C****************************************************************************** C This file contains a log of changes made to (P)ADCIRC. * C * C mod history C v51.52.34 - 01/25/2016 - jgf - Fixed issue where nodecode was being tested C on output arrays where wet/dry state should be considered, but C this test was being erroneously applied to stations by looking up C the nodecode on the station number itself, rather than the C nodes in the element containing the station; checking the C wet/dry state of a station was already being correctly handled C in subroutine stationArrayInterp(). C v51.52.33 - 01/14/2016 - tcm - Added compiler flags for Topaz, SGI ICE X. C v51.52.32 - 11/17/2015 - jgf - Remove stray debugging statement from adcprep. C v51.52.31 - 11/15/2015 - zc - Fixing bug where SWAN was writing the previous C time snap instead of the current by moving C writeOutput2D call out of timestep.F and into C adcirc.F after the SWAN time step has taken place C v51.52.30 - 11/11/2015 - jgf - Fixed output of time varying tau0 so that C ascii fort.90 files have a proper header; fort.90 and fort.90.nc C files are only written when tau0=-x.1 in fort.15. C v51.52.29 - 09/18/2015 - jgf - Fixed parsing and i/o error checking in C subroutine rsget (for reading fort.23). C v51.52.28 - 09/18/2015 - jgf - Fixed check that compares number of nodes C in mesh to number of nodes in fort.13 so that it occurs C immediately when fort.13 header is read. C v51.52.27 - 09/16/2015 - jgf - Added call to subroutine updateMetaData() C at end of subroutine writeNetCDFHotstartFile() to ensure that C netcdf metadata is updated in netcdf hotstart files when new C hotstart data are written to them. C v51.52.26 - 09/09/2015 - jgf - Removed stray debugging statements from C couple2swan.F and tweaked comments for NWS10GET. C v51.52.25 - 09/04/2015 - yf/jgf - Fixed initial guess for dradz in subroutine C glats() (for NWS10, i.e., GFS input) reducing it by a factor of 8. C v51.52.24 - 08/24/2015 - jgf - Modified loop extents from previous C change such that the QNMSG array is populated at indices C that correspond to the position in the LBCODE array where C the flux boundary conditions are specified. C v51.52.23 - 08/23/2015 - jgf - Fixed adcprep issues with array extents for C external flux boundaries and the LBCODE array for boundary C types at nodes in adcprep; the issues manifested as C too many flux values read from fort.15 for periodic river C flux boundary conditions. C v51.52.22 - 08/17/2015 - jgf - Added compiler flags for spirit at AFRL. C v51.52.21 - 08/14/2015 - jgf - Fixed NWS10 (NOAA GFS) meteorological input C subroutines so that they function properly in serial and parallel; C previously, there was no accommodation for NWS10 in adcprep, whereas C now adcprep recognizes this as a valid value; subroutines in C adcirc were nonfunctional due to outdated parameters and code C structure; adcirc code for NWS10GET was fixed and made functional C by bringing it in line with the overall structure of the wind C module; also fixed an issue with ramp value initialization when C nramp=1 and dramp=0.0. C v51.52.20 - 08/06/2015 - jgf - Fix to prevent NaNs in output files C in runs that use SWAN coupling with a vortex meteorological C model. This issue was a result of the parametric vortex C models not initializing the WVNX1 and WVNY1 arrays because C they do not use them but SWAN does. C v51.52.19 - 08/04/2015 - zc - Fixing bug that caused serial adcswan to not C write output files. descript % writeFlag was never set to .true. C v51.52.18 - 08/02/2015 - zc - Fixing bug in writing netCDF stations using C serial adcirc. Pointers to station coordinates were not defined. C v51.52.17 - 07/16/2015 - tcm - Fixed bug in time varying bathymetry C implementation, where the wrong amount of bathymetry change was being C used for adjusting the eta2 variable. C v51.52.16 - 07/12/2015 - jgf - Uncommented format statement that had C been accidentally commented out causing compilation issue. C v51.52.15 - 07/12/2015 - jgf - Removed consideration of noff value when C determining wet/dry status of a station for output purposes. C v51.52.14 - 06/12/2015 - jgf - Removed error message emitted by C adcprep and adcirc when NOLIFA=2 and NOLICAT=0 because it is C alarming and because it contradicts common practice. C v51.52.13 - 06/10/2015 - jgf - Added compiler flags for stampede at TACC. C v51.52.12 - 06/10/2015 - jgf - Fixed issue in adcprep where the EL_SHARE C array was dimensioned incorrectly, resulting in nearly all element C indices in the IMAP_EL_LG() mapping array being negative C (i.e., represented as ghost elements); the fix assigns a C positive element index in the mapping for the first subdomain C that contains the element, and a negative element index for C any subsequent subdomain that contains the element. This C fix will resolve issues with globalization of the NOFF array C for fulldomain hotstart files, which have been associated with C otherwise inexplicable numerical issues in prior versions. C v51.52.11 - 05/20/2015 - zc - Further fixes for multiple netCDF writer cores C and writing the fort.69 debug file. C v51.52.10 - 05/17/2015 - jgf - Fix for globalizing NOFF array that takes C the absolute value of the fulldomain element number when C retrieving it from the IMAP_EL_LG mapping array. This fixes C the issue where all, or nearly all, NOFF values were found C to be zero in the fulldomain hotstart file. C v51.52.09 - 05/14/2015 - zc - Bug fixes for compiling with XDMF libraries C v51.52.08 - 05/14/2015 - zc - Fixed netCDF writer processor which did not C have knowledge of NP_G and NE_G during C initialization. This is now read during the C readMetadata routine. C v51.52.07 - 05/06/2015 - jgf - Fixed two memory allocation issues for harmonic C analysis i/o in parallel; added command line option to direct C ADCIRC to simply report version number and exit. C v51.52.06 - 05/05/2015 - jgf - Fixed memory allocation issue when performing C globalio in parallel for ESLNodes.63 file. C v51.52.05 - 04/21/2015 - jgf - Fixed issue with NWS=12 (OWI) where barometric C pressure at nodes outside both met grids was set to zero instead C of the background pressure of 1013mb; the issue did not affect C situations where the OWI grid completely covered the ADCIRC mesh. C v51.52.04 - 04/21/2015 - jgf - Added compiler flags for LSU supermike2. C v51.52.03 - 04/20/2015 - jgf - Added compiler flags for LONI queenbee. C v51.52.02 - 04/13/2015 - jgf - Fixed compilation issue when compiling with C DEBUG_WARN_ELEV flag. C v51.52.01 - 04/13/2015 - jgf - Fixed compilation issue with harmonic analysis C data post processing in subdomain hotstart files in adcpost. C v51.52 - 03/28/2015 - jgf - Updated comments, log message integer formatting C (i0 instead of ix) and execution tracing for GFDL. C v51.51 - 03/25/2015 - jgf - In globalio module, added a check for C remainder if the maximum buffer size is not evenly divisible C by the rank of the globalized array; removed NOFF=2 handling C from globalio since it is already present in the calling routine; C fixed excessive logging for slope limiter now that the slope C limiter activity is reset every time step; added individual C spool counters for output types in write_output module that C previously shared spool counters, which caused them to interfere C with each other (e.g., the fort.72 was being written but not the C fort.71). C v51.50 - 03/21/2015 - jgf - Added back the clearing of the slope C limiter flag for all nodes at each time step prior to C re-evaluating whether to activate it again; adjusted the USE C statements at the top of the new slope limiter subroutines C to refer to the mesh module. C v51.49 - 03/16/2015 - tcm - Modifications to make slope limiting a callable C subroutine. C v51.48 - 03/16/2015 - jgf - Merged aaltuntas' fix on r1072 of the C ncsu-subdomain branch to deactivate geoidOffset for subdomain C boundary conditions. C v51.47 - 03/13/2015 - jgf - Consolidated the elemental slope limiter C into a single block of code at the end of GWCE_New() to C prevent non-limited slopes from being used in the momentum C equation solution. C v51.46 - 03/12/2015 - jgf - Fixed the allocation/initialization of wave C radiation stress arrays in global.F, hstart.F, and cstart.F C to prevent NaNs from uninitialized values at hotstart from C creeping into the solution for water surface elevation and C causing SWAN coupled simulations to terminate immediately; C applied a fix from Casey to eliminate a floating point C exception in SWAN when total water depth in ADCIRC is negative; C changed allocation and initialization of ESLONOFF array so that C memory is not allocated twice and so that the slope limiter C at a node is only activated at hotstart if the current value C of the nodal attribute at that node would allow it. C v51.45 - 03/09/2015 - jgf - Merged aa's latest fixes for subdomain modeling. C v51.44 - 03/09/2015 - jgf - Merged cs/kd changes to support writing and C post processing subdomain harmonic analysis output files as well C as corrections for writing harmonic analysis data to subdomain C hotstart files. These changes enable ADCIRC to perform harmonic C analysis for a larger number of harmonic constituents (e.g., C more than 30) while executing in parallel by avoiding the C i/o memory demand associated with globalio for harmonics. C v51.43 - 02/10/2014 - jgf - Fixed log messages in mesh.F because each of C the subdomains was dumping its chatter to the screen in parallel; C fixed the splash header (with version and copyright info) so that C it appears at the very top of the fort.16; removed erroneous C log message about the JCG solver. C v51.42 - 01/28/2014 - jgf - Merged subdomain modeling capability; the C changes to existing source code are minimal and limited to a few C additional subroutine calls; nearly all the code to support the C new feature is contained in the new modules subprep.F (for C adcprep) and subdomain.F (for adcirc/padcirc/padcswan). C v51.41 - 01/21/2014 - jgf - Added more error checking to hstime C utility to detect netcdf hotstart files that have a mesh C but no data. C v51.40 - 12/29/2014 - jgf - Fixed issue in parsing header of nodal C attributes file where the reading of the default value of an C unrecognized nodal attribute was not being skipped as expected. C v51.39 - 12/24/2014 - jgf - Added execution tracing to mesh.F module. C corrected log message routing in mesh module to reduce the C messages it prints to the screen, particularly in parllel. C v51.38 - 12/22/2014 - jgf - Merged GAHM_jie branch (v51.33.x). C v51.33.14 - 12/12/2014 - jgf - Fixed duplicate execution tracing calls. C v51.33.13 - 12/09/2014 - jgf - Merged trunk changes r1032:1050; fixed some C issues related to the move of output type identifiers like C OFF, ASCII, BINARY, etc from global.F to sizes.F. C v51.33.12 - 11/18/2014 - jgf - Added comma to hstime.F to fix syntax when C not compiling with NETCDF support. C v51.33.11 - 11/12/2014 - Decided to adjust compiler flags for debugging C to remove trapping on underflow floating point exceptions, C since default compiler behavior is likely to handle such C exceptions correctly (flush to zero) during production. C v51.33.10 - 11/07/2014 - jgf - Corrected stray subroutine call from C adcirc_terminate() to terminate(). C v51.33.09 - 11/07/2014 - jgf - Added option to compile with debugging to C cmplrflags.mk for intel-lonestar; also added SWAN=enable C option on the make command line for same (instead of C hardcoding -DADCSWAN into the DPRE flags); added dependency C of wind module on mesh module into makefile; fixed misplaced C END SELECT in couple2swan that was erroneously inside #ifdef C causing compilation issue if compiled without netcdf support; C fixes to enable compilation with debugging using intel11 C compilers on lonestar at TACC. C v51.33.08 - 11/06/2014 - jgf - Fixes to enable compilation in debug mode C with ftn compilers on garnet at ERDC. C v51.33.07 - 11/06/2014 - jgf - Fixes to enable compilation with ftn C compilers on garnet at ERDC; added execution tracing to C write_output module. C v51.33.06 - 10/31/2014 - jgf - Fixes to enable compilation with intel14 C compilers on hatteras at RENCI. C v51.33.05 - 10/30/2014 - jgf - Fixed issue with adcprep: mesh coordinates C were not suitably initialized, causing floating point exception C problems in the setup_kdt_search subroutine; eliminated C redundant specification of deg2rad, rad2deg, and Rearth C parameters as well as CPP projection sub between adcprep and C adcirc, including CPP projection of station coordinates; C removed INVCPP subroutine from adcprep since it was not C called from anywhere. C v51.33.04 - 10/26/2014 - jgf - Updated this file by adding entries for C each commit that has been made on this branch; merged trunk C changes that occurred from r1014 through r1031; minor fixes C to USE statements in aswip.F to allow the aswip executable C to compile; tested and confirmed that all executables C compile successfully. C v51.33.03 - 08/25/2014 - gjie - Fixed arrays. C v51.33.02 - 08/01/2014 - gjie - Added geofactor to control the geostrophic C balance. C v51.33.01 - 07/21/2014 - gjie - Added GAHM vortex wind model with NWS=20. C v51.37 - 11/11/2014 - zc - Fixing allocation of mpistat in C writer_main subroutine. Setting ESLDescript % writeFlag C flag to false so code does not attempt to write C except at end of simulation. Adding commands to clean C SWAN directory temporary fortran files when 'make clean' C or 'make clobber' is run. Adding attribute to C OutputDataDescript_t to flag writing of min/max file C ascii time stamp. Reorganizing SWAN output routines in C SwanOutput subroutine within couple2swan.F into C writeOutput2D so they use Jason`s pointer structure for C output variables and two sets of output routines are no C longer carried. This also allows Swan output to occur even C during non-coupling time steps. C v51.36 - 11/07/2014 - zc - Adjusting makefile for adcprep missing C dependencies. Removing R,DEG2RAD,RAD2DEG from pre_global.F C so that global.F versions are used. Renaming R in C read_global.F throughout REarth. Fixing bugs in mesh.F and C write_output.F C v51.35 - 10/25/2014 - jgf - Merged all xdmf2 developments (v51.21.xx) C to trunk. C v51.21.41 - 10/16/2014 - jgf - Continuing to fix issues discovered in C automated testing with ASGS: fixed issues related to constructing C file names when using dedicated writer processors to write C netcdf files; fixed log messages created by dedicated writer C processors so the messages will go into the subdomain C fort.16 files associated with those processors; added C capability for the code to echo the command line that was C used to execute it to document the associated command line C options (e.g., the number of writer processors). C v51.21.40 - 10/13/2014 - jgf - Fixed issue with dedicated writers when C there are many writers: the writers_write signal was not C being honored unless all writers had received some data. C v51.21.39 - 10/10/2014 - jgf - Fixed issue in adcprep for generation C of netcdf hotstart files containing harmonic analysis C info for means and variances calculations. C v51.21.38 - 10/03/2014 - jgf - Fixes related to merge: prevented C netcdf files from being initialized twice in parallel (once C in adcprep and once in padcirc --- now only in adcprep); C fixed issue where fort.90.nc was not being initialized in C adcprep. C v51.21.37 - 10/02/2014 - jgf - Improvements to makefile statements that C record the compiler flags that were used to compile each C executable; fixes that allow fort.90 (tau0 output) file to C be written in netcdf if the fort.63 was set to write in netcdf; C fixes for adcprep to repair the setup of netcdf files (was C broken by the merge of the trunk and branch but is now fixed). C v51.21.36 - 09/29/2014 - jgf - Fixes from functional testing: fixed C initialization of netcdf min/max files. C v51.21.35 - 09/26/2014 - jgf - Fixes uncovered during functional testing: C fixed memory allocation for time of max variable in min/max C output files; fixed naming of netcdf files including min/max C netcdf files; added code to makefile to record the actual C compiler flags to a file for later use in determining e.g. C whether the code was compiled with debugging turned on. C v51.21.34 - 09/24/2014 - jgf - Fixes to merged and updated branch for C compiling without XDMF support. C v51.21.33 - 09/23/2014 - jgf - Merging changes that have been made C in the trunk from v51.21 (trunk revision 936) to the C latest trunk v51.33 (trunk revision 1016) into this C xdmf2 branch. C v51.21.32 - 03/25/2014 - jgf - Added control.F and xdmfio.F to version C control, since they were overlooked in previous commits. C v51.21.31 - 03/25/2014 - jgf - Fixes resulting from compliance testing. C v51.21.30 - 03/21/2014 - jgf - Fixes from testing and preparations for C merging back to trunk. C v51.21.29 - 03/14/2014 - jgf - Fixes to dedicated writer code for C initializing and writing XDMF output files. C v51.21.28 - 03/13/2014 - jgf - Refactored XDMF output routines into their C own module so they can be called from write_output.F as well as C writer.F for use with dedicated writers. C v51.21.27 - 03/13/2014 - jgf - Updated adcirc2xdmf to convert only a C limited range of datasets if desired; significantly refactored C and simplified the dedicated writer code in writer.F for C reliability and expandability; modified adcprep to enable it C to read mesh data in a variety of formats. C v51.21.26 - 03/05/2014 - jgf - Had to rearrange some variables between C mesh module and messenger module to avoid circular dependencies; C updated makefile so that all executables build and link C properly; updated couple2swan module to reflect the changes C (improvements) in the write_output module. C v51.21.25 - 02/24/2014 - jgf - Made corrections to nodal attribute C conversion utilities xdmf2adcirc and adcirc2xdmf; completed C merge of xdmf nodal attribute reading routines into adcirc. C v51.21.24 - 02/23/2014 - jgf - Significant enhancements to write_output.F, C including promoting it to a module; completed merge of XDMF C output capabilities into write_output.F; merge of nodal C attributes XDMF capabilities into nodalattr.F in process. C v51.21.23 - 02/20/2014 - jgf - Placed mesh and boundaries modules in their C own source files and updated makefile accordingly. Builds and C runs successfully. C v51.21.22 - 02/20/2014 - jgf - ADCIRC successfully reads and runs with C an XDMF formatted mesh as well as a conventional ADCIRC ascii C mesh. C v51.21.21 - 02/19/2014 - jgf - All adcirc2xdmf and xdmf2adcirc tests pass. C v51.21.20 - 02/19/2014 - jgf - Added spatial grid collection for visualizing C levee boundaries. C v51.21.19 - 02/17/2014 - jgf - Refactored adcmesh.f90 code so that memory C allocation for mesh data and boundaries is performed using a C single set of subroutines, regardless of the file format of C the mesh data (ascii, xdmf or other). C v51.21.18 - 02/16/2014 - jgf - The xdmf2adcirc utility now successfully reads C the attributes of internal barrier boundaries and internal C barrier boundaries with cross barrier pipes correctly. C v51.21.17 - 02/14/2014 - jgf - The xdmf2adcirc utility now writes the mesh C and boundaries to the ascii adcirc fort.14 file correctly. C v51.21.16 - 02/12/2014 - jgf - Fixed data types for writing boundary data to C XDMF; added code for reading mesh and boundaries from XDMF for C use in xdmf2adcirc utility as well as reading data into ADCIRC C and adcprep. C v51.21.15 - 02/07/2014 - jgf - Additional progress on xdmf2adcirc utility; C added makefile for utility compilation. C v51.21.14 - 02/06/2014 - jgf - Added control.f90 for reading and writing C control file (fort.15) as metadata (or for any other purpose); C added code for reading and writing nodal attribute data to/from C xdmf; added beginnings of code to read XDMF data for C xdmf2adcirc.f90 utility. C v51.21.13 - 01/27/2014 - jgf - Tested and subsequently slightly rearranged C memory allocation and the initialization of the node table, C element table, and boundaries for self consistency; verified C that the refactored code runs properly and gives the right C answers in serial. C v51.21.12 - 01/24/2014 - jgf - Propagated the changes in the mesh and C boundary modules throughout the code; checked to make sure C that the code compiles and links successfully. C v51.21.11 - 01/23/2014 - jgf - Refactored all remaining mesh related C and boundary related data and subroutines from the global C module into the new mesh and boundary modules; wrote a new C subroutine for readin ascii fort.14 file sans initialization; C created stub subroutine for reading mesh file in XDMF format; C added support for specifying the name and file format of C and ADCIRC mesh file through command line options; moved C mesh-related and boundary related initialization to separate C subroutine that can be used regardless of the original file C format of the mesh data. This refactoring paves the way for C expansion and diversification in the way ADCIRC reads mesh C data, e.g., it is now much easier to support the reading C of mesh files in NetCDF format, or the reading of files C originally developed for use with other unstructured coastal C ocean models. C v51.21.10 - 01/03/2014 - jgf - Refactored the arrays nneigh, neitab, C neitabele, mju, nodele, areas, totalarea, nodes_lg, and C imap_el_lg out of the global module and into the new mesh C module along with the variables slam0 and sfea0 and the C subroutines allocmain1b, neighb, and invcp; created a new C subroutine totalAreaCalc() to encapsulate the calculation C of total area around each node from code previously located C in adcirc_init(); added code for new boundaries module. C v51.21.09 - 01/02/2014 - jgf - Refactored the variables ne, np, ne_g, C np_g, nm, x, y, slam, sfea, dp, and ics as well as the C associated memory allocation and the cpp subroutine into C the new mesh module. C v51.21.08 - 12/26/2013 - jgf - Added new mesh module, currently placed C in global.F source file, for holding data and subroutines C related to ADCIRC mesh. C v51.21.07 - 12/22/2013 - jgf - Finished testing code to write control C metadata to XDMF xml file. C v51.21.06 - 12/21/2013 - jgf - Finished code for writing fort.15 control C metadata to XDMF xml file. C v51.21.05 - 12/19/2013 - jgf - Finished testing and fixing code for C reading ascii fort.15 for use in construction of metadata. C v51.21.04 - 12/18/2013 - jgf - Finished code for reading ascii fort.15 C for use in xml metadata; added code for writing ascii fort.15 C for use in troubleshooting as well as conversion from xdmf C to adcirc ascii format. C v51.21.03 - 12/13/2013 - jgf - Added variables and initial code to read C fort.15 file for use in creating metadata for xml file. C v51.21.02 - 12/13/2013 - jgf - Added grid collection to hold a time series C of output data; refactored the storage of boundaries to its own C subroutine; added code to save variable-specific metadata to C xml files. C v51.21.01 - 12/11/2013 - jgf - Added adcirc2xdmf.f90 code to util subdir. C C v51.33.14 - 12/12/2014 - jgf - Fixed duplicate execution tracing calls. C v51.33.13 - 12/09/2014 - jgf - Merged trunk changes r1032:1050; fixed some C issues related to the move of output type identifiers like C OFF, ASCII, BINARY, etc from global.F to sizes.F. C v51.33.12 - 11/18/2014 - jgf - Added comma to hstime.F to fix syntax when C not compiling with NETCDF support. C v51.33.11 - 11/12/2014 - Decided to adjust compiler flags for debugging C to remove trapping on underflow floating point exceptions, C since default compiler behavior is likely to handle such C exceptions correctly (flush to zero) during production. C v51.33.10 - 11/07/2014 - jgf - Corrected stray subroutine call from C adcirc_terminate() to terminate(). C v51.33.09 - 11/07/2014 - jgf - Added option to compile with debugging to C cmplrflags.mk for intel-lonestar; also added SWAN=enable C option on the make command line for same (instead of C hardcoding -DADCSWAN into the DPRE flags); added dependency C of wind module on mesh module into makefile; fixed misplaced C END SELECT in couple2swan that was erroneously inside #ifdef C causing compilation issue if compiled without netcdf support; C fixes to enable compilation with debugging using intel11 C compilers on lonestar at TACC. C v51.33.08 - 11/06/2014 - jgf - Fixes to enable compilation in debug mode C with ftn compilers on garnet at ERDC. C v51.33.07 - 11/06/2014 - jgf - Fixes to enable compilation with ftn C compilers on garnet at ERDC; added execution tracing to C write_output module. C v51.33.06 - 10/31/2014 - jgf - Fixes to enable compilation with intel14 C compilers on hatteras at RENCI. C v51.33.05 - 10/30/2014 - jgf - Fixed issue with adcprep: mesh coordinates C were not suitably initialized, causing floating point exception C problems in the setup_kdt_search subroutine; eliminated C redundant specification of deg2rad, rad2deg, and Rearth C parameters as well as CPP projection sub between adcprep and C adcirc, including CPP projection of station coordinates; C removed INVCPP subroutine from adcprep since it was not C called from anywhere. C v51.33.04 - 10/26/2014 - jgf - Updated this file by adding entries for C each commit that has been made on this branch; merged trunk C changes that occurred from r1014 through r1031; minor fixes C to USE statements in aswip.F to allow the aswip executable C to compile; tested and confirmed that all executables C compile successfully. C v51.33.03 - 08/25/2014 - gjie - Fixed arrays. C v51.33.02 - 08/01/2014 - gjie - Added geofactor to control the geostrophic C balance. C v51.33.01 - 07/21/2014 - gjie - Added GAHM vortex wind model with NWS=20. C v51.33 - 07/16/2014 - tcm - Updated compiler flags for diamond, C to replace no longer supported options -xT with C -xSSSE3. While diamond has been removed from service, C another SGI Spirit is still in service. C v51.32 - 07/16/2014 - tcm - Added a feature that will kill all MPI C processes when an elevation greater than error elevation is C encountered during a coupled ADCIRC+STWAVE simulation, C NRS=4. This is needed to stop all MPI processes, not just C the ones that ADCIRC owns. C v51.31 - 07/07/2014 - tcm - Fixed a bug in prepping fort.141 files C that was caused by only adding 105 to the processor number C to create a unit number, since 141 would be used for both C a local unit number and a global unit number. Now addes 505 C to the processor number, so that local PE####/fort.141 will C have unit numbers from 505 to 760. C v51.30 - 07/02/2014 - tcm - Removed the conditional pre-processor C flags related to max/min file generation (-DNELMAX -DNUVMAX C -DNPRMAX -DNWVMAX -DNRSMAX). Added initialization of C max/min variables to handle an issue with a compiler that C didn't initalize them to zero. C v51.29 - 06/30/2014 - jgf - Updated compiler flags for lonestar C machine at TACC. C v51.28 - 05/12/2014 - jcd -- Merging swan4101 branch into trunk. C v51.27.02 - 05/11/2014 - jcd -- Removing our old hack to limit infinite C sweeping on some meshes. SWAN no longer needs it. C v51.27.01 - 05/07/2014 - jcd -- Added the latest SWAN release version 41.01 C to the ADCIRC code. Initial testing confirms it gives C identical results to the previous version. C v51.27 - 05/02/2014 - tcm -- FUNCTIONALITY CHANGE ALERT C Removed the use of fort.88 files for initializing river C elevations at the time of cold start. Initial river C elevations are now specified as a nodal attribute in the C fort.13 file with the attribute name of C initial_river_elevation. In addition to removing the fort.88 C file from the adcirc and adcprep stages, how initial river C elevations values are set has also changed. Previously if C initial river elevation values and GeoidOffSet values were C both used, then the larger of the two values was used provided C initial river elevations were greater than zero. This has C now changed so that initial river elevations and GeoidOffSet C values are now added together (provided initial river elevations C are greater than zero, otherwise only the GeoidOffSet value C is used.). C v51.26 - 03/18/2014 - tcm -- Added NHSTAR=-1 option which generates C individual binary hot-start files named fort.68_######### C where ######### is the time step iteration pre-padded with 0. C Also, corrected some bugs in the hot start writer through C hs-writers related to fort.16 and screen output information. C v51.25 - 03/14/2014 - tcm -- Merging the v51.20.01-07 changes C into the trunk. C v51.20.07 - 03/07/2014 - tcm - Adding trunk updates to this branch C prior to reintegration. C v51.20.06 - 03/07/2014 - tcm - Finished additions for max/min records C where the time the max/min occurs is also recorded. The C max/min files now have two "time records" instead of one. C The first "time record" records the actual max/min value C and the second "time record" records the time in seconds C when the max/min occurred. The NetCDF files for the max/min C for ADCIRC information has also been updated and includes a C new set of variables that prefix to the original names with C time_of_oldvariable name. NetCDF files are still not C used to initialize the max/min values at hotstart which C was a capability to start with. However, if the ascii max/min C files are present then the max/min values and their associated C times will be initialized from the file. If a legacy max/min C file is encountered without the times data, then the times C associated with the max/min is set to -99999.d0 to flag C unknown. C v51.20.05 - 02/28/2014 - tcm -Finished adding the use of external files C for specifying recording stations locations. Also added C some debugging checks for reading fort.18 files during C parallel execution. C v51.20.04 - 02/28/2014 - tcm - Adding the use of external files for C specifying recording stations into the ADCIRC/PADCIRC C codes. C v51.20.03 - 02/27/2014 - tcm - Adding the use of external files for C specifying recording station (elev, vel, conc, met) locations C Updated the adcprep routines at this time. Triggered by a C negative value for (NSTAE,NSTAV,NSTAC, or NSTAM) in the C fort.15 file. External station files are named: C elev_stat.151, vel_stat.151, conc_stat.151, met_stat.151. C The first line of this file contains the number of stations C the remaining lines of the file are the same format as the C stations are listed in the fort.15 now. Not that the actual C value of the NSTAE etc in the fort.15 is reset if the value C in the external file is different. C v51.20.02 - 01/24/2014 - tcm - Continued adding capability to track the C time step number that a max/min value occurs at. Updated C how max/min files are read at hstart so that now every proc C has the max/min value at hot start time. Write the values C and the time stamp in the max/min files now. Still not C finished. C v51.20.01 - 11/19/2013 - tcm - Begin adding capability to track the C time step number that a max/min value occurs at. C v51.24 - 01/31/2014 - tcm - Fixed bug in processing fort.88 file in C prep.F by using a maximum of 256 subdomains open at a time. C v51.23 - 12/16/2013 - jgf - Moved CPP subroutine from adcirc module to C global module, making it available to the get_gfdl subroutine C thereby fixing an issue with compilation of aswip. C v51.22 - 12/16/2013 - jgf - Added compiler flags for compiling padcswan C on garnet at ERDC. C v51.21 - 12/04/2013 - jgf - Fixed incorrect station coordinates C in all station netcdf files, 2D and 3D, serial and parallel, C where Cartesian coordinates were erroneously being written C to station netcdf files, coordinates are now correctly C written in degrees latitude and longitude. C v51.20 - 11/19/2013 - tcm - Merging changes from the trunk into this C branch ahead of re-integration. C v51.06.10 - 08/01/2013 - tcm - Added insertion of blank wind/pressure C snaps when wind files start after the beginning of the C simulation or not enough wind files to complete the C simulation for NWS=16 GFDL wind file formats. C v51.06.09 - 05/09/2013 - tcm - Corrected bug in hstart.F where C fort.75 files were errornously being opened. C v51.06.08 - 05/08/2013 - tcm - Added setMessageSource and C unsetMessageSource to writer.F and wind.F for GFDL. C v51.06.07 - 05/08/2013 - tcm - Changed GFDL ASCII for read, this format C is still in flux. C v51.06.06 - 04/01/2013 - tcm - Fixed a bug in GFDL interpolation. C v51.06.05 - 04/01/2013 - tcm - Updated GFDL interpolation to perform a C check to make sure points aren't too far away from the grid. C If GFDL_max_extrap_dist = -99999.d0 in the fort.22 file, then C an estimated maximum distance is computed to be 6*grid cell C spacing in the GFDL grid. C v51.06.04 - 02/25/2013 - tcm - Added pressure conversion from mb to C MH20 for nws = 16, GFDL wind/pressure C v51.06.03 - 02/21/2013 - tcm - Fixed bugs in nws=16 option. C v51.06.02 - 02/20/2013 - tcm - Added nws=+-16 (GFDL Met Data) C v51.06.01 - 02/08/2013 - tcm - Added a subroutine for reading and C interpolating GFDL winds/pressures onto ADCIRC nodes. C v51.19 - 11/12/2013 - pck - Update CF-UGRID Compliance to v.0.9.0. Also C decreased lower limit from 0.03 to 0.02 for Manning's n seen by C SWAN (per Hope et al. 2013 validation of Hurricane Ike). C v51.18 - 11/11/2013 - jgf - Fixed f90 compilation rule and arguments to C switch.pl (replacing -unix -f95 with swch) for compiling adcswan C and padcswan targets. C v51.17 - 11/01/2013 - tcm - Corrected a bug in adcprep for nws=2 where C wtiminc was not being correctly read from the fort.15 file. C v51.16 - 10/11/2013 - tcm - Enlarged the integer format width for C node numbers being written on the run status updates in C timestep.F that report maximum elevation and maximum C velocity for the grid for that time step. Width increased C from I7 to I8. C v51.15 - 09/10/2013 - jgf, tcm, jcd - Fixed an issue caused by C boundaries that are only two nodes long, causing C a floating point exception in read_input.F when dividing by C the distance between nodes (division by zero). C v51.14 - 07/25/2013 - jgf - Corrected a bug in NWS8 (symmetric holland) C where the background pressure was being erroneously applied in C mb whereas NWS8 requires Pa; fixed met-only mode by forcing C a recalculation of starting and ending timesteps for output C whenever the ADCIRC timestep has changed after a hotstart; C fixed bug that was causing NWS8 to run out of met data one C met time step too early. C v51.13 - 05/30/2013 - jgf - Updated compiler flags for compiling with C netcdf support on Diamond at ERDC. C v51.12 - 05/14/2013 - jgf - Corrected 3D netcdf hotstart issue in C parallel where typos in netcdfio.F were causing density3D to be C read erroneously in place of velocity3D and turbulence3D data. C v51.11 - 05/09/2013 - tcm - Corrected bug in hstart.F where C fort.75 files were erroneously being opened. C v51.10 - 05/06/2013 - jgf - Moved the logical variable writerReadMetaData C from global.F to netcdfio.F to avoid compilation issue on IBM C Power7 (AIX 6.1). C v51.09 - 05/06/2013 - jgf - Updated developers documentation to include C instructions for building ADCIRC (no code changes). C v51.09 - 04/22/2013 - tcm - Added message to fort.16 file on the number C of hotstart writers. Also, fixed bug were "-H" on the command C line first showed up as an invalid option during the first C parsing of the command line. C v51.08 - 03/05/2013 - jcd - Small change to the SWAN Makefile to include C the kdtree2 module, in case it is needed by some compilers. C v51.07 - 02/25/2013 - zc - Correcting loading of ESL on hot start C in serial. Corrected SwanMain to have one call to SWMAN C commented out (using NADC) if compiling with ADCIRC so C subroutine call not detected as missing arguments by some C compilers. Without ADCIRC, this is left uncommented. C v51.06 - 02/06/2013 - zc - Ensuring that SWAN is sent blended winds for C NWS=+/-29. Previously it only saw vortex generated winds. C v51.05 - 02/03/2013 - zc - Corrected variable FoundEye not being set in C NWS15GET, so wind drag always defaulted to Garratt formulation. C v51.04 - 01/16/2013 - jgf - Fixed issue found by JieGao in initialization C of quadrant angles that affected Vr calculations and caused C differences between calculated quadrant wind speeds and isotach C values from the track file. C v51.03 - 01/10/2013 - zc - Correcting bug that would cause error if C elemental_slope_limiter was not last attribute in fort.13. C Added check for FoundEleSlopeLim and LoadEleSlopeLim as well. C v51.02 - 10/23/2012 - zc - Reversing change from 9/28/2012. This is C actually a leftover debugging statement in NetCDF and is C corrected in versions > 4.2. Also, this seemed to cause C segfaults under some compilers C v51.01 - 10/19/2012 - jcd - Added a WaveWindMultiplier to allow the user C to adjust the wind speeds before they are passed to the wave model. C The conversion factor has a default value of 1.0, but it can be C adjusted in the fort.15 file with a command like: C &waveCoupling WaveWindMultiplier=1.09 / C v51.00 - 10/19/2012 - jgf - Extended the code for a hotstarting a run with C a time step size that is different from the time step size C in the hot start file to all run types (not just 3D baroclinic); C also cleaned up some stray debug statements. C v50.99 - 10/08/2012 - jgf - Added if/then statements so that minpr.63 C and maxwvel.63 won't be written if there was no wind forcing; C also, so that maxrs.63 won't be written