CHANGELOG

7.0.4 (2019-06-20)

    GUI
    - Fix creation of log file directory at startup
    - Fix time-lag dialog double popup during file uploading

7.0.3 (2019-06-08)

    Engine
    - Add: new output file conforming FLUXNET and AmeriFlux's FP-In Standards
      forlabels, units and format FLUXNET output features a large amount of
      new output variables, including many intermediate results, statistics on all
      high-frequency variables, covariances between all high-frequency variables,
      biomet mean values and metadata, as well as results of new implementations
      (see below)
    - Add: new data filtering tool allowing user to eliminate raw records
      corresponding to istantaneous wind directions from specified sectors
      (up to 16 sectors)
    - Add: new Stationarity test after Mahrt (1998)
    - Add: new Random Uncertainty estimation method after Mahrt (1998)
    - Add: new QC test named KID (Kurtosis Index on Differenced variable) test
      (Vitale et al., in preparation)
    - Add: new test to identify period of repeated data and drop-outs
      (Vitale et al., in preparation)
    - Add: new supported anemometers:  Metek uSonic-3 Class A MP and
      Metek uSonic-3 Cage MP
    - Add: computation of wind direction variance
    - Add: ET is now treated as a first-class flux (results only available in
      FLUXNET output)
    - Refine: improve computation of statistical moments which is now fully
      robust to presence of missing data
    - Refine: more accurate computation of mean wind direction
    - Refine: Potential Radiation is now computed used timestamp of END of
      period instead of START
    - Remove: obsolete GHG-Europe and AmeriFlux outputs
    - Fix: minor fixes to Random Error and Footprint settings

    GUI
    - Add: Fluxnet output settings
    - Add: visual tool for wind directions filtering settings to specify up to
      16 sectors
    - Add: random uncertainty estimation method after Mahrt (1998)
    - Add: new supported anemometers: Metek uSonic-3 Class A MP
      and Metek uSonic-3 Cage MP
    - Change: move random uncertainty settings from
      [RawProcess_RandomUncertainty_Settings] to [Project] managing backward
      compatibility
    - Remove: obsolete GHG-Europe and AmeriFlux outputs
    - Remove: previous data results setting

6.2.2 (2018-05-11)

    Engine
    - Add: support for Young 81000V, 81000RE, 81000VRE sonic anemometers.

    GUI
    - Add: support for Young 81000V, 81000RE, 81000VRE sonic anemometers.
    - Add: references to the Tovi platform.
    - Fix: Raw File Settings persistency.
    - Fix: local help not showing.

6.2.1 (2017-10-09)

    Engine
    - Add: support for new LI-7500DS instruments.

    GUI
    - Add: support for new LI-7500DS instruments.
    - Fix: magnetic declination retrieval from NOAA.

6.2.0 (2016-08-23)

