Log Crib Check Dam Performance under Multiple Debris-Flow Loadings – East Gate Landslide, British Columbia, Canada

. Check dams are used in gullies to prevent vertical downcutting of the thalweg. Check dams built as log crib structures are common in steep creeks prone to floods or debris floods. Recent experience with performance of log crib check dams subject to debris flow loadings is less common, as those check dams are often made from reinforced concrete, steel or masonry. This paper summarizes our experience with six log crib check dams built in a gully at East Gate Landslide, between Revelstoke and Golden, BC, Canada. The log crib check dams are subjected to multiple debris flow events every year. The project started as a pilot study and the design is adjusted before the next construction season, based on performance experience with previously built structures.


What is a Log Crib Check Dam?
The term Log Crib Check Dam applies to a log crib structure arranged transverse to the creek with "Konsolidierung" (German: consolidation) functionality. In the context of torrent control, Konsolidierung means preventing thalweg downcutting and supporting adjacent, unstable slopes with the sediment retained upstream of the barrier. Log crib check dams (Fig. 1) are three-dimensional lattice structures, where transverse and longitudinal logs are connected by steel nails in current practice (notching was traditionally used). The crib is filled with compacted soil providing stiffness and weight to the structure. Geotechnically, it functions like a gravity retaining wall.
Double-or multi-walled log crib systems are called "Krainerwände (crib walls)" in Austria, while in Switzerland they are typically referred to as "Holzkasten" (wooden box).

Mitigation Approach
The opportunity to study log crib check dams under debris flow loading arose at the East Gate Landslide (EGL) within the Glacier National Park, located 3.5 hours west of Calgary along the Trans-Canada Highway (TCH). The EGL is a 2,000,000 m 3 rock slump that occurred in 1997, travelled about 1,000 m, but did not reach the TCH at the valley bottom. Downstream of the feeder gully confluence, the Main Gully developed. This led to a self-amplifying cycle of downcutting and sideslope landslides (Fig. 2). Since 1999, several debris flows reached the TCH annually. About three quarters of the sediment is mobilized within the Main Gully located downstream of the EGL deposit.
The TCH is administered by Parks Canada Agency (PCA). Debris is retained immediately east of the TCH by two moustache-shaped retention basins (MRB) (see Fig. 2). Annual clean-out volumes are typically about 40'000 m 3 but may range up to 80'000 m 3 (i.e. the storage capacity of the MRB).
To reduce the sediment removal efforts and cost, A three-component gully erosion mitigation strategy was developed, consisting of; 1) diversions to reduce surface  flows in the Main Gully, 2) log crib check dams to control gully grade, limit undermining of gully side slopes caused by downcutting, and 3) revegetation of side slopes and eventually part of the slumped landslide mass to improve slope stability and reduce surface erosion.
Since the project is located in Glacier National Park, PCA wanted to use locally (and nearby) sourced timber for check dam construction rather than concrete or other materials. Since experience with log crib check dams in large debris flow prone gullies is very limited in Canada, PCA choose an adaptive implementation approach, which began with a pilot construction in 2020. Per end of the 2021 construction season 6 check dams were built (Fig. 1). As part of the multi-year erosion mitigation program in the Main Gully we anticipate construction of roughly 60 log crib check dams and several diversions, as well as re-vegetating a significant portion of the currently 60,000 m 2 of disturbed gully side slopes.

Project Settings
The flows in the Main Gully incise into colluvium deposited by pre-historic landslides. The over steepened side slopes creep, flow or slide into the Main Gully, supplying sediment for transport downstream.
The Main Gully starts where several feeder gullies coalesce ( The catchment area reporting to the MRB inlet is 0.68 km 2 , with a catchment relief of 1,269 m.
Detailed change detection using sequential LiDAR and photogrammetry data between 2015 and 2019, showed that the creek alignment in the Main Gully shifted by as much as 17 m horizontally and eroded 14 m vertically.

Performance Experience
The experience described below focuses on the performance of the check dams under debris-flow-and gully side slope loadings. For a more general description of other lessons learned refer to [1].

