The desire to consider additional ecological information in management planning has become a pressing concern in the field of forest ecology and management. While intensively managed forest stands provide ecological benefits, these can be different from the services and values supported by native ecosystems and plant communities. To better understand the implications of management for biological diversity, ecosystem services, timber production and other interests, an ecological classification methodology matched with existing forest inventory and management operations is proposed and developed. This methodology makes use of nearly 17,000 native plant community (NPC) observations provided by the Minnesota Department of Natural Resources (MNDNR) and others. These observations cover the period from 1964 – 2015, and coincide with stands monitored by the MNDNR Division of Forestry. The proposed imputation model (Chapter 1) represents an improvement over randomForest based methods in terms of accuracy, coverage, and the ability to consider complex categorical variables with essentially unlimited levels of detail. Extension of the methodology to include United States Department of Agriculture (USDA) Forest Inventory and Analysis (FIA) plot observations and additional predictive characteristics further improves classification results (Chapter 2). The net predictive capability is sufficient to produce estimates of the areal extent of major forested NPCs occurring in Minnesota. These estimates are derived from a process utilizing the spatial overlay of FIA plots with MNDNR stands having NPC observations. These “observed” FIA plots serve as training data to classify the full set of FIA plots observed in Minnesota. Finally, FIA data augmented with imputed NPC classifications are used to assess relationships between NPC classifications and growth and yield characteristics of the forests in each community (Chapter 3). Results indicate that NPC classification often corresponds to meaningful distinctions between different growth patterns and eventual yield of forested stands. Imputation can provide us with timely and accurate knowledge of NPC distribution, abundance, successional state, demographic, and economic relationships. This enhanced understanding of landscape-scale ecological conditions can, in turn, lead to better informed management decisions based on the extrapolation of observed ecological conditions and growth parameters to very similar, nearby management units.
University of Minnesota Ph.D. dissertation. November 2016. Major: Natural Resources Science and Management. Advisor: Alan Ek. 1 computer file (PDF); v, 88 pages.
Imputation of ecological detail using associated forest inventory, plant community and physiographic data.
Retrieved from the University of Minnesota Digital Conservancy,
Content distributed via the University of Minnesota's Digital Conservancy may be subject to additional license and use restrictions applied by the depositor.