RELEASE NOTES for CAMx v6.50, 04/30/18 The Comprehensive Air Quality Model with extensions (CAMx) is available at http://www.camx.com. Overview of Version 6.50 ------------------------ V6.50 includes new features, modifications, and bug fixes from the previous release (v6.40) * Use the chemistry parameters files specifically labeled for v6.5. New files with new formats are available to support chemistry updates. * The CAMx control namelist file has changed to support certain modifications. * CAMx can now generate output files in netCDF (v3 or v4 with optional compression). The original Fortran binary I/O format remains the default option. * The CAMx makefile now supports: 1) the option to compile with netCDF libraries 2) the free Sun Oracle Fortran compiler with OMP support Type "make" or "make help" for more information. NOTE: RECENT VERSIONS OF THE GFORTRAN COMPILER WILL NOT COMPILE CAMx WITH OMP PARALLELIZATION BECAUSE OF AN INCOMPATIBILITY WITH BLOCK DATA IN ISORROPIA. THIS IS NOT AN ISSUE WITH OTHER SUPPORTED COMPILERS. OPTIONS: a) USE MPI ONLY WITH GFORTRAN b) REVERT BACK TO AN OLDER VERSION OF GFORTRAN c) USE A DIFFERENT COMPILER * There are no changes to other I/O file formats from v6.40. v6.50 New Features and Major Updates ------------------------------------ 1. CAMx now supports netCDF for gridded output files, including the core model average concentration and deposition files, and Probing Tool output. Implications: netCDF is an option to the original Fortran binary I/O format. CAMx can be built with netCDF-3 or netCDF-4/HDF5; the latter allows for data compression using "data chunking" -- use default chunking settings, or test variations to check compression impacts on your system. See the Makefile instructions on how to build CAMx with netCDF libraries. Our implementation of uncompressed netCDF (v3 or v4) is compatible with Models-3 I/O API convention without the need to install and link with I/O API libraries. 2. The EQSAM ("4clim" version) for inorganic gas-aerosol partitioning has been added as an alternative to ISORROPIA. Thanks to Swen Metzger at The Cyprus Institute and ResearchConcepts for supplying the code and assisting in its evaluation. Implications: EQSAM is based on analytical parameterizations, resulting in slightly faster CAMx run times in our tests (4-7%). PM concentration differences from ISORROPIA will depend on specifics of each application. We encourage tests comparing results from both. 3. Chemical Process Analysis (CPA) has been extended to all gas-phase mechanisms. Implications: The inclusion of CPA calculations into the gas-phase mechanisms was originally done by hand, which was time-intensive and subject to coding errors. These calculations are now integrated into the Chemical Mechanism Compiler, which automates the generation of all mechanism solver code and CPA variable calculations. Many CPA variables are now derived using the same calculations as OSAT, improving the consistency between these Probing Tools. 4. The CAMx Makefile now supports the Oracle/Sun Fortran compiler with OMP. Implications: This free compiler adds to the list of supported compilers for CAMx (PGF, IFORT, GFORTRAN, ABSOFT, ORACLE). v6.50 Modifications ------------------- 1. A dry deposition parameter for ammonia (NH3) has been changed to more appropriately estimate NH3 removal rates. Implications: The parameter "Rscale" in the CAMx chemistry parameters files has been reset from "0" to "1". The value strongly influences NH3 deposition velocity by including (1) or ignoring (0) surface resistance. Several tests with Rscale = 1 have shown large increases (improvements) in NH3 concentrations. We recommend conducting sensitivity tests with Rscale = 0/1 for NH3 to understand its impact on your particular application. 2. Precursor gas emissions for SOAP organic aerosol chemistry are now automatically assigned from input CB05, CB6 or SAPRC07 VOC emissions as the model runs, removing the burden on the user to externally duplicate VOC emissions to make SOA precursor emissions. Implications: Up to now, CAMx required a duplicate set of VOC emissions for certain species for use in the SOAP scheme (TOL->TOLA, ISOP->ISP, etc.). This step was burdensome and often errantly skipped, leading to zero SOA production. Now the burden is removed and consistency between VOC and SOA precursor emissions is ensured. 3. The treatment of Fe and Mn catalyzed aqueous SOx reactions has been reverted back to the original approach (prior to v6.40). Implications: The updated metal-catalyzed approach introduced in v6.40 (scaling Fe and Mn from FRCS+FPRM, etc.) significantly reduced SO4 production, leading to under predictions. The original approach (setting Fe and Mn concentrations to a constant value) is now the default until updates can be made and tested for a later CAMx release. The scaling approach in v6.40 remains in the code and can be used with a simple parameter change. 4. SOAP2 has been updated with a new photolysis reaction (SOAP2.1). Only SOAP2 yields, saturation concentrations and polymerization rates introduced in v6.4 are available (original SOAP1 parameters have been removed). Implications: Particle-phase SOA photolysis was implemented assuming a photolysis lifetime of approximately one day. The new removal pathway, together with stronger SOA production based on wall- corrected chamber yields introduced with SOAP2, led to better agreement with observations in our tests. 5. Aerosol pH is now available as a part of CPA probing tool outputs providing better accessibility for the user. Implications: Aerosol pH, calculated by the thermodynamic equilibrium model ISORROPIA, is now tracked through the model and output as a CPA output variable (AER_PH). 6. ISORROPIA now employs the "metastable" liquid aerosol state assumption, instead of allowing a solid phase based on the deliquescence relative humidity. Implications: Under many atmospheric conditions, aerosols are more likely in the metastable liquid state due to the hysteresis effect, and recent ISORROPIA evaluation studies have shown that the metastable assumption produced good agreement with field observations. At very low RH, however, this assumption may not apply. 7. The chemistry parameters file header has been restructured to more clearly specify choice of gas-phase mechanisms, aerosol size scheme, and inorganic PM and SOA schemes/versions. Implications: These updates more clearly define the choice among the many new gas and PM chemistry treatments. 8. Chemistry mechanisms CB6r2h and SAPRC07TC have been extended to include source apportionment. Implications: SA may now be run with CB6r2h and SAPRC07TC. For CB6r2h, ozone decay from reactions with oceanic halogens has been added. For SAPRC07TC, SA has been updated from SAPRC99, but the SOA tracer class is not available and will be updated in a later CAMx release. v6.50 Bug Fixes --------------- 1. Fixed a bug in the RTRAC wet deposition algorithm. Implications: A check for the RTRAC surface model was missing, which led to CAMx crashing with a segmentation fault if the surface model was not invoked when running RTRAC. 2. Fixed a bug in the RTRAC dry deposition algorithm. Implications: The land/ocean mask update in v6.40 was not extended to the RTRAC dry deposition algorithm, leading to CAMx crashing with a segmentation fault when running RTRAC. 3. A check on the memory size of Probing Tool tracers has been added to avoid model crashes and cryptic fortran error messages. Implications: This check stops the model with an error message when the Probing Tool tracer array dimension exceeds 2^31 values, the maximum size of a single-precision integer. Since this value is much smaller than available machine memory, we plan to extend the Probing Tool array dimension to a double-precision integer in future releases. 4. Two infrequent bugs were fixed in MPI passing of data among MPI slices. Implications: (1) In certain rare circumstances and for very large applications, message passing of boundary conditions was impacted, leading to incorrect results relative to non-MPI runs. (2) A fix alleviated segmentation violations when running MPI with a large number of PiG puffs.