Australian Drought Predictions

Below are the results of applying the EMD algorithm (Empirical Mode Decomposition) to Australian Rainfall, and predicting the future rainfall with a VAR model (Vector Autoregression).

First, EMD splits the rainfall into IMF’s (Intrinsic Mode Functions) that are cyclical but variable in amplitude and frequency.

austr

You can see each of the modes, their strength and phase, and the uncertainty. The initial modes are very uncertain, not worth including in the model, so only a few IMF are used in the fit and prediction below.

austr1

You can see the drought conditions prevailing in the last few years, and the strong rebound in precipitation predicted from the model. Note the uncertainty limits expand very quickly, indicating that only a decade or so ahead is predictable before errors in each of the IMF’s overwhelms. See Demetris’s excellent paper on this for more information.

This model has a Nash-Sutcliffe Efficiency of around 0.1, better than the mean value, but not huge. Another way to evaluate it is to do the prediction to a specific date, leaving out some data. The following was done to 1980.

austr80

Clearly there is some skill at predicting peaks, but should not be relied on at this stage.

I am still working on the code for this, and will post it shortly, but I call it with a function:

>myemd2(Ar,stationary=T,N=3,end=1990)

The Ar is the time series for the rainfall, stationary=T lets me exclude any trend from the model (as trends in rainfall are very uncertain), N is the number of IMF’s to use, and end=1990 is where the data is used to.

Advertisements

0 thoughts on “Australian Drought Predictions

  1. David, the work you are doing here and the paper by Demetris Koutsoyiannis are streets ahead of the IPCC effort. If you are allowing quotes from the CRU hack (which Prof Jones has confirmed as accurate), then you can see in the following (long) post by a very experienced climatologist of the old school, the need for the type of analysis that you guys are doing. (I have removed an email mailing list of about 100 IPCC authors for brevity. There are about 10 Australians among them. One wonders why they should not have done better).1202939193.txtFrom: J Shukla <shukla@cola.iges.org>To: IPCC-Sec <IPCC-Sec@wmo.int>Subject: Future of the IPCC:Date: Wed, 13 Feb 2008 16:46:33 -0500Dear All,I would like to respond to some of the items in the attached text on issues etc. in particular to the statement in the section 3.1.1 (sections 3: Drivers of required change in the future).”There is now greater demand for a higher level of policy relevance in the work of IPCC, which could provide policymakers a robust scientific basis for action”.1. While it is true that a vast majority of the public and the policymakers have accepted the reality of human influence on climate change (in fact many of us were arguing for stronger language with a higher level of confidence at the last meetings of the LAs), how confident are we about the projected regional climate changes?I would like to submit that the current climate models have such large errors in simulating the statistics of regional (climate) that we are not ready to provide policymakers a robust scientific basis for “action” at regional scale. I am not referring to mitigation, I am strictly referring to science based adaptation.For example, we can not advise the policymakers about re-building the city of New Orleans – or more generally about the habitability of the Gulf-Coast – using climate models which have serious deficiencies in simulating the strength, frequency and tracks of hurricanes.We will serve society better by enhancing our efforts on improving our models so that they can simulate the statistics of regional climate fluctuations; for example: tropical (monsoon depressions, easterly waves, hurricanes, typhoons, Madden-Julian oscillations) and extratropical (storms, blocking) systems in the atmosphere; tropical instability waves, energetic eddies, upwelling zones in the oceans; floods and droughts on the land; and various manifestations (ENSO, monsoons, decadal variations, etc.) of the coupled ocean-land-atmosphere processes.It is inconceivable that policymakers will be willing to make billion-and trillion-dollar decisions for adaptation to the projected regional climate change based on models that do not even describe and simulate the processes that are the building blocks of climate variability. Of course, even a hypothetical, perfect model does not guarantee accurate prediction of the future regional climate, but at the very least, our suggestion for action will be based on the best possible science.It is urgently required that the climate modeling community arrive at a consensus on the required accuracy of the climate models to meet the “greater demand for a higher level of policy relevance”.2. Is “model democracy” a valid scientific method? The “I” in the IPCC desires that all models submitted by all governments be considered equally probable. This should be thoroughly discussed, because it may have serious implications for regional adaptation strategies. AR4 has shown that model fidelity and model sensitivity are related. The models used for IPCC assessments should be evaluated using a consensus metric.3. Does dynamical downscaling for regional climate change provide a robust scientific basis for action?Is there a consensus in the climate modeling community on the validity of regional climate prediction by dynamical downscaling? A large number of dynamical downscaling efforts are underway worldwide. This is not necessarily because it is meaningful to do it, but simply because it is possible to do it. It is not without precedent that quite deficient climate models are used by large communities simply because it is convenient to use them. It is self-evident that if a coarse resolution IPCC model does not correctly capture the large-scale mean and transient response, a high-resolution regional model, forced by the lateral boundary conditions from the coarse model, can not improve the response. Considering the important role of multi-scale interactions and feedbacks in the climate system, it is essential that the IPCC-class global models themselves be run at sufficiently high resolution.Regards,Shukla

