Sensitivity of resource selection and connectivity models to landscape definition

Context

The definition of the geospatial landscape is the underlying basis for species-habitat models, yet sensitivity of habitat use inference, predicted probability surfaces, and connectivity models to landscape definition has received little attention.

Objectives

We evaluated the sensitivity of resource selection and connectivity models to four landscape definition choices including (1) the type of geospatial layers used, (2) layer source, (3) thematic resolution, and (4) spatial grain.

Methods

We used GPS telemetry data from pumas (Puma concolor) in southern California to create multi-scale path selection function models (PathSFs) across landscapes with 2500 unique landscape definitions. To create the landscape definitions, we identified seven geospatial layers that have been shown to influence puma habitat use. We then varied the number, sources, spatial grain, and thematic resolutions of these layers to create our suite of plausible landscape definitions. We assessed how PathSF model performance (based on AIC) was affected by landscape definition and examined variability among the predicted probability of movement surfaces, connectivity models, and road crossing locations.

Results

We found model performance was extremely sensitive to landscape definition and identified only seven top models out of our suite of definitions (<1%). Spatial grain and the number of geospatial layers selected for a landscape definition significantly affected model performance measures, with finer grains and greater numbers of layers increasing model performance.

Conclusions

Given the sensitivity of habitat use inference, predicted probability surfaces, and connectivity models to landscape definition, out results indicate the need for increased attention to landscape definition in future studies.

     

Characterising landscape connectivity for conservation planning using a dispersal guild approach

Context

Land use changes have modified the extent and structure of native vegetation, resulting in fragmentation of native species habitat. Connectivity is increasingly seen as a requirement for effective conservation in these landscapes, but the question remains: ‘connectivity for which species?’.

Objective

The aim of this study was to develop and then apply a rapid, expert-based, dispersal guild approach where species are grouped on similar fine-scale dispersal behaviour (such as between scattered trees) and habitat characteristics.

Methods

Dispersal guilds were identified using clustering techniques to compare dispersal and habitat parameters elicited from experts. We modelled least-cost paths and corridors between patches and individual movement probabilities within these corridors for each of the dispersal guilds using Circuitscape. We demonstrate our approach with a case study in the Tasmanian Northern Midlands, Australia.

Results

The dispersal guild approach grouped the 12 species into five dispersal guilds. The connectivity modelling of those five guilds found that broadly dispersing species in this landscape, such as medium-sized carnivorous mammals, were unaffected by fragmentation while from the perspective of the three dispersal guilds made up of smaller mammals, the landscape appeared highly fragmented.

Conclusions

Our approach yields biologically defensible outputs that are broadly applicable, particularly for conservation planning where data and resources are limited. It is a useful first step in multi-species conservation planning which aims to identify those species most in need of conservation efforts.

     

The importance of functional connectivity in the conservation of a ground-dwelling mammal in an urban Australian landscape

The distribution of the northern brown bandicoot (Isoodon macrourus), a medium-sized ground-dwelling marsupial, was examined in habitat fragments within the urban landscape of the city of Brisbane, Australia. From surveys conducted in 68 fragments, bandicoots were found to be present in 33 (49%) despite widespread habitat loss and fragmentation. Logistic regression analysis revealed that of 13 measured independent variables, functional connectivity was the only factor that significantly predicted the presence of bandicoots within fragments, with connectivity positively correlated with the likelihood of occupation. Functional connectivity was equated to the likelihood of bandicoot immigration into the focal fragment from the nearest occupied fragment, based on the estimated resistance to movement offered by the intervening matrix. Within Brisbane, riparian habitat fragments typically have a relatively high level of functional connectivity, as thin strips of vegetation fringing waterways serve as corridors between larger riparian areas and facilitate the movement of bandicoots between patches. Analyses based on the Akaike Information Criterion revealed that the optimal model based on landscape context variables was convincingly better supported by the data than the optimal model produced from fragment characteristics. However, it is important to examine both internal attributes of habitat fragments and external features of the surrounding landscape when modelling the distribution of ground-dwelling fauna in urban environments, or other landscapes with a highly variable matrix. As urban centres throughout the world expand, it is crucial that the ecology of local wildlife be considered to ensure functional connection is maintained between habitat patches, especially for the conservation of species that are highly susceptible to fragmentation.     

