Avian fitness consequences match habitat selection at the nest‐site and landscape scale in agriculturally fragmented landscapes

相似文献   

The nest predator community of grassland birds responds to agroecosystem habitat at multiple scales

Nest predation is the leading cause of reproductive failure for grassland birds of conservation concern. Understanding variation in nest predation rates is complicated by the diverse assemblage of species known to prey on nests. As part of a long‐term study of grassland bird ecology, we monitored populations of predators known to prey on grassland bird nests. We used information theoretic approach to examine the predator community's association with habitat at multiple scales, including local vegetation structure of grassland patches, spatial attributes of grassland patches (size and shape), and landscape composition surrounding grassland patches (land cover within 400 and 1600 m). Our results confirmed that nest predators respond to habitat at multiple scales and different predator species respond to habitat in different ways. The most informative habitat models we selected included variability in local vegetation (CV in the density of forbs), local patch (area and edge‐to‐interior ratio), and landscape within a 1600 m buffer around grasslands (percent of land covered by human structures and development). As a separate question, we asked if models that incorporated information from multiple scales simultaneously might improve the ability to explain variation in the predator community. Multi‐ scale models were not consistently superior to models derived from variables focused at a single spatial scale. Our results suggest that minimizing human development on and surrounding conservation land and the management of the vegetation structure on grassland fragments both may benefit grassland birds by decreasing the risk of nest predation.     

Sage-Grouse Habitat Selection During Winter in Alberta

ABSTRACT Greater sage-grouse (Centrocercus urophasianus) are dependent on sagebrush (Artemisia spp.) for food and shelter during winter, yet few studies have assessed winter habitat selection, particularly at scales applicable to conservation planning. Small changes to availability of winter habitats have caused drastic reductions in some sage-grouse populations. We modeled winter habitat selection by sage-grouse in Alberta, Canada, by using a resource selection function. Our purpose was to 1) generate a robust winter habitat-selection model for Alberta sage-grouse; 2) spatially depict habitat suitability in a Geographic Information System to identify areas with a high probability of selection and thus, conservation importance; and 3) assess the relative influence of human development, including oil and gas wells, in landscape models of winter habitat selection. Terrain and vegetation characteristics, sagebrush cover, anthropogenic landscape features, and energy development were important in top Akaike's Information Criterion-selected models. During winter, sage-grouse selected dense sagebrush cover and homogenous less rugged areas, and avoided energy development and 2-track truck trails. Sage-grouse avoidance of energy development highlights the need for comprehensive management strategies that maintain suitable habitats across all seasons.     

Home range and habitat selection of northern bobwhite coveys in an agricultural landscape

Northern bobwhites thrive in fine-grained landscapes with a diversity of early succession woodland, grassland, and agriculture-associated habitat types. Bobwhite conservation has proved challenging in the increasingly coarse-grained Midwestern landscape as simplified agricultural cropping systems are implemented at larger spatial scales. Regardless, managing agricultural landscapes on private lands is the primary opportunity to restore bobwhite populations in the Midwestern United States. Although bobwhite habitat requirements are well understood, habitat selection in contemporary Midwestern landscapes is not well understood, especially on private lands where populations are declining. We used compositional analysis to investigate second- (study area) and third- (home range) order habitat selection by radiomarked bobwhite coveys on 4 private land study areas in southwestern Ohio. Mean covey home range size was 26.1 ± 2.2 ha (n = 48). Although home ranges were established in areas with more grassland cover, bobwhites most strongly selected early succession woody habitat (e.g., fencerows and ditches) at all scales, and selection for grassland diminished between the study area and home range scales. Grassland selection varied among sites and was strongest on sites with more row crop area. Woodlots were avoided at the study area scale, but were selected within home ranges. Grassland cover, like that provided by contemporary conservation programs, is an essential component of bobwhite habitat in the Midwest, but our results suggest more emphasis should be placed on early succession woody cover. Woody cover associated with fencerows, ditches, and woodlots adjacent to food sources and breeding habitat will likely improve non-breeding season survival, which is an influential vital rate in northern populations. © 2012 The Wildlife Society.     

