A controlled,hierarchical study of habitat fragmentation: responses at the individual,patch, and landscape scale

We compared the performance of individuals and whole populations of meadow voles, Microtus pennsylvanicus, within and between experimentally created habitat fragments of three sizes (1.0, 0.25, and 0.0625 ha) and between a 20-ha fragmented and a 20-ha continuous habitat landscape. We recorded 10,020 captures of 3946 individuals over 17 censuses between June 1993 and October 1994. Five demographic parameters showed significantly different population responses between the two landscapes but no difference in tests comparing fragment size: i.e., mean and peak population densities (the latter, in each of the two growing seasons) averaged 149 to 172% higher, population growth rate averaged 219% higher, and adult recruitment 170% higher in fragmented than in the continuous control landscape. Observations at the individual level (body sizes, rates of reproduction, residence times) suggested that these landscape differences involved enhanced performance of adult females associated with edge habitats rather than differential immigration or emigration. If this turns out to be a common response to fragmentation, the detection of such responses will be greater when comparing fragmented and unfragmented landscapes with qualitatively different structure than for fragments of varied size with differing proportions of edge. That responses to habitat fragmentation may be more evident at the very small (individual) and very large (landscape) scales, but may be obscured at the intermediate spatial scale of fragments, is a proposition that clearly requires more attention.     

The role of landscape structure in species richness distribution of birds,amphibians, reptiles and lepidopterans in Mediterranean landscapes

The parameters referring to landscape structure are essential in any evaluation for conservation because of the relationship that exists between the landscape structure and the ecological processes. This paper presents a study of the relationships between landscape structure and species diversity distribution (estimated in terms of richness of birds, amphibians, reptiles and butterflies) in the region of Madrid, Spain. The results show that the response of species richness to landscape heterogeneity varies depending on the group of species considered. For birds and lepidopterans, the most important factor affecting the distribution of richness of species is landscape heterogeneity, while other factors, such as the specific composition of land use, play a secondary role at this scale. On the other hand, richness of amphibians and reptiles is more closely related to the abundance of certain land-use types. The study highlights the importance of heterogeneity in Mediterranean landscapes as a criterion for landscape planning and for definition of management directives in order to maintain biodiversity.     

The influence of management history and habitat on plant species richness in a rural hemiboreal landscape,Sweden

We explored patterns of plant species richness at different spatialscales in 14 habitats in a Swedish rural landscape. Effects of physicalconditions, and relationships between species richness and management historyreaching back to the 17^th century were examined, using old cadastralmaps andaerial photographs. The most species-rich habitats were dry open semi-naturalgrasslands, midfield islets and road verges. Alpha diversity (species richnesswithin sites) was highest in habitats on dry substrates (excluding bedrock withsparse pines) and beta diversity (species richness among sites) was highest inmoist to wet habitats. Alpha and beta components of species richness tended tobe inversely related among habitats with similar species richness. Managementhistory influenced diversity patterns. Areas managed as grasslands in the17^th and 18^th century harboured more species than areasoutside the villages. We also found significant relationships between speciesrichness and soil type. Silt proved to be the most species-rich topsoil(10–20 cm) in addition to thin soils top of on green- orlimestone bedrock. The variation in species richness due to local relief orform of thesite also showed significant relationships, where flat surfaces had the highestnumber of species. In contrast, no significant relationship was found betweenspecies richness and aspect. Our study suggests that present-day diversitypatterns are much influenced by management history, and that small habitat,e.g., road verges and midfield islets, are important for maintaining speciesrichness.This revised version was published online in May 2005 with corrections to the Cover Date.     

Estimation and prediction of plant species richness in a mosaic landscape

Traditional agricultural mosaic landscapes are likely to undergo dramatic changes through either intensification or abandonment of land use. Both developmental trends may negatively affect the vascular plant species richness of such landscapes. Therefore, sustainable land-use systems need to be developed to maintain and re-establish species richness at various spatial scales. To evaluate the sustainability of specific land-use systems, we need approaches for the effective assessment of the present species richness and models that can predict the effects on species richness as realistically as possible. In this context, we present a methodology to estimate and predict vascular plant species richness at the local and the regional scale. In our approach, the major determinants of vascular plant species richness within the study area are taken into consideration: These are according to Duelli's mosaic concept the number of habitat types and of habitat patches within area units. Furthermore, it is based on the relative frequencies of species within habitat types. Our approach comprises six steps: (i) the determination of present habitat patterns within an observation area, (ii) the creation of a land-use scenario with simulated habitat patterns, (iii) the determination of species frequencies within habitat types of this area, (iv) a grouping of habitat-specific species, (v) the estimation of the probabilities for all species (or habitat specialists) to occur, either in stepwise, exponentially enlarged landscape tracts (local scale), or in the entire observation area (regional scale), and (vi) the validation of the estimated species numbers. The approach will be exemplified using data from the municipal district of Erda, Lahn-Dill Highlands, Germany. The current species numbers to be expected on the basis of probability calculations were compared with those recorded on the basis of extensive field work. This comparison shows that, on the basis of our simple calculations, the current local plant species richness can be predicted well, with a slight underestimation. This revised version was published online in May 2005 with corrections to the Cover Date. This revised version was published online in July 2006 with corrections to the Cover Date.     

