Land Information System (LIS) Public Release 6.2rp1 (r4421:4606)

This patch release includes updates r4421:4606.

These updates consist of:

• Merge 6/development into 6/public6.2

  I merged r4604:4605 from 6/development to 6/public6.2

  These updates consist of:

  • Update arch/configure.lis.linux_ifc because JasPer library was moved

• Merge 6/development into 6/public6.2

  I merged r4581:4603 from 6/development to 6/public6.2

  These updates consist of:

  • Update RELEASE_NOTES.txt for LIS 6.1rp7

• Merge 6/development into 6/public6.2

  I merged r4566:4580 from 6/development to 6/public6.2

  These updates consist of:

  • Merge 6/public6.1 into 6/development

    I merged r4431:4579 from 6/public6.1 to 6/development.

    These updates consist of:

    • ProTeX updates

    • Merge NU-WRF updates into 6/public6.1

      I manually merged commits r1412:1720 from the trunk of the NU-WRF repository on progress, svn/nu-wrf/code/trunk, into LIS' 6/public6.1 branch.

      Please note that even though some of these updates are for gfortran support, gfortran is not yet supported in the public release of LIS 6. These particular updates are included for completeness.

      These updates consist of:

      • Merged in updates from v6_macosx branch through r1716.

        • Replaced several pause statements in LIS code. Pause will cause LIS or WRF/LIS to hang indefinitely if executed in batch mode.
        • Changed several C files and a yacc file to #include stdlib.h instead of malloc.h. (malloc.h is non-standard and doesn't exist on OS X.)
        • Changed LIS and LVT Makefiles to greatly trim number of include paths. OpenMPI mpif90 barfed with the original list, but it turns out most paths are unnecessary.
        • Changed LIS C function mangling for gfortran so only single underscore is appended.
        • Changed LIS lib Makefiles to pass optimization flags for gfortran. Also, changed LIS_ARCH from Darwin_intel to Darwin_gfortran.
        • Tweaked ProTex comments to trim end spaces to appease gfortran.

      • Modified Fortran code to avoid use of system function to create directories.

      • Calls to 'system' fail when LIS is run with SGI MPT. Now uses customized C function that uses the POSIX mkdir function under the hood. C code includes macro to appropriately mangle the name for linking with Fortran.

      • Added new create_subdirs function to create directories.

        This is a replacement for invoking the operating system 'mkdir' command via function system. The system call fails when LIS or WRF/LIS are compiled with SGI MPT. This new C function calls the POSIX mkdir function instead.

        TODO: Handle name suffix conventions across different operating systems.

• Merge 6/development into 6/public6.2

  I merged r4432:4563 from 6/development to 6/public6.2

  These updates consist of:

  • Add Yuqiong Liu's ANSA SCF DI updates

    The following modifications are made to LIS6.2 to enable assimilation of snow cover fraction data using direct insertion:

    • added the observation plugin for assimilating ANSA SCF

      dataassim/obs/ANSA_SCF

    • modified the following file to allow assimilating SCF uisng either the standard (Rodell & House 2004 JHM) or customized (Liu et al 2013 AWR) direct insertion approaches:

      lsms/noah.3.2/da_snow/noah32_map_snow.F90

    • modified the following file in Noah3.2 to save snow density value for use in assimilation

      lsms/noah.3.2/noah32_module.F90

      lsms/noah.3.2/module_sf_noah32lsm.F90

      lsms/noah.3.2/noah32_main.F90

    • modified the following files in the plugins directory to allow ANSA SCF assimilation for Noah.3.2

      plugins/LIS_pluginIndices.F90

      plugins/LIS_DAobs_pluginMod.F90

      plugins/LIS_lsmda_pluginMod.F90

    • modified the follwoing files to allow reading in parameters for assimilating SCF using direction insertion

      core/LIS_PRIV_rcMod.F90

      core/LIS_readConfig.F90

    • modified Filepath to reflect new ANSA SCF observation plugin

    • remove the observation plugin for assimilating ANSA SCA. The new ANSA SCF is more accurate.

  • Add "#NETCDF, no HDF, GRIBAPI" flags to arch/configure.lis.linux_ifc

  • Correcting the latitude of origin value for NARR forcing.

    The correct values for the NARR grid can be found here: http://www.esrl.noaa.gov/psd/data/narr/format.html

  • Remove debugging code from Noah 3.3 read/write restart routines

  • Update length of file name

  • Add Eric Kemp's deep soil temperature TMN updates to Noah 3.3.

  • Bug fix: correct logic regarding evaluating certain conditional statements

    Fortran (nor C) does not guarantee the order in which expressions in a conditional statement will be executed. In a statement like:

    if ( X .and. Y ) then ...
    

    One does not know whether X will be evaluated first or Y. This will be a problem when Y depends on X. For example,:

    if ( present(X) .and. (X .eqv. .true.) ) then ...
    

    (X .eqv. .true.) may be evaluated first, but if X is not present, then X has no value, resulting in either a crash or undefined behavior.

    This example should be written as:

    if ( present(X) ) then
       if ( X .eqv. .true. ) then ...
    

• Merge 6/development into 6/public6.2

  I merged r4419:4431 from 6/development to 6/public6.2

  These updates consist of:

  • Bug fix: WRFout reader should read RAINNC+RAINC when RAINC is present, otherwise just RAINNC

  • Add Eric Kemp's deep soil temperature TMN updates to Noah 2.7.1 and Noah 3.1

  • Add check to determine whether LIS_rc%tbot_update_lag was enabled

  • Merge NU-WRF updates into 6/public6.1

    I manually merged commit r1411 from the trunk of the NU-WRF repository on progress, svn/nu-wrf/code/trunk, into LIS' 6/public6.1 branch.

    These updates consist of:

    • Merged in v6_lisplugins branch

    • Bug fix: Moved call to LIS_createTmnUpdate after LIS_domain_init.

      This fixes a bug in coupled WRF-LIS where the TmnUpdate pointer is allocated before the number of tiles is set by LIS. It does not affect uncoupled LIS.

    • Bug fixes to reading dynamic deep soil temperature data from LIS restart file.