Eastern Hog-Nosed Snake Habitat Selection at Multiple Spatial Scales in Ontario,Canada

Habitat loss is the greatest contributor to the decline of species globally. To prioritize protection of imperiled species, it is important to examine habitat use at multiple spatial scales because the availability of different resources and habitat features is scale dependent. We conducted a radio-telemetry study in the Long Point region of Ontario, Canada, in 2009 and 2010 to examine habitat selection at multiple spatial scales by eastern hog-nosed snakes (Heterodon platirhinos), a species at risk in Canada. We documented the habitat composition of home ranges compared to the surrounding landscape, the selection of locations within home ranges based on classified satellite imagery, and the use of microhabitat features based on site characterization in the field. At the scale of the home ranges, hog-nosed snakes avoided areas of agriculture and selected sand barrens. Within home ranges, hog-nosed snakes selectively used areas altered by humans (e.g., residential sites, openings in tree plantations). Microhabitats used by hog-nosed snakes had more woody debris, logs, and lower vegetative coverage than adjoining random sites. Because hog-nosed snakes prefer open areas and require sandy soils for nesting, management efforts should focus on the conservation and maintenance of sand barrens and patches of early successional forest. © 2021 The Wildlife Society.     

Waterfowl distribution and productivity in the Prairie Pothole Region of Canada: tools for conservation planning

Species conservation requires an understanding of the factors and interactions affecting species distribution and behavior, habitat availability and use, and corresponding vital rates at multiple temporal and spatial scales. Opportunities to investigate these relationships across broad geographic regions are rare. We combined long-term waterfowl population surveys, and studies of habitat use and breeding success, to develop models that identify and incorporate these interactions for upland-nesting waterfowl in the Prairie Pothole Region (PPR) of Canada. Specifically, we used data from the annual Waterfowl Breeding Population and Habitat Survey (1961–2009) at the survey segment level and associated habitat covariates to model and map the long-term average duck density across the Canadian PPR. We analyzed nest location and fate data from approximately 25,000 duck nests found during 3 multi-year nesting studies (1994–2011) to model factors associated with nest survival and habitat selection through the nesting season for the 5 most common upland nesting duck species: mallard (Anas platyrhynchos), gadwall (Mareca strepera), blue-winged teal (Spatula discors), northern shoveler (Spatula clypeata), and northern pintail (Anas acuta). Duck density was highly variable across the Canadian PPR, reflecting positive responses to local wetland area and count, and amounts of cropland and grassland, a regional positive response to latitude, and a negative response to local amounts of tree cover. Nest survival was affected by temporal and spatial variables at multiple scales. Specifically, nest survival demonstrated interactive effects among species, nest initiation date, and nesting cover type and was influenced by relative annual wetness, population density, and surrounding landscape composition at landscape scales, and broad geographic gradients (east-west and north-south). Likewise, species-specific probability of nest habitat selection was influenced by timing of nest initiation, population density, relative annual wetness, herbaceous cover, and tree cover in the surrounding landscape, and location within the Canadian PPR. We combined these models, with estimates of breeding effort (nesting, renesting, and nest attempts) from existing literature, in a stochastic conservation planning model that estimates nest distribution and success given spatiotemporal variation in duck density, habitat availability, and influential covariates. We demonstrate the use of this model by examining various conservation planning scenarios. These models allow estimation of local, landscape, and regional influence of conservation investments and other landscape changes on the productivity of breeding duck populations across the PPR of Canada. These models lay the groundwork for the incorporation of conservation delivery costs for full return-on-investment analyses and scenario analyses of climate, habitat, and land use change in regional and continental population models.     

A Hierarchical Analysis of Habitat Selection by Raccoons in Northern Indiana

Abstract: Although numerous studies have examined habitat use by raccoons (Procyon lotor), information regarding seasonal habitat selection related to resource availability in agricultural landscapes is lacking for this species. Additionally, few studies using radiotelemetry have investigated habitat selection at multiple spatial scales or core-use areas by raccoons. We examined seasonal habitat selection of 55 (31 M, 24 F) adult raccoons at 3 hierarchical orders defined by the movement behavior of this species (second-order home range, second-order core-use area, and third-order home range) in northern Indiana, USA, from May 2003 to June 2005. Using compositional analysis, we assessed whether habitat selection differed from random and ranked habitat types in order of selection during the crop growing period (season 1) and corn maturation period (season 2), which represented substantial shifts in resource availability to raccoons. Habitat rankings differed across hierarchical orders, between seasons within hierarchical orders, and between sexes within seasons; however, seasonal and intersexual patterns of habitat selection were not consistent across hierarchical orders of spatial scale. When nonrandom utilization was detected, both sexes consistently selected forest cover over other available habitats. Seasonal differences in habitat selection were most evident at the core-area scale, where raccoon selection of agricultural lands was highest during the maturation season when corn was available as a direct food source. Habitat use did not differ from availability for either sex in either season at the third-order scale. The selection of forest cover across both seasons and all spatial orders suggested that raccoon distribution and abundance in fragmented landscapes is likely dependent on the availability and distribution of forest cover, or habitats associated with forest (i.e., water), within the landscape. The lack of consistency in habitat selection across hierarchical scales further exemplifies the need to examine multiple biological scales in habitat-selection studies.     

