[S-GC37] Volatiles in the Earth - from Surface to Deep Mantle

[SGC37-04] The distribution and cycling of water and carbon in the global asthenospheric mantle revealed by machine learning

*Liu [e-mail]¹, Jingjun Zhou¹, Tianting Lei¹, Qunke Xia¹ (1.School of Earth Sciences, Zhejiang University)

Keywords:volatile, distribution, cycling, asthenosphere, machine learning

Water and carbon are important volatiles in the Earth mantle, and play significant roles in many physical and chemical features of the earth interior, and the habitability of the earth surface. Determining the amount of water and carbon in the upper mantle through investigation of global mid-ocean ridge basalts and ocean island basalts is the critical way to quantify the distribution and budget of water and carbon in the mantle, and the cycling of these volatiles between earth interior and surface. However, the reliable water and CO2 content data for the global oceanic basalts are not abundant enough, especially for CO2, due to the heavy degassing during the magmatic processes. The lack of data is not easy to resolved by the normally used calculation from assumed H2O/Ce, or CO2/Ba ratios. Here, based on the compiled complete data sets for the global oceanic basalts with reliable H2O an CO2 content, we trained reliable models by machine learning approaches that can predict the original H2O and CO2 content of the magma through the trace and major elemental concentrations of basalts or melt inclusions. We applied these models to the global mid-ocean ridge basalts, and revealed the updated amount and distribution of water and CO2 content along the global ridges. The predicted results of MORB glasses from this model (n = 1,931) expand the water content database of global MORBs and indicate a broad existence of high-H2O MORBs. This new approach allows us to investigate the water content of MORBs from some ridges lacking previous water content measurements (e.g., the Chile Ridge and the Pacific-Antarctic Ridge) and infer changes in the water content of MORB sources through applying the model to transform fault samples. The results also reveal that predicted CO2 contents and CO2/Ba ratios of global MORBs are highly variable, highlighting the significance of mantle heterogeneity, which can be attributed to the interactions with deep-sourced plumes or recycled crust (oceanic crust with or without sediments). Our findings underscore the potential of machine learning as a powerful tool for investigating the intricate interplay between carbon, mantle composition, and Earth’s long-term geological processes.