Modelling Ecological Presence–absence Data along an Environmental Gradient: Threshold Levels of the Environment

A methodology for estimating environmental thresholds of binary presence–absence data is presented where the level of the threshold is parameterised. Presence–absence data is fitted to three complementary different models: an independent null-model, a monotonically increasing or decreasing model, and an optimum model. The range of the three models is strictly between zero and one and the models are therefore well suited for modelling presence probabilities. The results of the three models may be combined by using Bayesian model selection methodologies. The proposed methodology is exemplified on observed binary presence–absence data of Bauera rubioides along an elevation gradient. Received: May 2005 / Revised: July 2005 An erratum to this article is available at.     

Avoiding Pitfalls of Using Species Distribution Models in Conservation Planning

Abstract:   Museum records have great potential to provide valuable insights into the vulnerability, historic distribution, and conservation of species, especially when coupled with species-distribution models used to predict species' ranges. Yet, the increasing dependence on species-distribution models in identifying conservation priorities calls for a more critical evaluation of model robustness. We used 11 bird species of conservation concern in Brazil's highly fragmented Atlantic Forest and data on environmental conditions in the region to predict species distributions. These predictions were repeated for five different model types for each of the 11 bird species. We then combined these species distributions for each model separately and applied a reserve-selection algorithm to identify priority sites. We compared the potential outcomes from the reserve selection among the models. Although similarity in identification of conservation reserve networks occurred among models, models differed markedly in geographic scope and flexibility of reserve networks. It is essential for planners to evaluate the conservation implications of false-positive and false-negative errors for their specific management scenario before beginning the modeling process. Reserve networks selected by models that minimized false-positive errors provided a better match with priority areas identified by specialists. Thus, we urge caution in the use of models that overestimate species' occurrences because they may misdirect conservation action. Our approach further demonstrates the great potential value of museum records to biodiversity studies and the utility of species-distribution models to conservation decision-making. Our results also demonstrate, however, that these models must be applied critically and cautiously.     

Predicting tree distributions in an East African biodiversity hotspot: model selection, data bias and envelope uncertainty

The Eastern Arc Mountains (EAMs) of Tanzania and Kenya support some of the most ancient tropical rainforest on Earth. The forests are a global priority for biodiversity conservation and provide vital resources to the Tanzanian population. Here, we make a first attempt to predict the spatial distribution of 40 EAM tree species, using generalised additive models, plot data and environmental predictor maps at sub 1 km resolution. The results of three modelling experiments are presented, investigating predictions obtained by (1) two different procedures for the stepwise selection of predictors, (2) down-weighting absence data, and (3) incorporating an autocovariate term to describe fine-scale spatial aggregation. In response to recent concerns regarding the extrapolation of model predictions beyond the restricted environmental range of training data, we also demonstrate a novel graphical tool for quantifying envelope uncertainty in restricted range niche-based models (envelope uncertainty maps). We find that even for species with very few documented occurrences useful estimates of distribution can be achieved. Initiating selection with a null model is found to be useful for explanatory purposes, while beginning with a full predictor set can over-fit the data. We show that a simple multimodel average of these two best-model predictions yields a superior compromise between generality and precision (parsimony). Down-weighting absences shifts the balance of errors in favour of higher sensitivity, reducing the number of serious mistakes (i.e., falsely predicted absences); however, response functions are more complex, exacerbating uncertainty in larger models. Spatial autocovariates help describe fine-scale patterns of occurrence and significantly improve explained deviance, though if important environmental constraints are omitted then model stability and explanatory power can be compromised. We conclude that the best modelling practice is contingent both on the intentions of the analyst (explanation or prediction) and on the quality of distribution data; generalised additive models have potential to provide valuable information for conservation in the EAMs, but methods must be carefully considered, particularly if occurrence data are scarce. Full results and details of all species models are supplied in an online Appendix.     

