An official website of the United States government
Abstract Climate and land use change are two of the primary threats to global biodiversity; however, each species within a community may respond differently to these facets of global change. Although it is typically assumed that species use the habitat that is advantageous for survival and reproduction, anthropogenic changes to the environment can create ecological traps, making it critical to assess both habitat selection (e.g. where species congregate on the landscape) and the influence of selected habitats on the demographic processes that govern population dynamics.We used a long‐term (1958–2011), large‐scale, multi‐species dataset for waterfowl that spans the United States and Canada to estimate species‐specific responses to climate and land use variables in a landscape that has undergone significant environmental change across space and time. We first estimated the effects of change in climate and land use variables on habitat selection and population dynamics for nine species. We then hypothesized that species‐specific responses to environmental change would scale with life‐history traits, specifically: longevity, nesting phenology and female breeding site fidelity.We observed species‐level heterogeneity in the demographic and habitat selection responses to climate and land use change, which would complicate community‐level habitat management. Our work highlights the importance of multi‐species monitoring and community‐level analysis, even among closely related species.We detected several relationships between life‐history traits, particularly nesting phenology, and species' responses to environmental change. One species, the early‐nesting northern pintail (Anas acuta), was consistently at the extreme end of responses to land use and climate predictors and has been a species of conservation concern since their population began to decline in the 1980s. They, and the blue‐winged teal, also demonstrated a positive habitat selection response to the proportion of cropland on the landscape that simultaneously reduced abundance the following year, indicative of susceptibility to ecological traps.By distilling the diversity of species' responses to environmental change within a community, our methodological approach and findings will help improve predictions of community responses to global change and can inform multi‐species management and conservation plans in dynamic landscapes that are based on simple tenets of life‐history theory.
more »
« less
Leveraging relationships between species abundances to improve predictions and inform conservation
Abstract Many management and conservation contexts can benefit from understanding relationships between species abundances, which can be used to improve predictions of species occurrence and abundance.We present conditional prediction as a tool to capture information about species abundances via residual covariance between species. From a fitted joint species distribution model, this framework produces a species coefficient matrix that contains relationships between species abundances. The species coefficients allow co‐observed species to be treated as a second set of predictors supplementing covariates in the model to improve prediction. We use simulations to demonstrate the potential benefits and limitations of conditional prediction across data types and species covariance before applying conditional prediction to two management contexts with real data.Simulations demonstrate that conditional prediction provides the largest benefits to continuous data and when there is residual covariance between many species.In our first application, we show that conditioning on other species improves in‐sample and out‐of‐sample predictions of fish and invertebrate species, including Atlantic cod. In our second application, we show that the species coefficient matrix can be used to identify bird species at risk of nest parasitism by Brown‐headed Cowbirds.Synthesis and applications. We present guidelines for using conditional prediction, which can help understand relationships between species abundances, improve predictions and inform conservation in a variety of contexts.
more »
« less
- Award ID(s):
- 2211764
- PAR ID:
- 10633528
- Publisher / Repository:
- Wiley
- Date Published:
- Journal Name:
- Journal of Applied Ecology
- Volume:
- 61
- Issue:
- 7
- ISSN:
- 0021-8901
- Page Range / eLocation ID:
- 1662 to 1672
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Summary Plant metabolites from diverse pathways are important for plant survival, human nutrition and medicine. The pathway memberships of most plant enzyme genes are unknown. While co‐expression is useful for assigning genes to pathways, expression correlation may exist only under specific spatiotemporal and conditional contexts.Utilising > 600 tomato (Solanum lycopersicum) expression data combinations, three strategies for predicting memberships in 85 pathways were explored.Optimal predictions for different pathways require distinct data combinations indicative of pathway functions. Naive prediction (i.e. identifying pathways with the most similarly expressed genes) is error prone. In 52 pathways, unsupervised learning performed better than supervised approaches, possibly due to limited training data availability. Using gene‐to‐pathway expression similarities led to prediction models that outperformed those based simply on expression levels. Using 36 experimental validated genes, the pathway‐best model prediction accuracy is 58.3%, significantly better compared with that for predicting annotated genes without experimental evidence (37.0%) or random guess (1.2%), demonstrating the importance of data quality.Our study highlights the need to extensively explore expression‐based features and prediction strategies to maximise the accuracy of metabolic pathway membership assignment. The prediction framework outlined here can be applied to other species and serves as a baseline model for future comparisons.more » « less
