Detecting outliers in species distribution data: Some caveats and clarifications on a virtual species study

Liu et al. (2018) used a virtual species approach to test the effects of outliers on species distribution models. In their simulations, they applied a threshold value over the simulated suitabilities to generate the species distributions, suggesting that using a probabilistic simulation approach would have been more complex and yield the same results. Here, we argue that using a probabilistic approach is not necessarily more complex and may significantly change results. Although the threshold approach may be justified under limited circumstances, the probabilistic approach has multiple advantages. First, it is in line with ecological theory, which largely assumes non‐threshold responses. Second, it is more general, as it includes the threshold as a limiting case. Third, it allows a better separation of the relevant intervening factors that influence model performance. Therefore, we argue that the probabilistic simulation approach should be used as a general standard in virtual species studies.     

African Hepatics XXIX. Some new or little-known species and extensions of range

Abstract

Calypogeia afrocaernlea E. W. Jones, sp. nov. is described (type, from Kilimanjaro, in BM). Critical notes are provided on Cylindrocolea chevalieri (Steph.) Schuster, Simia S. Arnell, Syzygiella ruwenzorensis Steph., and Taxilejeunea pulchriflora Pearson. The following are new combinations: Cheilolejeunea surrepens (Mitt.) (basionym Lejeunea surrepens), Stictolejeunea balfourii (Mitt.) (basionym Lejeunea balfouri). Calypogeia fusca (Lehm.) Steph. is recorded from Cameroun and Ruanda; C. longifolia Steph. from Sierra Leone.     

Some extensions of the use of matrices in population theory

This paper extends Leslie's vector and matrix treatment of populations. A simple matrix is given for species in which adult mortality and fertility are independent of age, but in which the juvenile mortality rate differs from the adult. The population vector can be changed into a population matrix. This should allow equations using functions of the size of the population to be developed. Genetic variables such as sex or other polymorphisms can be introduced, and the notation allows different systems of selection or non-random mating to be specified.     

Detecting the macroevolutionary signal of species interactions

Species interactions lie at the heart of many theories of macroevolution, from adaptive radiation to the Red Queen. Although some theories describe the imprint that interactions will have over long timescales, we are still missing a comprehensive understanding of the effects of interactions on macroevolution. Current research shows strong evidence for the impact of interactions on macroevolutionary patterns of trait evolution and diversification, yet many macroevolutionary studies have only a tenuous relationship to ecological studies of interactions over shorter timescales. We review current research in this area, highlighting approaches that explicitly model species interactions and connect them to broad‐scale macroevolutionary patterns. We also suggest that progress has been made by taking an integrative interdisciplinary look at individual clades. We focus on African cichlids as a case study of how this approach can be fruitful. Overall, although the evidence for species interactions shaping macroevolution is strong, further work using integrative and model‐based approaches is needed to spur progress towards understanding the complex dynamics that structure communities over time and space.     

Some extensions of a linear model for categorical variables

The Grizzle-Starmer-Koch (GSK) model is extended to include the traditional log-linear model and a general class of Poisson and conditional Poisson distributions. Estimators of the model parameters are defined under general exact and stochastic linear constraints.     

Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation

Accurate modeling of geographic distributions of species is crucial to various applications in ecology and conservation. The best performing techniques often require some parameter tuning, which may be prohibitively time‐consuming to do separately for each species, or unreliable for small or biased datasets. Additionally, even with the abundance of good quality data, users interested in the application of species models need not have the statistical knowledge required for detailed tuning. In such cases, it is desirable to use “default settings”, tuned and validated on diverse datasets. Maxent is a recently introduced modeling technique, achieving high predictive accuracy and enjoying several additional attractive properties. The performance of Maxent is influenced by a moderate number of parameters. The first contribution of this paper is the empirical tuning of these parameters. Since many datasets lack information about species absence, we present a tuning method that uses presence‐only data. We evaluate our method on independently collected high‐quality presence‐absence data. In addition to tuning, we introduce several concepts that improve the predictive accuracy and running time of Maxent. We introduce “hinge features” that model more complex relationships in the training data; we describe a new logistic output format that gives an estimate of probability of presence; finally we explore “background sampling” strategies that cope with sample selection bias and decrease model‐building time. Our evaluation, based on a diverse dataset of 226 species from 6 regions, shows: 1) default settings tuned on presence‐only data achieve performance which is almost as good as if they had been tuned on the evaluation data itself; 2) hinge features substantially improve model performance; 3) logistic output improves model calibration, so that large differences in output values correspond better to large differences in suitability; 4) “target‐group” background sampling can give much better predictive performance than random background sampling; 5) random background sampling results in a dramatic decrease in running time, with no decrease in model performance.     

