Global lithospheric thickness reconstruction using machine learning
Data files
Nov 17, 2023 version files 46.42 MB
-
README.md
8.58 KB
-
Supplementary_Data.xlsx
46.41 MB
Abstract
The lithosphere, as the outermost solid layer of our planet, preserves a progressively more fragmentary record of geological events and processes from Earth's history the further back in time one looks. Thus, the evolution of lithospheric thickness and its cascading impacts on Earth’s tectonic system are presently unknown. Herein, we track the lithospheric thickness history using machine learning based on lithogeochemical big data of basalt. Our results demonstrate that four dramatic lithospheric thinning events occurred during the Paleoarchean, early Paleoproterozoic, Neoproterozoic, and Phanerozoic with intermediate thickening scenarios. These events respectively correspond to supercontinent breakup and assembly periods. Causality investigation further indicates that crustal metamorphic and deformation styles are the feedback of lithospheric thickness. Cross-correlation between lithospheric thickness and metamorphic thermal gradients record the transition from intra-oceanic subduction systems to continental margin plus intra-oceanic in the Paleoarchean and Mesoarchean, and a progressive emergence of large thick continents that allow supercontinent growth, which promoted assembly of the first supercontinent during the Neoarchean.
This is the code and dataset for：
Zhen-Jie Zhang, Guo-Xiong Chen, Timothy Kusky, Jie Yang, Qiu-Ming Cheng. Lithospheric thickness records tectonic evolution by controlling crustal metamorphic and deformation conditions. Science Advances.
Author @Zhenjie Zhang, China University of Geosciences (Beijing)
zjzhang@cugb.eu.cn
Description of the Data and file structure
In this dataset, we include a data file.
Supplementary Data.xlsx
data S1. Lithogeochemical database of global basalts and results for the lithospheric thickness reconstructions.
data S2. Experimental database of basaltic liquid compositions to train the random forest (RF) regression model.
data S3. Global metamorphic thermal gradient database.
data S4. Lithogeochemical data of the Cenozoic basalts and lithospheric thickness reconstruction results in eastern China.
data S5. Metamorphic thermal gradient after Neogene and corresponding lithospheric thickness obtained by Litho1.0.
data S6. Normalized reconstructed PF and T/P series, results for the global and local Pearson correlation coefficient (PCC), time-lagged cross-correlation coefficients (TLCC), and windowed TLCC analyses.
Data description
Our compiled global basaltic rocks database (see data S1) contains 24,194 chronological and oxide analyses of whole-rock, aged between 3.8 Ga to the present. The chronological and oxide data were obtained from the EarthChem data repository (http://portal.earthchem.org/).
Our compiled experimental basaltic database (see data S2) contains 1,392 experimental basaltic liquid compositions in equilibrium with olivine and orthopyroxene, compiled from the LEPR database (https://lepr.earthchem.org/) and published literature.
Our compiled global metamorphic thermal gradient database (see data S3, n = 564) is after Brown et al. (2020, Geology).
Our compiled lithogeochemical data of the Cenozoic basalts (see data S4, n = 60) is after Guo et al. (2020, Geology).
Our compiled global metamorphic thermal gradient after Neogene database (see data S5, n = 15) is after Brown et al. (2020, Geology).
The missing values or empty cells in all data sheets are “n/a” (not applicable).
The wavelet toolbox (Code 7) used in this study is from: http://www.glaciology.net/wavelet-coherence.
