Central Lake_2026_brGDGT

This repository contains the R workflow developed to reconstruct and bias-correct mean annual air temperature (MAAT) from lacustrine branched glycerol dialkyl glycerol tetraethers (brGDGTs) using a probabilistic source-mixing approach. The workflow integrates machine-learning estimates of brGDGT provenance derived using the classification framework of Cromartie et al. (2025) with lake- and soil-specific temperature calibrations to correct for mixed-source brGDGTs. Calibration uncertainty is propagated through Monte Carlo simulations to generate confidence intervals for the blended temperature reconstruction. The machine-learning classifier itself is not included; users should refer to Cromartie et al. (2025) for details of the provenance classification framework. Ref. https://doi.org/10.5194/egusphere-2025-526 Cromartie et al. (2025)