Spatial occupancy models applied to atlas data show Southern Ground Hornbills strongly depend on protected areas

The expansion process of the Iberian ibex in the Sierra de Guadarrama National Park, Madrid (Spain)

In this paper we explore the usefulness of MaxEnt to predict the most suitable areas for a wildlife species, the Iberian ibex (Capra pyrenaica). For two decades (1990–2010), the species was established in a small part of the National Park Sierra de Guadarrama (Spain) and there has been a process of expansion to other areas of this protected area since 2010. However, almost two decades have elapsed since the modeling methods (MaxEnt) were proposed and no studies have tested their effectiveness using real distribution data, i.e. data from past predictions, to see if they fit the current distribution. We generated a model with presence– only data from 2007 and verified accuracy from 2017 data concerning real presence. Our results show a relationship between models and the species' current presence. The generated model can be useful to define the preferred locations of the species. We detected several differences between males and females of the species. This work not only shows the importance of selecting climatic and ecological variables for the construction of models but also indicates that they must be adjusted, at least for some species, to each sex and period of the year.

Community confounding in joint species distribution models

Joint species distribution models have become ubiquitous for studying species-environment relationships and dependence among species. Accounting for community structure often improves predictive power, but can also affect inference on species-environment relationships. Specifically, some parameterizations of joint species distribution models allow interspecies dependence and environmental effects to explain the same sources of variability in species distributions, a phenomenon we call community confounding. We present a method for measuring community confounding and show how to orthogonalize the environmental and random species effects in suite of joint species distribution models. In a simulation study, we show that community confounding can lead to computational difficulties and that orthogonalizing the environmental and random species effects can alleviate these difficulties. We also discuss the inferential implications of community confounding and orthogonalizing the environmental and random species effects in a case study of mammalian responses to the Colorado bark beetle epidemic in the subalpine forest by comparing the outputs from occupancy models that treat species independently or account for interspecies dependence. We illustrate how joint species distribution models that restrict the random species effects to be orthogonal to the fixed effects can have computational benefits and still recover the inference provided by an unrestricted joint species distribution model.

Why a landscape view is important: nearby urban and agricultural land affects bird abundances in protected areas

Protected areas are one of the primary conservation tools used worldwide. However, they are often embedded in a landscape that is intensely used by people, such as for agriculture or urban development. The proximity of these land-use types to protected areas can potentially affect the ecological effectiveness (or conservation effectiveness) of protected areas. In this article, we examine to what degree adjacent agricultural and urban land uses affect the ecological effectiveness of protected areas over the greater Gauteng region of South Africa. We selected 198 common, resident bird species, and analysed detection/non-detection data for these species collected over regular grid cells (approximately 61 km² in area). For each species, we estimated abundance per grid cell with the Royle-Nichols model in relation to the proportion of protected area as a covariate. Our study focused on how this relationship between proportion of protected area and abundance (which we term the ‘protection–abundance relationship’) changed as a function of other land-use types in the grid cell. Specifically, we examined the interaction effects between protected area and both urban and agricultural land-use type per grid cell on bird abundance. We assigned each species to one of seven guilds, namely: frugivores, gleaners, granivores, ground-feeders, hawkers, predators and vegivores, and examined how the protection–abundance relationship varied across guilds in relation to agriculture and urban area. As urban area within a grid cell increased, the protection–abundance relationship became more positive for 58% of all species. At the level of guilds, the protection–abundance relationship became more positive for two guilds (granivores and ground-feeders), more negative for frugivores, and remained unchanged for the other four guilds (gleaners, hawkers, predators and vegivores). As agricultural area within a grid cell increased, the protection–abundance relationship became more positive for 49% of all species. At the guild level, the protection–abundance relationship became more positive for six guilds (frugivores, gleaners, ground-feeders, hawkers, predators and vegivores) and remained unchanged for the granivores. Our results show land-use type near protected areas modified the effect protected areas had on bird abundances, and hence the ecological effectiveness of protected areas. Our results suggest that protected areas should be viewed as constituents within the landscape, rather than islands of protection.

