2011/06/07

The Total Cycle thing repeated

"With four parameters I can fit an elephant, and with five I can make him wiggle his trunk".
Attributed to von Neumann

The spreadsheet is available on request (11MB)


A plot of all the amplitudes and frequencies used. The 11 year cycle is TSI The scatter of low frequency low amplitude signals gives the noise



300 year and 60 year cycles superimposed on the constructed temperature record


A check to see if volcanic activity can be seen in the difference between actual HADCRUT3 and reconstruct



Remaider HADCRUT3-reconstruct with the used trailing moving average filtered and offset TSI

The reconstructed signal extended beyond the 2 ends of hadcrut3 note that the little ice age becomes a warm period using the 300year cycle

Acomparison of reconstruced cosines with no tren and reconstructed cosines + trend
Frequency/amplitude and phases of cosines + tsi used
Period
years
amplitudeOffset
months
0.50120.0360548.3
0.83950.006242.9
0.89100.0117731.23
1.76920.0115060.95
1.93730.0125122.7
1.99110.0203755.8
2.04090.0127129.55
2.10910.01076910.55
2.25690.0151959.55
2.31390.0173136.75
2.54180.0177424.4
2.67140.0171622.35
2.76080.0110355
2.87460.0272852.55
3.16380.0197670.96
3.27340.0154636.55
3.46990.014799.64
3.57940.023023.15
3.76840.0257723.35
4.01500.0191246.75
4.72400.0285626.68
5.10610.0201935.5
5.24660.0208889.55
5.72740.012946.65
5.97820.0249631.85
6.25160.0242294.35
6.61530.0167643
7.56930.0218273.2
8.42720.0199453.25
10.06930.0300616.7
11.00000.038659-69
11.75320.0134853.85
12.633604.55
14.94970.0300695
21.20830.0484144.45
35.68520.0264643.75
60.08330.1194323.05
101.83330.0266996.15
110.216705.35
290.27780134.3
317.05330.3957012.3
2508.166104.15
2508.333304.1

3 comments:

  1. Hey TFP,
    Have a look at our reconstruction of Lean TSI for some ideas on cycle periods.
    http://tallbloke.wordpress.com/2012/10/23/2011/02/21/tallbloke-and-tim-channon-a-cycles-analysis-approach-to-predicting-solar-activity/

    Cheers

    TB

    ReplyDelete
    Replies
    1. Hi TB
      You seem to have reconstructed a discredited TSI series. I believe even Lean now says this is not valid!

      However, the post on which you have replied usilises the usual Leif TSI plot (shown as 11.0000 in the table above) reconstructs HADCRUT temperatures.

      If you look at the latest plots (click on simulation) these have been made with around 5 frequencies The 60 year cycle is important as it is the one controlling the current flatlining. Unless you have a polynomial(*) to simulate the rising trend you have to include a large amplitude 317 year cycle. This then has the problem of fitting with temperatures in the 1800s

      I make no claim for the future predictions this is simply curve fitting!

      (*)
      240.080E-9*x^3
      -1.340E-3*x^2
      2.493E+0*x
      -1.547E+3

      Delete
    2. Hi TFP,
      It doesn't affect our result much if we use Lean 2K or Lean 2005, because the background shape of the curve is the same, just the amplitudes of the 11 year cycles change in relation to it. The TSI scale on the Y-axis is pretty much immaterial as there are other ways of suitably calibrating that.

      Delete