Summary of the Discussions at the DOME Workshop February 22 and 23, 2001 Draft of April 25, 2001, prepared by R. Hallberg General Considerations Following the presentation of a variety of new approaches to the description of the fundamental dynamics of overflows, there was an extended discussion of the general strategy of DOME. Several key points emerged from that discussion. * The results of DOME must be clearly relevant for the climate simulation and other end-user communities. Related studies (notably the FLAME work presented at this workshop and the MICOM group's North Atlantic simulations) do provide a compelling indication that improved treatments of the bottom boundary layer do provide a significant improvement in simulations of the large-scale ocean circulation. * The DOME testcases must evaluate whether numerical approaches to the description of bottom boundary layer dynamics are universal. The ideal parameterization would apply well to the full range of oceanographically significant overflows and bottom boundary layer processes. Equally important, there must be no substantial degradation in simulations due to the inclusion of BBL parameterizations. The equator is a particular location of concern with some proposed approaches. * DOME must have a clear enough focus that there is a high likelihood of unambiguous results. Other, clearly important processes, such as flow through channels, might be addressed in subsequent, focused studies. There was a general endorsement of the three-stage approach in DOME. Namely, idealized studies of overflows, dambreak and more realistic regional studies, and global ocean model intercomparisons complement each other, and ensure that the key objectives (listed above) will likely be met. The Idealized Experiments Many details of the idealized experiments were discussed at some length. There was substantial progress, although there are still several significant issues that must be resolved. Verification of the experiments is one of the more delicate issues. The true flows will be difficult to determine. Nonlinear effects will be significant in most interesting overflows, limiting the possibility of analytic solutions. Laboratory experiments will likely have limited value because of the difficulties of obtaining sufficiently inviscid flows with small enough slopes to be unambiguously oceanographically relevant. The current hope is that high-resolution simulations, some of them nonhydrostatic, with different models will exhibit enough convergence of results that a plausible "truth" can be obtained. In comparison with the "truth", integral properties or quantities that are observable in the ocean will probably be the most illuminating. It is important that a respectable range in parameter space should be sampled in order to assess the robustness and universality of any BBL parameterizations. But the number of experiments should be carefully limited to avoid unnecessary complication in the interpretation of the results. Of order 5 primary test cases will probably give the right balance. Secondary test cases might be considered to answer specific questions with specific models. Although nondimensional parameters will be carefully examined in determining the primary test cases, all results will be presented in dimensional values and with parameters that are typical of ocean overflows to assist in the interpretation of the DOME results. There will be two target resolutions which sample the limits where the deformation radius is and is not resolved. The resolutions of 50 km and 12.5 km were chosen; a suggested deformation radius of about 20 km might be reasonable in some overflows, meeting the criterion for the relative resolution of the deformation radius. Also, 50 km resolution will be typical of many coupled climate simulations over the next decade, while 12.5 km will be typical of eddy-permitting global simulations. In addition, higher resolutions will be of interest, but will be chosen at the discretion of participating groups. Canonical values will be suggested for all key parameters (for example, the bottom drag law), but variations that approximate the way the models would be used in climate simulations will be acceptable (and clearly documents). Similarly, there will be several suggested numbers of vertical degrees of freedom, although the distribution of resolution should approximate that used in real applications with each participating model. Specification of the forcing is a critical issue for the success of DOME. Ideally, the forcing should be imposed in such a way that the volume, density, and velocity of the inflowing dense plume can all be specified, independent of resolution or model type, and that the flow should be in reasonable balance (e.g. geostrophic) at the top of the slope. No such specification is known. There will be further work amongst a smaller subgroup to try to come up with the best compromise specification. One point that was decided is that any net inflow should be exactly balanced by an outflow somewhere in the domain to accommodate rigid lid models. 11 groups indicated a likelihood of participation in the idealized intercomparison. At least one contact person for each group, an institution, and the model are listed below: H. Arango Rutgers U. ROMS B. Barnier LEGI OPA T. Ezer Princeton U. POM D. Haidvogel Rutgers U. SEOM R. Hallberg NOAA/GFDL HIM (MOM also?) M. Hecht / W. Large NCAR NCOM D. Holland NYU MICOM (different BBL from above) P. Killworth/G. Nurser SOC MOMBLE S. Legg WHOI MIT Model T. Özgökmen Miami/RSMAS MICOM P. Schopf COLA Poseidon The More Realistic Phases Tentative plans for regional studies of the Denmark Strait and surrounding areas were presented by J. Dengg, and these were discussed at some length. A separate summary of these discussions is available elsewhere. It is also envisioned that a similar approach may be taken to simulation of the Mediterranean water plume in the Atlantic. Global intercomparisons were not discussed in detail, except to note that they are a fundamental part of the DOME strategy. Details of these intercomparisons should be coordinated with broader ocean model intercomparison projects. Future Actions There will be a closed (i.e. by invitation to active participants) workshop in approximately 6 months time to hammer out details of the experiments and analysis strategy, and to begin to draw conclusions. GFDL was chosen as the location for this meeting, both because of its proximity to many of the participants, and because of the ease of travel to New Jersey from Europe. Meanwhile, work will continue by correspondence to settle on the details of the forcing and experimental design of the idealized cases. A second, open meeting will be held subsequently (in the spring or summer of 2002), with broader contributions from anyone interested in overflows, bottom boundary layers, flow through cataracts and related ocean processes.