Recommendations for depletion modelling of granivorous birds
|
|  Yüklə 6.33 Mb.
|
6.3.2 Other invertebratesNon-arthropod invertebrate food of farmland birds include earthworms and molluscs (snails and slugs). These can also be sampled using the methods listed in Table 6.1, with the addition of soil sampling and extraction for sub-surface invertebrates and chemical extraction techniques for earthworms (Edwards & Bohlen, 1972). Tucker (1992) measured soil macro-invertebrate densities as an estimate of availability to surface picking or bill probing species of farmland birds. He took 15 soil cores per field, each core being 120mm diameter, 50mm deep, extracted with a sharpened steel cylinder. Cores were frozen and stored for up to 3 months. After thawing, invertebrates were extracted by washing and flotation and sorted into broad taxonomic groups and biomass estimated from wet weight. The invertebrate species measured using this method included Lumbricids, Molluscs, Coleopterans, Myriapods and Diptera larvae. In a study of habitat use of lapwings in rough grazing and arable areas in Scotland, Galbraith (1989) used soil cores and pitfall traps to sample surface-living and subsurface-living invertebrates. Each soil sample was 20 cm square and 10 cm deep, and was hand-sorted into earthworms, leatherjackets and other (mainly beetles and their larvae). Pitfall traps were plastic beakers (circular in cross-section, 8.5 cm deep and 7 cm in neck diameter) containing 5% formalin solution and buried up to their rims. Sheldon et al. (2002) used soil cores of 10 cm diameter and 10 cm depth to estimate earthworm food for foraging lapwings, in different crop types. Ten samples were taken from the central area of each field and hand-sorted. Village & Westwood (1994) measured earthworm abundance for foraging lapwings by formalin extraction at eight 50 cm x 50 cm quadrats per field on 22 occasions from October through January. In an investigation of stone curlew habitat selection, Green et al. (2000) measured the abundance of earthworms by applying a solution of formaldehyde to an area within a 0.25 m2 wire quadrat frame. For or five quadrats within a 30 m diameter area were treated at each site. They also measured invertebrates using pitfall traps (five traps in a line at intervals of 5 m at each site). The traps were 10 cm in diameter, 15 cm deep, rim flush with the soil surface, containing a preservative. Table 6.2 lists the most common methods used in the above studies on farmland birds. For each method, the taxa sampled are listed, and comments made on ease of use and processing time. Table 6.2 Methods used to sample the invertebrate food of farmland birds.
6.3.3 WeedsMethods of estimating plant cover and numbers are straight forward, and have been well reviewed in numerous textbooks, e.g. Greig-Smith (1983) and Moore and Chapman (1986). The main methods are summarised in Goldsmith (1991) and those useful for measuring abundance listed: density (plants often spread vegetatively making this measure less useful except for certain species, such as bulbs, orchids, annuals and trees), cover (proportion of ground covered by a species, often recorded with pins of very narrow diameter located at random, or estimated by eye and placed in a range, e.g. the Domin 1 to 10 range), biomass or yield (accurate but destructive), basal area (appropriate to trees or tussocky plants), frequency (the proportion of quadrats which contain a particular species, tends to combine abundance with distribution). It has recently been appreciated that the structure of vegetation may affect food availability to farmland birds (see Objective 1). Additional methods to measure vegetation structure include the following: drop disc, in which a disc of standard weight and diameter is dropped on the vegetation from a height of 1 m down a vertically held ruler, in order to provide an indication of leaf and stem density within the sward canopy, point quadrat, in which pins are slotted within a frame placed randomly within a plot, and the number of vegetation contacts made at each height interval recorded, enabling the 3-D structure of the sward to be determined, and graduated board, in which estimates of the proportion of the board obscured when viewed from 1 m are made at different heights, in order to build up a profile of vegetation density. Quadrat size and sampling pattern (random, systematic or stratified) should also be appropriate to the species and habitats being considered. The GCT estimates the overall general abundance of broad-leaved weeds and grass weeds by scoring them from zero (none present) to five (complete infestation of the crop by weeds). In the case of weed occurrence (presence/absence) most weeds are identified to species level (Ewald & Aebischer, 1999; Moreby et al., 1994; Moreby & Southway, 1999) Green (1984) took ten 0.1 m2 quadrat weed samples, 5 and 50 m from each field boundary. He gives no further details, but results are presented as ‘quadrat samples of Poa and Agrostis spikelets 1.0 m-2’. Watson & Rae (1997), in a study of corn buntings in north-east Scotland, scored weeds on a scale from 0 (none) to 5 (>75% of the ground covered). In surveys of cereal weeds undertaken from June - August, occurrence and abundance were estimated visually on a crude scale, while walking through fields and scanning from vantage points (Froud-Williams & Chancellor, 1982; Chancellor & Froud-Williams, 1984). |