Sensitivity analysis and predictability of deep propagating gravity waves

It is well known that the prediction of topographically-forced phenomena such as mountain waves is very sensitive to the properties of the upstream flow, namely the cross-barrier wind speed and static stability. Mountain waves are an example of a class of threshold phenomena that may develop or change regimes (e.g., breaking, nonlinear, etc.) when the basic properties of the upstream flow change through relatively small perturbations induced by the synoptic-scale or mesoscale flow. In this study, we focus on mountain waves generated by the Southern Andes, in part motivated by satellite observations that suggest this region in winter contains the largest stratospheric gravity wave (GW) amplitudes on the planet. Additionally, a new field program is being proposed, the Southern Andes – ANtarctic GRavity wave InitiAtive (SAANGRIA), that would employ the new NSF/NCAR GV (NGV) research aircraft from a base near the southern tip of South America in a 10-week field measurement campaign from late June to early September, 2012. This region, spanning the southern Andes, Drake Passage, and Antarctic Peninsula, is unique in that strong surface and upper-level winds in winter permit GWs to propagate to very high altitudes. Given the large-amplitude GWs that propagate routinely into the middle atmosphere, the region offers an ideal natural laboratory for studying the multi-scale predictability of GWs and their upscale influence on the larger-scale circulation, which is one of the objectives of SAANGRIA. Ensemble- and adjoint-based tools will be used to guide targeted observations over the Pacific Ocean upstream of the Andes in order to test predictability theories and methods of predictability analysis during SAANGRIA.

In this study, the nonlinear, adjoint and tangent linear models for the atmospheric portion of the nonhydrostatic Coupled Atmosphere/Ocean Mesoscale Prediction System (COAMPS) are used to explore the mesoscale sensitivity and predictability characteristics associated with airflow impinging on the Southern Andes. We analyze results from real world simulations during Aug.-Sep. 2008 and 2009, as well as idealized simulations. Results indicate that the 24-h forecast cross-mountain winds and mountain wave response are very sensitive to the initial state and in particular to synoptic-scale and mesoscale characteristics of mid-latitude cyclones. The mountain waves and cross-mountain winds are most sensitive to upstream features in the initial state that are present in the lower troposphere. The results indicate that predictability of mountain waves are limited by rapid perturbation growth on multiple scales. On the synoptic-scale, rapid growth associated with baroclinic waves impact the stability and cross barrier wind speed upstream of the Andes. On the smaller scales, the mountain waves are predominantly influenced by the wave launching conditions (upstream stratification, crest-level wind shear). We also examine the gravity wave sensitivity to the horizontal and vertical shear in both the real data and idealized frameworks.