Multi-scale predictive habitat suitability modeling based on hierarchically delineated patches: an example for yellow-billed cuckoos nesting in riparian forests,California, USA

The discipline of landscape ecology recognizes the importance of measuring habitat suitability variables at spatial scales relevant to specific organisms. This paper uses a novel multi-scale hierarchical patch delineation method, PatchMorph, to measure landscape patch characteristics at two distinct spatial scales and statistically relate them to the presence of state-listed endangered yellow-billed cuckoos (Coccyzus americanus occidentalis) nesting in forest patches along the Sacramento River, California, USA. The landscape patch characteristics calculated were: patch thickness, area of cottonwood forest, area of riparian scrub, area of other mixed riparian forest, and total patch area. A third, regional spatial variable, delineating the north and south portions of study area was also analyzed for the effect of regional processes. Using field surveys, the landscape characteristics were related to patch occupancy by yellow-billed cuckoos. The area of cottonwood forest measured at the finest spatial scale of patches was found to be the most important factor determining yellow-billed cuckoo presence in the forest patches, while no patch characteristics at the larger scale of habitat patches were important. The regional spatial variable was important in two of the three analysis techniques. Model validation using an independent data set of surveys (conducted 1987–1990) found 76–82% model accuracy for all the statistical techniques used. Our results show that the spatial scale at which habitat characteristics are measured influences the suitability of forest patches. This multi-scale patch and model selection approach to habitat suitability analysis can readily be generalized for use with other organisms and systems.     

The relative impact of forest patch and landscape attributes on black howler monkey populations in the fragmented Lacandona rainforest, Mexico

Land-use change is forcing many animal populations to inhabit forest patches in which different processes can threaten their survival. Some threatening processes are mainly related to forest patch characteristics, but others depend principally on the landscape spatial context. Thus, the impact of both patch and landscape spatial attributes needs to be assessed to have a better understanding of the habitat spatial attributes that constraint the maintenance of populations in fragmented landscapes. Here, we evaluated the relative effect of three patch-scale (i.e., patch size, shape, and isolation) and five landscape-scale metrics (i.e., forest cover, fragmentation, edge density, mean inter-patch isolation distance, and matrix permeability) on population composition and structure of black howler monkeys (Alouatta pigra) in the Lacandona rainforest, Mexico. We measured the landscape-scale metrics at two spatial scales: within 100 and 500 ha landscapes. Our findings revealed that howler monkeys were more strongly affected by local-scale metrics. Smaller and more isolated forest patches showed a lower number of individuals but at higher densities. Population density also tended to be positively associated to matrices with higher proportion of secondary forests and arboreal crops (i.e. with greater permeability), most probably because these matrices can offer supplementary foods. The immature-to-female ratio also increased with matrix permeability, shape complexity, and edge density; habitat characteristics that can increase landscape connectivity and sources availability. The prevention of habitat loss and isolation, and the increment of matrix permeability are therefore needed for the conservation of this endangered Neotropical mammal.     

Partitioning the multi-scale effects of human activity on the occurrence of riparian forest birds

Conservationists, managers, and land planners are faced with the difficult task of balancing many issues regarding humans impacts on natural systems. Many of these potential impacts arise from local-scale and landscape-scale changes, but such changes often covary, which makes it difficult to isolate and compare independent effects arising from humans. We partition multi-scale impacts on riparian forest bird distribution in 105 patches along approximately 500 km of the Madison and Missouri Rivers, Montana, USA. To do so, we coupled environmental information from local (within-patch), patch, and landscape scales reflecting potential human impacts from grazing, invasive plant species, habitat loss and fragmentation, and human development with the distribution of 28 terrestrial breeding bird species in 2004 and 2005. Variation partitioning of the influence of different spatial scales suggested that local-scale vegetation gradients explained more unique variation in bird distribution than did information from patch and landscape scales. Partitioning potential human impacts revealed, however, that riparian habitat loss and fragmentation at the patch and landscape scales explained more unique variation than did local disturbances or landscape-scale development (i.e., building density in the surrounding landscape). When distribution was correlated with human disturbance, local-scale disturbance had more consistent impacts than other scales, with species showing consistent negative correlations with grazing but positive correlations with invasives. We conclude that while local vegetation structure best explains bird distribution, managers concerned with ongoing human influences in this system need to focus more on mitigating the effects of large-scale disturbances than on more local land use issues. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

The influence of forest harvesting on landscape spatial patterns and old-growth-forest fragmentation in southeast British Columbia