Using indicator species to assess the state of macrobenthic communities

Environmental impact assessments are often followed by the continuous monitoring needed to determine community change. This long-term monitoring can be time-consuming and expensive. The concept of indicator species attempts to use their presence in a sample or area to characterise a certain degree of community change or pollution effects. This approach has been widely applied to benthic monitoring studies. However, many studies develop their own list of `indicators' in cases without having a prior knowledge of the area or any long-term data. This can result in the production of circular arguments. We carry out a meta-analysis on data sets from 5 of the 20 designated United Kingdom's sewage sludge dumping grounds and the data set from the classic study of Pearson & Rosenberg (1978). We construct a number of indices to examine this robustness across studies. Having refined our criteria for an `indicator taxa' we examine the spatial and temporal changes in macrobenthic communities occurring at the Tyne sewage sludge dumpsite to examine the utility of this approach. Of the total pool of 123 taxa, 81 taxa responded in one study only. While Spio filicornis(O. F. Müller), Spiophanes bombyx(Claparède), Lagis koreni (Malmgren) and Nephtys cirrosa (Ehlers) showed directly contradictory patterns in different locations. The Spearman's rank correlation test showed a significant negative relationship between the density of macro-litter per station found at the Tyne dumping ground and the abundance of Abra alba(Wood) (r _s=0.462, n= 6, P=0.1) and Amphiura filiformis(O. F. Müller)(r _s=0.493, n= 6, P=0.1). These were the only indicator taxa, which showed a strong relationship to sewage contamination. We therefore conclude that while the concept of indicators may be widely applicable, the actual indicator taxa are not. This demonstrate that the used of indicators must be continually developed providing prior information of the study area.     

Multiple equilibria in complex chemical reaction networks: extensions to entrapped species models

In two earlier papers, means were provided to decide the capacity of complex chemical reaction networks, taken with mass-action kinetics, to admit multiple equilibria in the context of the isothermal homogeneous continuous flow stirred tank reactor (CFSTR). In such a reactor, all species are deemed to be in the outflow, a fact which has an important bearing on the nature of the governing equations. In contrast, one can imagine CFSTR-like models of the cell in which certain large molecules (e.g., enzymes) remain entrapped within the cell, whereas smaller ones (e.g., metabolites) are free to diffuse through the cell boundary. Although such models bear a strong physical resemblance to the classical CFSTR picture, there are substantive differences in the corresponding mathematics. Without a presumption of mass-action kinetics, this research is intended to indicate a general way in which results about uniqueness of equilibria in the classical CFSTR context extend to entrapped species models.     

Detecting emergent effects of multiple predator species

When foraging together, multiple predator species that share a single prey often cause prey mortality that cannot be predicted based on knowledge of predation by each species separately. Modeling and managing the effects of multiple predator species depend on accurately assessing these combined effects. Two methods are currently used to experimentally examine combined predation by multiple predator species: the additive and substitutive experimental designs. I simultaneously employed both experimental designs to examine predation by two crab species on shared mussel prey. I show that the two methods yield results that disagree both quantitatively and qualitatively, leading to very different conclusions about the way that predator species combine to affect prey mortality. This discrepancy occurred because the two methods examine complimentary, but not interchangeable questions. I advocate using an experimental design that incorporates both additive and substitutive designs to achieve a more complete understanding of the combined effects of multiple predator species.     

