Transferability of species distribution models: The case of Phytophthora cinnamomi in Southwest Spain and Southwest Australia

Duque-Lazo, Joaquín; Gils, Hein van; Groen, T.A.; Navarro-Cerrillo, R.M.

doi:10.1016/j.ecolmodel.2015.09.019

Cited by 134 publications

(105 citation statements)

References 52 publications

(88 reference statements)

Supporting

Mentioning

101

Contrasting

Order By: Relevance

“…MaxEnt is considered to be a robust and transferable OCC (Duque‐Lazo et al. ) that yield high performances compared to other OCC algorithms in remote sensing applications (Stenzel et al. ).…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

How canopy shadow affects invasive plant species classification in high spatial resolution remote sensing

Lopatin

Dološ

Kattenborn

2019

Remote Sens Ecol Conserv

View full text Add to dashboard Cite

Plant invasions can result in serious threats for biodiversity and ecosystem functioning. Reliable maps at very‐high spatial resolution are needed to assess invasions dynamics. Field sampling approaches could be replaced by unmanned aerial vehicles (UAVs) to derive such maps. However, pixel‐based species classification at high spatial resolution is highly affected by within‐canopy variation caused by shadows. Here, we studied the effect of shadows on mapping the occurrence of invasive species using UAV‐based data. MaxEnt one‐class classifications were applied to map Acacia dealbata, Ulex europaeus and Pinus radiata in central‐south Chile using combinations of UAV‐based spectral (RGB and hyperspectral), 2D textural and 3D structural variables including and excluding shaded canopy pixels during model calibration. The model accuracies in terms of area under the curve (AUC), Cohen's Kappa, sensitivity (true positive rate) and specificity (true negative rate) were examined in sunlit and shaded canopies separately. Bootstrapping was used for validation and to assess statistical differences between models. Our results show that shadows significantly affect the accuracies obtained with all types of variables. The predictions in shaded areas were generally inaccurate, leading to misclassification rates between 65% and 100% even when shadows were included during model calibration. The exclusion of shaded areas from model calibrations increased the predictive accuracies (especially in terms of sensitivity), decreasing false positives. Spectral and 2D textural information showed generally higher performances and improvements when excluding shadows from the analysis. Shadows significantly affected the model results obtained with any of the variables used, hence the exclusion of shadows is recommended prior to model calibration. This relatively easy pre‐processing step enhances models for classifying species occurrences using high‐resolution spectral imagery and derived products. Finally, a shadow simulation showed differences in the ideal acquisition window for each species, which is important to plan revisit campaigns.

show abstract

Section: Discussionmentioning

confidence: 99%

“…) and enhances model transferability (Duque‐Lazo et al. ). First, MaxEnt classifications using all available variables were performed using a 10‐fold cross‐validation.…”

Section: Methodsmentioning

confidence: 99%

How canopy shadow affects invasive plant species classification in high spatial resolution remote sensing

Lopatin

Dološ

Kattenborn

2019

Remote Sens Ecol Conserv

View full text Add to dashboard Cite

show abstract

“…Analysis of citizen science datasets usually requires modeling algorithms for presence‐only data because absences are not recorded. Though a variety of options are available (e.g., boosted regression trees, environmental niche factor analysis, random forest; García‐Callejas and Araújo , Shabani et al ), we used one of the most robust and widely used methods, maximum entropy (MaxEnt) models, to correlate spatial variables with roadkill reports (Phillips et al , Elith et al , Duque‐Lazo et al ). MaxEnt is a machine‐learning approach that optimizes species‐environment associations using multiple function types including quadratic and product functions (Merow et al ) and is thus especially useful in mapping ecological phenomena that may have complex, non‐linear correlations, such as roadkill patterns.…”

Section: Methodsmentioning

confidence: 99%

“…MaxEnt is a machine‐learning approach that optimizes species‐environment associations using multiple function types including quadratic and product functions (Merow et al ) and is thus especially useful in mapping ecological phenomena that may have complex, non‐linear correlations, such as roadkill patterns. It has been shown to outperform alternative methods especially when there is a sufficient sample size and wide species distribution (Kasampalis et al , Duque‐Lazo et al ), and has been used in analyzing citizen science data (Crall et al , Fournier et al ) and mapping roadkill hotspots (Ha and Shilling , Garrote et al ). Because absence data are unavailable, MaxEnt compares presence points with a sample of points, known as background or pseudo‐absence points, from the study area of interest.…”

Section: Methodsmentioning

confidence: 99%

Informing snake roadkill mitigation strategies in Taiwan using citizen science

2018

View full text Add to dashboard Cite

Despite their prevalence in roadkill in East Asia, there has been little research on snake mortality along roads, even though the region's fauna contains some of the highest proportions of threatened or data‐deficient reptiles. We analyzed citizen‐collected roadkill data from Taiwan, comprising >11,000 records of snake roadkill from 2006–2017. We used maximum entropy environmental niche modeling to predict roadkill sighting patterns across snake species differing in habitat use, foraging behavior, and taxonomic group. Roadkill sightings were highest in low to mid‐elevation (i.e., 0–2,000 m) forests and strips of farmland or shrubland that cut through forests; these areas likely support high snake abundances or dispersal activity. Sightings were lowest in urban areas and at high elevations (i.e., >2,000 m), likely because of unfavorable habitat conditions. Road density had little influence on roadkill sightings; areas with dense roads may be of lower habitat quality and contain fewer snakes. Roadkill sighting patterns differed among snake species with different habitat use and behavior. Natural history and landscape factors should be considered in roadkill mitigation design to reduce snake roadkill effectively. We recommend the use of similar citizen‐science projects elsewhere to supplement conservation planning. © 2018 The Wildlife Society.

show abstract

“…The transferability of predictive models can be dataset, question and algorithm‐specific (Duque‐Lazo et al., ; Elith et al., ). It is often thought that model simplicity and parsimony should be preferred over complexity.…”

Section: Choice Of Modelling Algorithms Can Affect Model Transferabilitymentioning

confidence: 99%

Transferring biodiversity models for conservation: Opportunities and challenges

et al. 2018

View full text Add to dashboard Cite

After decades of extensive surveying, knowledge of the global distribution of species still remains inadequate for many purposes. In the short to medium term, such knowledge is unlikely to improve greatly given the often prohibitive costs of surveying and the typically limited resources available. By forecasting biodiversity patterns in time and space, predictive models can help fill critical knowledge gaps and prioritise research to support better conservation and management. The ability of a model to predict biodiversity metrics in novel environments is termed “transferability,” and models with high transferability will be the most useful in this context. Despite their potentially broad utility, little guidance exists on what confers high transferability to biodiversity models. We synthesise recent advances in biodiversity model transfers to facilitate increased understanding of what underpins successful model transferability, demonstrating that a consistent approach has so far been lacking but is essential for achieving high levels of repeatability, transparency and accountability of model transfers. We provide a set of guidelines to support efficient learning and the improvement of model transferability.

show abstract

Transferability of species distribution models: The case of Phytophthora cinnamomi in Southwest Spain and Southwest Australia

Cited by 134 publications

References 52 publications

How canopy shadow affects invasive plant species classification in high spatial resolution remote sensing

How canopy shadow affects invasive plant species classification in high spatial resolution remote sensing

Informing snake roadkill mitigation strategies in Taiwan using citizen science

Transferring biodiversity models for conservation: Opportunities and challenges

Contact Info

Product

Resources

About