4
$\begingroup$

How can I split all the climate modes of variability that you can find here, into positive, negative and neutral phases?

I know that depending on the climate mode there exist different methods of achieving this.

Do you know any general literature that can help me?

Thanks

$\endgroup$
  • $\begingroup$ I don't have any references to give, but wanted to leave a comment. Almost all the common modes of variability have been defined and refined many times over. Some have been optimized for predictability, some are defined for easy model calculation, and many have purely heuristic roots. I suggest starting by looking at the PDFs of all the indices. How many are ~Gaussian? Unimodal, bimodal, trimodal? Do all of them even have a continuous range? Which ones have overlapping inputs, like ones based on SSTs in overlapping domains? $\endgroup$ – Jareth Holt Aug 7 '18 at 18:02
  • $\begingroup$ Thanks. I plotted the PDFs and almost all of them are normally distributed. $\endgroup$ – aaaaa Aug 8 '18 at 6:37
1
$\begingroup$

I think thats a lot of information on the link you pasted. and the indexes shown there are mostly in time (time-series). Please note two important steps. FIRST: the index you want to use is already produced in scientific way and published so how is the index produce please read the relevant paper for that which can help. Apart from that these indexes are mostly (Not all of them) are usually developed to capture/shows standardized/normalized/Anomalous behaviors of the system. So these indexes are representation of the systems with mean of zero and standard deviation of usually one in case of standardized index, and so does varies for each one of them. The anomaly is different story and i won't go there.

SECOND: Now the deviation of the index from mean (which is zero) shows the deviation from mean conditions (that means from neutral conditions). So positive and negative will show the modes of system based on intended interpretation. like for strength positive deviation will be termed as strong and vice versa.

Finally you can select the positive and negative years as modes of climate with positive and negative deviation from mean. Further to dig deep you can look for mechanism and principals that led to these deviations.

If you want to use your data and look for modes of climate, then i will suggest PCA, SVD and EOF analysis which may not be applicable in this case. I hope that will help

$\endgroup$
  • $\begingroup$ Thanks. Indeed..I think (but not 100% sure) that for normally distributed climate modes common practice guides to assign positive (negative) phases to anything greater (less) than 1std of the mean. But for ENSO this is not the case! As El Niño/La Nina events are detected in a different way. $\endgroup$ – aaaaa Aug 3 '18 at 16:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.