Carabid beetle assemblages at Silwood Park, UK

Anthropogenically created habitat edges have long been known to affect communities of adjacent habitats. This data is curated from a collection of studies where pitfall-trapping were used to compare changes in ground beetle (Coleoptera Carabidae) assemblages along transects marking the transition between different types of habitats: grassland to woodland, grassland to farmland and farmland to woodland.  

Studies were part of undergraduate or Master degree projects, thus the methods used between years are not the same. Students original data and reports are available.

The study site is the Silwood Park Campus from Imperial College London, Buckhurst Road, Ascot, Berkshire SL5 7PY, United Kingdom. The campus, with 75 ha, contains grasslands, scrubland, ancient woodlands and few decades old oak-dominated woodlands. Woodlands in Silwood Park used in these studies developed from natural succession from about 1960’s. They are the typical species-poor type found in lowland southern Britain, classified as (W10a) with common tree species like Pedunculated Oak (Quercus robur), Silver birch (Betula pendula) and with Bramble (Rubus fruticosus), Common Nettle (Urtica dioica) and Bracken (Pteridium aquilinum) dominating the field layer. Grasslands are species-poor on sandy, acidic soil of the Bagshot Series, MG5 Cynosurus cristatus/ Centaurea debeauxii, Danthonia decumbens subcommunity (United Kingdome National Vegetation Classification, https://jncc.gov.uk/). Farmland used are former arable plots abandoned since the 1970's and maintained by cutting vegetation, ploughing and rotovating the soil in October every year. 

Silwood Park experiences an average annual rainfall of 698mm with little seasonal pattern.  Mean hourly temperature is 10oC with July max of 23 oC and January min of 1.4 oC (Means from Silwood Weather Station,1987-2022, https://www.imperial.ac.uk/silwood-park/research/field-experiments/silwood-weather/).

Methods

Data 2016: Three grassland to woodland 64m transects were placed perpendicular to the habitats edge at each site; 32m extending into the woodland, 32m extending into the grassland and the midpoint of the transect at the edge. One pitfall trap was placed at each of the following distances each side of the edge: 2m, 4m, 8m, 16m, 32m, and another was placed at 0m. Overall carabids were collected over a total of 21 trap-days divided in seven sampling periods of three days each. Abundance of carabid species collected in each pitfall trap is reported for each three-day collection period. 

Data 2017: Two transects were set in each habitat transition: grassland to woodland and grassland to farmland. Three 5x5m2 plots were placed at 5m intervals, along a 25m transect, across each site. In each plot, five dry pitfall traps were set 1m apart in the diagonal line of the plot (Fig.1).  Traps were plastic cups of 7 cm diameter buried at ground level. Pitfall traps were set 14 times in each plot and examined after 24h. Carabid beetles found were identified, marked with nail polish using color codes and released in the same location. After 24h pitfall traps were examined again, and the number of captures and recaptures recorded. Data are number of carabid beetles of each species found in pitfall traps for the whole plot in a given 48 sampling period.

Data 2019: Three replicate transects were set up along two habitat edges: Agriculture field to woodland and acid grassland to woodland. Transects were at a minimum distance of 20m from any other edge. The transects were separated by the maximum possible distance of 14m. Each replicate transect was composed of five plots ranging from 20m within the woodland interior to 20m into the matrix (grassland/agricultural field), using the central traps as reference for measures of distance. Each plot contained nine pitfall traps arranged in 2x2m grids resulting in a total of 270 traps. Simple “Barber” pitfall traps were composed of 2 stacked 200ml plastic cups with an open diameter of 7.2cm and buried at ground level. Beetle collection occurred over a period of 24 days (April 14 to May 8, 2019). Traps were emptied at three-day intervals for six of the eight sampling events, with one sampling event occurring after two days, and another after four days due to organizational complications. Samples were grouped by plot rather than by individual trap, resulting in the collection of 30 samples per sampling event. Specimens were conserved in 40% ethanol solution and stored for later identification.