Abstract
Farmland biodiversity may have declined because agricultural intensification has reduced ecological heterogeneity at multiple spatial scales. If so, agri-environment schemes (AES) that enhance farmland heterogeneity may be a way to restore farmland biodiversity. The aim of this project was to test the hypothesis that bird species richness and abundance in farmland depend on ecological heterogeneity at multiple spatial scales and that that consequence of habitat heterogeneity at one scale interact with that at other scales. To this end I investigated how ecological heterogeneity at multiple scales affected species richness, density, population dynamics and foraging of farmland birds. Material and methods We used information about farmland bird communities at several spatial scales. First, the Swedish Bird Survey, with yearly updated data on breeding birds across Sweden. Second, territory mapping of 136 1-km2 farmed landscapes in southern Sweden. Third, point-counts of birds in farmland of different heterogeneity (homogeneous plains, mixed farmland, forest-dominated farmland), with within-landscape variation in the amount and character of small biotopes. Fourth, point-counts of birds from organic and conventional farms in a matched-pairs design in landscapes of different ecological heterogeneity. To classify land-use we incorporated land use data and spatially explicit land parcel information (from the Swedish Integrated Administration and Control System, IACS) into an efficient and versatile database and used a GIS-tool to extract land use data with high flexibility. We performed large-scale analyses of land-use from aerial photographs of landscapes in Scania and compared and combined this information with the IACS to determine descriptors of farming intensity, farmland heterogeneity and their relationship across multiple scales. We performed population studies on some declining farmland birds, the house sparrow, the starling and the grey partridge. For these species we determined population data from targeted inventories and monitoring. We studied house-sparrow and starling habitat use and foraging in contrasting landscapes. We evaluated mitigation measures for partridges at landscape scales. Results By relating data from the Swedish Bird Survey to spatially explicit land-use data, we found that sometimes high numbers (up to 90%) of farmland birds (defined by the Farmland Bird Index) actually breed outside farmland (9). However, temporal trends were similar in farmland and non-farmland and unrelated to the proportion of a species that resided in farmland. Trends of species highly dependent on farmland were more variable. In southern Sweden, landscape heterogeneity and farming intensity, although correlated, represented different dimensions (4). Diversity and species richness of birds in those landscapes were primarily related to habitat structure and less to farming intensity per se (Olsson et al., in prep). Ordination showed species to be ordered along a gradient from those found in strict crop production (e.g. skylark) to those found mostly in pasture and shrub dominated landscapes (e.g. yellowhammer), suggesting that not all species benefit from increased heterogeneity. Bird species richness was higher in more heterogeneous farmland, but within landscape types the amount of small biotopes did not explain species richness (7, Olsson et al., in prep.). However, the amount of field borders with trees and bushes positively predicted species richness, demonstrating the importance of hedgerow vegetation. Only one species, the skylark, significantly benefitted from lack of hedgerows. Bird species richness was positively related to organic farming, but only in homogeneous landscapes (5). This pattern particularly depended on invertebrate feeding passerines. Similarly the richness of migrants was higher on organic farms only in homogeneous landscapes (1). Migrant passerines and non-passerines were differently affected by landscape heterogeneity, because the latter seemed to benefit from landscape simplification. However, not only may the effect of AES depend on landscape context, but also on the targets for biodiversity conservation, i.e. whether local or regional biodiversity is to be maximized (7). House sparrow occurrence on 204 farms was higher when heterogeneity was higher at the landscape and farmstead scale (von Post et al., ms1). House sparrow density was positively affected by animal husbandry but, surprisingly, negatively related to winter-feeding. A behavioral indicator, giving-up density (GUD; a measure of the amount of food left in a foraging patch once the forager chose to feed elsewhere, reflecting the food available in the surroundings) at artificial food-patches showed habitat quality to be lower at farms in homogeneous landscapes (von Post et al., ms2). Studies on habitat selection, population dynamics and experimental winter-feeding are under evaluation. Data from 33 starling colonies in Sweden over more than two decades, demonstrated spatially heterogeneous decline rates (8). Local population changes were positively related to local changes in reproductive output, suggesting that the decline of Swedish starlings is related to processes at small spatial scales. Detailed demographic and foraging studies are in progress. We manipulated apparent habitat quality for the declining grey partridge to determine the feasibility of simple mitigation measures for farmland birds (3), but are still in the early days. Discussion and conclusion We show that ecological heterogeneity at multiple spatial scales (landscape, field, and farmstead) is important for species richness and population persistence of farmland birds. Hence, restoring farmland heterogeneity may be an efficient way to recreate biodiversity in contemporary agricultural landscapes. However, homogeneous landscapes contained a partially unique composition of species, such that we should avoid setting a single target for increased landscape heterogeneity. AES that target ecological heterogeneity may most efficiently increase bird diversity in intensively farmed landscapes, but it is not clear how efficient these schemes are at targeting regionally or nationally rare species.
Sweden