Sensitivity of Systematic Reserve Selection to Decisions about Scale, Biological Data, and Targets: Case Study from Southern British Columbia

Abstract:  The identification of conservation areas based on systematic reserve-selection algorithms requires decisions related to both spatial and ecological scale. These decisions may affect the distribution and number of sites considered priorities for conservation within a region. We explored the sensitivity of systematic reserve selection by altering values of three essential variables. We used a 1:20,000–scale terrestrial ecosystem map and habitat suitability data for 29 threatened vertebrate species in the Okanagan region of British Columbia, Canada. To these data we applied a reserve-selection algorithm to select conservation sites while altering selection unit size and shape, features of biodiversity (i.e., vertebrate species), and area conservation targets for each biodiversity feature. The spatial similarity, or percentage overlap, of selected sets of conservation sites identified (1) with different selection units was ≤40%, (2) with different biodiversity features was 59%, and (3) with different conservation targets was ≥94%. Because any selected set of sites is only one of many possible sets, we also compared the conservation value (irreplaceability) of all sites in the region for each variation of the data. The correlations of irreplaceability were weak for different selection units (0.23 ≤ r ≤ 0.67), strong for different biodiversity features ( r = 0.84), and mixed for different conservation targets ( r = 0.16; 0.16; 1.00). Because of the low congruence of selected sites and weak correlations of irreplaceability for different selection units, recommendations from studies that have been applied at only one spatial scale must be considered cautiously.     

Identifying Conservation Areas on the Basis of Alternative Distribution Data Sets

Abstract:  Distribution data on biodiversity features is a major component of conservation planning that are often inaccurate; thus, the true distribution of each feature is commonly over- or underrepresented. The selection of distribution data sets may therefore lead to variability in the spatial configuration and size of proposed reserve networks and uncertainty regarding the extent to which these networks actually contain the biodiversity features they were identified to protect. Our goals were to investigate the impact on reserve selection of choosing different distribution data sets and to propose novel methods to minimize uncertainty about target attainment within reserves. To do so, we used common prioritization methods (richness mapping, systematic reserve design, and a novel approach that integrates multiple types of distribution data) and three types of data on the distribution of mammals (predicted distribution models, occurrence records, and a novel combination of the two) to simulate the establishment of regional biodiversity reserves for the state of Arizona (U.S.A.). Using the results of these simulations, we explored variability in reserve placement and size as a function of the distribution data set. Spatial overlap of reserve networks identified with only predicted distribution data or only occurrence distribution data never exceeded 16%. In pairwise comparisons between reserves created with all three types of distribution data, overlap never achieved 50%. The reserve size required to meet conservation targets also varied with the type of distribution data used and the conservation goal; the largest reserve system was 10 times the smallest. Our results highlight the impact of employing different types of distribution data and identify novel tools for application to existing distribution data sets that can minimize uncertainty about target attainment.     

The Boundary-Quality Penalty: a Quantitative Method for Approximating Species Responses to Fragmentation in Reserve Selection

Abstract:  Aggregation of reserve networks is generally considered desirable for biological and economic reasons: aggregation reduces negative edge effects and facilitates metapopulation dynamics, which plausibly leads to improved persistence of species. Economically, aggregated networks are less expensive to manage than fragmented ones. Therefore, many reserve-design methods use qualitative heuristics, such as distance-based criteria or boundary-length penalties to induce reserve aggregation. We devised a quantitative method that introduces aggregation into reserve networks. We call the method the boundary-quality penalty (BQP) because the biological value of a land unit (grid cell) is penalized when the unit occurs close enough to the edge of a reserve such that a fragmentation or edge effect would reduce population densities in the reserved cell. The BQP can be estimated for any habitat model that includes neighborhood (connectivity) effects, and it can be introduced into reserve selection software in a standardized manner. We used the BQP in a reserve-design case study of the Hunter Valley of southeastern Australia. The BQP resulted in a more highly aggregated reserve network structure. The degree of aggregation required was specified by observed (albeit modeled) biological responses to fragmentation. Estimating the effects of fragmentation on individual species and incorporating estimated effects in the objective function of reserve-selection algorithms is a coherent and defensible way to select aggregated reserves. We implemented the BQP in the context of the Zonation method, but it could as well be implemented into any other spatially explicit reserve-planning framework .     

