Since the PDO is a long term index, and the phase it's in can be accurately predicted (76% of the time) by looking the OLR patterns of the world (see my OLR thread), I did some research to find out the following....
Is the PNA effected by the Earth's long term OLR phase, or by the state of the PDO? How can we use the PDO/OLR to predict the PNA better?
To find out the answer to these questions I had to do some research to find the similarities between the PDO and PNA both in general, and how well they related in different OLR periods. Since it would be very time consuming to do all 12 months of the year, I focused more on the 4 months where the PNA has arguably the biggest impact on the US. These are the winter months of Dec, Jan, Feb, and March.
Since I only have PNA data dating back to 1950, my research consists only of PNA/PDO data from 1950-2002. Another thing to note is that in this research, I used the old formula PNA (more accurate one) that we have been using for the past 55 years (before they changed it this May), and is currently shown on the CDC site.
First, lets break down the PNA/PDO relationship month by month...
DECEMBER
Number of times the PNA and PDO have had an identical December value (greater then 0.00 being positive, less then 0.00 being negative)
35/53 (66%)
When the PDO is positive (greater then 0.00)
16/22 (73%) of the time the PNA is positive (above 0.00)
6/22 (27%) of the time the PNA is negative (below 0.00)
When the PDO is negative (less then 0.00)
19/31 (61%) of the time the PNA is negative
12/31 (39%) of the time the PNA is positive
Since a PDO/PNA with a value of 0.00 to + 0.50 doesnt do much, we will call that neutral. Anything above +0.50 is positive. Anything below -0.50 is negative. This will give us a more accurate idea on the correlation between the two. Since some of the values were so close, and the purpose of this is to show correlation, I also used a .05 difference round system (for example... if the PDO is 0.53, and the PNA is 0.49 I will call that PDO and PNA positive)
When the PDO is positive...
7/12 (58%) of the time the PNA is positive
3/12 (25%) of the time the PNA is neutral
2/12 (17%) of the time the PNA is negative
When the PDO is negative...
13/21 (62%) of the time the PNA is negative
3/21 (14%) of the time the PNA is neutral
5/21 (24%) of the time the PNA is positive
When the PDO is neutral...
7/21 (33%) of the time the PNA is neutral
9/21 (43%) of the time the PNA is positive
5/21 (24%) of the time the PNA is negative
When the PDO is strongly positive... (above 2.00)
1/1 (100%) of the time the PNA is positive
When the PDO is strongly negative... (below 2.00)
3/3 (100%) of the time the PNA is negative
JANUARY
Number of times the PNA and PDO have had an identical January value (greater then 0.00 being positive, less then 0.00 being negative)
40/53 (75%)
When the PDO is positive (greater then 0.00)
17/25 (68%) of the time the PNA is positive (above 0.00)
8/25 (32%) of the time the PNA is negative (below 0.00)
When the PDO is negative (less then 0.00)
23/28 (82%) of the time the PNA is negative
5/28 (18%) of the time the PNA is positive
Since a PDO/PNA with a value of 0.00 to + 0.50 doesnt do much, we will call that neutral. Anything above +0.50 is positive. Anything below -0.50 is negative. This will give us a more accurate idea on the correlation between the two. Since some of the values were so close, and the purpose of this is to show correlation, I also used a .05 difference round system (for example... if the PDO is 0.53, and the PNA is 0.49 I will call that PDO and PNA positive)
When the PDO is positive...
8/14 (57%) of the time the PNA is positive
4/14 (29%) of the time the PNA is neutral
2/14 (14%) of the time the PNA is negative
When the PDO is negative...
18/22 (82%) of the time the PNA is negative
3/22 (14%) of the time the PNA is neutral
1/22 (4%) of the time the PNA is positive
When the PDO is neutral...
5/17 (12%) of the time the PNA is neutral
2/17 (29%) of the time the PNA is positive
10/17 (59%) of the time the PNA is negative
The has never been strongly positive (above 2.00) in the month of Jan
When the PDO is strongly negative... (below 2.00)
5/5 (100%) of the time the PNA is negative
FEBRUARY
Number of times the PNA and PDO have had an identical January value (greater then 0.00 being positive, less then 0.00 being negative)
38/53 (72%)
When the PDO is positive (greater then 0.00)
17/23 (74%) of the time the PNA is positive (above 0.00)
6/23 (26%) of the time the PNA is negative (below 0.00)
When the PDO is negative (less then 0.00)
21/30 (70%) of the time the PNA is negative
9/30 (30%) of the time the PNA is positive
Since a PDO/PNA with a value of 0.00 to + 0.50 doesnt do much, we will call that neutral. Anything above +0.50 is positive. Anything below -0.50 is negative. This will give us a more accurate idea on the correlation between the two.