  2. Yes sure, lets talk about the science.Straight answers to three questions: 1. How confident are we about the projected regional climate changes? [Not at all].2. Is “model democracy” a valid scientific method? [No]3. Does dynamical downscaling for regional climate change provide a robust scientific basis for action? [No]Climate modellers have failed those who need climate forecasts most by chasing after a small human signal, that from the emails, they are not sure rises above natural variation.

  3. David, the work you are doing here and the paper by Demetris Koutsoyiannis are streets ahead of the IPCC effort. If you are allowing quotes from the CRU hack (which Prof Jones has confirmed as accurate), then you can see in the following (long) post by a very experienced climatologist of the old school, the need for the type of analysis that you guys are doing. (I have removed an email mailing list of about 100 IPCC authors for brevity. There are about 10 Australians among them. One wonders why they should not have done better).

    1202939193.txt
    From: J Shukla
    To: IPCC-Sec
    Subject: Future of the IPCC:
    Date: Wed, 13 Feb 2008 16:46:33 -0500
    Dear All,

    I would like to respond to some of the items in the attached text on
    issues etc. in particular to the statement in the section 3.1.1
    (sections 3: Drivers of required change in the future).

    “There is now greater demand for a higher level of policy relevance in
    the work of IPCC, which could provide policymakers a robust scientific
    basis for action”.

    1. While it is true that a vast majority of the public and the
    policymakers have accepted the reality of human influence on climate
    change (in fact many of us were arguing for stronger language with a
    higher level of confidence at the last meetings of the LAs), how
    confident are we about the projected regional climate changes?

    I would like to submit that the current climate models have such large
    errors in simulating the statistics of regional (climate) that we are
    not ready to provide policymakers a robust scientific basis for “action”
    at regional scale. I am not referring to mitigation, I am strictly
    referring to science based adaptation.

    For example, we can not advise the policymakers about re-building the
    city of New Orleans – or more generally about the habitability of the
    Gulf-Coast – using climate models which have serious deficiencies in
    simulating the strength, frequency and tracks of hurricanes.

    We will serve society better by enhancing our efforts on improving our
    models so that they can simulate the statistics of regional climate
    fluctuations; for example: tropical (monsoon depressions, easterly
    waves, hurricanes, typhoons, Madden-Julian oscillations) and
    extratropical (storms, blocking) systems in the atmosphere; tropical
    instability waves, energetic eddies, upwelling zones in the oceans;
    floods and droughts on the land; and various manifestations (ENSO,
    monsoons, decadal variations, etc.) of the coupled ocean-land-atmosphere
    processes.

    It is inconceivable that policymakers will be willing to make
    billion-and trillion-dollar decisions for adaptation to the projected
    regional climate change based on models that do not even describe and
    simulate the processes that are the building blocks of climate
    variability. Of course, even a hypothetical, perfect model does not
    guarantee accurate prediction of the future regional climate, but at the
    very least, our suggestion for action will be based on the best possible
    science.

    It is urgently required that the climate modeling community arrive at a
    consensus on the required accuracy of the climate models to meet the
    “greater demand for a higher level of policy relevance”.

    2. Is “model democracy” a valid scientific method? The “I” in the IPCC
    desires that all models submitted by all governments be considered
    equally probable. This should be thoroughly discussed, because it may
    have serious implications for regional adaptation strategies. AR4 has
    shown that model fidelity and model sensitivity are related. The models
    used for IPCC assessments should be evaluated using a consensus metric.

