FROG: A global machine-learning temperature calibration for branched GDGTs in soils and peats

Véquaud, Pierre; Thibault, Alexandre; Derenne, Sylvie; Anquetil, Christelle; Collin, Sylvie; Contreras, Sergio; Nottingham, Andrew T.; Sabatier, Pierre; Werne, Josef P.; Huguet, Arnaud

doi:10.1016/j.gca.2021.12.007

Cited by 42 publications

(10 citation statements)

References 92 publications

(206 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We found that they have similar trends (Figure 2 and Figure S5 in Supporting Information ), although the average temperature reconstructed by the former is about 2°C higher than the later, and a higher deviation was observed during colder periods, which is consistent with the modern observation data (Wang et al., 2020). A Bayesian calibration (BayMBT) (Crampton‐Flood et al., 2020) and a random Forest Regression for PaleOMAAT using brGDGTs (FROG) calibration (Véquaud et al., 2022) in soils were also proposed to calibrate the relationship of the 5‐methyl brGDGTs and the MAAT. However, in the Lingtai loess‐paleosol sequence, the IIb and IIc compounds (Figure S1 in Supporting Information ) required in these two models are usually below the detection limit, so, these two models may be currently inapplicable.…”

Section: Methodsmentioning

confidence: 99%

Decoupled Land and Ocean Temperature Trends in the Early‐Middle Pleistocene

Liu

Hong

et al. 2022

Geophysical Research Letters

View full text Add to dashboard Cite

The current global warming trend has exacerbated our concerns about future changes in land surface temperature (LST), a parameter that has profound impact on terrestrial ecosystems including many aspects of human life (IPCC, 2019). In the quest of better understanding continental climatic changes, documenting and deciphering underlying physical principles of long-term LST evolution in historical periods are critical. However, due to the lack of suitable tools, long-term terrestrial paleotemperature reconstruction is extremely rare, which increases the uncertainty of our projection of future trend for temperature changes. It remains uncertain whether land temperature should have followed the same long-term cooling trend as inferred from marine records over the

show abstract

Section: Methodsmentioning

confidence: 99%

Decoupled Land and Ocean Temperature Trends in the Early‐Middle Pleistocene

Liu

Hong

et al. 2022

Geophysical Research Letters

View full text Add to dashboard Cite

show abstract

“…In the field of biomarker‐based paleoclimatology, classification algorithms have been applied to identify sources of alkenones (Zheng et al., 2019), as well as plant waxes (Peaple et al., 2021). Machine learning regression algorithms, as well, as deep neural network applications, have also been applied to GDGTs as in order to generate temperature calibrations (Dunkley Jones et al., 2020; Véquaud et al., 2022; Zheng et al., 2022). Here, we use a compilation of GDGT distributions in 1,153 globally distributed soils, peats, and sediments from diverse depositional environments to train a classification algorithm which is capable of identifying the environment in which a sample was formed based on the distribution of both branched and isoGDGTs.…”

Section: Introductionmentioning

confidence: 99%

Development and Application of the Branched and Isoprenoid GDGT Machine Learning Classification Algorithm (BIGMaC) for Paleoenvironmental Reconstruction

