Son of Mulder.

So you now see what I mean about 'changes in scientific notation', et cetera.

Regards, suricat.

quote:

What is clear to me is that no matter how simple and obvious it seems to Steve_M & Cobbly Worlds - it aint.

I've never claimed that it was simple.

I've had another look at the diagram you linked to (you've linked to it before). This is my understanding of it:

Ignoring reflection, primarily it shows energy passing between three different "things": the earth's surface, the atmosphere and space/sun. The energy from evapotranspiration and thermals is convective energy into the atmosphere. If surface temperatures rise then convection increases, which increases the amount of energy in the atmosphere - the atmosphere warms (obvious).

The amount of energy radiated to space is a function of the temperature of the upper troposphere, so depends on how the extra energy is distributed in the layers of the atmosphere. The majority of models show the upper troposphere warming slightly more than the surface, so if a 1C rise in the upper troposphere is required to reach balance, a smaller than 1C rise is required at the surface.

If there is an extra watt per metre squared of convection at the surface, the effect on outgoing longwave depends on how high the extra energy gets. And the more convection there is, the less radiation there will be (because the convection cools the surface).

Convection is parametrized to match the weather observations, including those of the vertical structure of the atmosphere.

Furthermore, satellite observations of outgoing radiation can be compared with surface temperatures to check this.

(As an aside greenhouse gases also increase the opaqueness of the atmosphere, which means the radiation is emitted into space from a higher and colder layer).

Steve_M

quote:

Convection is parametrized to match the weather observations, including those of the vertical structure of the atmosphere.

Thanks for the narrative on the diagram. Just one question - how does your statement above differ from "Convection is parametrised to ensure the models of the climate are consistent with the weather observations"? ie certain component theories are assumed true and modelled any differences with reality are patched by the parametrisations. Which would mean that AGW is a self fulfilling prophesy.

quote:

"Convection is parametrised to ensure the models of the climate are consistent with the weather observations"? ie certain component theories are assumed true and modelled any differences with reality are patched by the parametrisations. Which would mean that AGW is a self fulfilling prophesy.

That could be true if the convection parametrization were a function of CO2 concentration. But they aren't.

Convection is (in the model with which I'm familiar) parametrized based on an understanding of the physics of water vapour, ice, thermodynamics etc. and on current observations of current weather phenomena - ie. lots of obs including folk getting into aircraft and sampling the insides of clouds etc. The development of convection parametrizations are driven by the need to get the weather forecasts right.

The current climate offers a wide range of conditions under which convection schemes can be tested, so you can assume that under a slightly warmer climate the scheme will behave just as well.

Steve_M

quote:

The current climate offers a wide range of conditions under which convection schemes can be tested, so you can assume that under a slightly warmer climate the scheme will behave just as well.

But that says they are tested in a place, at a temperature under those local conditions. You then have to integrate over the whole earth to get the total convection contribution. But as the whole earth's climate is changing because of anthropic CO2 how can that integration be achieved without building some sort of relationship between CO2 and convection (albeit via say temperature distribution). So if you want CO2 not to be a self fulfilling prophesy how can convection still be parametrised without making assumptions about the system to which it is fundamentally linked?

1) For two centuries medieval temperatures were higher than present, such that vinyards flourished in the North of England. There were of course no cars or planes then. Greenhouse gases due to man’s activities were infinitesimal at that time.

2) During the period 1645–1715, right in the middle of the Little Ice Age, solar activity as seen in sunspots was extremely low, with some years having no sunspots at all. This period of low sunspot activity is known as the Maunder Minimum. The precise link between low sunspot activity and cooling temperatures has not been established, but the coincidence of the Maunder Minimum with the deepest trough of the Little Ice Age is suggestive of such a connection

3) The current period of global warming started 150 years ago, 50 years before cars and planes were invented. Global temperature actually dropped between 1940 and 1975 at a time when CO2 levels continued to rise. However during that period levels of solar eruptions fell. Graphs of solar activity correlate better with global temperatures than graphs of levels of CO2.

4) The troposphere should be warming if greenhouse gases are trapping the sun’s heat. It isn’t.

5) Long term records going back 600 million of years show rises in CO2 levels lag global temperature rises by 800 years thus showing rising CO2 levels are an effect of warming not a major cause. Reason being it takes 800 years to warm the oceans. Warmer oceans emit more CO2.

6) CO2 emissions from human activity are only a small proportion compared to emissions from natural sources of CO2 & 95% of greenhouse gas is water vapour.

SoM

I can assure you that convection schemes do not consider the gaseous constituents of the atmosphere at all (apart from water vapour).

Convection is a local phenomenon. Convection in the UK is not affected by global temperature distributions. Similarly, most other weather phenomena are local enough not to be influenced by a global trend. So if your convection scheme works well in UK winter and summer, tropical winter and summer and desert winter and summer then you've probably covered most of the situations that are likely to happen for the next 100 years. Some of these convection schemes have also been tested on Martian weather.

Welcome Mambit_giggler

Yes, that's a reasonably correct summary of the Great Global Warming Swindle programme, which unfortunately is wrong in a large number of respects. See lots of threads in this forum. Or it's even on Wikipedia now!

http://en.wikipedia.org/wiki/The_Great_Global_Warming_Swindle

quote:

Convection is a local phenomenon. Convection in the UK is not affected by global temperature distributions. Similarly, most other weather phenomena are local enough not to be influenced by a global trend. So if your convection scheme works well in UK winter and summer, tropical winter and summer and desert winter and summer then you've probably covered most of the situations that are likely to happen for the next 100 years.

But any climate change means that the distribution of convection rates will change and so the total heat removed globally by convection may change significantly with little increase in overall average temperature.

Found this, laughed my socks off.

In an intriguing Climate Change report in Science, Wentz et al. (2007) note that the Coupled Model Intercomparison Project, as well as various climate modeling analyses, predict an increase in precipitation on the order of 1 to 3% per °C of surface global warming. Hence, they decided to see what has happened in the real world in this regard over the last 19 years (1987-2006) of supposedly unprecedented global warming, when data from the Global Historical Climatology Network and satellite measurements of the lower troposphere have indicated a global temperature rise on the order of 0.20°C per decade.
Using satellite observations obtained from the Special Sensor Microwave Imager (SSM/I), the four Remote Sensing Systems scientists derived precipitation trends for the world's oceans over this period; and using data obtained from the Global Precipitation Climatology Project that were acquired from both satellite and rain gauge measurements, they derived precipitation trends for the earth's continents. Appropriately combining the results of these two endeavors, they then derived a real-world increase in precipitation on the order of 7% per °C of surface global warming, which is somewhere between 2.3 and 7 times larger than what is predicted by state-of-the-art climate models.

How was this horrendous discrepancy to be resolved?

Based on theoretical considerations, Wentz et al. concluded that the only way to bring the two results into harmony with each other was for there to have been a 19-year decline in global wind speeds. But when looking at the past 19 years of SSM/I wind retrievals, they found just the opposite, i.e., an increase in global wind speeds. In quantitative terms, in fact, the two results were about as opposite as they could possibly be, as they report that "when averaged over the tropics from 30°S to 30°N, the winds increased by 0.04 m s-1 (0.6%) decade-1, and over all oceans the increase was 0.08 m s-1 (1.0%) decade-1," while global coupled ocean-atmosphere models or GCMs, in their words, "predict that the 1987-to-2006 warming should have been accompanied by a decrease in winds on the order of 0.8% decade-1."

