Predicting the Spread of Purple Loosestrife (Lythrum salicaria) in the Prairies

Authors

  • Cory J. Lindgren Department of Environment and Geography, University of Manitoba, Winnipeg, Manitoba R3T 2N2
  • David Walker Department of Environment and Geography, University of Manitoba, Winnipeg, Manitoba R3T 2N2

DOI:

https://doi.org/10.22621/cfn.v126i4.1376

Keywords:

Purple Loosestrife, Lythrum salicaria, invasive weed, ecological niche modelling, genetic algorithm for rule-set prediction, GARP, Manitoba, Saskatchewan, Alberta

Abstract

Purple Loosestrife (Lythrum salicaria) is an invasive plant introduced into North America in the early 1800s. It has since spread into the prairie provinces of Canada (Manitoba, Saskatchewan, and Alberta). It invades wetland habitats, marshes, riparian areas, and natural areas, and it outcompetes native wetland vegetation. In this study we modelled the potential distribution of Purple Loosestrife in the Prairies, explored which suites of predictive variables produced the best ecological niche models, and explored two different approaches to the partitioning of data in evaluating models. We used a number of performance measures and expert evaluation to select our best models. The best model was developed using a suite of climate variables and growing degree-days as the predictive variables and by partitioning testing and training data using stratified random sampling. The model indicated that Purple Loosestrife has not yet reached its full potential distribution in the Prairies. The modelling techniques presented in this paper may be used to predict the potential distribution of other emerging invasive plants, and the results can be used to optimize early detection and surveillance strategies for Purple Loosestrife in areas of the Prairies.

Downloads

Published

2013-04-22

Issue

Section

Articles