When the PDO is positive...
11/16 (69%) of the time the PNA is positive
2/16 (12%) of the time the PNA is neutral
3/16 (19%) of the time the PNA is negative
When the PDO is negative...
14/21 (67%) of the time the PNA is negative
5/21 (24%) of the time the PNA is neutral
2/21 (9%) of the time the PNA is positive
When the PDO is neutral...
8/15 (53%) of the time the PNA is neutral
4/15 (27%) of the time the PNA is positive
3/15 (20%) of the time the PNA is negative
The has never been strongly positive (above 2.00) in the month of Feb
When the PDO is strongly negative... (below 2.00)
2/2 (100%) of the time the PNA is negative
MARCH
Number of times the PNA and PDO have had an identical January value (greater then 0.00 being positive, less then 0.00 being negative)
38/53 (72%)
When the PDO is positive (greater then 0.00)
15/25 (60%) of the time the PNA is positive (above 0.00)
10/25 (30%) of the time the PNA is negative (below 0.00)
When the PDO is negative (less then 0.00)
23/27 (85%) of the time the PNA is negative
4/27 (15%) of the time the PNA is positive
Since a PDO/PNA with a value of 0.00 to + 0.50 doesnt do much, we will call that neutral. Anything above +0.50 is positive. Anything below -0.50 is negative. This will give us a more accurate idea on the correlation between the two. Since some of the values were so close, and the purpose of this is to show correlation, I also used a .05 difference round system (for example... if the PDO is 0.53, and the PNA is 0.49 I will call that PDO and PNA positive)
When the PDO is positive...
7/17 (41%) of the time the PNA is positive
6/17 (35%) of the time the PNA is neutral
4/17 (24%) of the time the PNA is negative
When the PDO is negative...
16/21 (76%) of the time the PNA is negative
5/21 (24%) of the time the PNA is neutral
0/21 (00%) of the time the PNA is positive
* Note that the PNA has never been positive when the PDO was negative in March
When the PDO is neutral...
8/14 (57%) of the time the PNA is neutral
1/14 (07%) of the time the PNA is positive
5/14 (36%) of the time the PNA is negative
When the PDO is strongly positive... (above 2.00)
2/4 (50%) of the time the PNA is positive
2/4 (50%) of the time the PNA is neutral
0/4 (00%) of the time the PNA is negative
When the PDO is strongly negative... (below 2.00)
1/2 (50%) of the time the PNA is negative
1/2 (50%) of the time the PNA is neutral
0/2 (00%) of the time the PNA positive
DEC, JAN, FEB, MAR COMBINATION
Number of times the PNA and PDO have had an identical January value (greater then 0.00 being positive, less then 0.00 being negative)
151/212 (71%)
When the PDO is positive (greater then 0.00)
65/95 (68%) of the time the PNA is positive (above 0.00)
35/95 (32%) of the time the PNA is negative (below 0.00)
When the PDO is negative (less then 0.00)
86/116 (74%) of the time the PNA is negative
30/116 (26%) of the time the PNA is positive
Since a PDO/PNA with a value of 0.00 to + 0.50 doesnt do much, we will call that neutral. Anything above +0.50 is positive. Anything below -0.50 is negative. This will give us a more accurate idea on the correlation between the two. Since some of the values were so close, and the purpose of this is to show correlation, I also used a .05 difference round system (for example... if the PDO is 0.53, and the PNA is 0.49 I will call that PDO and PNA positive)
When the PDO is positive...
33/59 (56%) of the time the PNA is positive
15/59 (25%) of the time the PNA is neutral
11/59 (19%) of the time the PNA is negative
When the PDO is negative...
61/85 (72%) of the time the PNA is negative
16/85 (19%) of the time the PNA is neutral
8/85 (09%) of the time the PNA is positive
When the PDO is neutral...
