2011 Global Temperatures

[skierinvermont]

I guess my point is the real "significant" cooling would commence during a weaker solar MAX, not the minimum. We should expect very little in the way of solar cooling at this time.

http://www.scribd.co...bal-Temperature

We are probably experiencing the maximum INTRA-cycle cooling effect at this point which is typically estimated to be .1-.2C.

[tacoman25]

This is Jan 1998 to Jan 2010. As we know, these years ended up being close to each other, and both spiked in the first half of the year from their respective Ninos. Skiier's claim of a .08 to .09C/decade trend in UAH since 1998 is completely false.

[tacoman25]

Indeed. In fact, those are century long error estimates. As you know, error estimates for shorter time periods are usually much longer. For a single decade, IPCC estimates of temperature change would have much more quantifiable error. Which is exactly why even using the wildly inaccurate and likely cold biased UAH data the IPCC decadal trends are not invalidated.

(Although it is worth noting that the century long error in IPCC estimates comes from uncertainty in the actual climate sensitivity, while uncertainty in decade long trends comes from both short term variability and uncertainty about the actual climate sensitivity).

But they have quantifiable decadal ranges too, do they not? And wouldn't they have to be at least as large as UAH/RSS?

And your characterizations of UAH now as "wildly inaccurrate" only make you come across as desperate here. Especially when RSS has a cooler trend over the past decade.

[skierinvermont]

This is Jan 1998 to Jan 2010. As we know, these years ended up being close to each other, and both spiked in the first half of the year from their respective Ninos. Skiier's claim of a .08 to .09C/decade trend in UAH since 1998 is completely false.

Yes, that is probably why you were confused, the trend line zero or even negative a year or two ago. However, starting Jan 1998 and ending Jan 2010 is an unfair comparison, because you are including a full warm year right at the start (1998) but you are not including a full warm year at the end (2010). A pretty fair comparison in terms of ENSO is 1998-2010 (full year of 2010).

[skierinvermont]

But they have quantifiable decadal ranges too, do they not? And wouldn't they have to be at least as large as UAH/RSS?

And your characterizations of UAH now as "wildly inaccurrate" only make you come across as desperate here. Especially when RSS has a cooler trend over the past decade.

According to UAH and RSS themselves.. they are wildly inaccurate. (Unless you want to call an error range for a 30 year trend of -.02K/decade to +.21K/decade as something else).

The IPCC decadal error estimates dont have to be as large as UAH and RSS .. it may be that our ability to predict surface warming may actually exceed our ability to measure it accurately (at least by satellite). This would be similar to how there is a theoretical temperature for the surface of the earth (I think theoretically it is like 15C), but we can't actually measure this temperature very well... we are much better at measuring relative changes (anomalies).

I don't think that's the case though.. the 95% confidence interval for a single decade is pretty large for IPCC models.

Of course these models dont know what ENSO or solar will actually do so if you plugged those values in, the error ranges would shrink.

[tacoman25]

Yes, that is probably why you were confused, the trend line zero or even negative a year or two ago. However, starting Jan 1998 and ending Jan 2010 is an unfair comparison, because you are including a full warm year right at the start (1998) but you are not including a full warm year at the end (2010). A pretty fair comparison in terms of ENSO is 1998-2010 (full year of 2010).

If you start with Jan 1998, you are starting in the middle of an El Nino, and ending with Jan 2010 in the middle of El Nino. How is this not fair? The starting point is before the main El Nino warming kicked in, and the ending point is before the main El Nino warming kicked in. It's no different than if you did a 1999 to 2008 trend range.

Anyway, any fair comparison will yield a basically flat trend. No way you can claim a nearly .1C/decade warming trend since 1998 with UAH.

[tacoman25]

According to UAH and RSS themselves.. they are wildly inaccurate. (Unless you want to call an error range for a 30 year trend of -.02K/decade to +.21K/decade as something else).

