Habitat use by a Neotropical desert marsupial (Thylamys pallidior): A multi-scale approach

In South America, most didelphid marsupials are associated with tropical and moist forests; however, a few species are adapted to more extreme environments such as desert systems. The desert mouse opossum, Thylamys pallidior, is a small marsupial which inhabits the arid regions of South America and has the broadest distribution of the genus. The objective of this research was to assess the habitat use and selection by the desert marsupial T. pallidior considering different scales and habitat levels of analysis, as well as seasonality. We considered two habitat levels (macro and microhabitat) and the microhabitat was analyzed at two spatial scales (1 ha and 4 m²). We recorded 95 captures of mouse opossums in 27,600 trap-nights. Our study showed that habitat type was not a significant factor for selection; instead mouse opossums selected areas with low complexity and high grass cover. The scale of analysis was also important since the selection was stronger in larger areas. Habitat selection by T. pallidior occurred mainly during the dry season. During the wet season, the species used all habitat areas and types indistinctly. The present study highlights the importance of performing analyses which include multiple spatial and temporal scales for a better understanding of the patterns involved in habitat use and species distribution.     

High-use movement pathways and habitat selection by ungulates

The cumulative movements of large mammals are expressed in many areas as semi-permanent wildlife trails. The mapping of semi-permanent trail networks offers a direct approach to assess habitat selection of high-use movement routes at relatively fine spatial scales across a landscape. Here we examine an ungulate trail network in north-central Utah created and maintained by the repeated movements of mule deer (Odocoileus hemionus) and elk (Cervus elaphus). In a resource selection analysis using multivariable spatial regression analysis, we show that at a spatial scale of 70 m open and low cover and distance to water are important predictors of movement pathway density. We also demonstrate at a scale of 10 m that elk and deer movement pathways are less steep than adjacent terrain. The mapping of trail networks should be a particularly useful technique for examining functional connectivity among resource patches across a landscape and identifying important high-use movement routes.     

Habitat suitability modelling for species at risk is sensitive to algorithm and scale: A case study of Blanding's turtle,Emydoidea blandingii,in Ontario,Canada

Species distribution modelling (SDM) can help conservation by providing information on the ecological requirements of species at risk. We developed habitat suitability models at multiple spatial scales for a threatened freshwater turtle, Emydoidea blandingii, in Ontario as a case study. We also explored the effect of background data selection and modelling algorithm selection on habitat suitability predictions. We used sighting records, high-resolution land cover data (25 m), and two SDM techniques: boosted regression trees; and maximum entropy modelling. The area under the receiver characteristic operating curve (AUC) for habitat suitability models tested on independent data ranged from 0.878 to 0.912 when using random background and from 0.727 to 0.741 with target-group background. E. blandingii habitat suitability was best predicted by air temperature, wetland area, open water area, road density, and cropland area. Habitat suitability increased with increasing air temperature and wetland area, and decreased with increasing cropland area. Low road density and open water increased habitat suitability, while high levels of either variable decreased habitat suitability. Robust habitat suitability maps for species at risk require using a multi-scale and multi-algorithm approach. If well used, SDM can offer insight on the habitat requirements of species at risk and help guide the development of management plans. Our results suggest that E. blandingii management plans should promote the protection of terrestrial habitat surrounding residential wetlands, halt the building of roads within and adjacent to currently occupied habitat, and identify movement corridors for isolated populations.     

Tree visitation and seed dispersal of wild cherries by terrestrial mammals along a human land-use gradient

The role of terrestrial mammals as seed dispersers of fleshy-fruited plants has only rarely been investigated in temperate regions although recent studies underline the importance of these animals for long-distance seed dispersal. Here we examine the potential role of mammals as seed dispersers of wild cherry (Prunus avium L.) along a gradient of human land-use intensity. We placed camera traps at 21 wild cherry trees to identify the mammal species that visited the trees. We conducted feeding trials to test if the recorded species were legitimate seed dispersers or seed predators and to assess gut passage times. We tested the influence of human land-use intensity by quantifying habitat and landscape structure around the study trees at different spatial scales and analyzing its influence on visitation rates of tree visitors. Red fox (Vulpes vulpes), roe deer (Capreolus capreolus), wild boar (Sus scrofa), marten (Martes spp.) and badger (Meles meles) were identified as seed dispersers of wild cherry, of which wild boar was largely a seed predator. Habitat and landscape structure at local spatial scales (70 m, 500 m radius) had no effect on the total visitation rates of mammals. At larger spatial scales (1.0–10.0 km radius) total visitation rates increased with increasing proportion of extensively used farmland and seminatural habitat in the area. The proportion of forest had no influence on visitation rates. The results suggest that high proportions of extensively used farmland and seminatural habitat increase the visitation and seed dispersal rates of large mammals. Comparing gut passage times with home range sizes and daily travel distances suggests that these mammals have the potential to disperse seeds over long distances and can provide gene flow in fragmented agricultural landscapes.     

