Composite of JER speleothem

---

Consultancy project aimed at reconstructing a paleoclimate time series composite based on 14 proxy records.

How to run

---

The script that creates the composites is the notebook JER_composite in the main directory.

To execute the notebook, follow these steps:

1. Clone the repository

git clone https://github.com/your-username/your-repo-name.git
cd your-repo-name

2. (Optional) Create and activate a virtua environment

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install Required packages

pip install -r requirements.txt

4. Launch Jupyter Notebook

jupyter notebook

---

Description of the Problem

The JER proxy consists of 14 speleothems, spanning from 0.4 to 11 thousand years before present (ka BP), thus covering the majority of the Holocene epoch. Time series derived from nearby speleothems exhibit a pronounced insolation signal. Nevertheless, earlier attempts to construct a composite JER time series have resulted in a detrended time series. Specifically a time series not presenting the insolation signal.

Given the strong hypothesis that insolation drives the isotopic signals, the central challenge of this consultancy is to produce a composite or reconstructed time series in which the insolation signal is preserved.

---

Methodology

To address this challenge, the following steps were undertaken:

• Monte Carlo age-depth modeling was performed on the dating data from the speleothems.

• For each Monte Carlo simulation, the ages corresponding to isotopic values were interpolated, generating an ensemble of time series.

• Probability density estimates of time resolution across the speleothem-interpolated ages were evaluated to determine the mean time resolution for each speleothem.

• A unified time vector, encompassing the full range of the time series, was constructed. The time resolution was selected so as to preserve the majority of information present in the isotopic records.

• The ensemble time series were then interpolated onto this new time vector, resulting in ensembles of isotopic time series aligned in time.

• The ensemble time series from the 14 speleothems were combined in pairs by computing the average isotopic value at each corresponding time point. Unlike prior methodologies that calculate the mean over all overlapping periods and adjust the series to account for bias before averaging, this approach does not include this adjustment. Our observations indicate that such adjustments can eliminate the insolation signal.

• No significant variance changes were observed across the individual speleothem time series; therefore, no normalization or standardization procedures were applied.

As a result, we obtained a composite time series from the 14 speleothems that retains the insolation signal.