Skip to main content

Spatial population dynamics of invasive non-native species (re)invasions at low density

Final Report Summary - DEPENSATION (Spatial population dynamics of invasive non-native species (re)invasions at low density)

DEPENSATION aimed to deliver both a better understanding of the underpinning processes of invasive non-native species (INNS) spread, and improved management. The project focused on the spatial population dynamics of species at low density in a spatial heterogeneous habitat, using the non-native invasive species American mink as a study system. The key ecological concepts that I have being investigating are: the scope of compensatory processes (those causing an improvement in fitness of individuals at low density due to e.g. when the amount of resources per capita is greater), depensatory processes (those causing a decrease in fitness of individuals at low density due e.g. to difficulty in locating a mate), dispersal (the movement of individuals from site of birth to the site of reproduction) and multi-species interactions in a heterogeneous landscape (e.g. prey availability and productivity). Understanding of key ecological processes associated with the spread of INNS is fundamental for management geared toward halting or reversing their advance. Concepts such as dispersal, species interactions, and the role of compensatory and depensatory processes define the demographic spread of invasive species, but have been crucially overlooked in applied management scenarios.

The research objectives of the project were thus to:
RO1) Determine the demographic and dispersal dynamics of culled American mink populations in relation to the rate of recolonisation at low (post-control) density and in a spatially heterogeneous prey landscape.
RO2) Summarise and integrate this developing knowledge into a range of demographic Bayesian“state-space models” representing both the “unseen” mink spatial dynamics and the “observation process” (using presence absence data, age structure, mating status, dispersal pattern, and volunteer motivation and effectiveness) gathered as part of the conservation project. Use these models to test hypotheses about the roles of compensation and depensation in dispersal in culled mink population dynamics.
RO3) Apply this understanding and update empirical support using the “active adaptive management” approach to reduce uncertainty about the optimal range expansion strategies in relation to the wishes.

In relation to this research, the main scientific results are:

Result 1) By means of literature research I have demonstrated a positive relation between mink density and prey availability through the facilitation of mink carrying capacity (maximum density a population can sustain) by American crayfish availability (Melero et al. 2013) (RO1).

Result 2) In relation to RO2, I tested the prediction that mate finding failure (Allee effect component) increases with the decrease of density impacting the population growth rate (demographic Allee effect). To test the hypothesis, I analysed the fecundity of female mink by means of scar staining that provide counts of the number of implanted foetus per female and study its density-dependence relationship. Unexpectedly, I did not find any reduction of the probability of mating or of the reproductive output (litter size) per female with the reduction of the density. Instead, there was some compensation, with females’ probability of conceiving and litter size slightly increasing with the decrease of the density of females but not of males. In addition, I found senescence in reproduction, with younger females giving up to 4 more pups than older ones. As my research has always followed an active adaptive approach with science guiding management actions and this updating the scientific research (RO3), I studied the synergetic effect of senescence and compensation in reproduction in relation to the control program. By doing so I could observe that the ongoing culling of mink has provoked a reduction of the age of the mink populations, with all adults having long been removed and only juveniles survive. After six years of control, 80% of the mink population was younger than 2 years old. This implied an increase in their reproductive output that in synergy with the compensation effect due to the density reduction predicts an increase in the population growth rate. Crucially, this increase has not overcome the efforts of the volunteers but advocates for a continued longer-term control (Melero, Robinson and Lambin 2014; under review).

Result 3) The reconstruction of the genealogy of culled mink used to estimate individuals’ origin, settlement and dispersal patterns. I found that mink population is more connected than expected all over the study area (20,000km2) with mink dispersing 50-60 km on average and a maximum of 100 km. This demonstrates the ability of mink to reinvade previously mink-free areas and the need of a large-scale control program. This information was used to predict mink settlement patterns and reinvasion and compensation ability through dispersal in relation to mink density, area productivity (mink produced in previous years) and habitat suitability (RO1) (Melero, Cornulier and Lambin; in preparation).

Result 4) In addition, I have also tested the reliability of the kinship assignment using microsatellites and maximum likelihood methods (method followed by most software, e.g. COLONY) on a wild population with partial sampling of individuals and genome. I found that there only direct maternal kinship relationships can be considered as reliable, followed by full siblings and half siblings. This is of high importance since currently spatial dynamics are often inferred trough kinship relationships without any evaluation of their reliability. I thus proposed a methodology to evaluate the reliability of the relationships by means of simulation of the studied system by means of simulated populations with known genetic structure to which I added different levels of spatial structure, overlapping generations and migration. My results show that the reliability of kinship assignment obtained with maximum likelihood methods in is not reliable when population data is incomplete and that uncertainty is higher when one or more relatives are inferred genotypes not included in the sample. The results are being use to correct our own kinship results described above but, in addition, my methodology also provides a new technique for kinship results evaluation (RO1) (Melero, Oliver and Lambin; in preparation).

Result 5) I have analysed mink dispersal ability in relation to density and the connectivity between individuals, and discovered that 94% of individuals leave their natal area (emigrate) irrespective of density in the natal area. However, their settlement in new areas (immigration) is density dependent, especially for males. This work thus provides important information the factors attracting reinvading mink and thus, target areas to control (Melero, Cornulier and Lambin; in preparation).

Result 6) Finally, as part of my project, I have also being involved been in the MinkApp project, a partnership between ecology, social and computer scientists aiming to understanding and influencing the behaviour of volunteers in nature through Natural Language Generation (full information in http://www.dotrural.ac.uk/minkapp/). As a result we have managed to increase the frequency of reporting of the volunteers involved in DEPENSATION (Tintarev, Melero et al, 2012; Arts et al, 2013; Webster et al, 2014; Arts, Melero et al, in prep.).

All our results have a direct impact on applied research since they have feedback into the management organization that updates the control program accordingly to the scientific information provided by our research program.


Currently, I have secured a prestigious fellowship at the Autonomous University of Barcelona. For further information on the project please contact Dr. Yolanda Melero at her current addresses: y.melero@creaf.uab.es or melero@ub.edu and/or Professor Xavier Lambin at x.lambin@abdn.ac.uk