Habitat fragmentation is considered one of the major conservation issues of recent decades. We tested predictions of landscape patterns in a 352,253-ha managed forest area in southeast British Columbia. We did this by focussing on forest fragmentation concerns among old-growth, harvest, and wildfire patches in 44 delineated landscapes using patch indices as measures of landscape pattern. We found no significant association between amount of harvesting and 15 old-growth patch indices. Comparisons among patch types revealed that amounts and spatial patterns of harvest patches differed little from amounts and spatial patterns of old-growth patches in control landscapes. Variability indices revealed similar variability between harvest patches and old-growth patches, and more variability between harvest patches and wildfire patches. Little of the evidence gathered in this study supported predictions of fragmentation of old-growth spatial patterns, or predicted differences between harvest spatial patterns and more naturally occurring spatial patterns. We suggest these results could be due to the relatively small amounts of harvesting and old-growth forest in these landscapes, and therefore habitat amount may be a more important factor than spatial configuration of patches in these landscapes.     

Landscape Fragmentation and Ice Storm Damage in Eastern Ontario Forests

With return times between 20 and 100 years, ice storms are a primary disturbance type for temperate forests of eastern North America. Many studies have been conducted at the forest patch and plot scales to examine relations between damage and variables describing site, composition and structure. This paper presents results from a landscape scale study of fragmentation relations with damage in eastern Ontario forests. Data previously collected for two independent and spatially non-overlapping patch level damage studies were used. A Generalized Linear Model (GLM) was used to analyse relations between damage and fragmentation metrics representing patch isolation, edge density, and the relative size and distribution of patches in the landscape. The metrics were applied using spatial extents of 1 × 1 km and 4 × 4 km, following analyses of the variability of numbers of patches and of the lacunarity of forest patterns over a range of extents. The results showed that patch isolation, as measured by the mean Euclidean distance between patches (ENN) was significantly related to damage.     

Effects of increasing landscape heterogeneity on local plant species richness: how much is enough?

Contemporary landscape ecology continues to explore the causes and consequences of landscape heterogeneity across a range of scales, and demands for the scientific underpinnings of landscape planning and management still remains high. The spatial distribution of resources can be a key element in determining habitat quality, and that in turn is directly related to the level of heterogeneity in the system. In this sense, forest habitat mosaics may be more affected by lack of heterogeneity than by structural fragmentation. Nonetheless, increasing spatial heterogeneity at a given spatial scale can also decrease habitat patch size, with potential negative consequences for specialist species. Such dual effect may lead to hump-backed shape relationships between species diversity and heterogeneity, leading to three related assumptions: (i) at low levels of heterogeneity, an increase in heterogeneity favours local and regional species richness, (ii) there is an optimum heterogeneity level at which a maximum number of species is reached, (iii) further increase in spatial heterogeneity has a negative effect on local and regional species richness, due to increasing adverse effects of habitat fragmentation. In this study, we investigated the existence of a hump-shaped relationship between local plant species richness and increasing forest landscape heterogeneity on a complex mosaic in the French Alps. Forest landscape heterogeneity was quantified with five independent criteria. We found significant quadratic relationships between local forest species richness and two heterogeneity criteria indicators, showing a slight decrease of forest species richness at very high heterogeneity levels. Species richness–landscape heterogeneity relationships varied according to the heterogeneity metrics involved and the type of species richness considered. Our results support the assumption that intermediate levels of heterogeneity may support more species than very high levels of heterogeneity, although we were not able to conclude for a systematic negative effect of very high levels of heterogeneity on local plant species richness.     

Identification of Landscape Elements Related to Local Declines of a Boreal Grey-sided Vole Population

Several studies indicate a long-term decline in numbers of different species of voles in northern Fennoscandia. In boreal Sweden, the long-term decline is most pronounced in the grey-sided vole (Clethrionomys rufocanus). Altered forest landscape structure has been suggested as a possible cause of the decline. However, habitat responses of grey-sided voles at the landscape scale have never been studied. We analyzed such responses of this species in lowland forests in Västerbotten, northern Sweden. Cumulated spring densities representing 22 local time series from 1980–1999 were obtained by a landscape sampling design and were related to the surrounding landscape structure of 2.5×2.5 km plots centred on each of the 22 1-ha trapping plots. In accordance with general knowledge on local habitat preferences of grey-sided voles, our study supported the importance of habitat variables such as boulder fields and old-growth pine forest at the landscape scale. Densities were negatively related to clear cuts. Habitat associations were primarily those of landscape structure related to habitat fragmentation, distance between habitat patches and patch interspersion rather than habitat patch type quantity. Local densities of the grey-sided vole were positively and exponentially correlated with spatial contiguity (measured with the fragmentation index) of old-growth pine forest, indicating critical forest fragmentation thresholds. Our results indicate that altered land use might be involved in the long-term decline of the grey-sided vole in managed forest areas of Fennoscandia. We propose two further approaches to reveal and test responses of this species to changes in landscape structure.     