Quantifying determinants contributing to plant species richness in mosaic landscapes: a single- and multi-patch perspective

Despite good theoretical knowledge about determinants of plant species richness in mosaic landscapes, validations based on complete surveys are scarce. We conducted a case study in a highly fragmented, traditional agricultural landscape. In 199 patches of 20 representative multi-patch-plots (MPPs, 1 ha) we recorded a total of 371 plant species. In addition to an additive partitioning of species diversity at the (a) patch- and (b) MPP-scale, we adopted the recently proposed ‘specificity’ measure to quantify the contribution of a spatial subunit to landscape species richness (subunit-to-landscape-contribution, SLC). SLC-values were calculated at both scales with respect to various spatial extents. General regression models were used to quantify the relative importance of hypothesis-driven determinants for species richness and SLC-values. At the patch scale, habitat type was the main determinant of species richness, followed by area and elongated shape. For SLC-values, area was more important than habitat type, and its relevance increased with the extent of the considered landscape. Influences of elongated shape and vegetation context were minor. Differences between habitat types were pronounced for species richness and also partly scale-dependent for SLC-values. Relevant predictors at the MPP-scale were nonlinear habitat richness, the gradient from anthropogenic to seminatural vegetation, and the proportions of natural vegetation and rare habitats. Linear elements and habitat configuration did not contribute to species richness and SLC. Results at the MPP-scale were in complete accordance with the predictions of the mosaic concept. Hence, our study represents its first empirical validation for plant species diversity in mosaic landscapes.     

Comparing the role of site disturbance and landscape properties on understory species richness in fragmented periurban Mediterranean forests

We hypothesized that the spatial configuration and dynamics of periurban forest patches in Barcelona (NE of Spain) played a minor role in determining plant species richness and assemblage compared to site conditions, and particularly to both direct (measured at plot level) and potential (inferred from landscape metrics) human-associated site disturbance. The presence of all understory vascular plants was recorded on 252 plots of 100 m² randomly selected within forest patches ranging in size from 0.25 ha to 218 ha. Species were divided into 6 groups, according to their ecology and conservation status. Site condition was assessed at plot level and included physical attributes, human-induced disturbance and Quercus spp. tree cover. Landscape structure and dynamics were assessed from patch metrics and patch history. We also calculated a set of landscape metrics related to potential human accessibility to forests. Results of multiple linear regressions indicated that the variance explained for non-forest species groups was higher than for forest species richness. Most of the main correlates corresponded to site disturbance variables related to direct human alteration, or to landscape variables associated to indirect human effects on forests: Quercus tree cover (a proxy for successional status) was the most important correlate of non-forest species richness, which decreased when Quercus tree cover increased. Human-induced disturbance was an important correlate of synanthropic and total species richness, which were higher in recently managed and in highly frequented forests. Potential human accessibility also affected the richness of most species groups. In contrast, patch size, patch shape and connectivity played a minor role, as did patch history. We conclude that human influence on species richness in periurban forests takes place on a small scale, whereas large-scale effects attributable to landscape structure and fragmentation are comparatively less important. Implications of these results for the conservation of plant species in periurban forests are discussed.     

Pattern of Kirtland's warbler occurrence in relation to the landscape structure of its summer habitat in northern Lower Michigan

Studies of the endangered Kirtland's warbler in relation to landscape ecosystems were conducted from 1986–1988 on a large wildfire-burn surrounding Mack Lake in southeastern Oscoda County, Michigan. A landscape ecosystem approach was used to distinguish low- and high-elevation segments of the landscape, as well as 11 local ecosystem types. The ecosystems were distinguished by physiography, microclimate, soil, and vegetation. The early occurrence of the warblers was strongly related to landscape structure, i.e., to the broad low- and high-elevation areas and the local ecosystem types within them. Territories of male warblers were observed in 5 of the 11 ecosystems. The five ecosystem types where warblers were observed were characterized by (1) a physiography of level or rolling terrain; (2) soil series of Grayling, Graycalm, Montcalm, or Rubicon; (3) uplands with relatively warm temperature during the breeding season; (4) vegetation dominated by low sweet blueberry, bearberry, wintergreen, northern pin oak, blue stem grasses, and hair cap moss; and (5) canopy of relatively tall, dense, and patchy jack pine and oak. Landscape structure appears to be an important factor affecting the occurrence of the warbler in its summer habitat in northern Lower Michigan.     

Threshold levels of habitat composition for the presence of the long-tailed tit (Aegithalos caudatus) in a boreal landscape

We assessed the habitat patch occupancy of a deciduous-mixed forest specialist, the long-tailed tit (Aegithalos caudatus), in a 1000 km² conifer dominated landscape in relation to two landscape parameters, namely proportion and isolation of suitable habitat. Data from five consecutive spring seasons were used and within habitat variation controlled for. The occurrence of long-tailed tits was positively related to the amount of habitat within 1 km² (p=0.0007) and negatively related to the distance between habitat patches (p<0.0001). When combined, the two variables explained >78% of the variation in local patch occupancy. There were distinct thresholds in these landscape variables for the probability of local long-tailed tit presence. In the model the probability increased from 0.1 to 0.8 when interpatch distance decreased from 500 to 100 m with 5% total habitat coverage. With a total proportion of 15% suitable habitat, the same probability jump occurred when interpatch distance changed from 900 to 500 m. The general importance of defined measurements and quantified threshold levels for species conservation and landscape management is discussed.