Headers for all data sheets
Data S1:
Index: Sample index
Source: Database for the data source
Reference: Original data reference
Latitude: Sample location latitude (degree)
Longitude: Sample location longitude (degree)
Age (Ma): Sample age
Material: Analyzed materials
Type: Sample type (e.g., volcanic, plutonic)
SiO2: SiO2 content (%)
TiO2: TiO2 content (%)
Al2O3: Al2O3 content (%)
FeOT: FeOT content (%)
MnO: MnO content (%)
MgO: MgO content (%)
CaO: CaO content (%)
Na2O: Na2O content (%)
K2O: K2O content (%)
P2O5: P2O5 content (%)
LOI: Loss on ignition (%)
Total: Total content (%)
RF predicted PF: Predicted final pressure (GPa)
Lithospheric thickness (km): Lithospheric thickness (km)
Std (GPa): Standard deviation for predicted final pressure (GPa)
Std (km): Standard deviation for lithospheric thickness (km)
predict-i: The (i+1)th predicted lithospheric thickness (km) [Note: i = 0-99]
Model precision: Model precision for each prediction
average (km): Average lithospheric thickness for 100 times prediction
std: Standard deviation for 100 times prediction of lithospheric thickness (km)
SD (%): Standard deviation for 100 times prediction of lithospheric thickness (%)
Data S2:
Index: Sample index
Experiment: Experiment number in the original reference
Citation: Original data reference
Label: Type of the data source (LEPR Database or Literature)
P (GPa): Experimental pressure (GPa)
T (°C): Experimental temperature (Celsius degree)
SiO2: SiO2 content (%)
TiO2: TiO2 content (%)
Al2O3: Al2O3 content (%)
FeOT: FeOT content (%)
MnO: MnO content (%)
MgO: MgO content (%)
CaO: CaO content (%)
Na2O: Na2O content (%)
K2O: K2O content (%)
Data S3:
Age (Ga): Age of metamorphism (Billion years)
Age (Ma): Age of metamorphism (Million years)
T/P (°C/GPa): Ratio between temperature and pressure (°C/GPa)
Data S4:
Sample: Sample index
Longitude (°E): Sample location longitude (degrees)
SiO2: SiO2 content (%)
TiO2: TiO2 content (%)
Al2O3: Al2O3 content (%)
FeOT: FeOT content (%)
MnO: MnO content (%)
MgO: MgO content (%)
CaO: CaO content (%)
Na2O: Na2O content (%)
K2O: K2O content (%)
P2O5: P2O5 content (%)
RF predicted PF: Predicted final pressure (GPa)
Lithospheric thickness (km): Lithospheric thickness (km)
Data S5:
Index: Sample index
Location: Location of sample
Latitude: Sample location latitude (degrees)
Longitude: Sample location longitude (degrees)
Age (Ga): Age of metamorphism (Billion years)
Peak P/P-T/peak T: Data are Peak P or P-T or peak T
Type: Type of metamorphism (HAG or HUHT or I or LUHPOE or LHPCE or LHPOE or LUHPCE or LHPOB)
Path: Path type (U or CW)
T (°C): Experimental temperature (Celsius degree)
P (GPa): Experimental pressure (GPa)
T / P (°C/GPa): Ratio between temperature and pressure (°C/GPa)
References: Data reference
Lithospheric thickness (km) obtained by Litho1.0: Lithospheric thickness (km) obtained by Litho1.0
Data S6:
Age (Ga): Age of metamorphism (Billion years)
Normalized P (GPa): Normalized pressure
Normalized T/P (°C/GPa): Normalized T/P ratio
Global PCC R2: Global Pearson correlation coefficient
Global PCC P-value: P-value for global Pearson correlation coefficient
Local PCC R2 (window=1.2Ga): Local Pearson correlation coefficient (window=1.2Ga)
Local PCC R2 (window=0.6Ga): Local Pearson correlation coefficient (window=0.6Ga)
TLCC offset (Ga): Offset time (Billion years) for time-lagged cross-correlation coefficients
TLCC Pearson R2: Pearson correlation coefficient for time-lagged cross-correlation coefficients
Windowed TLCC: Windowed time-lagged cross-correlation coefficients (offset from -1Ga to 1Ga with a step = 10 Ma)
Seven code files in our code.zip
1, DML_Thickness_a.py
Causality analysis by causal forests.
treatment = 'Pressure'
outcome = 'Temperature'
2, DML_Thickness_b.py
Causality analysis by causal forests.
treatment = 'Temperature'
outcome = 'Pressure'
3, RF_for_lid.py
Reconstructed lithospheric thickness based on random forest by using whole-rock oxides analyses of global basalts.
