- Top of page
For species that rely on visual cues to detect prey items, increasing the structural complexity of a patch can greatly influence forager behaviour through consequent reductions in prey detectability and accessibility. These effects are likely to manifest themselves in terms of foraging site selection and there is plentiful evidence for preferential site selection for a suite of taxa. However, the underlying effects of habitat structure on foraging behaviour, which are likely to drive these observed site selections, are much less well understood. We present the results of two studies designed to quantify the effects of vegetation structure on prey detectability and accessibility to avian invertebrate feeders and granivores on farmland. There was a significant negative relationship between potential prey detectability and both distance and vegetation height in cereal crops and stubbles for Northern Lapwings Vanellus vanellus. The interscan distance travelled by Lapwings differed significantly between habitats, with longer distances travelled in cereal crops and harrowed compared with ploughed soil and grasses. The peck rate, head-up rate and mean search period of foraging Chaffinches Fringilla coelebs were not affected by increasing vegetation structure but forager mobility was significantly reduced. We hope that by quantifying the effects of vegetation structure on prey detectability and accessibility we can highlight the importance of considering these factors, as well as prey abundance, when developing management strategies for farmland birds.
A thorough knowledge of the relationships between animals and their food resources is fundamental to our understanding of population dynamics and foraging patch selection (Lack 1954, Watson 1970, Hassell & May 1985). Prey abundance, detectability and accessibility can all influence functional and aggregative responses and depletion rates (e.g. Zwarts & Wanink 1993); and incorporating their effects is likely to significantly improve the accuracy of models predicting foraging behaviour and distribution. However, although food abundance is relatively straightforward to measure, few studies have directly examined prey detection or the effect of habitat variation on prey detectability and accessibility.
For species that rely on visual cues to detect prey items, increasing the structural complexity of a patch can greatly influence foraging behaviour through consequent reductions in prey detectability and accessibility (e.g. Eriksson 1985, Zwarts & Wanink 1993, Nystrand & Gronström 1997, Whittingham & Markland 2002), reduced forager mobility (Brodman et al. 1997) and increased energetic demands, for example via chilling effects when cold or wet (Dawson et al. 1992). The influence of these factors is most clearly demonstrated by foraging site selection. Preferential selection of relatively sparsely vegetated foraging sites has been demonstrated for a wide taxonomic range of ground-feeding birds (Whittingham & Evans 2004). Much of this work has been based on arable (e.g. Diaz & Telleria 1994, Moorcroft et al. 2002, Morris et al. 2002) and grassland (e.g. Milsom et al. 1998, Perkins et al. 2000, Johnstone et al. 2002) systems. However, although there is plentiful evidence that habitat structure influences foraging site selection, the underlying effects of habitat structure on foraging behaviour, which are likely to drive these observed site selections, are much less clear.
Here we present the results of two studies designed to quantify the influence of changing vegetation structure on the behaviour of birds foraging within agricultural systems. The first focuses on invertebrate feeders, using Northern Lapwing Vanellus vanellus L. as the model species, and the second on granivores, using Chaffinch Fringilla coelebs L. as the model species. Both these species rely predominantly on visual cues to detect prey items and move across the substrate to access them. Their foraging behaviour is therefore likely to be strongly influenced by habitat structure.
- Top of page
These studies have shown that changes in vegetation structure can directly influence foraging behaviour through its impact on prey detectability and accessibility and forager mobility. As vegetation height increases, foraging efficiency is likely to decline, probably because prey-locating cues become less detectable and forager mobility is impeded.
The laser pen method for measuring visibility through different habitats shows promising results for applications to a range of studies (Gillings in press). It demonstrates that the detectability of prey-locating cues is strongly affected by both distance and vegetation height and that these two factors interact, with distance effects being greater in taller vegetation. It has also enabled us to quantify the temporal decline in visibility caused by crop growth. This relationship between vegetation height and cue detectability probably accounts for the preference by Lapwings for short turf (Milsom et al. 1998) and the abandonment of cereal fields with tall vegetation (Gregory 1987).
The distance an individual moves between scanning points is a measure of how far they are capable of seeing (Metcalfe 1985). Although distance to prey captures and interscan distances did differ significantly between habitats, it was not possible directly to relate foraging behaviour with visibility profiles owing to the highly mobile nature of the plover flock. Besides logistical problems, a number of other factors may have disturbed the evidence for any relationship between visibility and foraging behaviour. For instance, it was not possible to incorporate the effects of prey abundance into these models. Prey abundance was difficult to measure but was generally low and spatially variable (Gillings 2003). Variation in prey abundance is likely to have had a significant effect on pacing behaviour (Anderson et al. 1997) and may have overridden any detectability effects.