Conservation Goals and the Relative Importance of Costs and Benefits in Reserve Selection

Abstract: Including both economic costs and biological benefits of sites in systematic reserve selection greatly increases cost‐efficiency. Nevertheless, limited funding generally forces conservation planners to choose which data to focus the most resources on; therefore, the relative importance of different types of data must be carefully assessed. We investigated the relative importance of including information about costs and benefits for 3 different commonly used conservation goals: 2 in which biological benefits were measured per site (species number and conservation value scores) and 1 in which benefits were measured on the basis of site complementarity (total species number in the reserve network). For each goal, we used site‐selection models with data on benefits only, costs only, and benefits and costs together, and we compared the efficiency of each model. Costs were more important to include than benefits for the goals in which benefits were measured per site. By contrast, for the complementarity‐based goal, benefits were more important to include. To understand this pattern, we compared the variability in benefits and in costs for each goal. By comparing the best and the worst possible selection of sites with regard to costs alone and benefits alone for each conservation goal, we introduced a simple and consistent variability measure that is applicable to all kinds of reserve‐selection situations. In our study, benefit variability depended strongly on how the conservation goal was formulated and was largest for the complementarity‐based conservation goal. We argue that from a cost‐efficiency point of view, most resources should be spent on collecting the most variable type of data for the conservation goal at hand.     

Use of Coarse‐Resolution Models of Species’ Distributions to Guide Local Conservation Inferences

Abstract: Distribution models are used increasingly for species conservation assessments over extensive areas, but the spatial resolution of the modeled data and, consequently, of the predictions generated directly from these models are usually too coarse for local conservation applications. Comprehensive distribution data at finer spatial resolution, however, require a level of sampling that is impractical for most species and regions. Models can be downscaled to predict distribution at finer resolutions, but this increases uncertainty because the predictive ability of models is not necessarily consistent beyond their original scale. We analyzed the performance of downscaled, previously published models of environmental favorability (a generalized linear modeling technique) for a restricted endemic insectivore, the Iberian desman (Galemys pyrenaicus), and a more widespread carnivore, the Eurasian otter (Lutra lutra), in the Iberian Peninsula. The models, built from presence–absence data at 10 × 10 km resolution, were extrapolated to a resolution 100 times finer (1 × 1 km). We compared downscaled predictions of environmental quality for the two species with published data on local observations and on important conservation sites proposed by experts. Predictions were significantly related to observed presence or absence of species and to expert selection of sampling sites and important conservation sites. Our results suggest the potential usefulness of downscaled projections of environmental quality as a proxy for expensive and time‐consuming field studies when the field studies are not feasible. This method may be valid for other similar species if coarse‐resolution distribution data are available to define high‐quality areas at a scale that is practical for the application of concrete conservation measures.     

Habitat Suitability Models and the Shortfall in Conservation Planning for African Vertebrates

Abstract:  Ongoing loss of biodiversity requires identifying large-scale conservation priorities, but the detailed information on the distribution of species required for this purpose is often missing. We present a systematic reserve selection for 1223 African mammals and amphibians in which habitat suitability models are used as estimates of the area occupied by species. In the framework of the World Conservation Union (IUCN) Global Amphibian Assessment and IUCN Global Mammal Assessment, we collected the geographic range (extent of occurrence) and habitat preferences for each species. We used the latter to build species-specific habitat suitability models inside geographic ranges, and for 181 species we verified the models by comparing suitability levels to presence-absence data collected in the field. We then used the suitable areas as estimators of the area of occupancy and compared the results of systematic reserve selection based on geographic ranges to those based on estimated areas of occupancy. Our results showed that the reserve system would need a 30-100% expansion to achieve minimal conservation targets, concentrated in the tropics, where species richness reaches a maximum. Comparative analyses revealed that using geographic ranges, which overestimate the area occupied by species, underestimates the total amount of area that needs to be conserved. The area selected for conservation doubled when we used the estimated area of occupancy in place of the geographic ranges. This happened because the suitable areas potentially occupied by each species overlapped less than their geographic ranges. As a result, any given protected area contained fewer species than predicted by the analysis of ranges. Because species are more specialized than our estimates of distribution based on extent of occurrence suggest, we propose that this is a general effect in systematic conservation planning.     