Indigenous knowledge of South African bird and rangeland ecology is effective for informing conservation science

South Africa's communal rangelands constitute ~25% of the country's land cover and are largely managed for livestock grazing. These habitats play an important role in rural livelihoods and cultural practices. Using semi-structured interviews, we documented indigenous local ecological knowledge (LEK) held by rural dwellers linked to natural resource utilisation, environmental health and cultural keystone indicator species (CKIS) in the grassland communities of southern KwaZulu-Natal, South Africa. Our main objective was to examine the ability for LEK to inform conservation management. We found that people who were heavily reliant on natural resources attained a higher LEK score, indicating a greater breadth of ecological knowledge, which in turn shaped their perceptions of environmental change. Community members confirmed the presence of conservation concern species within this area, highlighting the limitations of only using citizen science databases for conservation management, as their observations within these databases are biased towards major road routes and protected or urban areas. LEK can play an important role in identifying habitats crucial to species' persistence and delineating population trends over time. Our surveys highlighted the importance of the Southern Ground-hornbill Bucorvus leadbeateri as a CKIS that acts as an early warning system of changing weather, notably rain. However, LEK is context-specific, and some CKIS species such as the Southern Ground-hornbill have wide distribution ranges. Consequently, the cultural associations and implications differ based on local belief systems that are often defined by the language spoken and the community's geographical location. Our study demonstrated the importance of including indigenous LEK in conservation planning for threatened species and habitats and the importance of traditional family values responsible for transferring oral knowledge.

Utilizing bycatch camera-trap data for broad-scale occupancy and conservation: a case study of the brown hyaena Parahyaena brunnea

With human influences driving populations of apex predators into decline, more information is required on how factors affect species at national and global scales. However, camera-trap studies are seldom executed at a broad spatial scale. We demonstrate how uniting fine-scale studies and utilizing camera-trap data of non-target species is an effective approach for broadscale assessments through a case study of the brown hyaena Parahyaena brunnea. We collated camera-trap data from 25 protected and unprotected sites across South Africa into the largest detection/non-detection dataset collected on the brown hyaena, and investigated the influence of biological and anthropogenic factors on brown hyaena occupancy. Spatial autocorrelation had a significant effect on the data, and was corrected using a Bayesian Gibbs sampler. We show that brown hyaena occupancy is driven by specific co-occurring apex predator species and human disturbance. The relative abundance of spotted hyaenas Crocuta crocuta and people on foot had a negative effect on brown hyaena occupancy, whereas the relative abundance of leopards Panthera pardus and vehicles had a positive influence. We estimated that brown hyaenas occur across 66% of the surveyed camera-trap station sites. Occupancy varied geographically, with lower estimates in eastern and southern South Africa. Our findings suggest that brown hyaena conservation is dependent upon a multi-species approach focussed on implementing conservation policies that better facilitate coexistence between people and hyaenas. We also validate the conservation value of pooling fine-scale datasets and utilizing bycatch data to examine species trends at broad spatial scales.

Using citizen science data to inform the relative sensitivity of waterbirds to natural versus human-dominated landscapes in China

Habitat loss is widely regarded as one of the most destructive factors threatening native biodiversity. Because migratory waterbirds include some of the most globally endangered species, information on their sensitivity to landscape would benefit their conservation. While citizen science data on waterbird species occurrence are subjected to various biases, their appropriate interpretation can provide information of benefit to species conservation. We apply a bootstrapping procedure to citizen science data to reduce sampling biases and report the relative sensitivity of waterbird species to natural versus human-dominated landscapes. Analyses are performed on 30,491 data records for 69 waterbird species referred to five functional groups observed in China between 2000 and 2018. Of these taxa, 30 species (43.5%) are significantly associated with natural landscapes, more so for cranes, geese, and ducks than for shorebirds and herons. The relationship between land association and the threat status of waterbirds is significant when the range size of species is considered as the mediator, and the higher the land association, the higher the threat status. Sensitive species significantly associated with natural landscapes are eight times more likely to be classified as National Protected Species (NPS) Classes I or II than less sensitive species significantly associated with human-dominated landscapes. We demonstrate the potential for citizen science data to assist in conservation planning in the context of landscape changes. Our methods might assist others to obtain information to help relieve species decline and extinction.

