|
|
| Jan Pompe - 15/11/2007 07:27
Charles,
One of the reasons I mentioned acknowledging the data source is that not everybody will go back over to find it and some even forget if they had seen it.
Your model output shows banding where the charts produced by DeWitt show none after the 15 micron CO2 absorption band.
So who is right?
If you have left nothing out from the HITRAN data and I read the Boisoles pater correctly that area does in fact show banded transmission.
I would presume the hitran data is correct at least to a few percent. Now whether there is some other factor going on that is not covered by the hitran line data which I have not yet put into my model is of course a real possibility. It's also possible there is still some fundamental error in interpreting or implementing the hitran compatible code. However, I've been through it quite a bit recently and believe that it should be clean.
Assuming that Kiehl and Trenberth and their referenced sources are fairly close in their values, it would appear that my numbers are pushing almost 50% too high for outgoing and not much better for solar insolation.
I am assuming at present that the problem may be my water vapor numbers which undoubtedly include what's in clouds. Also my revised new model efforts are only clear sky so far. I would think things will go more towards Dewitt's notions when the model is corrected for clear sky h2o vapor and when cloud droplets with some sort of BB spectrum of liquid h2o are dealt with. Not to mention there is scattering - at least for shorter wavelengths not accounted for.
Currently, I'm in the middle of being bogged down with work projects and other events and the computer is really getting bogged down number crunching. And, my schedule promises to get much worse early next year. The computer just doesn't have the memory to deal with some of these huge excel files and the way excel works is not helping. It wasn't designed to work with spreadsheets containing over 2.5 million numbers, much less that many formulae cells.
So far though, it's less time consuming to keep at it this way than to write additional processing programs in c and then try to get them debugged.
|
|
| |
|
Location: Tennessee, USA | cba - 15/11/2007 11:57
Dewitt,
I'm doing both up and down radiation. I was concentrating on downward radiation while attempting to correct problems but both are equally important. There is still more absorption on the downward insolation than the accepted standard values, something like 120W/m^2 for clear sky versus a 70 to 90 W/m^2. Also, the values for surface radiated clear sky - as shown in my graph above total just under or over 50% absorption, something like 190 W/m^2. The values shown on your chart indicate transmission totals of between 15% and 30% for outward LW. Assuming a mix of clouds invoked, the total average absorption would rise significantly so it would depend upon the conditions assumed for that chart as to whether I'm above or below the absorption level it presents. Also, my graphs are using the exp(-tau)extinction rather than taking the optical depth Tau directly and lopping it off at 1.0. I'm not sure what your provided chart is doing or what was the resolution of the chart data used.
Looking at raw numbers, my line data is indicating there are tiny windows of several nm widths all around, surrounded by serious extinction areas.
Referencing Keihl and Trenberth's 1997 paper, they offer clear sky LW radiation absorption to the TOA as being 125 W/m^2 versus my 187 W/m^2 absorption. Their paper is a conglomeration of measured data from other researchers and calculated data they performed.
Charles
Charles,
The long wave flux measured at the top of the atmosphere of 265 W/m2 in Kiehl and Trenberth includes atmospheric emission as well as surface emission. That's why your absorption is too large. The graphs I posted are absorption only and probably include the effects of clouds. I'm posting the MODTRAN 3 calculated emission spectrum at 70 km for the 1976 standard atmosphere clear sky. Calculated emission is 259 W/m2. Emission at the surface is 350 W/m2 at 288 K instead of 390 W/m2 because this program uses a surface emissivity of about 0.92. As far as CIA for nitrogen and oxygen, their peak absorption is at 100 cm-1 (100,000 nm on your graph), which is where the MODTRAN graph ends.
