- Reexamine nesting BC algorithm for Uhat - scheme 0: as is: Interpolate Ubaro(n+1). Use in vertically integrated momentum equation to generate boundary values for Uhat(n+1) - scheme 1: Directly interpolate Uhat(n+1) from big to small domain. (requires big domain to interpolate different variable. Requires small domain to by-pass constructing boundary values from momentum eq.) - scheme 2: Also update large domain cell value that only contains fine domain boundary cells. (e.g. 2/3 original-coarse-value + 1/3 average-of-fine-values) - can be combined with other schemes - scheme 3: use scheme 0 but with interpolated Uhat(n-1). No longer certain what that means. - The Uhats in schemes 0 & 1 are, in the coarse grid, the same to 3rd order. scheme 0 advantage: it contains fine scale data in the BCs scheme 1 advantage: it contains Large-scale data outside domain at time n+1. Scheme 0 only has partial large scale Uhat data at time (n+1). - Construct a list of the various simplifications to the nesting scheme and what it would entail to code them for comparison in paper. - one-way - "explicit" two-way nesting need to change order of operations. 1 man-hr to code. - not updating RHS of coarse barotropic momentum before constructing surface pressure. need to skip a section of code. 30 man-minutes to code. - Going more complicated - Update eta(n+1) in interior of large domain instead of eta(n) Include all necessary supporting updates. - Instead of interpolating eta(n+1) at boundaries, pass coarse grid data used for the interpolation. Formulate deta BC's so boundary data will be consistent with updated coarse domain. - Incorporate volume change updates with nesting updates - Merge what we have now. - see if need similar updates on nesting (pass volume integrals, rescale with final, updated, volumes).