unless there was C wave forcing. C v50.98 - 10/04/2012 - jgf - Added subroutine to write LHS of harmonic C analysis matrix for debugging purposes; fixed issue in the C way that the LHS matrix is read from the binary hotstart C file; fixed issue with the way that harmonic analysis C load vectors are written to the netcdf hotstart file; fixed C an issue with the on-the-fly decomposition of harmonic C analysis load vectors from netcdf hotstart files; fixed C issue with reading harmonic analysis load vectors for C station velocity harmonics from binary hotstart file. C v50.97 - 09/28/2012 - zc - Corrected how the element array is written C to NetCDF files to prevent unnecessary screen output for some C builds of NetCDF. Also made sure negative values are treated C correctly in couple2swan max file writing. C v50.96 - 09/27/2012 - jgf - Removed stray debugging messages from C netcdfio.F that were causing segfault upon hotstart with netcdf C files. C v50.95 - 09/27/2012 - zc - Added SwanMaxDescript%Specifier = NOUTGW to C ensure NetCDF format is written for SWAN max files. C Corrected SWAN output routines to always use writer processors C if MNWPROC > 0. Behavior is consistent with write_output.F C v50.94 - 09/24/2012 - jgf - Added DA compiler flags (-DREAL8 etc) to C compile hstime, aswip, and adccmp; fixed read format issues in C nws15get; fixed cycle time dimension issue, globaldir C initialization issue, and lat/lon initialization in aswip C (these issues only affect post storm analysis for hwind). C v50.93 - 09/24/2012 - zc - Fixed a bug when both writing and reading C netcdf hot start files where timenc was deallocated but not C marked for reinitialization later. C v50.92 - 09/24/2012 - jcd - Fixed a bug in the NetCDF output of the sigma C layers in 3D. The variable NFEN was declared in pre_global.F, C and thus was not available to the FEGRIDS routine within the C global_3dvs.F module. Changed pre_global.F to use the NFEN C declared within the global_3dvs.F module. C v50.91 - 09/21/2012 - jcd - Merged the swan40.91 branch into the trunk. C v50.90 - 09/18/2012 - tcm Fixed a typo bug in adcprep.F for the windows C builds. C v50.89 - 09/16/20120 - zc - Merging netcdf_writer branch into trunk, C revisions 808:836 and resolved conflicts. Ensured that C rads.64.nc could be created without CSWAN flag. C v50.74.06 09/12/2012 - zc - Corrected SWAN output for max wind C velocities in netcdf to scalar format. C v50.74.05 09/12/2012 - zc - Corrected a bug in the SWAN output C when not using NetCDF writers. Allow the writer C processors to read metadata from ADCPREP created C output files during first pass since they do not read C fort.15. Tested for NETCDF3 and NETCDF4. Also, corrected C SWAN max files in NetCDF which were still writing in C ASCII. C v50.74.04 07/01/2012 - zc - Removing unused variables added for C original implementation of NetCDF writer cores C v50.74.03 06/29/2012 - zc - Remove now unused MPI_COMM for C inter-writer only communication. Adjust makefile C for correct dependencies for netcdfio.F and writer.F C v50.74.02 06/29/2012 - zc - NetCDF writers seem to be working C properly for ADCIRC and SWAN. Continued testing needed C before merge to main trunk. C v50.74.01 06/20/2012 - zc - Initial edits for NetCDF writer cores. C Writer cores are recognized under ADCIRC writes only C at this point. Need to ensure that writer cores have C knowledge of all NetCDF record numbers. C v50.88 - 09/14/2012 - zc - Corrected a single processor arriving at C LBCAST in couple2swan.F. Made a similar change for C consistency in write_output.F, but using ADCIRC_TERMINATE(). C v50.87 - 09/14/2012 - zc - Corrected the size of the netcdf_error array C in couple2swan.F to prevent message passing errors. C v50.86 - 09/12/2012 - tcm - Adding an empty directory "windows" where C MS Visual Studio projects for building metis/adcprep/adcirc/ C and padcirc can be stored. C v50.85 - 09/11/2012 - tcm - Added a few changes related to compiling C on Windows PC using MS Studio and Intel MPI. C v50.84.05 - 09/11/2012 - jcd - Corrected the use of MPI_COMM_WORLD C when it should instead have been COMM in SwanPunCollect. C v50.84.04 - 09/09/2012 - jcd - Two small changes to the existing files C for src/couple2swan.F and work/makefile. C v50.84.03 - 09/09/2012 - jcd - Added the newest release version 40.91 C in the swan sub-directory. Carried forward our macros.inc C files with the compiler flags for different machines. C v50.84.02 - 09/09/2012 - jcd - Trying to delete the existing SWAN. C v50.84.01 - 09/09/2012 - jcd - Fixed a small bug in write_output.F with C the writing of the new NetCDF output files. C v50.84 - 08/16/2012 - jgf - Fixed aswip compilation dependencies to enable C the building of aswip on garnet at ERDC. C v50.83 - 08/16/2012 - jgf - Fixed (removed or changed) stray references C to screen() subroutine in ADCIRC code. C v50.82 - 08/15/2012 - jgf - Adjusted routines related to a single C subdomain initiating code exit: if a subdomain experiences C a fatal error, it calls adcirc_terminate() as usual, and C sets a flag that causes mpi_abort() to be called in msg_fini() C (after writer processors have been shut down); also added C execution tracing to messenger.F and fixed a conflict between C ADCIRC and SWAN by renaming ADCIRC's screen() subroutine to C screenMessage(). C v50.81 - 08/01/2012 - zc - Added routines so a single processor can C initiate code exit. ADCIRC_LOCALTERMINATE and C ADCIRC_CHECKLOCALTERMINATE. All processors use the check routine C and the processor with the error uses LOCALTERMINATE and waits C until the others catch up. Added this for error checking in C read_input.F where mesh or fort.15 errors that apply to a single C processor only will hang the code since other processors have C no knowledge of the error and hang. Readily extendable to C other areas of the code as needed. C v50.80 - 07/31/2012 - jgf - Added underscore to mesh topology attribute C in netcdf files for NcML compliance. C v50.79 - 07/20/2012 - tcm - There was a bug in the prep.F routine in the C writing of the metControl namelist that was causing it to not C be written correctly with default values. The writing of this C namelist was corrected to remove the bugs. Then logic was C added so that the metControl namelist is only written in the C partitioned fort.15 files if it was present in the global C fort.15. Changed some float() to real(,kind=sz) to remove C warning messages. C v50.78 - 07/20/2012 - tcm - Changed real*8 to real(8) to remove C warning messages. C v50.77 - 07/20/2012 - tcm - Fixed a bug in prep.F where the default C filename was not initialized to blanks for certain compilers C and therefore could have extra characters that were not a part C of the acutal name, for example a file name might have been C fort.13(. This error happened when default name lengths C changed from 7 to 8 characters to allow for larger unit C numbers. Added the Utility Server (utils) compiler options C back to the cmplrflags.mk file. C v50.76 - 07/19/2012 - jgf - Fixed missing parenthesis on line C 7062 in prepNetCDF() where the code was checking if C NRS eq 3 or 4; this issue was preventing compilation. C v50.75 - 07/16/2012 - tcm - Implemented Casey's writing of rads.64 file C whenever nrs = 3 or nrs = 4. Removed the ifdef cswan and in C most cases replaced with testing of nrs values. C v50.74 - 06/11/2012 - jgf - Merged f90_netcdf branch revisions 764:806 C and resolved conflicts. Fixed two previously undetected C minor issues: (a) SL1 and SF1 arrays were not populated in C read_global.F for ICS=1, but they are required for netcdf; and C (b) subdomain to fulldomain station mappings were using wrong C dimensions during decomposition of harmonic means and variances C when hotstarting from a netcdf hotstart file. C v50.60.21 05/24/2012 - jgf - Fixed a floating point exception when C writing 3D velocity stations in parallel. C v50.60.20 05/24/2012 - jgf - Fixed two issues with the new metControl C and SWANOutputControl namelists so that (a) they are not C required in the fort.15, and (b) so that the metControl C namelist is written with the draglawstring properly enclosed C in single quotes by adcprep upon write to subdomain fort.15s C so that padcirc can read it successfully. C v50.60.19 05/22/2012 - casey - Corrected a bug with the output of the C SWAN max files. C v50.60.18 05/22/2012 - jgf - Fixed dependence of presizes.F on wind C module; made changes to support gfortran without underscoring; C fixed issue with 3D bottom roughness lengths, where Z0B C was declared in both nodalattr.F and global_3dvs.F, resulting C in Z0B being used uninitialized in the apply3DBottomFriction() C subroutine; fixed numerical issue in wetting and drying C due to DELETA sometimes being negative; fixed issue so that C maxrs.63 file is still produced, even without compiling in C SWAN coupling (purely to match past behavior); replaced some C stray f77-style subroutines in netcdf module; fixed C numerical issue with translation speed calculation in C symmetric vortex (NWS8) subroutine. C v50.60.17 05/17/2012 - corbitt - Fixed the vector max netcdf output so C it was outputing scalar magnitudes instead of vectors. C v50.60.16 05/17/2012 - jgf - Added tracing to writer.F and C couple2swan.F; removed references to netcdff library from C makefile in favor of adding -lnetcdff to FLIBS in C cmplrflags.mk for platforms that require it; fixed C array out-of-bounds in netcdf_error array by increasing C array length to 14. C v50.60.15 05/17/2012 - corbitt - Added netcdf output for maxele.63, C maxvel.63, minpr.63, maxrs.63, maxwvel.63. Format is dependent C on NOUTGE, NOUTGV, NOUTGW, so max files will be formatted same C as fort.63,64,73,and 74. If NOUTGE, NOUTGV, NOUTGW are set C to 0, standard ascii is output. Population of maxes from C hotstarts are still dependent on ascii format. In addition, C made the 2D baroclinic and swan netcdf output CF compliant. C Also preliminary set up netcdf max files for swan output. C v50.60.14 05/15/2012 - jgf - Fixed typo that was causing array C out-of-bounds when writing fort.45 in netcdf format. C v50.60.13 05/15/2012 - jgf - Moved metControl namelist for specification C of air density, drag law, and wind drag limit to end of C fort.15, instead of reconstructing the NWS line; this C brings this data structure into line with the C TimeBathyControl and SWANOutputControl namelists. Also fixed C an issue in checking NWS values that was rejecting negative C values of NWS. C v50.60.12 05/14/2012 - jgf - Fixed issue with compiler flags on C blueridge at RENCI to use the path to netcdf4.1 (required C when using NETCDF4=enable). C v50.60.11 05/14/2012 - jgf - Fixed issue in makefile to resolve C reference to global.F in presizes.F. C v50.60.10 05/14/2012 - jgf - Added code to prep.F to calculate and C save sigma coordinates to 3D netcdf file; moved FEGRIDS C subroutine from vsmy.F to global_3dvs.F so that it could C be used by adcprep; added an IF statement in couple2swan.F C so that only processor 0 would make the subroutine calls C to initialize the SWAN-related netcdf output files; also C changed SWAN_VOQR(IVTYPE+1) to SWAN_VOQR(IVTYPE)+1 in C couple2swan.F to remedy an array-out-of-bounds issue C (this appeared to be a typo). C v50.60.09 05/12/2012 - jgf - Added autocompression for netcdf4 C formatted SWAN output files; fixed issue that was causing C station coordinates to be reported in CPP rather than C geographic projection in netcdf output files; added optional C namelist to bottom of fort.15 for reading zc's logical C variables for turning SWAN output on and off; fixed an C issue in adcprep that was trying to open the SWAN command C file twice if the SWAN command file was named fort.26; C removed assumptions in symmetric vortex model (nws=8) about C northern and western hemispheres so it can now simulate C storms in eastern and southern hemispheres; fixed file C reading issue in symmetric vortex model where it was C trying to read subdomain fort.22 files; removed stop C keyword from symmetric vortex subroutine that could C cause problems in parallel operation. C v50.60.08 05/10/2012 - zc - Added netcdf output to SWAN output files C and rads.64. The swan output files are now turned on and off C using logicals in the global.F file. Output files are set C to the defaults that were previously hardwired. C v50.60.07 05/06/2012 - jgf - Corrected issue with writing 3D fulldomain C velocity files in netcdf in serial; added storage of sigma C array to 3D netcdf files. C v50.60.06 05/05/2012 - jgf - Fixed issues in netcdfio.F related to C setting vertical dimension when initializing and writing 3D C station and 3D fulldomain output files in netcdf format. C v50.60.05 05/04/2012 - jgf - Simplified structure of netcdfio.F C to reduce complexity and take advantage of commonalities C in the i/o of the various station and full domain files. C Tested netcdf output in serial and in parallel, including C serial and parallel hotstarting with netcdf hotstart files C on 2D quarter annulus case. All smoke tests pass at this C point; a few more tests are required. C v50.60.04 05/03/2012 - jgf - Eliminated memory allocation and copying of C 3D data and 3D hotstart data in netcdfio.F that is no longer C required when using f90-style subroutine calls. Added code C that pulls 3D data from netcdf hotstart files directly into C adcirc variables in serial and in parallel. This code C compiles; next step is testing. C v50.60.03 05/02/2012 - jgf - Fixed an incomplete implementation of C hotstarting in parallel with netcdf that was uncovered C in testing previously committed changes. Applied logging C fix from trunk (v50.65). C v50.60.02 05/01/2012 - jgf - Migration of adcirc netcdfio module C to f90-style subroutine calls is complete. It now compiles too. C The next step is to make corresponding changes to C write_output.F. C v50.60.01 04/29/2012 - jgf - Halfway through changing the subroutine C calls in netcdfio.F to f90-style. This is a work in progress. C v50.73 - 05/08/2012 - tcm - Removed use of reserved work (scale) from the C file diffmerge.F. C v50.72 - 05/07/2012 - tcm - Incorporated time varying bathymetry changes C into ADCIRC. These came from the time_varying_bathy branch. C If time varying bathymetry is not used, then no changes to the C fort.15 occur. To use time varying bathymetry, then add the C "timebathycontrol" namelist at the end of the fort.15 file. C Time varying bathymetry uses a fort.141 file to supply the C different bathymetry values and as output writes a fort.75 C (station) and fort.76 (global) that are controled by the C fort.61 and fort.63 file options. C v50.71 - 05/01/2012 - zc - Reorganized ps2dots and ps3dots routines C to improve code performance. Changed ADVECTLOCAL to only C be called when advection is turned on, which improves code C speed. Moved call to updater to after slope limiter C routine to avoid calling it twice. C v50.70 - 04/22/2012 - zc - Allowed the values from ESLNodes.63 to C be used in a hotstart run to reactivate slope limiter at C previously active nodes. Fix hot start allocation issues. C v50.69 - 04/16/2012 - jgf - Applied fix from Kendra Dresback for C erroneous format in writeOutHarmonicArrays() subroutine C that was causing phase values to be off by a factor of 10. C v50.68 - 04/15/2012 - jgf - Removed water surface elevation smoother C nodal attribute. C v50.67 - 04/15/2012 - zc - Corrected the elemental slope limiter to C apply smoothing if fort.13 max gradient is set to zero C instead of skipping the node. Also set so that a negative C fort.13 value triggers gradient warnings but no slope C limiting C v50.66 - 04/09/2012 - jgf - Corrected issue in reading NWS namelist C that caused a compiler warning about out-of-block goto. C v50.65 - 04/09/2012 - jgf - Corrected issue with logging where all C subdomains were logging to an erroneous fulldomain fort.16. C v50.64 - 04/08/2012 - jcd - Updated a few SWAN source files to reflect C the latest changes by the SWAN developers. C v50.63 - 04/07/2012 - zc - corrected the elemental slope limiter C output statements so they only appear once per node instead of C continuously until the gradient is below the threshold C v50.62 - 04/07/2012 - zc - corrected my previous update to use -99999 C when a node is dry instead of what resides in eta2 C v50.61 - 04/05/2012 - zc - fixed a bug when writing the sparse format C in the fort.69 debug file. Switched to full format for C simplicity. C v50.60 - 04/05/2012 - jgf - Merged newvortex branch revisions 705 to 760 C into the trunk (the newvortex branch was originally created C from v50.38 of the ADCIRC trunk). This merge contains the new C NWS19 asymmetric vortex model, new aswip met preprocessing and C analysis code, and new NWS15 hwind input capability. Detailed C changelog is shown below (v50.38.01 through v50.38.09): C v50.38.09 04/05/2012 - jgf - Removed extraneous use of globaldir from C aswip code. C v50.38.08 04/05/2012 - jgf - Corrected calculation of wind magnitude C for nws15 (hwind) in timestep.F; changed log level to echo C for logging drag law string from nws line in fort.15. C v50.38.07 03/25/2012 - jgf - Changed windMultiplier variable to C hWindMultiplier to avoid variable conflict between hwind and C owiwind modules. C v50.38.06 03/10/2012 - jgf - Added vmax and rmax vars to hWindData_t C block; added support for Powell wind drag to NWS15 (HWind); C added support to NWS15 for finding vmax and rmax; reorganized C aswip code to make it modular; added capabilities to the C wind analysis features of aswip to enable analysis of HWind C data via ADCIRC's new HWind support. C v50.38.05 03/05/2012 - jgf - Created NWS=15 for reading HWind data, C including nws15init, nws15get, loadHWindFile, and C hWindInterpolateVelocity subroutines, as well as supporting C changes in adcirc and in various routines in adcprep as needed; C code has been tested and appears to work correctly thus far. C v50.38.04 01/24/2012 - jgf - Removed dependency on precision.F, constants.F C fitpack.F, and lsq.F from code; removed NWS9; fixed issue C with percent Rmax specification in aswip. C v50.38.03 01/16/2012 - jgf - Changes to simplify the aswip code and C integrate it more closely with ADCIRC; made air density a C variable that can be specified by the user in the fort.15; C made the default value of the air density 1.15. C v50.38.02 01/11/2012 - jgf - Latest changes to account for storm motion C on Vmax and isotach wind speeds for NWS19; addition of code to C aswip for reading/writing additional data for visualizing C wind output on autogenerated unstructured mesh; addition of C nearest neighbor weighted interpolation to vortex.F for C interpolating Rmax between quadrants; some further simplification C of