129 Scientists sign Letter to UN Security General Ban Ki-Moon

[skierinvermont]

What is wrong with what they state? That there hasn't been any statistically significant warming since 1997? That is not incorrect.

I find, using HadCRUT that there has been warming over the last 16 years of .06C/decade (starting in 1997, starting in 1996 yields much more warming of course). .06C/decade is statistically significant.

Also, "The hypothesis that our emissions of CO₂ have caused, or will cause, dangerous warming is not supported by the evidence."

Complete nonsense that ignores the peer-reviewed literature. Denier propaganda.

[The_Global_Warmer]

2012 is going to finish like 5th to 8th warmest on record on GISS and UAH globally, yeah, cooling since 1997, yet we are no where near as cool as pre-1997, yeah wrap your head around that.

SLR at modern records, showing no sign of slowing.

OHC at record highs

Land ice at record lows

Ice albedo at record lows in modern times

Global and arctic ice near or at record lows

Not gonna be seeing this on denier blogs much.

[Snow_Miser]

neither of those links are peer-reviewed literature. they are responses to an article in Eos; basically online letters to the editor.

and IIRC the original Eos article you are contesting was indeed peer-reviewed.

why is his continual nonsense allowed, unchecked? it took 1 minute to debunk his nonsense.

You didn't debunk anything the papers brought up.

The actual doran and Zimmerman study wasn't peer reviewed either. Interesting that you choose to not harp on the doran and Zimmerman study because of that.

[Snow_Miser]

I find, using HadCRUT that there has been warming over the last 16 years of .06C/decade (starting in 1997, starting in 1996 yields much more warming of course). .06C/decade is statistically significant.

Also, "The hypothesis that our emissions of CO₂ have caused, or will cause, dangerous warming is not supported by the evidence."

Complete nonsense that ignores the peer-reviewed literature. Denier propaganda.

According to the met office, the trend in temperatures per decade is 0.03 degrees c since 1997, which is not statistically significant.

http://metofficenews.wordpress.com/2012/10/14/met-office-in-the-media-14-october-2012/

[Snow_Miser]

The mere 6 months required to see a positive trend on OHC was for if we had perfect OHC data. In fact, it might even be zero, although a strong El Nino or volcano might temporarily make it negative for 6 months maybe a year. I think the OHC data sampling is good enough and error bars small enough that the longest we could go without seeing a rise is 4 or 5 years. A flatline could arise from a combination of sample error, and short term variability particularly a decline in TSI. That was just a guess though, maybe the sample error is large enough that it takes 6 or 7 years to get a positive trend 95% of the time.

Also, I agree the warming trend has recently been on the low side of modeled, but I do not believe it has fallen outside the confidence interval of a +.15C/decade trend. It may have fallen outside the confidence interval of a .20C/decade trend. Given the record low solar minimum for modern times, and the increased frequency of El Ninos, it is not surprising that we are on the low side of the confidence interval.

The fact that we have multiple heat content datasets that disagree with each other on ocean heat content makes the heat content trends more uncertain. While one heat content trend may have small error bars, the overall heat content trends remain uncertain because of this disagreement.

It's one of the reasons why every ocean heat content dataset should be taken with a grain of salt.

The state of the climate report link above states that 15 years with a trend are needed to create a discrepancy between the forecasted and observed trend. The trend is now 0.03 degrees c per decade since 1997 on hadcrut4. That is more than 15 years, close to 16 years, actually.

We can say that there is a discrepancy between the forecasted and observed trends in temperatures since 1997, based off of the criteria needed to claim a discrepancy in the NOAA link above.

[skierinvermont]

According to the met office, the trend in temperatures per decade is 0.03 degrees c since 1997, which is not statistically significant.

http://metofficenews...4-october-2012/

That is mid 97 to mid 12 which is 15 years not 16. I did jan 97 to present, which is nearly 16 years. A full 16 years starting in Nov 96 would yield an even higher trend.

[skierinvermont]

The fact that we have multiple heat content datasets that disagree with each other on ocean heat content makes the heat content trends more uncertain. While one heat content trend may have small error bars, the overall heat content trends remain uncertain because of this disagreement.

It's one of the reasons why every ocean heat content dataset should be taken with a grain of salt.

The state of the climate report link above states that 15 years with a trend are needed to create a discrepancy between the forecasted and observed trend. The trend is now 0.03 degrees c per decade since 1997 on hadcrut4. That is more than 15 years, close to 16 years, actually.

