Dataset Information

Evaluation metrics and validation of presence-only species distribution models based on distributional maps with varying coverage.

ABSTRACT: We examine how different datasets, including georeferenced hardcopy maps of different extents and georeferenced herbarium specimens (spanning the range from 100 to 85,000 km²) influence ecological niche modeling. We check 13 of the available environmental niche modeling algorithms, using 30 metrics to score their validity and evaluate which are useful for the selection of the best model. The validation is made using an independent dataset comprised of presences and absences collected in a range-wide field survey of Carpathian endemic plant Leucanthemum rotundifolium (Compositae). Our analysis of models' predictive performances indicates that almost all datasets may be used for the construction of a species distributional range. Both very local and very general datasets can produce useful predictions, which may be more detailed than the original ranges. Results also highlight the possibility of using the data from manually georeferenced archival sources in reconstructions aimed at establishing species' ecological niches. We discuss possible applications of those data and associated problems. For the evaluation of models, we suggest employing AUC, MAE, and Bias. We show an example of how AUC and MAE may be combined to select the model with the best performance.

SUBMITTER: Konowalik K

PROVIDER: S-EPMC7811024 | biostudies-literature | 2021 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Evaluation metrics and validation of presence-only species distribution models based on distributional maps with varying coverage.

Konowalik Kamil K Nosol Agata A

Scientific reports 20210115 1

We examine how different datasets, including georeferenced hardcopy maps of different extents and georeferenced herbarium specimens (spanning the range from 100 to 85,000 km<sup>2</sup>) influence ecological niche modeling. We check 13 of the available environmental niche modeling algorithms, using 30 metrics to score their validity and evaluate which are useful for the selection of the best model. The validation is made using an independent dataset comprised of presences and absences collected ...[more]

PMID: 33452285

Similar Datasets

Project description:AimAvailability of uniformly collected presence, absence, and abundance data remains a key challenge in species distribution modeling (SDM). For invasive species, abundance and impacts are highly variable across landscapes, and quality occurrence and abundance data are critical for predicting locations at high risk for invasion and impacts, respectively. We leverage a large aquatic vegetation dataset comprising point-level survey data that includes information on the invasive plant Myriophyllum spicatum (Eurasian watermilfoil) to: (a) develop SDMs to predict invasion and impact from environmental variables based on presence-absence, presence-only, and abundance data, and (b) compare evaluation metrics based on functional and discrimination accuracy for presence-absence and presence-only SDMs.LocationMinnesota, USA.MethodsEurasian watermilfoil presence-absence and abundance information were gathered from 468 surveyed lakes, and 801 unsurveyed lakes were leveraged as pseudoabsences for presence-only models. A Random Forest algorithm was used to model the distribution and abundance of Eurasian watermilfoil as a function of lake-specific predictors, both with and without a spatial autocovariate. Occurrence-based SDMs were evaluated using conventional discrimination accuracy metrics and functional accuracy metrics assessing correlation between predicted suitability and observed abundance.ResultsWater temperature degree days and maximum lake depth were two leading predictors influencing both invasion risk and abundance, but they were relatively less important for predicting abundance than other water quality measures. Road density was a strong predictor of Eurasian watermilfoil invasion risk but not abundance. Model evaluations highlighted significant differences: Presence-absence models had high functional accuracy despite low discrimination accuracy, whereas presence-only models showed the opposite pattern.Main conclusionComplementing presence-absence data with abundance information offers a richer understanding of invasive Eurasian watermilfoil's ecological niche and enables evaluation of the model's functional accuracy. Conventional discrimination accuracy measures were misleading when models were developed using pseudoabsences. We thus caution against the overuse of presence-only models and suggest directing more effort toward systematic monitoring programs that yield high-quality data.

Dataset Information

Evaluation metrics and validation of presence-only species distribution models based on distributional maps with varying coverage.

Publications

Evaluation metrics and validation of presence-only species distribution models based on distributional maps with varying coverage.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets