https://dx.doi.org/10.1007/s10980-015-0175-7">
 

A comparison of landscape fragmentation analysis programs for identifying possible invasive plant species locations in forest edge

Abstract

Context: When predicting locations of invasive plant species, mapping habitat fragmentation can be an important part of the prediction process. There are many different fragmentation mapping programs, each computing a unique set of fragmentation metrics that can be used in modeling probabilities of invasive species presence.

Objectives: In this study, we compare the results from five freely available fragmentation programs: FRAGSTATS; the Landscape Fragmentation Tool; Shape Metrics; Patch Analyst; and PolyFrag. We compare these programs quantitatively on their ability to predict invasive plant presence and qualitatively for ease of use.

Methods: The programs were compared using invasive plant inventories completed by The Nature Conservancy on parcels within the Coastal Watershed in New Hampshire, USA. Known locations of invasive plants, pseudo-absence locations, and metrics derived from each of the fragmentation programs were used to create maps of predicted presence for the parcels. The maps were compared and assessed for accuracy.

Results: FRAGSTATS and PolyFrag created prediction maps with the highest accuracies and were relatively easy to use. The other programs had lower accuracies or were more difficult to implement. Both FRAGSTATS and PolyFrag compute similar fragmentation metrics and the models found similar metrics significant in predicting presence. Both programs predicted that invasive plants were less likely to be found in deciduous forests than in either mixed or coniferous forests.

Conclusions: At the parcel level, some fragmentation programs result in metrics with more predictive power. Based on this analysis, we recommend FRAGSTATS for use with raster datasets and PolyFrag for vector datasets.

Publication Date

2-20-2015

Journal Title

Landscape Ecology

Publisher

Springer

Digital Object Identifier (DOI)

https://dx.doi.org/10.1007/s10980-015-0175-7

Scientific Contribution Number

2592

Document Type

Article

Rights

© Springer Science+Business Media Dordrecht 2015

Share

COinS