vortex.F. C v50.38.01 12/02/2011 - jgf - Created branch to hold ongoing changes that C have been made to assess the meteorological fields generated C by the asymmetric vortex model (NWS19), including the following: C very heavy modification and simplification of vortex.F, C reducing it from 2200 lines of code to 800 lines of code; C addition of code to aswip to support the generation of radial C and spatial profiles of wind speed; and modification of the C interface between NWS9/NWS19 and vortex.F. C v50.59 - 04/04/2012 - jgf - Added preprocessor directives so that C netcdfio.F can be compiled with either netcdf3 libraries or C netcdf4 libraries; made adcprep read the flux time increment C (FTIMINC) in fort.20 as a real instead of an integer; corrected C the swan version numbers in the changelog below for version C 50.55. C v50.58 - 04/02/2012 - jcd - Two changes to adcprep: (1) to avoid an C endless loop when prepall does not find a default file, C now it will ask the user for the file name or to skip; and C (2) now it will allow the SWAN control file to have its default C name of INPUT, even if the swaninit file is not found. C v50.57 - 03/28/2012 - pck - Added two swan files missing from the update C v50.56 - 03/28/2012 - pck - Added another ramp option: NRAMP=8, where C the new variable DUnRampMete is the time to begin METE ramping. C Also added the nodal attribute (advection_state) that allows C advection to be initiated based on depth. If the depth is C greater than the advection_state depth, advection will be C applied (assuming that advection is turned on); otherwise C advection will be off. C v50.55 - 03/25/2012 - jgf - Merged Casey Dietrich's changes for the C latest version of SWAN (moving from v40.72 to v40.85 of SWAN) C as well as the capability to limit the spectral propagation C velocities in SWAN; the merged code also contains Taylor Asher's C new capability for modifying wave results in SWAN to account C for ice coverage. C v50.54 - 03/25/2012 - jgf - Made changes to enable compilation of netcdfio C on intel compilers; upgraded hstime utility to new ADCNETCDF C preprocessor variable; added preprocessor variables to turn off C automatic netcdf4 compression if the system netcdf4 library C does not support it (apparently support for compression was C added in netcdf v4.1, but is not available in netcdf 4.0.x). C v50.53 - 02/29/2012 - jgf - Removed massbal.F from code and makefile; C restored -i-dynamic compiler flag for intel to avoid C feupdateenv message; added swan macros.inc for g95 compiler; C turned on shuffle filter for compressed netcdf4 files. C v50.52 - 02/21/2012 - jgf - Merged the following changes from revisions C r728:r738 of the netcdf4 branch; this branch was developed to C enable compressed native netdf4 output from ADCIRC using an C output specifier of 5, i.e., setting NOUTGE=5 will cause ADCIRC C to produce fulldomain elevation output in NetCDF Classic Model, C NetCDF4-formatted, compressed output files: C v50.43.01 02/20/2011 - jgf - Changes related to netcdf4 C output files (output specifier = 5); changed NETCDF C preprocessor definition to ADCNETCDF to avoid collision C with the official NETCDF module, which was introduced C to allow access to the fortran 90 interface for netcdf; C automatically enabled compression for netcdf4 output format; C used f90 interface for saving text attributes to netcdf C files to avoid potential mismatches in character counts C in nf_put_att_text(); added tracing to vsmy.F and C timestep.F to track down memory issues; initialized C BPG to zero in cstart.F; added a convenience subroutine C ADCIRC_Terminate() to adcirc.F. C v50.51 - 02/13/2012 - jgf - Merged ADCIRC compiler flags and SWAN C macros.inc for Kraken. C v50.50 - 02/08/2012 - jgf - Changed organization of smoother and slope C limiter calculations slightly for sustainability; made negative C value of threshold gradient signify unlimited slope at a node C (setting threshold gradient to 0.0 now gives continuously limited C slope). C v50.49 - 02/08/2012 - zc - Created new nodal attribute for C elemental_slope_limiter which takes as a value the maximum C gradient to allow before applying the smoothing function which C is exactly the same as that used in water_surface_elevation_smooother. C ESLNodes.63 is written out at the end of the run showing the C nodes which triggered the function. C v50.48 - 02/08/2012 - jgf - Made stations allocate memory with MNST* C rather than NST* dimensions; moved statement that sets 1 station C as the minimum to allocate for concentration and meteorology C stations to avoid potential memory issues. C v50.47 - 02/07/2012 - jgf - Fixed tab handling issued in subroutine C parse() for station names that caused a segfault on pgi compiler. C v50.46 - 02/04/2012 - jgf - Created new nodal attribute called C water_surface_elevation_smoother and added code to GWCE_New C to apply the smoother if it has been loaded; the value of this C nodal attribute is the time step increment between applications C of the smoother; i.e., 0 represents no smoothing at that node, C 1 represents smoothing at every time step, 2 indicates smoothing C every other time step, etc. C v50.45 - 02/04/2012 - jgf - Cleaned up stray references to CGWCE_Original. C v50.44 - 02/04/2012 - jgf - Set namelist scratch file to write to localdir C to avoid collisions when running in parallel; removed the C GWCE_Original() subroutine from timestep.F per Rick's request, C fixed 3 instances in timestep.F when STOP was called without C shutting down mpi first. C v50.43 - 02/03/2012 - jgf - Changed rewind() on scratch file to open() C and close() to appease g95 compiler (for nws namelist i/o). C v50.42 - 02/02/2012 - jgf - Changed reading of NWS line so that namelist C is written and then read from a scratch file rather than C an internal character array (fixes issue caught by pgif95). C v50.41 - 12/23/2011 - tcm - Added compiler options for ERDC's Utility C server (compiler=utils) C v50.40 - 12/21/2011 - jgf - Merged the r364:365 of the debug_support C branch (v49.06.01), as follows: C v49.0601 - 12/09/09 - jgf - Added capability to produce fulldomain fort.69 C file (i.e., fulldomain elevation file on timesteps where the C warning elevation has been exceeded) in addition to existing C subdomain fort.69 files. C v50.39 12/08/2011 - tcm - Fixed a possible bug where eta2 was supposed C to be zeroed out after setting eta1 = eta2 but was not. C v50.38 - 11/28/2011 - jgf - Fixed CF compliance of time metadata in netcdf C by changing attribute text from seconds since base_date to C e.g., seconds since 2010-05-01 00:00:00 UTC (the date is C simply copied verbatim from the existing base_date metadata C field in the fort.15 file). C v50.37 - 09/23/2011 - jgf - Added extra checks to the debugging flags C for diamond in cmplrflags.mk; widened some integer format C for logging when reading fort.13, fort.14 and fort.15; C partial reorganization of aswip code including modification C to correct an issue that occasionally resulted in Rmax=0 C used as input to NWS19 subroutine in adcirc. C v50.36 - 09/13/2011 - tcm - Added tabs and comma delimiters to the parse C subroutine to handle issues with parsing station locations C in the fort.15 file. Utlitmately the station locations should C be read in as two real numbers followed by a ! and then a comment C line and the adcirc documention updated to reflect a standard C way of inserting comments. C v50.35 - 09/09/2011 - jgf - Fixed issue in adcprep when using prep13 C option: allocated and populated imap_nod_gl in relocalize(). C v50.34 - 09/01/2011 - jgf - Merged the following changes from r689:HEAD C of the nws_namelist branch: C v50.31.03 08/25/2011 - jgf - Checked for stationary storm when finding C storm direction in wind.F; added NWS value to read_input.F. C v50.31.02 08/19/2011 - jgf - Made different capitalizations of drag C law formulations valid; removed stop statement that was inserted C after reading of nws line for debugging. C v50.31.01 08/18/2011 - jgf - Reorganized reading of NWS line so that it C could be read as a namelist and still be backward compatible with C existing fort.15 files; made WindDragLimit and DragLawString C optionally selectable from the fort.15 with default value of C 0.0035 for WindDragLimit and default values of Garratt and C IceCube (depending on whether ice drag is active) for C DragLawString; added storm center recording to NWS=8, 9, and 19 C to make them compatible with Powell (sector-based) drag C formulation; made wind drag limit in SWAN the same as in ADCIRC. C v50.33 - 09/01/2011 - jgf - Fixed compiler flags for garnet at ERDC. C v50.32 - 09/01/2011 - jgf - Fixed compiler flags for diamond at ERDC. C v50.31 - 08/16/2011 - jgf - Added call to ApplyCanopyCoefficient C subroutine to all wind subroutines. C v50.30 - 08/03/2011 - jgf - Modified NWS19 so that it won't try to C interpolate beyond the end of the fort.22 file if the run C ends right at the last line of the file. C v50.29 - 08/01/2011 - jgf - Modified aswip so that if a storm becomes C stationary, its direction will be persisted and speed reduced C to 1 ... feeding 0 and 0 to ADCIRC NWS19 caused NaNs. C v50.28 - 07/26/2011 - jgf - Removed vestigial DO/ENDDO surrounding call to C SWMAIN in couple2swan, purely to enhance readability. C v50.27 - 07/12/2011 - tcm - Added the option to use dedicated writer C processors for full in addition to compact formats for ascii C global field files (fort 63,64,73,74,93,164 (cswan) ). Renamed C local variable "default" to "default_val". C v50.26 - 07/06/2011 - jcd - Changed owiwind.F so that blank snaps are C inserted when the end of file is reached. C v50.25 - 06/28/2011 - jgf - Fixed debug flags for Jade; removed -Mextend C from macros.inc for Jade; corrected typo in hstime.F. C v50.24 - 06/27/2011 - jgf - Made NWS=19 look in the fulldomain directory C for fort.22 in parallel rather than the subdomain directories. C v50.23 - 06/27/2011 - jgf - Added NWS=319 to list of allowable NWS values C in read_input.F; fixed initialization in hstime.F; fixed debug C flags for intel compiler. C v50.22 - 06/27/2011 - jcd - Set the unit number swiLUN = 26 in prep.F C v50.21 - 06/10/2011 - tcm - Changed I8 to I12 format specifications in C the fort.18 files (prep.F, messenger.F, SwanReadfort18.ftn90) C v50.20 - 06/10/2011 - jgf - Fixed an issue in prepUnSWAN that was causing C the opening of the fulldomain fort.26 file to fail. C v50.19 - 06/07/2011 - tcm - Fixed a bug in the makefile for the C library build of parallel adcirc. C v50.18 - 06/08/2011 - rjw - Added CASE(bottom_roughness_length) to prep.F C v50.17 - 06/07/2011 - jgf - Added -freal-loops to g95 compiler flags C for padcswan; changed -module to $(IMODS) in swan compilation C rules in makefile; added specific format lengths for I C to placate g95 compiler. C v50.16 - 06/06/2011 - jgf - Fixed issue with netcdf hotstart where the C time (in seconds) was being read from the netcdf hotstart file C but not communicated back to hstart.F for use in initialization; C added enhanced logging and i/o error checking to owiwind.F, as C well as creating a subroutine to remove some repeated code for C i/o; widened the format for NSPOOLG* in globalio.F from C I5 to I8 and widened format for NHSINC from I5 to just I in C read_input.F to avoid format overflows. C v50.15 - 05/31/2011 - jgf - Modified prepUnSWAN routine in adcprep so that C it loads the name of the swan command file from the swaninit C file; it will then copy that command file to the subdomain C directories without changing the name. C v50.14 - 05/27/2011 - RJW - Fixed 3D hotstart bug in the turb subroutine, C that resulted in variables not being initialized upon return to C the program from hotstart.(vsmy.F) C Explicitly allow subroutines to access variables using the "USE" C statement. This was needed to ensure MYPROC, MNPROC, SZ were C being properly used in subroutines (write_output.F,globalio.F, C global.F). C v50.13 - 05/27/2011 - jgf - Fixed issue caused by requesting netcdf C station output while specifying zero stations. C v50.12 - 05/26/2011 - tcm - Fixed bug that caused output to overflow the C formats on some compilers, for example, an output of C -0.1234567891E-123 would be written as -0.1234567891-123, when C using the Ew.d format. The FORTRAN 90 standard is not clear C on how the exp part is displayed when using Ew.d so we area C switching to the Ew.dEe format. Changes were made in C globalio.F, write_output.F and timestep.F for these. Also, C made the output formats consistent between globalio.F and C write_output.F. C v50.11 - 05/26/2011 - jgf - Added capability to adcprep to prep just C the fort.88 file; add capability to just prep the fort.20 file; C added command line argument processing to adcprep to bypass C menu driven interface; the menu driven interface is still C the default, and is triggered if there aren't any command C line options. C v50.10 - 05/25/2011 - jgf - Merged rivers50 branch r643:660, which C contained the changes for initialization of rivers above msl. C v50.05.05 05/23/2011 - jgf - Removed code that had been commented out b/c C it was no longer being used. C v50.05.04 05/13/2011 - jgf - Updated hstime utility to remove hardcoding C of the hotstart file name; added support for netcdf hotstart C files. C v50.05.03 05/13/2011 - jgf - Commented out more leftover debugging C statements. C v50.05.02 05/12/2011 - jgf - Commented out leftover debugging statements. C v50.05.01 05/11/2011 - jgf - Merged and resolved conflicts for the C following changes from the rivers_init branch: C ------------------------------------------------------------------------ C r484 | kdresback | 2010-06-28 12:19:14 -0400 (Mon, 28 Jun 2010) | 1 line C Changed paths: C M /adcirc/branches/rivers_init/prep/prep.F C M /adcirc/branches/rivers_init/prep/presizes.F C M /adcirc/branches/rivers_init/prep/read_global.F C M /adcirc/branches/rivers_init/src/hstart.F C M /adcirc/branches/rivers_init/src/read_input.F C M /adcirc/branches/rivers_init/src/timestep.F C Updates for the river initialization - adding the option for the river C boundary condition to not use time==0 as the start time. C ------------------------------------------------------------------------ C r460 | kdresback | 2010-06-07 18:29:40 -0400 (Mon, 07 Jun 2010) | 1 line C Changed paths: C M /adcirc/branches/rivers_init/prep/pre_global.F C Updates for river initialization when rivers are above MSL C ------------------------------------------------------------------------ C r454 | kdresback | 2010-06-02 15:52:24 -0400 (Wed, 02 Jun 2010) | 1 line C Changed paths: C M /adcirc/branches/rivers_init/prep/adcprep.F C M /adcirc/branches/rivers_init/prep/prep.F C M /adcirc/branches/rivers_init/prep/presizes.F C M /adcirc/branches/rivers_init/prep/read_global.F C M /adcirc/branches/rivers_init/src/cstart.F C M /adcirc/branches/rivers_init/src/global.F C M /adcirc/branches/rivers_init/src/global_3dvs.F C M /adcirc/branches/rivers_init/src/hstart.F C M /adcirc/branches/rivers_init/src/init_cond.F C M /adcirc/branches/rivers_init/src/read_input.F C M /adcirc/branches/rivers_init/src/transport.F C M /adcirc/branches/rivers_init/src/vsmy.F C Updates for river initialization when rivers are above MSL C ------------------------------------------------------------------------ C r453 | kdresback | 2010-06-02 14:10:49 -0400 (Wed, 02 Jun 2010) | 1 line C Changed paths: C A /adcirc/branches/rivers_init (from /adcirc/trunk:447) C Creating a branch for river initialization when rivers are above MSL C ------------------------------------------------------------------------ C v50.09 - 05/18/2011 - jcd - Incorporated Powell's sector-based wind drag C scheme, which is controlled by the -DPOWELL compiler flag. C Also corrected a few bugs. C v50.08 - 05/18/2011 - jgf - Changed node to rawNode in subroutine that C reads in min/max files ... not sure if this caused any issues C but the Intel compiler complained; fixed format statement for C log message reporting the value of IHOT. C v50.07 - 05/18/2011 - jgf - Added missing subroutine call to C readNetCDFHotstart3D in hstart.F; cleaned up typos in the C netcdf 3D hotstart reading subroutine. C v50.06 - 05/16/2011 - jgf - Fixed issue where 3D netcdf hotstart file C was requested but no other fulldomain 3D netcdf output was C requested (had to initialize num_v_nodes to NFEN in netcdfio.F). C Added some check_err calls to netcdfio.F where they were C previously missing. Fixed an issue with turbulence length C scale output in netcdf (nodal_data_id was referenced rather than C v_nodal_data_id). Fixed format issue with NHSTAR in read_input.F. C Initialized E0XGrad2A and E0YGrad2A in timestep.F to zero C to appease Intel compiler. C v50.05 - 05/10/2011 - tcm - added support in adcprep for compilation C on Window's based PC. Fixed a couple of stray bugs related C to character lengths in prep.F and added the NOFF variable to C hs_2_ascii.F. C v50.04 - 05/10/2011 - jgf - Fixed stray metadata in netcdf file due to C nvelnc not being defined correctly. C v50.03 - 05/09/2011 - jgf - Applied fixes to timestep.F and ipackv.F from C st to fix issue when the NeiMax node is on a boundary. C v50.02 - 05/06/2011 - tcm - Added a list of pre-processor definitions to C use in order to build Metis on a Window's based PC. C v50.01 - 05/06/2011 - tcm - Added a readme file about the changes to metis. C v50.00 - 05/06/2011 - tcm - Made a couple of modifications to the packages C called by metis to get it to work on a Windows based machine. C Added the gettimeofday.h module which is not available on a C Windows machine. A person wanting to compile metis for a Windows C based PC would select all the routines in the Lib folder and C would also include the gettimeofday.h routine in the Windows C folder and would use a -WINDOWS pre-processor flag in addition to C the other pre-processor flags in the normal makefile. C NOTE: Only use the gettimeofday.h for Windows machines, C Linux/Unix machines typically already have it installed. C These changes to metis do not change the usual building or C functionality of metis on Linux/Unix machines. C v49.82 - 04/26/2011 - jgf - Fixed initialization issue in netcdf when C netcdf hotstart file was requested in 3D but no other fulldomain C netcdf output was requested; removed warning message about C ITITER from adcprep; initialized previously uninitialized C variable in station parsing routine. C v49.81 - 04/25/2011 - rjw - Fixed bug in ./swan/Makefile now it will be able C to read the variables from global_3dv C v49.80 - 04/25/2011 - jgf - Removed stray debug msg from read_input.F. C v49.79 - 04/21/2011 - rjw - Fixed bug in ./prep/read_global.F for IEVC=51 C v49.78 - 04/21/2011 - rjw - Modified files in ./prep to allow for IEVC=51 