Inventory statistics meet big data: complications for estimating numbers of species

We point out complications inherent in biodiversity inventory metrics when applied to large-scale datasets. The number of units of inventory effort (e.g., days of inventory effort) in which a species is detected saturates, such that crucial numbers of detections of rare species approach zero. Any rare errors can then come to dominate species richness estimates, creating upward biases in estimates of species numbers. We document the problem via simulations of sampling from virtual biotas, illustrate its potential using a large empirical dataset (bird records from Cape May, NJ, USA), and outline the circumstances under which these problems may be expected to emerge.

Sampling effort determination in bird surveys: do current norms meet best-practice recommendations?

A critical design component of studies measuring diversity is sampling effort. Allocation of sampling effort dictates how many sites can be sampled within a particular time-frame or budget, as well as sample duration, frequency and intensity, thereby determining the resolution and reliability of emergent inferences. Conventional survey techniques use fixed-effort methods that assume invariant detectabilities among sites and species. Several approaches have been developed in the past decade that account for variable detectability by using alternative sampling methods or by adjusting standard counts before analysis, but it is unclear how widely adopted these techniques have been or how current bird surveying norms compare with best-practice recommendations. I conducted a systematic search of the primary literature to ascertain how sampling effort is determined, how much effort is devoted to sampling each site and how variation in detectability is dealt with. Of 225 empirical studies of bird diversity published between 2004 and 2016, five used results-based stopping rules (each derived independently), 54 used proportional sampling, and 159 (71%) used implicit effort-based stopping rules (fixed effort). Effort varied widely, but 61% of studies used samples of 10min or less and 62% of studies expended total effort per datum of 2h or less, with 78% providing no justification for sampling efforts used and just 15% explicitly accounting for estimated detectability. Given known variation in detectability, relying on short-duration fixed-effort approaches without validation or post hoc correction means that most bird diversity studies necessarily under-sample some sites and/or species. Having identified current bird surveying norms and highlighted their shortcomings, I provide five practical solutions to improve sampling effort determination, urging contributors and consumers of empirical ecological literature to consider survey data in terms of sample completeness.

Novel application of species richness estimators to predict the host range of parasites

Host range is a critical life history trait of parasites, influencing prevalence, virulence and ultimately determining their distributional extent. Current approaches to measure host range are sensitive to sampling effort, the number of known hosts increasing with more records. Here, we develop a novel application of results-based stopping rules to determine how many hosts should be sampled to yield stable estimates of the number of primary hosts within regions, then use species richness estimation to predict host ranges of parasites across their distributional ranges. We selected three mistletoe species (hemiparasitic plants in the Loranthaceae) to evaluate our approach: a strict host specialist (Amyema lucasii, dependent on a single host species), an intermediate species (Amyema quandang, dependent on hosts in one genus) and a generalist (Lysiana exocarpi, dependent on many genera across multiple families), comparing results from geographically-stratified surveys against known host lists derived from herbarium specimens. The results-based stopping rule (stop sampling bioregion once observed host richness exceeds 80% of the host richness predicted using the Abundance-based Coverage Estimator) worked well for most bioregions studied, being satisfied after three to six sampling plots (each representing 25 host trees) but was unreliable in those bioregions with high host richness or high proportions of rare hosts. Although generating stable predictions of host range with minimal variation among six estimators trialled, distribution-wide estimates fell well short of the number of hosts known from herbarium records. This mismatch, coupled with the discovery of nine previously unrecorded mistletoe-host combinations, further demonstrates the limited ecological relevance of simple host-parasite lists. By collecting estimates of host range of constrained completeness, our approach maximises sampling efficiency while generating comparable estimates of the number of primary hosts, with broad applicability to many host-parasite systems.

Digital Accessible Knowledge and well-inventoried sites for birds in Mexico: baseline sites for measuring faunistic change

Background

Faunal change is a basic and fundamental element in ecology, biogeography, and conservation biology, yet vanishingly few detailed studies have documented such changes rigorously over decadal time scales. This study responds to that gap in knowledge, providing a detailed analysis of Digital Accessible Knowledge of the birds of Mexico, designed to marshal DAK to identify sites that were sampled and inventoried rigorously prior to the beginning of major global climate change (1980).

Methods

We accumulated DAK records for Mexican birds from all relevant online biodiversity data portals. After extensive cleaning steps, we calculated completeness indices for each 0.05° pixel across the country; we also detected ‘hotspots’ of sampling, and calculated completeness indices for these broader areas as well. Sites were designated as well-sampled if they had completeness indices above 80% and >200 associated DAK records.

Results

We identified 100 individual pixels and 20 broader ‘hotspots’ of sampling that were demonstrably well-inventoried prior to 1980. These sites are catalogued and documented to promote and enable resurvey efforts that can document events of avifaunal change (and non-change) across the country on decadal time scales.

Conclusions

Development of repeated surveys for many sites across Mexico, and particularly for sites for which historical surveys document their avifaunas prior to major climate change processes, would pay rich rewards in information about distributional dynamics of Mexican birds.