We can say that there is a discrepancy between the forecasted and observed trends in temperatures since 1997, based off of the criteria needed to claim a discrepancy in the NOAA link above.

The Met office link says that flat periods can last up to 15 years in the simulations. There is not a flat period, there is a slight rise at 15 years and a moderate statistically significant rise for 16 years.

The discrepancies in OHC from ARGO arise between 0-700m and 0-2000m or because you are using old datasets that did not properly account for some measurement errors. Recent data sets for 0-700m or 0-2000m agree well.

Inferred estimates from sea level rise and mass balance tend to have error bars and do not always agree well with the ARGO estimates. This is why scientists sometimes have a difficult time "closing the sea level budget."

[skierinvermont]

If ENSO were one of the main reasons for why the temperatures have flatlined, and the sun had little influence on this recent flatlining of temperatures over the last 10 years, we should be able to see that with Ocean Heat Content. Unfortunately, that's not the case. The 0-700 meter OHC dataset has slowed down dramatically. With the 0-2000 meter dataset only having seen good quality data over only the last 9 years, it is unclear with regard to whether the heat gain there has slowed there as well.

It is not unclear what 0-2000m OHC was in the past. Very good estimates of historical OHC can be inferred from sea level rise. As well as from climate models. It is likely that 0-2000m OHC has slowed somewhat, but is still rising rapidly.

A recent dataset on OHC called the UKMO EN3 Ocean Heat Content dataset shows that OHC has slowed even moreso than the NODC OHC dataset.

The current monthly anomaly for Sea Surface Salinity and Temperature anomalies for the upper portion of the ocean are shown above for October 2012.

The dataset shows a slight decline in OHC in the 0-700 meter range over the last 8-9 years or so.

Of course, neither dataset comes even close to the Heat Content trend projected by Hansen et al. 2005's models.

There is also literally no trend in the 0-2000 meter OHC dataset as well.

The first thing I notice going to the UKMO EN3 page is that they specifically say XBT corrections have not been applied. They say they are working on it. XBT corrections are required to correct a large known cold bias. Shame on you. This is what happens when you go LOOKING for data that fits your obvious bias and then use it without any regard for its accuracy.

The 0-2000 meter dataset is of course less reliable than the 0-700 meter dataset, and has seen quality data for less than a decade, so take this chart for what it's worth.

It's interesting that 17 years is required for the criteria for the temperature records to establish a new trend, but only 9 years is sufficient enough for the Ocean Heat Content records for establishing a new trend. Interesting double standard.

As explained above, and without any objection from you, this is not a double standard. Short term variability effects surface temperature trends more. If we had perfect OHC data we would likely witness a linear rise every single year. This is not true for surface temperature which can be highly perturbed by ENSO, convection, etc.

Your previous claim was that thermal expansion was likely responsible for the Sea Level Rise that occured over the last decade. However, as Cazenave et al. showed, since Heat Content did not increase significantly during this period, the Sea Level Rise during this period was caused not through thermal expansion but by ice sheets melting, which fits well, considering that the Arctic has continued to warm over the las decade, even though the Global Mean Temperature has either flatlined or declined slightly.

Cazenave did not find no warming. They find an inferred trend of .25mm/yr. This is slightly less than measured by ARGO, but still corroborates a significant rise. Given the estimates of ice mass balance have recently been revised downwards slightly, that would yield an increased estimate of inferred OHC rise.

There is not an immediate response in Ocean Heat Content to a change in a forcing. There is a lag because of the thermal inertia of the oceans.

Mutliple studies have found this lag between solar activity and climate change.

There is no physical justification for saying that there will not be a lag between a forcing and OHC.

First, as I have already said, there IS a lag to surface temperature response because of the heat built up in the oceans that must dissipate before surface temperature responds.

Second, the studies you post are likely garbage and not actually peer-reviewed, like every other study I have read that you have posted. For example that garbage paper that claims to find an 11-yr periodicity in sea level rise. As I've said before, all you have to do is look at a chart of SLR to see there is no such 11-yr periodicity. He cooked the data.

