Efficient Modelling of Presence-Only Species Data via Local Background Sampling

Daniel, Jeffrey; Horrocks, Julie; Umphrey, Gary J.

doi:10.1007/s13253-019-00380-4

Cited by 5 publications

(5 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[51]). However, for Garcia et al [8], the habitats with a high distribution potential for the genus Cinchona were found to a greater extent in the regions of Junín, Huánuco and San Martín, with 36.3% (15,953.46 km 2 ), 47.60% (17,718.96 km 2 ) and 33.0% (16,813.68 km 2 ) of their respective territories. A plausible explanation for these differences is that they did not integrate edaphic variables as we did in the present work.…”

Section: Discussionmentioning

confidence: 89%

“…They are useful tools to address and optimize species management [14] as they rely heavily on bioclimatic variables to predict habitat suitability [15] and use records of the presence of the individuals of the species being modeled. These models are often represented as a process of a spatial point whose intensity is a function of environmental covariates [16,17]. The reliability of these models is obtained by performing the validation stage which consists of contrasting the presence and absence of individuals, based on records.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Current and Future Spatial Distribution of the Genus Cinchona in Peru: Opportunities for Conservation in the Face of Climate Change

Vergara,

Cieza-Tarrillo,

Ocaña

et al. 2023

Sustainability

View full text Add to dashboard Cite

The genus Cinchona belongs to the Rubiaceae family and comprises native Peruvian tree species distributed in tropical areas. It is currently endangered due to human disturbance and overexploitation for medicinal, forestry and food uses. To date, the current and future distribution of Cinchona spp. under the climate change scenario is unknown. Here, we modeled the present and future spatial distribution of the genus Cinchona using bioclimatic, edaphic and topographic variables using the maximum entropy algorithm (MaxEnt). The results indicate that 8.08% (103,547.89 km2) and 6.02% (77,163.81 km2) of the surface of Peru possesses areas with high and moderate distribution probabilities, respectively, to host the genus Cinchona, distributed mainly in the departments of Cusco, Amazonas, San Martín and Cajamarca. Furthermore, according to future climate scenarios, the areas of high suitability will increase their extension for the years 2050 and 2070 by 3.65% and 3.9%, respectively. Since Peru seeks to promote the forest sector to be the other force for its development, this study can be considered as a basis for the establishment of priority zones for the conservation, restoration, reforestation and sustainable management of Cinchona spp. species in Peru.

show abstract

Section: Discussionmentioning

confidence: 89%

Section: Introductionmentioning

confidence: 99%

Current and Future Spatial Distribution of the Genus Cinchona in Peru: Opportunities for Conservation in the Face of Climate Change

Vergara,

Cieza-Tarrillo,

Ocaña

et al. 2023

Sustainability

View full text Add to dashboard Cite

show abstract

“…Categories that exemplify this aggregated use are “Abundance,” “Habitat, distribution, niche and occupancy modeling,” “Phenology modeling,” “Population modeling,” and “Machine learning and AI identification.” These categories were less utilized than others because they generally require a large number of data points, often for a single species, and such a critical mass of data has only been reached for many species on various platforms within the last few years. Additionally, advances in statistical methods that can account for biases in community science data have only recently made such data useful for modeling (Bird et al 2014, Isaac et al 2014, Fithian et al 2015, Renner et al 2015, Pacifici et al 2017, Daniel et al 2020, van Eupen et al 2021). “Habitat, distribution, niche and occupancy modeling” was the most common category with 129 publications (6.1%).…”

Section: Resultsmentioning

confidence: 99%

Online community photo-sharing in entomology: a large-scale review with suggestions on best practices

Skvarla,

Fisher

2023

Annals of the Entomological Society of America

View full text Add to dashboard Cite

Over the past 2 decades, digital photography has grown increasingly accessible. This has ushered in a golden age of community science, where nonspecialists share natural history observations from across the globe via digital media. Importantly, these observations are accessible to researchers, who can readily share expertise directly with the community and connect observations and observers with research projects. This has fueled a rising number of publications combining specialist and nonspecialist observations, which occasionally make national and international headlines. Entomology has embraced this trend, but the scope and impact are not clear. A review of its effect on the field is therefore warranted. Herein, we review and analyze publications that incorporated information from photographs shared on photo-sharing websites. In total, 2,123 publications that incorporated information from 77 photo-sharing websites were examined. Seven websites accounted for 66% of the publication citations. 84.6% of publications focused on data from the Holarctic ecoregion (56.2% Palearctic and 28.4% Nearctic). Forty-six arthropod orders were represented, but the Big Five—Lepidoptera, Coleoptera, Hymenoptera, Hemiptera, and Diptera—accounted for 75% of all publications. We divide the publications into 28 discovery and use categories, review how community photograph data has been used within each category, and provide examples of categories utilized in nonentomology natural history fields, which we hope will provide inspiration and spur future research. We also discuss benefits and considerations when using such data—including the accuracy of identifications, inherent biases, and digital data impermanence—and suggest best practices to follow.

show abstract

“…When citizen science data consist of presence‐only, or only nonzero counts, researchers have often created pseudo‐absences, or zeros in the data, in an attempt to model where individuals do not occur (Conn et al, 2015a; Pearce and Boyce, 2006). A simple approach is to create zeros at random from all plots other than those with observed values (Stockwell and Peterson, 2002), but better approaches correct for sampling bias (Conn et al, 2017; Phillips et al, 2009) and include case–control methods (Fithian and Hastie, 2014) and local background sampling (Daniel et al, 2020), among others. In what follows, we will propose new ideas for creating pseudo‐absences based on spatial considerations.…”

Section: Introductionmentioning

confidence: 99%