I’ve moved to WordPress. This post can now be found at El Ninos Create Step Changes in TLT of the Northern Hemisphere Mid Latitudes
###############The mid latitudes are segmented by longitudes into continental and oceanic subsets in this post, making it easier to illustrate the short-term impacts of volcanic eruptions and the long-term step changes caused by the two significant El Nino events since 1979 that were not impacted by volcanic eruptions.
In agreement with the title of the first part of this series on Lower Troposphere Temperature (TLT), ENSO and Volcanic Aerosols Explain Most of the Arctic Volatility, that post shows that natural events, El Nino events and volcanic eruptions, explain most of the Arctic TLT variability, with the rest a function of high latitude SST. This post examines the mid latitudes of the Northern Hemisphere.
It complements the following two posts that showed how, since 1976, El Nino events (that were not impacted by volcanic eruptions) caused step changes in global SST anomalies.
Can El Nino Events Explain All of the Global Warming Since 1976? – Part 1
Can El Nino Events Explain All of the Global Warming Since 1976? – Part 2
But because satellite data for TLT has only been available since late 1978, only two significant El Nino events caused step changes since then. These were the 1986/87/88 and 1997/98 El Nino events. The steps are very obvious.
A PRELIMINARY LOOK AT THE DATA
Many presentations of global temperature anomalies (or some subset thereof) include linear trend lines to show the current rate of rise. By doing so, they imply that the temperature anomalies will continue at that pace into the future. Figure 1 is an illustration of TLT for the mid latitudes (30N-60N) of the Northern Hemisphere. The trend is approximately 0.27 deg C/decade, inferring a rise of 2.7 deg C over the next century. But does it really? The assumption is that the driver or drivers of the rise will remain constant.
http://i41.tinypic.com/i4e134.jpgFigure 1
I’ve divided the mid latitudes of the Northern Hemisphere into subsets as shown in Figure 2 to reflect the TLT over the oceans and continents. The KNMI Climate Explorer webpage (http://climexp.knmi.nl/selectfield_obs.cgi?someone@somewhere) allows time-series data defined by user-selected coordinates to be viewed and downloaded. I used it as the source of the AHU MSU TLT data in the following.
http://i42.tinypic.com/14uhu8h.jpgFigure 2
And as you can see, the trends for the subsets are not that different.
Northeast Pacific
http://i40.tinypic.com/199yrq.jpgFigure 3
************
Atlantic
http://i40.tinypic.com/2jbk09v.jpgFigure 4
************
Northwest Pacific
http://i43.tinypic.com/b5heyw.jpgFigure 5
************
North America
http://i41.tinypic.com/ayra6x.jpgFigure 6
************
Old World West
http://i39.tinypic.com/a581oh.jpgFigure 7
************
Old World East
http://i41.tinypic.com/qz5ap3.jpgFigure 8
************
PRELIMINARY NOTES
Note 1: Many of the statistical analyses performed on global temperatures anomalies and their subsets extract the short-term impacts of volcanic eruptions and El Nino events. Some address the minor variations caused by changes in Total Solar Irradiance (TSI). By examining the data and accounting for time lags, these analyses subtract the effects of those natural climate forcings based on coefficients found in scientific studies. During the periods (months or years) of the El Nino, La Nina or volcano, the known effects are subtracted from or added to the temperature anomaly data. This works fine for volcanic eruptions and TSI. But El Nino events (that are not impacted by volcanic eruptions) take warm subsurface water and through a detailed natural process bring it to the surface, where it increases the SST for a significant part of the global ocean over multiyear periods. These long-term changes in temperature appear as upward steps in global temperature or subset. The steps are very obvious in some datasets, less so in others, as you will see in the following.
Note 2: Climate models, the few that attempt to model El Nino events, do not simulate the historical amplitude, frequency, or processes well. But as you’ll see, without this ability, their use as a tool to predict or project future climate is severely limited.
Note 3: The Lower Troposphere Temperature (TLT) anomaly data identified as AHU MSU was accessed using the coordinate-based system available through the KNMI website, as noted above. They cover the period of January 1979 to November 2008. The SST data in Figures 9, 12, and 15 is OI.v2 SST data available through the NOAA NOMADS system. Monthly OI.v2 SST data is available from November 1981 to present. The NINO3.4 SST anomaly data used in the other comparative graphs is ERSST.v2 data, which is available through the NOMADS system on a monthly basis from January 1854 to present. The Northern Hemisphere Sato Index data is available from GISS. All data in the graphs (with the exception of Figure 1) have been smoothed with a 12-month running-average filter. The data is raw in Figure 1.
