Did The 97/98 El Nino Cause A Step Change In Global Temperature and At Least One Other Index?

And the second question: Does each El Nino cause a step change in global temperature? If they’re not counteracted by a La Nina that has an equal impact on global temperature, the answer should be yes.

INITIAL NOTE

This is a very simple analysis, using extremely simple techniques. It is intended is to illustrate that the effects of the 97/98 El Nino on global and Arctic temperatures need to be evaluated by others with better tools.

INTRODUCTION USING RSS MSU DATA

Several times per month, the major climate change blogs post global temperature anomaly updates. The most common graph shows global temperature anomalies since December 1978 or January 1979, the start of satellite measurement. An example using RSS MSU data is shown in Figure 1.

http://i37.tinypic.com/24z98p1.jpg
Figure 1

Depending on the AGW leaning of the site, or the analysis at hand, the illustration may include a linear trend line (Figure 2) to point out the severity, or lack thereof, of the warming, which depends on the start date.
http://i33.tinypic.com/wir8s6.jpg
Figure 2

But each time I see one of the short-term global temperature anomaly graphs, I see the 97/98 El Nino and two trends. (I’ve got a good imagination.) Those trends include the periods before and after the “El Nino of the Century”, with a step change between the two periods. To illustrate this in the following, I segmented the data. The end of the first period was the month with the lowest temperature in 1997, which is April 1997 for the RSS MSU data shown. Refer to Figure 3. EXCEL added the linear trend.
http://i36.tinypic.com/2l8z34k.jpg
Figure 3

The chosen month for the beginning of the later period was the one that corresponded to the first global temperature minimum following the 97/98 El Nino, June 1999. Refer to Figure 4. Again, I used EXCEL for the trend.
http://i35.tinypic.com/169geuq.jpg
Figure 4

Since I am not well versed in the intricacies of EXCEL*, I had to employ Windows PAINT to add the linear trend lines from Figures 3 and 4 to the complete record, as shown in Figure 5. Using the end of the first trend line and the beginning of the second as references, I connected the two and noted the difference in temperature. Based on the preceding, there was a 0.12 deg C step change in global temperature as a result of the 97/98 El Nino.

*Note: Refer to the update at the end of this post for a further discussion of the problem I had with trend lines when I prepared Figure 5 and similar graphs. http://i35.tinypic.com/110drw6.jpg
Figure 5

RSS MSU also divides their data by latitude. Figure 6 illustrates Arctic (60 to 82.5N) temperature anomalies from Jan 1979 to July 2008. Using the same months to divide the data set, Figure 7, the 97/98 El Nino caused more than a 0.37 deg C step change in Arctic temperature. It also caused a steep increase in the slope of the trend line for the latter period.
http://i34.tinypic.com/o9j0v5.jpg
Figure 6

PREFACE TO THE REMAINDER

I followed the same process used for the RSS Global Temperature Anomaly in the following examinations. To minimize verbiage, I’ve simply identified the data set, the months used to isolate the 97/98 El Nino, and the step change in the linear trend lines from the end of the pre-97/98 El Nino period to the beginning of the post-97/98 El Nino period.

GISS GLOBAL TEMPERATURE ANOMALY

Data Set: GISS Global Temperature Anomaly
Pre-97/98 El Nino Period End Month: July 1997
Post -97/98 El Nino Period Start Month: May 1999
Approximate Step Change in Global Temperature From 97/98 El Nino: 0.08 Deg C
http://i35.tinypic.com/2ajs8ib.jpg
Figure 7

NCDC GLOBAL TEMPERATURE ANOMALY

Data Set: NCDC Global Temperature Anomaly
Pre-97/98 El Nino Period End Month: May 1997
Post -97/98 El Nino Period Start Month: Nov 1998
Approximate Step Change in Global Temperature From 97/98 El Nino: 0.1 Deg C
http://i36.tinypic.com/2ijhvdw.jpg
Figure 8

UAH MSU GLOBAL TEMPERATURE ANOMALY

Data Set: UAH MSU Global Temperature Anomaly
Pre-97/98 El Nino Period End Month: April 1997
Post -97/98 El Nino Period Start Month: June 1999
Approximate Step Change in Global Temperature From 97/98 El Nino: 0.14 Deg C
http://i36.tinypic.com/14scl08.jpg
Figure 9

UAH MSU ARCTIC TEMPERATURE ANOMALY

Data Set: UAH MSU Arctic Temperature Anomaly
Pre-97/98 El Nino Period End Month: April 1997
Post -97/98 El Nino Period Start Month: June 1999
Approximate Step Change in Arctic Temperature From 97/98 El Nino: 0.52 Deg C
http://i38.tinypic.com/2cpx63c.jpg
Figure 10

HADCRUT GLOBAL TEMPERATURE ANOMALY

Data Set: HADCRUT Global Temperature Anomaly
Pre-97/98 El Nino Period End Month: April 1997
Post -97/98 El Nino Period Start Month: Mar 1999
Approximate Step Change in Arctic Temperature From 97/98 El Nino: 0.16 Deg C
http://i36.tinypic.com/1emam9.jpg
Figure 11

THE MEDITERREAN SEA - A Remarkable Step Change

Figure 12 illustrates SST anomaly for the Mediterranean Sea from January 1854 to May 2008. It’s ERSST.v2 data through NOMADS. The coordinates used to download the data were 30 to 45N and 5W to 35E. In this evaluation, I simply divided the data into two sets separated at 1998. Based on the trend lines, Mediterranean Sea SSTs rose more than 0.42 deg C as a result of the 97/98 El Nino.
http://i33.tinypic.com/6rpjzm.jpg
Figure 12

Update November 18, 2009: I wrote above, “Since I am not well versed in the intricacies of EXCEL, I had to employ Windows PAINT to add the linear trend lines from Figures 3 and 4 to the complete record, as shown in Figure 5…” In that paragraph, I revealed a difficulty with EXCEL I had in the preparation of this post. It had nothing to do with the content of the post. It pertained to the presentation of trend lines by EXCEL when two datasets with separate time periods are plotted on the same graph. When EXCEL adds the linear trends to two short-term datasets, it presents the linear trends over the combined terms of both datasets, from the start of the first dataset to the end of the latter dataset, as shown in the following graph. In other words, the trend lines overlap. I didn’t want the overlapping trends.


http://i48.tinypic.com/i6cv1y.png

The question, how would I snip the trend lines to only the shorter-term periods so that they don’t overlap, similar to what is shown in Figure 5?

H/T to blogger sod.