C v49.77 - 04/17/2011 - jgf - Removed erroneous second read of WTIMINC line C for NWS19. C v49.76 - 04/14/2011 - tcm - Cleaned up more reserved words in post.F C and adcpost.F (INDEX,LOC,SCALE). C v49.75 - 04/14/2011 - tcm - Commented out debugging flags for cray_xt4, C Cleaned up us of FORTRAN RESERVED WORDS (TIME,INDEX,SCALE, C ISIGN,LOC,IIDIM). Changed to (TIME --> TIMELOC,INDEX --> INDX, C ISIGN --> ISGN, LOC --> LOC2, SCALE --> SCALEVAL, C IIDIM --> IDIM2). The use of reserved words can lead to C unintended consequences and conflicts. C v49.74 - 04/13/2011 - tcm - Fixed bug in call to FEGRIDS in read_input.F C v49.73 - 04/11/2011 - jgf - Merged vortexowi branch (r421:r536) to trunk, C bringing in the changes from v49.08.01 to v49.08.04 to create C NWS=29 option for blending vortex winds (NWS19) with C OWI winds (NWS12). C v49.0804 - 12/20/10 - jgf - Made allowance for NWS29 in read_global.F, C prep.F and adcprep.F; cstart.F now looks in GBLINPUTDIR for C the fort.22 for NWS29; corrected missing parenthesis in wind.F; C added reading of WTIMINC line for NWS29 in presizes.F; added C blend factor for blending vortex winds and OWI winds. C v49.0803 - 04/16/10 - jgf - Added code to NWS19 to properly export the C vortex center coordinates and current Rmax. C v49.0802 - 04/08/10 - jgf - Fixed array out of bounds in log message C in timestep.F. Eliminated duplicate open of fort.22 in C NWS19. Fixed file name to open for NWS29 in hstart.F. C v49.0801 - 03/19/10 - jgf - Added code for new NWS29 option that embeds an C asymmetric vortex (NWS19) into an OWI formatted (NWS12) C basin scale met field (derived from NAM). Also separated C the application of wind reduction factors from the application C of the canopy coefficient, and made several key variables C from NWS19 available at the wind module level for use by C other wind-related subroutines. C v49.72 - 04/11/2011 - jgf - Removed init_cond.F and related comments C from makefile since they are no longer needed. C v49.71 - 04/11/2011 - kmd - Updated the initial condition file for C baroclinic to be evaluated in parallel similar to the binary C hotstart file; updated station information for 3D. C v49.70 - 04/11/2011 - rjw - Updated KSlip formulation fro KSlip=2, 3 C Added subroutine "Apply3DBottomFriction", to compute TK for 3D C simulations. This fixes Tau0 bug where time varying tau0 is C computed as function of TK. Added a call to "CalculateTimeVaryingTau0" C in timestep.F for C3D. C v49.69 - 04/07/2011 - jgf - Fixed typo (missing .) in netcdfio.F. C v49.68 - 04/07/2011 - tcm - Enlarged the field width (w) on several C format specifications for Ew.d fields. Note that w should be C set greater than or equal to 7 more than d according to the C FORTRAN 90 standards. For Fw.d, w should be greater than or C equal to 3 more than d. C v49.67 - 04/06/2011 - tcm - making additions for compiling adcirc on a C Windows based PC. In vortex.F resolved a conflict with the C variable "error" used as an integer in global.F and as a C real in constants.F. C v49.66 - 04/06/2011 - tcm - merged ice coverage into the trunk from C branch versions v49.64.01 through v49.64.01. C v49.64.02- 04/01/2011 - tcm - Fixed bug in Harmonic Analysis output when C output was called to be generated but the model never reached C the time interval between THAS and THAF during which the data C is to be created. Added a test for NWS and NCICE C compatability. Included Ice support in the init_cond.F file. C v49.64.01- 03/31/2011 - tcm - Added ice field data for use in modifying C wind drag coefficients (Enabled by setting NWS >= 1000 C uses files (fort.25 all ice types, fort.225 and fort.227 C only for ice types (NCICE) = 12 ). Added writing of global and C station ice concentration files (fort.93 and fort.91). C Added a cubic drag coefficient function based on ice C concentration levels and the typical Garratt wind speed C computations. Ice concentration files are only written C when met. files are written. C v49.65 - 04/06/2011 - tcm - Fixed typo bug in transport.F. C v49.64 - 04/06/2011 - rjw - Modified 3D code to allow user to select C standard Mellor-Yamada 2.5 closure (IVEC=50), or modified C Mellor-Yamada 2.5 with enhanced mixing (IVEC=51). (Note: for C IEVC=51, the alpha parameter for Z0S is read in from EVCON) C Added ability for 3D code use Manning's n values to compute C bottom roughness values based on Casey's work taken from C Bretschneider(1986). Modified 3D wetting scheme to simplify the C calculation of BSX/Y at newly wetted node. C v49.63 - 03/29/2011 - rjw - Removed debugging write stament from vsmy.F C v49.62 - 03/28/2011 - rjw - Modified the 3D code to include a hard wired C enhancement to the Mellor-Yamada 2.5 Turbulence Closure scheme C to account for increased mixing at the surface based on magnitude C of surface stress. Additionally, added nodal attribute, C "bottom_roughness_length" to be used in the 3D simulations. C included a couple of bug fixes found when implementing the C IM=611112 run type, and added back in the calculation of BSX and BSY C to the 3D wetting code in timestep.f. C v49.61 - 03/17/2011 - jgf - Corrected compiler flags for IBM p6 and C removed extra -DNETCDF for IBM p6 platform; corrected double C deallocation issue when writing ascii fort.44 file in parallel. C v49.60 - 03/09/2011 - jgf - Removed IStart3D variable from global_3dvs.F C to fix initialization issues with evtot; explicitly initialized C wsx1, wsy1, wsx2, wsy2 to avoid NaNs in vssol upon hotstart C of 3D runs in serial; added error handling to code that skips C down to the proper location in 3D output files; added flags for C gfortran compiler to take advantage of floating point exception C trapping for debugging. C v49.59 - 03/09/2011 - tcm - Changed "COMPILER" back to "compiler" in the C cmplrflags.mk file. C v49.58 - 03/08/2011 - jgf - Fixed NeiTabEle handling to account for its C semi-sparseness in subroutine bpg3D and in transport.F. C v49.57 - 03/08/2011 - jgf - Added private declaration to subroutines C in netcdf module to fix compilation on intel v9.1 compiler C (on kittyhawk); also tweaked intel compiler flag setup. C v49.56 - 03/03/2011 - jgf - Merged in netcdf3D branch revisions 562:597 C to trunk (branch version numbers v49.48.01 through v49.48.09). C v49.48.09- 03/03/2011 - jgf - Added compiler flags for IBM P6; added C capability to activate comprehensive trace logging with a single C compiler flag; changed trace logging so that all processors C write trace messages to local fort.16, which makes debugging C easier during parallel execution; added conditions C to netcdf memory deallocation for harmonic analysis hotstart; C fixed issue in binary hotstart in parallel by mapping to C NNODECODE instead of NODECODE; removed IHOT as argument to C initHarmonicParameters and eliminated duplicate calculations C of nfreq in harm.F; added doc directory with rough draft C for an ADCIRC Developers Guide. C v49.48.08- 02/25/2011 - jgf - Fixed record accounting in netcdf module C so that appends place data in proper position within the netcdf C data file, rather than always appending at the end; fixed issue C where 'it' was missing from call to hstart3D_out; added C call to prepnetcdf() to PREP15 option in adcprep. C v49.48.07- 02/24/2011 - jgf - Brought over vparr's adccmp.F from ERDC_esmf C branch for use in testing; added experimental new autotest code C for use in automated testing. C v49.48.06- 02/24/2011 - jgf - Fixed initialization issues for 3D netcdf C output files and 3D netcdf hotstart files in parallel; added C code to adcprep to capture 3D station names; added initialization C of 3D station coordinates to adcprep and adcirc/write_output.F. C v49.48.05- 02/23/2011 - jgf - Added dedicated arrays for storing coordinates C of 3D stations for use in 3D netcdf station output files; fixed C several typos and redundancy issues in netcdf module that C were interfering with 3D netcdf output. C v49.48.04- 02/23/2011 - jgf - Added new code to adcprep to set up C 3D netcdf output files for parallel runs and to set up netcdf C hotstart file file with 3D data for parallel runs; further C modularized mapping of full domain data to subdomains on the C fly during hotstart; modularized setting of output timing C during 3D hotstart. C v49.48.03- 02/11/2011 - jgf - Added new subroutines to netcdf module for C reading 3D data from netcdf hotstart file and for mapping C this 3D data from fulldomain to subdomains, combining reading C and mapping for each variable to conserve memory to the C greatest extent possible. C v49.48.02- 02/09/2011 - jgf - Added new subroutines to netcdf module for C allocating memory for 3D data in netcdf hotstart files; C initializing 3D data in netcdf hotstart files; and writing 3D C data to netcdf hotstart files; rearranged HSTART3D_OUT in C vsmy.F to handle both binary and netcdf output and to globalize C each 3D variable all at once rather than in horizontal layers C (although the layered output format is preserved for binary C output) and to allocate the memory for each 3D variable just C before it is needed and deallocating memory immediately after C its use is complete; the option of writing either subdomain or C fulldomain hotstart files was preserved for binary output; C the writing of data to binary hotstart files was also modularized C with the addition of a new subroutine. C v49.48.01- 02/03/2011 - jgf - Added new subroutines to netcdf module for C initializing and writing 3D data to netcdf output files; C modularized code for reading 3D stations in read_input.F and C added capability for reading 3D station names. C v49.55 - 03/03/2011 - tcm - Merged the last of the changes from the C ERDC tightly coupled with STWAVE branch. C v49.54.04 -03/03/2011 - tcm - Bring updates from the trunk getting C v49.54.03 -02/24/2011 - tcm - Added the option to make parallel adcirc C as a library back to the makefile. This library option is C needed for the external coupler. C v49.54.02- 02/34/2011 - jgf - Fixed minor compilation issues in adcprep.F C and writer.F. C v49.54.01- 02/22/2011 - tcm - Modifications to adcirc.F and writer.F to C allow for coupling with STWAVE via external couplers. C v49.54 - 02/22/2011 - tcm - Fixed a bug with hot start writers related C to the NOFF varible in writer.F. Cleaned up use of reserved C (dim, index, and scale) in the other files changed in this C commit. (dim --> iidim, index --> iindx, scale --> scaleval) C v49.53 - 02/18/2011 - jgf - Cleaned up logging for netcdf in adcprep C and fixed minor bug with v-velocity metadata in netcdf hotstart C files. C v49.52 - 02/17/2011 - jgf - Declared several subroutines in netcdf C module as private to enable compilation on Intel v9 compiler C on kittyhawk at RENCI. C v49.51 - 02/09/2011 - tcm - Merging fast search algorithms from the C merge_erdc_to49 branch. Incorporated the new search tree C searches into the prep routines and read_input.F. Updated the C readstations subroutine to accept a description variable for C allowing for the inputing of the correct station type. Added C the writer_pause subroutine but not using it currently. C v49.50 - 02/09/2011 - jgf - Fixed an issue in hstart.F where harmonic C analysis data were checked, even though harmonic analysis was C not specified as part of the simulation. C v49.49 - 02/02/2011 - jgf - Rearranged memory allocation for 3D globalio C to minimize the total amount of RAM required at any given time C for writing 3D output files. C v49.48 - 01/25/2011 - tcm - Added support for NRS=4 tight coupling with C STWAVE in timestep.F. C v49.47 - 01/25/2011 - rjw - Changed format statement 2393 to 2395 on C cmassey STWAVE routine in Updated read_input.F C v49.46 - 01/25/2011 - tcm - Updated read_input.F and adcprep routines C to allow for NRS=4 which is tightly coupled ADCIRC+STWAVE. C v49.45 - 01/24/2011 - jgf - Merged harmio branch r527:556 to trunk. C v49.43.19 01/23/2011 - jgf - Finished code to globalize all 3D output C files as well as code to write fulldomain 3D files in ascii C format in serial and in parallel; fixed 3D output file C initializations in cstart.F and hstart.F so they work in C parallel; corrected issues with reading 3D data from binary C hotstart file in parallel and mapping to subdomains on the fly. C v49.43.18 01/21/2011 - jgf - Added code to prep18 to write 3D station C mappings into fort.18 for 3D runs; added code to messenger.F C and global_3dvs.F to read 3D station mappings from fort.18 for C 3D runs; rearranged code to write out 3D turbulence stations C (fort.43) so that it is organized by output format; also added C code to use the new mappings to globalize and write out C the resulting fulldomain fort.43 files in ascii. C v49.43.17 01/19/2011 - jgf - Added globalization of 3D density (fort.44) C and 3D velocity (fort.45) to write_output.F; also added code C to write out these fulldomain 3D data in ascii; rearranged C existing 3D output code so that it is organized according to C output format. C v49.43.16 01/18/2011 - jgf - Added new globalio subroutines packNPbyM C and unpackNPbyM for globalizing 3D output; made changes to C enable globalization of fulldomain turbulence output file C (fort.46). C v49.43.15 01/18/2011 - jgf - Corrected several issues revealed by internal C testing of reading and writing harmonic analysis data in netcdf C format. C v49.43.14 01/15/2011 - jgf - Finished code for reading means and variance C calculation data for harmonic analysis from netcdf hotstart file; C code includes capability to read data in serial or in parallel C (mapping to subdomains in parallel on the fly). C v49.43.13 01/14/2011 - jgf - Finished code for reading harmonic analysis C data from netcdf hotstart file to continue harmonic analysis C during the hotstarted run; code includes capability to read C in h.a. data in in parallel as well as serial. C v49.43.12 01/12/2011 - jgf - Made adcprep reliant on harm.F for harmonic C variables; made adcprep reliant on global.F for station names; C completed changes to netcdfio.F and adcprep to allow harmonic C data to be written to hotstart files in parallel. C v49.43.11 01/10/2011 - jgf - Added code to netcdf module to reduce the C size of the 2D matrix that holds the boundary type arrays C so that they are dimensioned by (nope,max(nvdll)) and C (nbou,max(nvell)). C v49.43.10 01/09/2011 - jgf - Fixed initialization issues that were C preventing log and console messages from showing up during C adcprep; straightened out initialization of mesh boundaries C in netcdf to fix seg fault issue during adcprep of sl16 grid; C added a few missing calls to check_err in netcdf. C v49.43.09 12/24/2010 - jgf - Fixed issue in prep.F/prepnetcdf() that C caused compilation with netcdf to fail. C v49.43.08 12/23/2010 - jgf - Fixed issue where netcdf harmonic analysis C hotstart data were not being initialized in write_output.F; C added icha parameter to netcdf harmonic analysis hotstart file. C v49.43.07 12/22/2010 - jgf - Fixed numerous small issues in netcdf C module that were preventing it from writing harmonic analysis C load vectors to netcdf hotstart files. C v49.43.06 12/21/2010 - jgf - Corrected issue with initialization of netcdf C mesh data for hotstart file; added check for NHSTAR=3 before C initializing netcdf hotstart file; forced a rewrite of the C netcdf hotstart file each time it is called to write; moved C initialization of netcdf element data to initNetCDFCoord(). C v49.43.05 12/21/2010 - jgf - Added subroutines to netcdf module to C free memory that was allocated for writing netcdf data; added C subroutine calls to prep.F to free memory from netcdf module C once the data had been written; added initialization to C screenUnit i/o parameter in adcprep which corrected a memory C fault issue that occurred when using netcdf; moved call to C prepNetCDF() to just before the call to DECOMP() in adcprep. C v49.43.04 12/21/2010 - jgf - Added code to reduce memory use of netcdf C module by only retaining a single copy of the mesh and its C boundaries in memory and reusing it for each netcdf output file; C added new data types for storing harmonic station data C as well as fulldomain harmonic data to netcdf files; added C new variables, metadata, and subroutines for storing harmonic C analysis load vectors to netcdf output files; added subroutine C calls for tracing progress of subroutine calls to netcdf C module. C v49.43.03 12/09/2010 - jgf - Added code to write out means and variance C calculation results for harmonic analysis to full domain output C file in parallel; fixed issues in globalio for new packMbyNP; C fixed issues in output data descriptors for harmonic data; C added general subroutines for writing log and screen messages C to aid in debugging; fixed descrepancies between reading C and writing of binary hotstart file. C v49.43.02 12/08/2010 - jgf - Added new subroutines to write_output.F C to write harmonic analysis data to full domain output files C in parallel; finished separating harmonic analysis solution C from writing harmonic analysis output files. C v49.43.01 11/30/2010 - jgf - Preserved the reading of local hotstart C files as an option, but made the reading of fulldomain hotstart C files the default. Made changes to enable reading of harmonic C analysis hotstart data and 3D hotstart data from fulldomain C hotstart file in parallel by mapping the nodal data from C fulldomain to subdomain on each PE on the fly. Made changes C to globalio by adding new subroutines that can globalize harmonic C analysis data and 3D data. Reorganized harmonic analysis module C in the following ways: (1) moved variables from global.F to C harm.F if they pertained only to harmonic analysis; (2) separated C reading from initialization when harmonic analysis hotstart C file is read; (3) separated solution from writing when harmonic C analysis data is written to output; (4) moved code that updates C load vectors from timestep.F to harm.F; (5) moved code that C calls various subroutines to solve harmonic analysis at end of C a run from adcirc.F to harm.F. Added new code to C write_output.F to write globalized harmonic analysis data to C fulldomain hotstart file. Added subroutines to global.F C to streamline reading of hotstart arrays from binary files. C Removed subroutines from globalio.F that were not called C from anywhere. Made canopy coefficient default to 1 in nodal C attributes. Added fix for case where NRAMP is 1 but DRAMP is 0. C Updated hot2asc to read harmonic analysis data. C v49.44 - 01/19/2011 - tcm -- For owiwind.F changed the way the header C line of the file was read so that if changes in the text C portion occur in a file, these would not cause read failures. C This line is an 80 character line with two 10 digit date C fields which are the only data used off that line. The first C date field is contained in positions 56-65, and the second C date field is in positions 71-80. C v49.43 - 11/05/2010 - jgf - Removed netcdf remnants from hotlocalize C in prep.F and corrected a seg fault bug in hotlocalize caused C by six-digit IM values. C v49.42 - 11/05/2010 - jgf - Added compiler flags and library flags for C compiling with netcdf support on sapphire, jade, and diamond C at ERDC. C v49.41 - 11/04/2010 - jgf - Fixes for reading and writing netcdf hotstart C files in serial and in parallel; fix to allocate memory for C boundary arrays more efficiently in netcdf module. C v49.40 - 10/25/2010 - jgf - Fixes for writing netcdf metadata to trim C off trailing spaces when writing netcdf files from adcirc or C adcprep, as well as when writing subdomain fort.15 files from C adcprep. C v49.39 - 10/22/2010 - jgf - Updates to netcdf module for hotstarting, C including alternation of netcdf 67 and 68 files; additions to C support fulldomain integer arrays for nodecode and noff; updating C it and im in hotstart file; fix for ordering write in prep15 C (write metadata after 3D data in fort.15). C v49.38 - 10/06/2010 - jgf - Significant rearrangement of hstart.F to C consolidate the reading of the binary hotstart file all in C one section of the hotstart() subroutine; this was necessary C b/c the subroutine mingled the reading of the file with C initialization and logging. In order to have a choice of C file formats, the reading of the hotstart file must be a C separate step from initialization, logging and other tasks. C A call to read a netcdf hotstart file if IHOT.eq.367 or C IHOT.eq.368 was added. C v49.37 - 10/05/2010 - jgf - Removed utility program called init b/c it C was not being used; made compilation of prep.F and read_global.F C dependent on adcirc modules to facilitate sharing of common C code (this fixes an issue with compiling adcprep w/o netcdf); C finished reorganization and redevelopment of netcdf hotstart C code in netcdf module (netcdf hotstart testing still required). C v49.36 - 10/04/2010 - rjw - 1) Modifications to 3D turbulence: addition of C enhanced surface mixed layer. Variety of chioces hardcoded in C vsmy.f. Surface roughness in fort.15 is now the minimum C surface roughness. 