Evaluation of landscape connectivity at community level using satellite-derived NDVI

Landscape connectivity, defined as the degree to which the landscape facilitates or impedes movement among resource patches, has been considered to be a key issue for biodiversity conservation. However, the use of landscape connectivity measurements has been strongly criticised due to uncertainties in the methods used and the lack of validation. Moreover, measurements are typically restricted to the population level, whereas management is generally carried out at the community level. Here, we used satellite imagery and network metrics to predict the landscape connectivity at community level for semi-natural herbaceous patches in an urban area near Paris (France). We tested different measurement methods, both taking into account and ignoring the spatial heterogeneity of matrix resistance estimated by the normalised difference vegetation index (NDVI), and quantifying the link strength between patches with the shortest path and flow metrics. We assessed the fit of these connectivity predictions with empirical data on plant communities embedded in an urban matrix. Our results indicate that the best fit with the empirical data is obtained when the connectivity is estimated with the flow metric and takes into account the matrix heterogeneity. Overall, our study helps to estimate the landscape connectivity of urban areas and makes recommendations for ways in which we might optimise landscape planning with respect to conservation of urban biodiversity.     

Carabids (Coleoptera: Carabidae) in a forest patchwork: a connectivity analysis of the Bereg Plain landscape graph

For many species, one important key to persistence is maintaining connectivity among local populations that allow for dispersal and gene flow. This is probably true for carabid species (Coleoptera:Carabidae) living in the fragmented forests of the Bereg Plain (NE Hungary and W Ukraine). Based on field data, we have drafted a landscape graph of the area representing the habitat network of these species. Graph nodes and links represented two kinds of landscape elements: habitat (forest) patches and corridors, respectively. The quality of habitat patches and corridors were ranked (from low (1) to high (4)), reflecting local population sizes in the case of patches and estimated permeability in the case of corridors. We analysed (1) the positional importance of landscape elements in maintaining the connectivity of the intact network, (2) the effect of inserting hypothetical corridors into the network, (3) the effects of improving the quality of the existing corridors, and (4) how to connect every patch in a cost-effective way. Our results set quantitative priorities for conservation practice by identifying important corridors: what to protect, what to build and what to improve. Several network analytical techniques were used to account for the directed (source-sink) and highly fragmented nature of the landscape graph. We provide conservation priority ranks for the landscape elements and discuss the conditions for the use of particular network indices. Our study could be of extreme relevance, since a new highway is being planned through the area.     

Considering landscape connectivity and gene flow in the Anthropocene using complementary landscape genetics and habitat modelling approaches

Landscape Ecology - A comprehensive understanding of how rapidly changing environments affect species gene flow is critical for mitigating future biodiversity losses. While recent methodological...     

Modelling forest recolonization by an epiphytic lichen using a landscape genetic approach

The process of recolonization after disturbance is crucial for the persistence and dynamics of patch-tracking metapopulations. We developed a model to compare the spatial distribution and spatial genetic structure of the epiphytic lichen Lobaria pulmonaria within the perimeter of two reconstructed 19th century disturbances with a nearby reference area without stand-level disturbance. Population genetic data suggested that after stand-replacing disturbance, each plot was colonized by one or a few genotypes only, which subsequently spread clonally within a local neighborhood. The model (cellular automaton) aimed at testing the validity of this interpretation and at assessing the relative importance of local dispersal of clonal propagules vs. long-distance dispersal of clonal and/or sexual diaspores. A reasonable model fit was reached for the empirical data on host tree distribution, lichen distribution, and tree- and plot-level genotype diversity of the lichen in the reference area. Although model calibration suggested a predominance of local dispersal of clonal propagules, a substantial contribution of immigration of non-local genotypes by long-distance dispersal was needed to reach the observed levels of genotype diversity. The model could not fully explain the high degree of clonality after stand-replacing disturbance, suggesting that the dispersal process itself may not be stationary but depend on the ecological conditions related to disturbance.     

Individual-based modeling highlights the importance of mortality and landscape structure in measures of functional connectivity

Landscape Ecology - Functional landscape connectivity is vital for the conservation of wildlife species. Landscape connectivity models often overlook factors such as mortality and asymmetry in...     

Evaluating static and dynamic landscape connectivity modelling using a 25-year remote sensing time series

Context

Despite calls for landscape connectivity research to account for spatiotemporal dynamics, studies have overwhelmingly evaluated the importance of habitats for connectivity at single or limited moments in time. Remote sensing time series represent a promising resource for studying connectivity within dynamic ecosystems. However, there is a critical need to assess how static and dynamic landscape connectivity modelling approaches compare for prioritising habitats for conservation within dynamic environments.

Objectives

To assess whether static landscape connectivity analyses can identify similar important areas for connectivity as analyses based on dynamic remotely sensed time series data.

Methods

We compared degree and betweenness centrality graph theory metric distributions from four static scenarios against equivalent results from a dynamic 25-year remotely sensed surface-water time series. Metrics were compared at multiple spatial aggregation scales across south-eastern Australia’s 1 million km² semi-arid Murray–Darling Basin and three sub-regions with varying levels of hydroclimatic variability and development.

Results

We revealed large differences between static and dynamic connectivity metric distributions that varied by static scenario, region, spatial scale and hydroclimatic conditions. Static and dynamic metrics showed particularly low overlap within unregulated and spatiotemporally variable regions, although similarities increased at coarse aggregation scales.

Conclusions

In regions that exhibit high spatiotemporal variability, static connectivity modelling approaches are unlikely to serve as effective surrogates for more data intensive approaches based on dynamic, remotely sensed data. Although this limitation may be moderated by spatially aggregating static connectivity outputs, our results highlight the value of remotely sensed time series for assessing connectivity in dynamic landscapes.

     

Hierarchical relationships between landscape structure and temperature in a managed forest landscape

Management may influence abiotic environments differently across time and spatial scale, greatly influencing perceptions of fragmentation of the landscape. It is vital to consider a priori the spatial scales that are most relevant to an investigation, and to reflect on the influence that scale may have on conclusions. While the importance of scale in understanding ecological patterns and processes has been widely recognized, few researchers have investigated how the relationships between pattern and process change across spatial and temporal scales. We used wavelet analysis to examine the multiscale structure of surface and soil temperature, measured every 5 m across a 3820 m transect within a national forest in northern Wisconsin. Temperature functioned as an indicator – or end product – of processes associated with energy budget dynamics, such as radiative inputs, evapotranspiration and convective losses across the landscape. We hoped to determine whether functional relationships between landscape structure and temperature could be generalized, by examining patterns and relationships at multiple spatial scales and time periods during the day. The pattern of temperature varied between surface and soil temperature and among daily time periods. Wavelet variances indicated that no single scale dominated the pattern in temperature at any time, though values were highest at finest scales and at midday. Using general linear models, we explained 38% to 60% of the variation in temperature along the transect. Broad categorical variables describing the vegetation patch in which a point was located and the closest vegetation patch of a different type (landscape context) were important in models of both surface and soil temperature across time periods. Variables associated with slope and microtopography were more commonly incorporated into models explaining variation in soil temperature, whereas variables associated with vegetation or ground cover explained more variation in surface temperature. We examined correlations between wavelet transforms of temperature and vegetation (i.e., structural) pattern to determine whether these associations occurred at predictable scales or were consistent across time. Correlations between transforms characteristically had two peaks; one at finer scales of 100 to 150 m and one at broader scales of ^>300 m. These scales differed among times of day and between surface and soil temperatures. Our results indicate that temperature structure is distinct from vegetation structure and is spatially and temporally dynamic. There did not appear to be any single scale at which it was more relevant to study temperature or this pattern-process relationship, although the strongest relationships between vegetation structure and temperature occurred within a predictable range of scales. Forest managers and conservation biologists must recognize the dynamic relationship between temperature and structure across landscapes and incorporate the landscape elements created by temperature-structure interactions into management decisions.     

Spatial dynamics of coastal forest bird assemblages: the influence of landscape context,forest type,and structural connectivity

Context

Complex structural connectivity patterns can influence the distribution of animals in coastal landscapes, particularly those with relatively large home ranges, such as birds. To understand the nuanced nature of coastal forest avifauna, where there may be considerable overlap in assemblages of adjacent forest types, the concerted influence of regional landscape context and vegetative structural connectivity at multiple spatial scales warrants investigation.

Objectives

This study determined whether species compositions of coastal forest bird assemblages differ with regional landscape context or with forest type, and if this is influenced by structural connectivity patterns measured at multiple spatial scales.