Determinants of smooth-coated otter occupancy in a rapidly urbanizing coastal landscape in Southeast Asia

Urbanization often has negative impacts on wildlife, nevertheless many species can persist in heavily modified habitats. Understanding factors that promote species persistence in urbanizing landscapes is therefore important for maintaining biodiversity in changing landscapes and may inform more biodiversity-friendly development. We investigated effects of landscape-scale variables on habitat occupancy of Smooth-coated otter (Lutrogale perspicillata) in the Inner Gulf of Thailand. In this internationally important wetland complex, 86% of natural habitats have been altered and are now highly urbanized. We conducted track and sign surveys in 60, 25 km² grid cells encompassing 1,474 km². Within each cell, we quantified the landscape cover types including urban areas, natural habitat (predominately thin ribbons of mangrove along waterways), agriculture, aquaculture, and potential prey availability using a GIS and field surveys. We used occupancy models to identify habitat variables that affected probability of detection and occupancy. Estimated otter occupancy, based on the top model, was 0.33 ± 0.07 (95% CI 0.18–0.48) and detection probability was 0.50 ± 0.05 (95% CI 0.41–0.60). Otter occupancy was positively associated with the proportion of natural habitat and the cover of traditional aquaculture ponds, but negatively associated with agriculture and urban cover. The remaining natural patches appear to act as critical refuges for otter, allowing them to persist in an otherwise heavily transformed landscape. Because aquaculture ponds are likely important prey sources for otters, it may lead to conflict with aquaculture farmers. Further studies of feeding and movement patterns in cooperation with aquaculture farmers would be beneficial for developing detailed management plans for the species in this human dominated landscape.     

Vertical habitat segregation as a mechanism for coexistence in sympatric rodents

Coexistence has been widely studied in small mammals and frequently is assumed to be facilitated by habitat segregation. Using live trapping and spool-and-line experiments, we analyzed habitat selection and segregation across multiples scales for Peromyscus leucopus and Ochrotomys nuttalli. At the habitat scale of a forest stand (∼1–100s ha) P. leucopus co-occurred at all sites where O. nuttalli was found, and we did not detect evidence of positive or negative associations or habitat segregation. However, O. nuttalli was restricted to early successional forests, and P. leucopus had significantly lower abundances in early successional forests than in other habitats. We found similar patterns at the mesohabitat scale of the study site (400 m²). O. nuttalli abundance increased with increasing shrub and tree densities, while increases in P. leucopus abundance were associated with open understories. At the microhabitat scale of the individual movement trail, we found vertical segregation. Movement trails for O. nuttalli were at significantly higher elevation (mean height = 142.93 ± 37.10 cm) than P. leucopus trails (mean height = 15.4 ± 4.98 cm; F-value = 35.29, p < 0.001). We concluded that microhabitat segregation was driven by differential use of vertical space for movement and foraging. We suggest O. nuttalli is superior to P. leucopus in its ability to acquire food especially in shrubby subcanopies where few acorns are available. However, P. leucopus is superior in its ability to forage and avoid predators.     

Winter foraging habitat selection of brown-eared pheasant (Crossoptilon mantchuricum) and the common pheasant (Phasianus colchicus) in Huanglong Mountains, Shaanxi Province

The foraging habitat selections of brown-eared pheasant (Crossoptilon mantchuricum) and the common pheasant (Phasianus colchicus) were studied in Huanglongshan Nature Reserve Shaanxi, China. Foraging habitat characteristics were measured on the basis of expected differences between species at 183 sites from November to December 2006 and January 2007. The results showed that both species selected foraging habitats with altitude (<1200 m), conifer forest, half sunny and half shady slope, sunny slope, density of trees (<5 individuals/100 m²), cover of shrub (>50%), visibility class (<10%) and distance to water source (<300 m). However, the brown-eared pheasant selected habitats with cover of trees (30–50%), middle or lower slope location, distance to edge of woods (<300 m) and human disturbance (<500 m), and the selection on density of shrub was not observed, compared to the selections on cover of trees (<30%), lower slope location, distance to edge of woods (<500 m) and human disturbance (<300 m), and density of shrub (>500 individuals/100 m²) for common pheasant. We also found that the common pheasant avoid predators by concealment whereas brown-eared pheasant evade predations by running away strategy.     