The IPCC decadal error estimates dont have to be as large as UAH and RSS .. it may be that our ability to predict surface warming may actually exceed our ability to measure it accurately (at least by satellite). This would be similar to how there is a theoretical temperature for the surface of the earth (I think theoretically it is like 15C), but we can't actually measure this temperature very well... we are much better at measuring relative changes (anomalies).

I don't think that's the case though.. the 95% confidence interval for a single decade is pretty large for IPCC models.

Of course these models dont know what ENSO or solar will actually do so if you plugged those values in, the error ranges would shrink.

And that is exactly my point. If you are going to use the error range for UAH/RSS to call them wildly innaccurate, that same logic must then be applied to IPCC range estimates.

[BethesdaWX]

We are probably experiencing the maximum INTRA-cycle cooling effect at this point which is typically estimated to be .1-.2C.

Yes.

However, Intra-cycle cooling is minor and a smaller issue.....the concern is, as we head into the Solar Maximum, a weak Max is what gives the REAL cooling......Take away the energy of a Solar Max....ouch.

Yes, agree it started probably sometime mid-2010,. Issue is, we had the huge El Nino buck the trend. Aside from the Nino, the trend has been downward when removing the 2008 Nina dip and 2010 Nino spike.

[BethesdaWX]

If you start with Jan 1998, you are starting in the middle of an El Nino, and ending with Jan 2010 in the middle of El Nino. How is this not fair?

Anyway, any fair comparison will yield a basically flat trend. No way you can claim a nearly .1C/decade warming trend since 1998 with UAH.

Was about to say the same thing, there is no warming trend from 1998-2011. Its flat at best.

The 2010 El Nino is a blip in what will be a long term cooling trend, in my view.

[WeatherRusty]

Indeed. In fact, those are century long error estimates. As you know, error estimates for shorter time periods are usually much longer. For a single decade, IPCC estimates of temperature change would have much more quantifiable error. Which is exactly why even using the wildly inaccurate and likely cold biased UAH data the IPCC decadal trends are not invalidated.

(Although it is worth noting that the century long error in IPCC estimates comes from uncertainty in the actual climate sensitivity, while uncertainty in decade long trends comes from both short term variability and uncertainty about the actual climate sensitivity).

Given the current state of knowledge it is not possible to pinpoint a given global temperature to a specific time. The Charney sensitivity in a likely range of 2-4.5C is what we might expect following the system reaching equilibrium (the TOA imbalance would be neutralized) with a doubling of CO2. It may take decades to reach that equilibrium due to the thermal inertia of the oceans. So if a doubling from 280ppm to 560ppm occurs by 2060 it may be late century before 2-4.5C is realized.

[skierinvermont]

If you start with Jan 1998, you are starting in the middle of an El Nino, and ending with Jan 2010 in the middle of El Nino. How is this not fair? The starting point is before the main El Nino warming kicked in, and the ending point is before the main El Nino warming kicked in. It's no different than if you did a 1999 to 2008 trend range.

Anyway, any fair comparison will yield a basically flat trend. No way you can claim a nearly .1C/decade warming trend since 1998 with UAH.

There is a lag... so the first year of your record will be exceedingly warm due to the strong Nino warming in 1998. You are not including the last year of warmth.

You need to start with a warm YEAR and end with a warm YEAR.. not just the first month of warmth. You are starting with a full year of warmth, but ending with only a couple months of warmth. The fair comparison is Jan 1998 to Dec 2010, for which there is a trend of .1K/decade.

[skierinvermont]

Was about to say the same thing, there is no warming trend from 1998-2011. Its flat at best.

The 2010 El Nino is a blip in what will be a long term cooling trend, in my view.

This is just plain wrong. If you start with a warm ENSO year (1998) you need to end with a warm ENSO year (2010) to be fair. If you do this, the trend is +.09C/decade.

[BethesdaWX]

This is just plain wrong. If you start with a warm ENSO year (1998) you need to end with a warm ENSO year (2010) to be fair. If you do this, the trend is +.09C/decade.