The pattern of species turnover resulting from stochastic population dynamics: The model and field data

The model of random population dynamics in a sampling site returns geometric distribution of longevities of continuous presence (=persistence) and Poisson distribution of the presence–absence transitions. This discrete-time stochastic process describes the presence–absence pattern observed in the beetles surveyed 6 years on Mount Carmel, Israel. Homogeneous pools of species mostly on the Families rank, exhibit the predicted by the model patterns. Conformity to an ergodic hypothesis is the criterion of ecological homogeneity. This criterion assumes the equivalence of short-term behavior of entire pool and long-term behavior of any species from this pool. The pool of all 801 species of Order Coleoptera does not match the model. Thus a taxon of an arbitrary rank may not be considered a priory as a unit of ecological study. Determined from field data parameters of the model are biased and magnitude of the bias depends on longevity of the survey. Parameter of distribution depends also on species tolerance, which is the level adaptation of given species to given environment in given time interval. Random process of species turnover may be considered as a game of species to gain their presence against deteriorative fluctuations of environmental conditions.     

Incorporating consumer–resource spatial interactions in reserve design

The persistence of species in reserves depends in large part on the persistence of functional ecological interactions. Despite their importance, however, ecological interactions have not yet been explicitly incorporated into conservation prioritization methods. We develop here a general method for incorporating consumer–resource interactions into spatial reserve design. This method protects spatial consumer–resource interactions by protecting areas that maintain the connectivity between the distribution of consumers and resources. We illustrate our method with a conservation planning case study of a mammalian predator, American marten (Martes americana), and its two primary prey species, Red-backed vole (Clethrionomys rutilus) and Deer mouse (Peromyscus maniculatus). The conservation goal was to identify a reserve for marten that comprised 12% of a forest management unit in the boreal forest in Québec, Canada. We compared reserves developed using analysis variants that utilized different levels of information about predator and prey habitat distributions, species-specific connectivity requirements, and interaction connectivity requirements. The inclusion of consumer–resource interactions in reserve-selection resulted in spatially aggregated reserves that maintained local habitat quality for the species. This spatial aggregation was not induced by applying a qualitative penalty for the boundary length of the reserve, but rather was a direct consequence of modelling the spatial needs of the interacting consumer and resources. Our method for maintaining connectivity between consumers and their resources within reserves can be applied even under the most extreme cases of either complete spatial overlap or complete spatial segregation of consumer–resource distributions. The method has been made available via public software.     

Effectiveness of Vegetation Surrogates for Parasitoid Wasps in Reserve Selection

Abstract:  Selecting suitable nature reserves is a continuing challenge in conservation, particularly for target groups that are time-consuming to survey, species rich, and extinction prone. One such group is the parasitoid Hymenoptera, which have been excluded from conservation planning. If basic characteristics of habitats or vegetation could be used as reliable surrogates of specific target taxa, this would greatly facilitate appropriate reserve selection. We identified a range of potential habitat indicators of the species richness of pimpline parasitoid communities (Hymenoptera: Ichneumonidae: Pimplinae, Diacritinae, Poemeniinae) and tested their efficiency at capturing the observed diversity in a group of small woodlands in the agricultural landscape of the Vale of York (United Kingdom). Eight of the 18 vegetation-based reserve-selection strategies were significantly better at parasitoid species inclusion than random selection of areas. The best strategy maximized richness of tree species over the entire reserve network through complementarity. This strategy omitted only 2–3 species more (out of 38 captured in the landscape as a whole) than selections derived from the parasitoid survey data. In general, strategies worked equally well at capturing species richness and rarity. Our results suggest that vegetation data as a surrogate for species richness could prove an informative tool in parasitoid conservation, but further work is needed to test how broadly applicable these indicators may be.     