In discussing these embarrassing results, Wentz et al. correctly state that "the reason for the discrepancy between the observational data and the GCMs is not clear." They also rightly state that this dramatic difference between the real world of nature and the virtual world of climate modeling "has enormous impact," concluding that the questions raised by the discrepancy "are far from being settled." We agree. And until these "enormous impact questions" are settled, we wonder how anyone could conceivably think of acting upon the global energy policy prescriptions of the likes of Al Gore and James Hansen, who speak and write as if there was little more to do in the realm of climate-change prediction than a bit of fine-tuning.

I'm sure the usual suspects will say that their models are oh so good and oh so accurate. Well if that was my job I would to. Don't make it right though.

quote:

But any climate change means that the distribution of convection rates will change and so the total heat removed globally by convection may change significantly with little increase in overall average temperature.

I'm not sure what the confusion is.

Convection is a local phenomenon based on local weather conditions. If global temperatures increase, then every local instance of convection is affected. But the local changes are unlikely to be outside the bounds within which the convection scheme has already been tested. You can therefore sum up all the local changes to get a global change.

I don't know how important such a change is, but convection is just another method of getting energy absorbed by the surface of the earth into the atmosphere. The more energy moved by convection, the less radiation there will be and vice versa. The solar energy is absorbed by the earth, is transferred to and among the layers of the atmosphere by long-wave radiation or convection, then gets to space by radiation.

Steve_M,

quote:

I'm not sure what the confusion is.

No confusion on my part. I'm essentially making the point that in a non-linear, multicoupled open system it is dangerous to make assumptions that certain variables will behave like they have always behaved on a global scale.

You're implying that there is some built-in limiting process that is understood, that bounds convection overall to a predictable finite range. Having just read Billy Bumbly's posting I can add no more. Nice one Billy.
above

quote:

But any climate change means that the distribution of convection rates will change and so the total heat removed globally by convection may change significantly with little increase in overall average temperature.

I'm not sure what the confusion is.

Convection is a local phenomenon based on local weather conditions. If global temperatures increase, then every local instance of convection is affected. But the local changes are unlikely to be outside the bounds within which the convection scheme has already been tested. You can therefore sum up all the local changes to get a global change.

I don't know how important such a change is, but convection is just another method of getting energy absorbed by the surface of the earth into the atmosphere. The more energy moved by convection, the less radiation there will be and vice versa. The solar energy is absorbed by the earth, is transferred to and among the layers of the atmosphere by long-wave radiation or convection, then gets to space by radiation.

quote:

You're implying that there is some built-in limiting process that is understood, that bounds convection overall to a predictable finite range.

No I'm not. I'm saying that convection schemes have been tested in all domains that are likely to be experienced over the next 100 years.

wrt the Wentz paper. It's an interesting result, but precipitation is more strongly affected by natural variability than temperature, so to assume that the increase over the last 20 years (related to a 0.2C temperature increase) is entirely related to temperature may be incorrect, and the authors acknowledge this possibility.

If you look at the precipitation changes over the last century in, say, the IPCC report (Chapter 3 page 254) you will see little upward trend as compared with the stronger temperature trend, and there is plenty of science, including within Wentz's references, linking precipitation more strongly to changes in incoming shortwave radiation and volcanoes.

ipcc-wg1.ucar.edu/wg1/wg1-report.html

PS. ignore the duplicate post - I got a connection error when I submitted, and when I pressed refresh it got submitted again.

Steve_M

quote:

But the local changes are unlikely to be outside the bounds within which the convection scheme has already been tested. You can therefore sum up all the local changes to get a global change.

As you've repeated the post I shall assume it is intentional on your part and not a glitch in the system.

I conclude from your repeated assertion above that despite the fact that the models look forward 100 years you do not expect any parts of the earth to move outside of their local climate experience to date in any significant way.

quote:

I conclude from your repeated assertion above that despite the fact that the models look forward 100 years you do not expect any parts of the earth to move outside of their local climate experience to date in any significant way.

Erm...not quite. I do not expect any parts of the earth to move outside of the normal range for the earth in any significant way. Eg. If the temperature in Hyderabad today is 40C, humidity 50%, wind minimal, then the weather/convection will be similar to London in 2100 in the event that the temperature there hits 40C, humidity 50% and wind minimal.

