Statistical Emulation with Dimension Reduction for Complex Forward Models in Remote Sensing

• Activity Number: 28 - Computation, Design, and Quality Assurance of Physical Science and Engineering Applications
• Type: Contributed
• Date/Time: Sunday, August 8, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Gang Yang* and Emily Kang and Bledar Alex Konomi and Jonathan Hobbs
• Companies: University of Cincinnati and University of Cincinnati and University of Cincinnati and Jet Propulsion Laboratory, California Institute of Technology
• Keywords: Remote sensing; Forward model; Statistical emulation; Gaussian process; Functional principal component analysis; Gradient-based kernel dimension reduction

Abstract

The retrieval algorithms in remote sensing generally involve complex physical forward models that are nonlinear and computationally expensive to evaluate. Statistical emulation provides an alternative with cheap computation and can be used to quantify uncertainty, calibrate model parameters and improve computational efficiency of the retrieval algorithms. Motivated by this, we introduce a framework of building statistical emulators by combining dimension reduction of input and output spaces and Gaussian process modeling. The functional principal component analysis (FPCA) via a conditional expectation method is chosen to reduce the dimension of the output space of the forward model. In addition, the gradient-based kernel dimension reduction method is applied to reduce the dimension of input space when the gradients of the complex forward model are unavailable or computationally prohibitive. A Gaussian process with feasible computation is then constructed at the low-dimensional input and output spaces. Theoretical properties of the resulting statistical emulator are explored, and the proposed method is illustrated by application to NASA’s Orbiting Carbon Observatory-2 (OCO2) data.