Are slump folds reliable indicators of downslope flow in recent mass transport deposits?

Despite the widespread use of slump folds as indicators of palaeoslope orientation, there is a lack of detailed analysis of variations in fold geometries and orientations down the length of individual slump profiles within mass transport deposits (MTDs). To address this gap in knowledge, we have systematically recorded more than 500 structural measurements of fold hinges and axial planes along a 25 m section through a mesoscopic slump profile. Our case study is performed in wet unconsolidated (late Holocene) sediments, which are only recently exposed due to falling water levels in the Dead Sea. In this situation, the modern slope is exposed and directly visible, slumping having occurred in the past few centuries. Fold hinges define broad arcs at high angles to flow in the downslope toe of the slump and progressively swing to become sub-parallel to flow in the upslope region. Greatest amounts of shortening (-35%) are recorded at the toe, suggesting that the swing in trends of fold hinges and axial planes is a consequence of differential layer-normal shear rather than downslope strain gradients. Significant variations of >90 degrees occur in the orientation and vergence of slump folds on either side of a 10 m wide gully, which cuts the slump sheet. In some instances, folds have nucleated around longer (>10 cm) wooden sticks that were incorporated into the slump, whereas shorter wooden fragments align parallel to the flow direction. The differences in orientations of wooden sticks and wooden fragments are consistent with differential layernormal shear on each side of a flow cell. Evaporite concretions grew within the sediments during slumping and influenced the geometry and kinematics of slump folds, suggesting that slope failure may have been a slow `creep' event generated by slope instability, rather than a result of catastrophic failure associated with large earthquakes. Our work illustrates the problems associated with using partial datasets, where classical structural analysis of transects <10 m apart would incorrectly suggest slump directions opposed to one another by 90 degrees. This study thereby highlights the extreme variability within a downslope profile of a single slump. It may therefore help explain discrepancies in regional datasets where slumps, sporadically sampled at different stratigraphic levels, may provide apparently diverse flow directions.