September 2008 Update on Global Temperature – RSS
Posted by The Diatribe Guy on September 7, 2008
The August month-end temperature anomaly for RSS, from this data source comes in at 0.146. This is at about the same level as the July anomaly (0.147), but a full 0.2 degrees Celsius below the August 2007 anomaly (0.367). This extends the run of consecutive year-over-year anomalies below the previous year to 12 consecutive months. The last time such a run occurred was the period ending February 2000. That period ended a run of 15 consecutive months of year-over-year decreases. Like that time period, the current anomalies coincide with La Nina on the heels of El Nino. So, while indicative of the cooling trends we’ve been seeing in the recent years, it is worth noting that there are natural cycles at play here, as well. The current La Nina has waned and we are now in a neutral ENSO state, so things will get a little more interesting. Will the quiet sun compensate for the waning La Nina?
Well, I have finally put RSS anomalies into my modeling spreadsheets. However, there is far less data to work with than the surface temperature data sets, so I have to focus on the changing trends of the shorter periods, since there isn’t yet a single 30-year term (though that will occur at the end of this year). But before I get to the charts, let’s note a few observations that can be made from the raw data.
First, the current 12-month average anomaly is now 0.086. (Reminder that this is stated in terms of degrees Celsius.) The last time the average reached this level was the period ending December 2000, when it was .077. The minimum 12-month average reached during that time frame was the period ended February 2000, at .042. Barring a remarkably cold September, that average will not be reached next month. The anomaly would need to reach -0.28. In fact, the projected anomaly for September is 0.201, which would be the highest anomaly since last October, yet would be below last September and continue the run of lower-year-over-year anomalies.
For those keeping score, here is the last 12-month series of anomalies (in 0.01 C):
2007 09: 26.40
2007 10: 23.00
2007 11: 12.10
2007 12: 8.50
2008 01: -7.00
2008 02: -0.20
2008 03: 7.90
2008 04: 8.00
2008 05: -8.20
2008 06: 3.50
2008 07: 14.70
2008 08: 14.60
There have been ten consecutive anomalies less than 15. Prior to that stretch, there had been one anomaly under 15 in the last 34 such readings. The last time there were 10 consecutive anomalies at a level less than 15 was the period ending October 1994 (The highest anomaly during that stretch was 11.1).
One thing I was interested in seeing was whether or not there is obvious seasonality in the average anomalies for given months. I ran a simple test of averages over the course of the data set by month, and found that the deviations from overall average range from a high of 0.0171 degrees Celsius in the month of February to a low of -.0213 in June. I had wondered whether or not distance from the Sun, or difference in axis orientation relative to the Sun might matter. The lower average anomaly in June suggests that there may be a slight effect, although the opposing effect seems to be shifted to January/February, rather than December. However, the deviations are within the statistical range of error, so I won’t be using these averages to adjust my analysis. It is an interesting element to consider, though.
With that, let’s check out the charts. The first chart shows the overall trend since inception of the satellite readings. The almost 30 years of anomalies, based on a linear trend, corresponds to 1.689 degrees per Century.
The current cooling period (as determined by linear trend) can extend back to March 1997. This chart specifically dates back to that date, simply to illustrate the furthest back one can go to demonstrate our current period of no warming. As one person pointed out to me, the 1998 El Nino was a historically high one, and certainly helps to drive a negative trend line. That is true and needs to be considered. On the other hand, the temperatures are the temperatures, and it can also be argued with ease that if natural cycles outside of Carbon Dioxide drive the respective spikes/troughs in temperature, then the 1998 El Nino represents a peak point in that activity and shouldn’t necessarily be discounted in terms of trending. There are arguments either way. I prefer to present the data as it is and let the reader decide.
The more recent trends are decidedly negative. One could argue this is due to the recent La Nina, and we will see whether or not that is true in the near future, as we are now ENSO neutral. If temperatures persist in a lower state, I suppose the blame will then shift to the quiet sun. That may well be valid, but it seems to point to a hypocrisy in the embracing of natural cycles when temperatures cool while discarding their impact when temperatures rise. Just my personal opinion being interjected there. Others likely disagree.
I present the 60-month trend of the raw data, but also present to you how the 60-month slope has changed over the last few years. Clearly, this has been driven lower for some time now:
I present the same charts for 10 years. Interestingly, even though the longest negative trend goes back more than 10 years, the current 120-month trend line is positive, since there were some intermittent low points that drive the beginning part of the trend line lower, thus creating a positive slope. You can see from the chart how this happens, and barring very low anomalies in the near future, the next few early points that drop off, being higher, will likely increase the 10-year slope for a time. Still, the general trend of the changing slopes is still negative in recent periods.
Without comment, I also offer the 180-month and 240-month charts:
My model projects a continual plunging of the 60-month slope value into the foreseeable future. I am not as confident in the robustness of this model, given the relative sparseness of data compared to the other models. That said, the model projects a lightly higher anomaly in September, and then if the model is to be believed, get out your sweaters. Within 6 months, the model predicts a significant plunge in anomalies. I, for one, am the first to hope the model’s predictive value is sub-par.