Reliable assessments of future climate development require an improvedunderstanding of natural climate variability on time-scales fromseasons to decades. For this, non-linear interactions betweenprocesses at different spatio-temporal scales play a fundamentalrole. To study such multi-scale interactions the atmospheric modelPLASMA (Parallel LArge-scale Self-adaptive Model of the Atmosphere)has been developed in this DEKLIM project. This model enables therepresentation of two-way feedbacks between global and regional scalesand thus bridges the gap between the resolution of global and regionalmodels. The principal feasibility of our approach, which includes theapplication of new numerical methods and parallelisation techniques,has been proven by implementing a barotropic version of PLASMA andperforming several sensitivity studies. The progress within thisproject is based on the very close and efficient joint effort of threesubprojects resident at AWI Potsdam (AWIP), TU Munich (TUM) and AWIBremerhaven (AWIB).The adaptive model PLASMA has been implemented using an unstructuredtriangular grid. For the numerical implementation of the shallow-waterequations on a parallel high performance computing architecture theadaptive Lagrange-Galerkin method has been applied. The mesh generatoramatos (Adaptive Mesh generator for Atmospheric and OceanicSimulation) provides the complex adaptive data structure. The packageFoSSI (Family of Simplified Solver Interfaces) solves the largesystems of linear equations. For the validation and numerical testingof the model a new analytical test case has been developed whichprovides an analytical solution for unsteady flow. Tests performedwith PLASMA demonstrate the numerical convergence of the modeltowards analytical solutions. Sensitivity studies of the influence ofthe orography on a balanced initial flow field demonstrate thesuccessful modelling of multi-scale interaction processes.
AWI Organizations > Infrastructure > Scientific Computing
Helmholtz Research Programs > MARCOPOLI (2004-2008) > MAR1-Decadal Variability and Global Change