Archive

Archive for the ‘Modelling’ Category

Annual report on GFS GCM performance

January 30, 2015 Leave a comment

A mailing list I receive from UCAR brought news of an addition to the GFS forecast archive “New Dataset: NCEP GFS 0.25 Degree Global Forecast Grids Historical Archive” and links to a web page.

On browsing an excellent general interest snippet appeared

Image

Page 16/17, performance over time for a few parameters of various GCM and with the observation on the GFS useful horizon “Increase is about one day per decade”.

Review of GFS Forecast Skills in 2013
Fanglin Yang
IMSG – Environmental Modeling Center
National Centers for Environmental Prediction

Link is on the performance review page, 2013 rev. is PDF (2.7MB)

Read more…

Advertisements
Categories: analysis, Modelling, weather

Strong evidence for linear law removal of atmospheric carbon dioxide

September 25, 2013 Leave a comment
Image

Figure 1

Image

Figure 2

Figures 1 and 2 are demonstrating both northern[1] and southern[2] hemisphere decay from a Dirac injection[3] of a test signal. The consequent effect is very close to perfectly linear, proportionality between pressure and effect of pressure over more than an order of magnitude of data variation (hence linear law). This seems to destroy IPCC claim of a non-simple law. Deviation is <1%

In addition the effect is a simple low pass filter on all kinds of atmospheric carbon dioxide. A later article might cover this in detail.

Read more…

Anders Angstrom: Nocturnal radiation at various altitudes

September 20, 2013 Leave a comment
Image

The curves bring out some interesting facts that deserve special consideration.
For ordinary values of the humidity, the effective radiation has a maximum at 1 to 4 km. altitude. An increase of the humidity or a decrease of the temperature
gradient shifts this maximum to higher altitudes. The effective radiation gradient
is consequently positive at low altitudes and negative at high altitudes. — A. Angstrom

Fig. 14 from paper 100 years ago by Anders Angstrom on LWIR emission. This shows families of curves grouped by two different lapse rates.

Note: This Daedal Earth blog article was rapidly produced to make the PDF available for citation elsewhere. Content may change later.

Sub-extract from Smithsonian Miscellaneous Collections, Volume 65, Number 3, published 1915. is A study of the Radiation of the Atmosphere

Physicist Dr Anders Angstrom was the son of physicist Dr Knut Anstrom (radiation instrument inventor) the son of physicist Dr Anders Angstrom after whom the wavelength unit the Angstrom is named. Confusion is understandable.

Read more…

Trenberth’s missing variable

August 16, 2013 1 comment

Trenberth is widely alleged to have written things about missing heat, good headlines but not really what he wrote.

He was upset over being unable to explain a divergence between what he expected from models and actual data, complaining the data is inadequate, presumably something which was either not measured or inaccurate. The reader can find the precise words if they want.

I hold the opinion the data we have is poor, hence things such as US Standards body questions absolute accuracy of TSI instruments, we increasingly are unsure about what solar factors affect earth and other more technical issues.

I put it like this: the calculations involved do not cross check. Trenberth is admitting for all the vast amount of money spent, man lifetimes of effort, the data is still too poor to figure out what is wrong.

Since then hiding heat, the missing heat problem has been discussed widely, papers published and still no answer, at least that I have heard about.

The assumption is that AGW theory is correct but the earth as a grand calorimeter says no.

Read more…

Categories: controversy, entropy, Modelling

Length of Day, modelling the lunar and annual effect

April 14, 2011 1 comment

As a result of helping someone out with lunar effects in Earth length of day I wondered if a slightly more comprehensive version would work.

The result  is an interactive spreadsheet which might be useful.

I would not usually  produce such a monster, this is an xls >30Mbyte but it does include reconstruction of the lunar LoD signal, subtraction from the raw LoD and decimation for plotting. It is live, you can turn on or off individual terms/factors and see the effect.

This needs a great deal of explanation, codas, and so on.

For now here is the file lod-work

Do not try this unless you have a fairly large computer. Checked it works with gnumeric, openoffice/libreoffice. Excel should not have a problem but recursion is used.

Categories: Lunar, Modelling

ERBS TSI

March 6, 2011 3 comments

As part of an ongoing investigation I looked at the ERBS TSI data. This dataset is not particularly interesting and  as with all satellite data is far too short to say much. One interesting snippet did appear.

Creating a rough model posed some problems but in practice was simple.

First was the data has a Y2K corrupted date which was kindly sorted out by V.

The sampling is irregular (scattered time points) which the software can usually handle if slowly. The result is unremarkable.

I’ve shown a fore and hindcast which indicates the model is stable and sane. It will be somewhat wrong.

A surprise came when I looked at a paper associated with the dataset where the final sentence of the conclusion is “The fact that the measurements increased with time relative to the proxies  suggests the existence of a second TSI variability component with an amplitude greater than 0.04% (0.5 Wm-2), and with a period greater or equal to approximately 20 years.”
A minor model component: 20.0432969909    3.69480925972    0.124480125667

Period just over 20 years, amplitude 0.125 * 2 * sqrt(2) = 0.35W p-p

With such short data that will be way out and TSI certainly is far more complex in the long term.

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.81.834&rep=rep1&type=pdf

I’ve shown SORCE as well. This is not really TSI, is narrowband/is in my opinion incomplete [disputed, see comments]. It does however show the wide variation is measured values. Given the extreme difficulty in making these measurements the usual rule of thumb is sensible, the instrument must be at least an order more accurate and so for absolute maybe we know +-40W sq/m

Putting that into context 40W in 1360W is +-3% of absolute, not as easy as many will imagine and for a remote instrument is an intensely hostile environment would not be a surprise. Lets hope the instrument is returned to earth for post mission calibration checks. See the point?

“Validation of spacecraft active cavity radiometer total solar irradiance  [TSI] long-term measurement trends using proxy TSI least squares analyses

“aRobert Benjamin Lee III and bRobert S. Wilson
aNASA Langley Research Center, Atmospheric Sciences, MS 420, Hampton, VA 23681-2199
Science Application International Corp. (SAIC), One Enterprise Parkway, Hampton, VA 23666 ”

http://www.acrim.com/

A lot of interesting material on TSI  can be found here http://www.leif.org/research/

Categories: analysis, Modelling, solar