    • Added bug fixes to output tempbot for earlier Noah LSM versions when requested.

    • Bug fix. Now allows LIS to write tempbot to history file when requested.

    • Bug fixes to dynamic deep soil temperature option.

    • Added soil temperature adjustment module. Fixes to restart code.

    • Initial implementation of WRF deep soil temperature adjustments in WRF.

      Option #1 adjusts climatological deep soil temperature to model terrain height using standard atmospheric lapse rate (6.5 K / km). Always used in REAL.

      Option #2 parameterizes the deep soil temperature as the weighted average of (a) the annual mean skin temperature from the previous calendar year, and (b) the mean daily skin temperature from the most recent N days of the simulation, where N is set by the user (140 days is recommended). The equation is from Salathe et al, J. Climate, 2008:

      Tbot = alpha*mean_annual_skinT_prev_year +
             (1-alpha)*mean_daily_skinT_prev_N_days
      

      where alpha is 0.6. These values of alpha and N are tuned to published observed values of 30% amplitude damping and 70-day lag from skin temperature (per Salathe paper).

      Time series and sums are saved to update these terms at the end of each day (right term) or at the end of the year (left term). I added code to write and read these fields to/from LIS restart files. Note that if LIS is cold started, the climatological deep soil temperature is initially used for both terms and for the N-sized daily time series.

      NOTES:

      • This code compiles but is not test run yet.
      • Only supports Noah 3.2 LSM.
      • Should support either LIS retrospective or WRF/LIS mode.

    • Cherry-pick r4123 from LIS 6/development branch

    • Bug fix regarding interpolating precip for WRFout forcing reader

      WRF writes precipitation accumlated from the beginning of the simulation. Therefore compute the difference between the bookends to find the accumulated precipitation for the hour.

    • Changed code for setting wrfout file names.

      First version of this code was hardwired for original IFLOODS configuration, where some wrfout files were created 9 or 18 seconds past the hours. This is now changed to remove "special rules" and instead just look for files at the expected standard times (e.g., 60 minutes at top of the hour for WRF).

      Users with time steps that prevent wrfout writes at those exact time should create symbolic links to match the date/time filenames expected by LIS.

    • Modified time steps in LIS LSM options to pick correct domain value.

      Note that time step logic for writing history or restart files is undisturbed.

    • Several bug fixes to address WRF/LIS coupling with nests.

      1. CRITICAL Changed LIS Noah.3.2 LSM to use proper domain timestep rather than the smallest of the timesteps for all domains.

      2. TMN (Deep soil temperature) is now passed from WRF to LIS. Requires following entry to forcing_variables.txt:

         TMN:        1  1  K       # Soil Temperature at Lower Boundary
         

         Further work is needed on this, as the TMN used in a LIS spin-up is not necessarily the same as that provided by REAL.

      3. Minor bug fix to cosz (cosine of zenith angle) calculation input to LIS. Currently cosz is not used, but it is passed to the LIS Noah LSM and presumably will be used in the future.

    • Merge LIS 6.1 updates (r4067:4090) into NU-WRF

      • This brings the version of LIS within NU-WRF up to 6/public6.1 r4090.

      • Change build system to include LIS water points again.

      • Now use landusef and the LIS tiling algorithm to pick dominant land cover.

      • Tweak WPS LANDMASK to reconcile with dominant soil type and land cover.

      • Restored bug fix for writing netCDF output header.

      • Bug fix to read shdmin and shdmax from WPS file.

      • Added routine to fetch MMINLU attribute from WPS file.

      • Restored implementation of WRFout suppforcing plugin.

      • Added WPS shdmax and shdmin plugins. Compiles.

      • Merge LIS 6.1 updates (r3885:4066) into NU-WRF

        This brings the version of LIS within NU-WRF up to 6/public6.1 r4066.

      • Second round of fixes. LIS now runs with WRFout plugin.

      • Added LIS as a separate compile target.

        This code is a bit unsettled -- LIS is built, but I don't know what will happen if both WRF and LIS are requested as targets.

      • Added WRFout plugin to LIS.

• Update RELEASE_NOTES.txt for LIS 6.2r public release

  Highlight that Cable 1.4b is restricted distribution.

Land Information System (LIS) Public Release 6.2r (r3524:4420)

This is the initial release of the public version of LIS 6.2.

This release includes updates r3524:4420.

• Includes VIC 4.1.1.
• Includes CABLE 1.4b --- restricted distribution.
• Includes Catchment F2.5.
• Includes Noah 3.3.
• Includes SiB2.
• Includes WRSI.
• Support for North American Mesoscale Forecast System (NAM) "242 AWIPS Grid -- Over Alaska" product.
• Support for USGS potential evapotranspiration (PET) data (for use in WRSI).
• Support for Climate Prediction Center's (CPC) Rainfall Estimates version 2 (RFE2) daily precipition (for use in WRSI).
• Support to apply lapse-rate correction to bottom temperature field (for use in Noah).

Land Information System (LIS) Public Release 6.1rp7 (r3884:4599)

• Disable LIS_FORC_CO2

  Support for this forcing variable is not available in LIS public 6.1.

• Revert r4578 update to Makefile

• Add "#NETCDF, no HDF, GRIBAPI" flags to arch/configure.lis.linux_ifc

• Correcting the latitude of origin value for NARR forcing.

  The correct values for the NARR grid can be found here: http://www.esrl.noaa.gov/psd/data/narr/format.html

• ProTeX updates

• Merge NU-WRF updates into 6/public6.1

  I manually merged commits r1412:1720 from the trunk of the NU-WRF repository on progress, svn/nu-wrf/code/trunk, into LIS' 6/public6.1 branch.

  Please note that even though some of these updates are for gfortran support, gfortran is not yet supported in the public release of LIS 6. These particular updates are included for completeness.

  These updates consist of:

  • Merged in updates from v6_macosx branch through r1716.

    • Replaced several pause statements in LIS code. Pause will cause LIS or WRF/LIS to hang indefinitely if executed in batch mode.

    • Changed several C files and a yacc file to #include stdlib.h instead of malloc.h. (malloc.h is non-standard and doesn't exist on OS X.)

    • Changed LIS and LVT Makefiles to greatly trim number of include paths. OpenMPI mpif90 barfed with the original list, but it turns out most paths are unnecessary.

    • Changed LIS C function mangling for gfortran so only single underscore is appended.

    • Changed LIS lib Makefiles to pass optimization flags for gfortran. Also, changed LIS_ARCH from Darwin_intel to Darwin_gfortran.

    • Tweaked ProTex comments to trim end spaces to appease gfortran.

    • Modified Fortran code to avoid use of system function to create directories.

      Calls to 'system' fail when LIS is run with SGI MPT. Now uses customized C function that uses the POSIX mkdir function under the hood. C code includes macro to appropriately mangle the name for linking with Fortran.

    • Added new create_subdirs function to create directories.

      This is a replacement for invoking the operating system 'mkdir' command via function system. The system call fails when LIS or WRF/LIS are compiled with SGI MPT. This new C function calls the POSIX mkdir function instead. TODO: Handle name suffix conventions across different operating systems.

• Wrap long lines to appease Portland Fortran compiler

• Update testcases for new "TBOT terrain adjustment option:"

• Add Eric Kemp's bottom temperature lapse-rate correction

• Add Soni Yatheendradas' TRMM 3B42V6 updates to TRMM 3B42RT

• Add Soni Yatheendradas' TRMM 3B42V6 updates

  From Soni Yatheendradas:

  Thanks to Yudong and Jonathan Case for helping out with debugging.

  Per my e-mail 06/20/13:

  TRMM data file time stamp is every 3Z, e.g., 6Z, 12Z etc, so for any file, it’s time validity is from 1.5 hrs before to 1.5 hrs after.

  Different tests on the earlier TRMM reader revealed the following:

  1. For runs where (a) time steps had their edges touching the breakpoints, e.g., 30 min and 15 min, and (b) the LIS clock at the end of the 1st time step was not 0Z and was not within the time starting at the breakpoint before 0Z (e.g. not 23.5Z, meaning the lis.config did not start @23Z for a 30-min time step): your run is fine.

  2. For runs where (a) time steps had the breakpoints inside them at their mid-point, e.g., 1 hr and 20 min, and (b) the LIS clock at the end of the 1st time step was not 0Z and was not within the time starting at the breakpoint before 0Z (e.g. not 23Z, meaning the lis.config did not start @22Z for a 1-hr time step): your run was incorrect, since only alternate TRMM files were read, e.g, 0Z, 6Z, 12Z etc. instead of correctly reading 0Z, 3Z, 6Z, 9Z, 12Z etc.

  3. In 1.(b) above, if now the LIS clock at the end of the 1st time step was not 0Z but was within the time starting at the breakpoint before 0Z (e.g. 23.5Z, meaning the lis.config started @23Z for a 30-min time step): that and only that 1st time step actually read a TRMM data file that was exactly one day behind than what it should have read.

  4. In 2.(b) above, if now the LIS clock at the end of the 1st time step was not 0Z but was within the time starting at the breakpoint before 0Z (e.g. 23Z, meaning the lis.config started @22Z for a 1-hr time step): In addition to the error mentioned above in 2 where every alternate file was read, the 1st time step would also have read a TRMM data file one day back as in 3. above.

  Now, as mentioned in 2. above, there were time steps that had the breakpoints inside them at their mid-point: these should logically have taken the averaged value of the TRMM data files before and after. The current code always read in and implemented only one TRMM file data, never an average.

  The entire complex decision structure in get* is now replaced with a much simpler one: for every time step, get the time instants of the beginning and end of that time-step, or time-marks. These time-marks combined with the TRMM data time stamps at or beyond either end of these time-marks now give the weights to average as necessary.

  Per Jon Case & my respective e-mails 04/22/13 & 06/24/13: Inside the corrected glbprecip_TRMM3B42RT.F90 file that John Case sent, the precip_regrid was initialized to -1.0, and towards the end of the glbprecip_TRMM3B42RT subroutine, the suppdata was set to precip_regrid conditional on "(precip_regrid(i,j) .ne. -1.0)". Yudong thinks that it's probably better to have both the initialization and the condition have LIS undef value instead of 1.0. This difference has not been tested, and John Case may have done these to make the min/max values came out right in the GRIB1 files.

  Finally, the averaging of rain that changes in the midst of a time step implemented here might need to be implemented later for other similar precip readers where only value of rain @ the beginning or end of such time step is applied over whole time step.

• Adjust SNODEP reader alarm Set the SNODEP reader alarm to read the hour 12 SNODEP observations at hour 12 not at 11:45.

• Update the legalese at the top of each file

• Bug fix: correct type of local variable

• Update Noah 3.1 Bondville test-case to emphasize that it is a point comparison

• Usersguide updates

• Bug fix regarding interpolating precip for WRFout forcing reader

  WRF writes precipitation accumlated from the beginning of the simulation. Therefore compute the difference between the bookends to find the accumulated precipitation for the hour.

• Update testcases to reflect updates to LIS_histDataMod

• Bug fix: referencing an array out of bounds

• Bug fix: WRFout reader should read RAINNC+RAINC when RAINC is present, otherwise just RAINNC

• Clean up comments

• Add Eric Kemp's deep soil temperature TMN updates to Noah 2.7.1 and Noah 3.1

• Add check to determine whether LIS_rc%tbot_update_lag was enabled

• Merge NU-WRF updates into 6/public6.1

  I manually merged commit r1411 from the trunk of the NU-WRF repository on progress, svn/nu-wrf/code/trunk, into LIS' 6/public6.1 branch.

  These updates consist of:

  • Merged in v6_lisplugins branch

    • Bug fix: Moved call to LIS_createTmnUpdate after LIS_domain_init. This fixes a bug in coupled WRF-LIS where the TmnUpdate pointer is allocated before the number of tiles is set by LIS. It does not affect uncoupled LIS.
    • Bug fixes to reading dynamic deep soil temperature data from LIS restart file.
    • Added bug fixes to output tempbot for earlier Noah LSM versions when requested.
    • Bug fix. Now allows LIS to write tempbot to history file when requested.g
    • Bug fixes to dynamic deep soil temperature option.
    • Added soil temperature adjustment module. Fixes to restart code.

