SDMtune: An R package to tune and evaluate species distribution models

Ecology and Evolution, Год журнала: 2020, Номер 10(20), С. 11488 - 11506

Опубликована: Сен. 30, 2020

Abstract Balancing model complexity is a key challenge of modern computational ecology, particularly so since the spread machine learning algorithms. Species distribution models are often implemented using wide variety algorithms that can be fine‐tuned to achieve best prediction while avoiding overfitting. We have released SDMtune , new R package aims facilitate training, tuning, and evaluation species in unified framework. The main innovations this its functions perform data‐driven variable selection, novel genetic algorithm tune hyperparameters. Real‐time interactive charts displayed during execution several help users understand effect removing or varying hyperparameters on performance. supports three different metrics evaluate performance: area under receiver operating characteristic curve, true skill statistic, Akaike's information criterion corrected for small sample sizes. It implements four statistical methods: artificial neural networks, boosted regression trees, maximum entropy modeling, random forest. Moreover, it includes display outputs create final report. therefore represents new, user‐friendly framework still‐growing field modeling.

Язык: Английский

---

Major restructuring of marine plankton assemblages under global warming

Nature Communications, Год журнала: 2021, Номер 12(1)

Опубликована: Сен. 1, 2021

Marine phytoplankton and zooplankton form the basis of ocean's food-web, yet impacts climate change on their biodiversity are poorly understood. Here, we use an ensemble species distribution models for a total 336 524 to determine present future habitat suitability patterns. For end this century, under high emission scenario, find overall increase in plankton richness driven by ocean warming, poleward shift species' distributions at median speed 35 km/decade. Phytoplankton is projected more than 16% over most regions except Arctic Ocean. In contrast, slightly decline tropics, but strongly temperate subpolar latitudes. these latitudes, nearly 40% assemblages replaced shifting species. This implies that threatens contribution communities plankton-mediated ecosystem services such as biological carbon sequestration.

Язык: Английский

---

flexsdm: An r package for supporting a comprehensive and flexible species distribution modelling workflow

Methods in Ecology and Evolution, Год журнала: 2022, Номер 13(8), С. 1661 - 1669

Опубликована: Апрель 20, 2022

Abstract Species distribution models (SDM) are widely used in diverse research areas because of their simple data requirements and application versatility. However, SDM outcomes sensitive to input methodological choices. Such sensitivity applications mean that flexibility is necessary create SDMs with tailored protocols for a given set model use. We introduce the r package flexsdm supporting flexible species modelling workflows. functions arguments serve as building blocks construct specific protocol user's needs. The main features flexibility, integration other tools, simplicity objects returned function speed. As an illustration, we define complete workflow California red fir Abies magnifica . This provides by incorporating comprehensive tools structured three steps: (a) Pre‐modelling prepare input, example, sampling bias correction, pseudo‐absences background points, partitioning, reducing collinearity predictors. (b) Modelling allow fitting evaluating different approaches, including individual algorithms, tuned models, ensembles small ensemble models. (c) Post‐modelling include related models' predictions, interpolation overprediction correction. Because comprises large part process, from outlier detection users can delineate partial or workflows based on combination meet

Язык: Английский

---

Escarpment evolution drives the diversification of the Madagascar flora

Science, Год журнала: 2024, Номер 383(6683), С. 653 - 658

Опубликована: Фев. 8, 2024

Madagascar exhibits high endemic biodiversity that has evolved with sustained and stable rates of speciation over the past several tens millions years. The topography is dominated by a mountainous continental rift escarpment, highest plant diversity rarity found along steep, eastern side this geographic feature. Using process-explicit model, we show precipitation-driven erosion landward retreat high-relief creates transient habitat organization through multiple mechanisms, including catchment expansion, isolation highland remnants, formation topographic barriers. Habitat reconnection on million-year timescale serves as an allopatric pump creating observed biodiversity.

Язык: Английский

---

Macroecology in the age of Big Data – Where to go from here?

Journal of Biogeography, Год журнала: 2019, Номер 47(1), С. 1 - 12

Опубликована: Июль 17, 2019

Abstract Recent years have seen an exponential increase in the amount of data available all sciences and application domains. Macroecology is part this “Big Data” trend, with a strong rise volume that we are using for our research. Here, summarize most recent developments macroecology age Big Data were presented at 2018 annual meeting Specialist Group Ecological Society Germany, Austria Switzerland (GfÖ). Supported by computational advances, has been rapidly developing field over years. Our highlighted important avenues further progress terms standardized collection, integration, method development process integration. In particular, focus on (a) gaps new initiatives to close them, example through space‐ airborne sensors, (b) how various sources types can be integrated, (c) uncertainty assessed data‐driven analyses (d) machine learning approaches opened ways investigating processes rather than simply describing patterns. We discuss opens up opportunities, but also poses challenges macroecological future, it will essential carefully assess quality, reproducibility compilation analytical methods, communication uncertainties. Major depend definition standards workflows macroecology, such scientific quality integrity guaranteed, collaboration research projects made easier.

Язык: Английский

---

A quantitative review of abundance‐based species distribution models

Ecography, Год журнала: 2021, Номер 2022(1)

Опубликована: Дек. 15, 2021

The contributions of species to ecosystem functions or services depend not only on their presence but also local abundance. Progress in predictive spatial modelling has largely focused occurrence rather than As such, limited guidance exists the most reliable methods explain and predict variation We analysed performance 68 abundance‐based distribution models fitted 800 000 standardised abundance records for more terrestrial bird reef fish species. found a large amount models. While many performed poorly, subset consistently reconstructed range‐wide patterns. best predictions were obtained using random forests frequently encountered abundant within same environmental domain as model calibration. Extending outside conditions used training generated poor predictions. Thus, interpolation abundances between observations can help improve understanding patterns, our results indicate extrapolated under changing climate have much greater uncertainty. Our synthesis provides road map key property distributions that underpins theoretical applied questions ecology conservation.

