Trenberth Debunks Himself

[flacaltenn]

Don't be stupid. That they reset them every ten days means they can be no more than 10 days away from the reality of the measurement. And for what it's worth, I can see at least three of BTK's four volcanoes in the NOAA data. And these comments about NOAA's opinions of BTKs work: how about some quotes supporting that accusation? Got any?

Think I already told you that you don't have to "reset" the model to truth data. You use the truth data to "revise" the run periodically. Does NOT mean that the model is only 10 days from truth. (See Uncle Westwall's comments).. And if you see volcanic signatures in the NOAA data, it's only because YOU WANT to see them..

NOAA has a severe disagreement the BTK explanation of how the heat got there and when.
Seeing as how they have REAL data without screwy artifacts, they most likely are looking critically at 0.1degC temp diffs that appear ONLY in the BTK modeling.

 

[Abraham3]

Again, if NOAA has made such comments about BTK 2013, you should be able to provide a link to them.

 

[flacaltenn]

Again, if NOAA has made such comments about BTK 2013, you should be able to provide a link to them.

OK Abe.. Its on my list of stuff to do. But I would suggest that if you value this study so much, you should be following like the rest of us do.. or is it just that SkS hasnt given you the talking points yet.

 

[SwimExpert]

Don't be stupid. That they reset them every ten days means they can be no more than 10 days away from the reality of the measurement. And for what it's worth, I can see at least three of BTK's four volcanoes in the NOAA data. And these comments about NOAA's opinions of BTKs work: how about some quotes supporting that accusation? Got any?

Think I already told you that you don't have to "reset" the model to truth data. You use the truth data to "revise" the run periodically. Does NOT mean that the model is only 10 days from truth. (See Uncle Westwall's comments).. And if you see volcanic signatures in the NOAA data, it's only because YOU WANT to see them..

NOAA has a severe disagreement the BTK explanation of how the heat got there and when.
Seeing as how they have REAL data without screwy artifacts, they most likely are looking critically at 0.1degC temp diffs that appear ONLY in the BTK modeling.

Actually -- and maybe I've gotten myself a little lost somewhere -- but if I understand what Abby is claiming correctly, doesn't the model predictions cease to be model predictions and instead become a historical record?

 

[flacaltenn]

Don't be stupid. That they reset them every ten days means they can be no more than 10 days away from the reality of the measurement. And for what it's worth, I can see at least three of BTK's four volcanoes in the NOAA data. And these comments about NOAA's opinions of BTKs work: how about some quotes supporting that accusation? Got any?

Think I already told you that you don't have to "reset" the model to truth data. You use the truth data to "revise" the run periodically. Does NOT mean that the model is only 10 days from truth. (See Uncle Westwall's comments).. And if you see volcanic signatures in the NOAA data, it's only because YOU WANT to see them..

NOAA has a severe disagreement the BTK explanation of how the heat got there and when.
Seeing as how they have REAL data without screwy artifacts, they most likely are looking critically at 0.1degC temp diffs that appear ONLY in the BTK modeling.

Actually -- and maybe I've gotten myself a little lost somewhere -- but if I understand what Abby is claiming correctly, doesn't the model predictions cease to be model predictions and instead become a historical record?

It WOULD --- if you absolutely restarted the model with rel data values every 10 days. But what im saying is it makes more sense to use the real data to reweight where the model is takin you. So the restart MODIFIES and DIRECTS the output, but is not actually assuming the data values. In other words, the truth data is an input like every other variable in the model and is given a weighting that guarantees general agreement with the data, but allows the model to find its own contributions to the result..

Like those suspicious volcanic events for instance....

 

[SwimExpert]

Okay, I'm back up to speed now.

 

[Abraham3]

Think I already told you that you don't have to "reset" the model to truth data. You use the truth data to "revise" the run periodically. Does NOT mean that the model is only 10 days from truth. (See Uncle Westwall's comments).. And if you see volcanic signatures in the NOAA data, it's only because YOU WANT to see them..

NOAA has a severe disagreement the BTK explanation of how the heat got there and when.
Seeing as how they have REAL data without screwy artifacts, they most likely are looking critically at 0.1degC temp diffs that appear ONLY in the BTK modeling.

Actually -- and maybe I've gotten myself a little lost somewhere -- but if I understand what Abby is claiming correctly, doesn't the model predictions cease to be model predictions and instead become a historical record?

It WOULD --- if you absolutely restarted the model with rel data values every 10 days. But what im saying is it makes more sense to use the real data to reweight where the model is takin you. So the restart MODIFIES and DIRECTS the output, but is not actually assuming the data values. In other words, the truth data is an input like every other variable in the model and is given a weighting that guarantees general agreement with the data, but allows the model to find its own contributions to the result..