Habitat use by endangered Japanese crayfish (<Emphasis Type="Italic">Cambaroides japonicus</Emphasis>) in low-gradient streams of southern Hokkaido,Japan: reach and microhabitat-scale analysis

The Japanese crayfish (Cambaroides japonicus), the only native crayfish in Japan, is endangered and has experienced rapid population declines. We surveyed the habitat requirements of Japanese crayfish at the reach and microhabitat scales in semi-natural low-gradient streams. Habitat use by Japanese crayfish differed between the spatial scales. Reach-scale analysis revealed that the bed slope was the only positive predictor of crayfish density. This finding indicates that relatively high-gradient reaches, such as headwater reaches, are an important habitat for the conservation of Japanese crayfish in low-gradient streams. Microhabitat-scale analysis showed that crayfish density was positively affected by substrate coarseness and the presence of instream vegetation cover (bank vegetation, woody debris, and leaf patches), whereas it was negatively affected by distance from the stream edge. Coarse substrates and vegetation cover may function as shelters from water flows and predators during low flow periods. The use of stream-edge areas may allow quick access to refugia and enable the crayfish to avoid unexpected flood disturbance and predation. These habitat characteristics should therefore be preserved for the conservation of Japanese crayfish, and scale-dependent habitat characteristics should be considered in future conservation plans.     

A multi-scale approach for identifying conservation needs of two threatened sympatric steppe birds

Habitat selection is an inherently scale-sensitive process in which detected selection patterns frequently depend on the scale of analysis employed. We used a multi-scale modelling approach to identify how the distributions of two sympatric birds are shaped by differential selection at the landscape, land use and microhabitat scales and by human infrastructures as possible sources of disturbance. We studied two threatened steppe birds, the pin-tailed sandgrouse (PTS) and black-bellied sandgrouse (BBS) in central Spain. Land use gradients explained most of the variation in PTS and BBS occurrence, but there was cross-scale interdependence between the lower (microhabitat) and upper (landscape) spatial scales for the PTS. Synergies between the three scales highlighted the importance of integrating habitat scales in a single modelling framework. The process of habitat selection was also modulated by human disturbance. Both species selected ploughs of large size distant from houses, tracks and other infrastructures, although BBS exhibited broader habitat tolerance than the PTS, and was more sensitive to human disturbance. At microhabitat scale, PTS selected ploughs with greater green vegetation cover and insect abundance and fallows with lower dry vegetation cover and height but greater stone cover. This might reflect a trade-off between camouflage (vegetation and stone cover for concealment) and visibility for predator detection and escape. Ploughs and fallows should be maintained by means of traditional 2-year rotations and low management during the breeding season. Ongoing urbanization trends and infrastructure development inside protected areas should be limited. Multi-scale models were key to identify scale-specific factors that determine sandgrouse habitat preferences and conservation requirements at appropriate levels, and are recommended to better guide regional and local conservation efforts of threatened species.     

Flight morphology corresponds to both surrounding landscape structure and local patch conditions in a highly specialized peatland butterfly (Lycaena epixanthe)

1. Movement mediates the response of populations and communities to landscape and habitat spatial structure, yet movement capability may itself be modified by selection pressures accompanying landscape change. Insect flight morphology can be affected by both the landscape surrounding habitat patches and the distribution of resources within habitat patches. 2. This study investigated the relative influence of local habitat patch conditions and surrounding landscape structure on variation in morphological traits associated with flight in the bog copper (Lycaena epixanthe), a butterfly endemic to temperate Nearctic peatlands. 3. Eight habitat patches were sampled to assess the influence of the surrounding landscape (connectivity of potential habitat and matrix composition) and patch size (an integrated proxy of resource density and spatial distribution) on investment into flight, measured by thorax and abdomen mass, and wing area. 4. The results revealed an effect of both local habitat conditions and landscape structure on flight‐related morphological traits. Increasing forest cover in the surrounding landscape, indicative of increased habitat patch isolation, corresponded with less mobile phenotypes in both sexes. Surrounding landscapes with more water were also generally associated with less mobile phenotypes. Investment into flight was greater in smaller peatlands in which host plant density is higher and more homogeneously distributed. 5. The present study highlights that morphological traits associated with mobility may be responding to both local habitat patch characteristics and surrounding landscape structure. It also supports the hypothesis that local habitat conditions contribute to morphological variation in butterflies.     