Martínez‐Sosa

Tierney

Pérez‐Angel

et al. 2023

Paleoceanog and Paleoclimatol

View full text Add to dashboard Cite

Glycerol dialkyl glycerol tetraethers (GDGTs), both archaeal isoprenoid GDGTs (isoGDGTs) and bacterial branched GDGTs (brGDGTs), have been used in paleoclimate studies to reconstruct environmental conditions. Since GDGTs are produced in many types of environments, their relative abundances also depend on the depositional setting. This suggests that the distribution of GDGTs also preserves useful information that can be used more broadly to infer these depositional environments in the geological past. Here, we combined existing iso‐ and brGDGT relative abundance data with newly analyzed samples to generate a database of 1,153 samples from several modern sedimentary settings. We observed a robust relationship between the depositional environment and the relative abundances of GDGTs in our samples. This data set was used to train and test the Branched and isoGDGT Machine learning Classification (BIGMaC) algorithm, which identifies the environment a sample comes from based on the distribution of GDGTs with high precision and recall (F1 = 0.95). We tested the model on the sedimentary record from the Giraffe kimberlite pipe, an Eocene maar in subantarctic Canada, and found that the BIGMaC reconstruction agrees with independent stratigraphic and palynological information, provides new information about the paleoenvironment of this site, and helps improve its paleotemperature reconstruction. In contrast, we also include an example from the PETM‐aged Cobham lignite as a cautionary example that illustrates the limitations of the algorithm. We propose that in cases where paleoenvironments are unknown or are changing, BIGMaC can be applied in concert with other proxies to generate more refined paleoclimate records.

show abstract

“…Another type of GDGTs with branched alkyl chains – so-called branched GDGTs (brGDGTs) – were suggested to be produced by bacteria and are ubiquitous in aquatic and terrestrial environments ( Schouten et al, 2013 ). Their analysis in soils/peats ( Véquaud et al, 2022 ) and lake sediments ( Martínez-Sosa et al, 2021 ) distributed worldwide showed that their structure varies mainly with air temperature and soil pH, making them increasingly used as temperature and pH paleoproxies. BrGDGTs are the only microbial organic proxies which can be used for temperature reconstructions in both aquatic and terrestrial settings.…”

Section: Introductionmentioning

confidence: 99%

“…BrGDGTs are the only microbial organic proxies which can be used for temperature reconstructions in both aquatic and terrestrial settings. Nevertheless, paleoenvironmental data derived from these molecules have to be interpreted with care, as (i) their source microorganisms remain unknown, although some might belong to the phylum Acidobacteria ( Sinninghe Damsteé et al, 2011 ; Sinninghe Damsté et al, 2018 ) and (ii) high uncertainty is associated with global brGDGT-mean annual air temperature (MAAT) calibrations (>4°C; Véquaud et al, 2022 ). The development of new environmental proxies, independent and complementary to brGDGTs, is crucial to improve the reliability and accuracy of continental reconstructions.…”

Section: Introductionmentioning

confidence: 99%

Membrane lipid adaptation of soil Bacteroidetes isolates to temperature and pH

et al. 2023

Self Cite

View full text Add to dashboard Cite

3-hydroxy fatty acids (3-OH FAs) are characteristic components of the Gram-negative bacterial membrane, recently proposed as promising temperature and pH (paleo) proxies in soil. Nevertheless, to date, the relationships between the 3-OH FA distribution and temperature/pH are only based on empirical studies, with no ground truthing work at the microbial level. This work investigated the influence of growth temperature and pH on the lipid composition of three strains of soil Gram-negative bacteria belonging to the Bacteroidetes phylum. Even though non-hydroxy FAs were more abundant than 3-OH FAs in the investigated strains, our results suggest that 3-OH FAs are involved in the membrane adaptation of these bacteria to temperature. The strains shared a common adaptation mechanism to temperature, with a significant increase in the ratio of anteiso vs. iso or normal 3-OH FAs at lower temperature. In contrast with temperature, no common adaptation mechanism to pH was observed, as the variations in the FA lipid profiles differed from one strain to another. We suggest that models reconstructing environmental changes in soils should include the whole suite of 3-OH FAs present in the membrane of Gram-negative bacteria, as all of them could be influenced by temperature or pH at the microbial level.

show abstract

FROG: A global machine-learning temperature calibration for branched GDGTs in soils and peats

Cited by 42 publications

References 92 publications

Decoupled Land and Ocean Temperature Trends in the Early‐Middle Pleistocene

Decoupled Land and Ocean Temperature Trends in the Early‐Middle Pleistocene

Development and Application of the Branched and Isoprenoid GDGT Machine Learning Classification Algorithm (BIGMaC) for Paleoenvironmental Reconstruction

Membrane lipid adaptation of soil Bacteroidetes isolates to temperature and pH

Contact Info

Product

Resources

About