Methods

Three replicate bird surveys were conducted in four coastal forest types at ten survey locations across two regional landscape contexts in northeast Australia. Structural connectivity patterns of 11 vegetation types were quantified at 3, 6, and 12 km spatial scales surrounding each survey location, and differences in bird species composition were evaluated using multivariate ordination analysis.

Results

Bird assemblages differed between regional landscape contexts and most coastal forest types, although Melaleuca woodland bird assemblages were similar to those of eucalypt woodlands and rainforests. Structural connectivity was primarily correlated with differences in bird species composition between regional landscape contexts, and correlation depended on vegetation type and spatial scale.

Conclusions

Spatial scale, landscape context, and structural connectivity have a combined influence on bird species composition. This suggests that effective management of coastal landscapes requires a holistic strategy that considers the size, shape, and configuration of all vegetative components at multiple spatial scales.

     

Is landscape connectivity a dependent or independent variable?

With growing interest in landscape connectivity, it is timely to ask what research has been done and what re mains to be done. I surveyed papers investigating landscape connectivity from 1985 to 2000. From these papers, I determined if connectivity had been treated as an independent or dependent variable, what connectivity metrics were used, and if the study took an empirical or modeling approach to studying connectivity. Most studies treated connectivity as an independent variable, despite how little we know about how landscape structure and organism movement behaviour interact to determine landscape connectivity. Structural measures of connectivity were more common than functional measures, particularly if connectivity was treated as an independent variable. Though there was a good balance between modeling and empirical approaches overall – studies dealing with connectivity as a dependent, functional variable were mainly modeling studies. Based on the research achieved thus far, fu ture landscape connectivity research should focus on: (1) elucidating the relationship between landscape struc ture, organism movement behaviour, and landscape connectivity (e.g., treating connectivity as a dependent variable), (2) determining the relationships between different measures of connectivity, particularly structural and functional measures, and (3) empirically testing model predictions regarding landscape connectivity.This revised version was published online in May 2005 with corrections to the Cover Date.     

Genetic analysis of landscape connectivity in tree populations

Genetic connectivity in plant populations is determined by gene movement within and among populations. When populations become genetically isolated, they are at risk of loss of genetic diversity that is critical to the long-term survival of populations. Anthropogenic landscape change and habitat fragmentation have become so pervasive that they may threaten the genetic connectivity of many plant species. The theoretical consequences of such changes are generally understood, but it is not immediately apparent how concerned we should be for real organisms, distributed across real landscapes. Our goals here are to describe how one can study gene movement of both pollen and seeds in the context of changing landscapes and to explain what we have learned so far. In the first part, we will cover methods of describing pollen movement and then review evidence for the impact of fragmentation in terms of both the level of pollen flow into populations and the genetic diversity of the resulting progeny. In the second part, we will describe methods for contemporary seed movement, and describe findings about gene flow and genetic diversity resulting from seed movement. Evidence for pollen flow suggests high connectivity, but it appears that seed dispersal into fragments may create genetic bottlenecks due to limited seed sources. Future work should address the interaction of pollen and seed flow and attention needs to be paid to both gene flow and the diversity of the incoming gene pool. Moreover, if future work is to model the impact of changing landscapes on propagule movement, with all of its ensuing consequences for genetic connectivity and demographic processes, we will need an effective integration of population genetics and landscape ecology.     

Landscape drivers of connectivity for a forest rodent in a coffee agroecosystem

Landscape Ecology - The majority of remaining tropical forests exist as fragments embedded in a matrix of agricultural production. Understanding the effects of these agricultural landscapes on...     

Yards increase forest connectivity in urban landscapes

Landscape Ecology - Tree canopy connectivity is important for supporting biodiversity. In urban landscapes, empirical examinations of habitat connectivity often overlook residential land, though...     

Examining speed versus selection in connectivity models using elk migration as an example

Context

Landscape resistance is vital to connectivity modeling and frequently derived from resource selection functions (RSFs). RSFs estimate relative probability of use and tend to focus on understanding habitat preferences during slow, routine animal movements (e.g., foraging). Dispersal and migration, however, can produce rarer, faster movements, in which case models of movement speed rather than resource selection may be more realistic for identifying habitats that facilitate connectivity.

Objective

To compare two connectivity modeling approaches applied to resistance estimated from models of movement rate and resource selection.