The Chaffinch experiment offers the opportunity to investigate the effects of vegetation structure on foraging behaviour while controlling for variation in food abundance. We found that increasing vegetation structure had no effect on foraging behaviour apart from on forager mobility; foragers crossed fewer rows on the long : long substrate, where vegetation structure was more complex. A similar reduction in mobility in tall cover has been demonstrated for a number of other species (e.g. Brownsmith 1977, Brodman et al. 1997, Devereux et al. 2004).
The lack of any effect of increasing vegetation structure on peck rate, head-up rate and mean search period may be due, in part at least, to the density of available food. The density used represented the upper end of natural densities found on cereal stubble fields (Moorcroft et al. 2002). The difference in vegetation structure between short : long and long : long would not have affected mobility along rows, only across rows, but with high food abundance individuals are less likely to be forced to cross rows to maintain intake rates. This was highlighted by the fairly one-dimensional foraging pattern of individuals during long : long trials, moving up and down a single row. However, forager mobility appeared less restricted on the short : long substrate. Individuals exhibiting a more two-dimensional foraging pattern, crossing rows more freely. It is therefore likely that at lower seed densities, with individuals required to move across rows to maintain intake rates, any effect of vegetation structure on foraging behaviour would become more apparent. We predict that these effects would manifest themselves in terms of an increase in search period and a decrease in peck rate and therefore foraging efficiency. If increasing vegetation structure reduces foraging efficiency and increases mobility costs, it follows that the giving up densities should be higher in patches with greater vegetation structure (Brown et al. 1997). Foraging theory suggests that individuals will leave a patch when the harvest rate no longer exceeds the metabolic, predation risk and missed opportunity costs of foraging there (Brown 1988).
The observed relationships between vegetation structure, in particular vegetation height, and both prey detectability and accessibility are likely to hold for most species that rely on visual cues to detect prey and move across the ground to access them. These relationships could explain some of the observed preferential selection of relatively sparsely vegetated foraging sites by ground-feeding birds (Whittingham & Evans 2004). For instance, Moorcroft et al. (2002) demonstrated that Linnet Carduelis cannabina, Yellowhammer Emberiza citrinella, Reed Bunting Emberiza schoeniclus and Corn Bunting Miliaria calandra all preferred stubble fields with larger areas of bare ground and proposed that these relationships could be influenced by factors including physiological costs, prey detectability and/or predator detection ability. Indeed, Whittingham and Evans (2004) suggest that shorter vegetation is likely to enhance foraging efficiency for at least 11 of the 20 species that make up the UK Government's ‘Farmland Bird Index’, designed to measure the wildlife value of agricultural systems (Gregory et al. 2004). Of these, five (Linnet, Corn Bunting, Reed Bunting, Tree Sparrow Passer montanus and Skylark Alauda arvensis) are formally common and widespread farmland bird species for which UK Biodiversity Action Plans (BAP) have been developed. However, there is little reference to vegetation structure and its impact on prey detectability, prey accessibility and forager mobility in any of these BAPs.
Although shorter vegetation may offer increased foraging efficiency for a number of ground-feeding birds, it should be noted that it is unlikely to be beneficial to all. Depending on a species’ predator avoidance strategy, within-field cover can be either obstructive or protective (Lima 1990). Those species that fly to cover to avoid predation, such as Yellowhammers, Reed Buntings and Chaffinches, are likely to benefit from the reduced visual obstruction in short vegetation. However, for those species that rely on crypsis to avoid predation, such as Grey Partridge Perdix perdix, any benefits arising from increased foraging efficiency in short stubble are likely to be outweighed by the costs of increased predation risk (Whittingham & Evans 2004). It therefore appears that mosaics of short and long vegetation are likely to hold the maximum benefit for many farmland birds (Benton et al. 2003). This is also likely to be the case in grassland systems where, although the issues of prey detectability and forager mobility are likely to be of similar importance (e.g. Whitehead et al. 1995, Devereux et al. 2004), sward structure can have profound effects on invertebrate communities and therefore on food abundance for ground-feeding insectivores (McCracken et al. 1995, Benton et al. 2003).
We hope that by quantifying the potential effects of vegetation structure on prey detectability and accessibility we have highlighted the importance of considering these factors, as well as prey abundance, when developing management strategies designed to improve the habitat for these species. Although the techniques available for creating a suitable mosaic of vegetation structure are numerous, e.g. topping, flailing, mowing and scarification (Wakeham-Dawson & Aebischer 1998), further research is necessary to determine the most effective technique and the most appropriate scale of any such management.