Abstract. Although in situ measurements observed on modern frequently occurring turbidity currents have been performed, the flow characteristics of turbidity currents that occur only once every hundreds of years and deposit turbidites over a large area have not yet been elucidated. In this study, we propose a method for estimating the paleo-hydraulic conditions of turbidity currents from ancient turbidites by using machine learning. In this method, we hypothesize that turbidity currents result from suspended sediment clouds that flow down a steep slope in a submarine canyon and into a gently sloping basin plain. Using inverse modeling, we reconstruct seven model input parameters including the initial flow depth, the sediment concentration and the basin slope. Repeated numerical simulation using one-dimensional shallow water equations under various input parameters generates a dataset of the characteristic features of turbidites. This artificial dataset is then used for supervised training of a deep learning neural network (NN) to produce an inverse model capable of estimating paleo-hydraulic conditions from data of the ancient turbidites. Only 3,500 datasets are needed to train this inverse model. The performance of the inverse model is tested using independently generated datasets. Consequently, the NN successfully reconstructs the flow conditions of the test datasets. In addition, the proposed inverse model is quite robust to random errors in the input data. Judging from the results of subsampling tests, inversion of turbidity currents can be conducted if an individual turbidite can be correlated over 10 km at approximately 1 km intervals. These results suggest that the proposed method can sufficiently analyze field-scale turbidity currents.

This study proposes a new method of inverse analysis from ancient turbidites to the non-steady turbidity currents with consideration of multiple grain-size classes. The forward model employed in this study is based on the shallow water equation, and the initial condition of flows are assumed to be the lock-exchange type condition. To obtain a solution of the inverse problem, this study employed the genetic algorithm for finding the optimized initial conditions. The present method successfully estimated the true given initial conditions of the turbidity currents from the artificial data sets of deposits created by the calculation of the forward model. The author also applied the method to the turbidite bed in the Kiyosumi Formation. As a result of inverse analysis, the obtained solution fits well to the observed data of the individual turbidite, providing estimates of the flow velocity, the flow thickness and the sediment concentration of the turbidity current. The flow thickness and velocity when the turbidity current reached at the downstream end of the study area were reconstructed to be 334.6 m, 0.98 m/s respectively at the location of the downstream end. Result of our analysis is the first example of reconstructing a reasonable conditions of the turbidity current from an ancient turbidite observed in the field, and the method is expected to be applied in various regions in the future.

This study proposes a new method of inverse analysis from ancient turbidites to the non-steady turbidity currents with consideration of multiple grain-size classes. The forward model employed in this study is based on the shallow water equation, and the initial condition of flows are assumed to be the lock-exchange type condition. To obtain a solution of the inverse problem, this study employed the genetic algorithm for finding the optimized initial conditions. The present method successfully estimated the true given initial conditions of the turbidity currents from the artificial data sets of deposits created by the calculation of the forward model. The author also applied the method to the turbidite bed in the Kiyosumi Formation. As a result of inverse analysis, the obtained solution fits well to the observed data of the individual turbidite, providing estimates of the flow velocity, the flow thickness and the sediment concentration of the turbidity current. The flow thickness and velocity when the turbidity current reached at the downstream end of the study area were reconstructed to be 334.6 m, 0.98 m/s respectively at the location of the downstream end. Result of our analysis is the first example of reconstructing a reasonable conditions of the turbidity current from an ancient turbidite observed in the field, and the method id expected to be applied in various regions in the future.

This study proposes a new method of inverse analysis of ancient turbidites to represent non-steady turbidity currents and account for multiple grain-size classes. The forward model employed in this study is based on the shallow water equation, and the initial conditions of flows are assumed as a lock-exchange type. To obtain a solution to the inverse problem, a genetic algorithm is employed to determine the optimal initial conditions. The present method successfully estimated the true initial conditions of the turbidity currents from the artificial data sets of deposits created by the forward model. The method is also applied to a turbidite bed of the Kiyosumi Formation. The results of the inverse analysis yield solutions that fits well with the observed data of the individual turbidite, and provide estimates of the flow velocity, flow thickness and sediment concentration of the turbidity current. The flow thickness and velocity when the turbidity current reached the downstream end of the study area were reconstructed to be 334.6 m, 0.98 m/s, respectively.

Abstract. Although in situ measurements in modern frequently occurring turbidity currents have been performed, the flow characteristics of turbidity currents that occur only once every 100 years and deposit turbidites over a large area have not yet been elucidated. In this study, we propose a method for estimating the paleo-hydraulic conditions of turbidity currents from ancient turbidites by using machine learning. In this method, we hypothesize that turbidity currents result from suspended sediment clouds that flow down a steep slope in a submarine canyon and into a gently sloping basin plain. Using inverse modeling, we reconstruct seven model input parameters including the initial flow depth, the sediment concentration, and the basin slope. A reasonable number (3500) of repetitions of numerical simulations using a one-dimensional layer-averaged model under various input parameters generates a dataset of the characteristic features of turbidites. This artificial dataset is then used for supervised training of a deep-learning neural network (NN) to produce an inverse model capable of estimating paleo-hydraulic conditions from data on the ancient turbidites. The performance of the inverse model is tested using independently generated datasets. Consequently, the NN successfully reconstructs the flow conditions of the test datasets. In addition, the proposed inverse model is quite robust to random errors in the input data. Judging from the results of subsampling tests, inversion of turbidity currents can be conducted if an individual turbidite can be correlated over 10 km at approximately 1 km intervals. These results suggest that the proposed method can sufficiently analyze field-scale turbidity currents.