Multiscale resistant kernel surfaces derived from inferred gene flow: An application with vernal pool breeding salamanders

The importance of assessing spatial data at multiple scales when modelling species–environment relationships has been highlighted by several empirical studies. However, no landscape genetics studies have optimized landscape resistance surfaces by evaluating relevant spatial predictors at multiple spatial scales. Here, we model multiscale/layer landscape resistance surfaces to estimate resistance to inferred gene flow for two vernal pool breeding salamander species, spotted (Ambystoma maculatum) and marbled (A. opacum) salamanders. Multiscale resistance surface models outperformed spatial layers modelled at their original spatial scale. A resistance surface with forest land cover at a 500‐m Gaussian kernel bandwidth and normalized vegetation index at a 100‐m Gaussian kernel bandwidth was the top optimized resistance surface for A. maculatum, while a resistance surface with traffic rate and topographic curvature, both at a 500‐m Gaussian kernel bandwidth, was the top optimized resistance surface for A. opacum. Species‐specific resistant kernels were fit at all vernal pools in our study area with the optimized multiscale/layer resistance surface controlling kernel spread. Vernal pools were then evaluated and scored based on surrounding upland habitat (local score) and connectivity with other vernal pools on the landscape, with resistant kernels driving vernal pool connectivity scores. As expected, vernal pools that scored highest were in areas within forested habitats and with high vernal pool densities and low species‐specific landscape resistance. Our findings highlight the success of using a novel analytical approach in a multiscale framework with applications beyond vernal pool amphibian conservation.     

Improving conservation strategies of raptors through landscape ecology analysis: The case of the endemic Cuban Black Hawk

相似文献   

Habitat selection of breeding riparian birds in an urban environment: untangling the relative importance of biophysical elements and spatial scale

Aim Urbanization is a leading threat to global biodiversity, yet little is known about how the spatial arrangement and composition of biophysical elements – buildings and vegetation – within a metropolitan area influence habitat selection. Here, we ask: what is the relative importance of the structure and composition of these elements on bird species across multiple spatial scales? Location The temperate metropolitan area of Cincinnati, Ohio, USA. Methods We surveyed breeding birds on 71 plots along an urban gradient. We modelled relative density for 48 bird species in relation to local woody vegetation composition and structure and to tree cover, grass cover and building density within 50–1000 m of each plot. We used an information‐theoretic approach to compare models and variables. Results At the proximate scale, native tree and understory stem frequency were the most important vegetation variables explaining bird distributions. Species’ responses to landscape biophysical features and spatial scales varied. Most native species responded positively to vegetation measures and negatively to building density. Models combining both local vegetation and landscape information represented best or competitive models for the majority of species, while models containing only local vegetation characteristics were rarely competitive. Smaller spatial scales (≤ 500 m) were most important for 36 species, and eight species had best models at larger scales (> 500 m); however, several species had competitive models across multiple scales. Main conclusions Habitat selection by birds within the urban matrix is the result of a combination of factors operating at both proximate and broader spatial scales. Efforts to manage and design urban areas to benefit native birds require both fine‐scale (e.g., individual landowners and landscape design) and larger landscape actions (e.g., regional comprehensive planning).     

Landscape genetic connectivity in European wildcat (Felis silvestris silvestris): a matter of food,shelters and demographic status of populations