NO...the problem is...when you start from the 1998 El Nino Spike, you get the cooling second half of the spike Into the Dip...using the 2010 El Nino spike as a fnish-point..you DON'T FINISH the Dip at the end....thus it is decieving to do that!

The 1998 El Nino spike was warmer than the 2010 El Nino spike...In between the two, there was no trend, and cooling from 2002-2008.

[tacoman25]

There is a lag... so the first year of your record will be exceedingly warm due to the strong Nino warming in 1998. You are not including the last year of warmth.

You need to start with a warm YEAR and end with a warm YEAR.. not just the first month of warmth. You are starting with a full year of warmth, but ending with only a couple months of warmth. The fair comparison is Jan 1998 to Dec 2010, for which there is a trend of .1K/decade.

Then how do you explain these?

Also, would this not be a fair comparison?

[skierinvermont]

NO...the problem is...when you start from the 1998 El Nino Spike, you get the cooling second half of the spike Into the Dip...using the 2010 El Nino spike as a fnish-point..you DON'T FINISH the Dip at the end....thus it is decieving to do that!

You still don't understand how a linear regression works... The dive down from the 1998 El Nino doesn't produce a cooling trend from 1998-2010. As long as the temperature is above the trend line it is "pulling" it upwards at that end.

Think of it as an upside down see-saw with a fulcrum in the middle... this is a very good way to approximate how a linear regression works. You have a big spike in 1998 which is very heavy and pulls the see saw up at that end. Unless you have a similar spike at the end, it is not a fair comparison.

[skierinvermont]

Then how do you explain these?

Also, would this not be a fair comparison?

Well first of all, all of those show some positive trend between .02-.06C/decade. That's less than Jan '98 -Jan '11 which shows .09C/decade, but more than Jan '98- Jan '10 which shows no trend.

The best way to tell if you are making a fair comparison with regard to ENSO is to look at the ENSO trend over the same period you are doing a temperature trend. If you are using satellite temps you should lag ENSO by 5 months, surface temps more like 2 months.

So let's look at the ENSO trends over the 5 periods in question (the three graphs above that you just posted, as well as Jan '98 - Jan '10, and Jan '98 - Jan '11). I have previously stated that Jan '98-Nov '10 is a fair comparison, while Jan '98 - Jan '10 is not. I don't know what any of the trends for these periods will be, I'm just going to generate the graphs and that should allow us to decide which period is the most "fair" with regards to ENSO. We want to find a period where ENSO (5 months lagged) shows no trend. I don't know what the results will be so here goes nothing. This is going to take me a while so I hope it is appreciated:

Your first graph appears to be Jan '97 to Jan '09 (IE August '06 to August '08). Here is the ENSO trend for the same period using a 5 month lag. As you can see, it shows a strong negative trend, indicating ENSO would have led to a cooling influence over the course of this period. Despite this fact, temperature looked to have warmed at a rate of +.02C/decade over the period.

Your second graph appears to be Jan '97 to Jan '10. Here is the ENSO trend for the same period using a 5 month lag. It still shows a negative trend although not as much as the first. This explains why the warming trend increased from +.02C/decade to +.05C/decade.

Your third graph appears to be Jan '00 to Jan '09. Here is the ENSO trend for the same period using a 5 month lag. The trend is positive indicating ENSO had a warming effect over this period, which is somewhat surprising to me given the temperature trend is only +.06C/decade. It is of note that this is the shortest of the 5 graphs examined and a slight adjustment to the lag effect could alter the results slightly.

Now for the original two periods we were discussing, Jan '98 - Jan '10 and Jan '98 - Nov '10

Here is Jan '98 - Jan '10 ENSO using 5 month lag. As you can see, the ENSO trend is negative which is why temperatures show no trend over the period.

Here is Jan '98 - Jan '11 ENSO using 5 month lag. This is the period I suggested one use since it starts with a warm year following a strong Nino, and ends in a strong year following a Nino. It also is a fairly long period (12 years) and I'm not a big fan of trends much less than a decade because the measurements are not that accurate or precise and there are too many factors at play. I wasn't quite correct, the period actually shows a slight increase in ENSO, although it is the closest to zero of any of the 5 graphs (look at equation for trend line in top right of each graph). The temperature trend over the same period is +.09C/decade.