Bird diversity and distribution in relation to urban landscape types in northern Rwanda

Using the point count method, linear mixed models, Shannon's diversity index, and Bray-Curtis cluster analysis, we conducted a study of the effect of urban fabric layout on bird diversity and distribution in northern Rwanda. The results showed a significant effect of city landscapes on bird richness and relative abundance; residential neighborhoods, institutional grounds, and informal settlements had the highest species diversity in comparison to other microlandscape types. Riversides were characterized by specialized bird species, commonly known to be restricted to wetland environments. Built-up areas and open field landscapes had comparable results. One Albertine Rift endemic bird species, the Ruwenzori Double-collared Sunbird (Cinnyris stuhlmanni), was recorded. Three migratory birds were found in Musanze city for the first time: the Common Sandpiper (Actitis hypoleucos), the Spotted Flycatcher (Muscicapa striata), and the Willow Warbler (Phylloscopus trochilus). Two bird species have not been previously reported in Rwanda: the Garden Warbler (Sylvia borin) and the Lesser Spotted Eagle (Aquila pomarina). The implications of this study are particularly relevant to urban decision makers who should consider the existence of a great diversity of avian fauna when developing and implementing master plans, especially when villages and cities are in proximity of protected areas or natural reserves.

Patterns of host and flea communities along an elevational gradient in Colorado

Patterns in community composition across a landscape are the result of mechanistic responses and species interactions. Interactions between hosts and parasites have additional complexity because of the contingency of host presence and interactions among parasites. To assess the role of environmental changes within host and parasite communities, we surveyed small mammals and their fleas over a dynamic elevational gradient in the Front Range in Colorado, USA. Communities were characterized using several richness and diversity metrics and these were compared using a suite of frequentist and randomization approaches. We found that flea species richness was related to the number of host species based upon rarefaction, but no patterns in richness with elevation were evident. Values of diversity measures increased with elevation, representing that small-mammal and flea communities were more even upslope, yet turnover in composition was not related to examined variables. The results suggest there are strong local effects that drive these small-mammal and flea communities, although the breadth of flea species is tied to host availability.

Use of approximate inference in an index of completeness of biological inventories

To assess the completeness of a floristic or faunal inventory, one may use the ratio of the observed number of species to the "true number" of species (C). If the inventory is complete, C =1. The estimate of the true number can be obtained from accumulation curves, nonparametric methods, or other techniques. We devised a simple method for computing confidence intervals (CI) for C and for evaluating the null hypothesis that the inventory is complete. The method is based on the assumptions that an estimate of the variance of the true number of species is known and that the distribution of the estimator of the true number of species is approximately normal. We applied our method to bird inventories in the Balsas Basin of Mexico. The completeness index for subtransects were lower (84.0, 85.4, and 89.9%) than for the whole transect (91.6%) (all significantly different from 100%). Thus, these particular inventories were incomplete at 2 spatial resolutions. Our method of estimating CI for C can be used to estimate species richness obtained from databases of different sites or to test the null hypothesis that an inventory derived from a database is complete.

Evaluating the performance of species richness estimators: sensitivity to sample grain size

1. Fifteen species richness estimators (three asymptotic based on species accumulation curves, 11 nonparametric, and one based in the species-area relationship) were compared by examining their performance in estimating the total species richness of epigean arthropods in the Azorean Laurisilva forests. Data obtained with standardized sampling of 78 transects in natural forest remnants of five islands were aggregated in seven different grains (i.e. ways of defining a single sample): islands, natural areas, transects, pairs of traps, traps, database records and individuals to assess the effect of using different sampling units on species richness estimations. 2. Estimated species richness scores depended both on the estimator considered and on the grain size used to aggregate data. However, several estimators (ACE, Chao 1, Jackknifel and 2 and Bootstrap) were precise in spite of grain variations. Weibull and several recent estimators [proposed by Rosenzweig et al. (Conservation Biology, 2003, 17, 864-874), and Ugland et al. (Journal of Animal Ecology, 2003, 72, 888-897)] performed poorly. 3. Estimations developed using the smaller grain sizes (pair of traps, traps, records and individuals) presented similar scores in a number of estimators (the above-mentioned plus ICE, Chao2, Michaelis-Menten, Negative Exponential and Clench). The estimations from those four sample sizes were also highly correlated. 4. Contrary to other studies, we conclude that most species richness estimators may be useful in biodiversity studies. Owing to their inherent formulas, several nonparametric and asymptotic estimators present insensitivity to differences in the way the samples are aggregated. Thus, they could be used to compare species richness scores obtained from different sampling strategies. Our results also point out that species richness estimations coming from small grain sizes can be directly compared and other estimators could give more precise results in those cases. We propose a decision framework based on our results and on the literature to assess which estimator should be used to compare species richness scores of different sites, depending on the grain size of the original data, and of the kind of data available (species occurrence or abundance data).