Distribution patterns of carabid beetle species at the landscape-level

Aim Carabid beetles are known to be affected by abiotic habitat conditions, both at a local microscale and in their distribution between habitat types. We explored whether the non-random distribution at micro- and macroscales is reflected in non-random distribution patterns at an intermediate mesoscale. The aim of this paper is: (1) to identify and to characterize carabid distribution patterns at the meso- or landscape scale; (2) to analyse if patterns can be confirmed between years; and (3) to propose environmental effects that explain the observed patterns. Location We studied a topographically varied area of c. 4 km(2) within a landscape of continuous beech forests. The study area is situated close to Gottingen, Central Germany, on a limestone plateau and its south- and southeast-facing slopes, at 280-420 m a.s.l. Methods Carabid beetles were sampled with pitfall traps during four consecutive years. Each year 10, 20 or 50 traps km(-2) were operated for 4-6 months. We interpolated continuous distribution maps from pitfall sites and correlated species maps between separate years. Spatial autocorrelation was analysed with Mantel tests and correlograms of Moran's I. Results 1. We analysed the spatial distribution of twelve carabid species. All interpolations produced more or less patchy map patterns. No pattern was exactly reproduced between years. 2. Carabid species distribution at the landscape level could be differentiated into four types: random, weak gradient, distinct gradient and restricted area. Maps for species with distinct gradients or restricted distributions were correlated between all years, while maps for species with random patterns were mostly uncorrelated. 3. Mantel tests detected overall spatial autocorrelation (SA) in most species. Correlograms of Moran's I supported the characterization of distribution types. The correlograms indicated patch sizes of 0.8-1.2 km extent for all species but three that were randomly distributed. 4. Abax parallelus (Duftschmid), Carabus coriaceus L., and Cychrus caraboides (L.) appeared to be randomly distributed within the study area. Also A. parallelepipedus (Piller & Mitterpacher) had a random areal distribution, but the patch structure was reproduced between years. 5. Abax ovalis (Duftschmid), Carabus auronitens Fabricius, C. nemoralis Muller, and Pterostichus burmeisteri Heer displayed rather weak distribution gradients that were reproduced at least between some years. 6. Pterostichus madidus (Fabricius) and P. melanarius (Illiger) displayed distinct distribution gradients. Their contrasting distribution could be explained by a topoclimatic gradient: P. madidus density was shifted to dry and warm parts of the study area, P. melanarius density to moist and cool parts. 7. Cychrus attenuatus Fabricius was restricted to a northern, high-elevated area, whereas Carabus problematicus Herbst was restricted to patches along the warm edge of the limestone plateau. Main conclusions Within the beech forest landscape, we identified reproducible mesoscale distribution patterns of carabid beetles. Distinct distribution gradients and restricted distribution areas could be attributed to a topoclimatic differentiation of the landscape. Thus, for some carabid species, microclimatic habitat associations scale up to distributions within a landscape.