Modelling mitigation of bird population declines in the UK through landscape-scale environmental management.

Abstract

Biodiversity is declining on a global scale despite efforts to the contrary. Birds are effective indicators of ecosystem health, occurring in almost every habitat on Earth. However, many UK birds have declined since the 1960s, and are now classified as endangered or rare. Knowledge of factors influencing the presence and abundance of such species is therefore vital for their conservation. Habitat diversity affects avian diversity attesting that birds are a vital resource to conservationists. Not only are breeding birds influenced directly by their immediate habitat, they are also indirectly affected by the surrounding landscape, indicating the need for local and landscape-level studies and management. This study takes a multi-scale approach to examine the consequences of habitat and landscape changes on bird populations in two contrasting and mixed land-use sites: heathland and woodland in the New Forest (Hampshire) and arable farmland with scattered woodlands in Cambridgeshire. Recently acquired, high resolution airborne remote sensing datasets (Light Detection and Ranging, LiDAR) were used to develop metrics that quantified vegetation structure within the two study landscapes. These variables, together with vegetation composition (recorded from field surveys) were examined in relation to a series of bird indices (density, species richness, diversity, number of declining species, conservation priority, and rarity), as species richness and diversity alone can mask effects on more vulnerable species. Relationships with bird community composition and the habitat variables were also investigated using Multidimensional Scaling (MDS). Although bird communities are known to differ between broad habitat classifications, this has not explicitly been quantified. The results from these two studies were used to predict the effects of landscape change on the bird indices and to identify the bird species affected, with a view to providing management recommendations for the relevant authorities. The most diverse habitat in the New Forest for bird species was the scrubland (despite low bird density), represented by a positive relationship with scrubby vegetation variables, such as the percent cover of vegetation at 2-5 m and height Vertical Distribution Ratio (VDR). Beech woodlands supported the greatest number of declining bird species. Pine was overall poor bird habitat, signified by a negative relationship of the percent cover of pine with the majority of the bird indices. Other conifer was positively related to the Index of Relative Rarity (IRR), and supported the rare Firecrest. Heathland also had a high IRR value on account of the rare Dartford Warbler, supported by a positive relationship with the percent cover of heather, indicating that alternative habitats to those that increase diversity were extremely important to habitat specialists. The habitat associations of these bird species were confirmed by the MDS analysis. Furthermore, the MDS also showed that although poor in terms of the bird indices, pine provided habitat for other rare and declining birds, including Common Crossbills. Woodland edges in Cambridgeshire were the most diverse for bird species (a ‘classic’ edge effect), but which conversely resulted in adjacent fields being poor bird habitat. The MDS analysis showed that corvids were strongly associated with these edge habitats creating an exclusion zone. An increase in the proportional length of woody hedge vegetation in field boundaries supported more declining bird species than the other habitats. Hedges also increased (and were positively related with) the majority of the bird indices in the field-only analysis. Rarity and IRR were positively related to variables depicting woodland vegetation (percent cover of oak and vegetation height), suggesting that rare birds, such as Marsh Tits or Ravens, were in taller oak woodlands. Furthermore, a negative relationship of rarity with wood area suggests that the woodlands were sufficiently interconnected over the Cambridgeshire landscape to allow populations to persist. Overall, the MDS results showed that in both landscapes, bird community composition was more similar between the woodlands and most dissimilar between the non-woodland habitats. However, once separated, the woodlands were found to vary by vegetation composition (and habitat class) in the New Forest and by particular vegetation species and structure (scrubbier vs taller woodlands) in Cambridgeshire. Predictions of landscape change, such as scrub removal, in the New Forest, reduced bird density, and would also reduce bird diversity, and affect scrub preferring species such as Willow Warblers. Pine removal would increase many of the bird indices, but would affect conifer specialists, Common Crossbills and Wood Warblers. Beech decline locally was predicted to reduce the number of declining bird species supported, affecting the Hawfinch population. In Cambridgeshire, declines in hedge length would reduce the number of declining bird species supported (e.g. Yellowhammers), and most of the bird indices over this agricultural landscape. The spread of improved grass would reduce species richness and diversity, and increase corvid density. Declines in oak and tree height, through tree disease or felling, would reduce the number of rare species in the woodlands, including Marsh Tits. Contrasting habitat composition, structure and configuration of both the woodland and non-woodland habitats in these two landscapes, results in contrasting bird indices and community composition. Unsurprisingly, the New Forest was overall better for birds, however, Cambridgeshire supported bird species that were absent from the New Forest, such as the extremely rare and declining Turtle Dove. Bird species habitat preferences also differed between the landscapes, for example, the Goldfinch was associated with conifer in the New Forest, but with hedges in agricultural Cambridgeshire. These two landscape studies had the same conclusions; biodiversity should not be taken alone to measure habitat health as this often masks trends in rare and declining species, as represented by metrics detailing the number of declining bird species, species priority, rarity, IRR and community composition, being related to different habitat variables. This leads on to the second conclusion; that landscape heterogeneity is vital to maintain gamma diversity by providing habitat for as many species as possible. Thus, conservation should be targeted at a landscape scale and incorporate all bird measures, including conservation priority, rarity and community composition as well as diversity.