Relating tradable credits for biodiversity to sustainability criteria in a dynamic landscape

Tradable biodiversity credit systems provide flexible means to resolve conflicts between development and conservation land-use options for habitats occupied by threatened or endangered species. We describe an approach to incorporate the influence of habitat fragmentation into the conservation value of tradable credits. Habitat fragmentation decreases gene flow, increases rates of genetic drift and inbreeding, and increases probabilities of patch extinction. Importantly, tradable credit systems will change the level of fragmentation over time for small and/or declining populations. We apply landscape equivalency analysis (LEA), a generalizable, landscape-scale accounting system that assigns conservation value to habitat patches based on patch contributions to abundance and genetic variance at landscape scales. By evaluating habitat trades using two models that vary the relationship between dispersal behaviors and landscape patterns, we show that LEA provides a novel method for limiting access to habitat at the landscape-scale, recognizing that the appropriate amount of migration needed to supplement patch recruitment and to offset drift and inbreeding will vary as landscape pattern changes over time. We also found that decisions based on probabilities of persistence alone would ignore changes in migration, genetic drift, and patch extinction that result from habitat trades. The general principle of LEA is that habitat patches traded should make at least equivalent contributions to rates of recruitment and migration estimated at a landscape scale. Traditional approaches for assessing the “take” and “jeopardy” standards under the Endangered Species Act based on changes in abundance and probability of persistence may be inadequate to prevent trades that increase fragmentation.     

Temporal and spatial changes in an area of the New Jersey Pine Barrens landscape

In order to document the extent of landscape fragmentation for a section of the New Jersey Pine Barrens region, we have used satellite image and spatial analysis to monitor landscape change between 1972 and 1988. Land-cover patterns were quantified by mean, number, and size of patches; and amount of edges between land cover types. During the intervening sixteen year period, fractal dimension, diversity, and contagion generally decreased while dominance, disturbance and edges increased, indicating a trend to a more dissected and disturbed landscape. There was an increase in the number of forest patches and a significant decrease in the average size of forest patches. In contrast, the mean patch size for the non-forest category has increased as a result of a coalescence of patches. The landscape fragmentation is shown by a downward shift in the distribution of forest patches by size class. These changes in landscape pattern have implications for many ecological processes and resources. Management practices need to consider landscape fragmentation in the Pinelands National Reserve in order to preserve the essential character of the Pine Barrens landscape.     

Separating the effects of habitat amount and fragmentation on invertebrate abundance using a multi-scale framework

Context

Herbicide treatments in viticulture can generate highly contrasting mosaics of vegetated and bare vineyards, of which vegetated fields often provide better conditions for biodiversity. In southern Switzerland, where herbicides are applied at large scales, vegetated vineyards are limited in extent and isolated from one another, potentially limiting the distribution and dispersal ability of organisms.

Objectives

We tested the separate and interactive effects of habitat amount and fragmentation on invertebrate abundance using a multi-scale framework, along with additional environmental factors. We identified which variables at which scales were most important in predicting patterns of invertebrate abundance.

Methods

We used a factorial design to sample across a gradient of habitat amount (area of vegetated vineyards, measured as percentage of landscape PLAND) and fragmentation (number of vegetated patches, measured as patch density PD). Using 10 different spatial scales, we identified the factors and scales that most strongly predicted invertebrate abundance and tested potential interactions between habitat amount and fragmentation.

Results

Habitat amount (PLAND index) was most important in predicting invertebrate numbers at a field scale (50 m radius). In contrast, we found a negative effect of fragmentation (PD) at a broad scale of 450 m radius, but no interactive effect between the two.

Conclusions

The spatial scales at which habitat amount and fragmentation affect invertebrates differ, underpinning the importance of spatially explicit study designs in disentangling the effects between habitat amount and configuration. We showed that the amount of vegetated vineyards has more influence on invertebrate abundance, but that fragmentation also contributed substantially. This suggests that efforts for augmenting the area of vegetated vineyards is more beneficial for invertebrate numbers than attempts to connect them.

