Effect of long‐term drainage on plant community, soil carbon and nitrogen contents and stable isotopic (δ13C, δ15N) composition of a permanent grassland

This study compares data statistically that were collected from both long‐term drained and undrained plots to test hypotheses concerning the effect of drainage on plant community, soil total nitrogen (TN), soil total carbon (TC) and stable isotopic (δ15N, δ13C) contents in a permanent grassland. In addition, the effects of soil depth, topography (elevation, slope, aspect and compound topographic index (CTI)) and spatial autocorrelation were taken into account. Data were collected in 2010 at Rowden Moor, North Wyke, Devon, UK, where, for the plots of this study, subsurface drainage was introduced in 1987. The results of a set of six linear mixed models showed that: (i) plant community did not depend on drainage, but on elevation and spatial effects, (ii) both TN and TC not only depended on drainage, but also topography and sample depth, (iii) the TC to TN ratio did not depend on drainage, but on elevation, CTI and sample depth only, (iv) δ15N values did not depend on drainage, but on topography and sample depth and (v) δ13C values depended on drainage together with topography and sample depth. Thus, drainage represented a significant effect for only TN, TC and δ13C. Furthermore, changes in soil physicochemical conditions, following the introduction of drainage in the clay soil 24 years previously, induced a shift in the plant community from a Lolium perenne L. dominated grassland with numerous patches of Juncus species, towards one with Lolium perenne and Trifolium repens L.