Are there invariant and specific characteristic spatial scales of diversity patterns in grasslands?

The concept of characteristic scales (e.g. minimum area) is fundamental in community ecology. Scales where community state-variables such as diversity measures show maximum or minimum variability are optimal for detecting patterns in plant coexistence and plant-environmental relationships. While there is a consensus among field ecologists about the approximate magnitude of the related optimum scales, few systematic studies exist for assessing the variability/invariance of these characteristics. The spatial variance of selected vegetation state variables (diversity measures) was determined in contrasting grasslands. More than 60 vegetation stands representing four dry grassland types were studied with minimum of 10 replicates per types, applying a standardized sampling protocol. Meadows steppes (dry and semi-dry grasslands on loess), open sand steppes of humus-poor sand dunes, short grass salt steppes on solonetz soils and rock grasslands on dolomite and limestone hills were sampled at various locations in Hungary, Italy, Romania and Slovenia. A list of species present in 5 × 5 cm contiguous micro-quadrats was sampled along 52 m long transects of 1040 units arranged within uniform (homogenous) vegetation patches in each grassland stand. From each transect, diversity measures (variance of species richness, number and diversity of species combinations) were estimated at increasing plot sizes from cm × 5 cm to 5 cm × 10 cm … to 5 cm × 2500 cm. Spatial scales where vegetation characteristics reached maximum and minimum values were assessed in each transect. Spatial scale of maximum variability ranged between 5cm and 500 cm. Maximum scales of different diversity measures were close to each other. The smallest maximum scales were found in meadow steppes and maximum scales increased gradually from dolomite to sand grasslands with the largest values in salt steppes. However, estimated maximum scales varied from stand to stand within a grassland type. No significant differences were found in medians, however, the variability of estimates within a grassland type increased from meadow steppe to salt steppe. Within a certain type, medians were significantly larger in disturbed stands and they tended to decrease in recovering stands. Estimates of maximum scales were robust while estimates of minimum scales were highly variable without apparent trends. Minimum scales were strongly influenced by the effect of rare species. In contrast, maximum scales showed clear trends reflecting types of the spatial organization rather than some syntaxonomic features.