Finding all optimal solutions to the reserve site selection problem: formulation and computational analysis

The problem of selecting nature reserves has received increased attention in the literature during the past decade, and a variety of approaches have been promoted for selecting those sites to include in a reserve network. One set of techniques employs heuristic algorithms and thus provides possibly sub-optimal solutions. Another set of models and accompanying algorithms uses an integer programming formulation of the problem, resulting in an optimization problem known as the Maximal Covering Problem, or MCP. Solution of the MCP provides an optimal solution to the reserve site selection problem, and while various algorithms can be employed for solving the MCP they all suffer from the disadvantage of providing a single optimal solution dictating the selection of areas for conservation. In order to provide complete information to decision makers, the determination of all alternate optimal solutions is necessary. This paper explores two procedures for finding all such solutions. We describe the formulation and motivation of each method. A computational analysis on a data set describing native terrestrial vertebrates in the state of Oregon illustrates the effectiveness of each approach.     

How decisions about fitting species distribution models affect conservation outcomes

Species distribution models (SDMs) are increasingly used in conservation and land-use planning as inputs to describe biodiversity patterns. These models can be built in different ways, and decisions about data preparation, selection of predictor variables, model fitting, and evaluation all alter the resulting predictions. Commonly, the true distribution of species is unknown and independent data to verify which SDM variant to choose are lacking. Such model uncertainty is of concern to planners. We analyzed how 11 routine decisions about model complexity, predictors, bias treatment, and setting thresholds for predicted values altered conservation priority patterns across 25 species. Models were created with MaxEnt and run through Zonation to determine the priority rank of sites. Although all SDM variants performed well (area under the curve >0.7), they produced spatially different predictions for species and different conservation priority solutions. Priorities were most strongly altered by decisions to not address bias or to apply binary thresholds to predicted values; on average 40% and 35%, respectively, of all grid cells received an opposite priority ranking. Forcing high model complexity altered conservation solutions less than forcing simplicity (14% and 24% of cells with opposite rank values, respectively). Use of fewer species records to build models or choosing alternative bias treatments had intermediate effects (25% and 23%, respectively). Depending on modeling choices, priority areas overlapped as little as 10–20% with the baseline solution, affecting top and bottom priorities differently. Our results demonstrate the extent of model-based uncertainty and quantify the relative impacts of SDM building decisions. When it is uncertain what the best SDM approach and conservation plan is, solving uncertainty or considering alterative options is most important for those decisions that change plans the most.     

Traditional occupancy–abundance models are inadequate for zero-inflated ecological count data

Traditional occupancy–abundance and abundance–variance–occupancy models do not take into account zero-inflation, which occurs when sampling rare species or in correlated counts arising from repeated measures. In this paper we propose a novel approach extending occupancy–abundance relationships to zero-inflated count data. This approach involves three steps: (1) selecting distributional assumptions and parsimonious models for the count data, (2) estimating abundance, occupancy and variance parameters as functions of site- and/or time-specific covariates, and (3) modelling the occupancy–abundance relationship using the parameters estimated in step 2. Five count datasets were used for comparing standard Poisson and negative binomial distribution (NBD) occupancy–abundance models. Zero-inflated Poisson (ZIP) and zero-inflated negative binomial (ZINB) occupancy–abundance models were introduced for the first time, and these were compared with the Poisson, NBD, He and Gaston's and Wilson and Room's abundance–variance–occupancy models. The percentage of zero counts ranged from 45 to 80% in the datasets analysed. For most of the datasets, the ZINB occupancy–abundance model performed better than the traditional Poisson, NBD and Wilson and Room's model. He and Gaston's model performed better than the ZINB in two out of the five datasets. However, the occupancy predicted by all models increased faster than the observed as density increased resulting in significant mismatch at the highest densities. Limitations of the various models are discussed, and the need for careful choice of count distributions and predictors in estimating abundance and occupancy parameter are indicated.     

