Dataset Information

The effects of sampling bias and model complexity on the predictive performance of MaxEnt species distribution models.

ABSTRACT: Species distribution models (SDMs) trained on presence-only data are frequently used in ecological research and conservation planning. However, users of SDM software are faced with a variety of options, and it is not always obvious how selecting one option over another will affect model performance. Working with MaxEnt software and with tree fern presence data from New Zealand, we assessed whether (a) choosing to correct for geographical sampling bias and (b) using complex environmental response curves have strong effects on goodness of fit. SDMs were trained on tree fern data, obtained from an online biodiversity data portal, with two sources that differed in size and geographical sampling bias: a small, widely-distributed set of herbarium specimens and a large, spatially clustered set of ecological survey records. We attempted to correct for geographical sampling bias by incorporating sampling bias grids in the SDMs, created from all georeferenced vascular plants in the datasets, and explored model complexity issues by fitting a wide variety of environmental response curves (known as "feature types" in MaxEnt). In each case, goodness of fit was assessed by comparing predicted range maps with tree fern presences and absences using an independent national dataset to validate the SDMs. We found that correcting for geographical sampling bias led to major improvements in goodness of fit, but did not entirely resolve the problem: predictions made with clustered ecological data were inferior to those made with the herbarium dataset, even after sampling bias correction. We also found that the choice of feature type had negligible effects on predictive performance, indicating that simple feature types may be sufficient once sampling bias is accounted for. Our study emphasizes the importance of reducing geographical sampling bias, where possible, in datasets used to train SDMs, and the effectiveness and essentialness of sampling bias correction within MaxEnt.

SUBMITTER: Syfert MM

PROVIDER: S-EPMC3573023 | biostudies-literature | 2013

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

The effects of sampling bias and model complexity on the predictive performance of MaxEnt species distribution models.

Syfert Mindy M MM Smith Matthew J MJ Coomes David A DA

PloS one 20130214 2

Species distribution models (SDMs) trained on presence-only data are frequently used in ecological research and conservation planning. However, users of SDM software are faced with a variety of options, and it is not always obvious how selecting one option over another will affect model performance. Working with MaxEnt software and with tree fern presence data from New Zealand, we assessed whether (a) choosing to correct for geographical sampling bias and (b) using complex environmental response ...[more]

PMID: 23457462

Similar Datasets

Project description:Predicting species' potential geographical range by species distribution models (SDMs) is central to understand their ecological requirements. However, the effects of using different modeling techniques need further investigation. In order to improve the prediction effect, we need to assess the predictive performance and stability of different SDMs.We collected the distribution data of five common tree species (Pinus massoniana, Betula platyphylla, Quercus wutaishanica, Quercus mongolica and Quercus variabilis) and simulated their potential distribution area using 13 environmental variables and six widely used SDMs: BIOCLIM, DOMAIN, MAHAL, RF, MAXENT, and SVM. Each model run was repeated 100 times (trials). We compared the predictive performance by testing the consistency between observations and simulated distributions and assessed the stability by the standard deviation, coefficient of variation, and the 99% confidence interval of Kappa and AUC values.The mean values of AUC and Kappa from MAHAL, RF, MAXENT, and SVM trials were similar and significantly higher than those from BIOCLIM and DOMAIN trials (p<0.05), while the associated standard deviations and coefficients of variation were larger for BIOCLIM and DOMAIN trials (p<0.05), and the 99% confidence intervals for AUC and Kappa values were narrower for MAHAL, RF, MAXENT, and SVM. Compared to BIOCLIM and DOMAIN, other SDMs (MAHAL, RF, MAXENT, and SVM) had higher prediction accuracy, smaller confidence intervals, and were more stable and less affected by the random variable (randomly selected pseudo-absence points).According to the prediction performance and stability of SDMs, we can divide these six SDMs into two categories: a high performance and stability group including MAHAL, RF, MAXENT, and SVM, and a low performance and stability group consisting of BIOCLIM, and DOMAIN. We highlight that choosing appropriate SDMs to address a specific problem is an important part of the modeling process.

Project description:Loss of habitat, specifically deforestation, is a major driver of biodiversity loss. Species-area relationship (SAR) models traditionally have been used for estimating species richness, species loss as a function of habitat loss, and extrapolation of richness for given areas. Sampling-species relationships (SSRs) are interrelated yet separate drivers for species richness estimates. Traditionally, however, SAR and SSR models have been used independently and not incorporated into a single approach. We developed and compared predictive models that incorporate sampling effort species-area relationships (SESARS) along the entire Atlantic Forest of South America, and then applied the best-fit model to estimate richness in forest remnants of Interior Atlantic Forest of eastern Paraguay. This framework was applied to non-volant small mammal assemblages that reflect different tolerances to forest loss and fragmentation. In order to account for differences in functionality we estimated small mammal richness of 1) the entire non-volant small mammal assemblage, including introduced species; 2) the native species forest assemblage; and 3) the forest-specialist assemblage, with the latter two assemblages being subsets of the entire assemblage. Finally, we geospatially modeled species richness for each of the three assemblages throughout eastern Paraguay to identify remnants with high species richness. We found that multiple regression power-law interaction-term models that only included area and the interactions of area and sampling as predictors, worked best for predicting species richness for the entire assemblage and the native species forest assemblage, while several traditional SAR models (logistic, power, exponential, and ratio) best described forest-specialist richness. Species richness was significantly different between assemblages. We identified obvious remnants with high species richness in eastern Paraguay, and these remnants often were geographically isolated. We also found relatively high predicted species richness (in relation to the entire range of predicted richness values) in several geographically-isolated, medium-size forest remnants that likely have not been considered as possible priority areas for conservation. These findings highlight the importance of using an empirical dataset, created using sources representing diverse sampling efforts, to develop robust predictive models. This approach is particularly important in geographic locations where field sampling is limited yet the geographic area is experiencing rapid and dramatic land cover changes. When combined, area and sampling are powerful modeling predictors for questions of biogeography, ecology, and conservation, especially when addressing habitat loss and fragmentation.

Dataset Information

The effects of sampling bias and model complexity on the predictive performance of MaxEnt species distribution models.

Publications

The effects of sampling bias and model complexity on the predictive performance of MaxEnt species distribution models.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets