Landscape evaluation in conservation: molecular sampling and habitat modeling for the Iberian lynx.

Conservation of endangered species requires comprehensive understanding of their distribution and habitat requirements, in order to implement better management strategies. Unfortunately, this understanding is often difficult to gather at the short term required by rapidly declining populations of many rare vertebrates. We present a spatial habitat modeling approach that integrates a molecular technique for species detection with landscape information to assess habitat requirements of a critically endangered mammalian carnivore, the Iberian lynx (Lynx pardinus), in a poorly known population in Spain. We formulated a set of model hypotheses for habitat selection at the spatial scale of home ranges, based on previous information on lynx requirements of space, vegetation, and prey. To obtain the required data for model selection, we designed a sampling protocol based on surveys of feces and their molecular analysis for species identification. After comparing candidate models, we selected a parsimonious one that allowed (1) reliable assessment of lynx habitat requirements at the scale of home ranges, (2) prediction of lynx distribution and potential population size, and (3) identification of landscape management priorities for habitat conservation. This model predicted that the species was more likely to occur in landscapes with a higher percentage of rocky areas and higher cover of bushes typical of mature mediterranean shrubland mosaics. Its accuracy for discriminating lynx presence was approximately 85%, indicating high predictive performance. Mapping model predictions showed that only 16% of the studied areas constitute potential habitat for lynx, even though the region is dominated by large extents of well-preserved native vegetation with low human interference. Habitat was mostly clumped in two nearby patches connected by vegetation adequate for lynx dispersal and had a capacity for 28-62 potential breeding territories. The lynx population in Sierra Morena is probably the largest persisting today, but it is still critically small for optimism about its long-term persistence. Model results suggest habitat conservation and restoration actions needed for preserving the species, including reconciliation of hunting management with preservation of mature shrubland over large areas (particularly in rocky landscapes). The approach presented here can be applied to many other species for which the ecological information needed to develop sound habitat conservation strategies is lacking.     

Habitat selection increases the isolating effect of habitat fragmentation on the gopher tortoise

Habitat selection requires choice, which differentiates it from habitat use, and choice, in turn, is dependent upon the responses of organisms to the environmental, social, and other cues that they perceive. Habitat selection by the gopher tortoise (Gopherus polyphemus) was investigated by translocating tortoises and monitoring their movements within two sites in central Florida. The first site supported a stable preponderance of high-quality habitat, and tortoises avoided areas with a dense tree canopy cover caused by fire exclusion. The second site was badly invaded by an introduced weed, and tortoises avoided areas where the weed had formed a dense monoculture. At both sites, individuals appeared to be responding to visual cues to avoid areas that were relatively dark. In landscapes with relatively large amounts of high-quality habitat, this avoidance behavior serves the gopher tortoise well by keeping individuals within the dominant habitat type. In degraded areas, high-quality habitat often becomes increasingly uncommon, and the avoidance behavior exhibited by the tortoises will result in individuals becoming confined to small patches, causing a significant reduction in fitness and hence questioning their long-term survival in such areas. The results from our study show that in order to maintain viable tortoise populations in areas increasingly subjected to human fragmentation and degradation, it is crucial not only to suppress tree canopy cover continually and prevent invasion by exotic weeds, but also to be mindful that the avoidance behavior of the gopher tortoise could prevent individuals from fully occupying a high-quality habitat in response to restoration and management efforts.     

Multistate modeling of habitat dynamics: factors affecting Florida scrub transition probabilities

