Abstract
Aim
Estimating environmental suitability from species distribution data is crucial in defining spatial conservation measures. To this end, species distribution models (SDMs) are commonly applied, but seldom validated by completely independent data. Here, we use data on individual tracks derived from electronic tags as an alternative means of validating SDM outputs.
Location
West coast of Scotland, NE Atlantic.
Methods
We used a binomial generalized additive model (GAM) to predict the environmental suitability for flapper skate (Dipturus cf. intermedia) in Scottish waters. The GAM modelled relative habitat usage as a function of environmental variables using presence–absence data obtained from scientific trawl surveys. Additional data obtained from electronic tags attached to six individual flapper skates were used to estimate individual tracks using a tidal-based geolocation model. Concordance between individual tracks and GAM-predicted maps of relative habitat usage (RHU) was tested by comparing predicted RHU between estimated tracks and randomly generated tracks.
Results
Environmental suitability for the flapper skate was driven by depth and distance from the coast in the SDM. We found high spatial concordance between the estimated tracks of the six tagged individuals and regions of high RHU predicted by the SDM.
Main conclusions
Integrating outputs from an independent data source allowed us to validate predictions from a species distribution model (SDM). The integration of individual- and population-level data sources increases confidence in the outputs being used to define spatial conservation measures. The information on flapper skate distribution provided by this study provides a useful framework for considering spatial conservation measures for this species.
Estimating environmental suitability from species distribution data is crucial in defining spatial conservation measures. To this end, species distribution models (SDMs) are commonly applied, but seldom validated by completely independent data. Here, we use data on individual tracks derived from electronic tags as an alternative means of validating SDM outputs.
Location
West coast of Scotland, NE Atlantic.
Methods
We used a binomial generalized additive model (GAM) to predict the environmental suitability for flapper skate (Dipturus cf. intermedia) in Scottish waters. The GAM modelled relative habitat usage as a function of environmental variables using presence–absence data obtained from scientific trawl surveys. Additional data obtained from electronic tags attached to six individual flapper skates were used to estimate individual tracks using a tidal-based geolocation model. Concordance between individual tracks and GAM-predicted maps of relative habitat usage (RHU) was tested by comparing predicted RHU between estimated tracks and randomly generated tracks.
Results
Environmental suitability for the flapper skate was driven by depth and distance from the coast in the SDM. We found high spatial concordance between the estimated tracks of the six tagged individuals and regions of high RHU predicted by the SDM.
Main conclusions
Integrating outputs from an independent data source allowed us to validate predictions from a species distribution model (SDM). The integration of individual- and population-level data sources increases confidence in the outputs being used to define spatial conservation measures. The information on flapper skate distribution provided by this study provides a useful framework for considering spatial conservation measures for this species.
| Original language | English |
|---|---|
| Pages (from-to) | 682-693 |
| Number of pages | 12 |
| Journal | Diversity and Distributions |
| Volume | 22 |
| Issue number | 6 |
| Early online date | 23 Mar 2016 |
| DOIs | |
| Publication status | Published - Jun 2016 |
Bibliographical note
The authors would like to thank the College of Life Sciences of Aberdeen University and Marine Scotland Science which funded CP's PhD project. Skate tagging experiments were undertaken as part of Scottish Government project SP004. We thank Ian Burrett for help in catching the fish and the other fishermen and anglers who returned tags. We thank José Manuel Gonzalez-Irusta for extracting and making available the environmental layers used as environmental covariates in the environmental suitability modelling procedure. We also thank Jason Matthiopoulos for insightful suggestions on habitat utilization metrics as well as Stephen C.F. Palmer, and three anonymous reviewers for useful suggestions to improve the clarity and quality of the manuscript.Keywords
- data integration
- Dipturus cf. intermedia
- generalized additive model
- individual movment
- model validation
- species distribution model
- tidal geolocation model