     

Landscape contrast: a solution to hidden assumptions in the metacommunity concept?

The metacommunity concept provides a spatial perspective on community dynamics, and the landscape provides the physical template for a metacommunity. Several aspects of landscape heterogeneity, such as landscape diversity and composition, and characteristics of the matrix between habitat patches such as habitat connectivity, and geometry of habitat patches, may moderate metacommunity processes. These aspects of landscape heterogeneity are rarely considered explicitly in the metacommunity discussion, however. We propose landscape contrast (i.e., the average dissimilarity in habitat quality between neighboring patches) as a key dimension of landscape heterogeneity. The concept of landscape contrast unifies discrete and continuous landscape representations (homogeneous, gradient, mosaic and binary) and offers a means to integrate landscape heterogeneity in the metacommunity concept. Landscape contrast as perceived by the organisms affects several fundamental metacommunity processes and may thus constrain which metacommunity models may be observed. In a review of empirical metacommunity studies (n = 123), only 22 % of studies were explicit about their underlying landscape model assumptions, with striking differences among taxonomic groups. The assumed landscape model constrained, but did not determine, metacommunity models. Integration and explicit investigation of landscape contrast effects in metacommunity studies are likely to advance ecological theory and facilitate its application to real-world conservation problems.     

How do local habitat management and landscape structure at different spatial scales affect fritillary butterfly distribution on fragmented wetlands?

Habitat fragmentation, patch quality and landscape structure are important predictors for species richness. However, conservation strategies targeting single species mainly focus on habitat patches and neglect possible effects of the surrounding landscape. This project assesses the impact of management, habitat fragmentation and landscape structure at different spatial scales on the distribution of three endangered butterfly species, Boloria selene, Boloria titania and Brenthis ino. We selected 36 study sites in the Swiss Alps differing in (1) the proportion of suitable habitat (i.e., wetlands); (2) the proportion of potential dispersal barriers (forest) in the surrounding landscape; (3) altitude; (4) habitat area and (5) management (mowing versus grazing). Three surveys per study site were conducted during the adult flight period to estimate occurrence and density of each species. For the best disperser B. selene the probability of occurrence was positively related to increasing proportion of wetland on a large spatial scale (radius: 4,000 m), for the medium disperser B. ino on an intermediate spatial scale (2,000 m) and for the poorest disperser B. titania on a small spatial scale (1,000 m). Nearby forest did not negatively affect butterfly species distribution but instead enhanced the probability of occurrence and the population density of B. titania. The fen-specialist B. selene had a higher probability of occurrence and higher population densities on grazed compared to mown fens. The altitude of the habitat patches affected the occurrence of the three species and increasing habitat area enhanced the probability of occurrence of B. selene and B. ino. We conclude that, the surrounding landscape is of relevance for species distribution, but management and habitat fragmentation are often more important. We suggest that butterfly conservation should not focus only on a patch scale, but also on a landscape scale, taking into account species-specific dispersal abilities.     

Small urban woodlands as biodiversity conservation hot-spot: a multi-taxon approach

To evaluate the importance of urban woodlands to serve as potential sites for biodiversity conservation, we analysed bird, carabid beetle and small mammal community responses to urbanisation at different spatial scales. We analysed the relationships between the variations of the structure (species richness S, diversity H′ and dominance D) of animal communities of woodlands distributed along a rural–urban gradient, and the variations along this same gradient of (1) the vegetation within woodlands, (2) the landscape at 100 m and (3) 600 m around the woodlands. We identified the spatial scales whose variations along the gradient most affected each animal community structure, and characterised community responses to these variations. Our results showed that urbanisation affected taxa differently according to their dispersal ability. Carabid beetles, less mobile, seem to be sensitive to increasing fragmentation and built surfaces from periurban to town centre which could make their movement within the urban landscape difficult. Birds, mobile species, seem to be more sensitive to variations of the vegetation structure within woodlands from periurban to town centre that could affect their capacity to maintain in habitat patches. Although our study did not allow relating the small mammal community structure to urbanisation, it suggests that this taxa is sensitive to urban local disturbances. A relevant management scale of woodlands can be specified for each taxa conservation. Urban woodlands accommodate over 50% of the species present in periurban woodlands, and effective management could enhance this number. Woodlands seem to be a good choice for promoting biodiversity conservation in towns.     