NOTE 4: NINO3.4 SST anomaly and Sato Index (mean optical thickness of stratospheric volcanic aerosols) data are provided to aid in the illustration of the timing of the ENSO and volcanic eruptions only. They have not been scaled for any other purpose. The NINO3.4 SST anomalies are scaled by factor of 0.24. The Sato Index data has been inverted using a factor of -3.
NOTE 5: The comparative graphs of TLT and SST for the ocean subsets are provided only as a reference, to show that the TLT and SST anomaly curves do resemble one another and that the trends and the timing of their perturbations are similar.
MID LATITUDE NORTHEAST PACIFIC TLT
Figure 9 is a comparative graph of mid latitude Northeast Pacific TLT and SST anomalies. They correlate reasonably well. The impacts of the two major volcanic eruptions (El Chichon in 1982 and Mount Pinatubo in 1991) are visible, as are the aftereffects of the 1986/87/88 and 1997/98 El Nino events.
http://i41.tinypic.com/14buptf.jpgFigure 9
These impacts on mid latitude Northeast Pacific TLT are more easily seen when Northeast Pacific TLT anomalies are compared to NINO3.4 anomaly and Sato Index data. Refer to Figure 10. Note how well the timing of the volcanic eruptions and the dips in Northeast Pacific TLT anomalies agree. Note also how the volcanic eruptions have no long-term impact. That is, Northeast Pacific TLT anomalies return near to the values they were at before the volcanic eruption. This effect is repeated in all of the comparative graphs that follow as well.
http://i41.tinypic.com/2cohg74.jpgFigure 10
The responses to the El Nino events are also easily seen in Figure 10. From 1979 to 1988, the Northeast Pacific TLT anomalies are relatively flat, excluding the dip and rebound associated with the El Chichon eruption. Then, lagging the 1986/87/88 El Nino by approximately 18 months, Northeast Pacific TLT rises from ~-0.3 deg C to ~+0.5 deg C. But Northeast Pacific TLT does not react to the subsequent La Nina; the Northeast Pacific TLT takes almost 9 years to decrease to its minimum levels. NINE YEARS. After a delayed response to the 1997/98 El Nino (another but smaller step change), Northeast Pacific TLT anomalies remain relatively flat. The exception is the rise and fall from 2003 to 2006, which appears to be anomalous, but if the scaling factor of the NINO3.4 SST anomaly data is considered (0.24) the reaction is not that unusual.
************
Update: While examining the TLT anomalies of the Southeast Pacific (mid and low latitudes) for upcoming posts, something was illustrated clearer. The low latitudes (North and South) of the eastern Pacific react in synch with the sign of the ENSO event immediately, which is fitting since the NINO areas make up a portion of those datasets. There may be minor lags, but in general when NINO3.4 SST anomalies rise during an El Nino, so do the TLT anomalies for the low latitudes of the eastern Pacific, North and South. During La Ninas, TLT anomalies in low latitudes of the eastern Pacific decrease. Nothing surprising there.
BUT, mid latitude TLT anomalies, North and South, oppose the changes in the low latitudes during an ENSO event. Note the dip in the TLT anomalies of the mid latitude North Pacific, Figure 10, during the 1997/98 El Nino. That dip does not appear at the same time in the SST data, Figure 9. And, referring again to Figure 10, the significant rise in the mid latitude North Pacific TLT anomalies after the 1986/87/88 El Nino appears actually as an opposing response to the 1988/89 La Nina. There was also an apparent shift due to the accumulation of warm water in the northwest Pacific at the same time, adding to that rise. I’ll discuss this effect in more detail in an upcoming post.