Third, you have still failed to propose one plausible mechanism by which a change in solar forcing does not begin to effect OHC until 6+ years after the forcing begins. It really is analogous to the pot on stove analogy (which you have apparently abandoned because you don't like it anymore). When you turn the burner off, the pot of water begins to cool immediately. When a strong negative forcing is applied, the energy balance of the earth goes negative, and this should be IMMEDIATELY apparent in OHC which makes up ~90% of the earth's energy balance at any given time.

By DEFINITION a forcing alters the earth's energy balance, and the oceans represent ~90% of the earth's energy balance at any given time. By DEFINITION, there can be no lag between a forcing and the response of the earth's energy balance represented by OHC.

This is really quite basic, and if you thought about it for more than 5 seconds, instead of ignoring it because you don't like the answer, you would realize this.

[TerryM]

neither of those links are peer-reviewed literature. they are responses to an article in Eos; basically online letters to the editor.

and IIRC the original Eos article you are contesting was indeed peer-reviewed.

why is his continual nonsense allowed, unchecked? it took 1 minute to debunk his nonsense.

Trix

I sometimes think that their only goal is to generate the maximum # of hits. Allowing this stuff drives serious people away but ensures that some of us stick around to refute the outrageous claims. I've pretty well given up any hope that a moderator will step in and bring a halt to such bogus threads.

It would be nice to find a good, well moderated climate forum. Neven's site has become very well respected regarding Arctic sea ice so I know the concept is viable. If the aim is simply to generate traffic I think we're in for more of the same.

Terry

[Snow_Miser]

That is mid 97 to mid 12 which is 15 years not 16. I did jan 97 to present, which is nearly 16 years. A full 16 years starting in Nov 96 would yield an even higher trend.

Yes, since 1996 was of course a La Nina.

But the trend analysis from the Met Office does confirm no statistically signficiant warming since 1997 when looking at the sks trend calculator.

[Snow_Miser]

The Met office link says that flat periods can last up to 15 years in the simulations. There is not a flat period, there is a slight rise at 15 years and a moderate statistically significant rise for 16 years.

The discrepancies in OHC from ARGO arise between 0-700m and 0-2000m or because you are using old datasets that did not properly account for some measurement errors. Recent data sets for 0-700m or 0-2000m agree well.

Inferred estimates from sea level rise and mass balance tend to have error bars and do not always agree well with the ARGO estimates. This is why scientists sometimes have a difficult time "closing the sea level budget."

Yes, and we are approaching the threshold for where a discrepency can finally be evident between the modeled and observed temperature increases since 1997. One more year, and there will be a discrepency evident with the rate at which temperatures have been changing compared to modelled temperature changes over the last 15-16 years.

I'm not sure what you mean by your second point. The UKMO EN3 dataset that I presented that shows OHC declining in the upper 700 meters of the oceans is not an old dataset. It was just updated in October of this year.

[Snow_Miser]

It is not unclear what 0-2000m OHC was in the past. Very good estimates of historical OHC can be inferred from sea level rise. As well as from climate models. It is likely that 0-2000m OHC has slowed somewhat, but is still rising rapidly.

We have absolutely zero data before 2003 for the data for the deep ocean in the 1000-2000 m region, so we really can't say either way if it has slowed or not. Sea Level Rise can be influenced by many factors, so while it is a good proxy for heat content changes, it is not an absolutely certain measurement of historical oceanic heat content.

The first thing I notice going to the UKMO EN3 page is that they specifically say XBT corrections have not been applied. They say they are working on it. XBT corrections are required to correct a large known cold bias. Shame on you. This is what happens when you go LOOKING for data that fits your obvious bias and then use it without any regard for its accuracy.

They also write this on the same page:

Use the links below to download data with some of the recently proposed XBT bias corrections applied. The first two also include bias corrections for mechanical bathythermograph (MBT) data.

Firstly, how could you have missed that if the first thing you noticed were the passages about the XBT bias corrections?

Secondly, I am not favoring that dataset moreso than the NODC OHC dataset. I was showing it to show that all Heat Content datasets should be taken with a grain of salt, and uncertainties with these datasets need to be stressed when they are being used.

As explained above, and without any objection from you, this is not a double standard. Short term variability effects surface temperature trends more. If we had perfect OHC data we would likely witness a linear rise every single year. This is not true for surface temperature which can be highly perturbed by ENSO, convection, etc.

Perhaps I was mistaken about you having a double standard. Apologies.

