A spatial approach to soil-ecological experimentation at landscape scale

Weinheim

Wiley VCH Verlag

2008

S. 338-343

7 Abb., 9 Lit. Ang.

Artikel aus einer ZeitschriftUnselbständiges Werk

200148487

Journal of Plant Nutrition and Soil Science 2008, 171(3), S. 338-343

The upscaling of soil-ecological processes to larger landscape units represents a special challenge to soil ecology. Results from micro- or mesocosms cannot easily be transferred to other scales because effects are often scale-dependent. In this context, field experiments which take into account the heterogeneity of the landscape may be promising. Therefore, we carried out an experiment based on a transect study in the agrolandscape of NE Germany on heterogeneous sandy soil in which the feeding activity of the soil-organism community was assessed by means of the bait-lamina test at each of the 101 transect locations. At every 4th position, prior to the measurement the soil biota were stimulated by a treatment consisting of adding easily available C and water to the soil. Our aim was to test whether this kind of spatial approach enables to separate effects induced by treatments from landscape effects. The results showed a highly variable feeding activity along the transect after 4 weeks. Despite this variability, a basic trend could be identified which was related to a landscape factor, i.e., the relief. On upper-slope positions, the feeding activity tended to be less in comparison to positions down-slope. At every 4th position of the transect, the stimulating effect of the substrate and water addition could be clearly detected and quantified with spectral and cross-spectral analysis. It is concluded that effects of treatments in heterogeneous landscapes may be distinguished from site effects when the signal-to-noise ratio is high and soil and treatment effects on the variable of interest are sufficiently different from one another. In a heterogeneous landscape with gradients of site properties, a treatment based on the frequency domain and applied in regular intervals can be distinguished with spectral analysis techniques.