  • Initial implementation of WRF deep soil temperature adjustments in WRF.

    • Option #1 adjusts climatological deep soil temperature to model terrain height using standard atmospheric lapse rate (6.5 K / km). Always used in REAL.

    • Option #2 parameterizes the deep soil temperature as the weighted average of (a) the annual mean skin temperature from the previous calendar year, and (b) the mean daily skin temperature from the most recent N days of the simulation, where N is set by the user (140 days is recommended). The equation is from Salathe et al, J. Climate, 2008:

      Tbot = alpha*mean_annual_skinT_prev_year +
             (1-alpha)*mean_daily_skinT_prev_N_days
      

      where alpha is 0.6. These values of alpha and N are tuned to published observed values of 30% amplitude damping and 70-day lag from skin temperature (per Salathe paper).

      Time series and sums are saved to update these terms at the end of each day (right term) or at the end of the year (left term). I added code to write and read these fields to/from LIS restart files. Note that if LIS is cold started, the climatological deep soil temperature is initially used for both terms and for the N-sized daily time series.

      NOTES:

      • This code compiles but is not test run yet.
      • Only supports Noah 3.2 LSM.
      • Should support either LIS retrospective or WRF/LIS mode.

  • Cherry-pick r4123 from LIS 6/development branch

  • Bug fix regarding interpolating precip for WRFout forcing reader

  • WRF writes precipitation accumlated from the beginning of the simulation. Therefore compute the difference between the bookends to find the accumulated precipitation for the hour.

  • Changed code for setting wrfout file names.

    • First version of this code was hardwired for original IFLOODS configuration, where some wrfout files were created 9 or 18 seconds past the hours. This is now changed to remove "special rules" and instead just look for files at the expected standard times (e.g., 60 minutes at top of the hour for WRF). Users with time steps that prevent wrfout writes at those exact time should create symbolic links to match the date/time filenames expected by LIS.

  • Modified time steps in LIS LSM options to pick correct domain value.

    • Note that time step logic for writing history or restart files is undisturbed.

  • Several bug fixes to address WRF/LIS coupling with nests.

    1. CRITICAL Changed LIS Noah.3.2 LSM to use proper domain timestep rather

       than the smallest of the timesteps for all domains.

    2. TMN (Deep soil temperature) is now passed from WRF to LIS. Requires

       following entry to forcing_variables.txt:

       TMN:        1  1  K       # Soil Temperature at Lower Boundary
       

       Further work is needed on this, as the TMN used in a LIS spin-up is not necessarily the same as that provided by REAL.

    3. Minor bug fix to cosz (cosine of zenith angle) calculation input to LIS. Currently cosz is not used, but it is passed to the LIS Noah LSM and presumably will be used in the future.

  • Merge LIS 6.1 updates (r4067:4090) into NU-WRF

    • This brings the version of LIS within NU-WRF up to 6/public6.1 r4090.

  • Change build system to include LIS water points again.

  • Now use landusef and the LIS tiling algorithm to pick dominant land cover.

  • Tweak WPS LANDMASK to reconcile with dominant soil type and land cover.

  • Restored bug fix for writing netCDF output header.

  • Bug fix to read shdmin and shdmax from WPS file.

  • Added routine to fetch MMINLU attribute from WPS file.

  • Restored implementation of WRFout suppforcing plugin.

  • Added WPS shdmax and shdmin plugins. Compiles.

  • Merge LIS 6.1 updates (r3885:4066) into NU-WRF

    • This brings the version of LIS within NU-WRF up to 6/public6.1 r4066.

  • Second round of fixes. LIS now runs with WRFout plugin.

  • Added LIS as a separate compile target.

    This code is a bit unsettled -- LIS is built, but I don't know what will happen if both WRF and LIS are requested as targets.

  • Added WRFout plugin to LIS.

• Update SiB2 testcase

  Write snowt as an instantaneous value.

• Bug fix for writing netCDF output header

• Changes to formatting of the testcases for consistency.

  No output changes to the testcases will occur.

• Enable WRFout supplemental forcing reader

• Increase FT_NUM_SUPPFORCE to 40 to support new supplemental forcing readers

• Add WRFout supplemental forcing reader

  Add the WRFout supplemental forcing reader from 6/development r4021.

  This adds the support for using WRF output as the LIS forcing (in support of iFloods project).

• Merge NU-WRF updates into 6/public6.1

  I manually merged commits r1266 and r1349 from the trunk of the NU-WRF repository on progress, svn/nu-wrf/code/trunk, into LIS' 6/public6.1 branch.

  These updates consist of:

  • Modified WRF and LIS code to pass snow depth, snow liquid water equivalent, and snow coverage from LIS back to WRF. Currently only the LIS Noah LSMs are supported.
  • Changed a number of variables from scalars to 1-D arrays.
  • Subroutine sib2_routines expects a number of arguments to be arrays. Compiling with '-gen-interfaces -warn interfaces' triggered errors. Note that the answers should not change -- when sib2_routines is called, the arrays are assumed to be of size 1, equivalent to scalars. This is only technically a bug, not a major problem.

• Modified the setting of the data in the "mndy" array in SiB-2's astro routine. No change to the output. The change was made to satisfy very strict compile settings.

• Update testcases to reflect new lis.config option: "TBOT data source:"

• Add support to write TBOT to output file when running TEMPLATE lsm

  This requires defining the TBOT_TESTING macro found in the template_main.F90 file.

• Add WPS testcase

• Update LIS history writer

  Add support to write albedo and max snow albedo parameters into LIS history files.

• Add support to read WPS-based parameter data

  These updates were manually copied from branches/untested/p61_NDHMS r3763.

• Bug fix: set q for both monthly and quarterly albedo

• Bug fix: correct read_WPStexture routine

  This routine was incorrectly using SOILCTOP to determine soil texture.

  SOILCTOP is 16-category top-layer soil type fractions. This routine was treating it as a 0-1 map of each category.

• Added updated SiB-2 LSM testcase.

• Small fix to the CLSM-F2.5 testcase README file and MODEL_OUTPUT_LIST.TBL file for consistency.