Many ecosystems are influenced by disturbances that create specific successional states and habitat structures that species need to persist. Estimating transition probabilities between habitat states and modeling the factors that influence such transitions have many applications for investigating and managing disturbance-prone ecosystems. We identify the correspondence between multistate capture-recapture models and Markov models of habitat dynamics. We exploit this correspondence by fitting and comparing competing models of different ecological covariates affecting habitat transition probabilities in Florida scrub and flatwoods, a habitat important to many unique plants and animals. We subdivided a large scrub and flatwoods ecosystem along central Florida's Atlantic coast into 10-ha grid cells, which approximated average territory size of the threatened Florida Scrub-Jay (Aphelocoma coerulescens), a management indicator species. We used 1.0-m resolution aerial imagery for 1994, 1999, and 2004 to classify grid cells into four habitat quality states that were directly related to Florida Scrub-Jay source-sink dynamics and management decision making. Results showed that static site features related to fire propagation (vegetation type, edges) and temporally varying disturbances (fires, mechanical cutting) best explained transition probabilities. Results indicated that much of the scrub and flatwoods ecosystem was resistant to moving from a degraded state to a desired state without mechanical cutting, an expensive restoration tool. We used habitat models parameterized with the estimated transition probabilities to investigate the consequences of alternative management scenarios on future habitat dynamics. We recommend this multistate modeling approach as being broadly applicable for studying ecosystem, land cover, or habitat dynamics. The approach provides maximum-likelihood estimates of transition parameters, including precision measures, and can be used to assess evidence among competing ecological models that describe system dynamics.     

Using forest inventory data to assess fisher resting habitat suitability in California.

The fisher (Martes pennanti) is a forest-dwelling carnivore whose current distribution and association with late-seral forest conditions make it vulnerable to stand-altering human activities or natural disturbances. Fishers select a variety of structures for daily resting bouts. These habitat elements, together with foraging and reproductive (denning) habitat, constitute the habitat requirements of fishers. We develop a model capable of predicting the suitability of fisher resting habitat using standard forest vegetation inventory data. The inventory data were derived from Forest Inventory and Analysis (FIA), a nationwide probability-based sample used to estimate forest characteristics. We developed the model by comparing vegetation and topographic data at 75 randomly selected fisher resting structures in the southern Sierra Nevada with 232 forest inventory plots. We collected vegetation data at fisher resting locations using the FIA vegetation sampling protocol and centering the 1-ha FIA plot on the resting structure. To distinguish used and available inventory plots, we used nonparametric logistic regression to evaluate a set of a priori biological models. The top model represented a dominant portion of the Akaike weights (0.87), explained 31.5% of the deviance, and included the following variables: average canopy closure, basal area of trees <51 cm diameter breast height (dbh), average hardwood dbh, maximum tree dbh, percentage slope, and the dbh of the largest conifer snag. Our use of routinely collected forest inventory data allows the assessment and monitoring of change in fisher resting habitat suitability over large regions with no additional sampling effort. Although models were constrained to include only variables available from the list of those measured using the FIA protocol, we did not find this to be a shortcoming. The model makes it possible to compare average resting habitat suitability values before and after forest management treatments, among administrative units, across regions and over time. Considering hundreds of plot estimates as a sample of habitat conditions over large spatial scales can bring a broad perspective, at high resolution, and efficiency to the assessment and monitoring of wildlife habitat.     

Hierarchical modeling of count data with application to nuclear fall-out

We propose a method for a Bayesian hierarchical analysis of count data that are observed at irregular locations in a bounded domain of R². We model the data as having been observed on a fine regular lattice, where we do not have observations at all the sites. The counts are assumed to be independent Poisson random variables whose means are given by a log Gaussian process. In this article, the Gaussian process is assumed to be either a Markov random field (MRF) or a geostatistical model, and we compare the two models on an environmental data set. To make the comparison, we calibrate priors for the parameters in the geostatistical model to priors for the parameters in the MRF. The calibration is obtained empirically. The main goal is to predict the hidden Poisson-mean process at all sites on the lattice, given the spatially irregular count data; to do this we use an efficient MCMC. The spatial Bayesian methods are illustrated on radioactivity counts analyzed by Diggle et al. (1998).     

Comparing geographic boundaries in songbird demography data with vegetation boundaries: a new approach to evaluating habitat quality