    3. Does dynamical downscaling for regional climate change provide a
    robust scientific basis for action?

    Is there a consensus in the climate modeling community on the validity
    of regional climate prediction by dynamical downscaling? A large number
    of dynamical downscaling efforts are underway worldwide. This is not
    necessarily because it is meaningful to do it, but simply because it is
    possible to do it. It is not without precedent that quite deficient
    climate models are used by large communities simply because it is
    convenient to use them. It is self-evident that if a coarse resolution
    IPCC model does not correctly capture the large-scale mean and transient
    response, a high-resolution regional model, forced by the lateral
    boundary conditions from the coarse model, can not improve the response.
    Considering the important role of multi-scale interactions and feedbacks
    in the climate system, it is essential that the IPCC-class global models
    themselves be run at sufficiently high resolution.

    Regards,
    Shukla

    • Yes sure, lets talk about the science.

      Straight answers to three questions:
      1. How confident are we about the projected regional climate changes? [Not at all].
      2. Is “model democracy” a valid scientific method? [No]
      3. Does dynamical downscaling for regional climate change provide a robust scientific basis for action? [No]

      Climate modellers have failed those who need climate forecasts most by chasing after a small human signal, that from the emails, they are not sure rises above natural variation.

  4. Yes, the science that comes through the hacked emails is not sophicticated. While there are some bright authors, the mangle mill of usual suspects reduces it to quite a mess. Recent science advances desperately need to be brought to the attention of emerging researchers. To the extent that blogging will help, that is excellent. The rambling email I quoted is far from the leading edge. I keep coming back to the concept of a series of neutral-ground conferences whose objectives are to mutually reveal where the cutting edge is and where the funds need to be directed. The tone of the hacks leaves little hope for that with the current generation of IPCC organisers.Your example above using rainfall has some advantages over temperature, which is a more prevalent test bed. Rainfall is bounded at the lower end and observationally, if a rain gauge is not read for a few days the signal has accumulated, not disappeared, and can be spread with little error. Is there a way in which these 2 properties can be used to further refine the graphs (which you might label a bit more explicitly for us) or has that implicitly been done already? Measurement uncertainty at a site should be less for rainfall than for temperature and the signal should show better through the noise.

  5. Yes, the science that comes through the hacked emails is not sophicticated. While there are some bright authors, the mangle mill of usual suspects reduces it to quite a mess. Recent science advances desperately need to be brought to the attention of emerging researchers. To the extent that blogging will help, that is excellent. The rambling email I quoted is far from the leading edge. I keep coming back to the concept of a series of neutral-ground conferences whose objectives are to mutually reveal where the cutting edge is and where the funds need to be directed. The tone of the hacks leaves little hope for that with the current generation of IPCC organisers.

    Your example above using rainfall has some advantages over temperature, which is a more prevalent test bed. Rainfall is bounded at the lower end and observationally, if a rain gauge is not read for a few days the signal has accumulated, not disappeared, and can be spread with little error. Is there a way in which these 2 properties can be used to further refine the graphs (which you might label a bit more explicitly for us) or has that implicitly been done already? Measurement uncertainty at a site should be less for rainfall than for temperature and the signal should show better through the noise.

  6. Pingback: list your startup

  7. Pingback: zobacz

  8. Pingback: bateria do laptopa toshiba

  9. Pingback: kliknij tutaj

  10. Pingback: warsztaty kulinarne

  11. Pingback: Link Indexr by josh zamora

  12. Pingback: socially responsible fashion

  13. Pingback: singers

  14. Pingback: google sniper

  15. Pingback: click link

  16. Pingback: strona

  17. Pingback: link do strony

  18. Pingback: massage body body paris

  19. Pingback: click for more

  20. Pingback: take a look at the site here

  21. Pingback: click for more

  22. Pingback: sell your bitcoin

  23. Pingback: strona firmy

  24. Pingback: click for more details

  25. Pingback: pick up artist london

  26. Pingback: link

  27. Pingback: link

  28. Pingback: kamerzysta Lublin

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s