So listed in order, the most ENSO neutral periods with their corresponding ENSO and temperature trends are:

1) Jan '98 - Jan '11

ENSO trend: +.0008C/yr

Temperature trend: +.09C/yr

2) Jan '98 - Jan '10

ENSO trend: -.0013C/yr

Temperature trend: -.02C/yr

3) Jan '97 - Jan '10

ENSO trend: - .0014C/yr

Temperature ternd: +.05C/yr

4) Jan '97 - Jan '09

ENSO trend: -.0018C/yr

Temperature trend: +.02C/yr

5) Jan '00 - Jan '09

ENSO trend: +.0029C/yr

Temperature trend: +.05C/yr

Based on all of the above I would estimate the ENSO neutral trend to be around +.07C/yr since 1998. I give heavy weight to the first one on the list since it is by far the most ENSO neutral, and it is the longest period on the list. Most of the examples you selected showed a declining trend in ENSO over the same period.

[nzucker]

Given all these graphs and ENSO trends, it sounds as if a fair trend for UAH is about .06C-.07C/decade, talking about recent years.

[tacoman25]

Well first of all, all of those show some positive trend between .02-.06C/decade. That's less than Jan '98 -Jan '11 which shows .09C/decade, but more than Jan '98- Jan '10 which shows no trend.

The best way to tell if you are making a fair comparison with regard to ENSO is to look at the ENSO trend over the same period you are doing a temperature trend. If you are using satellite temps you should lag ENSO by 5 months, surface temps more like 2 months.

So let's look at the ENSO trends over the 5 periods in question (the three graphs above that you just posted, as well as Jan '98 - Jan '10, and Jan '98 - Jan '11). I have previously stated that Jan '98-Nov '10 is a fair comparison, while Jan '98 - Jan '10 is not. I don't know what any of the trends for these periods will be, I'm just going to generate the graphs and that should allow us to decide which period is the most "fair" with regards to ENSO. We want to find a period where ENSO (5 months lagged) shows no trend. I don't know what the results will be so here goes nothing. This is going to take me a while so I hope it is appreciated:

Your first graph appears to be Jan '97 to Jan '09 (IE August '06 to August '08). Here is the ENSO trend for the same period using a 5 month lag. As you can see, it shows a strong negative trend, indicating ENSO would have led to a cooling influence over the course of this period. Despite this fact, temperature looked to have warmed at a rate of +.02C/decade over the period.

Your second graph appears to be Jan '97 to Jan '10. Here is the ENSO trend for the same period using a 5 month lag. It still shows a negative trend although not as much as the first. This explains why the warming trend increased from +.02C/decade to +.05C/decade.

Your third graph appears to be Jan '00 to Jan '09. Here is the ENSO trend for the same period using a 5 month lag. The trend is positive indicating ENSO had a warming effect over this period, which is somewhat surprising to me given the temperature trend is only +.06C/decade. It is of note that this is the shortest of the 5 graphs examined and a slight adjustment to the lag effect could alter the results slightly.

Now for the original two periods we were discussing, Jan '98 - Jan '10 and Jan '98 - Nov '10

Here is Jan '98 - Jan '10 ENSO using 5 month lag. As you can see, the ENSO trend is negative which is why temperatures show no trend over the period.

Here is Jan '98 - Jan '11 ENSO using 5 month lag. This is the period I suggested one use since it starts with a warm year following a strong Nino, and ends in a strong year following a Nino. It also is a fairly long period (12 years) and I'm not a big fan of trends much less than a decade because the measurements are not that accurate or precise and there are too many factors at play. I wasn't quite correct, the period actually shows a slight increase in ENSO, although it is the closest to zero of any of the 5 graphs (look at equation for trend line in top right of each graph). The temperature trend over the same period is +.09C/decade.

