17:15 〜 19:15
[MIS01-P01] 平面2次元混濁流モデリングの野外スケールへの適用―1929年グランドバンクス地震の例―
キーワード：タービダイト、混濁流モデリング、Grand Banks
Turbidity currents are gravity-driven flows of sediment particles, typically triggered by slope failures and floods. In paleo-seismology, recurrence intervals of large earthquakes are often inferred from the stratigraphy of turbidites, which are interpreted to have been deposited by strong seismic motions. However, the formation of turbidites involves both the generation and propagation of turbidity current. If this process can be quantitatively simulated, one could reconstruct the seismic intensity of paleoearthquakes from the spatial distribution of turbidites. Yet, there are few examples in which field-scale turbidity currents have been reproduced using two-dimensional physical models. This scarcity arises primarily from the lack of sufficiently detailed observational data for verification, as well as potential discrepancies between empirical formulas derived at laboratory scales and those in the field scales. This study estimates empirical coefficients for sediment transport at the field scale by reproducing the observed data from the 1929 Grand Banks event.
The 1929 Grand Banks earthquake generated one of the historical largest turbidity currents. During this event, flow velocities (ranging from 3 to 19 m/s) were inferred from the sequential timing of submarine cable failures. Additionally, marine cores obtained in this area provide information on turbidite thickness and grain size distribution. This study used a trial-and-error approach to identify model parameters that reproduce both the sediment distribution and hydrodynamic conditions reported in previous investigations.
We modeled turbidity currents using turb2d (Naruse, 2020), a three-equation model based on momentum conservation, fluid mass conservation, and sediment mass conservation (Parker et al., 1986). Dimensionless water entrainment rate was calculated using the empirical formula of Parker et al. (1987), which relates the enlargement of the turbidity current to the bulk Richardson number. We defined the first tuning coefficient for water entrainment as ef. We also included water detrainment, which counteracts water entrainment due to particle settling velocity, and defined a second tuning coefficient df. The sediment-transport closure equations are derived from tank experiments and river observations; dimensionless basal sediment entrainment rate follows the formula reanalyzed by Traer et al. (2012) based on Garcia and Parker (1991). In these closure equations, the maximum near-bed sediment concentration is set to 30%. We considered the maximum near-bed concentration as the third tuning coefficient Camax, which becomes relevant for flow velocities exceeding several m/s. We divided sediment grain sizes into three classes—fine sand, very fine sand, and medium silt—and used the 64 µm grain size class (which observed in the most core data) for primary verification.
We found that conditions producing results broadly consistent with the observed sediment distribution were ef = 0.5–1.0, df = 1.5–3, and Camax = 0.05–0.1. The parameters ef, df, and Camax primarily controled the spatial extent of deposition, the trend of thinning in sediment layer thickness, and the absolute layer thickness, respectively. The value of ef lied within the range supported by experimental data. The value of df was generally consistent with that from direct numerical simulations (Salinas et al., 2019), and including water detrainment was essential to reproduce the observed distribution of turbidites. It has previously been difficult to constrain Camax due to the scarcity of observations at flow velocities above 2 m/s. In the 1929 Grand Banks event, velocities remained above this threshold until reaching the abyssal plain, suggesting that a value of Camax exceeding 0.1 (10%) would cause excessive sedimentation and erosion.
By reproducing one of the largest known events—the 1929 Grand Banks earthquake—we obtained empirical coefficients that appear applicable at field scales. These results will be used for future numerical simulations of field-scale turbidity currents.
The 1929 Grand Banks earthquake generated one of the historical largest turbidity currents. During this event, flow velocities (ranging from 3 to 19 m/s) were inferred from the sequential timing of submarine cable failures. Additionally, marine cores obtained in this area provide information on turbidite thickness and grain size distribution. This study used a trial-and-error approach to identify model parameters that reproduce both the sediment distribution and hydrodynamic conditions reported in previous investigations.
We modeled turbidity currents using turb2d (Naruse, 2020), a three-equation model based on momentum conservation, fluid mass conservation, and sediment mass conservation (Parker et al., 1986). Dimensionless water entrainment rate was calculated using the empirical formula of Parker et al. (1987), which relates the enlargement of the turbidity current to the bulk Richardson number. We defined the first tuning coefficient for water entrainment as ef. We also included water detrainment, which counteracts water entrainment due to particle settling velocity, and defined a second tuning coefficient df. The sediment-transport closure equations are derived from tank experiments and river observations; dimensionless basal sediment entrainment rate follows the formula reanalyzed by Traer et al. (2012) based on Garcia and Parker (1991). In these closure equations, the maximum near-bed sediment concentration is set to 30%. We considered the maximum near-bed concentration as the third tuning coefficient Camax, which becomes relevant for flow velocities exceeding several m/s. We divided sediment grain sizes into three classes—fine sand, very fine sand, and medium silt—and used the 64 µm grain size class (which observed in the most core data) for primary verification.
We found that conditions producing results broadly consistent with the observed sediment distribution were ef = 0.5–1.0, df = 1.5–3, and Camax = 0.05–0.1. The parameters ef, df, and Camax primarily controled the spatial extent of deposition, the trend of thinning in sediment layer thickness, and the absolute layer thickness, respectively. The value of ef lied within the range supported by experimental data. The value of df was generally consistent with that from direct numerical simulations (Salinas et al., 2019), and including water detrainment was essential to reproduce the observed distribution of turbidites. It has previously been difficult to constrain Camax due to the scarcity of observations at flow velocities above 2 m/s. In the 1929 Grand Banks event, velocities remained above this threshold until reaching the abyssal plain, suggesting that a value of Camax exceeding 0.1 (10%) would cause excessive sedimentation and erosion.
By reproducing one of the largest known events—the 1929 Grand Banks earthquake—we obtained empirical coefficients that appear applicable at field scales. These results will be used for future numerical simulations of field-scale turbidity currents.