However, to claim that OHC changes are not influenced by a multitude of factors would be as silly as to ignore either the human or the natural factors that impact the climate. There is a very complicated process that involves vertical mixing that transports heat to different portions of the ocean, and there are also oceanic currents that transport heat content of the oceans as well. The heat content of the upper ocean is also influenced by changes in solar activity which was seen in a paper by White et al. 1997.

There are many things that influence OHC besides Greenhouse Gases.

Cazenave did not find no warming. They find an inferred trend of .25mm/yr. This is slightly less than measured by ARGO, but still corroborates a significant rise. Given the estimates of ice mass balance have recently been revised downwards slightly, that would yield an increased estimate of inferred OHC rise.

I don't think you understand why I cited the Cazenave study.

You made a claim that thermal expansion was contributing significantly to sea level rise, and I refuted that assumption with data that showed otherwise. The 0.25 mm/year that they extracted is SO small that it would equate to around 2.5 cm in the next 100 years. This small of a residual trend can be very easily explained by other factors that occur naturally like groundwater extractions.

Second, the studies you post are likely garbage and not actually peer-reviewed, like every other study I have read that you have posted. For example that garbage paper that claims to find an 11-yr periodicity in sea level rise. As I've said before, all you have to do is look at a chart of SLR to see there is no such 11-yr periodicity. He cooked the data.

It's not a garbage paper. None of the papers I posted are garbage papers. Unfortunately, it doesn't even look like you bothered to open the links to the papers I provided. Had you done so, you would have seen that both papers I posted were peer reviewed. Instead, you resort to ridiculous conspiracy theories about Shaviv manipulating the data. He didn't. That paper was published in a peer reviewed journal. Had he manipulated the data, the reviewers would called him out on this manipulation. Since he didn't manipulate any data, the paper was published in the Journal of Geophysical Research, a very reputable scientific journal.

Sorry, but I don't buy that.

Third, you have still failed to propose one plausible mechanism by which a change in solar forcing does not begin to effect OHC until 6+ years after the forcing begins. It really is analogous to the pot on stove analogy (which you have apparently abandoned because you don't like it anymore). When you turn the burner off, the pot of water begins to cool immediately. When a strong negative forcing is applied, the energy balance of the earth goes negative, and this should be IMMEDIATELY apparent in OHC which makes up ~90% of the earth's energy balance at any given time

I didn't say that the OHC changes are not immediately impacted by the change in the forcing. I said that there isn't an immediate cooling response to a decline in a forcing. Look at the graph I posted. While the numbers are arbitrary, it represents how the climate system, particularly the oceans, respond to a change in the forcing. They respond slowly, and do not equilibriate very quickly to a change in a forcing at all. This is the lag, and why so many researchers have discovered this lag between solar activity changes and temperature changes.

[skierinvermont]

Yes, since 1996 was of course a La Nina.

But the trend analysis from the Met Office does confirm no statistically signficiant warming since 1997 when looking at the sks trend calculator.

The 15 year trend is still positive.. it doesn't matter if it is statistically significantly positive or not. The standard is that there is 95% confidence that any given 15-yr period will have a positive trend. If you want to talk about statistically significant trend, it probably takes 17-19 years to be 95% confident of a statistically significant positive trend.

As I said before, the last 15 years are clearly on the low side of the confidence interval, however that is pretty well explained by the record solar minimum and increased La Nina frequency. The underlying trend is likely still close to .15C/decade. When the effect of ENSO and TSI are removed according to long-term statistical relationships, the underlying trend is still .15C/decade. I have demonstrated this many times on this forum.

[skierinvermont]

The UKMO EN3 OHC graphs you posted were created by Bob Tisdale, the noted WUWT AGW-denier contributor, and we have no way of knowing whether he used the corrected or uncorrected version. Since all the data is coming from ARGO and UKMO doesn't have its own independent source of data, we can be virtually certain that this graph is of the uncorrected version.

You need to be more careful about what source you use. You seem all too willing to blindly accept data that denies AGW.

[Vergent]

What is wrong with what they state? That there hasn't been any statistically significant warming since 1997? That is not incorrect.

0.? per decade is such a creeping rate that the statistics will be questionable. There will always be questions unanswered. So, in 15 yrs u choose this cherry pick to hang your hat on? Show me where the ice is not melting. Show me where the ocean is cooling. This is 98% of the issue. Show me where.............

edit

directed to "Snow, not "skier"

[The_Global_Warmer]

I think between now and 2020 we are going to see the yearly record broken 2 times at least.