Only the spectral range from 800 to 1200 cm-1 comes directly from the surface, as can be seen by the brightness temperature in that region above 280 K. The CO2 banded emission comes from about the tropopause at 220 K except for the spike in the middle at 667 cm-1 coming from CO2 emission in the warmer upper stratosphere. The same effect can be seen for ozone at 1000 cm-1. Note that emission from for wavenumbers less than 400 cm-1 doesn't come from the surface either. Water vapor partial pressure decreases much faster with altitude than CO2 partial pressure so emission comes from lower in the atmosphere where it's warmer. The data is also available in wavelength form, but you have to convert tables to plot it and I'm too lazy to do it.
Edited by DeWitt 16/11/2007 12:25
|
|
| |
|
| I'm not sure just how well all the analysis can really hold up completely. Beyond the surface (ignoring clouds and particulates - gas only) your BB curve of a good emissivity and .92 seems very low for the real emissivity at those wavelengths as virtually anything and everything radiates better than that - .92 is probably valid more for visible wavelength surface emissivity/albedo stuff on average. Oceans probably are more like .96 or higher at visible for higher angle insolation.
emissivity of gasses for a segment (small enough to be almost isothermal) of atmosphere should be quite low - meaning that very little is absorbed or emitted so emissions at higher altitudes are modulated by the line emission as well as by planck's law value for that T.
I've not yet gone through the process of determining absorption and emission at each section with my model so I don't know what the model will actually generate yet - except TOA from surface emissions. It will be time consuming to do that because there's too much data for the spreadsheet to function without getting into some virtual memory swap cycling disaster. I'm almost ready to do that but other obligations have impacted my time for the last week. I tried to do it in one spreadsheet but every manipulation of a cell resulted in a recalculation of about 5 to 10 seconds - and this was with almost all number values rather than formula values. It'd take 30 minutes to enter a new cell if they were all formula cells.
Consequently, I've got to make between 5 and 40 separate almost identical spreadsheets to break up the problem into something manageable by the computer so I'm probably looking at hours to do it.
cba
|
|
| |
|
Location: Tennessee, USA | Here are some links on continuum absorption by water vapor:
Theoretical Calculation of Water Vapor Continuum Absorption
Infrared water vapor continuum absorption at atmospheric temperatures (abstract only)
This one is mainly of interest because it has the dissociation energies for water-water and water-nitrogen dimers.
Water Vapor Continuum Absorption of Solar Radiation Tested Using ARM Data
Recent Developments in the Water Vapor Continuum
In order to compare the graphs in the above links to Boisselles et. al., you need to divide the Boisselles scale by the number of molecules in a cubic centimeter of gas or 2.7 x 1019.
I don't see 0.92 as 'very low'. An emissivity of 1 is only 40 W/m2 higher. The point is that for absorption only, your calculation is too low and atmospheric emission is, in fact, the majority of the total energy emitted at the TOA. This can only be calculated from a multi-layer model as the decline in temperature and water vapor concentration with altitude determines the emission characteristics.
Edited by DeWitt 16/11/2007 19:06
|
|
| |
|
| I expanded the line broadening wings out to 200nm on either side of a line from 10nm. H2O only absorption for absorption over the bottom 1 km of the atmosphere went from a total from 89 w/m^2 up to 180 w/m^2. Running all molecules again brought out a total of 219 w/m^2 for the bottom 1km of atmosphere and outbound radiation based on 288.2K surface temp. That provides 180 w/m^2 for h2o versus 39 W/m^2 additional absorption for all other molecules. Otherwise, temperatures, pressures and molecular concentrations are according to the std atm, 1976. It's taking 30-60 minutes for each run at present to get the raw data for a whole column and that doesn't include trying to combine and calculate for an entire column.
I have some concern over just what is going on with the line broadening as to whether it's putting in additional legitimate data or whether we're just adding a bunch of noise and error by exceeding limitations of the line broadening approach (hitran documentation).
I guess the next step is to calculate the total atm absorption and then if that is looking reasonable, try to calculate the layered absorption / emission and see how that looks.