************
For those having trouble visualizing the step change with 9-year decline and the flatness of the first and last periods, I’ve tried to highlight them in Figure 11, with some degree of success.
http://i40.tinypic.com/dlkrnm.jpgFigure 11
MID LATITUDE NORTH ATLANTIC TLT
Mid latitude North Atlantic SST and TLT anomalies are shown in Figure 12. The two datasets agree reasonably well. TLT shows a greater variability but the trends and major perturbations agree. From what we know of the North Atlantic, part of the upward trend is a result of the Atlantic Multidecadal Oscillation (AMO), a function of Thermohaline Circulation in the North Atlantic.
http://i41.tinypic.com/zl5g93.jpgFigure 12
In Figure 13, the NINO3.4 SST anomaly and Sato Index data are compared to mid latitude North Atlantic TLT anomalies. The North Atlantic TLT anomalies react to the 1991 volcanic eruption with an approximate 6-month lag. It’s more difficult to determine what transpired in 1982. The North Atlantic TLT anomalies were declining before the eruption, but then seem to be driven further down by the volcanic aerosols. They rebound quickly, but then cycle back down steeply in what seems to be a secondary reaction. By the middle of 2005, the North Atlantic TLT anomalies have bottomed out and are rebounding again.
http://i40.tinypic.com/14l2g74.jpgFigure 13
In some respects it appears that North Atlantic TLT anomalies then increase gradually with oscillations through their peak in 1998. In other respects, though, it appears there may be a step change. During the period of 1986 to 1998, the Mid Latitude North Atlantic TLT anomalies first lag the 1986/87/88 El Nino by 6 months, peaking twice in the same way as the NINO3.4 SST anomalies. NINO3.4 SST anomalies drop sharply, but the North Atlantic TLT anomalies don’t. They decline at a much slower rate (similar to the rate after the 1997/98 El Nino) before rising again in reaction to the next increase in NINO3.4 SST anomaly. Have the North Atlantic TLT anomalies risen in one or more steps over the period? If the average TLT anomalies from January 1979 through December 1985, and from January 1988 through December 1996, and from January 2000 to November 2008 (excluding the periods of the two significant El Ninos) are added to the graph, the steps appear. Refer to Figure 14.
http://i39.tinypic.com/2iu4x1j.jpgFigure 14
MID LATITUDE NORTHWEST PACIFIC TLT
Figure 15 is a comparative graph of mid latitude Northwest Pacific TLT and SST anomalies. Again, the two data sets correspond well, including the large anomalous rise and fall from 1988 to late 1991.
http://i42.tinypic.com/2058prq.jpgFigure 15
Comparing the mid latitude Northwest Pacific TLT anomalies to NINO3.4 SST anomalies and the Sato Index data, Figure 16, helps show the cause of the significant rise and fall in the Northwest Pacific TLT anomalies between 1988 and late 1991. The Northwest Pacific TLT anomalies appear to lag the 1986/87/88 NINO3.4 SST anomalies by two years.
http://i43.tinypic.com/73pllg.jpgFigure 16
In Figure 17, the average Northwest Pacific TLT anomalies for the periods before (January 1979 to December 1987) and after (January 1989 to November 2008) the lagged rise in TLT have been added to the graph. They illustrate the magnitude of the change in Northwest Pacific TLT anomalies that resulted from the 1986/87/88 El Nino--approximately 0.65 deg C. That increase in TLT anomaly lagged the NINO3.4 SST anomaly by 2 years and remained near its peak until the 1991 volcanic eruption dropped it. Keep in mind that the NINO3.4 data in Figure 17 has been scaled, that it is less than a quarter of the true value; the lagged increases in mid latitude Northwest Pacific TLT anomalies from 1988 to 1991 are less than the actual NINO3.4 SST anomalies. Note also that it takes 2 years for the mid latitude Northwest Pacific TLT anomalies to return to “normal” after the 1997/98 El Nino, while the lag on the increase is only 6 months.
http://i44.tinypic.com/k00aah.jpgFigure 17
MID LATITUDE NORTH AMERICAN TLT
Figure 18 compares mid latitude North American TLT to scaled NINO3.4 and scaled (inverted) Sato Index data. North American TLT is quite volatile. It is, therefore, very difficult to determine what caused the wide swings before 1986, especially with the volcanic eruptions and El Nino occurring near the same time in 1982. From 1986 to 1997, mid latitude North American TLT mimics NINO3.4 SST anomalies, with the exception of the few years after the 1991 volcanic eruption. Then the 1997/98 El Nino causes a significant shift in North American TLT.
http://i42.tinypic.com/2e0mq76.jpgFigure 18
Using the averages of the mid latitude North American TLT anomalies for the periods of January 1979 to December 1996 and January 1998 to November 2008 as reference, mid latitude North American TLT anomalies shifted more than 0.55 deg C as a result of the 1997/98 El Nino. Refer to Figure 19. Note that there may be a smaller intermediate step after the 1986/87/88 El Nino, but it is not highlighted in the illustration.