Pilot Study in 2020
A pilot check dam was constructed in 2020 to test several design parameters and performance under site conditions. Fig. 3 shows the 2020 Pilot Check Dam after construction and late in the 2021 freshet.

Performance Fall 2020 to Freshet 2021
Overall, the check dam performed well under the debris flow loading. The 2020 Pilot Check Dam experienced its first debris flow event (5'700 m 3 debris) on November 5, 2020 without damage; some minor deposition occurred on the right (in flow direction) wing wall. However, a ground sill made from logs across the channel and located several metres downstream to hold riprap in place was destroyed. This was likely due to insufficient lateral embedment. During the 2021 freshet, the check dam survived several debris flow events (individual volumes not surveyed). By end of May 2021, a total sediment volume of about 29,000 m 3 (surveyed) was retained in the MRB. Video camera recordings documented a range of debris flow hydrographs at the 2020 Check Dam. Some were fairly "smooth" surges, while other surges exhibited highly variable discharge (or velocity) even within a single surge. This is possibly due to large boulders causing temporary blockages until momentum pushed them forward again.
The following performance and key issues on the pilot check dam were noted: 1. Global Stability: The check dam performed satisfactorily and survived the debris flow loadings. No significant global displacements were noted. 2. Toe Scour: Fig. 3 shows significant toe scour occurred. In addition, the foundation soil was washed out under the front mud sill (i.e., lower most stretcher). This poses a stability concern. Toe scour may have been accelerated by loss of the ground sill during the November 5, 2020 event.
However, the scour hole may have formed regardless. 3. Wing wall overtopping: The pilot weir was intentionally sized relatively small because the narrow, V-shaped gully cross-section limited the weir dimensions that could practically be built. Mud deposits, a leaning log, and displaced roll of flexible pipe are clear evidence of wing wall overtopping by debris flows, indicating that the pilot weir is undersized. Based on review of video recordings, wing wall overtopping is interpreted to be of short duration (a few seconds) and did not result in abutment scour. 4. Abrasion: Significant abrasion and wear is evident on the logs (Fig. 3) show wear and abrasion but remained intact. c. Debris has abraded concave shapes into the stretchers and worn the front of exposed headers on the battered, downstream dam face. 5. Other Wear: A piece of wood was split of the lowermost front stretcher of the left wing wall (Fig.  3), presumably from a point load (boulder or log). Grade stabilization: The channel base above the weir has infilled with sediment as intended at about 3.5H:1V or 16⁰. 6. Wing wall protection: Backfilling the wing walls appears to have been a successful strategy to avoid direct, high dynamic loads on the back of the wing walls. The scour hole underneath the Pilot Check Dam was filled with rocks and the voids were filled with wet mix grout (200 mm slump number) supplied by a helicopter drop bucket. Toe scour protection was repaired by neatly placing large rip rap and constructing a ground sill made from laterally well embedded logs. For check dams #2 to #5 a "splashpad" was also made from grouted riprap.

Design Modifications in Summer 2021
• Steel plate armouring made from 12 mm thick sheets of 1.2 m by 2.4 m, was added on top of the corduroy logs of the weir base and parts of the weir sides (Fig. 6A) to protect against excessive wear from the abrasive debris flows. The weir crest was cantilevered far enough in downstream direction to project the flows beyond the dam face.

Performance Fall 2021 and Winter 2022
The check dams fared reasonably well through the fall 2021 and winter 2022, however in early spring landslides induced large loads on the check dams (Fig.  4). As anticipated, the uppermost check dam (#6 without wing walls) was buried by a 13,000 m 2 landslide entering from the left side slope (Fig. 5). By April 26, 2022 the thalweg was pushed to the right of check dam #6, because the right bank had melted earlier (southern exposure) than the centre and dense ice directed the thalweg to the right. To avoid check dam outflanking, the ice was removed and timber deflection logs were placed obliquely to the creek flow direction. Subsequently, the thalweg shifted back on the centre of check dam #6, although by May 11, 2022 #6 was still not visible under the debris.
Prior to the onset of the debris flow "season" in Spring 2022, another large landslide entered the Main Gully obliquely from the right bank (Fig. 6A). This landslide moved at least 10 m (visual estimate) and induced significant earth pressures on check dams #3, #4, and #5. Fig. 6B illustrates how the right wing walls were sheared off the lower portion of the check dam and rotated forward (in-plan view) by the landslide.
No significant changes were noted for check dams #1 and #2.

Performance Freshet 2022
The majority of check dam damage in the spring of 2022 was not attributable to debris flows but rather the landslide issues described above. The landslide from the right continued shearing off the right wing walls at #3, #4, #5. The large landslide entering further upstream from the left also sheared check dam #5 in the lower left side (Fig. 8).
During the 2022 freshet, debris flows deposited approximately 6'000 m 3 of debris (visual estimate) in the Moustache Retention Basin. This is significantly less than the typical annual clean-out volume of about 40'000 m 3 . Debris flow-related damage included: • Some damage to the weir steel armouring (deformations in general and a torn off sheet at #4). • Ryan Calder of PCA observed a damaging event during his May 27 th field inspection; knocking the top of the left wing-wall off at check dam #4 (compare Fig. 4 (before) and Fig. 7 (after)).

Conclusions
The key issues during the 2021 freshet were toe scour and abrasion from the very angular sediment particles carried by the debris flows. Extensive abrasion and wear on timber in direct contact with flowing debrisparticularly near the weir crest and on the 5V:1H battered downstream dam face -threatened to shorten the lifespan of the structures. This illustrates a limitation of log crib check dams in very active ‡ gullies. These issues were remedied by installation of steel plates over the weir openings and grouted riprap in the scour pools below the weirs. Key issues in 2022 were loads induced by landslides, entering from the side slopes shearing off wing walls, and hence causing deformations in the check dams. This illustrates the limited capacity to carry earth pressures from side slope movement of log crib structures. Although log crib check dams may have lower load carrying capacity compared to concrete check dams, experience at East Gate Landslide demonstrates a remarkably high tolerance for deformations before the deformed check dam loses its grade stabilization functionality.
Despite the significant, landslide induced damage, check dams #1 to #5 maintained their functionality: • The damaged wing walls at check dams #3 to #5 provided sufficient weir functionality to guide flows. However, partial blockage of the weir crosssectional area (Fig. 7) with debris was noted. This could be due energy losses at the check dam drops or simply more viscous rheology of a debris flow event occurring at the end of the freshet. • No significant toe scour issues occurred in the 2022 freshet, most likely due to the splash pad ‡ Note: This gully is very active and experiences multiple debris flow events per year with several 10'000 m 3 of sediment retained annually in the downstream retention basin.
design improvements with grouted riprap. The relatively small debris volume mobilized during the 2022 freshet may also be a factor. A further contributing factor could be that the debris flows reached their sediment carrying capacity and could not entrain more sediment.
• The steel plates provided good abrasion protection. To ascertain continued gully stabilization, it is now pertinent to continue check dam construction and interrupt the cycle of downcutting and side slope landsliding until stable side slopes have established.
The authors would like to acknowledge the support of Parks Canada Agency for this ongoing project.
The BGC Engineering Inc. team -in particular Alex Strouth -provided constructive feedback and inputs throughout this project.
Seppi Berwert-Lopes of Belop GmbH in Sarnen, Switzerland kindly offered his wealth of experience with log crib check dams.
McElhanney is the overall civil designer and their team around Shane Anderson, Patrick Zerr and Nick Guaran greatly contributed to the success of this project.
Tetra Tech EBA under the guidance of Charles Hunt was instrumental in the early gully assessments while Nigel Skermer encouraged the use of check dams.
The skilled crew at SpiderExcavators from Golden, BC built the structures in extremely challenging terrain.
This manuscript was prepared on the authors own time with support from BGC Engineering Inc. and Geoinfra Ingenieure AG.