Understanding landscape impacts on gene flow is necessary to plan comprehensive management and conservation strategies of both the species of interest and its habitat. Nevertheless, only a few studies have focused on the landscape genetic connectivity of the European wildcat, an umbrella species whose conservation allows the preservation of numerous other species and habitat types. We applied population and landscape genetics approaches, using genotypes at 30 microsatellites from 232 genetically-identified wildcats to determine if, and how, landscape impacted gene flow throughout France. Analyses were performed independently within two population patches: the historical north-eastern patch and the central patch considered as the colonization front. Our results showed that gene flow occurred at large spatial scales but also revealed significant spatial genetic structures within population patches. In both population patches, arable areas, pastures and permanent grasslands and lowly fragmented forested areas were permeable to gene flow, suggesting that shelters and dietary resources are among the most important parameters for French wildcat landscape connectivity, while distance to forest had no detectable effect. Anthropized areas appeared highly resistant in the north-eastern patch but highly permeable in the central patch, suggesting that different behaviours can be observed according to the demographic context in which populations are found. In line with this hypothesis, spatial distribution of genetic variability seemed uneven in the north-eastern patch and more clinal in the central patch. Overall, our results highlighted that European wildcat might be a habitat generalist species and also the importance of performing spatial replication in landscape genetics studies.

     

Multi-Scale Associations between Vegetation Cover and Woodland Bird Communities across a Large Agricultural Region

Improving biodiversity conservation in fragmented agricultural landscapes has become an important global issue. Vegetation at the patch and landscape-scale is important for species occupancy and diversity, yet few previous studies have explored multi-scale associations between vegetation and community assemblages. Here, we investigated how patch and landscape-scale vegetation cover structure woodland bird communities. We asked: (1) How is the bird community associated with the vegetation structure of woodland patches and the amount of vegetation cover in the surrounding landscape? (2) Do species of conservation concern respond to woodland vegetation structure and surrounding vegetation cover differently to other species in the community? And (3) Can the relationships between the bird community and the woodland vegetation structure and surrounding vegetation cover be explained by the ecological traits of the species comprising the bird community? We studied 103 woodland patches (0.5 - 53.8 ha) over two time periods across a large (6,800 km²) agricultural region in southeastern Australia. We found that both patch vegetation and surrounding woody vegetation cover were important for structuring the bird community, and that these relationships were consistent over time. In particular, the occurrence of mistletoe within the patches and high values of woody vegetation cover within 1,000 ha and 10,000 ha were important, especially for bird species of conservation concern. We found that the majority of these species displayed similar, positive responses to patch and landscape vegetation attributes. We also found that these relationships were related to the foraging and nesting traits of the bird community. Our findings suggest that management strategies to increase both remnant vegetation quality and the cover of surrounding woody vegetation in fragmented agricultural landscapes may lead to improved conservation of bird communities.     

Exploring effective conservation networks based on multi-scale planning unit analysis. A case study of the Balsas sub-basin,Maranhão State,Brazil

Nature conservation and restoration activities require delineation of effective conservation networks. This paper presents a methodology which allows a quick evaluation of different planning options for extensive areas. We analyzed the spatial structure of remaining patches of the natural Cerrado vegetation in the Balsás sub-basin, South of Maranhão State of Brazil (about 25,590 km²) in order to understand how the remaining habitats are distributed and spatially configured. Conservation area network scenarios are based on hexagonal cells, referred to as analysis unit (AU) cells. A multi-scale analysis of 10,000 ha and 50,000 ha AU cells was set up to represent local and regional scales, respectively. For each AU at both local and regional scales we computed landscape metrics of native vegetation: NATIVE VEGETATION COVER: percentage of native vegetation cover; (NV-NP): number of patches; (NV-MSI): mean shape index. Subsequently, five different conservation and restoration strategies were defined: (a) only enforce nature conservation within legally established units; (b) target nature conservation only within the local AU landscape; (c) target regional management by combining neighboring AU; (d) management of both local landscape and region; (e) protect the legal conservation areas and promote local and regional conservation. We also generated scenarios of habitat capacity for mammals and matched these results with the different vegetation conservation scenarios. Results indicate that only 12% of the study area is well conserved and that 43% of the region is in a very critical condition. The percentage of AU cells where native vegetation conservation actions are required differ for the five conservation strategies: These results allow policy makers and other stakeholders to target the locations and extent of conservation units required. We suggest that about 45% of the sub-basin could be managed at local, regional or both scales. Regarding a mammal species diversity scenario an even higher percentage of the average habitat capacity of the selected species occurs in open cerrado and valleys areas that coincide with critical areas. The proposed multi-scale analysis unit cell approach can support the planning process of extensive areas as necessary in Brazil.     