To plan for the habitat needs of forest songbirds of conservation concern, managers need to understand how spatial heterogeneity in forest conditions influences habitat quality. I used difference boundary detection (wombling) and spatially constrained clustering to delineate boundaries in various combinations of four forest vegetation variables (understory height, understory density, percent deciduous vs. conifer understory, and percent canopy closure) in two Michigan northern hardwood forests. My goal was to identify vegetation boundaries that corresponded with boundaries in an understory-dependent songbird’s distribution, and with boundaries in demographic measures for this songbird that indicate habitat quality (e.g., occupancy by older vs. yearling males, reproductive success). Both forests were actively-managed, mature stands: The first site (78 ha) was heavily deer-browsed (HB), with many browse-resistant conifers in the understory, and the second (62 ha) was less-browsed (LB), with deciduous-dominated understory. I compared the vegetation difference and cluster boundaries to difference boundaries based on 6 years of distribution and demographic data for black-throated blue warblers (Dendroica caerulescens). At the HB site, warbler boundaries overlapped strongly with vegetation boundaries that included all four variables, and clustering effectively divided the habitat into areas with different warbler occupancy and demographic characteristics. At the LB site, warbler distribution showed high overlap with difference and cluster boundaries based on just the height and density of understory vegetation, and cluster boundaries again effectively partitioned the study area into sites that varied in habitat quality. Thus, geographic boundary analysis is likely to be a useful tool for identifying key vegetation variables for management, and for delineating clusters (habitat patches) within sites that capture differences in habitat quality.

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.     

Improving seabed classification from Multi-Beam Echo Sounder (MBES) backscatter data with visual data mining

Multi-Beam Echo Sounders are often used for classification of seabed type, as there exists a strong link between sonar backscatter and sediment characteristics of the seabed. Most of the methods for seabed classification from MBES backscatter create a highly-dimensional data set of statistical features and then use a combination of Principal Component Analysis and k-means clustering to derive classes. This procedure can be time consuming for contemporary large MBES data sets with millions of records. This paper examines the complexity of one of most commonly used classification approaches and suggests an alternative where feature data set is optimised in terms of dimensionality using computational and visual data mining. Both the original and the optimised method are tested on an MBES backscatter data set and validated against ground truth. The study found that the optimised method improves accuracy of classification and reduced complexity of processing. This is an encouraging result, which shows that bringing together methods from acoustic classification, visual data mining, spatial analysis and remote sensing can support the unprecedented increases in data volumes collected by contemporary acoustic sensors.     

Hierarchical Bayesian strategy for modeling correlated compositional data with observed zero counts

This article proposes a hierarchical multivariate conditional autoregressive model applied to a compositional response vector. We particularly focus on situations when the composition is discrete occurring when observations are based on small multinomial counts. We address drawbacks that exist in current modeling approaches for such data. Our hierarchical model will be demonstrated with data used to help manage a commercial sockeye salmon fishery in the Fraser River of British Columbia.     

Impacts of cyclone Hamish at One Tree Reef: integrating environmental and benthic habitat data

The southern Great Barrier Reef (GBR), a region that rarely experiences cyclones, was impacted by tropical cyclone (TC) Hamish in March 2009. We documented on-reef physical and habitat conditions before, during and after the cyclone at One Tree Reef (OTR) using data from environmental sensor instrumentation and benthic surveys. Over 5 years of monitoring, ocean mooring data revealed that OTR experienced large swells (4–8 m) of short duration (10–20 min) not associated with a cyclone in the area. These swells may have contributed to the physical disturbance of benthic biota and decline in coral cover recorded prior to and after TC Hamish. During the cyclone, OTR sustained southeasterly gale force winds (>61.2 km h⁻¹) for 18.5 h and swells >6 m in height for 4 h. Benthic surveys of exposed sites documented a 20% drop in live coral cover, 30% increase in filamentous algae cover and the presence of dislodged corals and rubble after the storm. Leeward sites were largely unaffected by the cyclone. Benthic cover did not change in the lagoon sites. Significant rubble movement and infill of the lagoon occurred. Two years after the cyclone, algal cover remained high and laminar corals had not recovered. Total coral cover at impacted sites had continued to decline. Environmental conditions and habitat surveys supported Puotinen’s (Int J Geogr Inf Sci 21:97–120, 2007) model for cyclone conditions that cause reef destruction. While TC Hamish had a major impact on the reef, change in benthic cover over several years was due to multiple stressors. This on-reef scale integration of physical and biological data provided a rare opportunity to assess impacts of a major storm and other disturbances, showing the importance of considering multiple stressors (short-lived and sustained) in assessing change to reef habitats.     

