PS 40-209
Mapping the potential extent of native plant communities in Minnesota: Linking land use with ecosystem function

Tuesday, August 11, 2015
Exhibit Hall, Baltimore Convention Center
Seth Fore, Earth System Science and Policy, University of North Dakota, Grand Forks, ND
Michael J. Hill, Earth System Science and Policy, University of North Dakota, Grand Forks, ND

Global changes in land use and land cover (LULC) patterns have dramatically altered the spatial distribution of native plant communities globally. Moreover, in many historically undeveloped parts of the world LULC change is ongoing or actively accelerating primarily at the expense of native vegetation. The importance of changes in LULC relate to their impact on the ecosystem functions and subsequent services (EFS) that underpin many dimensions of human wellbeing. LULC effects EFS by altering the composition and abundance of resident species, or more specifically, the makeup of species traits. An approach using plant functional traits has developed and proven effective in linking land use decisions to changes in EFS.  

However, for areas with a long-term legacy of LULC change like Minnesota (MN) and much of US, an impediment to estimating changes EFS is obtaining a natural vegetation reference or baseline to compare current vegetation against. The primary objectives for our research were then twofold. First, create a potential natural vegetation map for MN with sufficiently fine spatial scale and floristic resolution. Second, attribute reference and current vegetation with key plant functional trait values obtained from the TRY Trait Database and assess the difference.


Using a statewide collection of native vegetation relevés, we used Random Forests to model potential natural vegetation at four levels of floristic detail following the MN Native Plant Community Classification System. Depending on vegetation class and level of floristic detail, model accuracy varied between 65-95% with kappa values between 45-95%. In conjunction with uncertainty maps, non-metric multi-dimensional scaling was used to explore likely tradeoff’s between similar vegetation classes. Apart from providing a baseline, fine-scale representation of potential natural vegetation can be used for a variety of conservation, restoration and ecosystem management purposes. 

For this study we’ve selected traits with well documented links to a variety of EFS including: specific leaf area, leaf dry matter content, leaf nitrogen content, vegetative height, and others. Briefly, these traits are linked to biomass production, carbon sequestration, nutrient turnover, and climate/water regulation. Using the trait data we’ve obtained to this point, preliminary results suggest meaningful differences in community-weighted mean trait values between reference and current vegetation states. While TRY trait data requests have obtained the proper approval, delivery is currently pending for a number of species. After these data are obtained, the final analysis will conducted and results formalized.