CALIBRATION OF CLIMOFUNCTIONS FROM MODERN SOILS IN SOUTHERN PERU USING LIBS (LASER-INDUCED BREAKDOWN SPECTROSCOPY) AND APPLICATIONS TO CLIMATE RECONSTRUCTION

LIBS (Laser-Induced Breakdown Spectroscopy) is a field capable, handheld tool that measures the elemental composition of solid materials. Recent advances in LIBS calibration allow quantitative analysis of geological materials for major elements. LIBS analysis of paleosols may be used to reconstruct climate history through the application of climofunctions - ratios of mobile major elements that covary with certain climate parameters. Application of climofunctions to the paleo-record requires calibration to the study area by analysis of local modern soils formed under known climate conditions. Here, we apply LIBS to the analysis of soils from the central Andes of southern Peru.

The central Andes host diverse tectonomorphic regions with distinct climate conditions - the hyper-arid, low-altitude forearc along the coast; the high alpine settings of the Western and Eastern Cordilleras with peak heights over 6000 m; the semi-arid, cool Altiplano Plateau with an average elevation of 4000 m; and the hot and wet lowland jungles of the Subandean zone and foreland basin. In the summer of 2022, 52 samples of Holocene–modern soils were sampled from transects spanning each of these settings. Climate conditions of soil formation are based on publicly available rainfall and temperature data from long-term monitoring stations. LIBS composition data collected for each soil is used to calculate values of common climofunctions for each soil (paleosol weathering index, chemical index of alteration, bases/alumina ratio), and these climofunctions are plotted against known climate conditions. Regression equations and residual values are calculated for each to identify the most robust climofunctions for use in LIBS analysis as well as provide calibrations to apply selected climofunctions to analysis of paleosols for deep-time climate reconstruction.

Ongoing work is focused on expanding the dataset with 55 additional soil samples from western Bolivia. Future work will apply machine learning algorithms to this expanded soil dataset to identify and calibrate novel climofunctions specific to the study area. Climofunction calibrations and data presented here will serve as a baseline for future work to provide a comparison for the efficacy of novel climofunctions identified by machine learning.