Multi-process Late Quaternary landscape evolution modelling reveals lags in climate response over small spatial scales

摘要

Landscapes evolve in complex, non-linear ways over Quaternary timespans. integrated geomorphological field studies usually yield plausible hypotheses about timing and impact of process activity. Landscape Evolution Models (LEMs) have the potential to test and falsify these landscape evolution hypotheses. [Despite this potential, LEMs have mainly been used with hypothetical data and rarely to simulate the evolution of an actual landscape. In this paper, we use a LEM (LAPSUS: LandscApe ProcesS modelling at mUlti dimensions and scaleS) to explore if it is possible to test and falsify conclusions of an earlier field study on 50 ka landscape evolution in Okhombe Valley, KwaZulu Natal, South Africa. In this LEM, five landscape processes interact without supervision: water driven erosion and deposition, creep, solifluction, biological weathering and frost weathering. Calibration matched model results to three types of qualitative fieldwork observations: individual process activity over time, relative process activity over time and net landscape changes over time. Results demonstrate that landscape evolution of the Okhombe valley can be plausibly simulated. A particularly interesting and persistent feature of model results are erosional and depositional phases that lag climatic drivers both by decades, and by several ka within a few hundred meters. The longer lag has not been reported for this spatial scale before and may be an effect of slow landscape-soil-vegetation feedbacks. The combined modelling and fieldwork results allow a more complete understanding of these responses to climate change and can fill in hiatuses in the stratigraphical record. Suggestions are made for methodological adaptations for future LEM studies.