Introduction to MPAS Dynamic Core, Adjoint Model and Their Future Extension and Applications

The dynamic framework of the Model for Prediction Across Scales-Atmosphere (MPAS-A) has unstructured Spherical Centroidal Voronoi Tesselation (SCVT) meshes with C-staggering, it represents one of the most important advances in numerical weather prediction (NWP) over the past few decades. Because of the openness, normative computer program and documents, scientific nature and progressiveness, the MPAS-A is chosen for developing an advanced global four-dimensional variational (4D-Var) data assimilation (DA) system. As the first step, the tangent linear model and adjoint model of MPAS-A are developed. The advanced global MPAS-A 4D-Var DA system not only avoids the errors arising from back-and-forth interpolations between MPAS-A forecasting analysis grids and non MPAS-A DA grids as well as reduces computational cost arising from conversions between MPAS-A model variables and non MPAS-A DA analysis variables at each iteration of minimization, but also provides new opportunities for across-scales global DA with unstructured spherical centroidal meshes for weather and climate studies.