2) Modifications to 3D wetting and drying C in timestep.f, modified wetting velocity formulation. C 3) Added ability to run 3D in GWCE explicit mode. GWCE is C solved outside 3D routine, IM type 611122 enables all of the C same GWCE explicit settings as 111122 except indicates to run C 3D instead of 2D. To implement also change GWCE time weighting C factors to 0.0 1.0 0.0 and change solver flag to 0. C v49.35 - 10/04/2010 - jgf - Fixes for netcdf: Fixed dimensioning of C NVELL during adcprep (was causing segfault on ec95d.grd among C others); streamlined netcdf module by removing redundant C dimensioning and ids for multidimensional nodal and station C data; fixed error handling and added graceful exit when C netcdf file from adcprep is not present when padcirc starts; C added and significantly rearranged code for writing netcdf C hotstart files. C v49.34 - 09/28/2010 - jgf - Fixes for netcdf: added call to update meta C data when hotstarting; eliminated ncmode variable, the C subroutine that sets it, and all references to it throughout C adcirc and adcprep. C v49.33 - 09/27/2010 - jgf - Fixes for netcdf: boundary data now written C correctly by adcprep now; read_global.F now uses the parse() C and a2f() subroutines from global.F instead of repeating them C in read_global.F ... this allows adcprep to pick up the fixes C that were made in global.F for parsing station names and also cuts C down on the incidence of cut-and-pasted code in ADCIRC. C v49.32 - 09/23/2010 - jgf - Fixes for netcdf: SLAM0 and SFEA0 C as well as mesh boundaries. Now works in serial. Parallel C should also work, but testing is required to confirm. C v49.31 - 09/23/2010 - jgf - Changes for writing netcdf in parallel: C added if/then to use array_g etc instead of array when writing C station data in parallel run; added a subroutine that is called C by adcprep to populate global variables so that they may be C written into the new initialized netcdf output file; moved C fort.15 parsing subroutines to global.F so that they can be C called from different places; fixed miscellaneous bug in NWS19 C in read_input.F by reading RSTIMINC when NRS.GE.1 instead of C just NRS.eq.1; fixed memory allocation for station names C in prep/read_global.F; moved the hotstart-relevant variables C from local declarations in prep.F to declarations in C pre_global.F; created new subroutine in prep.F to initialize C netcdf output files for parallel runs where they will be C required. C v49.30 - 09/16/2010 - jgf - Working on aswip: added capability to set C the Rmax to a constant or to a percentage of the calculated C value based on isotach data (if any) via command line options; C added capability to set method of selecting quadrant Rmax C via command line option; removed persistence of average Rmax C from hindcast into the forecast; corrected issue where C isotach speed can sometimes be zero in source data. C v49.29 - 08/25/2010 - jgf - Cleaned up parsing of metadata and station C names such that the netcdf module no longer assumes that a "!" C is present on each line, but is able to use it as a metadata C terminator if it is present. C v49.28 - 08/24/2010 - jgf - Declared variables in netcdf module as C private if the underlying type declaration was private. C v49.27 - 08/15/2010 - jgf - Fixes to storage of meta data, lat/lon C coordinates for stations and mesh nodes, and station names in C netcdf. C v49.26 - 08/15/2010 - jgf - Fixes to netcdf so it is now capable of C writing all data to the eight main output files in serial. C ------------------------------------------------------------------------ C r501 | stanaka | 2010-08-11 10:20:22 -0400 (Wed, 11 Aug 2010) | 1 line C Changed paths: C M /adcirc/trunk/src/timestep.F C Implemented Clint's zombie killing process for the error C elevation node. If there is a error elevation node, MSG_FINI is C called to finalize MPI environment. C (jgf: let's call this v49.25.) C v49.24 - 08/10/2010 - jgf - Fixed issues with the construction of arrays C for writing in netcdf files. Data in the reorganized module were C not being copied correctly from ADCIRC arrays to netcdf arrays. C v49.23 - 08/06/2010 - jgf - Fixes to reorganized NetCDF module. C Files now initialize properly, but there are still issues with C writing actual data. C v49.22 - 08/06/2010 - jgf - Commented out DEBUG statements in sizes.F C (related to command line option processing) and read_input.F C (related to opening files fort.14, fort.15, and fort.16). C v49.21 - jgf - Merged the subdomainhotstart branch into the trunk, C which included a massive reorganization of the netcdf module C for greater flexibility. The subdomainhotstart branch included C the following changes: C v49.17.04 08/04/2010 - jgf - Fourth set of changes for netcdf. C v49.17.03 07/29/2010 - jgf - Third set of changes for netcdf. C v49.17.03 07/27/2010 - jgf - Second set of changes for netcdf. C v49.17.02 07/23/2010 - jgf - First set of changes for working through C netcdf issues. C v49.17.01 07/14/2010 - jgf - Added an extra command line option -S to C write subdomain hotstart files during a run where fulldomain C hotstart files are specified. C------------------------------------------------------------------------ C r491 | stanaka | 2010-07-19 11:18:31 -0400 (Mon, 19 Jul 2010) | 1 line C Changed paths: C M /adcirc/trunk/src/adcirc.F C M /adcirc/trunk/src/global.F C M /adcirc/trunk/src/messenger.F C M /adcirc/trunk/src/sizes.F C M /adcirc/trunk/src/timestep.F C M /adcirc/trunk/src/writer.F C M /adcirc/trunk/work/makefile C C Introduced writer core approach for Hot Start file outputting. C This function works with '-H' option, for example C './padcirc -H 2'. We should use this option with less than two C hotstart writers, because we have only two hotstart files C (fort.67,68). Also, this option does not support NetCDF file C format. C (jgf: let's call this v49.20). C ------------------------------------------------------------------------ C r488 | stanaka | 2010-07-09 12:44:19 -0400 (Fri, 09 Jul 2010) | 1 line C Changed paths: C M /adcirc/trunk/src/couple2swan.F C M /adcirc/trunk/src/sizes.F C M /adcirc/trunk/src/writer.F C C Modified the writer cores algorithm to round robin outputting, C and adding spliting output file option. For safety outputting, C this round robin outputting works only with spliting option. If C you want to use the writer cores with this option, you should use C '-Ws' option instead of original writer cores option '-W'. C (jgf: let's call this v49.19). C ------------------------------------------------------------------------ C r487 | ndill | 2010-07-07 11:10:40 -0400 (Wed, 07 Jul 2010) | 1 line C Changed paths: C M /adcirc/trunk/wind/aswip_1.0.3.F C fixed format for reading in atcf when abs(longitude) > 100 C (jgf: let's call this v49.18). C ------------------------------------------------------------------------ C v49.17 - 06/28/2010 - jgf - Fixed a bug in hotlocalize where IM values of C greater than 10 triggered loading of 3D hotstart info. C v49.16 - 06/18/2010 - tcm -- Implemented pressure ramping differently, C as per Rick Luettich's suggestion: C We now set a background pressure at 1013 mb, then compute C the difference between the applied pressure and the C background pressure. Then apply the meterological ramp to C that difference and add it back to the background pressure. C Fixed a couple of typos in hstart.F where RampMete was C used instead of RampMete2. This fix, also fixes the problem C with the prmin for cold-started runs. C v49.15 - 06/16/10 - tcm Expanded the functionality of store63 and fixed C a bug in that same routine. store63 could now be used in C place of storeOne. C v49.14 - 06/04/10 - tcm Fixed a bug in write_output.F related to C globalizing NOFF (an elemental quantity) for parallel jobs C which caused hotstart runs to behave different than coldstart. C This bug was first worked on in v48.4638 and v48.4642. C v49.13 - 04/16/10 - jgf - Commented out debugging statements in adcprep.F, C prep.F, and read_global.F. C v49.12 - 04/16/10 - jgf - Commented out debugging statement in adcprep.F. C v49.11 - 03/23/10 - rjw - r431 -Changed KSlip formula in vsmy.F to use C q(n,1) for bottom velocity assuming that it occurs at height C delta z off the bed, where delta z is thickness of bottom C vertical layer C v49.10 - 03/17/10 - jgf - r424 -Changed WindDragLimit back to 0.0035 C from 0.0020 in preparation for merge back to trunk. C r419 | cdietrich | 2010-03-09 C Fixing an error in the passing of Manning's n values as Madsen C roughness lengths to SWAN. C r417 | cdietrich | 2010-03-02 C Changes to adcpost to allow for meshes with 10M+ elements. C r415 | cdietrich | 2010-02-16 C Added wave_refraction_in_swan as a nodal attribute in the C fort.13 file, so wave refraction can be enabled on a C vertex-by-vertex basis. C r412 | cdietrich | 2010-02-10 C One last change for meshes with 10M+ elements. C r411 | cdietrich | 2010-02-09 C Changing formatted writing and reading of element numbers C inside the fort.18 file from I8 to I12 to allow for meshes C with 10M+ elements. C r410 | cdietrich | 2010-02-08 C Small change to SWAN to allow for hot-starting of the coupled C model. C r408 | cdietrich | 2010-02-04 C Fixing bug with the output of the radiation stress gradients C to the rads.64 file. C r407 | cdietrich | 2010-02-02 C Fixing bug related to writing of last record in SWAN output C files. Needed to flush the writer cores from within the C SwanOutput routine. C r403 | cdietrich | 2010-01-27 C Turning off wave refraction for further testing. C r391 | cdietrich | 2010-01-20 C Implement coupling of bottom friction from ADCIRC to SWAN. C Our Manning's n values are used to compute a physical C roughness length, which is passed to SWAN and used with its C Madsen formulation. Now bottom friction in SWAN varies C spatially and is tied to the same roughness information used C by ADCIRC. C r390 | cdietrich | 2010-01-20 C Small change to wave-modified friction in ADCIRC. C r389 | cdietrich | 2010-01-20 C Lowering the cap on the wind drag coefficient from 0.0035 C to 0.002, to be in line with results from Mark Powell. C r388 | cdietrich | 2010-01-20 C Updating SWAN and ADCIRC compiler flags for the ERDC Diamond C machine. C r379 | cdietrich | 2010-01-11 C Updating SWAN compiler flags for the ERDC Sapphire machine. C r373 | cdietrich | 2010-01-06 C Removed SWAN's check for skinny elements. C r366 | cdietrich | 2009-12-16 C Committing changes related to the wave-modified friction. C Now the Manning n inside ADCIRC is increased in regions with C heavy wave breaking in SWAN. The wave-modified friction is C activated with an extra CSWANFRIC compiler flag. C r328 | cdietrich | 2009-10-19 C Committing SWAN half of coupling changes to repository. C r327 | cdietrich | 2009-10-19 C Committing ADCIRC half of coupling changes to repository. C r322 | cdietrich | 2009-10-13 C Creating new branch for coupling to the unstructured SWAN. C v49.09 - 03/17/10 - jgf - Removed stray cr from line endings in C makefile. C v49.08 - 03/08/10 - kmd - merged 3D_kmd branch into trunk C - 02/22/10 - kmd - modified the station output for the C 3D density file (fort.41) and updated the 2D compact C output for fort.63, fort.64, fort.73 and fort.74 C - 01/19/10 - kmd - updates to code for the spherical C factors in 3D and some of the output for 3D C r368 | kdresback | 2009-12-22 C created a branch for 3D baroclinic changes in this C version of the code (v49) - merged my 3D_baroclinic C branch with v49.06 that contained updates by rjw. C r332 | kdresback | 2009-10-22 C updates to the messenger.F file for bug C r323 | kdresback | 2009-10-13 C created initial branch for baroclinic code using v48 C of the code. This branch contains major updates C to the transport.F, vsmy.F, hstart.F and introduces C a new subroutine for initial conditions (init_cond.F). C v49.07 - 02/04/10 - jgf - Added entry for Cray-XT5 to C cmplrflags.mk. C v49.06 - 11/25/09 - rjw - merged 3D_rjw branch into trunk C - 10/30/09 - rjw - modifed turbulent lenght scale for newly wet C nodes and fixed bug in post processing fort.43 files and C commented out call to massbal3D C - 10/06/09 - rjw - Created a branch for 3D changes and merged C previous changes back into this branch to be debugged and C then remerged into the trunk C r285 | rweaver | 2009-09-01 12:31:04 -0400 (Tue, 01 Sep 2009) C modified vsmy.F and fixed bug in vsmy.F C r208 | rweaver | 2009-06-18 16:53:47 -0400 (Thu, 18 Jun 2009) C fix bug in Holland wind NWS=8 computation of translational velocity C r178 | rweaver | 2009-04-15 11:09:31 -0400 (Wed, 15 Apr 2009) C fixed format statements in post.f C r136 | rweaver | 2009-03-06 10:47:47 -0500 (Fri, 06 Mar 2009) C Commit changes in src and prep to 3D turbulent calculations C v49.05 - 11/24/09 - tcm - Renamed reserved words (count became kount, C index became indx, int became in2t and sum became summ). C Index in particular was being used as a user defined integer C and the intrinsic function in the same routine. C v49.04 - 11/24/09 - tcm - Fixed a bug with type NWS=3 and NWS=6 winds C in prep.F by removing them from the adcprep process and C instead of accessing local copies of the global wind file, C ADCIRC/PADCIRC now access the global wind file directly which C must now be in the global input directory. C V49.03 - 11/10/09 - tcm - Adding a couplers directory where coupler C routines and modules can be added to allow for coupling C ADCIRC to other models. Populated with an empty directory C to coupling with the wave model STWAVE. C V49.02 - 10/29/09 - tcm - Fixed bugs in hstart.F related to reading in C wind records for NWS = +2,+4,+5,+7,+12. This bug caused the C wind records to be behind by one record when hot starting. C Added an initialization for ITHS = 0 in adcirc.F for cold C start cases. Fixed format statement in owiwind.F. C v49.01 - 10/28/09 - TCM - Added else to initialize numSkipSnaps to C be 0 in owiwind.F (This fix did not get carried over with C the remerge from fastfloodplain to V49.) Also changed C reserved word count to kount in netcdfio.F to avoid C conflicts. C v49.00 - 09/29/09 - jgf - Incremented version number to v49.00 to C kick off the development of the next version of ADCIRC. C v48.4642 - 09/18/09 - jgf - Changed syntax for correcting NOFF in C write_output.F, since g95 choked on the original syntax. C Also fixed several bugs related to writing output C (particularly hotstart output) to subdomain files during C parallel execution. C v48.4641 - 09/09/09 - jgf - Fixed a potential issue in globalio.F where C the fortran intrinsic function 'size' was used as a variable C name in several older subroutines; replaced these instances C with the variable name 'size_this'. C v48.4640 - 09/04/09 - rjw - Changed aswip to persist Cent.press if a C zero valued for the central pressure is given C v48.4639 - 09/02/09 - jgf - Added feature that will read in min/max C files on a hotstart to preserve the full record of min and C max values; set NWS8 to choke if Rmax or Cp are set to zero C on any line of fort.22, including OFCL ("forecast"). C v48.4638 - 09/02/09 - tcm - Fixed a bug in writing hot start values for C NOFF which could have the wrong values after globalizing C because there are no rules for which processor owns an C element and Noff is an elemental value. The way around is to C take the minimum of (Noff, 1) as NOFF is supposed to be C either 0 or 1. C v48.4637 - 08/31/09 - rjw - Fixed out of bounds array in aswip C v48.4636 - 08/24/09 - rjw - Fixed Bug in NWS 19 code (first lon not C properly converted) C v48.4635 - 08/14/09 - tcm - Fixed bug in call to neighb within read_input C updated the adcirc makefiles for a pc C v48.4634 - 08/13/09 - tcm - Fixed bug in globalio where multiple variables C where repeated in use module line. C v48.4633 - 08/05/09 - rjw - Fixed Bug in translation direction in ASWIP output C v48.4632 - 08/05/09 - rjw - Fixed formatting error in ASWIP output C to make it compatible with wind.F and should still be C compatible with utilities that parse ATCF output. C v48.4631 - 07/30/09 - rjw - Modified aswip set background pressure to C hardwired value of 1013 C v48.4630 - 07/28/09 - jgf - Modified aswip output to include commas C to make it compatible with utilities that parse ATCF output. C v48.4629 - 07/28/09 - rjw - Added storm speed(kts) and direction(degrees) C information to aswip.F output fort.22 file C and the translation velocities (U and V) are now computed from C the speed and direction that is read in from C the new