• rockfall;
  • rockwall retreat;
  • mountain geomorphology;
  • magnitude–frequency relations;
  • sediment budgets;
  • non-linear dynamics;
  • ergodic relations


Rates of rockwall retreat and rockfall supply are fundamental components of sediment budgets in steep environments. However, the standard procedure of referencing rockwall retreat rates using only lithology is inconsistent with research findings and results in a variability that exceeds three orders of magnitude. The concept proposed in this paper argues that the complexity inherent in rockfall studies can be reduced if the stages of (i) backweathering, (ii) filling and depletion of intermediate storage on the rock face and (iii) final rockfall supply onto the talus slopes are separated as these have different response functions and controlling factors. Backweathering responds to preweathering and weathering conditions whereas the filling and depletion of intermediate storage in the rock face is mainly a function of internal and external triggers. The noise apparent in backweathering rates and rockfall supply can be reduced by integrating the relevant controlling factors in the response functions. Simple conceptual models for the three stages are developed and are linked by a time-dependent ‘rockfall delivery rate’, which is defined as the difference between backweathering and rockfall supply, thus reflecting the specific importance of intermediate storage in the rock face. Existing studies can be characterized according to their ‘rockfall delivery ratio’, a concept similar to the ‘sediment delivery ratio’ used in fluvial geomorphology. Their outputs can be qualified as trigger-dependent rockfall supply rates or backweathering rates dependent on (pre-)weathering conditions. It is shown that the existing quantitative backweathering and rockfall supply models implicitly follow the proposed conceptual models and can be accommodated into the uniform model. Suggestions are made for how best to incorporate non-linearities, phase transitions, path dependencies and different timescales into rockfall response functions.