Like those suspicious volcanic events for instance....

From BTK 2013:

2 The Ocean Reanalysis

[6] ORAS4 has been produced by combining, every 10 days, the output of an ocean model forced by atmospheric reanalysis fluxes and quality controlled ocean observations. These consist of temperature and salinity (T/S) profiles from the Hadley Centre's EN3 data collection [Ingleby and Huddleston, 2007], which include expendable bathythermographs (T only, with depth corrections from Table 1 of Wijffels et al. [2008]), conductivity-temperature-depth sensors (T/S), TAO/TRITON/PIRATA/RAMA moorings (T/S), Argo profilers (T/S), and autonomous pinniped bathythermograph (or elephant seals, T/S). Altimeter-derived along track sea level anomalies from AVISO are also assimilated. Gridded maps of SST from NOAA are used to adjust the heat fluxes via strong relaxation, and altimeter global mean sea-levels are used to constrain the global average of the fresh-water flux. The ocean model horizontal resolution is approximately 1°, refined meridionally down to 1/3° at the equator. There are 42 vertical levels with separations varying smoothly from 10 m at the surface to 300 m at the bottom, with partial cell topography.

[7] A model bias correction [BMW13] is used to reduce potential spurious variability resulting from changes in the observing system. The bias correction first guess—a seasonal cycle of 3-D model error—is estimated from the data-rich Argo period, and applied to ORAS4 from the beginning of the record. This is updated as the analysis progresses via an adaptive scheme (see BMW13 for details; see also Figure S3 of the auxiliary material). The five ensemble members of ORAS4 sample plausible uncertainties in the wind forcing, observation coverage, and the deep ocean. The uncertainty is probably underestimated in ORAS4, because the uncertainty in observations and their quality control [Lyman et al., 2010] is not sampled. Quality improvements in ORAS4 relative to earlier ocean reanalyses stem from the use of improved atmospheric surface fluxes, improved data assimilation, and more comprehensive quality-control of the observation data set, with important corrections to the ocean observations.

[8] The methods section S01 in the auxiliary material provides more specific information on the model, surface forcing, observation data sets, bias correction and ensemble generation. A detailed description and evaluation of ORAS4 is given in BMW13, and a discussion of the sensitivity of the reanalysis to several aspects not included in the ensemble generation.
*******************************************************************
The purpose of using the ORAS4 model is to EXTRAPOLATE from the actual measurements an estimation of the state of the remainder of the ocean and produce an accurate estimate of ocean heat content extending deeper than 2000m as well as into unsampled areas.

 

[Abraham3]

Backing up:

1 Introduction
[2] Increasing greenhouse gases have apparently led to an increasing radiative imbalance at the top of the atmosphere (TOA) of order 0.5–1 W m–2 in the past two decades, based on observational [Trenberth, 2009; Trenberth et al., 2009; Murphy et al., 2009] and model [Hansen et al., 2011] estimates. Over the past 50 years, the oceans have absorbed about 90% of the total heat added to the climate system [Bindoff et al., 2007], while the rest goes to melting sea and land ice, and warming the land surface and atmosphere. Although the Mount Pinatubo eruption in 1991 caused a short-term reduction in TOA radiation [Trenberth and Dai, 2007], increasing greenhouse gases should have led to increasing warming. However, sea surface temperature (SST) increases stalled in the 2000s and this is also reflected in upper ocean heat content (OHC) for the top 700 m in several analyses [Levitus et al., 2009, 2012; Lyman et al., 2010]. Although the energy imbalance from 1993 to 2003 could be accounted for, it was not possible to explain the energy imbalance from 2004–2008. This led to the concept of “missing energy” [Trenberth and Fasullo, 2010].

[3] Several recent modeling studies [Easterling and Wehner, 2009; Palmer et al., 2011; Katsman and van Oldenborgh, 2011; Meehl et al., 2011] advocate for the role of the deep ocean in the heat uptake. However, there are challenges in sampling the deeper reaches of the ocean although previous estimates [Purkey and Johnson, 2010; Kouketsu et al., 2011] found the deep ocean (below 1000 m) gaining heat at rates of less than 0.1 Wm−2 (95% confidence interval; global average). The most recent estimate of the ocean warming in 0–2000 m depth range of 0.39 W m–2 (per unit area of the World Ocean) from 1955 to 2010 emphasized the dominant role of the 700–2000 m depth range in the heat uptake (about one-third of the total warming) [Levitus et al., 2012; L12 hereafter] .