fort.22 file into NWS19 code C v48.4628 - 07/27/09 - jgf - Extended METONLY capability to cases where C 3D or harmonic output is requested, as well as the case where C execution begins as a hotstart. C v48.4627 - 07/27/09 - jgf - Created METONLY logical variable: if only C met output is requested, ADCIRC will skip solution of the C GWCE and the momentum equations. C v48.4626 - 07/27/09 - jgf - fixed bug in presizes.F if fort.15 file is C not found. C v48.4625 - 07/22/09 - rjw - fixed bugs in NWS19 code in wind.F and aswip.F C v48.4624 - 07/22/09 - rjw - fixed bugs in NWS19 code in wind.F and aswip.F C v48.4623 - 07/22/09 - rjw - fixed confusing lines and file names in C ASymmetric WInd Preprocessing (ASWIP) program C v48.4622 - 07/22/09 - rjw - Added NWS=19 Asymmetric Holland v2.0 C to src/ files C v48.4621 - 07/22/09 - rjw - Added NWS=19 Asymmetric Holland v2.0 C to prep/ files C v48.4620 - 07/21/09 - jgf - Modified makefile so that aswip executable C may be built with 'make aswip'. C v48.4619 - 07/21/09 - jgf - Fixed inclusion of MNMPROC in write_output.F; C merged lumped/explicit mode from Seizo Tanaka; rjw C merged aswip.F to generate metadata for asymmetric C vortex model. C v48.4618 -07/10/09 - tcm - Fixed bug in writing out global elevation C values using -99999.0 that exists in serial version C Also added 3 fields to the OutputDataDescript_t type C v48.4617 -06/19/09 - jgf - Made numeric constants in NWS8 double C precision; in NWS9 sub added initialization of told and C changed time.eq.time2 to time.gt.time2. C v48.4616 -06/18/09 - rjw - Modified the NWS=8 Holland wind to fix C bug in translational velocity to account for direction. C v48.4615 -06/02/09 - rjw - Removed erroneous call to C COMPUTE_RESELEM from read_input.F. C v48.4614 - 05/27/09 - mcf - Fixed implementation of Rick Luettich C V/Vmax tapering formulation (as in NWS=8) C for future experimentation - not activated C v48.4613 - 05/27/09 - jgf - Commented out debug lines in C write_output.F and globalio.F that were filling up C fort.16. C v48.4612 - 05/26/09 - rjw - Fix format statement in postprocess for 3D. C v48.4611 - 05/22/09 - mcf - Reverted damping of translational velocity C back to the original formulation C v48.4610 - 05/22/09 - jgf - Changed air density back to 0.001293 in C stress calc for NWS8 and NWS9 in timestep.F. C v48.4609 - 05/21/09 - mcf - Fixed seg-fault when writing netcdf hotstart C files, incorporated fix in the define portion of netcdfio C for grids with no specified boundary forcing segments or nodes C v48.4608 - 05/19/09 - mcf - Implemented an old bug fix: units conversion C in asymmetric wind code (NWS=9) C v48.4607 - 05/18/09 - jgf - Fixes to get adcprep and adcirc to compile C and link with netCDF on Jade (Cray-XT4) at ERDC. C v48.4606 - 05/13/09 - rjw - modify makefile to compile adcprep using C version.F C v48.4605 - 05/12/09 - rjw - modify asymmetric code to use damping of C translational velocity based on ratio of V:Vmax (pass Vmax C from wind.F into vortex.F subroutine uvp), and set (in vortex.F) C velocities at center of vortex (dist < 1m) to zero. C v48.4604 - 05/08/09 - jgf - Fixed typo in version.F C v48.4603 - 05/08/09 - rjw- updated the distance calculation formula C in Holland Model (NWS=8) to more accuarately calculate C distance along a shpere, and modified the coriolis formula C to compute a spatially varying coriolis term also for the C Symmetric Holland model (NWS=8) C v48.4602 - 05/08/09 - jgf - Changed air density for symmetric and C asymmetric vortex models (NWS=8 and 9) to 1.15 kg/m^3, C for Holland B calculation as well as conversion to wind C stress. Changed asymmetric far field wind damping so C that only the translation velocity is damped. C v48.4601 - 05/01/09 - jgf - Fixed improperly located check for C netcdfio support in presizes.F; removed FileFmt etc from USE C PRE_GLOBAL in prep.F (now uses VERSION instead). Merged C prep.F from v48.48 (trunk r161) b/c it has fastprep13. C v48.46 - 02/26/09 - jgf - Corrected backloaded tau0 averaging C formula: now just one time level. C v48.45 - 02/26/09 - jgf - Fixed backloaded tau0 formula by C removing an extra alpha factor from tau0(t-3) term. C v48.44 - 02/02/09 - mcf - fixed time unlimited index and time variable C in the fort.63.nc output file for hotstart cases in netcdfio C module C v48.43 - 01/29/09 - rjw - modified fixes to massbal.F (check for C ICS status) and vsmy.F (initialized turbulent length scale to C lmin= 0.001 in cstart.F (COLDSTART_3D)) and reformatted C declaration statements in global_3dvs.F. C v48.42 - 01/28/09 - jgf - Added two new time averaging options for tau0 C ... activated by tau0=-6.x or 7.x in fort.15. C v48.41 - 12/11/08 - rjw - fix massbal.F COMPUTE_RESELEM subroutine C In order to compare the global to local locations C Global x and y need to be in converted units not in Lat and Lon C use subroutine CPP to make conversion C v48.40 - 12/11/08 - rjw - fix version number in version.F C v48.39 - 12/11/08 - rjw - applied bug fixes in cstart.F and hstart.F C format statement, changed default il value to 2 in vsmy.F C v48.38 - 12/10/08 - jgf - Added #ifdef HAVE_MPI_MOD around mpif.h C in massbal.F to fix failed compile on Cray-XT3 (sapphire). C v48.37 - 12/10/08 - mcf - modified netcdfio metadata C v48.36 - 12/03/08 - jgf - Fixed a bug in PREP22 of adcprep, C combined CASE(3,6) into CASE(8,9) b/c the approach of C using the shell to copy files in CASE(3,6) had been C commented out. C v48.35 - 12/03/08 - jgf - Fixed a bug in adcprep that was causing C unprintable characters to appear in the names of C subdomain fort.22 files on the IBM p5 platform. C v48.34 - 12/02/08 - jgf - Fixed a bug in adcprep to look for C tau0min and tau0max values in fort.15 if tau0 is less C than -5.0 AND greater than -6.0. C v48.33 - 11/25/08 - jgf - Added a check for negative central pressure C deficit in Symmetric Holland code (NWS=8). Also fixed a bug C introduced in the merge of Seizo's fix to internal barrier C boundary handling in v48.32. Reimplemented tau0base C formulation when TAU0=-3.0, moved fulldomain time C varying TAU0 option to TAU0=-5.0 ... if tau0 output is C desired, the corresponding options are -3.1 or -5.1. C v48.32 - 09/16/08 - jgf - Merged fix from C.Mattocks to include C directives in metis, e.g., to "metis.h". C Fixed HollandGet so that lat is read in as four digits, thus C allowing storm locations west of 99.9W lon; integrated Seizo C fixes to decomp.F (mpi communication table) and in timestep.F C (wet/dry checks along internal barriers). C v48.31 - 09/16/08 - mcf - added hotstart file format identifiers in C from version.F, modified netCDF metadata, added time counters C and record numbers for hotstart files, fixed netCDF hotstart C flow for preprocessor, adcirc and padcirc. C v48.30 - 08/29/08 - rjw - merged CD's mods for 3D wetting and Drying C v48.29 - 08/07/08 - mcf - modified starting character counter in parser C v48.28 - 08/01/08 - mcf - included File Format Version parameters C in netCDF hotstart files and changed FileFmtMinor to 2 to account C for the removal of an extra empty record in hotstart files C v48.27 - 07/30/08 - mcf - corrected dt when writing initializing netCDF C hotstart files in prep.F C v48.26 - 07/30/08 - mcf - corrected noff range value when writing netCDF C hotstart files (range 0 to 1) C v48.25 - 07/30/08 - mcf - corrected memory/compiling/running issues C with arrays in the binary and netCDF hotstart capability C v48.24 - 07/25/08 - mcf - Added code in the hstart.F file to check if C hotstart binary file exists in subdomains for netcdf hotstart C input file option when running the parallel version. C v48.23 - 07/25/07 - jgf - Fixed hotstart file incrementation in C write_output.F, prep.F, hstart.F and harm.F. Updated hot2asc.F C utility with fixed file incrementation. Restored machdep.F. C Updated cmplrflags.mk for apple darwin (removed -DIBM). C v48.22 - 07/24/07 - jgf - Fixed hotstart file incrementation in C write_output.F and hstart.F. C v48.21 - 07/24/08 - mcf - Added code in the hstart.F file to only read C netCDF hotstart files when running the serial version. C v48.20 - 07/24/08 - mcf - Added code in the preprocessor to initialize C netCDF hotstart files. C v48.19 - 07/22/08 - mcf - Added code to assign the globalized C hotstart variables to netCDF variables only in processor 0. C v48.18 - 07/09/08 - mcf - Added code to adcprep and adcirc to C initialize and write netCDF hotstart files. C v48.17 - 06/09/08 - jgf - Added code to adcprep to decompose UnSWAN C boundaries and distribute the UnSWAN input file to each of the C subdomain directories. C v48.16 - 06/05/08 - jgf - Merged additional changes from cf's latest C (from adc48_15_4). C v48.11 - 03/14/08 - jgf - Added cf's latest write_output.F, corrected C Chezy conversion to quadratic in nodalattr.F, added C info messages to cmplrflags.mk and makefile, corrected issues C to allow use of g95 compiler. C v48.10 - 03/13/08 - jgf - Added cf's latest prep.F, read_global.F, C global.F, and netcdfio.F to enable boundaries and station C names in the metadata and to fix bugs. C v48.09 - 02/25/08 - jgf - Added cf's latest netcdfio.F and C write_output.F and synchronized arguments. Also added C ibtypenc variable (and allocation). Fixed bug in makefile C that prevented compilation of adcprep without netcdf. C v48.08 - 02/19/08 - jgf - Removed trailing space from "pgi " in C cmplrflags.mk. Removed p15 and other utilities from 'all' C target in makefile. Corrected bug in FEGRIDS sub in vsmy.F C where variables were declared as REAL rather than REAL(SZ). C Also made all constants into doubles in FEGRIDS sub. C In lsq.F, changed rinv to intent(inout) on line 538 to relieve C compiler issue on certain platforms. Added PREPNETCDF C subroutine to adcprep to initialize netcdf output files for C parallel runs. Updated makefile to reflect the linking of C the netcdf module with adcprep. C v48.07 - 02/07/08 - jgf - Merged changes from mcf for netCDF. C v48.04 - 11/19/07 - jgf - Fixed bug that was preventing NetCDF metadata C from being written to subdomains. Added ABS() to station C mappings when reading the fort.80 file in adcpost. Added C inverted barometer boundary condition (commented out pending C successful testing). C v48.03 - 10/19/07 - jgf - Created write_output.F, rearranged NetCDF C to use globalio. Merged kmd's 3D hotstart decomposition C in prep.F. C v48.02 - 10/04/07 - jgf - Fixed bug in UPDATEC3D in messenger.F. C v48.01 - 10/03/07 - jgf - Changed CrayXT3 and PGI compiler flags C (-c89 and -Mextend). Fixed bug that caused the min and max C variable tau0 to be read even if not needed. Fixed bug in C post52 that was writing NSTAE instead of NSTAV. Fixed missing C comma in FORMAT statement in globalio.F. Fixed missing right C paren in timestep.F. C v48.00 - 10/02/07 - jgf - Added ONLY keyword to USE NETCDF in C timestep.F to enable it to compile without conflicts. C v47.13 - 09/28/07 - jgf - Fixed two typo bugs in post.F for binary C fort.64 and fort.74 files. Fixed 3D stations bug that gave C incorrect output for stations inside ghost elements. Fixed C mixup between NOUTGE=3 and NOUTGE=4. C v47.12 09/05/07 - jgf - Merged code for netcdf from cf. C v47.11 - 09/03/07 - jgf - Added code to read in limits of time varying C tau0 directly from fort.15 file. C v47.10 - 08/28/07 - jgf - Merged code back in to create fort.80 file C for use in post processing 3D and harmonic analysis files. C Merged adcpost back in. C v47.09 - 07/26/07 - jgf - Fixed two bugs with new nodal attribute that C were causing SIGSEGV. C v47.08 - 07/24/07 - jgf - Added new nodal attribute for floor and C ceiling of variable tau0. Made "fully populated" output format C standard again for fort.63, fort.64, etc. Made "sparse" output C format selectable by setting NOUTGE(etc)=3. C v47.07 - 07/23/07 - jgf - Added feature to inflate to create chunked C output. Removed the hardwired 2nd floor for variable tau0. C v47.06 - 07/16/07 - jgf - Fixed bug in timestep.F where NSTAE was C used to write 2D velocity station file. Added global output C for time varying tau0. C v47.05 - 07/15/07 - jgf - Enhanced inflate.F so that it count the C number of datasets in the file and place the correct number C at the top in NDSETS. Updated 3D header and store subroutines C in globalio. C v47.04 - 07/05/07 - jgf - Merged latest asymmetric vortex wind model, C made CF from fort.15 into floor of equivalent quadratic C bottom drag when Manning's N is used, added hstime utility C and updated it for the latest hotstart file format. C v47.03 - 06/30/07 - jgf - Changed inflate to use infinite loop instead C of relying on NDSETS. C v47.02 - 06/28/07 - jgf - Rewrote the prep13 subroutine so that it C only uses one open file at a time to avoid exceeding the C Fortran compiler limitation on simultaneously open files; C changed to unformatted read of nodal attribute files to C make them more tolerant of leading spaces and end-of-line C characters. C v47.01 - 06/21/07 - jgf - Applied the fixes from the stable version C of ADCIRC to v46.57. C v47.00 - 06/04/07 - jgf - Applied the changes from the stable ADCIRC C series v46.32jgfxx to version 46.52 as follows: C v46.32jgf12- 02/20/07 - jgf - Fixed bug in adcpost where format string was C one field short. C v46.32jgf11- 02/19/07 - jgf - Fixed bug in 3D station post processing C in adcpost. Added hsconv program to convert v44.15 hot start C files to v46.32 hot start files for operational forecasting C system. C v46.32jgf10- 01/10/07 - jgf - Fixed bug by renaming LV_TEMP_A to TEMP_LV_A. C Fixed bug in Holland subroutine by using a backward difference C instead of forward. C v46.32jgf8 - 11/07/06 - jgf - Commented out section in adcprep that C reverses the sign of NOUTE, etc for hotstarts. Fixed bug in C code that found the end of output files for appending during C hotstart; now uses infinite loop rather than trusting the C number of datasets listed at the top of the file. Fixed bug C in logical if statement in nodal attributes by adding C parentheses because different compilers were treating C the logical statements differently. Added Shintaro's wet/dry C changes, including his check for infinitely narrow C bottlenecks. Modified compiler flags for SGI. Fixed bug in C makefiles to make all .Fs capital. C v46.32jgf6 - 10/19/06 - jgf - Updated PC makefiles. C ---------------------------------------------------------------------------- C C v46.57 - 02/15/07 - sb - The change on RS2GET made at v46.54 was * C removed. The original strategy (explained in the comment on * C v46.54) is now used. * C v46.56 - 02/09/07 - sb -Dedicated writer feature was implemented. * C v46.55 - xx/xx/xx - This version number was skipped. * C v46.54 - 01/03/07 - sb - RS2GET in rs2.F now takes maximum values if * C stwave domains overlap. (It used to take the value of the * C domain of the first appearance in a fort.23 file.) * C This requires more memory, * C - SB_WETDRY compilation flag to enable w/d fixes that had been * C implemented since v46.48 was intoduced. * C (NOT enabled by default.) * C - Made torelant to leading white spaces in attribute names * C in fort.13. * C v46.53 - 12/11/06 - sb - Fixed a minor bug. * C v46.52 - 11/17/06 - sb - Fixed a minor bug. * C v46.51 - 11/17/06 - sb - Compilation flags to disable some of max files* C have changed: * C -DNELMAX (to suppress maxele.63) * C -DNUVMAX (to suppress maxvel.63) * C -DNPRMAX (to suppress minpr.63) * C -DNWVMAX (to suppress maxwvel.63) * C -DNRSMAX (to suppress maxrs.63) * C v46.50 - 11/03/06 - sb - Two of the three w/d modifications introduced * C in v46.48 are now commented out. Only one remains. Maximum * C values of several variables in time at each node are now * C written out in fort.63 format at the last time step. The new * C output files are as follows: * C maxele.63, maxvel.63, maxwvel.63, minpr.63, maxrs.63 * C These outputs can be suppressed by adding some of the * C following compiler options: * C -DNEVMAX (to suppress maxele.63 and maxvel.63) * C -DNPWMAX (to suppress minpr.63 and maxwvel.63) * C -DNRSMAX (to suppress maxrs.63) * C v46.49 - 11/03/06 - sb - One of three wetting/drying algorithm * C modifications introduced in v46.48 was commented out. * C Sparse global output file writing was implemented. * C fort.73 file output was suppressed. * C v46.48 - 11/03/06 - sb - Modified wetting and drying algorithm * C - Fixed a bug in river ramping * C - Fort.63, 64, 73 and 74 files are now in sparse format. * C Use "inflate" program under "util" directory to inflate the * C outputs files. * C v46.41 - 10/04/06 - rtm removed system calls, now using * C c routines. * C v46.40 - 10/04/06 - sb added dynamic tau0 feature. This feature kick * C in when TAU0 = -3.0 in fort.15. * C v46.39 - 10/01/06 - vjp, rtm - added global output files for units * C 61,62,63,64,67-68,71,72,73,74, but 51,52,53,54,55 not yet. * C v46.33 - 09/21/06 - vjp, rtm, sb, jgf - Merged the following changes * C from: * C rtm, vjp: added option 'partmesh' to preprocessor to partition* C mesh only; added option 'prep13' to decompose the fort.13 * C file only; added subroutine relocalize to relocalize selected * C files after a prepall; added GLOBAL_IO module to write global * C output files in parallel on-the-fly; improvements to * C build13.F. Merged the following changes from sb: added * C OWIWIND module with NWS=12 for direct input of OWI wind data * C on rectangular grid (no preprocessing); added RS2 module with * C NWS=2xx for STWAVE output direct read; added buildstwave23.F. * C v46.32 - 09/15/06 - jgf - Made boundary layer adjustment in Holland * C model a parameter in the fort.15 file. * C v46.31 - 09/15/06 - jgf - Fixed integer divide in Manning's N * C exponent bug. Fixed bug in first read of Holland fort.22. * C Now subtracting translation speed before calculating Holland * C parameters. * C v46.30 - 09/07/06 - jgf - Fixed bug for reading new Holland reference * C time and allocating memory in Holland subroutine. * C v46.29 - 09/07/06 - jgf - Fixed compilation bugs in lsq.F and vortex.F.