Forest loss and matrix composition are the major drivers shaping dung beetle assemblages in a fragmented rainforest

Context

Identifying the drivers shaping biological assemblages in fragmented tropical landscapes is critical for designing effective conservation strategies. It is still unclear, however, whether tropical biodiversity is more strongly affected by forest loss, by its spatial configuration or by matrix composition across different spatial scales.

Objectives

Assessing the relative influence of forest patch and landscape attributes on dung beetle assemblages in the fragmented Lacandona rainforest, Mexico.

Methods

Using a multimodel inference approach we tested the relative impact of forest patch size and landscape forest cover (measures of forest amount at the patch and landscape scales, respectively), patch shape and isolation (forest configuration indices at the patch scale), forest fragmentation (forest configuration index at the landscape scale), and matrix composition on the diversity, abundance and biomass of dung beetles.

Results

Patch size, landscape forest cover and matrix composition were the best predictors of dung beetle assemblages. Species richness, beetle abundance, and biomass decreased in smaller patches surrounded by a lower percentage of forest cover, and in landscapes dominated by open-area matrices. Community evenness also increased under these conditions due to the loss of rare species.

Conclusions

Forest loss at the patch and landscape levels and matrix composition show a larger impact on dung beetles than forest spatial configuration. To preserve dung beetle assemblages, and their key functional roles in the ecosystem, conservation initiatives should prioritize a reduction in deforestation and an increase in the heterogeneity of the matrix surrounding forest remnants.

     

Communities in context: the influences of multiscale environmental variation on local ant community structure

We explored the ways in which environmental variation at multiple spatial scales influences the organization of ant species into local communities. Ground-dwelling ants were sampled in sandhill habitat at 33 locations throughout northern Florida, USA. Variance partitioning of local, landscape, and regional datasets using partial redundancy analysis indicates that ant community composition is significantly influenced by environmental variability across all scales of analysis. Habitat generalists appear to replace habitat specialists at sites with high proportions of matrix habitat in the surrounding landscape. Conversely, habitat specialists appear to replace habitat generalists at sites with more sandhill habitat in the surrounding landscape and greater amounts of bare ground locally. Local niche differentiation leading to species-sorting, combined with the effects of spatially structured dispersal dynamics at landscape scales, may explain this pattern of community structure. Regional influences on local ant communities were correlated with geographical and environmental gradients at distinct regional scales. Therefore, local ant communities appear to be simultaneously structured by different processes that occur at separate spatial scales: local, landscape, and regional scales defined by spatial extent. Our results illustrate the importance of considering multiscale influences on patterns of organization in ecological communities. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Influence of patch and landscape characteristics on the distribution of the subterranean rodent <Emphasis Type="Italic">Ctenomys porteousi</Emphasis>

The understanding and prediction of the responses of animal populations to habitat fragmentation is a central issue in applied ecology. The identification of habitat variables associated to patch occupancy is particularly important when habitat quality is affected by human activities. Here, we analyze the influence of patch and landscape characteristics on patch occupancy by the subterranean herbivorous rodent Ctenomys porteousi. Patch occupancy was monitored in a network of 63 habitat patches identified by satellite imagery analysis which extends along almost the whole distributional range for C. porteousi. Suitable habitat for the occurrence of C. porteousi is highly fragmented and represents <10% of the total area in its distributional range. The distribution of C. porteousi in the patch network is affected not only by characteristics of the habitat patches, but also by those of the surrounding landscape matrix. Significant differences between occupied and empty patches were found in several environmental variables. Overall, occupied patches were larger, less vegetated, more connected, and had larger neighbor patches than empty patches. A stepwise procedure on a generalized linear model selected four habitat variables that explain patch occupancy in C. porteousi; it included the effects of habitat quality in the matrix surrounding the patch, average vegetation cover in the patch, minimum vegetation cover in the matrix surrounding the patch, and the area of the nearest neighbor patch. These results indicate that patch occupancy in C. porteousi is strongly influenced by the availability and quality of habitat both in the patch and in the surrounding landscape matrix.     