[4] A key question is how recent and robust is the role of deep ocean in the heat uptake, because the advent of the Argo ocean observing system is known to have a profound impact in ocean state estimations [Balmaseda et al., 2007]. Another question is how much disruption there is of the warming trends by natural variability and from the volcanic eruptions, which contributed to the TOA imbalance but have not been factored into most analyses of OHC variations. Several decadal hiatus periods of the upper ocean warming associated with natural decadal and interannual variability, such as La Niña events, have been found in a model [Meehl et al., 2011]. Recently, the 2001–2010 interannual variations of TOA radiation and OHC have been associated with El Niño-Southern Oscillation (ENSO) [Loeb et al., 2012].

[5] Here we investigate the time evolution of the global OHC at different depth ranges for the period 1958–2009 using the latest European Centre for Medium-Range Weather Forecasts ocean reanalysis system 4 (ORAS4) [Balmaseda et al., 2013, BMW13 in what follows], which provides a continuous record of the history of the global ocean by combining a wealth of observational information (section 2). Section 3 presents the time evolution of ORAS4 OHC at different depth ranges, the geographical distribution of the warming trends, and several sensitivity experiments, including the impact of Argo. The main findings are summarized in section 4.
***************************************
BTK are thus not alone in finding that heat is preferentially moving into the deep ocean.

 

[flacaltenn]

Actually -- and maybe I've gotten myself a little lost somewhere -- but if I understand what Abby is claiming correctly, doesn't the model predictions cease to be model predictions and instead become a historical record?

It WOULD --- if you absolutely restarted the model with rel data values every 10 days. But what im saying is it makes more sense to use the real data to reweight where the model is takin you. So the restart MODIFIES and DIRECTS the output, but is not actually assuming the data values. In other words, the truth data is an input like every other variable in the model and is given a weighting that guarantees general agreement with the data, but allows the model to find its own contributions to the result..

Like those suspicious volcanic events for instance....

From BTK 2013:

2 The Ocean Reanalysis

[6] ORAS4 has been produced by combining, every 10 days, the output of an ocean model forced by atmospheric reanalysis fluxes and quality controlled ocean observations. These consist of temperature and salinity (T/S) profiles from the Hadley Centre's EN3 data collection [Ingleby and Huddleston, 2007], which include expendable bathythermographs (T only, with depth corrections from Table 1 of Wijffels et al. [2008]), conductivity-temperature-depth sensors (T/S), TAO/TRITON/PIRATA/RAMA moorings (T/S), Argo profilers (T/S), and autonomous pinniped bathythermograph (or elephant seals, T/S). Altimeter-derived along track sea level anomalies from AVISO are also assimilated. Gridded maps of SST from NOAA are used to adjust the heat fluxes via strong relaxation, and altimeter global mean sea-levels are used to constrain the global average of the fresh-water flux. The ocean model horizontal resolution is approximately 1°, refined meridionally down to 1/3° at the equator. There are 42 vertical levels with separations varying smoothly from 10 m at the surface to 300 m at the bottom, with partial cell topography.

[7] A model bias correction [BMW13] is used to reduce potential spurious variability resulting from changes in the observing system. The bias correction first guess—a seasonal cycle of 3-D model error—is estimated from the data-rich Argo period, and applied to ORAS4 from the beginning of the record. This is updated as the analysis progresses via an adaptive scheme (see BMW13 for details; see also Figure S3 of the auxiliary material). The five ensemble members of ORAS4 sample plausible uncertainties in the wind forcing, observation coverage, and the deep ocean. The uncertainty is probably underestimated in ORAS4, because the uncertainty in observations and their quality control [Lyman et al., 2010] is not sampled. Quality improvements in ORAS4 relative to earlier ocean reanalyses stem from the use of improved atmospheric surface fluxes, improved data assimilation, and more comprehensive quality-control of the observation data set, with important corrections to the ocean observations.

[8] The methods section S01 in the auxiliary material provides more specific information on the model, surface forcing, observation data sets, bias correction and ensemble generation. A detailed description and evaluation of ORAS4 is given in BMW13, and a discussion of the sensitivity of the reanalysis to several aspects not included in the ensemble generation.
*******************************************************************
The purpose of using the ORAS4 model is to EXTRAPOLATE from the actual measurements an estimation of the state of the remainder of the ocean and produce an accurate estimate of ocean heat content extending deeper than 2000m as well as into unsampled areas.

Yup.. That 1st bolded statement backs what I asserted. The keyword in COMBINING.. Not forcing the output to the data values. Like I said -- it's a reweighting of truth data just like any other weighted inputs to the model.