-
Abstract A key challenge in conservation biology is that not all species are equally likely to go extinct when faced with a disturbance, but there are multiple overlapping reasons for such differences in extinction probability. Differences in species extinction risk may represent extinction selectivity, a non‐random process by which species’ risks of extinction are caused by differences in fitness based on traits. Additionally, rare species with low abundances and/or occupancies are more likely to go extinct than common species for reasons of random chance alone, that is, bad luck. Unless ecologists and conservation biologists can disentangle random and selective extinction processes, then the prediction and prevention of future extinctions will continue to be an elusive challenge.We suggest that a modified version of a common null model procedure, rarefaction, can be used to disentangle the influence of stochastic species loss from selective non‐random processes. To this end we applied a rarefaction‐based null model to three published data sets to characterize the influence of species rarity in driving biodiversity loss following three biodiversity loss events: (a) disease‐associated bat declines; (b) disease‐associated amphibian declines; and (c) habitat loss and invasive species‐associated gastropod declines. For each case study, we used rarefaction to generate null expectations of biodiversity loss and species‐specific extinction probabilities.In each of our case studies, we find evidence for both random and non‐random (selective) extinctions. Our findings highlight the importance of explicitly considering that some species extinctions are the result of stochastic processes. In other words, we find significant evidence for bad luck in the extinction process.Policy implications. Our results suggest that rarefaction can be used to disentangle random and non‐random extinctions and guide management decisions. For example, rarefaction can be used retrospectively to identify when declines of at‐risk species are likely to result from selectivity, versus random chance. Rarefaction can also be used prospectively to formulate minimum predictions of species loss in response to hypothetical disturbances. Given its minimal data requirements and familiarity among ecologists, rarefaction may be an efficient and versatile tool for identifying and protecting species that are most vulnerable to global extinction.more » « less
-
Abstract Phylogenetic regression is a type of generalised least squares (GLS) method that incorporates a modelled covariance matrix based on the evolutionary relationships between species (i.e. phylogenetic relationships). While this method has found widespread use in hypothesis testing via phylogenetic comparative methods, such as phylogenetic ANOVA, its ability to account for non‐linear relationships has received little attention.To address this, here we implement a phylogenetic Kernel Ridge Regression (phyloKRR) method that utilises GLS in a high‐dimensional feature space, employing linear combinations of phylogenetically weighted data to account for non‐linearity. We analysed two biological datasets using the Radial Basis Function and linear kernel function. The first dataset contained morphometric data, while the second dataset comprised discrete trait data and diversification rates as response variable. Hyperparameter tuning of the model was achieved through cross‐validation rounds in the training set.In the tested biological datasets, phyloKRR reduced the error rate (as measured by RMSE) by around 20% compared to linear‐based regression when data did not exhibit linear relationships. In simulated datasets, the error rate decreased almost exponentially with the level of non‐linearity.These results show that introducing kernels into phylogenetic regression analysis presents a novel and promising tool for complementing phylogenetic comparative methods. We have integrated this method into Python package named phyloKRR, which is freely available at:https://github.com/ulises‐rosas/phylokrr.more » « less
-
Abstract A prominent challenge for managing migratory species is the development of conservation plans that accommodate spatiotemporally varying distributions throughout the year. Migratory networks are spatially‐explicit models that incorporate migratory assignment and seasonal abundance data to define patterns of connectivity between stages of the annual cycle. These models are particularly useful for widespread application because different types of migratory data can be used to quantify individual and population‐level movement across the annual cycle of migratory species. While there are clear benefits of combining migratory assignment and abundance data for the development of conservation strategies, there is a concurrent need for corresponding user‐friendly software to facilitate the integration of these data for conservation.Here, we presentmignette(migratory network tools ensemble), an R package for developing migratory network models to estimate network connectivity among migratory populations. We demonstrate the functionality ofmignettewith three empirical examples that highlight the use of different types of tracking data for migratory assignment.mignettefacilitates the modelling of migratory networks by providing R functions to: (1) define breeding and nonbreeding nodes, (2) assemble abundance and assignment data and (3) model the migratory network. Additionally,mignetteprovides R functions to visualize modelled migratory networks.With increasing availability of migratory assignment and abundance data,mignetterepresents a valuable tool for developing effective conservation strategies for migratory species.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1111/1365-2664.14670