Mapping multiple plant species abundance patterns - A multiobjective optimization procedure for combining reflectance spectroscopy and species ordination

Nature conservation and ecological restoration crucially depends on the knowledge about spatial patterns of plant species that control habitat conversion and disturbance regimes. Especially, species abundances are capable of indicating early development tendencies for setting habitat management strategies. This study demonstrates the transfer of field spectroscopy to hyperspectral imagery to map multiple plant species abundances in an open dryland area using two imaging spectrometers in two different phenological phases. We show that species abundances can partially be described by multiple gradients forming different coordinates in a contour map. For this purpose, species abundances were projected into an ordination space using non-metric multidimensional scaling and subsequent spatial interpolation. It was demonstrated that different gradients can be modeled in a Partial Least Squares regression framework resulting in distinct spectral features for certain gradient directions. We combine both objectives in a multiobjective NSGA-II procedure to maximize the quantitative determination of species abundance in ordination and spectral predictability in related field spectra, simultaneously. NSGA-II was finally used to select optimal spectral models for n = 35 single species that were transferred to hyperspectral imagery for mapping purpose. We can show that abundance predictabilities can be evaluated on the basis of individual model performances that hold different spectral features for each species in a designated phenological phase. Finally, we present spatially explicit multi-species maps for the best n = 18 and abundance maps for n = 8 models that could be linked to patterns of species richness, coexistence, succession stages and habitat type conditions.     

The effects of hoarding habitat selection of Eurasian red squirrels (Sciurus vulgaris) on natural regeneration of the Korean pines

Cone-cores discarded by Eurasian red squirrels were used to study the habitat selection of Korean pine-seeds hoarding, in forest patch Nos. 16 and 19 in Liangshui Nature Reserve, China. Ten transects with a total length of 15 km were uniformly set, and data from 343 valid samples were collected in a 369 hm² area. One hundred and eighty four were hoarding samples which were determined according to the cluster analysis based on the number of the cone-cores, while the other 159 were control samples. The principal component analysis, using 11 habitat factors, suggested that the distance from Korean pine forest, forest type, number of Korean pine seedlings, density and type of bush significantly influenced the habitat selection of hoarding by Eurasian red squirrels. The results of Bailey’s method indicated that the squirrels showed (1) preference for natural coniferous forest, natural fir and spruce forest and planted spruce forest; (2) avoidance of planted Korean pine forest and planted larch forest; and (3) random use of natural Korean pine forest. Moreover the distance from the Korean pines in the range of 150–600 m showed no effect on the habitat selection of hoarding by the Eurasian red squirrels. More than 50% of the cone-cores were discarded in either fringe or gap of the Korean pine forest with more cone-cores found at <300 m than at 300 m away (One-Way ANOVA; df = 3, 183, F = 5.76, p = 0.0009). This demonstrated that the Eurasian red squirrels could take the cone-cores out of the Korean pine forest. The density of bushes in samples of hoarding area was significantly lower than that in control samples (Kruskal–Wallis test; df = 1, χ² = 83.99, p < 0.0001). The number of the Korean pine seedlings in samples of hoarding area was significantly higher than that in the control samples (Kruskal–Wallis test; df = 1, χ² = 104.13, p < 0.0001). This illustrated that the hoarding habitat favoured the germination of the Korean pine seedlings. In conclusion the behavior of hoarding Korean pine seeds by the Eurasian red squirrels can promote the regeneration and dispersal of the Korean pines.     