Making more out of sparse data: hierarchical modeling of species communities

Community ecologists and conservation biologists often work with data that are too sparse for achieving reliable inference with species-specific approaches. Here we explore the idea of combining species-specific models into a single hierarchical model. The community component of the model seeks for shared patterns in how the species respond to environmental covariates. We illustrate the modeling framework in the context of logistic regression and presence-absence data, but a similar hierarchical structure could also be used in many other types of applications. We first use simulated data to illustrate that the community component can improve parameterization of species-specific models especially for rare species, for which the data would be too sparse to be informative alone. We then apply the community model to real data on 500 diatom species to show that it has much greater predictive power than a collection of independent species-specific models. We use the modeling approach to show that roughly one-third of distance decay in community similarity can be explained by two variables characterizing water quality, rare species typically preferring nutrient-poor waters with high pH, and common species showing a more general pattern of resource use.     

Waterfowls habitat modeling: Simulation of nest site selection for the migratory Little Tern (Sterna albifrons) in the Nakdong estuary

This paper aims to find patterns in nest site selection by Little Terns Sterna albifrons, in the Nakdong estuary in South Korea. This estuary is important waterfowl stopover and breeding habitat, located in the middle of the East Asia-Australasian Flyway. The Little Tern is a common species easily observed near the seashore but their number is gradually declining around the world. We investigated their nests and eggs on a barrier islet in the Nakdong estuary during the breeding season (May to June, 2007), and a pattern for the nest site selection was identified using genetic programming (GP). The GP generated a predictive rule-set model for the number of Little Tern nests (training: R² = 0.48 and test: 0.46). The physical features of average elevation, variation of elevation, plant coverage, and average plant height were estimated to determine the influence on nest numbers for Little Tern. A series of sensitivity analyses stressed that mean elevation and vegetation played an important role in nest distribution for Little Tern. The influence of these two variables could be maximized when elevation changed moderately within the sampled quadrats. The study results are regarded as a good example of applying GP to vertebrate distribution patterning and prediction with several important advantages compared to conventional modeling techniques, and can help establish a management or restoration strategy for the species.     

A decision-support system in ICZM for protecting the ecosystems: integration with the habitat directive

This paper describes a decision-support system based on landscape ecology and focused on the study of ecosystems’ health. System capabilities are illustrated with three cases of integrated coastal zone management (ICZM), in the Adriatic Sea (Italy): the lagoon of Venice and the Rimini and Ancona coastal areas. Indicators and indices are developed with a focus on sub-regional and local problems in coastal management, with a multi-scale approach based on landscape and seascape ecology. Land-use changes of the coastal areas were detected by analyzing two sets of satellite images. Indices combining satellite imagery, socio-economic and environmental indicators, and landscape and seascape maps were created, showing ecological changes, habitat loss and gaps in conservation policy. The approach used provides means for the identification of conflicts and for the assessment of sustainability. Results show that the lagoon of Venice plays an important role in mitigating and compensating the impacts of human activities, and needs to be protected and restored. The Rimini area shows high ecological footprint and development-intensity and low biocapacity. The Ancona area needs the protection of its natural coastal space from potential sources of anthropogenic impacts to maintain its sustainability. A model of environment changes is critical for formulating effective environmental policies and management strategies. The developed decision-support system provides a suitability map per each area analyzed, which can be used in order to maximize different policy objectives and reduce coastal conflicts.     

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.     

A new procedure to optimize the selection of groups in a classification tree: Applications for ecological data

Agglomerative cluster analyses encompass many techniques, which have been widely used in various fields of science. In biology, and specifically ecology, datasets are generally highly variable and may contain outliers, which increase the difficulty to identify the number of clusters. Here we present a new criterion to determine statistically the optimal level of partition in a classification tree. The criterion robustness is tested against perturbated data (outliers) using an observation or variable with values randomly generated. The technique, called Random Simulation Test (RST), is tested on (1) the well-known Iris dataset [Fisher, R.A., 1936. The use of multiple measurements in taxonomic problems. Ann. Eugenic. 7, 179–188], (2) simulated data with predetermined numbers of clusters following Milligan and Cooper [Milligan, G.W., Cooper, M.C., 1985. An examination of procedures for determining the number of clusters in a data set. Psychometrika 50, 159–179] and finally (3) is applied on real copepod communities data previously analyzed in Beaugrand et al. [Beaugrand, G., Ibanez, F., Lindley, J.A., Reid, P.C., 2002. Diversity of calanoid copepods in the North Atlantic and adjacent seas: species associations and biogeography. Mar. Ecol. Prog. Ser. 232, 179–195]. The technique is compared to several standard techniques. RST performed generally better than existing algorithms on simulated data and proved to be especially efficient with highly variable datasets.     