The value of indicator species for landscape characterization

Given that different types of landscape are affected by different environmental factors, is it possible to identify plant species indicative of these landscape types? An analysis of indicator species was made using 140 vegetation samples distributed over five landscape types. A total of 30 indicator species were identified, displaying varying indicator value. The variation disclosed the existence of a gradient between landscape types in terms of the prevailing ecological conditions. The analysis of the results also led to a reflection on the hierarchical level of which species are indicative (landscape vs. ecosystem). Further 20 samples were taken to check the validity of the indicator species. The identification of the indicator species present in these samples enabled each sample to be assigned to one of the landscape types; that assignment was subsequently checked against the actual landscape from which samples were taken. Seventy-five percent of the 20 validation samples were correctly assigned; the misassignment of the remaining 25% was due largely to the heterogeneous nature of the landscapes. On the basis of these results, we concluded that it is possible to characterize landscapes on the basis of indicative plant species, although characterization becomes more incomplete as the internally heterogeneous nature of the landscape increases.     

Constrained indicator species analysis (COINSPAN): an extension of TWINSPAN

Abstract. A new vegetation classification method, COnstrained-INdicator-SPecies-ANalysis (COINSPAN) is introduced as an elaboration of Two-Way-INdicator-SPecies-ANalysis (TWINSPAN). Instead of operating on the bisection of a primary Correspondence Analysis ordination axis, at each level of division, COINSPAN bisects a primary Canonical Correspondence Analysis axis. The associated computer program, of the same name, performs both analyses as options. COINSPAN is applied to a simple model data set, with two constraining gradients, and to a pine forest vegetation survey in order to illustrate the functioning and utility of the method.     

Application of detectability in the use of indicator species: A case study with birds

The use of indicator species is popular in ecological monitoring and management. In recent years, new methods to improve the quality and application of indicator data have been proposed and developed. Here we propose the use of detection probability in the selection and application of indicator species. We evaluated environmental and observer factors believed to affect detection of potential species. Observer effects were the most evident factor and may necessitate the greatest consideration in the use of indicator species. Our results call attention to the fact that raw counts are far from accurate and that the use of detection probability can and should be incorporated into sampling protocols, species selection, and the allocation of effort for projects that use indicator species as part of monitoring and management programs.     

A measure of species diversity related to the lognormal distribution of individuals among species

An investigation has been made of the relation between species diversity and the lognormal distribution of individuals among species, using phytoplankton samples from the Indian Ocean. The area under the truncated lognormal curve representing a sample gives the total number of species, S, while the logarithmic standard deviation, σ, gives a measure of the scatter of distribution of individuals among species, the other factor affecting diversity. Using a method described by Hald, truncated lognormal curves were fitted to the phytoplankton data and estimates obtained for the mean, ξ, the logarithmic standard deviation, σ, the number of species in the modal octave, N_o, and the number of species in the population universe, N. Since one is interested in a sample only in so far as it reflects the properties of the population from whence it came, estimated population parameters were used to define measures of diversity by which means it was hoped to obtain a diversity index independent of sample size, i.e., a diversity value related more to the population than to the sample.     

Phylogeny ofEncephalartos: Some eastern cape species

A group of eight species ofEncephalartos, comprisingE. altensteinii, E. arenarius, E. horridus,E. latifrons, E. lehmannii, E. longifolius, E. princeps, andE. trispinosus, from the Eastern Cape Province of South Africa, was studied by analysis of iso-enzymes, ribosome DNA, and ITS 1 and 2 genes. The reason for this investigation was that the morphology of the vegetative and reproductive parts, though very distinctive in this geographical region, do not correlate, and it was hoped that molecular data would elucidate evolutionary relationships. The three sets of molecular data were found to agree to a remarkable degree. It was concluded that the vegetative morphology was misleading but that the cone characteristics agree with molecular data and provide insight into the interrelationships of these species. Thus,E. princeps was concluded to be relatively remotely related to the vegetatively similarE. lehmannii, andE. arenarius is not at all close toE. latifrons even though the two species are easy to confuse when not in cone.