• Update testcases to reflect updates to LIS_histDataMod

• Final code modifications for correcting the SiB-2 implementation within LIS-6.

• Continued to update the SiB-2 LSM code for the latest version of LIS-6/development.

• Moved SiB2 directory to be consistent with other LSMs, and started to make the code more consistent with other LSMs as well. Starting to update code for latest LIS.

• Added sample lines to the MODEL TBL for the new ALMA variables.

• Reorganized LIS_histDataMod.F90 so the variables are closer to the ALMA output order and grouping.

  Added the following output variables to LIS_histDataMod.F90:

  WaterTableD, LWup, SAlbedo, SnowTProf, TotSoilCarb, TotLivBiom

  Added the follwing outputs to CLSM-F2.5:

  WaterTableD, LWup, Salbedo, SnowTprof, TWS (Terrestrial Water Storage)

• Fixed several lines of code to ensure that latitude and longitude are consistently used in the setting of the domains and tiles in many of the projections. Bugs were found and fixed that inconsistenly used array values that mixed between latitude and longitude.

  Many thanks to Kristi Arsenault for finding the issue and identifying a solution.

• Set some undefined variables in noah33_lsmMod.F90, and made some other small changes so all Noah-3.+ versions have consistent code.

• Change VIC forcing index

  I changed the VIC forcing index, LIS_vicforcingId, from 16 to 31.

  LIS_vicforcingId = 31
  

  The old value of 16 conflicted with the index for Stage IV precipitation, LIS_stg4Id.

• Change standard name of output variable snow cover from Snowcover to SnowCover

  This is used when creating the NetCDF header.

• Bug fixes for MERRA baseforcing.

• Adjust NLDAS2 forcing domain definition

• Fixed a bug which used undef data out of their respective domains (e.g., higher than 50 for 3B42V6).

• Correct output of SMLiqFrac

  Small fix to Noah 2.7.1 to correct the output of SMLiqFrac to make it consistent with Noah 3.2 and Noah 3.3.

• Update lis.config documentation

  Add documentation for the following NLDAS2 supplemental forcing options:

  NLDAS2 data center source:
  NLDAS2 use model based pressure:
  

Land Information System (LIS) Public Release 6.1rp6 (r3850:3883)

• Resolve DA plugin omissions

Land Information System (LIS) Public Release 6.1rp5 (r3735:3849)

• Correct LIS_MOC_ROOTMOIST output units

• Correct running domain specification for Noah 3.1 Bondville testcase

• Update catchment testcase

  The lis.config options "tile coord file" and "tile veg file" were renamed to "Catchment tile coord file" and "Catchment tile veg file", respectively.

• updates and cleanups to the perturbation schemes

• small updates to the DA subsystem (in preparation for the tutorial)

• manually add the lsms/noah.3.2/da_soilm files

• Add VU_AMSREsm, MODISsca, and ANSA_SNWD support

  Manually add VU_AMSREsm, MODISsca, and ANSA_SNWD support from r3816 6/development to 6/public6.1.

• Remove syntheticlst from 6/public6.1

• Update lis.config

  Changes to lis.config file for inclusion in Users Guide.

• Remove old data assimilation test cases from 6/public6.1

• Add synthetic soil moisture ensemble kalman filter test case

• Set gmaoenkf_struc%forecast_var = LIS_rc%udef

  The EnKF diagnostic variable gmaoenkf_struc%forecast_var was not set when Observations%assim was false. When Observations%assim is false set gmaoenkf_struc%forecast_var = LIS_rc%udef.

• Comment out syntheticlst.

  This was removed from 6/public6.1.

• Add additional GrADS descriptor file

• Add SWE DA support to Noah 3.2

• Add synthetic snow water equivalent (swe) ensemble kalman filter test case

• Apply dynamic observation error

  These are Yuqiong Liu's updates to apply dynamic observation error to GMAO ENKF.

Land Information System (LIS) Public Release 6.1rp4 (r3515:3734)

• Bug fix: upscaling code did not properly check mask and neighbours

• Bug fix: remove extraneous ssdev argument from call to LIS_readObsAttributes

• Bug fix: reset negative GDAS precip rates to 0.

  The hrf06 file is a 6 hour forecast from hr. So it contains the events that also occurred during the 3 hour forecast from hr, namely, hrf03. Thus hrf06 contains the precip from hrf03 plus any additional precip that occurred in the following 3 hours.

  Should (2*hrf06 - hrf03) yield negative values for precip, then we assume that there was no additional precip for the last 3 hours of that 6-hour forecast and that the run of the 6-hour forecast produced slightly different results for the first 3 hours (as compared against the hrf03 forecast).

  When LIS is using values from an hrf06 file, then LIS has already run over the period that corresponds to the first 3 hours of that 6-hour forecast. LIS needs data corresponding to the last 3 hours. Thus always reset negative precip rates to 0.

• Add Joe Santanello's porosity updates from NU-WRF r1177.

  Log message from NU-WRF:

  Porosity is now saved in LIS for NOAH 2.7.1, NOAH 3.1, NOAH 3.2, and CLM2 LSMs, exported back to WRF, and passed to GOCART dust emission module. Original porosity values from NOAH or RUC LSMS are used if LIS is not selected as the LSM.

• Bug fix: TRMM 3B42RT routines were not correctly interpolating and checking precip rates

  Set the logical bit mask, lb, used by the spatial interpolation routines. Check precip_regrid against the undefined value -9999.

• Bug fix: GDAS was producing negative precip rates

  pcp comes from the averaged GDAS forcing data. In some cases, it is derived from the hrf06 and hrf03 values ( 2*hrf06 - hrf03 ).

  One would expect the 2*hrf06 precip to be greater than or equal to the hrf03 precip, but in some cases it is slightly smaller. When the 2*hrf06 precip is slightly smaller than the hrf03, one gets a small negative precip rate. Trap this small rate and reset it to 0.

• Fixed a bug in the call to "SIBALB". The call incorrectly sent the snow depth variable "SNDZ" before the snow water equivalent variable "WESN" in the list of arguments. They have now been switched to the correct order as expected in the subroutine SIBALB.