Improving and evaluating the soil cover indicator for agricultural land in Canada

A soil cover days (SCD) model has been developed by Agriculture and Agri-Food Canada for use as an agri-environmental indicator to monitor the relationship between agricultural production activities and agri-environmental quality. The SCD indicator integrates information on crops, soils, climate, and field activities to estimate the total equivalent number of days that agricultural soils are covered by crop canopy, crop residue and snow in a given year. Daily cover fractions of plant and residue for a given crop in an ecoregion are simulated using typical crop calendar and field management practices, and the equivalent number of days that soil is covered by snow in winter is derived from long term climate normals. The equivalent SCD for a spatial unit is then derived as the area-weighted sum of different crops and different management practices within the unit. This paper presents the SCD framework, details an assessment of the accuracy of the model and outlines future improvements. Annual snow days derived from 30-year climate normals as used in the model was strongly correlated (excluding mountain areas) with that derived from satellite data (R² = 0.45, n = 48), even though the remote sensing product showed significant temporal and spatial variability. Crop residue fraction estimated by the model was strongly correlated with field data collected over major crop areas and crop types (R² = 0.74, n = 55), and modelled plant cover fraction was well correlated with that derived from remote sensing data (R² = 0.57, n = 57). Large discrepancies were observed for some samples due to deviation of the actual crop calendar from that estimated using climate normals. National map showing the change in the indicator from 1981 to 2011 reveals changes in crop and residue management practices.     

Changes in the surface roughness of snow from millimetre to metre scales

The roughness of snow influences the movement of air across the snow surface and resulting transfers of energy. Here we focus on the roughness of the snowpack surface to determine its range of variability for different snow conditions (e.g., time since last snowfall), across spatial scales that ranged from 0.01 cm (card) to more than 1000 cm (transect) or more than 5-orders of magnitude, and due to the deposition of aeolian constituents. Digital photogrammetry of snow surfaces was used to compute two roughness metrics at two mountain sites in north-central Colorado. These metrics are the random roughness (RR) that disregards the spatial structure and the fractal dimension (D) computed from variogram analysis.At the crystal scale, D is between 1.67 (card) and 1.60 (board), which increases to 1.77 between 0.1 and 1.0 cm. At longer scales, D is 1.53 (board) to 1.56 (transect). There was no significant change in surface roughness during the accumulation season, with RR values at about 0.002. During the melt season the surface roughness doubled, with the RR values increasing from about 0.002 to 0.004. Snow was more rough parallel to the wind when dunes were present, and roughness varied spatially. The average RR value computed for the white snow surface of 0.014 is substantially greater than the value computed for the red dust surface of 0.0032. Due to undulations of smaller amplitude and as a result of the dust itself, the red dust surface is more random (D is 2.62 versus 2.23 for the white snow). Our results show that there is consistency in roughness over different scales, yet large scale processes (e.g., wind and radiation activity) influence the magnitude of roughness metrics much more than small scale processes (e.g., crystal form and metamorphism).     

Habitat characteristics and trophic structure of benthic macroinvertebrates in a forested headwater stream

Three channel reaches with different habitat characteristics were selected to test the variability in community structure of benthic macroinvertebrates by comparing the relative abundance of functional feeding groups among the reaches. The important factors influencing the spatial and temporal organization of community structure were explored using nonmetric multidimensional scaling (NMS). The habitat characteristics in the reaches were different in terms of habitat type, hydrological factors, and substrate composition. The first headwater reach was classified as a step-pool reach with similar relative areas of riffle and pool habitats. The second mid-reach and the third down reach had greater areas of pool habitat followed by runs and riffles whose proportions were similar between the latter two reaches. The relative abundances of functional feeding groups were different among the surveyed reaches. Gammarid shredders predominated in the upper reach, and chironomid collector–gatherers and collector–filterers were in greater abundance in the two lower reaches. The proportions of gammarids were minor in the mid and downstream reaches. NMS ordination indicated that the proportion of substrates < 8 mm, discharge, and water depth mainly determined the spatial and temporal distribution of samples based on the macroinvertebrate community in the study reaches. These results suggest that different habitat characteristics result in a distinct community structure in each reach.     

Habitat characteristics from local to landscape scales combine to shape vertebrate scavenging communities

