Dimension Reduction for Gaussian Process Emulation: An Application to the Influence of Bathymetry on Tsunami Heights

Abstract

Abstract. High accuracy complex computer models, also called simulators, require large resources in time and memory to produce realistic results. Statistical emulators are computationally cheap approximations of such simulators. They can be built to replace simulators for various purposes, such as the propagation of uncertainties from inputs to outputs or the calibration of some internal parameters against observations. However, when the input space is of high dimension, the construction of an emulator can become prohibitively expensive. In this paper, we introduce a joint framework merging emulation with dimension reduction in order to overcome this hurdle. The gradient-based kernel dimension reduction technique is chosen due to its ability to drastically decrease dimensionality with little loss in information. The Gaussian process emulation technique is combined with this dimension reduction approach. Theoretical properties of the approximation are explored. Our proposed approach provides an answer to the dimension reduction issue in emulation for a wide range of simulation problems that cannot be tackled using existing methods. The efficiency and accuracy of the proposed framework is demonstrated theoretically and compared with other methods on an elliptic partial differential equation (PDE) problem. We finally present a realistic application to tsunami modeling. The uncertainties in the bathymetry (seafloor elevation) are modeled as high-dimensional realizations of a spatial process using a geostatistical approach. Our dimension-reduced emulation enables us to compute the impact of these uncertainties on resulting possible tsunami wave heights near-shore and on-shore. Considering an uncertain earthquake source, we observe a significant increase in the spread of uncertainties in the tsunami heights due to the contribution of the bathymetry uncertainties to the overall uncertainty budget. These results highlight the need to include the effect of uncertainties in the bathymetry in tsunami early warnings and risk assessments.

Rheological considerations for the modelling of submarine sliding at Rockall Bank, NE Atlantic Ocean

“…However, this may be due to the smooth aspect of a Gaussian shape: other irregular (and time-varying) shapes may have a different influence on the resulting wave. However, the investigation of the influence of a high dimensional shape as input is complex and was only recently addressed 41 …”

Dimension Reduction via Gaussian Ridge Functions

“…Before we define these coefficients, we first focus on the covariance term on the right-hand side in (24). In what follows, we employ the squared exponential covariance kernel (25) k…”

“…There are two key ideas that emerge from this discussion. The first is that the posterior mean of a Gaussian process computed on a dimension-reducing subspace is a good candidate for g. In fact, this idea has been previously studied in [36] and [25], albeit using different techniques than those presented here. The second idea is that the suitability of a dimensionreducing subspace is reflected by the posterior variance: should the posterior variance be too large, as in Figure 2(a), then one may need to opt for a different choice of k.…”

Data-Driven Polynomial Ridge Approximation Using Variable Projection

“…Bilionis et al [53], use a related approach from a Bayesian perspective in the context of uncertainty quantification, where the subspace defined by U enables powerful dimension reduction. And Liu and Guillas [32] develop a related low-dimensional model with linear combinations of predictors for Gaussian processes; their approach leverages gradient-based dimension reduction proposed by Fukumizu and Leng [20] to find the linear combination weights.…”