Efficient Bayesian analysis of occupancy models with logit link functions

Occupancy models (Ecology, 2002; 83: 2248) were developed to infer the probability that a species under investigation occupies a site. Bayesian analysis of these models can be undertaken using statistical packages such as WinBUGS, OpenBUGS, JAGS, and more recently Stan, however, since these packages were not developed specifically to fit occupancy models, one often experiences long run times when undertaking an analysis. Bayesian spatial single-season occupancy models can also be fit using the R package stocc. The approach assumes that the detection and occupancy regression effects are modeled using probit link functions. The use of the logistic link function, however, is algebraically more tractable and allows one to easily interpret the coefficient effects of an estimated model by using odds ratios, which is not easily done for a probit link function for models that do not include spatial random effects. We develop a Gibbs sampler to obtain posterior samples from the posterior distribution of the parameters of various occupancy models (nonspatial and spatial) when logit link functions are used to model the regression effects of the detection and occupancy processes. We apply our methods to data extracted from the 2nd Southern African Bird Atlas Project to produce a species distribution map of the Cape weaver (Ploceus capensis) and helmeted guineafowl (Numida meleagris) for South Africa. We found that the Gibbs sampling algorithm developed produces posterior samples that are identical to those obtained when using JAGS and Stan and that in certain cases the posterior chains mix much faster than those obtained when using JAGS, stocc, and Stan. Our algorithms are implemented in the R package, Rcppocc. The software is freely available and stored on GitHub (https://github.com/AllanClark/Rcppocc).

Effectiveness of protected areas for bird conservation depends on guild

AIM

Protected areas are key conservation tools intended to increase biodiversity and reduce extinction risks of species and populations. However, the degree to which protected areas achieve their conservation goals is generally unknown for many protected areas worldwide. We assess the effect of protected areas on the abundance of 196 common, resident bird species. If protected areas were beneficial to avian biodiversity, we expect landscapes with a higher proportion of protected areas will have higher densities of species compared to landscapes with no protection.

LOCATION

Greater Gauteng region, South Africa.

METHODS

We analysed bird survey data collected over regular grid cells across the study area. We estimated bird abundance in relation to the proportion of a grid cell that was protected with the Royle-Nichols model and fitted the model once for each of the species. We examined variation in estimated abundance as a function of avian guild (defined by the type of food a species preferentially ate and its foraging mode) with a regression tree analysis.

RESULTS

Abundance was significantly positively related to the proportion of protected areas in grid cells for 26% of the species, significantly negatively related in 15%, and not significantly related in 59% species. We found three distinct guild groups which differed in their average abundance, after accounting for associated variance. Group 1 consisted of guilds frugivores, ground-feeders, hawkers, predators, and vegivores and average abundance was strongly positively related to the proportion of protected areas. Group 2 included granivores, and average abundance was strongly negatively related to proportion of protected areas. Group 3 included gleaners only, and average abundance was not related to proportion of protected areas.

MAIN CONCLUSION

We conclude that the network of protected areas within the greater Gauteng region sustained relatively higher abundances of common birds and thus perform an important conservation role.

Identifying important conservation areas for the clouded leopard Neofelis nebulosa in a mountainous landscape: Inference from spatial modeling techniques