Steve_M,

quote:

No I'm not. I'm saying that convection schemes have been tested in all domains that are likely to be experienced over the next 100 years.

To me that's the same as saying we know what we expect to happen so we'll parametrise convection so the the models predict what we expect will happen and also it won't differ from what has happened in the past.

Steve_M,

quote:

Eg. If the temperature in Hyderabad today is 40C, humidity 50%, wind minimal, then the weather/convection will be similar to London in 2100 in the event that the temperature there hits 40C, humidity 50% and wind minimal.

Yes but to integrate that over the world you need a predicted distribution of local conditions and convection is a contributor to the conditions hence can't just be parametrised.

quote:

To me that's the same as saying we know what we expect to happen so we'll parametrise convection so the the models predict what we expect will happen and also it won't differ from what has happened in the past.

Yes, in a model, convection will always behave in exactly the same way (for a given local temperature, humidity wind speed, etc.) That is a physically realistic proposition.

In case you didn't notice my comment above, the duplicate post was an accident.

quote:

Yes but to integrate that over the world you need a predicted distribution of local conditions and convection is a contributor to the conditions hence can't just be parametrised.

Is there a misunderstanding as to what exactly a parametrisation is?

The inputs and outputs of a parametrisation such as convection are meteorological values (temperature, humidity), so the results from the parametrisation are physically meaningful, and are fed into the other physical formulations and parametrisations that make up a model. So a convection scheme does contribute to the conditions.

quote:

Yes, in a model, convection will always behave in exactly the same way (for a given local temperature, humidity wind speed, etc.) That is a physically realistic proposition.

So in iterating te model you take initial conditions in a cell, readoff convection values from a table then in second iteration receive input from adjacent cells and provide input to those cells, predict new conditions and do more table lookups. Iterate for 100 years over the whold planet and have faith that the predicted outcomes are realistic. Afraid my credulity is stretched beyond reality.

quote:

The inputs and outputs of a parametrisation such as convection are meteorological values (temperature, humidity), so the results from the parametrisation are physically meaningful, and are fed into the other physical formulations and parametrisations that make up a model. So a convection scheme does contribute to the conditions.

I appreciate the method I just don't believe the results will be physically realistic after many iterations.

Afraid I'm still stuck with my comments many months ago about models of this complexity vs reality.

Yes I'd seen your note about the duplicate posting. Thanks.

quote:

So in iterating te model you take initial conditions in a cell, readoff convection values from a table then in second iteration receive input from adjacent cells and provide input to those cells, predict new conditions and do more table lookups. Iterate for 100 years over the whold planet and have faith that the predicted outcomes are realistic. Afraid my credulity is stretched beyond reality.

Well, it's nowhere near as simple as that. I checked the documentation of the convection scheme of the model I'm familiar with and it runs to 50 pages or so. The code runs to 84461 lines in 141 subroutines.

With constant atmospheric constituents, models run for 100 years in a stable manner producing a reasonable climate. When forced by 20th century observations they predict 20th century temperatures. So the predicted outcomes are definitely realistic.

But the length of the model run is not that significant so I wouldn't get hung up on the 100 years. The IPCC runs are done that way because they are based on emissions scenarios.

Alternatively, you can view a model as being a very complex energy balance equation. If you set up a model of today, then immediately doubled levels of CO2 you would immediately see impacts of that in the model evolution (ie. the 4W/m^2 imbalance from the radiation model due to CO2 absorption of infrared). You don't need to trust in 100 years of stability to see the potential impacts. You could then "add" 1C warming to the whole temperature profile of the whole globe, and (if the model doesn't collapse in a numerically unstable heap) the 4W should go away for a bit. But in a few timesteps, the amount of water vapour will build up due to increased temperature and a radiative imbalance will return due to the increased absorption by the extra water vapour.