Autogenic masking of allogenic inputs: Numerical modelling of signal preservation in deep-water fan strata

Abstract

Listen

Translate article

Variations in climate, sea-level and tectonic processes can be recorded in various ways in strata, and recognising and measuring such external signals preserved in strata is an important aspect of deciphering Earth history. Key questions are what type of signals are likely to be preserved; how input signals may be shredded, over-printed or otherwise reduced before preservation, by stochastic or autogenic processes; and how allocyclic products may be mimicked and masked by autogenic processes. A reduced-complexity numerical forward model of down-slope sediment erosion, transport and dispersive deposition of submarine fan strata is the simplest-possible formulation that produces reasonably realistic strata that can be robustly analysed for signal content across model realisations with a range of external forcing amplitudes and periods. Three different initial topographic surfaces that include increasing influence of random noise allow comparison of the signal preserved with different levels of random influence on the deterministic model. Spectral analysis, with rigorous testing for statistical significance, shows that the signal recorded depends strongly on the level of noise present in the underlying topography, that short-period high-amplitude allocyclic signals are the most likely to be preserved in this case, and that autogenic dynamics in the modelled system can effectively mask even long-period high-amplitude input signals because they have similar periodicities. Autogenic processes in this model can also produce 5:1 bundling of cyclical strata commonly assumed to be strong evidence for orbital forcing. This analysis suggests that we might be substantially underestimating the complexity involved in extracting a signal of external forcing from strata where any autogenic processes operate, to the extent that the two different effects may often be practically indistinguishable, indicating that a multiple hypothesis approach is important to account for the resulting uncertainty.