Leung, Brian; Cacho, Oscar Jose; Spring, D

Author(s)

Leung, Brian

Cacho, Oscar Jose

Spring, D

Publication Date

2010

Abstract

'Aim' At first detection, little information is typically known about an invader's characteristics, true arrival date or spatial extent. Yet, before management options such as control or eradication can be considered, we need to know where a nuisance species has already spread. This is particularly difficult because of stochastic processes. Here, we develop an approach that requires little a 'priori' information, yet accurately delimits the range of a biological invader. 'Location' We used a simulated landscape, subjected to stochasticity inherent in establishment and spread, to test novel theory for delimiting locally spreading populations. 'Methods' We distinguish three stages to identify the boundary of an invasion, which we term Approach, Decline, Delimit (ADD). Our ADD algorithm uses general characteristics of the invasion pattern, obtained during a search for occupied sites, in combination with sampling and probability theory to delimit the invasion. We compare ADD against four naïve delimitation strategies, for long and normal dispersal kernels. 'Results' Our results illustrate the potential difficulty in delimiting invasions. Naïve strategies, such as stopping when the invader is absent, typically failed to properly delimit the invasion. In contrast, ADD operated relatively efficiently, and was robust to habitat heterogeneity and knowledge of the true epicentre, but was sensitive to the sparseness of the invasion. For long-distance dispersal kernels, ADD had 80% accurate delimitations when 'c'. 5% or more of the cells were occupied within the invasion boundary; for normal dispersal kernels, ADD had 95% accurate delimitations when 'c'. 2.5% or more of the cells were occupied. 'Main conclusions' There is virtually no existing theory for delimiting invasions. ADD is efficient and accurate, even with unknown time of invasion, unknown dispersal kernels, stochastic establishment dynamics and spatial heterogeneity, except for very low invasion densities.

Citation

Diversity and Distributions, 16(3), p. 451-460

ISSN

1472-4642

1366-9516

Link

https://hdl.handle.net/1959.11/5600

Publisher

Blackwell Publishing Ltd

Title

Searching for non-indigenous species: rapidly delimiting the invasion boundary

Type of document

Journal Article

Entity Type

Publication

Author(s)	Leung, Brian Cacho, Oscar Jose Spring, D
Publication Date	2010
Abstract	'Aim' At first detection, little information is typically known about an invader's characteristics, true arrival date or spatial extent. Yet, before management options such as control or eradication can be considered, we need to know where a nuisance species has already spread. This is particularly difficult because of stochastic processes. Here, we develop an approach that requires little a 'priori' information, yet accurately delimits the range of a biological invader. 'Location' We used a simulated landscape, subjected to stochasticity inherent in establishment and spread, to test novel theory for delimiting locally spreading populations. 'Methods' We distinguish three stages to identify the boundary of an invasion, which we term Approach, Decline, Delimit (ADD). Our ADD algorithm uses general characteristics of the invasion pattern, obtained during a search for occupied sites, in combination with sampling and probability theory to delimit the invasion. We compare ADD against four naïve delimitation strategies, for long and normal dispersal kernels. 'Results' Our results illustrate the potential difficulty in delimiting invasions. Naïve strategies, such as stopping when the invader is absent, typically failed to properly delimit the invasion. In contrast, ADD operated relatively efficiently, and was robust to habitat heterogeneity and knowledge of the true epicentre, but was sensitive to the sparseness of the invasion. For long-distance dispersal kernels, ADD had 80% accurate delimitations when 'c'. 5% or more of the cells were occupied within the invasion boundary; for normal dispersal kernels, ADD had 95% accurate delimitations when 'c'. 2.5% or more of the cells were occupied. 'Main conclusions' There is virtually no existing theory for delimiting invasions. ADD is efficient and accurate, even with unknown time of invasion, unknown dispersal kernels, stochastic establishment dynamics and spatial heterogeneity, except for very low invasion densities.
Citation	Diversity and Distributions, 16(3), p. 451-460
ISSN	1472-4642 1366-9516
Link	https://hdl.handle.net/1959.11/5600
Publisher	Blackwell Publishing Ltd
Title	Searching for non-indigenous species: rapidly delimiting the invasion boundary
Type of document	Journal Article
Entity Type	Publication

Searching for non-indigenous species: rapidly delimiting the invasion boundary

Files: