To quantify the erosion parameters and thus assess the erodibility of soils, hole erosion tests (HET) are used with methods of interpretation developed recently by researchers. The application of these methods requires some input data, and the most critical one is the final diameter of the eroded hole that is determined after completion of the test. The precision of the hole diameter after erosion leads to a better and more reliable evaluation of the soil's resistance against erosion, more precisely the soil erodibility characteristics, tau c (critical shear stress) and ker (erosion coefficient). In the present study, three methods for determining the final hole diameter are used: M1, direct weighing of the paraffin skeleton of the eroded hole; M2, differential mass of the sample before and after erosion; and M3, 3-D scanning technique of the paraffin skeleton. These methods were applied on five HET tests performed on soil samples containing the same percentage of Armorican kaolinite mixed with Hostun sand. The soil samples are compacted at different densities and to the optimum water content, as determined from a standard Proctor test. The results indicate that method M3 based on a 3-D scanner acquisition and processing of the skeleton volume after erosion compares very well with the more usual ones, M1 and M2, with the obvious advantage of providing the full 3-D reconstruction of the post-erosion hole, allowing further investigations (morphological analyses). Furthermore, an assessment of the uncertainty in predicting erodibility, particularly associated with the post-erosion diameter calculated through the three methods, was conducted and yielded satisfactory outcomes.

It is important to quantify the soil resistance against erosion cause by to the overflow of dikes and levees. Small-scale tests are excluded, due to the lack of similarity for a free-surface flow phenomenon on stepped slopes, with non-established flow and erosion of a cohesive soil. Moreover, using hydraulic laboratory flumes does not make it possible to have a correct representation of the soil in place, in terms of layer compaction. This is why we have developed an on site overflowing device. The device was deployed as part of the DigueELITE research project on a 3.5 m (9.8 ft.) high experimental dike in channels, 60 cm wide (≈2 ft.) and 15 m long (≈50 ft.), covering the downstream slope (1.5H/1V) and the downstream platform. The procedure followed is based on ASTM-D6460 standard. The test campaigns were carried out with flow rates up to 500 l/s per linear meter (0.5 m2/s), water velocity up to 5 m/s, and a discharge depth up to 30 cm (≈12 in.) at crest. Two soil types were studied: lime-treated soil, and untreated cohesive soil. The first phase of erosion is that of the surface layer. The second phase is that of the embankment constitutive soil. The erosion shows a stair-steps pattern, due to the layers of compaction. The results obtained show that lime-treated soil has better erosion resistance than untreated soil. Compared to the untreated soil, erosion in the lower part of the slope is 3 times less in lime-treated soil, and the scour depth development process at the downstream toe is 5 to 10 times smaller. This paper presents the experimental setup, the results obtained, and the perspectives. The most important findings are that overflowing experiments are feasible on site with the proposed test set-up.