Search for: All records

Abstract Accurate and cost-effective quantification of the carbon cycle for agroecosystems at decision-relevant scales is critical to mitigating climate change and ensuring sustainable food production. However, conventional process-based or data-driven modeling approaches alone have large prediction uncertainties due to the complex biogeochemical processes to model and the lack of observations to constrain many key state and flux variables. Here we propose a Knowledge-Guided Machine Learning (KGML) framework that addresses the above challenges by integrating knowledge embedded in a process-based model, high-resolution remote sensing observations, and machine learning (ML) techniques. Using the U.S. Corn Belt as a testbed, we demonstrate that KGML can outperform conventional process-based and black-box ML models in quantifying carbon cycle dynamics. Our high-resolution approach quantitatively reveals 86% more spatial detail of soil organic carbon changes than conventional coarse-resolution approaches. Moreover, we outline a protocol for improving KGML via various paths, which can be generalized to develop hybrid models to better predict complex earth system dynamics. 

Fires are an integral part of many terrestrial ecosystems and have a strong impact on soil properties. While reports of topsoil magnetic enhancement after fires vary widely, recent evidence suggests that plant ashes provide the most significant source of magnetic enhancement after burning. To investigate the magnetic properties of burnt plant material, samples of individual plant species from Iceland and Germany were cleaned and combusted at various temperatures prior to rock magnetic and geochemical characterization. Mass-normalized saturation magnetization values for burnt plant residues increase with the extent of burning in nearly all samples. However, when normalized to the loss on ignition, fewer than half of ash and charcoal samples display magnetic enhancement relative to intact plant material. Thus, while magnetic mineral concentrations generally increase, changes in the total amount of magnetic material are much more variable. Elemental analyses of Icelandic samples reveal that both total plant Fe and saturation magnetization are strongly correlated with Ti and Al, indicating that most of the Fe-bearing magnetic phases originate from inorganic material such as soil and atmospheric dust. Electron microscopy confirmed that inorganic particulate matter remains on most plant surfaces after cleaning. Plants with more textured leaf surfaces retain more dust, and ash from these samples tend to exhibit higher saturation magnetization and metal concentrations. Magnetic properties of plant ash therefore result from the thermal transformation of Fe in both organic compounds and inorganic particulate matter, which become concentrated on a mass basis when organic matter is combusted. These results indicate that the soil magnetic response to burning will vary among sites and regions as a function of 1) fire intensity, 2) the local composition of dust and soil particles on leaf surfaces, and 3) vegetation type and consequent differences in leaf morphologies.