Application of Manifold Methods for Data Assimilation in Integrated Earth System Models

Abstract:

A novel manifold-based method is presented to assimilate different types of spatio-temporal data in integrated land and oceanic models. This method has been developed based on the assumption that hydrologic, atmospheric and oceanic data can be mapped onto an underlying differential manifold. In this study, the proposed method is employed to reconstruct meteorological forcing dataset over Lake Michigan, bathymetry of Gull Lake, and precipitation over the Grand River watershed in the State of Michigan. In the first case study, hourly interpolated meteorological forcing parameters are used to run a three-dimensional hydrodynamic model of Lake Michigan to show the improvement that results from the use of the interpolation method by comparing model results with observed data for currents. In the second case study, the bathymetry of the Gull Lake is interpolated from the scatter point data using the manifold technique. A hydrodynamic model of Gull Lake has been developed and improved by using the interpolated bathymetry. In the last case study, 13-year daily participation data are interpolated over the Grand River watershed and used as input for an integrated, distributed watershed model. All three case studies illustrate the effectiveness of the presented manifold based interpolation algorithm.