the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Testing the reliability of global surface temperature reconstructions of the last glacial cycle
Abstract. Reconstructing past variations of the global mean surface temperature is used to characterise the Earth system response to perturbations as well as validate Earth system simulations. Reconstructing GMST beyond the instrumental period relies on algorithms aggregating local proxy temperature records. Here, we propose to establish standards for the evaluation of the performance of such reconstruction algorithms. Our framework relies on pseudo-proxy experiments. That is, we test the ability of the algorithm to reconstruct a simulated GMST, using artificially generated proxy data created from the same simulation. We apply the framework to an adapted version of the GMST reconstruction algorithm used in Snyder (2016), and the synthesis of marine proxy records for temperature of the last 130 kyr from Jonkers et al. (2020). We use an ensemble of 4 transient simulations of the last glacial cycle or the last 25 kyr for the pseudo-proxy experiments.
We find the algorithm to be able to reconstruct timescales longer than 4 kyr over the last 25 kyr. However, beyond 40 kyr BP, age uncertainty limits the algorithm capability to timescales longer than 15 kyr. The main sources of uncertainty are a factor, that rescales near global mean sea surface temperatures to GMST, the proxy measurement, the specific set of record locations, and potential seasonal bias. Increasing the number of records significantly reduces all sources of uncertainty but the scaling. We also show that a trade-off exists between the inclusion of a large number of records, which reduces the uncertainty on long time scales, and of only records with low age uncertainty, high accumulation rate, and high resolution, which improves the reconstruction of the short timescales.
Finally, the method and the quantitative results presented here can serve as a basis for future evaluations of reconstructions. We also suggest future avenues to improve reconstruction algorithms and discuss the key limitations arising from the proxy data properties.
- Preprint
(1909 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 06 Aug 2024)
-
RC1: 'Comment on egusphere-2024-1387', Bryan C. Lougheed, 17 Jun 2024
reply
see PDF
-
AC1: 'Reply on RC1', Jean-Philippe Baudouin, 11 Jul 2024
reply
Preliminary reply to the comments of the first reviewer.
Before replying in detail to the comments by Bryan C. Lougheed, we would like to clarify a few essential points that our study may have failed to clearly convey.
First, we do not compare the reconstruction data published in Snyder 2016 with synthetic, model-based GMST reconstructions. Instead, the goal of the study is to evaluate the methodology itself, at the base of the reconstruction. We consider therefore only one reconstruction methodology (also called algorithm in the study). The methodology used in our study includes minor changes from the one used by Snyder 2016, as explained in Section 3.1.2. This adaptation of the algorithm has a negligible influence on our results.
The core of our analysis is based on pseudo-proxy experiments. For this, synthetic timeseries (i.e. pseudo-proxy records) are created from climate model data. These timeseries are specific to a climate simulation. The reconstruction algorithm is applied to the synthetic timeseries to produce an estimate of the simulated GMST. This estimation is then compared with the GMST of the simulation, which we can directly compute from the raw model output. In this pseudo-proxy framework, the true GMST is known (it is given by the GMST of the climate simulation output), and we can quantify how different our estimation is from it. We draw our results from this comparison: the estimated GMST can deviate from the simulated GMST due to the perturbations that were added during the construction of the synthetic pseudo-proxy timeseries (e.g., random noise due to measurement uncertainty, or smoothing to represent bioturbation), or due to the way the reconstruction algorithm estimates the GMST from the synthetic timeseries. Some of these results can, with caution, be extrapolated to the real reconstruction published in Snyder 2016 as well as other studies using a similar methodology to reconstruct the GMST (e.g., Friedrich & Timmermann, 2018).
We understand that the use of a slightly different reconstruction algorithm and a new SST proxy database could confuse readers. They can indeed ask: why not use the same database and algorithm? This would indeed enable a stricto sensu evaluation of the reconstruction algorithm published in Snyder 2016. It is however not possible to compute the synthetic model-based timeseries with the database published by Snyder 2016 the same way as for the PalMod database. The data published by Snyder 2016 misses depths and age ensembles, which are needed to account for bioturbation (from accumulation rates) and age uncertainties in the pseudo-proxy records. In particular, simply sampling an age value given the age of a measurement and the uncertainty of the age (Snyder supposed a uniform 5kyr standard deviation) is not an option. This would lead to a break of the stratigraphic principle, stating that ages should always increase with depth. As we show that bioturbation and age uncertainty are two major processes influencing the accuracy of the reconstructed GMST, we cannot use the Snyder database to perform our pseudo-proxy experiments.
We hope this short reply can help readers to better understand our study. We will respond in more detail to all comments by Bryan C. Lougheed in a subsequent reply. We will also consider a more careful framing of our study in a future version of our manuscript to reduce the risk of readers misunderstanding the purpose and methodology of our paper.
Citation: https://doi.org/10.5194/egusphere-2024-1387-AC1 -
RC2: 'Reply on AC1', Bryan C. Lougheed, 12 Jul 2024
reply
Dear authors,
Thank you for your reply in the discussion forum, which has helped me and will also help other readers/reviewers.
It has become clearer to me now that your paper is indeed not data/model comparison, but seeks to calculate GMST from a climate model in two different ways:
(1) GMST directly from a climate model
(2) From the same climate model, create synthetic sediment core records at known core locations, and calculate GMST from those synthetic core records using Snyder GMST.
Subsequently, you compare the two GMSTs that you have calculated using the above two methods, whereby the second method mimics the current state-of-the-art.
If I understand correctly, you only use real-world data to provide realistic age-depth information (i.e. sedimentation rates) for the synthetic core records, which makes sense.
I did not get the above at first reading of the paper. Some revisions leading to a more concise and shorter introduction text, as well as the use of clear and consistent terms would help readers similar to myself better understand the experimental design. Perhaps it would also be useful to include a flowchart as a figure? I have attached an example under the big blue paperclip icon. Hopefully it is a correct summary of the experimental design!
Kind regards,
Bryan
-
RC2: 'Reply on AC1', Bryan C. Lougheed, 12 Jul 2024
reply
-
AC1: 'Reply on RC1', Jean-Philippe Baudouin, 11 Jul 2024
reply
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
249 | 74 | 18 | 341 | 10 | 11 |
- HTML: 249
- PDF: 74
- XML: 18
- Total: 341
- BibTeX: 10
- EndNote: 11
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1