As for your trailing comment. Let them publish a few verifications on the MEASURABLE and INTERESTING parts of the ocean --- and then we can talk about Davy Jones locker.

 

[flacaltenn]

Interesting Figure from earlier Balsmeda paper on ORAS4 reanalysis.. Compares the raw, corrected data versus the model running void of corrections to truth data.

Blue is the variance in the model output. Black is raw data, Red is Corrected Raw..
The model itself is a crap shoot below 2000m, So forget this excuse that's the only rationale for re-analysis.

 

[Abraham3]

It WOULD --- if you absolutely restarted the model with rel data values every 10 days. But what im saying is it makes more sense to use the real data to reweight where the model is takin you. So the restart MODIFIES and DIRECTS the output, but is not actually assuming the data values. In other words, the truth data is an input like every other variable in the model and is given a weighting that guarantees general agreement with the data, but allows the model to find its own contributions to the result..

Like those suspicious volcanic events for instance....

From BTK 2013:

2 The Ocean Reanalysis

[6] ORAS4 has been produced by combining, every 10 days, the output of an ocean model forced by atmospheric reanalysis fluxes and quality controlled ocean observations. These consist of temperature and salinity (T/S) profiles from the Hadley Centre's EN3 data collection [Ingleby and Huddleston, 2007], which include expendable bathythermographs (T only, with depth corrections from Table 1 of Wijffels et al. [2008]), conductivity-temperature-depth sensors (T/S), TAO/TRITON/PIRATA/RAMA moorings (T/S), Argo profilers (T/S), and autonomous pinniped bathythermograph (or elephant seals, T/S). Altimeter-derived along track sea level anomalies from AVISO are also assimilated. Gridded maps of SST from NOAA are used to adjust the heat fluxes via strong relaxation, and altimeter global mean sea-levels are used to constrain the global average of the fresh-water flux. The ocean model horizontal resolution is approximately 1°, refined meridionally down to 1/3° at the equator. There are 42 vertical levels with separations varying smoothly from 10 m at the surface to 300 m at the bottom, with partial cell topography.

[7] A model bias correction [BMW13] is used to reduce potential spurious variability resulting from changes in the observing system. The bias correction first guess—a seasonal cycle of 3-D model error—is estimated from the data-rich Argo period, and applied to ORAS4 from the beginning of the record. This is updated as the analysis progresses via an adaptive scheme (see BMW13 for details; see also Figure S3 of the auxiliary material). The five ensemble members of ORAS4 sample plausible uncertainties in the wind forcing, observation coverage, and the deep ocean. The uncertainty is probably underestimated in ORAS4, because the uncertainty in observations and their quality control [Lyman et al., 2010] is not sampled. Quality improvements in ORAS4 relative to earlier ocean reanalyses stem from the use of improved atmospheric surface fluxes, improved data assimilation, and more comprehensive quality-control of the observation data set, with important corrections to the ocean observations.

[8] The methods section S01 in the auxiliary material provides more specific information on the model, surface forcing, observation data sets, bias correction and ensemble generation. A detailed description and evaluation of ORAS4 is given in BMW13, and a discussion of the sensitivity of the reanalysis to several aspects not included in the ensemble generation.
*******************************************************************
The purpose of using the ORAS4 model is to EXTRAPOLATE from the actual measurements an estimation of the state of the remainder of the ocean and produce an accurate estimate of ocean heat content extending deeper than 2000m as well as into unsampled areas.

Yup.. That 1st bolded statement backs what I asserted. The keyword in COMBINING.. Not forcing the output to the data values. Like I said -- it's a reweighting of truth data just like any other weighted inputs to the model.

As for your trailing comment. Let them publish a few verifications on the MEASURABLE and INTERESTING parts of the ocean --- and then we can talk about Davy Jones locker.

Nowhere do I see BTK use the term "weighted".

As to your verifications, BTK seems to have been a verification of Easterling and Wehner, 2009; Palmer et al., 2011; Katsman and van Oldenborgh, 2011; and Meehl et al., 2011. Blesides, it's quite obvious you have no intention of ever accepting any data that goes against your TSI hypothesis or that supports the IPCC findings.

 

[Abraham3]

Abraham,

Let's say we're in 2017 and temperature is still controlled by the noise of the enso? How long will things have to be stalled before you question things.

I believe I already answered that. As long as our satellites show an radiative imbalance at the Top of the Atmosphere - as they do now and by an amount that has been growing for the last decade - the Earth is getting warmer. The primary candidate for the primary cause of that warming is human GHG emissions.

Nothing has changed.