http://i43.tinypic.com/315odmo.jpgFigure 19
MID LATITUDE WESTERN OLD WORLD TLT
A comparative graph of mid latitude Western Old World TLT anomalies, scaled NINO3.4 SST anomalies, and scaled (inverted) Sato Index data is shown in Figure 20. Large perturbations reflect lagged responses to ENSO events and volcanic eruptions.
http://i40.tinypic.com/29y4hly.jpgFigure 20
I had originally planned to go to great lengths to describe the year-to-year variations as I have in the earlier datasets, but this post is too long as it is, so I’ve marked up the comparative graph to highlight the changes in mid latitude Western Old World TLT anomalies and their causes in Figure 21. Note that, before 1991, there’s a 2-year lag between El Nino events (primarily the 1986/87/88 El Nino) and the response from Western Old World TLT anomalies. There’s also a 2-year lag between the 1982 volcanic eruption and the Western Old World TLT anomalies. Then from 1993 to the 1997/98 El Nino there’s little time lag. But the only thing that happened between those two periods, between 1991 and 1993, was the Mount Pinatubo eruption. What could that volcanic eruption have done to change the response times of TLT? And that’s not a rhetorical question. What happened?
http://i39.tinypic.com/29w752f.jpgFigure 21
But what’s very apparent, if mid latitude Western Old World TLT anomaly averages (before the 1986/87/88 El Nino, between the 1986/87/88 and 1997/98 El Ninos, and after the 1997/98 El Nino) are added to the comparative graph, there were two significant step changes of approximately 0.3 degrees C over the term of the data. Refer to Figure 22.
http://i43.tinypic.com/6784sp.jpgFigure 22
MID LATITUDE EASTERN OLD WORLD TLT
In Figure 23, mid latitude Eastern Old World TLT anomalies, scaled NINO3.4 SST anomalies, and scaled (inverted) Sato Index data are illustrated. The perturbations in the Eastern Old World TLT anomalies are very similar in timing to the Western Old World data, but most of them are exaggerated in the Eastern Old World data.
http://i44.tinypic.com/29ynlo6.jpgFigure 23
In Figure 24, the averages of the Eastern Old World TLT anomalies for the same periods as Figure 21 (before the 1986/87/88 El Nino, between the 1986/87/88 and 1997/98 El Ninos, and after the 1997/98 El Nino) were added to the comparative graph. The overall rise in the Eastern dataset, approximately 0.65 deg C, is greater than the Western.
http://i41.tinypic.com/s3252u.jpgFigure 24
CLOSING
The segments of the mid latitude TLT data showed different:
-Response Times to El Nino Events,
-Response Times to Volcanic Eruptions,
-Rates of Response to El Nino Events,
-Rates of Response to La Nina Events, and
-Rates of Response to Volcanic Eruptions.
And there appears to be little consistency in the responses with a given dataset. For example, in some data sets, the earlier (pre-1991) El Nino response is greater than the later (post-1993) response, and in others, the reverse is true. In some data sets, before 1991, the time lag between a natural forcing (El Nino event or a volcanic eruption) and the mid latitude data response is about 2 years, but then, after 1993, the response time drops to a few months.
To help illustrate this, Figure 25 is a gif animation of Mid Latitude TLT and each of the subsets discussed in this post.
As noted earlier, climate models, the few that attempt to model El Nino events, do not simulate their historical amplitude, frequency, or processes well. Without this ability, their use as a tool to predict or project future climate on a global or regional level is severely limited.
http://i40.tinypic.com/s0vfqe.jpgFigure 25
SOURCES
The AHU MSU Lower Troposphere Temperature data is available through the KNMI Climate Explorer website:
http://climexp.knmi.nl/selectfield_obs.cgi?someone@somewhere
The Optimally Interpolated Sea Surface Temperature Data (OISST) are available through the NOAA National Operational Model Archive & Distribution System (NOMADS).
http://nomads.ncdc.noaa.gov/#climatencdc
The Sato Index Data is available from GISS at:
http://data.giss.nasa.gov/modelforce/strataer/
Specifically:
http://data.giss.nasa.gov/modelforce/strataer/tau_line.txt
No comments:
Post a Comment