Commit by bueler :: r1206 /trunk/ (8 files in 4 dirs):

Fixed diagnostic diffusivity (of SIA) calculation: now it happens only in one place, namely velocitySIAStaggered(). We save diffusivity on staggered grid, in memory also used for 'hardav' in SSA; this is now IceModelVec2Stag vWork2dStag. Also a bit more reporting on bed smoother initialization.

Commit by bueler :: r1205 /trunk/ (3 files in 2 dirs):

Added and documented option -bed_smoother_range. Added slightly to performance stats put in history string.

Commit by bueler :: r1204 /trunk/ (8 files in 5 dirs):

Generally bed smoother/roughness-parameterizer ready to go. Thickness details (margin) fiddles done. Derived classes turn the mechanism off. Half-width of smoothing domain set to 5km, but not that there is always one grid point in each direction if the mechanism is on at all. Remains to add option adjusting this range.

Commit by bueler :: r1203 /trunk/src/base/ (iMinit.cc iMsia.cc):

Important fix inside velocitySIAStaggered() so that theta is correctly on the staggered grid point.

Commit by bueler :: r1202 /trunk/ (7 files in 5 dirs):

Improved bed roughness parameterization scheme by recognizing 4th order Taylor polynomial was convex like the function it is approximating. Updated tools to reflect that.

Commit by bueler :: r1201 /trunk/src/ (4 files in 2 dirs):

Clean up of r1200 commit: revert changes to iceModelVec2.cc (why can't global go to proc0?); fix bedrough_test.cc so test 26 passes; clobber 'possibly uninitialized poly_ice' compile warning for flowlaw_test.cc. Test 22 still fails, but this failure un-observable by eye.

Commit by bueler :: r1200 /trunk/ (12 files in 5 dirs):

Work on bed roughness parameterizer. Fails a couple of tests even when turned off.

Commit by bueler :: r1199 /trunk/ (10 files in 6 dirs):

Working PISMBedSmoother, still not attached to IceModel (that is next). Added min() and max() capability to IceModelVec2S, but have not used it in many possible applications in IceModel.

Commit by bueler :: r1198 /trunk/src/ (base/grid.cc trybedrough.cc):

bed roughness parameterization test program runs minimally, but PISMBedSmoother is not attached to IceModel

Commit by ckhroulev :: r1197 /trunk/ (4 files in 3 dirs):

Added flowlaw_test, an executable similar to flowTable. flowlaw_test accepts the -ice_type command-line option but does not compute GK parts. It is used by the flow law regression tests 24 and 25.

Commit by bueler :: r1196 /trunk/ (7 files in 4 dirs):

Initial implementation of bed smoother. Compiles, but PISMBedSmoother is not known to IceModel, and test program trybedrough.cc not complete.

Commit by bueler :: r1195 /trunk/ (4 files in 3 dirs):

Work on implementing Schoof (2003) strategy. New C++ code compiles but is incomplete. docs are pretty complete.

Commit by bueler :: r1194 /trunk/ (3 files in 3 dirs):

More work related to bed roughness, and the addition of option -paleo_precip into spinup script for SeaRISE Greenland, as it was used.

Commit by bueler :: r1193 /trunk/ (5 files in 5 dirs):

Cleanup and uniformity of names in src/base/ directory.

Commit by bueler :: r1192 /trunk/ (13 files in 2 dirs):

More tools and text related to implementing Schoof (2003) bed roughness parameterization technique. The Matlab/Octave codes in src/base/bedroughplay/ come from Ed's Karthaus/McCarthy pedagogical codes.

Commit by bueler :: r1191 /trunk/doc/ (browser_bedrough.txt doxyfile_browser):

a bit more docs re bed smoothing Schoof (2003) technique

Commit by bueler :: r1190 /trunk/doc/ (3 files):

Modest start on bed roughness improvements related to Schoof (2003).

Commit by bueler :: r1189 /trunk/ (5 files in 2 dirs):

Testing on marmaduke reveal that r1187 produces small differences on different machines for tests 22 and 23; may be libraries? In any case, tolerances lowered to generally 10^-7 with several exceptions down to 10^-5.

Commit by aaschwanden :: r1188 /trunk/src/ (4 files in 2 dirs):

Squashed commit of the following:

commit 61802c9a117ac38f035d4624130c73f145982ffb
Author: Andy Aschwanden <andy dot aschwanden at arsc dot edu>
Date:   Tue Jun 29 14:09:16 2010 -0800

    cold-ice methods related changes/additions in postprocessing, i.e. ice is treated as temperate if within a "cold_mode_is_temperate_ice_tolerance" of the pressure melting point

Commit by bueler :: r1187 /trunk/test/software/pism/ (17 files):

Completely superficial changes to look of software tests. Stop me before I fiddle again ...