Transposition of a Process-Based Model, Flumy: from Meandering Fluvial Systems to Channelized Turbidite Systems

Channelized turbidite systems are among the most important hydrocarbon reservoirs. Yet building realistic turbidite reservoir models is still a challenge. Flumy has been firstly developed to simulate the long-term evolution of aggrading meandering fluvial systems in order to build facies reservoir models. In this study, Flumy has been transposed to channelized turbidite systems. The channel migration linear model of Imran et al. (1999) dedicated to subaqueous flows has been implemented. The whole model has been calibrated taking into account the differences on channel morphology, avulsion frequency, and aggradation and migration rates. This calibration and the comparison of the model to natural systems rely on: i) the channel planform morphology characterized by the meander wavelength, amplitude, and sinuosity; ii) the channel trajectory and the resulting stratigraphic architecture described using Jobe et al. (2016) indexes. Flumy succeeds in reproducing turbidite channel planform morphology as shown by the mean sinuosity of 1.7, the wavelength to width and amplitude to width ratios around 4 and 1 respectively. First-order meander architecture, characterized by the ratios meander belt width versus channel width, meander belt thickness versus channel depth, and the deduced stratigraphic mobility number (the ratio between lateral versus vertical channel displacements), is also well reproduced: 2.5, 3.8, and 0.6 respectively. Both lateral and downstream channel normalized migrations are around 3.5 times lower than in fluvial systems. All these values are absolutely coherent with the observations. In the other hand, the channel trajectory observed on seismic cross sections (hockey stick geometry) is not fully reproduced: the local stratigraphic mobility number is divided upward by 3 whereas up to 10 is expected. This behavior is generally explained in the literature by an increasing aggradation rate through time and/or flow stripping at outer bend that decreases lateral migration rate (Peakall et al., 2000). These processes are not currently simulated in Flumy, and need to be implemented. This study shows that Flumy model reproduces quite well the first order characteristics observed in the nature and can be used to simulate channelized turbidite reservoirs.