I was wondering if anyone could explain to me the metric "correlation coefficient" as it is used in determining when a funnel cloud becomes a tornado. What I have read so far about this metric confuses me in how the metric is used. My background is as a research psychologist and "correlation coefficient" is used very differently in the social sciences.
I am continuing my search to learn more and I wanted to bounce this around in this forum to see if I could gain some additional insight to how it is used and if anyone has worked with data related to determining the thresholds with correlation coefficients. I would be really interested in discussing this.
"Correlation Coefficient" Metric
-
- Expert
- Posts: 712
- Joined: Tue Jan 12, 2021 8:03 am
- Location: Italy
- Station model: WH2650
- Software: WH2650 (direct upload)
- Contact:
Re: "Correlation Coefficient" Metric
I think, here is explained in simple terms: https://www.ustornadoes.com/2017/03/28/ ... ed-editon/
The correlation is between the two measure returned by the dual-polarization radar:
- high value = symmetric objects ( droplets )
- low value = irregular objects ( debris )
P.S.
This is the first time I read about this subject.
Could you link your readings?
The correlation is between the two measure returned by the dual-polarization radar:
- high value = symmetric objects ( droplets )
- low value = irregular objects ( debris )
P.S.
This is the first time I read about this subject.
Could you link your readings?
-
- Newbie
- Posts: 3
- Joined: Thu Jul 28, 2022 4:17 pm
- Location: Elizabethtown, PA, USA
- Station model: Ambient Weather WS-2902C
- Software: Ambient Weather
- Contact:
Re: "Correlation Coefficient" Metric
Thank you so much for your reply, I will add it to my extensive reading list. And, of course, I will share with this Forum and you any insights I gain from my research. Thanks again.
All my best, Rick.
All my best, Rick.
-
- Newbie
- Posts: 3
- Joined: Thu Jul 28, 2022 4:17 pm
- Location: Elizabethtown, PA, USA
- Station model: Ambient Weather WS-2902C
- Software: Ambient Weather
- Contact:
Re: "Correlation Coefficient" Metric
I wanted to follow up with a response based upon my earlier questions regarding the metric "correlation coefficient" and just some impressions from a non-meteorologist but someone really interested in data in whatever form it takes. The first impression has to do with the term correlation coefficient which for the non-meteorologist is slightly confusing, especially if you are coming from the social sciences. When I look at the intent and purpose of the correlation coefficient, a term like “debris coefficient” seems more appropriate. It doesn’t change the interpretation, I think it just better describes its intent. I found it especially confusing when I was reading about the correlation coefficient being correlated with some other metric or variable.
The second impression was in viewing radar scopes and their interpretation. I found that very fascinating and I think the raw data behind the radar scope displays would be even more interesting to analyze. It would be interesting to see if there are any thresholds or tipping points right before the TDS - Tornadic Debris Signature begins to establish itself. I have found in many assorted data distributions that there are trigger points, thresholds, or tipping points which have a profound impact on subsequent data. For example, are there statistical predictors in earlier data that predict the formation of the tornado. If it has not been done already, this could be an interesting analytical framework.
And lastly, an impression I had is the wealth of resources available, especially the number of online courses made available through NOAA/NWS. I would highly recommend to anyone interested in meteorology to consult the following site (https://www.meted.ucar.edu/index.php), you will not be disappointed in seeing what is at your fingertips.
The second impression was in viewing radar scopes and their interpretation. I found that very fascinating and I think the raw data behind the radar scope displays would be even more interesting to analyze. It would be interesting to see if there are any thresholds or tipping points right before the TDS - Tornadic Debris Signature begins to establish itself. I have found in many assorted data distributions that there are trigger points, thresholds, or tipping points which have a profound impact on subsequent data. For example, are there statistical predictors in earlier data that predict the formation of the tornado. If it has not been done already, this could be an interesting analytical framework.
And lastly, an impression I had is the wealth of resources available, especially the number of online courses made available through NOAA/NWS. I would highly recommend to anyone interested in meteorology to consult the following site (https://www.meted.ucar.edu/index.php), you will not be disappointed in seeing what is at your fingertips.