High resolution, quantitative reconstruction of erosion rates based on anatomical changes in exposed roots at Draix, Alpes de Haute-Provence — critical review of existing approaches and independent quality control of results

Extensive areas of the French Alps are underlain by Jurassic black marls. Wherever these “terres noires” crop out they become subject to intense erosion, causing major sedimentation in regional reservoirs and river systems. In the badlands near Draix (Alpes de Haute-Provence, France), measured sediment rates were obtained at the plot scale by surface elevation change-based methods and at the catchment scale by monitoring sedimentation in dams. In this study, we use a dendrogeomorphic approach based on anatomical changes in exposed roots of Pinus sylvestris L. to accurately quantify continuous denudation rates. A total of 123 cross sections (75 from buried and 48 from exposed roots of 23 trees) were sampled in the Moulin basin. The size and position of roots at the time of exposure was determined via anatomical variations in the annual growth rings of roots. In cross sections of buried roots, a sharp reduction of earlywood tracheid lumen area — a growth signature which has traditionally been used to determine the moment of root denudation — was observed as soon as erosion reduced soil cover to ≤ 3 cm. As a consequence, estimates of eroded soil thickness had to be adjusted to take account of this bias. Bias-adjusted, averaged, medium-term erosion rates derived from exposed roots vary between 6 and 7 mm y–1 at Moulin basin depending on the importance accorded to the uplift of roots after exposure. Values are significantly correlated to slope angle and match with erosion rates derived from monitored iron stakes (5.7 mm y–1) or measurements of sediment yield in retention dams (4.7 mm y–1) at the outlet of the Moulin basin. Besides demonstrating that the interpretation of anatomical signatures in tree roots to erosion have to be revised, this paper also shows that dendrogeomorphic analyses of roots are indeed a powerful tool for the quantification of minimal rates of soil erosion in environments where measurements of past activity are not available.