH samples were justified as SH data, but there are pitfalls.  Let's say a NH record matches a more direct SH record.  It's tempting to use it as a confirmation of the SH record or to fill in gaps in the SH record.  Humans so easily ignore everything that doesn't match, so there's a selection bias here.  Then, can we trust a physical explanation for the relationship proposed after the correlation was found as much as an explanation first found and then confirmed by data?  And again, even if the link was proposed first, we tend to ignore findings that don't support it until we find something that apparently does.  We must also consider the possibility that the SH record is wrong, and therefore also the NH confirmation.  And any actual NH proxies which predict SH would have been overlooked.

Yes, the general picture is that the global temperature has declined through the Holocene, but the disagreement is about whether we can confidently say that there hasn't (or has) been a more abrupt temperature change in that period than the past century, or temperatures higher than what we currently measure.  The details are in question.

I don't intend to dismiss anything but direct measurements.  But in many cases I prefer to call it an educated guess rather than solid science.
NIL DIFFICILE VOLENTI
 
User avatar
Watsisname
Science Officer
Science Officer
Posts: 973
Joined: 06 Sep 2016
Location: Bellingham, WA

General global warming / climate chage discussion

21 Jun 2017 02:25

midtskogen wrote:
Source of the post For instance, as an example of modern augury, "when the swallows fly low, the weather will turn wetter".  Obviously, that fails generalisation, but we can propose a physical explanation.  The swallows hunt insects and insects will fly lower closer to safety when the humidity and the risk for rain increase.  We can test this and get reasonably good results, but it doesn't always mean that it's the best science.  Humans are masters in finding patterns, but they do not always exist.  That's how our minds are wired.  So we must be extra careful.

This is a good analogue.  If we discover this relationship that works "reasonably well" (passes some confidence threshold in a statistical analysis) then we might begin to think the relationship is physical rather than random or uncorrelated.  It could be a fluke, but the stronger the result the less likely it is one, and we can be quantitative about it.

This methodology is a form of pattern matching, and your critique of it is correct.  If that's all there was to it (no underlying physical understanding, no correspondence to prior knowledge, etc), then it would be rather poor science.  But that isn't all there is to it.

In the swallow-rain example, the relationship makes a number of predictions.  You provide an example of insect hunting.  Test it, see if it works.  That's not bad science.  That's great science.  You observed something, hypothesized about the underlying mechanisms, built a model which made new predictions, and then you tested them.  The results of those tests will indicate the strength of the model.

With this methodology for northern proxies producing southern temperatures, the reconstruction itself is a prediction.  A different reconstruction using different methodology acts as a test.  How well do they agree?  We don't necessarily expect perfect agreement (that would be quite surprising), but the better it is, the more confidence we have. 

Then we may consider what the reconstructions show, or don't show, and how that fits with prior knowledge.  I'll frame this as a couple of questions to think about:

How would a global change, comparable in magnitude and timescale to this one, not show up in the record, when the global changes that we know about (like volcanic eruptions, solar activity, orbital-obliquity cycle) do?  Why is this record consistent with our understanding of the physics of radiative forcings and how they have changed, if we suppose the real record includes such unknown, comparable magnitude changes?
 
User avatar
midtskogen
Pioneer
Pioneer
Topic Author
Posts: 423
Joined: 11 Dec 2016
Location: Oslo, Norway
Contact:

General global warming / climate chage discussion

21 Jun 2017 06:31

Watsisname wrote:
Source of the post  If we discover this relationship that works "reasonably well" (passes some confidence threshold in a statistical analysis) then we might begin to think the relationship is physical rather than random or uncorrelated.

Yes, but we must not get carried away.  Let's say I want to use swallows as a proxy for relative humidity.  And I have a dataset showing swallow altitude and relative humidity measured by instruments.  I do some number crunching and find a suitable smoothing of the data and a conversion function giving me an accuracy with 0.1 percent point.  Perfect!  Better than most hygrometers.  But I got carried away and I didn't consider that the dataset was for African swallows, not European.  And that in order to get the 0.1 accuracy I overfitted when I chose a 163 degree polynomial conversion function.  And it might been a stretch to use the swallows in my backyard field to determine the current NH relative humidity.

Exaggerations aside, there is no clear line between good science and educated guesses.

Watsisname wrote:
Source of the post How would a global change, comparable in magnitude and timescale to this one, not show up in the record, when the global changes that we know about (like volcanic eruptions, solar activity, orbital-obliquity cycle) do?

In the AR5 figures the current warming shows up clearly, but it's not that striking, is it?  As I said, the hockey stick has been downplayed.  And that's just the past 2000 years (hardly).  I think the SH records are too thin.  It also seems a bit odd that the proxies tend to agree a lot more in the instrumental period - does it mean that they're better calibrated or otherwise adjusted for this period?  Which would imply an indirect splice with instruments.
NIL DIFFICILE VOLENTI
 
willi55
Observer
Observer
Posts: 6
Joined: 05 Sep 2017

General global warming / climate chage discussion

05 Sep 2017 08:19

Global warming is a serious issue nowadays. I think society should pay more attention to this issue, 
we should do all possible to prevent it
 
A-L-E-X
World Builder
World Builder
Posts: 804
Joined: 06 Mar 2017

General global warming / climate chage discussion

05 Sep 2017 10:11

Yes nuclear+renewables is definitely the way to go.
  • 1
  • 5
  • 6
  • 7
  • 8
  • 9

Who is online

Users browsing this forum: No registered users and 1 guest