With more abrupt warming like we saw in the 80s and 90s.

Obviously the big solar drop, even if it continues at the weaker TSI output hasn't caused the energy imbalance to reverse, slow it has. But the surface temp's have likely been effected quite a bit by the -PDO and ENSO with the solar to some extent. But ice loss keeps accelerating, land ice, Spring Snow Cover as well.

OHC is slower but still at record highs. The radiative effectiveness of Co2/Methane and all GHG's combined will grow quite fast as we go forward.

Co2 in 10 years will be around 430PPM with Methane likely much higher. It will win out and warming will be abrupt.

But we don't need to go 10 years forward to test anything.

The fact that the Sun's output has tanked and the energy imbalance keeps growing is pretty stout.

GISS, NCDC, and UAH have all had 2nd or 3rd warmest month's this year.

With GISS and NCDC doing it recently without the record SST's to drive it.

favorable conditions for cooler Earth have been on-going mostly for the better part of the last decade but we haven't cooled at all.

[Snow_Miser]

The 15 year trend is still positive.. it doesn't matter if it is statistically significantly positive or not. The standard is that there is 95% confidence that any given 15-yr period will have a positive trend. If you want to talk about statistically significant trend, it probably takes 17-19 years to be 95% confident of a statistically significant positive trend.

As I said before, the last 15 years are clearly on the low side of the confidence interval, however that is pretty well explained by the record solar minimum and increased La Nina frequency. The underlying trend is likely still close to .15C/decade. When the effect of ENSO and TSI are removed according to long-term statistical relationships, the underlying trend is still .15C/decade. I have demonstrated this many times on this forum.

The 15 year trend is positive on HadCtuT4, but it is substantially less than that of the modeled projections over this timeframe. We can claim that there is a discrepency because of the criteria that was given from NOAA when it can be claimed if a discrepency exists or not.

In the next 10 years, we will probably be able to see how strong of a response there is over the next 10 or so years to changes in solar activity and oceanic oscillations.

[Snow_Miser]

The UKMO EN3 OHC graphs you posted were created by Bob Tisdale, the noted WUWT AGW-denier contributor, and we have no way of knowing whether he used the corrected or uncorrected version. Since all the data is coming from ARGO and UKMO doesn't have its own independent source of data, we can be virtually certain that this graph is of the uncorrected version.

You need to be more careful about what source you use. You seem all too willing to blindly accept data that denies AGW.

There's an internal contradication with the bolded.

Once again, I have not "accepted" the UKMO EN3 dataset as fact. You on the other hand seem to be strugging to accept it as even a possibility of what might have happened with the ocean temperature and heat content over the last 10 years.

Why would Bob Tisdale choose the verison that has more errors in it? That doesn't make any sense. Bob Tisdale contrusts his charts from the data provided from the leading scientific organizations.

[TerryM]

Tisdale draws chart instead of using the originals for a good reason.

If you use the originals, it invalidates his point.

Terry

[Snow_Miser]

More nonsense from Tamino who tries and misrepresent what had actually been predicted by climate models.

http://bobtisdale.wo...de-the-incline/

[skierinvermont]

There's an internal contradication with the bolded.

Once again, I have not "accepted" the UKMO EN3 dataset as fact. You on the other hand seem to be strugging to accept it as even a possibility of what might have happened with the ocean temperature and heat content over the last 10 years.

Why would Bob Tisdale choose the verison that has more errors in it? That doesn't make any sense. Bob Tisdale contrusts his charts from the data provided from the leading scientific organizations.

Tisdale got the data from KNMI climate explorer and has no way of knowing which version of UKMO EN3 they used. The data was only on KNMI for a couple weeks before being taken down - possibly because UKMO informed them the data had not been corrected for errors yet. All very suspicious.

This is not a reliable source. End of story. There have been a half dozen studies on OHC derived from ARGO - they all yield a similar result to the NODC version.

[Snow_Miser]

Tisdale got the data from KNMI climate explorer and has no way of knowing which version of UKMO EN3 they used. The data was only on KNMI for a couple weeks before being taken down - possibly because UKMO informed them the data had not been corrected for errors yet.

This is not a reliable source. End of story. There have been a half dozen studies on OHC derived from ARGO - they all yield a similar result to the NODC version.