Living on the edge: quantifying the structure of a fragmented forest landscape in England

Forest ecosystems have been widely fragmented by human land use, inducing significant microclimatic and biological changes at the forest edge. If we are to rigorously assess the ecological impacts of habitat fragmentation, there is a need to effectively quantify the amount of edge habitat within a landscape, and to allow this to be modelled for individual species and processes. Edge effect may extend only a few metres or as far as several kilometres, depending on the species or process in question. Therefore, rather than attempting to quantify the amount of edge habitat by using a fixed, case-specific distance to distinguish between edge and core, the area of habitat within continuously-varying distances from the forest edge is of greater utility. We quantified the degree of fragmentation of forests in England, where forests cover 10 % of the land area. We calculated the distance from within the forest patches to the nearest edge (forest vs. non-forest) and other landscape indices, such as mean patch size, edge density and distance to the nearest neighbour. Of the total forest area, 37 % was within 30 m and 74 % within 100 m of the nearest edge. This highlights that, in fragmented landscapes, the habitats close to the edge form a considerable proportion of the total habitat area. We then show how these edge estimates can be combined with ecological response functions, to allow us to generate biologically meaningful estimates of the impacts of fragmentation at a landscape scale.     

Utility of landscape mosaics and boundaries in forest conservation decision making in the Atlantic Forest of Brazil

We evaluated changes in the Atlantic Forest landscape over the last 40 years based on changes in boundaries and mosaics, including the hypothetical landscape resulting from the application of Brazilian laws for forest protection. Mosaics were identified as sets of land-use patches with a similar pattern of boundaries. Landscapes of different years, therefore, can be distinguished by differences in mosaics. We developed a technique to identify boundaries between patches from land-use maps using ArcGis^® and to build the patch x boundary matrix required for mosaic identification by means of a factorial and cluster analysis. The mosaics were characterized by some key uses as well as by their boundaries with other land uses. The mosaics were scored for forest conservation according to five issues: landscape permeability, cover, availability, quality, and fragmentation of forest. The values were based on land use and boundary patterns. Although Brazilian laws regarding forest protection have promoted conservation and the hypothetical legal landscape has presented the highest forest habitat availability, this expansion perpetuates a boundary pattern that complicates conservation and management, thus increasing the pressure on forest patches and favoring the further fragmentation of protected forest patches. These conclusions cannot be reached by simply recording changes in land uses.     

Matrix is important for mammals in landscapes with small amounts of native forest habitat

Acknowledgment that the matrix matters in conserving wildlife in human-modified landscapes is increasing. However, the complex interactions of habitat loss, habitat fragmentation, habitat condition and land use have confounded attempts to disentangle the relative importance of properties of the landscape mosaic, including the matrix. To this end, we controlled for the amount of remnant forest habitat and the level of fragmentation to examine mammal species richness in human-modified landscapes of varying levels of matrix development intensity and patch attributes. We postulated seven alternative models of various patch habitat, landscape and matrix influences on mammal species richness and then tested these models using generalized linear mixed-effects models within an information theoretic framework. Matrix attributes were the most important determinants of terrestrial mammal species richness; matrix development intensity had a strong negative effect and vegetation structural complexity of the matrix had a strong positive effect. Distance to the nearest remnant forest habitat was relatively unimportant. Matrix habitat attributes are potentially a more important indicator of isolation of remnant forest patches than measures of distance to the nearest patch. We conclude that a structurally complex matrix within a human-modified landscape can provide supplementary habitat resources and increase the probability of movement across the landscape, thereby increasing mammal species richness in modified landscapes.     

The landscape matrix modifies the effect of habitat fragmentation in grassland butterflies

The landscape matrix is suggested to influence the effect of habitat fragmentation on species richness, but the generality of this prediction has not been tested. Here, we used data from 10 independent studies on butterfly species richness, where the matrix surrounding grassland patches was dominated by either forest or arable land to test if matrix land use influenced the response of species richness to patch area and connectivity. To account for the possibility that some of the observed species use the matrix as their main or complementary habitat, we analysed the effects on total species richness and on the richness of grassland specialist and non-specialist (generalists and specialists on other habitat types) butterflies separately. Specialists and non-specialists were defined separately for each dataset. Total species richness and the richness of grassland specialist butterflies were positively related to patch area and forest cover in the matrix, and negatively to patch isolation. The strength of the species-area relationship was modified by matrix land use and had a slope that decreased with increasing forest cover in the matrix. Potential mechanisms for the weaker effect of grassland fragmentation in forest-dominated landscapes are (1) that the forest matrix is more heterogeneous and contains more resources, (2) that small grassland patches in a matrix dominated by arable land suffer more from negative edge effects or (3) that the arable matrix constitutes a stronger barrier to dispersal between populations. Regardless of the mechanisms, our results show that there are general effects of matrix land use across landscapes and regions, and that landscape management that increases matrix quality can be a complement to habitat restoration and re-creation in fragmented landscapes.