  Further research revealed that this bug existed in the Catchment driver as well (outside of LIS). This bug was discovered/fixed in the Catchment driver in October 2010.

  This bug is an issue only when running on the Catchment domain, and does not affect simulations run in other domains (lat/lon, etc.).

• Changed code to require adding "Catchment" to the two following lines in the "lis.config" file:

  tile coord file:
  tile veg file:
  

  Now the lines in the "lis.config" file must be:

  Catchment tile coord file:
  Catchment tile veg file:
  

• Fixed the following three items in Catchment's code to transfer forcing to the tile:

  1. Catchment now correctly sets precipitation as snowfall when the forcing air temperature is less than freezing. Previously, precipitation would be set as snowfall when the air temperature was less than or equal to freezing. The change was made to be consistent with the Catchment offline driver.
  2. When the precipitation is snowfall, the code now sets the convective rainfall to zero. Previously, the code was not setting this value to zero, leading Catchment to calculate a negative value for large-scale rainfall.
  3. The code now checks that the convective rainfall is not more than the total rainfall.

• Add Jiarui Dong's core polar stereographic and hrap projections updates

• Bug fix: fix allocation for array param_dd

  I increased the size of param_dd from 6 to 8.

• Bug fix to cat_main.F90 by Hiroko Kato

  The LIS_gfrac(nid)%greenness(:) array is a 1-d array in tile-space. It was being looked up as if it were in 1-d grid-space.

  The catchment routine expects an array of catchment tile id's (cat_id) to be passed in. This cat_id array is not used, but it was not initialized. This has been corrected.

  Index into cat_struc(nid)%cat_param(:) for LIS_MOC_SOILMOIST and LIS_MOC_ROOTMOIST output was fixed.

• Small fixes to Noah3.1, Noah3.2, and Noah3.3 to correct the output of SMLiqFrac, and to fix the output warning message to the lisdiag file for the default minimum greenness value.

Land Information System (LIS) Public Release 6.1rp3 (r3124:3514)

This patch release includes updates r3124:3514.

These updates consist of:

• Bug fix: WRF-LIS coupling; only advance LIS' clock when running the innermost nest.

  The previous revision of this file advanced LIS' clock on every call. This caused time within LIS to run ahead of the time within WRF. This is a workaround that keeps the innermost nest in sync with WRF; however, outer nests in LIS will be one outermost timestep behind WRF.

  In WRF each domain/nest has its own clock. WRF first solves its outermost nest, advancing the outer clock time from t1 to t2. Then WRF solves its inner nest, advancing that clock from t1 to t2. When passing time from WRF into LIS, LIS sees time advancing forward and then receding backward.

  Attempts to correct this issue within LIS by resetting LIS' clock have failed. I am not able to get the ESMF alarms to ring at the proper times after I have reset LIS' clock. The alarms either do not ring, ring too soon, or ring too often.

• Add zlib to the LDFLAGS for Linux IFC systems

• Updates for compiling on IBM/AIX systems.

• Update NLDAS1 and NLDAS2 testcases

  I reset the "NLDAS1 data center source:" and the "NLDAS2 data center source:" options to be consistent with the nldas and nldas2 testcases currently available on the LIS website.

• Update Noah output

  I added Chris Franks' minimum rh and relative soil moisture output updates to Noah 3.1.

  They were already in Noah 2.7.1 and Noah 3.2.

• Update Noah output

  I have added the CH, CM, and mixing ratio output updates from Noah 2.7.1 into Noah 3.1 and Noah 3.2.

• Bug fix: noah31%q1 was used without being set.

• Bug fix: correct the label for Noah's q1.

  I incorrectly labelled Noah's q1 as relative humidity instead of mixing ratio.

• Bug fix: AIX does not like the function style system call.

• Broke several long lines into shorter lines to appease the PGI compiler.

• Enhancement to make reading 3B42RT compressed files (.gz) more reliable.

• Enhancement to make reading CMORPH compressed files (.Z) more reliable. Tested with parallel runs, with compressed and uncompressed data files producing identical results.

• Checking in Kristi's changes for the NLDAS-1 and NLDAS-2 supplemental forcing code to be able to read in forcing data from the GES DISC. The GES DISC version of NLDAS forcing datasets have a different naming convention, directory structure, and (one different) GRIB ID value; however, the values of the fields are the same as that currently on the LIS project disk and/or from the NCEP/EMC ftp site.

  Note that the lis.config also changes for running with NLDAS-1 forcing: "NLDAS" has been renamed as "NLDAS1" for the lines in the card file.

• Added support to read .Z file, and dealt with CMORPH file name change in Jun. 2008.

• Corrected the moving of bookends.

• Bug fix (Joe Santanello). Converted units of input precipitation. Precipitation from WRF is in mm. LIS, in particular Noah, expects precipitation to be a rate (mm/s).

  I convert the units from mm to mm/s during the import of WRF input data.

• Added modifications to output carbon related variables using the ALMA standard. Variables include: GPP, NPP, NEE, AutoResp, HeteroResp, and LeafResp.

  Also added the output of Rnet (net surface radiation).

• I've implemented LAI->GVF for real-time MODIS LAI data, using a scheme given by Niu et al. (2011). Comparing MODIS_LAI/AVHRR_GVF:

  https://lis.gsfc.nasa.gov/PMM/OUTPUT-CMEM3-1KM/lnsr-noah32-nldas2/E314-lsm-stats-sm.gif

  with MODIS_LAI/MODIS_GVF:

  https://lis.gsfc.nasa.gov/PMM/OUTPUT-CMEM3-1KM/lnsr-noah32-nldas2/E514-lsm-stats-sm.gif

  one can see the new scheme is more self-consistent, and the GVF range between the two is very similar.

  Note: A correction has to be made to urban LAI values, to make sure LAI over urban can not be less than 0.1. This is because Noah hard-coded a GVF of 0.05 to urban-class. This change will only be made when Noah3.2 is being used.

  Ref:

  Niu, G.-Y., et al. (2011), The community Noah land surface model with multiparameterization options (Noah-MP): 1. Model description and evaluation with local-scale measurements, J. Geophys. Res., 116, D12109, doi:10.1029/2010JD015139.