Chainsawing for conservation: Ecologically informed tree removal for habitat management

Summary In many ecosystems, increases in vegetation density and the resulting closure of forest canopies are threatening the viability of species that depend upon open, sunlight‐exposed habitats. Consequently, we need to develop management strategies that recreate open habitats while minimizing the impacts on non‐target areas. Selective logging creates canopy gaps, but may result in undesirable effects in other respects. Thus, chainsaws have not been a popular tool for conservation. We conducted a landscape‐scale experiment to test whether selective tree removal can restore patch‐level habitat quality for Australia’s most endangered snake (Hoplocephalus bungaroides) and its main prey (the lizard Oedura lesueurii). We selectively removed canopy trees surrounding 25 overgrown rock outcrops and compared the resultant habitat structure and abiotic conditions to 30 overgrown, shady outcrops and 20 open, sunny outcrops. Removing vegetation decreased canopy cover by 19% in experimental plots and increased incident radiation and thermal regimes. These changes increased the availability of suitable shelter sites for our target species by 131%. At the landscape scale, our manipulations had a trivial effect on forest habitat; by increasing the area of sun‐exposed outcrops, we decreased forest cover by <0.1%. Our results show that targeted canopy removal can increase the availability of sun‐exposed habitat patches for endangered species in biologically meaningful ways. Thus, selective tree felling may be an effective conservation tool for open‐habitat specialists threatened by vegetation overgrowth.     

Vegetation and songbird response to land abandonment: from landscape to census plot

While intensification of human activities and its ecological effects in many natural areas have recently received much attention, land abandonment in marginal areas is still the largely ignored side of a process rooted in the same socioeconomic context. Decreasing human impact in marginal rural areas often triggers a recovery of seminatural vegetation. Over a period of 25 years, we studied the changes in landscape and vegetation structure that followed land abandonment in a traditional Mediterranean mosaic of crops, grasslands, shrublands and woodlands, and assessed their effects on songbird occurrence and distribution. We combined an analysis of vegetation changes based on aerial photo interpretation with an analysis of bird censuses from 1978, 1992 and 2003 at two spatial scales: landscape and census plot (respectively 2800 and 3 ha). The perceived temporal changes in the vegetation were scale dependent. At the landscape scale, open habitats tended to disappear and woodlands matured. The contrasts in vegetation structure that defined habitat patches at the onset of the study tended to disappear. There was an overall shift of the bird community in favour of woodland species. At the scale of the census plot, however, the colonization by woody vegetation of patches formerly characterized by a homogeneous grass cover increased the local diversity of the vegetation, at least temporarily. Of seven species dependent on open habitats, the occurrence rate of five species significantly decreased, whereas it increased for two species: woodlark (Lulula arborea) and melodious warbler (Hippolais polyglotta). This increase was linked to the transitional increase in local vegetation diversity. In patches originally dominated by woodlands, local vegetation diversity decreased as woody vegetation expanded into clearings. The occurrence rate significantly increased for seven species relying on closed woodlands, while it decreased for two woodland species. As most species of high conservation profile in the Mediterranean are tied to open or to heterogeneous transitional habitats, these trends raise questions concerning their persistence in the future.     

The spatial scaling of habitat selection by African elephants

1. Understanding and accurately predicting the spatial patterns of habitat use by organisms is important for ecological research, biodiversity conservation and ecosystem management. However, this understanding is complicated by the effects of spatial scale, because the scale of analysis affects the quantification of species-environment relationships. 2. We therefore assessed the influence of environmental context (i.e. the characteristics of the landscape surrounding a site), varied over a large range of scales (i.e. ambit radii around focal sites), on the analysis and prediction of habitat selection by African elephants in Kruger National Park, South Africa. 3. We focused on the spatial scaling of the elephants' response to their main resources, forage and water, and found that the quantification of habitat selection strongly depended on the scales at which environmental context was considered. Moreover, the inclusion of environmental context at characteristic scales (i.e. those at which habitat selectivity was maximized) increased the predictive capacity of habitat suitability models. 4. The elephants responded to their environment in a scale-dependent and perhaps hierarchical manner, with forage characteristics driving habitat selection at coarse spatial scales, and surface water at fine spatial scales. 5. Furthermore, the elephants exhibited sexual habitat segregation, mainly in relation to vegetation characteristics. Male elephants preferred areas with high tree cover and low herbaceous biomass, whereas this pattern was reversed for female elephants. 6. We show that the spatial distribution of elephants can be better understood and predicted when scale-dependent species-environment relationships are explicitly considered. This demonstrates the importance of considering the influence of spatial scale on the analysis of spatial patterning in ecological phenomena.