Edited by cba 18/11/2007 01:06
|
|
| |
|
| Here is a TOA graph for radiation intensity from only the surface. It is output from my model as it presently exists. Changes include going out to +/- 200nm for line broadening and this is a true 1 dimensional model for line absorption. It is using the 1976 standard atmosphere values for about 35 of the Hitran molecules with pressure broadening and wavelength shifting done per the Hitran documentation. Temperatures, pressures, and molecular concentrations are from the 1976 Standard Atmosphere model, while isotope concentrations for particular molecules are from the Hitran data.
The atmosphere has been divided into quite a number of sections. There are 24 sections between the surface and 25km, each of 1km thickness. From 25km through 50km, there are 10 layers of 2.5 km thickness and from 50km through 120km, there are 14 layers. Calculations are done by line and are corrected for temperature and pressure and molecular concentrations at each layer.
The TOA graph is constructed by feeding the surface radiation attenuated by all previous sections to each additional surface.
The surface temperature is set for 288.2K with radiated output of 384.6 W/m^2 assuming emissivity for the surface = 1.0 and the final power total at the TOA is 107.4 W/m^2. This is over wavelengths from 0.200 microns to 65.5 microns. Resolution over this range is done at 1 nm (0.001 microns). The actual chart is graphed by taking the value at every 65th nm rather than taking min/max or average values.
best regards
cba
 (Surface-TOA1976.jpg)
Attachments
----------------
 Surface-TOA1976.jpg (70KB - 65 downloads)
|
|
| |
|
| Included below is the first result of a complete atmospheric column radiative transfer from surface to 120km from my one dimensional model. It includes the surface at 288.2k and the 1976 standard atmosphere generic average temperatures, pressures and molecular concentrations. It uses the Hitran molecular database. Calculations are done at 1 nm resolution from 200nm out to 65.5 microns. This is a clear sky calculation. Total outbound radiation amounts to 233.6 W/m^2 at the TOA.
 (totalTOAstdatm1976.jpg)
Attachments
----------------
totalTOAstdatm1976.jpg (78KB - 65 downloads)
|
|
| |
|
Location: Sydney Australia | Charles,
Nice chart. So it's the surface that is emitting at 288.2K what is emitting at ~220 K? |
|
| |
|
| Anywhere from 10km to 24km is in that vacinity although most of that is 217k in the middle. Even a little wider if one goes to 230k. I'm sure it's all contributing but just how much from each layer is something that would have to be specifically analyzed. The chart final calculation uses the radiative transfer function which is made by taking the surface emission, calculating the BB emission for the first km, added and attenuated by the first km, then fed into the 2nd km etc. on up to the top. Each layer depends upon all the previous layers. I would expect that the lower sections at a particular temperature contributes the most as the molecular density is the highest down lower.
I'm wondering how important it might be to regraph with x-axis using wavenumbers. That shouldn't be too difficult. It looks roughly like charts shown from ARM or nimbus on another thread.
It won't take nearly as long to do another calculation with differernt parameters but it is far from automated or quickly done. I'm anxious to try to do a co2 increase and compare output just to see what is going to happen.
|
|
| |
|
| cba - 22/11/2007 22:07
... below is the first result of a complete atmospheric column radiative transfer from surface to 120km from my one dimensional model. ... includes the surface at 288.2k and the 1976 standard atmosphere generic average temperatures, pressures and molecular concentrations. ... uses the Hitran molecular database. Calculations ... at 1 nm resolution from 200nm out to 65.5 microns. ... clear sky calculation. Total outbound radiation amounts to 233.6 W/m^2 at the TOA.
Are we at a point yet where we can calculate what increased CO2 levels does as to surface temperature due to 'bleed off' (via outbound longwave IR) of surface (earth) thermal energy?
This would be in pursuit of a sort of 'sensitivity' figure - looking at where *net* (surface received) inbound solar radiation minus CO2-influenced *net* outbound longwave IR would be affected by atmosphereic CO2 levels.
