Abundances of macrobenthic species were monitored twice yearly (March and September) at 6 locations in Tees Bay, UK, between 1973 and, and once yearly at 4 stations in the outer Tees estuary and 7 stations in the inner estuary between 1980 and 1999. In the Bay, multivariate analysis revealed a serial pattern of community change over years for all areas, but with a major shift in community composition between 1986 and 1988. Inter-annual variability in community composition was significantly greater after 1987 than before 1987 in all areas. Overall, inter-annual variability was greater in areas near the estuary mouth than in areas farther away, although the direction of community change and the timing of the discontinuity were the same in all areas. The serial nature of community change with time was also weaker in the areas close to the estuary mouth. Although there was no clear pattern of change in the number of species present over the sampling period, a dramatic increase in Shannon diversity (H ') occurred after 1987, due to an increase in evenness that resulted from the reduction of a few previously dominant species, notably the small polychaete Spiophanes bombyx. Although biodiversity measures describing the taxonomic breadth of the species assemblages also showed a marked step change in 1987, this was one of reduced diversity, with average taxonomic distinctness (∆ + ) decreasing and the variation in taxonomic distinctness (Λ + ) increasing. These abrupt, detrimental changes coincided with a well-documented change in a variety of components of the North Sea ecosystem during the same period. Traditional species diversity measures, such as H ', therefore gave a false impression of improving environmental quality over this period: given that the average taxonomic spread was reduced, certain taxa were under-represented with respect to others, and community composition as measured by a multivariate stability index (MSI) became less stable. H ' also failed to distinguish putatively impacted areas close to the estuary mouth compared with those more distant, despite clear differences in ∆ + , Λ + , and in community stability (MSI). Overall patterns of biodiversity and community composition in the Bay have thus been affected temporally by regional changes in the North Sea ecosystem, and spatially by the effects of the estuarine outflow. In the estuary itself, multivariate analysis also revealed a serial pattern of community change, with a major shift in composition in 1994 in both the outer and inner estuary which coincided with the construction of a barrage in the estuary. The numbers of both individuals and species began to increase at this time in the outer estuary. H ' showed no obvious changes over the period, but in the outer estuary a step change in ∆ + and Λ + occurred at the same time as that in the Bay. However, the direction of change was the reverse of that in the Bay, suggesting an improvement in environmental quality or a shift to more saline conditions.

The relative effects of tree clearing, increased livestock densities and nutrient enrichment have rarely been compared across markedly different organism types, but negative effects are generally predicted. In contrast, adoption of rotational grazing is thought to benefit biodiversity in pastures but there are few supporting data. We examined the response of native plants, birds and reptiles to livestock management in south-eastern Australia. We selected 12 pairs of rotationally and continuously grazed farms. Two 1-ha plots were established in native pastures on each farm, one cleared and the second still retaining woodland tree cover. Stocking rates, fertilizer histories and landscape tree cover varied among farms.The abundance and richness of all taxa was lower in cleared pastures.The less mobile organisms (reptiles and plants) were positively correlated with tree cover at landscape scales, but only when trees were present at the plot scale. This pattern was driven by a few observations in landscapes with approximately 50% tree canopy cover. Neither bird abundance nor richness was correlated with stocking rates or nutrient enrichment, but plant richness responded negatively to both.The response of reptiles varied, declining with nutrient enrichment but positively correlated with livestock densities. These responses may be partly interpreted within the context of prior filtering of species pools through long-term grazing pressure. No taxa responded positively to rotational grazing management. We predict that reductions in livestock density and soil nutrients will directly benefit plants and less so reptiles, but not birds. Indirect benefits are predicted for birds and reptiles if management increases persistence of trees within paddocks. Although some forms of rotational grazing can increase woodland tree recruitment, rotational grazing in itself is unlikely to enhance diversity.

A key part of tropical forest spatial complexity is the vertical stratification of biodiversity, with widely differing communities found in higher rainforest strata compared to terrestrial levels. Despite this, our understanding of how human disturbance may differentially affect biodiversity across vertical strata of tropical forests has been slow to develop. For the first time, how the patterns of current biodiversity vary between three vertical strata within a single forest, subject to three different types of historic anthropogenic disturbance, was directly assessed. In total, 229 species of butterfly were detected, with a total of 5219 individual records. Butterfly species richness, species diversity, abundance and community evenness differed markedly between vertical strata. We show for the first time, for any group of rainforest biodiversity, that different vertical strata within the same rainforest, responded differently in areas with different historic human disturbance. Differences were most notable within the canopy. Regenerating forest following complete clearance had 47% lower canopy species richness than regenerating forest that was once selectively logged, while the reduction in the mid-storey was 33% and at ground level, 30%. These results also show for the first time that even long term regeneration (over the course of 30 years) may be insufficient to erase differences in biodiversity linked to different types of human disturbance. We argue, along with other studies, that ignoring the potential for more pronounced effects of disturbance on canopy fauna, could lead to the underestimation of the effects of habitat disturbance on biodiversity, and thus the overestimation of the conservation value of regenerating forests more generally.

No abstract