Abiotic and biotic factors modulate carcass consumption by scavengers, affecting ecosystem functioning. Habitat structure is arguably a key factor influencing scavenging, but its role remains poorly understood, particularly at small spatial scales. We examine how habitat characteristics at landscape (50–1000 m radius) and local (≤10 m radius) scales around carrion affect the structure of vertebrate scavenging communities. We used remote cameras to monitor the consumption of 151 ungulate carcasses in one temperate (55 carcasses) and two Mediterranean (56 and 40 carcasses) study areas in Spain in 2011–2013. Our results showed complex habitat–scavenger relationships that mainly relied upon the spatial scale, the type of carcass and the study area. While the response of scavenger richness to habitat characteristics was consistent across study areas, the effects of diversity varied regionally at the landscape scale. Large and medium-sized carcasses in open landscapes had lower scavenger richness, likely because open habitats promote vulture dominance. At the local scale, shrub cover lowered scavenger richness and diversity, hindering carrion location by avian scavengers. Our results suggest that the structure of vertebrate scavenging assemblages at carcasses is driven by carcass and habitat characteristics operating as ecological filters at different scales (i.e. local, landscape, and biogeographical), which affect a species’ ability to locate, access and dominate carrion. Understanding the factors underlying the complex habitat–community relationships shown here has implications for managing key ecosystem functions and services. We propose a multi-scale conceptual framework to disentangle scavenger–carcass relationships.     

Summer fish communities in northern Gulf of Mexico estuaries: Indices of ecological condition

We used fish community data from trawl samples collected from >100 estuaries, bayous, and coastal lagoons of the Louisianan Biogeographic Province (Gulf of Mexico) to develop indicators of large-scale ecological condition. One data set, from which we derived reference values for fish community indicators, was based on bottom trawl samples collected from 367 randomly located sites during the summers of 1992–1994. A second trawl data set with similar geographic scope from 2000 to 2004 was used to test the robustness of the indicators derived from the reference data set to new data. We constructed a fish community index (FCI) from three basic indicators: number of species per trawl, total abundance per trawl, and an index of trophic balance among three common feeding guilds. The FCI was not correlated with salinity over a range from freshwater to marine and hypersaline conditions (0–52 psu). Direct correlations between the index and environmental variables generally were weak, although some were significant (p < 0.05). The FCI was negatively correlated with water clarity (secchi depth), water column depth, and sediment toxicity; correlations of the FCI with pH, sediment organic carbon, and sediment silt + clay content were positive. There was a hyperbolic relationship between dissolved oxygen and maximum values of the index, and no significant correlation with watershed land cover at the whole-estuary or estuary-complex scale. Values of all indicators increased between the two time periods. The FCI is a broad indicator of ecological condition for estuaries within the Louisianan Province, with data aggregated at scales ranging from large estuaries to the entire region. Sample density was insufficient to judge performance of the indicators or index at smaller scales.     

Validating Landsat-based landscape metrics with fine-grained land cover data

ContextModerate-grained data may not always represent landscape structure in adequate detail which could cause misleading results. Certain metrics have been shown to be predictable with changes in scale; however, no studies have verified such predictions using independent fine-grained data.ObjectivesOur objective was to use independently derived land cover datasets to assess relationships between metrics based on fine- and moderate-grained data for a range of analysis extents. We focus on metrics that previous literature has shown to have predictable relationships across scales.MethodsThe study area was located in eastern Connecticut. We compared a 1 m land cover dataset to a 30 m resampled dataset, derived from the 1 m data, as well as two Landsat-based datasets. We examined 11 metrics which included cover areas and patch metrics. Metrics were analyzed using analysis extents ranging from 100 to 1400 m in radius.ResultsThe resampled data had very strong linear relationships to the 1 m data, from which it was derived, for all metrics regardless of the analysis extent size. Landsat-based data had strong correlations for most cover area metrics but had little or no correlation for patch metrics. Increasing analysis areas improved correlations.ConclusionsRelationships between coarse- and fine-grained data tend to be much weaker when comparing independent land cover datasets. Thus, trends across scales that are found by resampling land cover are likely to be unsuitable for predicting the effects of finer-scale elements in the landscape. Nevertheless, coarser data shows promise in predicting fine-grained for cover area metrics provided the analysis area used is sufficiently large.     

Restocking a keystone species in a biodiversity hotspot: Recovering the European rabbit on a landscape scale