23/67 (34%) of the time the PNA is neutral
19/67 (28%) of the time the PNA is positive
25/67 (37%) of the time the PNA is negative
When the PDO is strongly positive... (above 2.00)
3/5 (60%) of the time the PNA is positive
2/5 (40%) of the time the PNA is neutral
0/5 (00%) of the time the PNA is negative
When the PDO is strongly negative... (below 2.00)
11/12 (92%) of the time the PNA is negative
1/12 (08%) of the time the PNA is neutral
0/12 (00%) of the time the PNA positive
WINTERS AS A WHOLE (THE DJFM AVERAGE)
Number of times the PNA and PDO have had an identical January value (greater then 0.00 being positive, less then 0.00 being negative)
40/53 (75%)
When the PDO is positive (greater then 0.00)
14/20 (70%) of the time the PNA is positive (above 0.00)
6/20 (30%) of the time the PNA is negative (below 0.00)
When the PDO is negative (less then 0.00)
26/33 (79%) of the time the PNA is negative
7/33 (21%) of the time the PNA is positive
Since a PDO/PNA with a value of 0.00 to + 0.50 doesnt do much, we will call that neutral. Anything above +0.50 is positive. Anything below -0.50 is negative. This will give us a more accurate idea on the correlation between the two.
When the PDO is positive...
10/14 (71%) of the time the PNA is positive
3/14 (21%) of the time the PNA is neutral
1/14 (07%) of the time the PNA is negative
* Note that the PNA has only once ever been negative when the PDO was positive for the winter
When the PDO is negative...
15/21 (71%) of the time the PNA is negative
6/21 (29%) of the time the PNA is neutral
0/21 (00%) of the time the PNA is positive
* Note that the PNA has never been positive when the PDO was negative for the winter
When the PDO is neutral...
8/18 (44%) of the time the PNA is neutral
2/18 (11%) of the time the PNA is positive
8/18 (44%) of the time the PNA is negative
When the PDO is strongly positive... (above 1.00)
5/7 (71%) of the time the PNA is positive
2/7 (29%) of the time the PNA is neutral
0/7 (00%) of the time the PNA is negative
* 100% of the time the PNA averaged out to be above 0.00 although some years were neutral
When the PDO is strongly negative... (below 1.00)
10/13 (77%) of the time the PNA is negative
3/13 (23%) of the time the PNA is neutral
0/13 (00%) of the time the PNA positive
* 100% of the time the PNA averaged out to be below 0.00 although some years were neutral
Key points:
- The PNA is only opposite of the PDO 13% of the time (using neutral)
- The PNA is has an equal value to the PDO 65% of the time (using neutral)
- When the PDO is strong (+2.00), the PNA has an equal value 82% of the time (using neutral)
- When the PDO is strong (+2.00), the PNA has never had an opposite value (using neutral)
- 71% of the time, the PNA has an equal value to the PDO in the winter as a whole (using neutral)
- Only once has the PNA ever been opposite of the PDO for the winter as a whole (using neutral)
- When the PDO averages +1.00 for the winter, the PNA has NEVER been opposite (using 0.00)
What does all this prove? It proves that there is a very strong relationship between the PDO and the PNA, and that they correlate very strongly to each other. Now comes the question... Is the PNA a decadel osciallation that runs on OLR phases like the NAO/PDO/ENSO does, or is the PNA a biproduct of the PDO?
As stated in an earlier thread, we have recorded 2 OLR phases since 1948. The negative phase occurred from 1948-1978. The positive phase occurred from 1980-present. To find out if the PNA runs on the OLR phases, I am going to take months in which the PDO ran opposite of what it normally would for its phase, and see what the PNA did in that timeframe.
After doing some research on this, I got very surprising results especially in the negative OLR phase...
In the negative OLR era (48-78) which favors a NEGATIVE PDO, I took a composite of ALL months in which the PDO averaged out to be positive. Here is what I found...
43/46 (93%) of those months featured a +PNA (above 0.00)
3/46 (07%) of those months featured a -PNA (below 0.00)
* Discounting the winter months (DJFM), 100% of the time the PDO was negative (30/30) in this period, as was the PNA.
In the positive OLR era (80-present) which favors a POSITIVE PDO, I took a composite of ALL months in which the PDO averaged out to be negative. Here is what I found...
33/51 (65%) of those months featured a -PNA (below 0.00)
18/51 (35%) of those months featured a +PNA (above 0.00)
All togeather in opposite phases....
76/97 (78%) of the months were a like.
21/97 (22%) of the months were different.
This means that the PNA is not a decadal oscillation, and is most likely effected and greatly influenced by the state of the PDO. In other words, now that we know how to accurately predict the PDO, the PNA should also be able to be forecast with accuracy in the long term as the correlation between the two is amazing.
Im sorry I couldnt include any pictures in this thread