Given cba's numeical techniques it should now be possible for a first order approximation due to the widening/deepening of various absorption 'bands' in the outbound spectrum.
|
|
| |
|
| I did a rather simplistic effort this weekend, to the tune of many hours of work. The graph shown is for the 1976 std. atmosphere which has 330ppm of co2. I doubled that to 660 ppm and reran the calculations. About 2.7 W/m^2 additional absorption occurred for this doubling which means at the TOA there is 2.7 W/m^2 radiating out into space for this additional 330ppm of co2 - assuming no changes anywhere in temperature. That implies there is a need for some increase in temperatures to rebalance what was assumed to be balanced.
Taking the simple approach of ascertaining what the surface temperature and bottom 1km of atmosphere would have to increase by to meet the change, the results turned out to be a 1.5 or 1.6 degrees K increase. In reality, I would expect that the rebalance would be distributed more over the atmosphere based upon where the additional energy (both inbound and outboud) would be absorbed, providing the power needed for increased temperatures and radiation rates. Note that in the lower atmosphere there are already additional energy transfer mechanisms at work transferring amounts of energy up well beyond 1km. Also, this model at the moment is only for clear skies and will be substantially affected by cloud cover variations.
The net result is this is showing just almost 235 w/m^2 which is just about the averaged solar flux absorbed by the earth/atmosphere system. Small variations in the least digit aren't to important as the accuracy of a few percent is all that is possible even with the Hitran database and there is also still some really far IR energy out past 65.5 microns which is being ignored and that will account for some of the missing energy although it will essentially be cancelled out in the differential values between 2xco2 and the std atm.
It would seem that validating the reasonableness of the graph is probably of prime importance as compared to experimental measurements. After that, attempting to do an energy budget by layer might offer a good path to take. Using the baseline 1976 std atm, it should be possible to determine total energy to each layer as achieved by radiant and other mechanisms. Assuming similar values for the 'other' it should be possible to then establish the new temperature/emissions function of altitude and ascertain a reasonable and more accurate result for balance in the 2xco2 realm.
On the other hand, since this is clear sky only, it might be more important to first create a suitable cloud model to incorporate an 'average' column or a corresponding average cloudy model prior to dealing with the energy budget and attempting to create a new lapse diagram. I haven't come to any conclusions yet as to which might be the most suitable approach.
As for sensitivity it would seem the model is indicating we'd have a clear sky world sensitivity of 0.6 K/ W/m^2 and a total for a doubling of 2.7W/m^2. That puts it to be rather insensitive compared to most of the 3d models, but not all that far off from a number of other analyses attempting to determine that value. It's also possible that after dealing with clouds, that number might rise into the lower levels of the generally accepted range which mostly consists of values that tend to be a bit outrageous and do not seem to be based in reality.
cba
|
|
| |
|
| Thanks cba; that's exactly what I was getting at.
I take it then that the "About 2.7 W/m^2 additional absorption occurred for this [CO2] doubling" figure is due to the difference of the areas that have been integrated 'under the curve' (out to 65.5 microns anyway) for each CO2 case?
Edit: spelling
Edited by _Jim 26/11/2007 13:38
|
|
| |
|
Location: The Netherlands, -2,3 m msl | Charles, nice work. Question, again the humidity which is of course directly affecting water vapor feedback. For ballpark figures, from slide 6 here let's assume average annual evaporation of a meter per year. That's 2.74 liters (2740 g) per m2 per day or 114 g per hour is 0.032 gram per second. It takes 2500 joule to evaporate one gram of water, so for 0.032 gram that's 79 joule per second per square meter or 79 W/m2 Now to keep relative humidity constant when increasing the ambient temperature of 15 C to 16 C, suppose a dewpoint of about 9 degrees we see here a decrease of 67% to 63%. Obviously we also have to raise the dewpoint one degree to get back to 67% Now the absolute humidity calculated here goes from 9 gram/m3 at a dewpoint of 9 degrees to 9.6 gram/m3 at a dewpoint of 10 degrees, an increase of 7%. To sustain an increase of 7% more water vapor in the atmospere it seems logical that the rate of evaporation also has to increase by 7% as well, which in turn requires 7% more energy. Hence I'd need 7% of 79 W/m2 or 5.5 W/m2 extra to maintain constant relative humidity. So how much excess energy is there to get that positive water feedback in? or did we just debunk that?