Methods

Using movement data from migrating elk, we evaluated continuous time Markov chain (CTMC) and movement-based RSF models (i.e., step selection functions [SSFs]). We applied circuit theory and shortest random path (SRP) algorithms to CTMC, SSF and null (i.e., flat) resistance surfaces to predict corridors between elk seasonal ranges. We evaluated prediction accuracy by comparing model predictions to empirical elk movements.

Results

All connectivity models predicted elk movements well, but models applied to CTMC resistance were more accurate than models applied to SSF and null resistance. Circuit theory models were more accurate on average than SRP models.

Conclusions

CTMC can be more realistic than SSFs for estimating resistance for fast movements, though SSFs may demonstrate some predictive ability when animals also move slowly through corridors (e.g., stopover use during migration). High null model accuracy suggests seasonal range data may also be critical for predicting direct migration routes. For animals that migrate or disperse across large landscapes, we recommend incorporating CTMC into the connectivity modeling toolkit.

     

Using step and path selection functions for estimating resistance to movement: pumas as a case study

Context

GPS telemetry collars and their ability to acquire accurate and consistently frequent locations have increased the use of step selection functions (SSFs) and path selection functions (PathSFs) for studying animal movement and estimating resistance. However, previously published SSFs and PathSFs often do not accommodate multiple scales or multi-scale modeling.

Objectives

We present a method that allows multiple scales to be analyzed with SSF and PathSF models. We also explore the sensitivity of model results and resistance surfaces to whether SSFs or PathSFs are used, scale, prediction framework, and GPS collar sampling interval.

Methods

We use 5-min GPS collar data from pumas (Puma concolor) in southern California to model SSFs and PathSFs at multiple scales, to predict resistance using two prediction frameworks (paired and unpaired), and to explore potential bias from GPS collar sampling intervals.

Results

Regression coefficients were extremely sensitive to scale and pumas exhibited multiple scales of selection during movement. We found PathSFs produced stronger regression coefficients, larger resistance values, and superior model performance than SSFs. We observed more heterogeneous surfaces when resistance was predicted in a paired framework compared with an unpaired framework. Lastly, we observed bias in habitat use and resistance results when using a GPS collar sampling interval longer than 5 min.

Conclusions

The methods presented provide a novel way to model multi-scale habitat selection and resistance from movement data. Due to the sensitivity of resistance surfaces to method, scale, and GPS schedule, care should be used when modeling corridors for conservation purposes using these methods.

     

Improving habitat and connectivity model predictions with multi-scale resource selection functions from two geographic areas

Landscape Ecology - Habitat loss and fragmentation are the most pressing threats to biodiversity, yet assessing their impacts across broad landscapes is challenging. Information on habitat...     

Temporal shifts in landscape connectivity for an ecosystem engineer,the roe deer,across a multiple-use landscape

Context

Routine movements of large herbivores, often considered as ecosystem engineers, impact key ecological processes. Functional landscape connectivity for such species influences the spatial distribution of associated ecological services and disservices.

Objectives

We studied how spatio-temporal variation in the risk-resource trade-off, generated by fluctuations in human activities and environmental conditions, influences the routine movements of roe deer across a heterogeneous landscape, generating shifts in functional connectivity at daily and seasonal time scales.

Methods

We used GPS locations of 172 adult roe deer and step selection functions to infer landscape connectivity. In particular, we assessed the influence of six habitat features on fine scale movements across four biological seasons and three daily periods, based on variations in the risk-resource trade-off.

Results

The influence of habitat features on roe deer movements was strongly dependent on proximity to refuge habitat, i.e. woodlands. Roe deer confined their movements to safe habitats during daytime and during the hunting season, when human activity is high. However, they exploited exposed open habitats more freely during night-time. Consequently, we observed marked temporal shifts in landscape connectivity, which was highest at night in summer and lowest during daytime in autumn. In particular, the onset of the autumn hunting season induced an abrupt decrease in landscape connectivity.

Conclusions

Human disturbance had a strong impact on roe deer movements, generating pronounced spatio-temporal variation in landscape connectivity. However, high connectivity at night across all seasons implies that Europe’s most abundant and widespread large herbivore potentially plays a key role in transporting ticks, seeds and nutrients among habitats.

     

Estimation of metademographic rates and landscape connectivity for a conservation-reliant anuran

Landscape Ecology - Amphibian conservation efforts commonly assume populations are tied to waterbodies that collectively function as a metapopulation. This assumption is rarely evaluated, and there...