So listed in order, the most ENSO neutral periods with their corresponding ENSO and temperature trends are:

1) Jan '98 - Jan '11

ENSO trend: +.0008C/yr

Temperature trend: +.09C/yr

2) Jan '98 - Jan '10

ENSO trend: -.0013C/yr

Temperature trend: -.02C/yr

3) Jan '97 - Jan '10

ENSO trend: - .0014C/yr

Temperature ternd: +.05C/yr

4) Jan '97 - Jan '09

ENSO trend: -.0018C/yr

Temperature trend: +.02C/yr

5) Jan '00 - Jan '09

ENSO trend: +.0029C/yr

Temperature trend: +.05C/yr

Based on all of the above I would estimate the ENSO neutral trend to be around +.07C/yr since 1998. I give heavy weight to the first one on the list since it is by far the most ENSO neutral, and it is the longest period on the list. Most of the examples you selected showed a declining trend in ENSO over the same period.

And if we did this same analysis with RSS, the trend would be less positive...despite UAH's "cold bias" .

It will be ineteresting to see the Jan 1999 to Dec 2011 trend. Although, that still wouldn't really be "fair", since a two year Nina followed Jan 1999, and the strong 1998 Nino would not be included in the analysis, but the strong 2010 one would. That is actually why I think Jan 1998 to Jan 2010 is perfectly fair, because right after the 1998 Nino, a 3 year Nina event followed, as opposed to prior to the 2010 Nino, we only had a 1 year Nina followed by a borderline second year event (not officially Nina).

You do know that when Real Climate did an ENSO corrected analysis of temperature trends 1998-2008, they found an essentially flat trend (slightly positive with GISS, no positive with HadCRU)? Not sure if that has changed much over the past 2 years.....

[skierinvermont]

And if we did this same analysis with RSS, the trend would be less positive...despite UAH's "cold bias" .

It will be ineteresting to see the Jan 1999 to Dec 2011 trend. Although, that still wouldn't really be "fair", since a two year Nina followed Jan 1999, and the strong 1998 Nino would not be included in the analysis, but the strong 2010 one would. That is actually why I think Jan 1998 to Jan 2010 is perfectly fair, because right after the 1998 Nino, a 3 year Nina event followed, as opposed to prior to the 2010 Nino, we only had a 1 year Nina followed by a borderline second year event (not officially Nina).

You do know that when Real Climate did an ENSO corrected analysis of temperature trends 1998-2008, they found an essentially flat trend (slightly positive with GISS, no positive with HadCRU)? Not sure if that has changed much over the past 2 years.....

The Jan 1998 to Jan 2010 has a more negative ENSO trend than Jan 1998 to Jan 2011 has positive almost by a factor of two. A good compromise would be to include half of 2010.. and you would end up with about .07C/decade I think vs the .09C you get by ending in Jan 2011.

You must have remembered wrong.. it looks about .1-.15C/decade since 1998 to me from RC ENSO-corrected (a little more than that on GISS, a little less on HadCRUT). Definitely a bit of a slowdown but decisively positive.

[skierinvermont]

Given all these graphs and ENSO trends, it sounds as if a fair trend for UAH is about .06C-.07C/decade, talking about recent years.

Yep.. agree. I would argue it is a bit more than that 1997/1998-present (.07-.08C/yr), and a bit less than that 2000/2002-present (.05-.06C/yr).

[skierinvermont]

Given all these graphs and ENSO trends, it sounds as if a fair trend for UAH is about .06C-.07C/decade, talking about recent years.

Yep.. agree. I would argue it is a bit more than that 1997/1998-present (.07-.08C/yr), and a bit less than that 2000/2002-present (.05-.06C/yr). You can see the real flattening on the realclimate ENSO corrected graph above really doesn't begin until around 2002. I would argue that this flattening is due to the declining solar activity.

We can see from the 4 start periods in 97 or 98, the best conclusion is probably +.07C/yr.. for example Jan '97 -Jan '10 had a +.05C/yr temp trend even though the ENSO trend was quite negative.