• Changes made:

  1. Replaced macro "#if (defined CRTM)" to "#if (defined RTMS)", to control compilation of all RTM codes. Updated make/LIS_misc.h accordingly.
  2. In "core/LIS_laiMod.F90", added consistency check between greenness fraction and LAI values: if gfrac > 0, LAI can not be < 0.1. This prevents crash in Noah's CANRES() function. Also since the check is done here, Noah lsm code does not need to be modified.

• Reverted to revision 3135. Moved the LAI fix to noah32_main.F90 as David suggested.

• The following changes were made, to support reading real-time MODIS 8-day LAI data:

  1. Added a minimal LAI value check in lsms/noah.3.2/module_sf_noah32lsm.F90 right before the call to "CANRES", to make sure LAI value is not 0. Otherwise, CANRES() will crash.
  2. Refactored LAI readers under params/lai/, so now each lai reader has its own directory, and a "setup" interface.
  3. Added and implemented real-time MODIS LAI reader: params/lai/MODIS_RT.
  4. Added support for intervalType < 0 in core/LIS_timeMgrMod.F90. If intevalType < 0, the interval will be set as -intervalType days. This is to support MODIS's 8-day interval, and maintain backward compatibility.
  5. Cleaned up those "#if (defined CRTM)" macros in rtms/CMEM3/ and rtms/CRTM2EM/.
  6. "Filepath" file updated to reflect changes made in 2 above.

• Added option to output SMFrozFrac in volumetric.

Land Information System (LIS) Public Release 6.1rp2 (r3056:3124)

This patch release includes updates r3056:3124.

These updates consist of

• Corrected bareclass and waterclass definitions in the routine to read in MODIS_IGBP landcover type dataset, and made consistent with UMD and USGS landcover type routines.

  Corrected the index number to determine the longitude used to calculate change of day and interpolation of the Noah 3.1/3.2 greenness fraction; this change only applies to simulations where the greenness values are specified in "lis.config". Simulations of LIS-Noah that use greenness maps from LIS are not affected by this change.

  Added code to set the greenness fraction to zero in the "dynsetup" routines for bare, snow, and water classes in the Noah 3.1/3.2 LSMs. This code already was contained in the "setvegparms" routines, and is now extended to every timestep (as also indicated in the NCAR Noah driver). Made the "dynsetup" and the "setvegparms" routines consistent where possible between the Noah 3.1 & 3.2 LSMs.

• Saving the background and the actual roughness length from previous timestep to use in the call to the SFCDIF routine. This change will only affect offline non-coupled runs when also not using exchange coefficients from the forcing data. Only a very minor change is found in the simulated output.

• Fixed a bug to check for using forcing height data from the forcing file for uncoupled offline runs. Does not affect runs where the forcing height is specified in the "lis.config" file.

• Corrected LIS_MOC_COUNT.

• adding two new variables to the LIS output (diagnostic T2 and diagnostic Q2)

• Updated configure.lis.linux_ifc file to use ESMF 3.1.0rp3 library compiled against Intel MPI 4.0.1.007-beta. This is the version that was recommended for the NU-WRF project.

Land Information System (LIS) Public Release 6.1rp1 (r2883:3055)

This patch release includes updates r2883:3055.

These updates consist of

• Fixed dlon/dlat issue. This only affects non-square lat/lon domains.

  For the lat/lon projection:

  griddesc(9) is dlon

  griddesc(10) is dlat

• Extended the LIS I/O to work beyond just the LSM output

• Updates to the gribwriter testcase.

• Bug fix. Min/max values were not correctly computed in diagnoseDataEntry. This is a recent bug. Min/max diagnosing was correct in Increment 2.

• Mainly cosmetic changes to the code, but also including changes to make more consistent between all Noah versions.

• Renaming Noah 3.1 ESMF coupling routines to avoid a naming conflict with the non-ESMF coupling routines.

• Bug fix. LIS was not writing NetCDF output when running in parallel.

• Bug fix. In coupled mode with WRF, Noah3.2 was overwritting the ch (and cm) exchange coefficient. This was fixed. In coupled mode, Noah 3.2 is using the ch (and cm) values from WRF.

• I am committing Hiroko Kato's ECMWF updates.

• Updated Noah 2.7.1 to be consistent with Noah 3.2 in its treatment of CH and CM.

  This only affects running Noah 2.7.1 in an offline/uncoupled running mode.

  CH and CM are initialized in either the coldstart or restart routine. They are not reset to 1e-4 before every call to sfcdif.

  CH is not reset to undef in every call to noah271_f2t.

• Bug fix.

  Increased the maximum size of the grid%subgrid_tiles array to LIS_rc%maxt*LIS_rc%nensem.

• Bug fix.

  Ryan Ruhge found an error regarding computing the weights for the conservative interpolation scheme.

  This bug fix corrects the computation of the weights for the conservative interpolation scheme.

• Noah 2.7.1, Noah 3.1, and Noah 3.2 must always call sfcdif, either to compute exchange coefficients or to compute ustar.

  This update corrects the issue where the there versions of Noah were not always calling sfcdif.

• Corrected slopetype check.

  The Noah 3.1 and 3.2 setsoils routines were checking the value of slopetype against -9999 to determine whether the point was land or water. This is incorrect. The correct check is to compare slopetype against 0 (0 == water).

• History support updates.

  Corrected standard deviation computation. Corrected minimum/maximum computation.

  Removed unused feature to specify minvalue and maxvalue.

  Added subgrid_tiles to the grid data structure. This allows a given grid-cell to know which tiles are contained within it.

• Added extra information (iv number) to an error message.

• Reverted an update to noah271_coldstart. The coldstart routine should initialize ch and cm to 1e-4.

• turned off the hardcoded soil moisture wilting point in Noah 2.7.1

• Catchment bug fixes. Thanks to Hiroko Kato.

  I replaced "LIS_rc%domain == 8" with "LIS_rc%domain == LIS_catdomainId"

  I corrected the setting of cat_struc(n)%cat_param%gnu. It should be 2.17 not 2.71.

• Cosmetic changes to the code to improve readability, in-line documentation, and consistency between the 2.7.1, 3.1, and 3.2 versions of the Noah LSM. Changes will have no affect on model results.