Edited by andre 26/11/2007 15:15
|
|
| |
|
| _Jim - 26/11/2007 13:33
Thanks cba; that's exactly what I was getting at.
I take it then that the "About 2.7 W/m^2 additional absorption occurred for this [CO2] doubling" figure is due to the difference of the areas that have been integrated 'under the curve' (out to 65.5 microns anyway) for each CO2 case?
Edit: spelling
For this simplistic scenario, the 2.7 is the difference between the resulting integrals which are in this case, sums over the wavlengths.
My guess is that taking a subtraction of one from the other would result in little spikes around existing bands and perhaps some residual spikes from lines that gained a bit.
|
|
| |
|
| Andre,
While it likely has some important effects, I'm concentrating at the moment only upon radiative transfer. Water vapor is going to be another complimentary heat transport mechanism which will have effects, assuming overall temperatures change. The water vapor will absorb energy at the surface and rise up, carrying the energy aloft to higher altitudes where it ultimately transfers out and permits the water to rain out as there must be balance with the amount of h2o in the atmosphere.
The ultimate consequence of this may well be to moderate the lower atmospheric temperatures while raising mid level atmospheric temperatures. This could result in lower surface temperatures than would otherwise be required. Ultimately, the temperature realignments must occur so that total incoming power is radiated back out and also must be according to where the extra power winds up as without the extra power absorption or transfer, the temperature increase to support the added radiative output cannot be maintained. Of course, the less warming at the surface, the less change in the water cycle so it will be some sort of trade-off which will result in some increase in surface temperature - presuming that the increase in absorption of incoming and outgoing radiation has an effect there - which is the most reasonable presumption.
cba
|
|
| |
|
Location: The Netherlands, -2,3 m msl | cba - 26/11/2007 22:40 . The water vapor will absorb energy at the surface and rise up, carrying the energy aloft to higher altitudes where it ultimately transfers out ... Two observations: The latent heat carried aloft will be radiated out at water vapor frequency bands, not at CO2 frequency bands. So apart from some effect in the fringes, the increase of CO2 is hardly affecting this process The much lower warming in the lower troposphere but especially the cooling in the tropics seems to contradict the idea of more warming in the lower troposhere due to an increased pace of the water cycle as can be seen in the radiosonde data in the lower diagram, defeating the models.
Edited by andre 27/11/2007 04:26
 (fig5-7.JPG)
Attachments
----------------
fig5-7.JPG (66KB - 61 downloads)
|
|
| |
|
| But there still has to be additional energy coming in that raises the temperature somewhat to achieve an increase in the water cycle. Any real 'feedback control system' is going to have some variation in order for there to be a 'control function' that responds to a change and a 'correction' applied. I personally think that while your watervapor mechanism is probably quite important, that much of the major feedback is actually clouds (specialized water vapor) which is going to have a more substantial effect than merely a different set of absorption/emission lines. The clouds will alter albedo for incoming solar as well as affect the outgoing radiation. These will also radiate in a different fashion from the water vapor gases, evidently more of a continuum. It's not something I've read about yet but with the possible/probable exception that the droplet size may have an effect on the spectrum, clouds should be radiating more than the gas or vapor form.
