Mapping black ash dominated stands using geospatial and forest inventory data in northern Minnesota, USA

TitleMapping black ash dominated stands using geospatial and forest inventory data in northern Minnesota, USA
Publication TypeJournal Article
Year of Publication2019
AuthorsEngelstad, Peder S., Falkowski Michael J., D’Amato Anthony W., Slesak Robert A., Palik Brian J., Domke Grant M., and Russell Matthew B.
JournalCanadian Journal of Forest Research
Pagination892 - 902
Date PublishedJun-03-2019
ISSN0045-5067
Keywordsblack ash, compound topographic index (CTI), emerald ash borer,"forest inventory, Remote sensing
Abstract

Emerald ash borer (EAB; Agrilus planipennis Fairmaire, 1888) has been a persistent disturbance for ash forests in the United States since 2002. Of particular concern is the impact that EAB will have on the ecosystem functioning of wetlands dominated by black ash (Fraxinus nigra Marsh.). In preparation, forest managers need reliable and complete maps of black ash dominated stands. Traditionally, forest survey data from the United States Forest Inventory and Analysis (FIA) Program have provided rigorous measures of tree species at large spatial extents but are limited when providing estimates for smaller management units (e.g., stands). Fortunately, geospatial data can extend forest survey information by generating predictions of forest attributes at scales finer than those of the FIA sampling grid. In this study, geospatial data were integrated with FIA data in a randomForest model to estimate and map black ash dominated stands in northern Minnesota in the United States. The model produced low error rates (overall error = 14.5%; area under the curve (AUC) = 0.92) and was strongly informed by predictors from soil saturation and phenology. These results improve upon FIA-based spatial estimates at national extents by providing forest managers with accurate, fine-scale maps (30 m spatial resolution) of black ash stand dominance that could ultimately support landscape-level EAB risk and vulnerability assessments.

URLhttp://www.nrcresearchpress.com/doi/10.1139/cjfr-2018-0481
DOI10.1139/cjfr-2018-0481
Short TitleCan. J. For. Res.