Of course it is silly to analyze the data on this scale because even the 30 year trend on UAH is .08-.2C/decade (95% confidence interval, mean .14C/decade). RSS is .08-.25C/decade mean .16C/decade. The point being the error bars on even the full 30 year trend are quite large, and they would be even larger for a 10 or 12 year trend.

1) Jan '98 - Jan '11

ENSO trend: +.0008C/yr

Temperature trend: +.09C/yr

2) Jan '98 - Jan '10

ENSO trend: -.0013C/yr

Temperature trend: -.02C/yr

3) Jan '97 - Jan '10

ENSO trend: - .0014C/yr

Temperature ternd: +.05C/yr

4) Jan '97 - Jan '09

ENSO trend: -.0018C/yr

Temperature trend: +.02C/yr

[BethesdaWX]

You still don't understand how a linear regression works... The dive down from the 1998 El Nino doesn't produce a cooling trend from 1998-2010. As long as the temperature is above the trend line it is "pulling" it upwards at that end.

Think of it as an upside down see-saw with a fulcrum in the middle... this is a very good way to approximate how a linear regression works. You have a big spike in 1998 which is very heavy and pulls the see saw up at that end. Unless you have a similar spike at the end, it is not a fair comparison.

huh?

No I understand that. The problems I see are in what you are counting as a "the" trend... If we remove the large El Nino spike of 2010.....what would our trend be? Not saying that should be done...however, look at the graph that Tacoman posted, why you are complaining over starting in the warm 1998 year, if we are ending our trend in a "warm" year? We need to equally weight each dataset.

[tacoman25]

Well first of all, all of those show some positive trend between .02-.06C/decade. That's less than Jan '98 -Jan '11 which shows .09C/decade, but more than Jan '98- Jan '10 which shows no trend.

The best way to tell if you are making a fair comparison with regard to ENSO is to look at the ENSO trend over the same period you are doing a temperature trend. If you are using satellite temps you should lag ENSO by 5 months, surface temps more like 2 months.

So let's look at the ENSO trends over the 5 periods in question (the three graphs above that you just posted, as well as Jan '98 - Jan '10, and Jan '98 - Jan '11). I have previously stated that Jan '98-Nov '10 is a fair comparison, while Jan '98 - Jan '10 is not. I don't know what any of the trends for these periods will be, I'm just going to generate the graphs and that should allow us to decide which period is the most "fair" with regards to ENSO. We want to find a period where ENSO (5 months lagged) shows no trend. I don't know what the results will be so here goes nothing. This is going to take me a while so I hope it is appreciated:

Your first graph appears to be Jan '97 to Jan '09 (IE August '06 to August '08). Here is the ENSO trend for the same period using a 5 month lag. As you can see, it shows a strong negative trend, indicating ENSO would have led to a cooling influence over the course of this period. Despite this fact, temperature looked to have warmed at a rate of +.02C/decade over the period.

Your second graph appears to be Jan '97 to Jan '10. Here is the ENSO trend for the same period using a 5 month lag. It still shows a negative trend although not as much as the first. This explains why the warming trend increased from +.02C/decade to +.05C/decade.

Your third graph appears to be Jan '00 to Jan '09. Here is the ENSO trend for the same period using a 5 month lag. The trend is positive indicating ENSO had a warming effect over this period, which is somewhat surprising to me given the temperature trend is only +.06C/decade. It is of note that this is the shortest of the 5 graphs examined and a slight adjustment to the lag effect could alter the results slightly.

Now for the original two periods we were discussing, Jan '98 - Jan '10 and Jan '98 - Nov '10

Here is Jan '98 - Jan '10 ENSO using 5 month lag. As you can see, the ENSO trend is negative which is why temperatures show no trend over the period.

