UC Berkeley

Soil organic matter (SOM) is the largest actively-cycling terrestrial reservoir of carbon (C) and an integral component of thriving natural and managed ecosystems. Climate­ and land­use­induced changes in plant inputs to soil may result in changes to soil organic carbon (SOC) storage with large implications for the global C balance. Despite the potential for large C feedbacks, the processes that dictate the response of SOC to changes in plant inputs are still poorly understood and inadequately represented in models, thus, limiting their predictive capability for policy and management decisions. The overarching goal of my Ph.D. research was to improve our process­level understanding of the response of SOC to changes in plant inputs, with interest in the competing microbial and mineral mechanisms that govern the decomposition and stabilization of SOC, respectively, and the capacity of soils to store C. To this end, I leveraged data collection and synthesis, meta-analysis, and process-based modeling to advance our understanding and provide predictive tools.

In Chapter 1, I provide an introduction and overview of the field of SOC modeling, and outline the trail map of my doctoral research. I then explore, in Chapter 2, the potential magnitude of the soil C sink over the last decade using remotely-sensed observations as a proxy for the rate of plant inputs. My findings suggest that soils have played a large role in the terrestrial carbon sink, especially in grassland ecosystems. This provides a strong motivation for better understanding the underlying mechanisms at play, since, as I show in Chapter 3, using simple SOC models to infer complex dynamics can lead to model artifacts and false mechanistic attribution. I highlight potential pitfalls and make recommendations for future modeling studies to include important mechanisms – e.g., microbial and mineral interactions – and to use more comprehensive data streams.

Accordingly, in Chapters 4 and 5, I focus on the representation of microbes and minerals, respectively, in process-based SOC models. Specifically, in Chapter 4, I diagnose unrealistic behaviors observed in recent mechanistic models and propose modifications by leveraging ecological theory and analysis of long-term litter (plant input) manipulation experiments. In Chapter 5, I investigate the role of minerals in the capacity of soils to store C, and propose model formulations that match an extensive observational synthesis of mineral-associated C across soil types.

With advancements in process-based SOC models, and their widespread application, comes the need for accurate numerical methods to efficiently solve such systems of equations. Thus, in Chapter 6, I propose a novel numerical integration method that is uniquely suited for solving coupled, depth-resolved equations that arise in advection-dominated environmental systems. This class of equations is common in SOC modeling, and I present an example to that effect and assess numerical performance.

Finally, in Chapter 7, I close by discussing the broad implications of this research in the context of society. A deep understanding of soil C and nutrient cycling is essential for building predictive tools to better inform land management and conservation policy. I end by describing gaps and scaling challenges that motivate my future work.