    Engine
    - Add: new supported variable "anemometer_diagnostic". If selected,
      it filters anemometer data if the diagnostic value is different from zero.
    - Add: now biomet variables that are provided without positional qualifiers
      are by default given a "null qualifier" in the form of _0_0_n, with n
      progressing from 1. This applies now also to external biomet files, which
      means that users no longer need to assign 'fake' positional qualifiers to
      variables
    - Add: support of Gill WindMaster and WindMaster Pro "w-boost"
      correction, necessary to fix firmware bug for firmware versions:
      [2329.x.y with x<700].
    - Refine: AoA correction policy. Now Nakai et al. 2006 is applied only to
      R3 family and R2. Nakai and Shimoyama 2012 is applied to WindMaster and
      WindMaster Pro. The combination of "w-boost" and AoA correction for
      the WindMaster/Pro is handled automatically by EddyPro. Effectively,
      the user as the options of (1) applying or not the w-boost correction
      alone for WindMaster/Pro; (2) applying or not Nakai's AoA corrections.
      However, EddyPro controls that both corrections are applicable to the
      current anemometer and overrides user's settings (and output
      corresponding warnings) if necessary.
    - Refine: now planar fit settings are not read if planar-fit file is used.
      The useless reading operation did not cause any problems.
    - Refine: changed "time lag" into "time-lag" everywhere
    - Fix: shut down BA/ZOH correction, which is currently implemented but not
      exposed via the GUI.
    - Fix: Kelvin/Celsius conversion factor harmonised everywhere (273.15 instead
      of some occurring 273.16)
    - Fix: now biomet data is not read in during planar fit (which was never
      useful anyway) and this fixes an array allocation issue that was causing
      a crash when both planar fit and biomet data were used.
    - Fix: eliminate 'allocatable' property for TempFact in drift_correction.f90,
      which is however not yet enabled, so it was causing no issues.
    - Fix: missing time-lag optimization file when requested caused tilt correction
      method to be set to inexistent value and thus default to no correction
    - Fix: flags for absolute limits is now correctly calculated. Flags for
      this test obtained so far are unreliable.
    - Fix: output of full (co)spectra. The bug was such that there was a
      mis-alignment of one data point between frequencies and associated
      (co)spectra values. The bug did not affect calculation and output of
      binned cospectra, nor usage of full cospectra in spectral corrections.
    - Fix: now files with timestamp referring to end of the dataset are
      processed correctly. The bug did not allow processing them in rather
      unusual (but not impossible) situations, such as when files shorter
      than the flux averaging interval were processed.
    - Fix: handle inconsistent selection of embedded biomet data and non-GHG
      file type. This pair of inconsistent settings is not possible with the GUI,
      only with manual editing of the .eddypro file.
    - Fix: definitions of Hamming and Hann windows were mistakenly inverted.
      Now they are correct. The difference between the two windows is anyway
      minimal and resulting binned (co)spectra are thus only minimally
      affected, and only at the low frequency end.

    GUI
    - Fix: make non blocking a zero flow rate gas analyzer description.
    - Fix: make sure that an ignored variable in the metadata raw file description
      table has almost all the fields disabled.
    - Fix: toolbar with no labels in initial maximized state.
    - Fix: inconsistent value of the data set available date on the detect
      range dialog showing when running in advanced mode.

    - Refine: make sure that on Mac is possible to load any kind of external
      biomet files.
    - Refine: remove italic from anemometer firmware version.
    - Refine: trim spaces reading variables in external biomet files.

    - Add: set the Angle of Attack checkbox default setting to disabled.
    - Add: "Build Continuous Dataset" to the Minimal Advanced Output Settings
      (the current default).
    - Add: some textual hint in the metadata tables.
    - Add: accept Gill Windmaster/Pro two or three fields firmware versions.
    - Add: change max plausibility range from 12 to 20 sigma as per user request.
    - Add: enable drag and drop tips on Windows.
    - Add: the registered trademark symbol to the main window title.
    - Add: master Anemometer Flags combo in Basic Settings page.
    - Add: support for Campbell CSAT-3B.
    - Add: raise variable flags decimals from 4 to 10 digits.
    - Add: make anemometer firmware version mandatory for Gill WindMaster/Pro.
    - Add: checkbox for Gill's w-boost bug.

6.1.0 (2015-12-14)

    Engine
    - Add: support for new LI-7500RS and LI-7200RS instruments.
    - Refine: introduced a mechanism to allow user to ask EddyPro to select
      the most appropriate AoA correction method, based on sonic anemometer
      model and serial number. User can also force a selection, or de-select
      the correction completely.
    - Fix: FLUXNET biomet output file now contains biomet results also when
      fast EC data are not available (similar to standard biomet output
      file). Also fixed initialization, so that if biomet data is not
      available, FLUXNET biomet output file reports -9999.
    - Fix: detection of EndIndex in the list of binned spectra files when a
      subperiod is selected, which coincides with the period covered by
      available spectra files. The bug caused EddyPro not to create ensemble
      averaged spectra/cospectra, nor to perform the spectral assessment.
    - Fix: SplitCount function, that caused biomet files not to be properly
      read (and engine to crash) when only one biomet variable is present in
      the external biomet files, besides the timestamp(s).

