Data-intensive ecological modeling and informatics tools

Huang, Xin (2023) Data-intensive ecological modeling and informatics tools. Doctoral thesis, Northern Arizona University.

Text Huang_2023_data-intensive_ecological_modeling_informatics_tools.pdf - Published Version Download (5MB)

Abstract

The advent of enormous observations and measurements offers tremendous potential to improve our predictive understanding of ecological processes under global climate change. However, despite the abundance of data, the ability of ecological prediction is limited. It remains challenging to translate the ultimate value of big data into actionable insights. The overall research question in my dissertation is how to develop informatics frameworks that can better integrate big data with process-based models to improve predictive understanding of ecological processes and ecological forecast of future ecosystem states. In Chapter 1, I introduce the motivation, data diversity, and current research state on ecological prediction. In Chapter 2, I present a model-independent data assimilation (MIDA) module. The easy-to-use capability of MIDA enables ecologists to conduct data model fusion in various applications. In Chapter 3, I apply MIDA in an 8-yr tundra warming experiment to advance our mechanistic understanding of how carbon cycle in the tundra ecosystem would change with global warming. I found that without considering elevated water table depth, the Terrestrial ECOsystem (TECO) model will underestimate carbon emissions. In Chapter 4, I develop an assessment method to evaluate the effects of forcing corrections on the improvement of forecast accuracy. Compared to the conventional method, this new assessment method can identify the decrease in forecast accuracy for pool-based variables (e.g., Leaf). I also found that only response variables with memory (e.g., carbon pools) can benefit from periodic forcing correction. In Chapter 5, I conclude all findings in my dissertation and provide my perspectives on data-model integration to advance our predictive understanding of the land carbon cycle. Overall, my dissertation work develops new informatic tools and methods to tackle three uncertainty sources (i.e., empirical parameter values, incomplete model structures, and unrealistic future forcings) which impede realistic ecological predictions. My dissertation indicates an avenue about how to translate big data to improvements in the predictive capability of terrestrial ecosystem carbon cycles. Furthermore, it will benefit policy making about mitigating global warming and the associated climate change.

Item Type:	Thesis (Doctoral)
Publisher’s Statement:	© Copyright is held by the author. Digital access to this material is made possible by the Cline Library, Northern Arizona University. Further transmission, reproduction or presentation of protected items is prohibited except with permission of the author.
Keywords:	cyberinfrastructure; data assimilation; ecological;carbon cycle; prediction; global change experiment; terrestrial; climate change; ecosystem models; uncertainty
Subjects:	T Technology > TD Environmental technology. Sanitary engineering
NAU Depositing Author Academic Status:	Student
Department/Unit:	Graduate College > Theses and Dissertations College of Engineering, Informatics, and Applied Sciences > School of Informatics, Computing, and Cyber Systems
Date Deposited:	12 May 2025 21:48
Last Modified:	12 May 2025 21:48
URI:	https://openknowledge.nau.edu/id/eprint/6131

Actions (login required)

IR Staff Record View

Downloads