• Bug fix for running nests in non-esmf coupled mode.

  This bug fix frees several allocated arrays and sets the LIS_initialized flag.

• small updates to Noah 3.2 for parameter estimation

Land Information System (LIS) Public Release 6.1r (r2702:2882)

This is the initial release of the public version of LIS 6.1.

This release includes updates r2702:2882.

• Includes Noah 3.1.
• Includes Noah 3.2.
• Support for SPoRT Daily GVF data.
• Support for North American Regional Reanalysis (3d) (NARR) data.
• Support for NCEP's modified IGBP MODIS landcover data.
• Support to specify direction for output variables.

Land Information System (LIS) Public Release 6.0rp8 (r2899:2966)

• Bug fix: retnldas2a was deallocating some variables twice.
• Also, this commit contains some ProTeX updates.

Land Information System (LIS) Public Release 6.0rp7 (r2702:2898)

• This is a back port of bug fixes from 6/public6.1 to 6/public6.0. It incorporates updates up to r2895.

  These updates consist of:

  • Fixes for compiling on IBM AIX.

  • Fixes for compiling with gfortran.

  • Fixes for compiling with Lahey.

  • Fix for dlon/dlat issue. This only affects non-square lat/lon grid-cells. For the lat/lon projection:

    griddesc(9) is dlon

    griddesc(10) is dlat

  • Fix for diagnosing Catchment snowdepth.

Land Information System (LIS) Public Release 6.0rp6 (r2593:2701)

• I manually merged NLDAS2 updates from 6/development into 6/public.

• I corrected the use of griddesc(9) and griddesc(10) for lat/lon projections.

  griddesc(9) specifies dlon.

  griddesc(10) specifies dlat.

  These array elements were incorrectly set in getgeos. I also reset the value of mi after the GEOS4 and GEOS5 grid changes in getgeos.

  These array elements were incorrecly used in gdasLSWG_forcingMod.F90.

  These array elements were incorrecly used in compute_grid_coord_latlon.F90.

• In-line documentation corrections.

• I added Clay Blankenship's bug-fix regarding setting dlon in get_fieldpos.

• I replaced the use of a hard-coded unit number with the use of LIS_getNextUnitNumber and LIS_releaseUnitNumber.

• The transition to the T574 grid occurred on 12z28jul2010.

Land Information System (LIS) Public Release 6.0rp5 (r2554:2592)

• I corrected a bug regarding writing GRIB output on a Mercator project.

Land Information System (LIS) Public Release 6.0rp4 (r2491:2553)

• Bug fix.

  The transition to the T574 grid occurred on 12z28jul2010.

• I corrected in-line documentation.

• Bug fix.

  LIS incorrectly specified the T574 grid change to be on 12z27July2010. The T574 grid change occurs on 12z28July2010.

Land Information System (LIS) Public Release 6.0rp3 (r2408:2489)

• Added a check to trap when czenith (cosine of zenith angle) becomes larger than 1.
• Redeclared several indices to be integer*8 to prevent overflows when running at high resolution.
• Updates and bug fixes to NASA AMSRE soil moisture reader for data assimilation Consolidated HDF and binary versions of NASA AMSRE
• Switched to using V06 of NASA L3 retrievals
• Updates to several testcases.
• Jon Case's T574 GDAS and GFS forcing updates.
• Tools for generating ensemble restarts (upscale and downscale)
• Bug Fix: The GRIB module was not correctly setting the time-stamp for years before 2000. This update fixes this problem.
• Some changes to CLM2 to remove some hardcodes
• Bug fix: I am adding Hiroko Kato Beaudoing's Princeton forcing update. This fixes a bug regarding interpolating the Princeton forcing data.

Land Information System (LIS) Public Release 6.0rp2 (r2397:2406)

• Bug fix. I corrected an error regarding writing output with nested domains.

Land Information System (LIS) Public Release 6.0rp1 (r2255:2396)

• Corrected the indexing into varname when LIS_FORC_Forc_Hgt has been selected.

  Commented out an "if (.false.)" block to facilitate compiling with Absoft.

• Eric Kemp's bug fix regarding an overflow error when reading landcover data at 2km resolution.

• Bug fix. Corrected the intent of a variable in suppforcing/nldas2/retnldas2a.F90.

• Added support to run makdep on .f90 files.

• I merged the Noah forcing height updates from 6/development to 6/public.

• I merge some NASA_AMSRE updates from 6/development.

• Added a new rule to the Makefile.

• changes to the code to support the use of bottom soil temperature as the WRF input in the ESMF coupled version

• updated GFS baseforcing implementation

• Protex updates.

  I have added files named 0Intro_*.txt. This are Protex files that serve as introductions to various sections of the Reference Manual.

• LIS6-WRF3.1.1 tight coupling updates.

• Bug fix for GRIB testcase.

• Bug fix. lwd was not properly set in time_interp_gdas.

  Fix to ensure LIS output contains forcing variables when run coupled to WRF

• changed the I/O interfaces to be more flexible.

  The forcing_variables.txt file and the MODEL_OUTPUT_LIST.TBL files need to contain only variables of interest (they do not need to contain the exhaustive list of variables)

• Bug fixes to GDAS for properly treating instantaneous and time averaged fields

• I have added Hiroko Kato Beaudoing's catchment update.

  Her update is "a minor bug fix for LIS6/catchment output variables. Units fixes from W/m2 to kg/m2/s for canopy evaporation and transpiration. Also, "runoff" which is a total of surface and subsurface runoff was output as surface runoff, so it's changed to "runsrf"."

• This is the first pass at cleaning up the output support.

  It creates an array of pointers to the output variables. This replaces much of the code, needed to explicitly list the output variables, with do loops. This makes it easier to add new output variables to LIS.

  TODO: NetCDF support still needs some more clean up.

• added a bug fix so that open unit numbers are properly closed even when errors are encountered in the file I/O

• added a bug fix for the monthly alarm settings (when day 15 is selected as the starting time of the LIS simulations, the monthly alarms do not ring)

• added Xice to the forcing_variables.txt file.

• added lat/lon fields to the netcdf output