Rabbit populations in Iberia remain at low densities in several areas in which their endangered predators still coexist, and the recovery of these populations is therefore urgent if the integrity of Iberian Mediterranean ecosystems is to be maintained. The enhancement of wild rabbit populations has been attempted through the use of in situ extensive rabbit captive breeding enclosures (restocking plots), which reduce mortality caused by terrestrial predators and dispersal movements and permit the breeding of young individuals which can then naturally disperse to settle in the surrounding areas. However, their effectiveness, the role of its size, the optimal habitat management that should be promoted around them and the habitat features remains uncertain. Here, we show results from a four year study of an ambitious rabbit restocking plan on a landscape scale. We measured rabbit abundance in a vast area in which thirty-two restocking plots were built to create an initial rabbit population for further dispersion, in addition to an intensive habitat management program. We also compared rabbit abundance between managed and unmanaged UTM cells of 2.5 km × 2.5 km. Our results showed that rabbit abundance was three times higher in managed cells, but four years after restocking, rabbit abundances had only reached the threshold needed to support stable Iberian lynx presence (at least 10 latrines per km^?1) in 9 of the 23 managed cells. Rabbit abundance was strongly affected by habitat treatment and scrub coverage. The increase of shelter was useful in low cover areas but ineligible in places with high scrub cover, where the increase of refuge plus scrub clearing to create pastures improve rabbit abundance more effectively. In the light of our results, restocking plots should be built only in places with suitable habitat, whereas pastures should be created in dense scrublands and refuge in low cover areas.     

Habitat selection by endangered Himalayan musk deer (Moschus chrysogaster) and impacts of livestock grazing in Nepal Himalaya: Implications for conservation

Habitat management within and outside protected areas is a key to effective conservation of wildlife. This is particularly vital for declining wildlife populations within the boundary of conservation areas, while sharing their potential habitat range with foraging livestock. In an effort to understand the habitat selection by Himalayan musk deer (Moschus chrysogaster) and explore any potential impacts of livestock grazing, we conducted the present study in a conservation area of central Nepal Himalaya. We recorded data on musk deer and livestock presence and absence (based on signs of fecal pellet, footprint, and resting site) along the elevational transect with associated topographic features (elevation, slope, aspect, distance to water, and vantage point distance) and vegetation features (tree spp., shrub spp., herb spp., and canopy-cover). Using logistic regression model we found that elevation, aspect, canopy-cover, and tree spp. in the area significantly affect the likelihood of habitat selection by musk deer. In particular, they selected the southern aspect of the area with elevation ≥ 3529 m, canopy-cover ≥ 42%, and with stands of Pinus spp. and Abies spp. Slope and canopy-cover significantly affected the foraging area selection by livestock. They selected the gentler slopes in the northern aspect of the area with altitude < 3529 m and canopy-cover < 42%. Also, presence of one group of herbivore (i.e. musk deer and livestock) was not found to affect the likelihood of habitat selection by the other group. These independent habitat selections are possibly the responses to morphological and behavioral adaptations than to impacts and interactions between these two groups of herbivores. We suggest to avoid any disturbances and livestock grazing on the area that disrupt the resources and conditions likely selected and occupied by musk deer population.     

Defining quantitative stream disturbance gradients and the additive role of habitat variation to explain macroinvertebrate taxa richness

Most studies dealing with the use of ecological indicators and other applied ecological research rely on some definition or concept of what constitutes least-, intermediate- and most-disturbed condition. Currently, most rigorous methodologies designed to define those conditions are suited to large spatial extents (nations, ecoregions) and many sites (hundreds to thousands). The objective of this study was to describe a methodology to quantitatively define a disturbance gradient for 40 sites in each of two small southeastern Brazil river basins. The assessment of anthropogenic disturbance experienced by each site was based solely on measurements strictly related to the intensity and extent of anthropogenic pressures. We calculated two indices: one concerned site-scale pressures and the other catchment-scale pressures. We combined those two indices into a single integrated disturbance index (IDI) because disturbances operating at both scales affect stream biota. The local- and catchment-scale disturbance indices were weakly correlated in the two basins (r = 0.21 and 0.35) and both significantly (p < 0.05) reduced site EPT (insect orders Ephemeroptera, Plecoptera, Trichoptera) richness. The IDI also performed well in explaining EPT richness in the basin that presented the stronger disturbance gradient (R² = 0.39, p < 0.001). Natural habitat variability was assessed as a second source of variation in EPT richness. Stream size and microhabitats were the key habitat characteristics not related to disturbances that enhanced the explanation of EPT richness over that attributed to the IDI. In both basins the IDI plus habitat metrics together explained around 50% of EPT richness variation. In the basin with the weaker disturbance gradient, natural habitat explained more variation in EPT richness than did the IDI, a result that has implications for biomonitoring studies. We conclude that quantitatively defined disturbance gradients offer a reliable and comprehensive characterization of anthropogenic pressure that integrates data from different spatial scales.