Predicting the effects of habitat homogenization on marine biodiversity.

Seafloor habitats throughout the world's oceans are being homogenized by physical disturbance. Even though seafloor sediments are commonly considered to be simple and unstructured ecosystems, the negative impacts of habitat homogenization are widespread because resident organisms create much of their habitat's structure. We combine the insight gained from remote sensing of seafloor habitats with recently developed analytical techniques to estimate species richness and assess the potential for change with habitat homogenization. Using habitat-dependent species-area relationships we show that realistic scenarios of habitat homogenization predict biodiversity losses when biogenic habitats in soft sediments are homogenized. We develop a simple model that highlights the degree to which the reductions in the number of species and functional diversity are related to the distribution across habitats of habitat-specific and generalist species. Our results suggest that, by using habitat-dependent species-area relationships, we can better predict variation in biodiversity across seafloor landscapes and contribute to improved management and conservation.     

Dispersal depression with habitat fragmentation in the bog fritillary butterfly

Habitat fragmentation is expected to impose strong selective pressures on dispersal rates. However, evolutionary responses of dispersal are not self-evident, since various selection pressures act in opposite directions. Here we disentangled the components of dispersal behavior in a metapopulation context using the Virtual Migration model, and we linked their variation to habitat fragmentation in the specialist butterfly Proclossiana eunomia. Our study provided a nearly unique opportunity to study how habitat fragmentation modifies dispersal at the landscape scale, as opposed to microlandscapes or simulation studies. Indeed, we studied the same species in four landscapes with various habitat fragmentation levels, in which large amounts of field data were collected and analyzed using similar methodologies. We showed the existence of quantitative variations in dispersal behavior correlated with increased fragmentation. Dispersal propensity from habitat patches (for a given patch size), and mortality during dispersal (for a given patch connectivity) were lower in more fragmented landscapes. We suggest that these were the consequences of two different evolutionary responses of dispersal behavior at the individual level: (1) when fragmentation increased, the reluctance of individuals to cross habitat patch boundaries also increased; (2) when individuals dispersed, they flew straighter in the matrix, which is the best strategy to improve dispersal success. Such evolutionary responses could generate complex nonlinear patterns of dispersal changes at the metapopulation level according to habitat fragmentation. Due to the small size and increased isolation of habitat patches in fragmented landscapes, overall emigration rate and mortality during dispersal remained high. As a consequence, successful dispersal at the metapopulation scale remained limited. Therefore, to what extent the selection of individuals with a lower dispersal propensity and a higher survival during dispersal is able to limit detrimental effects of habitat fragmentation on dispersal success is unknown, and any conclusion that metapopulations would compensate for them is flawed.     

Strategic climate policy with offsets and incomplete abatement: Carbon taxes versus cap-and-trade

This paper provides a first analysis of a “policy bloc” of fossil fuel importers which implements an optimal climate policy, faces a (non-policy) fringe of other fuel importers, and an exporter bloc, and purchases offset from the fringe. We compare a carbon tax and a cap-and-trade scheme for the policy bloc, in either case accompanied by an efficient offset mechanism for reducing emissions in the fringe. The policy bloc is shown to prefer a tax over a cap, since only a tax reduces the fuel export price and by more when the policy bloc is larger. Offsets are also more favorable to the policy bloc under a tax than under a cap. The optimal offset price under a carbon tax is below the tax rate, while under a cap and free quota trading the offset price must equal the quota price. The domestic carbon and offset prices are both higher under a tax than under a cap when the policy bloc is small. When the policy bloc is larger, the offset price can be higher under a cap. Fringe countries gain by mitigation in the policy bloc, more under a carbon tax since the fuel import price is lower.