* C Added build12.F utility and new version of build13.F utility * C from vjp. Added reference time to Holland model, subtracted * C storm trans. vel. from wind speed before calcs in Holland. * C v46.28 - 08/26/06 - jgf - Fixed IDENMSG bug (kd) in prep.F, updated * C preprocessing code to support baroclinic and transport runs, * C added PREP10 subroutine. Added call to timeconv in Holland * C subroutine to calculate time between hindcasts. Added * C StormNumber to replace WTIMINC for NWS=8 (Holland model) for * C wind multiplier support. * C v46.27 - 08/22/06 - jgf - Replaced IRType with IDen in header for * C fort.44 files. Added scheme to handle negative default values * C of tau0 in nodal attributes file. * C v46.26 - 08/22/06 - jgf - Moved new wind.F into src directory where it * C belongs. * C v46.25 - 08/21/06 - jgf - Implemented choice of tau0 schemes for * C negative tau0 in fort.15 file, updated asymmetric hurricane * C code from cf. * C v46.24 - 08/16/06 - jgf - Added updated p15.F and owi22.F from sb. * C widened output format in fort.63 to avoid truncation bug when * C elevations have negative three digit exponents. Lifted * C restriction on overlap of IBTYPEs 52 and 24. Fixed bug in * C prep22 for NWS=1,3,4 where outer loop was missing and only * C the first set of wind data was processed. * C v46.23 - 08/13/06 - jgf - Fixed a bug in the multiple ramp functions * C in timestep.F for the Katrina ramp. * C v46.22 - 08/09/06 - jgf - Added a 1/0.7 factor to max wind forecast in * C Holland wind model. * C v46.21 - 08/09/06 - jgf - Added code to support new IBTYPE=52. Changes * C to Holland wind model: added debugging code, fixed bug that * C was miscalculating wind time increment, added 0.88 factor to * C convert from 1 minute to 10 minute winds. Fixed bug in * C prep13 that was breaking long lines of default values onto * C multiple lines. * C v46.20 - 08/03/06 - jgf - Applied a 0.7 factor to Holland winds, fixed * C a bug in header for debug elevations. * C v46.19 - 07/26/06 - jgf - Discovered and fixed a bare #ifdef bug on * C periodic bcs in timestep.F. Added negative NSCREEN option to * C write screen output to a fort.999 file. * C v46.18 - 07/21/06 - jgf - Implemented horizontal eddy viscosity * C and horizontal eddy diffusivity as nodal attributes for * C research purposes, not to be used in production runs. * C v46.17 - 07/21/06 - jgf - Fixed empty arguments bug in WriteWarnElev() * C subroutine call. * C v46.16 - 07/20/06 - jgf - Added #ifdef IBM to INT() calls to * C work around bugs in non-IBM compilers that don't like the * C KIND argument. Fixed a col. 73 bug in in prep.F. Changed * C METIS1 to METIS in makefile to fix CrayXT3 build bug. * C Merged asymmetric hurricane wind model from cf and cm. * C v46.15 - 07/10/06 - jgf - Added a FORMAT statement to prep.F for * C fort.13 to keep attributes with many values on a single line. * C v46.14 - 07/07/06 - jgf - Fixed bug from Katrina testing in hstart.F * C where too many records were read from hotstart file. * C v46.13 - 07/06/06 - jgf - Fixed build bug in turb in vsmy.F where real * C exponentiation was used on a PARAMETER. * C v46.12 - 07/06/06 - jgf - Added debugging code for wet/dry. Removed * C the check on KEMAX from screen output logic. Merged fixes * C from vjp: modified two lines in src/wind.F for IBM, added a * C util subdir and put build13.F in it, added wind subdir and put* C p15.F and owi22.F in it, modified the makefile (but did not * C modify the pc oriented makefiles) to build new targets and * C reduced fflags to just 3 instead of 5 types, modified the * C cmplrflags.mk file. Fixed a bug in subroutine turb in vsmy.F * C where constants were initialized but not saved. * C v46.11 - 07/05/06 - jgf - Added debugging code for high elevations. * C v46.10 - 07/03/06 - jgf - Fixed bug (ral) that prevented friction * C flags from being set for spatially varying friction. * C v46.09 - 06/20/06 - jgf - Fixed bug in timestep.F when NRamp=0. Fixed * C handling of NOLIFA=3 in read_input.F. Fixed bugs in Holland * C subroutines of wind.F (mult. lon by -1, mult RMW by 1000). * C v46.08 - 06/14/06 - jgf - adcpost: Fixed bug in number of stations * C reported in fort.72 files. Upgraded file handling in Holland * C subroutine. Created fine-grained ramp functions by converting * C NRAMP in fort.15 to the number of ramp functions. * C v46.07 - 06/09/06 - jgf - Fixed bug in reading Holland files. * C v46.06 - 06/08/06 - jgf - Fixed bug in reading WTIMINC in adcirc * C and increased line size in adcprep for Holland wind model. * C Fixed compile bug on IBM p5. * C v46.05 - 06/07/06 - jgf - Added code to calculate Cd from Mannings N * C at every time step. Rewrote Holland subroutine to accept * C hybrid hindcast/forecast input files. Replaced metis.F with * C a new one from vjp with upgrades to weir node dual handling. * C Fixed bug that prevented linking with libmetis.a on IBM p5. * C v46.04 - 06/05/06 - jgf - Applied bde's fixes to compile on IBM p5. * C v46.03 - 05/30/06 - jgf - Implemented Holland Wind model (NWS=8) * C subroutine and added code to adcprep to copy NWS8 files * C to subdirectories. Commented out some features req'd for * C Katrina compliance. * C v46.02 - 05/25/06 - jgf - Implemented GeoidOffset code to allow a * C permanent water level offset from the geoid. Implemented * C wind drag cap function. Implemented directional wind * C roughness and canopy factor. Added a combined radiation / * C flux condition for river boundaries (IBTYPE=22). Added a * C hard-coded spatially varying tau0 scheme vbased on depth to * C match v42. Applied kd's BPG bug fix. Fixed major bugs in * C ADCPREP/ADCPOST to post-process the wind recording station * C files, changing the format of the fort.80 file. * C v46.00 - 04/18/06 - jgf - Adcirc: Created new nodal attributes module * C to hold node-by-node data that is time invariant. Applied * C vjp's fixes to itpackv.F and messenger.F, added kind parameter* C to int() calls, made nscreen the time step increment between * C screen outputs (except for error messages) to work around * C limited log file sizes on certain platforms, added option for * C NWS=7 (grid-independent wind stress). Adcprep: Added code to * C break up new nodal attributes file (unit 13). Removed code * C that decomposed the StartDry and spatially varying bottom * C friction files. Added implicit none to all subroutines, added * C trim() function to cut down files size from adcpost, made * C infinite loop over data sets to fix bug where the wrong number* C of data sets are listed at the top of subdomain files after * C hot start. Added capability to adcprep option 2 to skip * C processing of certain files. Build: added vjp's compiler * C flags, meb's small fixes to makefiles, updates to PC * C makefiles, new version of config.guess. * C v45.12 - 03/17/06 - kd/rk/jgf - Major additions by kd/rk: * C Added 3D baroclinic transport code in transport.F. Added new * C BPG subroutine in timestep.F. Updated cstart.F, global_3dvs.F * C and read_input.F to handle associated input data. Fixed bug * C in VSSOL that was causing inadvertent mixing of current and * C future timelevels. Fixed bug in VSDISP that caused an * C erroneous early return. 3D BAROCLINIC NOW WORKS. * C jgf: Fixed bug in read_global3dvs to handle 3D station * C locations properly (SL0 and SF0). Added code to adcprep to * C process new input data for 3D baroclinic cases. Added MPI * C calls in timestep.F in BPG3D() to pass BPGs to ghost nodes * C for 3D baroclinic in parallel (now works also). Added code * C to read in vert. 3D spacing and vert. eddy vis. in adcprep. * C Applied tjc's DIRNAME patch for ESMF compatibility. Rewrote * C adcprep routine for fort.20, as the original did not handle * C ghost nodes. Fixed another bug in fort.20 adcprep to widen * C the formatting for fluxes. Added jjw's 3 tier tau0 scheme. * C Added bde's bluegene cmplrflags. Increased modularity by * C subroutinizing station locations and adcprep file opening. * C v45.11 - 02/02/06 - jgf - Adcirc: Bugfix to remove "save" compiler * C flag: OnDiag/OffDiag calculation on every entrance now to * C GWCE_New. Bugfix(kd) to correct 3D elevations: flag IFNLFA * C mult. by eta2 in timestep.F. Bugfix to vsmy.F to use IFNLFA * C flag instead of NOLIFA in calculations. Adcprep: Added code to* C handle 3D recording stations defined by coordinates rather * C than node number (changed format of fort.80 file), added IDEN * C to fort.80 file, fixed routines for post processing fort.41, * C 42, 43, 44, 45 and 46 files, added compiler flags for x86_64 * C (contrib. tjc) and to "save" subroutine values, fixed NINT * C bug in read_input.F that made 3D output off by 1 time step. * C v45.10 - 01/12/06 - jgf - Adcprep: partially updated preprocessing * C routines to reflect changes in 3D input files. * C v45.09 - 01/09/06 - jgf - Adcirc: fixed typo bugs in global_3dvs.F that* C prevented 3D output (NOT3DSDS->NTO3DSDS). Adcprep: added new * C logical variable for aperiodic vs periodic flow boundaries to * C fix fort.20 breakup bug. * C v45.08 - 12/16/05 - jgf - Adcirc: eliminated 'formatted' statements * C for binary files, relying instead on compiler flags for * C record length, moved setting of NOLIFA flag in adcirc.F to * C fix hotstart bug, changed declarations to COMPLEX(SZ) to fix * C bug in 3D hotstart. * C v45.07 - 11/21/05 - jgf - Adcirc: added option to NHSTAR for testing * C adcirc's hot start capability. Prep/post: Wrote code to * C handle fort.71, fort.72, fort.73; upgraded REALs to proper * C precision; modified fort.80 file format to handle NOFF array * C in hotstart files; created post processing options to either * C prepall or just prep14/15. * C v45.06 - 09/07/05 - jgf - Found/fixed bug in timestep.f: TotalArea * C was erroneously zeroed. Changed names of some LOGICALs for * C greater consistency in naming of solution options. Selected * C set of default options, including corrected area integration. * C Discovered and fixed bug for parallel when NCCHANGE flag was * C set in one subdomain but not another. Updated pre and post * C processing routines, including merging changes written by * C MEB to handle fort.20 files. * C v45.05 - 08/1- rl - from 45.04 - for NWS=6, eliminated adding 360 to * C negative longitudes to match AVN model grid setup. User is * C now required to provide met and ADCIRC grid that correspond * C in space without adjusting the longitude. Also the input * C variable order has been changed to U,V,P to be consistent * C with other NWS input formats. * C Modules affected: wind.f * C v45.04 - 06/16-21/05 - rl - from 45.03 - compute baroclinic press grad * C from surf down, set = above value if into ground. * C read in sigma grid (IGC=0) or E.V. values (IEVC=0) from * C fort.15 file. * C setup initial condition files for baroclinic or transport * C 2DDI, 3D runs * C revised GWCE and 2D Momentum Eq. baroclinic calcs. * C made new GWCE and 2D ME formulations the default * C Modules affected: vssol.f, read_input.f global.f global_3D.f * C timestep.f * C v45.03 - 06/08/05 - rl - from 45.02 - fixed 3D bugs: compute QX, QY in * C 3D VSSOL, corrected name BcPressNN in ZSURFBuoy, added EP to * C fort.67/68 hot start files & eliminated fort.99 file, * C fixed roundoff error for ievc=10,11 * C Modules affected: vssol.f timestep.f hstart.f global_3dv.f * C v45.02 - 03/18/05 - rl - fixed EQUIVALENCE compile problem, changed * C alpha header inputs to 80 characters long, cleaned up Kendra's* C pred/corr algorithm, including re-localizing OffDiag, OnDiag * C Modules affected: global.f timestep.f read_input.f * C v45.01 - 10/ /04 - rl - v44.19t Kendra's pred/corr algorithm * C reinstated jjw's spatially variable Tau0 algorithm * C Modules affected: read_input.f, global.f * C v44.19a-t - 06/04 - jjw,rl - from 44.18 - new wetting and drying check * C NOFF added to hotstart file * C Modules affected: adcirc.f cstart.f hstart.f * C v44.18 - 05/26/04 - rl - from 44.17d - small change to EP check, * C overtopping barriers, baroclinic pressure gradient calc, and * C IMAG to AIMAG in messg. * C Modules affected: timestep.f messenger.f * C v44.17d - 05/24/04 - rl - from 44.17c - change variable declaration in * C preprocessor so they are not hardwired to real*4. Rather, * C they are hardwired to real*8 or value of SZ as determined by * C compiler flag. No change to ADCIRC. * C Modules affected: presizes.f pre_global.f * C v44.17c - 05/21/04 - rl - from 44.17b - change output format in post * C processing for fort.63, fort.64 files. No change to ADCIRC. * C Modules affected: post.f * C v44.17b - 05/18/04 - rl - from 44.17a - vic parrs fixes to prep22 and * C prep23 to write wind info to ghost nodes. No change to ADCIRC * C Modules affected: prep.f * C v44.17a - 05/18/04 - rl - from 44.17 - islands kept as islands when they* C are cut by subdomains in parallel run. No change to ADCIRC * C Modules affected: prep.f * C v44.17 - 05/03/04 - rl - from 44.16 - bug fix for 2 part lateral stress* C calculations in GWCE. Bug dates back to 44.15. New spatially * C variable Tau0 formulation. * C Modules affected: timestep.f read_input.f * C v44.16 - 04/30/04 - rl - from 44.15 - added symmetric lateral stress * C and Smagorinski lateral stress coefficient * C Modules affected: timestep.f read_input.f global.f * C v44.15 - 04/16/04 - rl - from 44.13 - cleaned up implementation of all * C conservative, lateral stress, ME area integration and mass * C lumping options. Added IM values to access these. * C Added CBaroclinic logical & keyed baroclinic calculations to * C it. Moved LOGICAL declarations from sizes.f to global.f * C Added local LOGICAL CHARMV call and declaration to subroutines* C LSQSOLEG & LSQSOLVG in harm.f * C Removed CLUMP logical from global_3dvs * C Fixed VSMY so that DASigT is computed for baroclinic run. * C Modules affected: global.f read_input.f timestep.f sizes.f * C harm.f adcirc.f global_3dvs vsmy.f * C messenger.f * C v44.13 - 03/16/04 - rl - from 44.12 - fixed bug in ME TotalArea that * C that was introduced in v44.09 * C v44.12 - 03/15/04 - rl - from 44.11 - fixed bug in GWCE lateral stress * C that was present in v44.10 & v44.11 * C v44.11 - 03/13/04 - rl - from 44.10 - revamped spatially varying Tau0 * C Modules affected: read_input.f * C v44.10 - 03/09/04 - rl - from 44.09 - fixed bug in GWCE lateral stress * C v44.09 - 03/08/04 - rl - from 44.08 - changed lateral stresses in GWCE * C from Randy's method to straight forward computation * C THIS VERSION HAS A BUG CORRECTED IN VERSION 44.10 * C v44.08 - 02/25/04 - rl - from 44.07 - moved GWCE to subroutine in * C timestep.f * C Modules affected: global.f read_input.f timestep.f * C v44.07 - 12/20/03 - rl - from 44.06 - major bug fix in conservative * C GWCE formulations * C Modules affected * C timestep.f * C v44.06 - 12/17/03 - rl - from 44.05 - added 2nd nonconservative momen- * C tum Eq. (correct area integration) * C Modules affected * C timestep.f * C v44.05 - 10/19/03 - rl - from 44.04 - implemented 2nd conservative * C momentum formulation using product * C rule. * C v44.04 - 10/09/03 - rl - from 44.03 - moved 2DDI momentum calculations * C to subroutines in timestep.f module* C Modules effected * C timestep.f * C v44.03 - 10/07/03 - rl - from 44.02 - implemented conservative momentum* C equation in main timestepping code * C continued removing local variables * C from global module. * C Modules affected * C adcirc.f cstart.f global.f * C hstart.f read_input.f timestep.f* C v44.02 - 09/30/03 - rl - from 44.01 - additional changes associated * C with removing local variables from * C global module. * C v44.01 - 10/06/03 - rl - from 43.03 - changed names of load vectors in * C 2D solution * C v43.03 - 05/20/03 - rl - from 43.02 - parallel wind stuff (m.brown) * C output buffer flush (m.cobb) * C 3D fixes (k.dresback) * C drop MNPROC in fort.15 (t.campbell)* C various bug fixes in RBCs * c ZSURFBUOY/BCPG calc * C v43.02 - 02/06/03 - rl - from 43.01 - code clean up & documentation * C v43.01 - 01/31/03 - jf - from 43.00 - reconcile with v42.07 2D code * C changed var. names: NNEIGH-> * C NNeigh, NEIGH->NeiTab, * C NEIGHELE->NeiTabEle, * C UBAR1->UU1, * C VBAR1->VV1, UBAR2->UU2, VBAR2-> * C VV2, TAUSX1->WSX1, TAUSY1->WSY1, * C TAUSX2->WSX2, TAUSY2->WSY2, * C TAUBX1->BSX1, TAUBY1->BSY1, * C TAUBXN1->BSXN1, TAUBXN2->BSXN2, * C TAUBXN3->BSXN3, TAUBYN1->BSYN1, * C TAUBYN2->BSYN2, TAUBYN3->BSYN3, * C TAUBXPP3->BSXPP3,TAUBYPP3->BSYPP3* C parallel unified 2D/3D source * C v43.00b - 12/11/02 - jf - from 43.00a - bug fixes * C v43.00a - sum /02 - tc - from 36.01 (3D) & 41.12? (2D), create F90/ * C parallel unified 2D/3D source * C v42.07 - 08/12/02 - rl - from 42.06 - combined discharge & wave rad bc * C v42.06 - 06/24/02 - rl - from 42.05 - redid no normal vel grad bc v1 * C v42.05 - 06/18/02 - rl - from 42.04 - combined v1 & v2 no normal vel * C grad bcs * C v42.04 - 06/08/02 - rl - from 42.02 - fixed bug in no normal vel grad * C bc v1 * C v42.03 - 06/06/02 - rl - from 42.01 - no normal velocity grad bc v2 * C v42.02 - 06/05/02 - rl - from 42.01 - no normal velocity grad bc v1 * C v42.01 - 04/12/02 - rl - from 41.11 - redid GWCE formulation * C v41.11 - 09/14/01 - rl - from 41.10 - added NWS = -2, eliminate MNWLON * C and MNWLAT from NWS11GET * C v41.10 - 07/25/01 - rl - from 41.09 - bug fix in GWCE lateral viscosity* C v41.09 - 06/30/01 - jw - from 41.08 - minor mods per vp version 41.05 * C v41.08 - 06/22/01 - rl - reconciled v41.07 and v41.05m009 * C changes to HABSMIN and ETA2 * C v41.07 - 04/09/01 - rl - from 41.06 - initialized PRN1(),PRN2() for * C NRS<>0 * C v41.06 - 04/02/01 - rl - corrected dimensioning on most of wind reading* C subroutines. Changed MNWP to MNP in wind * C forcing ALLOCATION statements. Note MNWLON * C and MNWLAT still need to be eliminated from * C NWS11GET subroutine. * C v41.05m009 - 06/12/01 - rl&jjw - updated NODECODE after initial drying sect* C v41.05m008 - 06/01/01 - jjw - changed code by adding output in wet/dry sect* C v41.05m007 - 05/28/01 - jjw - changed HABSMIN=0.8D0*H0 * C v41.05m006 - 05/22/01 - jjw - add writes to track wetting/drying bug * C v41.05m005 - 05/21/01 - jjw - modified wet/dry to allow barrier overtopping* C modified from version 41.05.original * C v41.05m004 - 02/15/01 - jjw - from 41.05.m003 - added logic to reset min * C depth behind overtopping barriers from wetting* C v41.05m003 - 02/14/01 - jjw - from 41.05.original - skipped over version * C 41.05m001/m002. Fixed bug in wet/dry * C algorithm which prevented overtopping * C internal barriers from wetting * C v41.03 09/15/00 - rl - added bridge piling friction, fixed several * C F90 bugs * C v41.02 09/07/00 - rl - fixed F90 hot start bug and consolidated with * C version 35.xx * C v40_02m004 - 05/02/00 - rl - changed so that contribution to RHS forcing is* C zeroed out for any element that contains a dry* C node. Note: LHS is taken care of automatically* C v40.02m003 - 04/28/00 - jjw - Changed wet/dry interface from essential no * C normal & tangential to natural no normal * C v40.02m002 - 12/22/99 - jjw/vjp - Vic change to avoid compiler problems * C v40.02m001 - 12/21/99 - jjw - add cross barrier pipes cjjwm001 * C * C * C 8/7/2000 modfied jcg_with_blas -- fixed bug in itjcg for single cpu case * C 8/7/2000 modfied wetting and drying section to fix bug and optimize * C 11/1/2000 added init of TIP2 in hstart and fixed Sun makefile for Jesse * C C******************************************************************************