Язык: Английский

---

Influence of climate, soil, and land cover on plant species distribution in the European Alps

Ecological Monographs, Год журнала: 2020, Номер 91(2)

Опубликована: Сен. 23, 2020

Abstract Although the importance of edaphic factors and habitat structure for plant growth survival is known, both are often neglected in favor climatic drivers when investigating spatial patterns species diversity. Yet, especially mountain ecosystems with complex topography, missing components may be detrimental a sound understanding biodiversity distribution. Here, we compare relative climate, soil land cover variables predicting distributions 2,616 vascular European Alps, representing approximately two‐thirds all flora. Using presence‐only data, built point‐process models (PPMs) to relate observations different combinations covariates. We evaluated PPMs through block cross‐validations assessed independent contributions soil, covariates predict using an innovative predictive partitioning approach. found climate most influential driver influence ~58.5% across species, decreasing from low high elevations. Soil (~20.1%) (~21.4%), overall, were less than but increased along elevation gradient. Furthermore, showed strong local effects lowlands, while contribution stabilized at mid‐elevations. The explained by increasing endemism, fact that becomes more homogeneous as diversity declines higher altitudes. In contrast, predictors follow opposite trend. Additionally, elevations, human‐mediated appear reduce predictors. conclude are, like principal distribution Alps. While disentangling their remains challenge, future studies can benefit markedly including distributions.

Язык: Английский

---

Identifying climate refugia for high‐elevation Alpine birds under current climate warming predictions

Global Change Biology, Год журнала: 2022, Номер 28(14), С. 4276 - 4291

Опубликована: Апрель 20, 2022

Abstract Identifying climate refugia is key to effective biodiversity conservation under a changing climate, especially for mountain‐specialist species adapted cold conditions and highly threatened by warming. We combined distribution models (SDMs) with forecasts identify high‐elevation bird ( Lagopus muta , Anthus spinoletta Prunella collaris Montifringilla nivalis ) in the European Alps, where ecological effects of changes are particularly evident predicted intensify. considered future (2041–2070) (SSP585 scenario, four models) identified three types refugia: (1) in‐situ potentially suitable both current conditions, ex‐situ (2) only according all or (3) at least out conditions. SDMs were based on very large, high‐resolution occurrence dataset (2901–12,601 independent records each species) collected citizen scientists. fitted using different algorithms, balancing statistical accuracy, realism predictive/extrapolation ability. selected most reliable ones consistency between training testing data extrapolation over distant areas. Future predictions revealed that (with partial exception A. will undergo range contraction towards higher elevations, losing 17%–59% their (larger losses L. ). ~15,000 km 2 Alpine region as species, which 44% currently designated protected areas (PAs; 18%–66% among countries). Our findings highlight usefulness spatially accurate scientists, importance model extrapolating Climate refugia, partly included within PAs system, should be priority sites habitats, habitat degradation/alteration human activities prevented ensure suitability alpine species.

Язык: Английский

---

Resolution in species distribution models shapes spatial patterns of plant multifaceted diversity

Ecography, Год журнала: 2022, Номер 2022(10)

Опубликована: Сен. 7, 2022

Species distribution models (SDMs) are statistical tools that relate species observations to environmental conditions retrieve ecological niches and predict species' potential geographic distributions. The quality robustness of SDMs clearly depend on good modelling practices including ascertaining the relevance predictors for studied choosing an appropriate spatial resolution (or ‘grain size'). While past studies showed improved model performance with increasing sessile organisms, there is still no consensus regarding how inappropriate can impede understanding mapping multiple facets diversity. Here, we modelled 1180 plant across European Alps two sets (climate soil) at resolutions ranging from 100‐m 40‐km. We assessed predictors' importance each resolution, calculated taxonomic (TD), relative phylogenetic (rPD) functional diversity (rFD) accordingly, compared resulting diversities space. In accordance previous studies, found predictive generally decrease decreasing predictor resolution. Overall, multifaceted was be strongly affected by particularly rPD, as exhibited weak average linear relationships between 1‐km (0.13 ≤ R 2 0.57). Our results demonstrate necessity using highly resolved explain distributions, especially in mountain environments. Using coarser might cause mispredicted, important consequences biodiversity management conservation.

Язык: Английский

---

Including imprecisely georeferenced specimens improves accuracy of species distribution models and estimates of niche breadth

Global Ecology and Biogeography, Год журнала: 2023, Номер 32(3), С. 342 - 355

Опубликована: Янв. 8, 2023

Abstract Aim Museum and herbarium specimen records are frequently used to assess the conservation status of species their responses climate change. Typically, occurrences with imprecise geolocality information discarded because they cannot be matched confidently environmental conditions thus expected increase uncertainty in downstream analyses. However, using only precisely georeferenced risks undersampling geographical distributions species. We present two related methods allow use imprecisely biogeographical analysis. Innovation Our procedures assign (1) locations or (2) climates that closest centroid precise a For virtual species, including alongside improved accuracy ecological niche models projected future, especially for c . 20 fewer occurrences. Using underestimated loss suitable habitat overestimated amount both future. Including also improves estimates breadth extent occurrence. An analysis 44 North American Asclepias (Apocynaceae) yielded similar results. Main conclusions Existing studies examining effects spatial imprecision typically compare outcomes based on against same error added them. real‐world cases, analysts possess mix must decide whether retain discard latter. Discarding can undersample lead mis‐estimation past future method, which we provide software implementation enmSdmX package R, is simple help leverage large number deemed “unusable” geolocation.

Язык: Английский

---