Active current projects

Forest Futures: Geospatial data science and artificial intelligence for sustainable forestry in Michigan

PI: Kai Zhu, co-PIs: Joshua P. Newell, Dimitrios Gounaridis
Sponsor: US Department of Agriculture, McIntire-Stennis Cooperative Forestry Research Program
Summary: This innovative research will integrate ground-level forest inventories and high-resolution satellite imagery with advanced ecological modeling to capture and predict forest dynamics. The project’s core objectives include an empirical assessment of historical forest changes using various datasets; projecting future forest dynamics through mechanistic models; and developing stakeholder engagement platforms featuring an AI-driven interface for interactive insights. By addressing pivotal knowledge gaps, the team aims to inform and transform forest management strategies, utilizing the maple syrup industry as a case study while extending benefits to the broader forestry sector.

Linking phenological change to range change in North American plant species

PI: Sydne Record, co-PIs: Linda Black Elk, Kai Zhu
Sponsor: National Science Foundation, Environmental Data Science Innovation & Inclusion Lab
Summary: This project aims to comprehensively understand and predict plant phenological dynamics, crucial for Indigenous communities and conservation efforts, by integrating diverse data sources such as community science observations, observatory network data, remotely sensed observations, and herbarium specimens. With a growing urgency due to climate change, the research will focus on how shifts in phenology influence species’ distributions and developmental timings, which could affect ecosystem balance and favor non-native species. Combining these varied data sets with species distribution information, the project will shed light on the resilience of different plant species—rare versus common, native versus invasive—in altering climates. A co-production approach with Indigenous scholars will ensure research methodologies and outcomes are valuable to their communities. The Environmental Science Innovation and Inclusion Lab (ESIIL), with its resources and expertise in interdisciplinary collaboration and computing, will serve as an ideal hub for this initiative.

CAREER: Advancing a macrosystems framework for climate-phenology coupling through integrated research and education

PI: Kai Zhu
Sponsor: National Science Foundation, Directorate for Biological Sciences
Summary: The project presents a novel framework to examine the intricate links between climate and phenology, the study of the timing of biological events. By developing dynamic models and utilizing robust multi-scale data, including that from NEON, this research aims to unravel the complexity of climate-phenology interactions on a global scale, examine plant distributional shifts, and analyze the underexplored concept of teleconnections. Emphasis will be placed on enhancing ecological forecasting and creating reproducible workflows for data integration, while simultaneously advancing climate change and data science education through innovative teaching programs and social media outreach. The ultimate goal is to deliver societal benefits, including better pollen outbreak prediction, while training the next generation of educators in environmental science.

MRA: Macroecology of microorganisms: Scaling fungal biodiversity from soil cores to the North American continent

PI: Kai Zhu, co-PI: Kabir G. Peay
Sponsor: National Science Foundation, Directorate for Biological Sciences
Summary: This research aims to fundamentally advance our understanding of soil fungal biodiversity and the factors shaping it across North America by analyzing two large-scale DNA sequencing datasets, NEON and Dimensions of Biodiversity. The team focuses on discerning the environmental variables and plant-fungi interactions that influence soil fungal communities, employing a novel tripartite model with top-down, bottom-up, and integrated approaches—surpassing the limitations of prior studies that overlooked species interactions. This work is poised to scale microbial insights to broader ecosystems, enabling robust predictions about changes in microbial diversity on a continental level amidst environmental shifts. Findings suggest significant potential alterations in fungal communities, and the project’s innovative techniques promise to enhance predictions of fungal responses to future environmental challenges, thus illuminating fungal biogeography with implications for ecological forecasting.

Life-cycle analysis synthesized with ecosystems and risk focused on sustainable forestry

PI: Benjamin P. Goldstein, co-PI: Kai Zhu
Sponsor: University of Michigan, Biosciences Initiative
Summary: The project develops a new framework, named LASER (Life-cycle Analysis Synthesized with Ecology and Risk), to better understand the impact of using forests to capture carbon dioxide and combat climate change. This method evaluates the full life cycle of carbon in forests, and how it affects the environment and local communities. It uses real-world case studies from Brazil, Canada, and the United States, aiming to improve current forest management practices and carbon offset programs. Additionally, the project will create user-friendly tools for the public and experts to see and experiment with the effects of forest-based climate strategies, helping to inform better decisions for the Earth’s future and local populations.

Bayesian modeling of multi-source phenology to predict airborne allergens

PI: Kai Zhu, co-PI: Kerby Shedden
Sponsor: University of Michigan, Michigan Institute for Data Science
Summary: Our project is set out to enhance both the near-term and long-term forecasting of airborne allergens—such as pollen and fungal spores—intensified by climate-driven shifts in plant phenology. As seasonal allergies, exacerbated by earlier and prolonged allergen seasons, are becoming a more severe public health concern, we aim to leverage Bayesian hierarchical models to assimilate diverse and currently siloed data sources, including continuous aerial sampling and citizen phenology observations. This integration seeks to bolster predictive model performance for immediate decision-making and deepen mechanistic understanding for future projections. This project will yield a trifecta of impacts: guiding further data collection strategies, engaging the public via citizen science, and developing an open-source model that can be adapted by fellow researchers, thereby propelling progress in environmental data science.

Synthesizing experiments and observations of soil fungal community responses to climate change

PI: Kai Zhu, co-PIs: Donald R. Zak and Peter B. Reich
Sponsor: University of Michigan, Institute for Global Change Biology
Summary: The study seeks to address how soil fungal communities, vital for ecosystem functions, are affected by climate change by integrating experimental manipulations of climate conditions and continental-scale observational studies. Utilizing a two-step synthesis approach, the project will first generate environmental niche models for soil fungi from large-scale molecular observations to pinpoint climate variables that most influence fungi, predict their range shifts, and identify those most at risk. It will then compare these model results with responses from long-running climate change experiments to assess the validity of predictions about fungal community changes. This study merges extensive DNA-based datasets, including functional and taxonomic information, with data from the B4WarmED and TeRaCON experiments. The objective is to test hypotheses on fungal associations, both among fungi and with plants, and the degree to which biodiversity is constrained by multiple species. The goal is to synthesize experimental and observational predictions to understand microbial community vulnerability to global change.

Selected past projects

  • Forecasting pollen high and low: Near-term ecological forecasting with state-of-the-art machine learning algorithm
  • DISSERTATION RESEARCH: Forest climate requirements change through species life history
    • PI: James S. Clark, co-PI: Kai Zhu
    • Sponsor: National Science Foundation, Directorate for Biological Sciences