Here is Jan '98 - Jan '11 ENSO using 5 month lag. This is the period I suggested one use since it starts with a warm year following a strong Nino, and ends in a strong year following a Nino. It also is a fairly long period (12 years) and I'm not a big fan of trends much less than a decade because the measurements are not that accurate or precise and there are too many factors at play. I wasn't quite correct, the period actually shows a slight increase in ENSO, although it is the closest to zero of any of the 5 graphs (look at equation for trend line in top right of each graph). The temperature trend over the same period is +.09C/decade.

So listed in order, the most ENSO neutral periods with their corresponding ENSO and temperature trends are:

1) Jan '98 - Jan '11

ENSO trend: +.0008C/yr

Temperature trend: +.09C/yr

2) Jan '98 - Jan '10

ENSO trend: -.0013C/yr

Temperature trend: -.02C/yr

3) Jan '97 - Jan '10

ENSO trend: - .0014C/yr

Temperature ternd: +.05C/yr

4) Jan '97 - Jan '09

ENSO trend: -.0018C/yr

Temperature trend: +.02C/yr

5) Jan '00 - Jan '09

ENSO trend: +.0029C/yr

Temperature trend: +.05C/yr

Based on all of the above I would estimate the ENSO neutral trend to be around +.07C/yr since 1998. I give heavy weight to the first one on the list since it is by far the most ENSO neutral, and it is the longest period on the list. Most of the examples you selected showed a declining trend in ENSO over the same period.

I appreciate this work you did. However, if you add it all together for those periods, the average is an essentially neutral ENSO forcing. The average temperature increase is .038C....which if you apply to the average period of 11.8 years, is .045C/decade. Less than half of the .1C/decade increase you originally claimed over this period.

[skierinvermont]

I appreciate this work you did. However, if you add it all together for those periods, the average is an essentially neutral ENSO forcing. The average temperature increase is .038C....which if you apply to the average period of 11.8 years, is .045C/decade. Less than half of the .1C/decade increase you originally claimed over this period.

Only if you include the 5th one on the list which I can show you is a fluke if you like.. it is very sensitive to changing the start and end point by a few months because it is such a short time period. You need to heavily weight the first 4 on the list, especially the first one. Additionally, it has a later start point (2000) and I think the later you make your start point the lower the ENSO-corrected trend will be. As I said in my last post, I think the ENSO-corrected trend since 2002, is more like .05C/yr which is the same number you're getting. Looking only at the 4 that start around 1997/1998 and end 2010/2011, I find a result more like .07-.08C/yr.

The RC ENSO-corrected graph also supports more like .07C/.08C/decade (much more than that on GISS).

I agree, in retrospect, it's not quite .1C/decade given that the original time period I chose showed a slight positive ENSO trend.

[BethesdaWX]

Nit-picking trends again? JAN 2000..come on now. You also leave out the +PDO until 2007 (-PDO in La Nina 1999-2001, increasing positive trend through 2006 would lead that datsset to warm), +AMO increasing through the dataset.....

All ENSO spikes are going to behave differently, so using it as a base of conlsusion comes to a weak solution.

[tacoman25]

Only if you include the 5th one on the list which I can show you is a fluke if you like.. it is very sensitive to changing the start and end point by a few months because it is such a short time period. You need to heavily weight the first 4 on the list, especially the first one. Additionally, it has a later start point (2000) and I think the later you make your start point the lower the ENSO-corrected trend will be. As I said in my last post, I think the ENSO-corrected trend since 2002, is more like .05C/yr which is the same number you're getting. Looking only at the 4 that start around 1997/1998 and end 2010/2011, I find a result more like .07-.08C/yr.

The RC ENSO-corrected graph also supports more like .07C/.08C/decade (much more than that on GISS).

I agree, in retrospect, it's not quite .1C/decade given that the original time period I chose showed a slight positive ENSO trend.

I think it is actually essential to include the 5th one on the list, since it is the only period that starts with a full Nina year and ends with a full Nina year (same logic you use for the Jan 1998 to Jan 2011 period, Nino to Nino). And it is a 9 year period, which is not statistically significantly different than an 11 or 12 year period.