The survival of large carnivores is increasingly precarious due to extensive human development that causes the habitat loss and fragmentation. Habitat selection is influenced by anthropogenic as well as environmental factors, and understanding these relationships is important for conservation management. We assessed the environmental and anthropogenic variables that influence site use of clouded leopard Neofelis nebulosa in Bhutan, estimated their population density, and used the results to predict the species’ site use across Bhutan. We used a large camera‐trap dataset from the national tiger survey to estimate for clouded leopards, for the first time in Bhutan, (1) population density using spatially explicit capture–recapture models and (2) site‐use probability using occupancy models accounting for spatial autocorrelation. Population density was estimated at D^Bayesian=0.40 (0.10 SD) and D^maximum−likelihood=0.30 (0.12 SE) per 100 km². Clouded leopard site use was positively associated with forest cover and distance to river while negatively associated with elevation. Mean site‐use probability (from the Bayesian spatial model) was ψ^spatial=0.448 (0.076 SD). When spatial autocorrelation was ignored, the probability of site use was overestimated, ψ^nonspatial=0.826 (0.066 SD). Predictive mapping allowed us to identify important conservation areas and priority habitats to sustain the future of these elusive, ambassador felids and associated guilds. Multiple sites in the south, many of them outside of protected areas, were identified as habitats suitable for this species, adding evidence to conservation planning for clouded leopards in continental South Asia.

The second Southern African Bird Atlas Project: Causes and consequences of geographical sampling bias

Using the Southern African Bird Atlas Project (SABAP2) as a case study, we examine the possible determinants of spatial bias in volunteer sampling effort and how well such biased data represent environmental gradients across the area covered by the atlas. For each province in South Africa, we used generalized linear mixed models to determine the combination of variables that explain spatial variation in sampling effort (number of visits per 5' × 5' grid cell, or "pentad"). The explanatory variables were distance to major road and exceptional birding locations or "sampling hubs," percentage cover of protected, urban, and cultivated area, and the climate variables mean annual precipitation, winter temperatures, and summer temperatures. Further, we used the climate variables and plant biomes to define subsets of pentads representing environmental zones across South Africa, Lesotho, and Swaziland. For each environmental zone, we quantified sampling intensity, and we assessed sampling completeness with species accumulation curves fitted to the asymptotic Lomolino model. Sampling effort was highest close to sampling hubs, major roads, urban areas, and protected areas. Cultivated area and the climate variables were less important. Further, environmental zones were not evenly represented by current data and the zones varied in the amount of sampling required representing the species that are present. SABAP2 volunteers' preferences in birding locations cause spatial bias in the dataset that should be taken into account when analyzing these data. Large parts of South Africa remain underrepresented, which may restrict the kind of ecological questions that may be addressed. However, sampling bias may be improved by directing volunteers toward undersampled regions while taking into account volunteer preferences.

Spatial models to account for variation in observer effort in bird atlases

To assess the importance of variation in observer effort between and within bird atlas projects and demonstrate the use of relatively simple conditional autoregressive (CAR) models for analyzing grid‐based atlas data with varying effort. Pennsylvania and West Virginia, United States of America. We used varying proportions of randomly selected training data to assess whether variations in observer effort can be accounted for using CAR models and whether such models would still be useful for atlases with incomplete data. We then evaluated whether the application of these models influenced our assessment of distribution change between two atlas projects separated by twenty years (Pennsylvania), and tested our modeling methodology on a state bird atlas with incomplete coverage (West Virginia). Conditional Autoregressive models which included observer effort and landscape covariates were able to make robust predictions of species distributions in cases of sparse data coverage. Further, we found that CAR models without landscape covariates performed favorably. These models also account for variation in observer effort between atlas projects and can have a profound effect on the overall assessment of distribution change. Accounting for variation in observer effort in atlas projects is critically important. CAR models provide a useful modeling framework for accounting for variation in observer effort in bird atlas data because they are relatively simple to apply, and quick to run.

