Stronger legacy effects of cropland than of meadows or pastures on soil conditions and plant communities in French mountain forests

QuestionDifferences in understorey vegetation between ancient and recent forests have been thoroughly explored; however, few studies have investigated the legacies of different former land uses in recent forests. Indeed, due to more intense agricultural practices (tillage and fertilisation), legacy effects are expected to be stronger in former cropland or meadows than in former pastures. Our objectives were to compare soil conditions, taxonomic composition and functional composition of understorey plant communities in recent forests located on former pastures, meadows or cropland, with ancient forests as a reference. LocationTarentaise Valley, Savoy, France. MethodsBased on land-use maps surveyed between 1862 and 1864, we selected 82 forest sites with different former land uses in mountain forests in the French Alps and carried out soil sampling and botanical surveys. To account for potential confounding factors (altitude, canopy cover, tree species composition), we applied multiple linear regressions to analyse soil properties, canonical correspondence analysis to analyse plant taxonomic composition and multispecies generalised linear mixed-effects models to analyse relationships between plant functional composition and former land uses. ResultsThe soils of former cropland were richer in nutrients and more alkaline compared to other past land uses, while soils on former pastures and meadows differed only slightly from ancient forests. Ancient forests were characterised by acidophilic, shade-tolerant, low-stature, forest-dependent species, whereas former cropland was characterised by calcicolous non-forest species. Former pasture and meadow communities displayed a distinct taxonomic composition compared to other past land uses, but a functional composition closer to ancient forest than to former cropland. ConclusionFormer cropland has a stronger legacy effect than former pastures or meadows. This could explain small differences between ancient and recent forests observed in previous studies conducted in mountain landscapes where former cropland was rare.