[skierinvermont]

I think it is actually essential to include the 5th one on the list, since it is the only period that starts with a full Nina year and ends with a full Nina year (same logic you use for the Jan 1998 to Jan 2011 period, Nino to Nino). And it is a 9 year period, which is not statistically significantly different than an 11 or 12 year period.

Well we are in close enough agreement I'm not going to argue this any further.. I can't definitively rule out the ENSO corrected trend is only .05C/yr. Like I said I think it is more like .07/.08C/yr since 1997/1998 which drops slightly the later you make your start data reaching around .05C/yr if you start in 2002-present. Since 1998 though I think it is more like .07/08C/yr because Jan 1998- present is +.09C/yr and is nearly ENSO neutral. Either way, a trend of zero can be definitively ruled out, even on UAH/RSS.

Every extra year of data helps to resolve the question of "how much if any warming since 1998" because each year is basically 10% more data to answer the question with and the ENSO fluctuations start to get drowned out the longer the period gets. So we can look back at this again at the end of this year. Of course the ENSO trends will be more negative which will have to be accounted for.

[skierinvermont]

I just realized that Jan97 - Dec10 is a good period to look at because it's the most ENSO neutral of any period we picked. And it's also the longest.

The ENSO trend (5mo lag) is just +.0005C/yr, which is the smallest of any of the 6 periods we have selected.

The temperature trend is +.14C/yr.

In retrospect, I probably should have used a 6 month lag on ENSO.

[tacoman25]

Well we are in close enough agreement I'm not going to argue this any further.. I can't definitively rule out the ENSO corrected trend is only .05C/yr. Like I said I think it is more like .07/.08C/yr since 1997/1998 which drops slightly the later you make your start data reaching around .05C/yr if you start in 2002-present. Since 1998 though I think it is more like .07/08C/yr because Jan 1998- present is +.09C/yr and is nearly ENSO neutral. Either way, a trend of zero can be definitively ruled out, even on UAH/RSS.

Every extra year of data helps to resolve the question of "how much if any warming since 1998" because each year is basically 10% more data to answer the question with and the ENSO fluctuations start to get drowned out the longer the period gets. So we can look back at this again at the end of this year. Of course the ENSO trends will be more negative which will have to be accounted for.

I think a fair compromise would be .05-.06/decade trend for UAH for the last 10-12 years. When you factor in RSS, though, the overall satellite trend over that period would be only very slightly positive.

It will be difficult to find a good period of comparison at the end of 2011. Jan 1998 to Dec 2011 certainly wouldn't be fair (Nino year to Nina year), Jan 1999 to Dec 2011 wouldn't be fair (eliminates the 1998 Nino, while including the big 2010 Nino, and two Nina years at the beginning ending with a big Nino and then one Nina year at the end). Jan 2000 to Dec 2011 will probably be a pretty fair comparison, but then we've shortened the period of record to just 11 years again. Jan 1997 to June 2011 might be the best combination of ENSO neutrality and longer period of record.

[skierinvermont]

I think a fair compromise would be .05-.06/decade trend for UAH for the last 10-12 years. When you factor in RSS, though, the overall satellite trend over that period would be only very slightly positive.

It will be difficult to find a good period of comparison at the end of 2011. Jan 1998 to Dec 2011 certainly wouldn't be fair (Nino year to Nina year), Jan 1999 to Dec 2011 wouldn't be fair (eliminates the 1998 Nino, while including the big 2010 Nino, and two Nina years at the beginning ending with a big Nino and then one Nina year at the end). Jan 2000 to Dec 2011 will probably be a pretty fair comparison, but then we've shortened the period of record to just 11 years again. Jan 1997 to June 2011 might be the best combination of ENSO neutrality and longer period of record.

Yes you can probably subtract off about .04C/yr from whatever we get for UAH to convert it to RSS for 1998-present trends.

Actually won't Jan 1998- Dec 2011 be pretty good? Jan 1998 Dec 2010 was slightly positive (+.0008C/yr) so a year of Nina might actually bring it down a bit to near zero maybe -.0002C/yr.