Using citizen science data to identify the sensitivity of species to human land use

Conservation practitioners must contend with an increasing array of threats that affect biodiversity. Citizen scientists can provide timely and expansive information for addressing these threats across large scales, but their data may contain sampling biases. We used randomization procedures to account for possible sampling biases in opportunistically reported citizen science data to identify species' sensitivities to human land use. We analyzed 21,044 records of 143 native reptile and amphibian species reported to the Carolina Herp Atlas from North Carolina and South Carolina between 1 January 1990 and 12 July 2014. Sensitive species significantly associated with natural landscapes were 3.4 times more likely to be legally protected or treated as of conservation concern by state resource agencies than less sensitive species significantly associated with human-dominated landscapes. Many of the species significantly associated with natural landscapes occurred primarily in habitats that had been nearly eradicated or otherwise altered in the Carolinas, including isolated wetlands, longleaf pine savannas, and Appalachian forests. Rare species with few reports were more likely to be associated with natural landscapes and 3.2 times more likely to be legally protected or treated as of conservation concern than species with at least 20 reported occurrences. Our results suggest that opportunistically reported citizen science data can be used to identify sensitive species and that species currently restricted primarily to natural landscapes are likely at greatest risk of decline from future losses of natural habitat. Our approach demonstrates the usefulness of citizen science data in prioritizing conservation and in helping practitioners address species declines and extinctions at large extents.

A Variational Bayes Approach to the Analysis of Occupancy Models

Detection-nondetection data are often used to investigate species range dynamics using Bayesian occupancy models which rely on the use of Markov chain Monte Carlo (MCMC) methods to sample from the posterior distribution of the parameters of the model. In this article we develop two Variational Bayes (VB) approximations to the posterior distribution of the parameters of a single-season site occupancy model which uses logistic link functions to model the probability of species occurrence at sites and of species detection probabilities. This task is accomplished through the development of iterative algorithms that do not use MCMC methods. Simulations and small practical examples demonstrate the effectiveness of the proposed technique. We specifically show that (under certain circumstances) the variational distributions can provide accurate approximations to the true posterior distributions of the parameters of the model when the number of visits per site (K) are as low as three and that the accuracy of the approximations improves as K increases. We also show that the methodology can be used to obtain the posterior distribution of the predictive distribution of the proportion of sites occupied (PAO).

Twenty-five years of change in southern African passerine diversity: nonclimatic factors of change

We analysed more than 25 years of change in passerine bird distribution in South Africa, Swaziland and Lesotho, to show that species distributions can be influenced by processes that are at least in part independent of the local strength and direction of climate change: land use and ecological succession. We used occupancy models that separate species' detection from species' occupancy probability, fitted to citizen science data from both phases of the Southern African Bird Atlas Project (1987-1996 and 2007-2013). Temporal trends in species' occupancy probability were interpreted in terms of local extinction/colonization, and temporal trends in detection probability were interpreted in terms of change in abundance. We found for the first time at this scale that, as predicted in the context of bush encroachment, closed-savannah specialists increased where open-savannah specialists decreased. In addition, the trend in the abundance of species a priori thought to be favoured by agricultural conversion was negatively correlated with human population density, which is in line with hypotheses explaining the decline in farmland birds in the Northern Hemisphere. In addition to climate, vegetation cover and the intensity and time since agricultural conversion constitute important predictors of biodiversity changes in the region. Their inclusion will improve the reliability of predictive models of species distribution.

Departures from the Energy-Biodiversity Relationship in South African Passerines: Are the Legacies of Past Climates Mediated by Behavioral Constraints on Dispersal?

Legacies of paleoclimates in contemporary biodiversity patterns have mostly been investigated with global datasets, or with weakly dispersive organisms, and as a consequence been interpreted in terms of geographical or physical constraints. If paleoclimatic legacies also occurred at the regional scale in the distributions of vagile organisms within biomes, they would rather suggest behavioral constraints on dispersal, i.e., philopatric syndromes. We examined 1) the residuals of the regression between contemporary energy and passerine species richness in South African biomes and 2) phylogenetic dispersion of passerine assemblages, using occupancy models and quarter-degree resolution citizen science data. We found a northeast to southwest gradient within mesic biomes congruent with the location of Quaternary mesic refugia, overall suggesting that as distance from refugia increased, more clades were lacking from local assemblages. A similar but weaker pattern was detected in the arid Karoo Biomes. In mobile organisms such as birds, behavioral constraints on dispersal appear strong enough to influence species distributions thousands of years after historical range contractions.