4, Lowess.py
Reconstructed PF and T/P series.
5, Thickness-TP_Pearson.py
Global and local Pearson correlation coefficients (PCCs) between PF and T/P.
6, Time-lagged_cross-correlation_coefficients.py
Time-lagged cross-correlation coefficients (TLCC) between the PF and T/P ratio.
7, main_wavelet.m
Wavelet analyze for PF and T/P.
Note:
Codes 1-6 are based on Python 3.8
Requirements:
scipy==1.5.0
numpy==1.21.0
pandas==1.0.5
sklearn==1.3.0
imblearn==0.6.2
dowhy==0.6
econml==0.13.0
seaborn==0.11.2
reg_resampler==2.1.1
matplotlib==3.2.2
Code 7 is based on Matlab.
Sharing/access Information
Data processing steps were conducted in this study:
(1) Causality analysis by causal forest demonstrates that the temperature variable can be disregarded in the estimation of final pressure (PF) (or lithospheric thickness) [by using Codes 1 and 2, and data S2]
(2) The random forest (RF) regression model was trained using experimental basalt data [by using Code 3, and data S1 and S2]
(3) Apply the trained RF regression model to calculate the PF for global basalts [by using Code 3, and data S1 and S2]
(4) Reconstruction of normalized PF and T/P serials since 3.8 Ga [by using Code 4, and data S1 and S3]
(5) Calculation for global and local Pearson correlation coefficients between PF and T/P serials [by using Code 5, and data S6]
(6) Wavelet analysis for the PF and T/P serials [by using Code 7, and data S6]
(7) Calculation for Time-lagged cross-correlation coefficients between the PF and T/P serials [by using Code 6, and data S6]
All data sheets can be read directly using our code files.
You can also easily fork my project by using the MyDDE platform at Global lithospheric thickness reconstruction. No software and environment installation is required.
Our compiled global basaltic rocks database (see data S1) contains 24,194 chronological and oxide analyses of whole-rock, aged between 3.8 Ga to the present. The chronological and oxide data were obtained from the EarthChem data repository (http://portal.earthchem.org/). The lithospheric thicknesses are caclulated by the method in the paper.
Our compiled experimental basaltic database (see data S2) contains 1,392 experimental basaltic liquid compositions in equilibrium with olivine and orthopyroxene, compiled from the LEPR database (https://lepr.earthchem.org/) and published literature. Samples were filtered to include only those basalts with SiO2 (43–55 wt%) and MgO (7–17 wt%). We also manually removed the samples with total oxides below 95 wt%, and missing pressure and temperature.
Our compiled global metamorphic thermal gradient database (see data S3, n= 564) is after Brown et al. (2020, Geology).
data S4 is the lithogeochemical data of the Cenozoic basalts (after Guo et al., 2020, Geology) and lithospheric thickness reconstruction results in eastern China.
data S5 is the metamorphic thermal gradient after Neogene and corresponding lithospheric thickness obtained by Litho1.0.
data S6 is the normalized reconstructed PF and T/P series, results for the global and local Pearson correlation coefficient (PCC), time-lagged cross-correlation coefficients (TLCC), and windowed TLCC analyses.
Codes 1-7 are the computational source codes used in this paper.
Codes 1-6 are based on Python 3.8
Requirements:
scipy==1.5.0
numpy==1.21.0
pandas==1.0.5
sklearn==1.3.0
imblearn==0.6.2
dowhy==0.6
econml==0.13.0
seaborn==0.11.2
reg_resampler==2.1.1
matplotlib==3.2.2
Code 7 is based on Matlab after http://www.glaciology.net/wavelet-coherence.
data S1-S6 can be read by Excel.