A Nationwide Assessment of the Biodiversity Value of Uganda's Important Bird Areas Network

Abstract:  BirdLife International's Important Bird Areas (IBA) program is the most developed global system for identifying sites of conservation priority. There have been few assessments, however, of the conservation value of IBAs for nonavian taxa. We combined past data with extensive new survey results for Uganda's IBAs in the most comprehensive assessment to date of the wider biodiversity value of a tropical country's IBA network. The combined data set included more than 35,000 site × species records for birds, butterflies, and woody plants at 86 Ugandan sites (23,400 km²), including 29 of the country's 30 IBAs, with data on additional taxa for many sites. Uganda's IBAs contained at least 70% of the country's butterfly and woody plant species, 86% of its dragonflies and 97% of its birds. They also included 21 of Uganda's 22 major vegetation types. For butterflies, dragonflies, and some families of plants assessed, species of high conservation concern were well represented (less so for the latter). The IBAs successfully represented wider biodiversity largely because many have distinctive avifaunas and, as shown by high cross-taxon congruence in complementarity, such sites tended to be distinctive for other groups too. Cross-taxon congruence in overall species richness was weaker and mainly associated with differences in site size. When compared with alternative sets of sites selected using complementarity-based, area-based, or random site-selection algorithms, the IBA network was efficient in terms of the number of sites required to represent species but inefficient in terms of total area. This was mainly because IBA selection considers factors other than area, however, which probably improves both the cost-effectiveness of the network and the persistence of represented species.     

Variance and Uncertainty in the Expected Number of Occurrences in Reserve Selection

Abstract:  Reserve selection often concerns the design of reserve networks for the long-term maintenance of biodiversity. We considered uncertainty in the context of three common reserve-selection formulations, the expected number of populations, proportional coverage of land-cover types, and the probability of having at least one population. By uncertainty, we mean variance in the outcome of any probability-based reserve selection formulation. A typical reserve-selection formulation might ask for the least expensive set of sites that contains n populations per species. It is implicit here that this requirement concerns the expected number of populations, which actually is obtained only with a 50% chance. If the requirement is changed to select the least expensive set of sites that gives n populations per species with a 95% probability, the number of sites required in the solution increases and the identity of the sites is changed toward sites that have high probabilities of persistence (or occurrence) and low associated binomial variance. Anthropogenic threat is one factor that may cause probabilistic uncertainty in the context of proportional area coverage.     

Assigning trend-based conservation status despite high uncertainty

Estimates of temporal trends in species’ occupancy are essential for conservation policy and planning, but limitations to the data and models often result in very high trend uncertainty. A critical source of uncertainty that degrades scientific credibility is that caused by disagreement among studies or models. Modelers are aware of this uncertainty but usually only partially estimate it and communicate it to decision makers. At the same time, there is growing awareness that full disclosure of uncertainty is critical for effective translation of science into policies and plans. But what are the most effective approaches to estimating uncertainty and communicating uncertainty to decision makers? We explored how alternative approaches to estimating and communicating uncertainty of species trends could affect decisions concerning conservation status of freshwater fishes. We used ensemble models to propagate trend uncertainty within and among models and communicated this uncertainty with categorical distributions of trend direction and magnitude. All approaches were designed to fit an established decision-making system used to assign species conservation status by the New Zealand government. Our results showed how approaches that failed to fully disclose uncertainty, while simplifying the information presented, could hamper species conservation or lead to ineffective decisions. We recommend an approach that was recently used effectively to communicate trend uncertainty to a panel responsible for setting the conservation status of New Zealand's freshwater fishes.     

Predictive vegetation modeling for conservation: impact of error propagation from digital elevation data.

The effect of digital elevation model (DEM) error on environmental variables, and subsequently on predictive habitat models, has not been explored. Based on an error analysis of a DEM, multiple error realizations of the DEM were created and used to develop both direct and indirect environmental variables for input to predictive habitat models. The study explores the effects of DEM error and the resultant uncertainty of results on typical steps in the modeling procedure for prediction of vegetation species presence/absence. Results indicate that all of these steps and results, including the statistical significance of environmental variables, shapes of species response curves in generalized additive models (GAMs), stepwise model selection, coefficients and standard errors for generalized linear models (GLMs), prediction accuracy (Cohen's kappa and AUC), and spatial extent of predictions, were greatly affected by this type of error. Error in the DEM can affect the reliability of interpretations of model results and level of accuracy in predictions, as well as the spatial extent of the predictions. We suggest that the sensitivity of DEM-derived environmental variables to error in the DEM should be considered before including them in the modeling processes.