Evaluation of OLAM performance in TWP-ICE and winter orographic precipitation in Colorado

The Ocean–Land–Atmosphere Model (OLAM) extends the capabilities of Regional Atmospheric Modeling System (RAMS) to a global model coupled to an ocean model. OLAM has several unique attributes that clearly depart from the norm of general circulation models. Among them, an unstructured adaptive grid that enables local mesh refinement to any degree and a representation of complex terrain uses a form of volume-fraction or cut-cell method in which model levels are strictly horizontal. State-of-the-art components of RAMS@CSU microphysics (e.g., explicit activation of aerosols and several bin-emulating approaches) complement the capability of explicitly resolving convection of this global modeling system. In order to evaluate these recent OLAM improvements that are especially relevant in the cloud-resolving scale, we chose two cases that differ in nature: monsoonal tropical convection over the north of Australia and orographic precipitation in Colorado. This first test corresponds to the Tropical Warm Pool-International Cloud Experiment (TWP-ICE) field experiment that took place over and around Darwin, Australia, from January 20 through February 13, 2006. A wide variety of observations from different platforms documented during TWP-ICE represent an almost ideal framework to evaluate the ability of this global model to simulate the evolution of tropical convection during a period governed by different regimes (monsoon, suppressed, and clear periods). The second test corresponds to a wintertime orographic mixed-phase cloud field previously studied in our group. This orographic snowfall case occurred in February 2007 over north-central Colorado. Both cases are simulated using a nearly identical model configuration, a global grid with seven degrees of refinement. Cell size is slightly above 1km within the regions of interest that cover approximately 30000 square kilometers. Comparisons are made between OLAM simulations and observations as well as between RAMS and OLAM output data.