Already, clouds account for 2/3 of the albedo. They are sporadic and will vary with temperature changes, throwing in serious uncertainties for those desiring to make detailed models. Also, when a situation arises that creates snow an glaciers, this will mimick the cloud formation 'control variable' and provide the mechanism for having ice ages last for long periods of time while reducing the available water vapor present so as to minimize its ability to provide the feedback mechanism. I guess this is pretty much Lindzen's iris concept and it also tends to stress the importance of that Cern Cloud experiment by Svensmark and others and provides links to the solar magnetic field variations, comsic rays and the consequences of the rest of the universe on earth's climate.
cba
|
|
| |
|
Moderator
Location: Graham, North Carolina | Hey Charles:
AH HA! And now you have uncovered one of the Earth's dirty little secrets that all the models are missing, Wind! The change in vapor content versus heat content will result in a change in the the air flow or wind. With higher temperatures lifting water vapor to higher altitudes you get high surface wind speeds which both change surface ocean currents and increase the radiative heat flow to a degree depending across the difference in the ambient temperature of the air source from which the wind is coming and the surface it crosses. (Keeping in mind that the the source temperature is gauged by the barometric region that is acting as the source at the time.)
Hence, if you look at the recent data for the Arctic there is a marked average wind speed increase north of the Arctic circle over the last 30 years. This increase has participated in both the increase of ice floe break up and pressure ridges forming in the Arctic Sea Ice. Wind also participates in the increase in the intrusion of the warmer ocean wind driven surface water towards the Eastern side of Greenland and the overturning of the Arctic ocean changing the sea current patterns, the biota concentration and distribution, along with the sea over turn characteristics we are seeing. Okay, enough said, all of this is BS or conjecture as it is not supported by actual observation, please disregard this rant...
Dave |
|
| |
|
| cba - 27/11/2007 14:09 But there still has to be additional energy coming in that raises the temperature somewhat to achieve an increase in the water cycle. Any real 'feedback control system' is going to have some variation in order for there to be a 'control function' that responds to a change and a 'correction' applied. No there doesn't have to be some extra energy coming in, because if there were the the temperature would continually rise. Only a little heat needs to come in for a short time to raise the temperature until input and output balance again. That means that if the incoming solar energy is not changing, a good approximation, then after we have increased the greenhouse gases the same energy must be emitted from the Earth's atmosphere as before. The problem is that the Earth-atmosphere is not a real 'feedback control system". There is no control feedback designed into it. (Intelligent Design is just a Creationist idea.) CO2 absorption is more or less saturated, and by Kirchhoff's then so is emission to space. Thus if we melt the sea ice and reduce the albedo, the surface temperature will keep rising forever until some else thing happens. This something will be a change in cloud cover. The additional clouds will have to match the lost sea ice, in order for the unchanged solar radiation to match the unchanged outgoing long wave radiation. This will only happen with a change in global circulation. Interesting eh? Cheers, Alastair. |
|
| |
|
Location: The Netherlands, -2,3 m msl | The problem is that the Earth-atmosphere is not a real 'feedback control system". There is no control feedback designed into it.
Nature is full of feedback systems, no need to design anything. No intelligence involved. Whenever the output of a natural system affects the input. Predator to prey ratios for instance, pure feedback. The latent heat in the water vapour cycle has three feedback properties. A positive feedback due to increased greenhouse effect of more water vapor by warming, which is happily quoted by warmers and negative feedback both by the exponential increasing energy (Clausius Clappeyron) taken by the evaporation of more water to keep the humidity up and diurnal day time cloud forming. But Charles is right, in a semi proportional feedback loop, (versus differentiating and intergrating feedback) there must be a deviating from equilibrium to set off the feedback. So the little warming of more CO2 will remain albeit reduced by the net negative feedback.
The additional clouds will have to match the lost sea ice,
In general I believe that the albedo effect of sea ice is a little exagarated. Mind that the incoming solar radiation has to be multiplied by the consine of the lattitude. Hence a cloud at the equator has double effect reflecting visual light than sea ice at 60 degrees. The albedo of sea ice at 90 degrees lattitude is obvliosuly irrelevant in wintertime. |
|
| |
Simple Radiative Models of the Atmosphere