I would imagine that they would try and put up the most reliable version of the dataset as they can. The data with the XBT corrections is clearly available from the link that I posted, so I can't see why they wouldn't have linked to the dataset with XBT corrections.

Let's see some of these studies. The UKMO EN3 dataset also has data from ARGO, so I'm not sure what your point here is.

[skierinvermont]

The 15 year trend is positive on HadCtuT4, but it is substantially less than that of the modeled projections over this timeframe. We can claim that there is a discrepency because of the criteria that was given from NOAA when it can be claimed if a discrepency exists or not.

In the next 10 years, we will probably be able to see how strong of a response there is over the next 10 or so years to changes in solar activity and oceanic oscillations.

Can you read?

The simulations rule out (at the 95% level) zero trends for intervals of 15 yr or more, suggesting that an observed absence of warming of this duration is needed to create a discrepancy with the expected present-day warming rate.

There is no 15 year period with a zero trend. It does NOT say that trends for 15 years must have statistically significant positive trends. It says that 95% of the time the trend is >0. That's it.

The number of years required to be 95% confident of a statistically significant positive trend is probably 17-20 years.

Or we could also say with only maybe 80% confidence that 15 year trends should have statistically significant positive trends.

You are either lying or cannot read. Which is it?

[skierinvermont]

I would imagine that they would try and put up the most reliable version of the dataset as they can. The data with the XBT corrections is clearly available from the link that I posted, so I can't see why they wouldn't have linked to the dataset with XBT corrections.

Let's see some of these studies. The UKMO EN3 dataset also has data from ARGO, so I'm not sure what your point here is.

You would imagine.. blah blah blah. Shut up.

Either show us some verification that this is a legitimate source other than a graphic Tisdale posted on WUWT which can no longer be verified, or shut up.

[Snow_Miser]

You would imagine.. blah blah blah. Shut up.

Either show us some verification that this is a legitimate source other than a graphic Tisdale posted on WUWT which can no longer be verified, or shut up.

Okay Andrew, whatever you say. Shows how much scientific support you really have to back up your points when questioned.

[skierinvermont]

The data with the XBT corrections is clearly available from the link that I posted, so I can't see why they wouldn't have linked to the dataset with XBT corrections.

OK so derive it from the XBT corrected version yourself. I would, except the data takes forever to download.

The only verification you've give us so far is "Tisdale said on WUWT it comes from KNMI" plus the assumption that KNMI used the correct version from the UKMO website (Plus the assumption that such a version was even available at the time on their website).

[Snow_Miser]

Can you read?

The simulations rule out (at the 95% level) zero trends for intervals of 15 yr or more, suggesting that an observed absence of warming of this duration is needed to create a discrepancy with the expected present-day warming rate.

There is no 15 year period with a zero trend. It does NOT say that trends for 15 years must have statistically significant positive trends. It says that 95% of the time the trend is >0. That's it.

The number of years required to be 95% confident of a statistically significant positive trend is probably 17-20 years.

Or we could also say with only maybe 80% confidence that 15 year trends should have statistically significant positive trends.

You are either lying or cannot read. Which is it?

If a trend is not statistically significant, you can not claim with certainty that the trend is not zero.

This is basic statistics.

[skierinvermont]

If a trend is not statistically significant, you can not claim with certainty that the trend is not zero.

This is basic statistics.

Nobody said you could.

All that was said was 95% of the time a 15-yr trend will be positive. The 15 year trend is positive. End of story.

Statistics makes very specific claims. The claim that "the 15-yr trend will be positive," has a very different probability than the claim "the 15 year trend will be statistically significantly positive."

[Snow_Miser]

OK so derive it from the XBT corrected version yourself. I would, except the data takes forever to download.

The only verification you've give us so far is "Tisdale said on WUWT it comes from KNMI" plus the assumption that KNMI used the correct version from the UKMO website (Plus the assumption that such a version was even available at the time on their website).

Rational thought process is what this is called. I don't see any reason why KNMI would select a version of the UKMO EN3 dataset with more errors. Maybe you would, and you can explain why.

[Snow_Miser]

Nobody said you could.

All that was said was 95% of the time a 15-yr trend will be positive. The 15 year trend is positive. End of story.

Statistics makes very specific claims. The claim that "the 15-yr trend will be positive," has a very different probability than the claim "the 15 year trend will be statistically significantly positive."

Which means if the trend is not significant, you can't claim that the trend is positive.