    GUI
    - Add: support for new LI-7500RS and LI-7200RS instruments.
    - Refine: add automatic selection for the AoA correction method.
    - Refine: set the default AoA correction method to automatic selection.
    - Add: embedded software version (firmware) field in the anemometer
      description table in the metadata editor.
    - Refine: enable embedded software version (firmware) field for any
      gas analyzer and change it to a free text form.
    - Refine: allow project file opening with double-click on Mac.
    - Fix: spectral file assessment load button not working properly.
    - Fix: essential file path not properly preserved during a run on Windows.
    - Refinement: Update QuaZIP to version 0.7.1 (2015-01-07).

6.0.0 (2015-08-05)

    Engine
    - Fix: now all dynamic metadata records are taken into account (including
      the last one that was incorrectly ignored before) and all dynamic
      metadata are retrieved correctly.
    - Fix: array deallocation missing. Bug signalled by Sebastien Lafont
      (INRA, France). Thanks for the contribution!
    - Fix: now processing.eddypro file is copied correctly in the output folder
      (Linux/Mac only; not a problem on Windows).
    - Fix: bug causing execution to stop with error if a 'fast temperature'
      measurement was used, and from sonic SoS was available, not sonic
      temperature.
    - Fix: now output directory is always created correctly and no spurious
      'none' folder is created within the folder containing the executables.
      This was not causing any problem in any likely scenario, but could cause
      it if running in embedded mode on a desktop.
    - Refine: criteria for (co)spectra filtering are now explicitly set by user.
    - Refine: settings for (co)spectra filtering are now independent from the
      selection of the spectral correction method and are the same for ensemble
      averaging, model fit and discrimination between model and direct method
      in spectral correction after Fratini et al. (2012).
    - Refine: support for biomet data. Now any suffix triplet is
      supported and there is no more limit to the occurrences of the same
      variable (within the maximum number of biomet supported, which is 250.000).
      Greatly simplified source code, which is now fully scalable.
    - Refine: now names of external biomet files no longer need to reflect the
      chronological order in their alphabetic order. The detection of biomet
      files is based only on the extension and the chronological order is
      derived by reading the data.
    - Refine: now perform a stricter test on raw file names. To be considered
      for processing, raw files must match the 'Raw file format' completely,
      with the only exception of asterisk ('*') in the raw file format, that is
      considered a place-holder for any character.
    - Refine: now, when both 'Select a sub-period' in Basic Settings
      page and 'Build continuous dataset' in Output page are selected,
      all output files will span the selected time period regardless of
      the period covered by raw data files.
    - Refine: now biomet output and full output always span the same
      time period. However, alignment row-by-row is guaranteed only when
      option 'Build continuous dataset' is selected. Otherwise, alignment
      will depend on data availability.
    - Refine: now empty spaces in labels of custom variables are replaces
      with underscores.
    - Refine: de-spiking algorithm of Vickers and Mahrt (1997) is now more
      robust against error values in raw data.
    - Refine: now WPL interface only allows to select whether to apply the
      correction or not. The way it is applied is decided internally,
      depending on the type of analyser (open or closed-path), the type of
      high-freq concentration data (molar density, mole fraction, mixing
      ratio) and the availability of other high-frequency data.
    - Refine: eliminated creation of unused 'log' sub-directory.
    - Refine: if run ends without processing any averaging period, eliminate
      temporary files initialized by current run in output folder.
    - Refine: files names of any length (with the file systems' limit of
      255 bytes) are now supported. Paths up to 1024 bytes (4 times the
      Microsoft recommended path lengths) are now supported.
    - Add: if multiple external biomet files are used in a biomet directory,
      a new test checks that variable number, order and units is
      consistent among the different files. If not, resorts to not
      using biomet data.
    - Add: in the context of validation of metadata file, introduce
      some sanity checks on numeric values such as altitude, canopy height etc.
      Introduce default values in case implausible values are found.
    - Add: now all output files contain a suffix indicating whether the run
      was performed in Express mode ('exp'), Advanced mode ('adv') or
      Metadata Retrieval mode ('mdr').
    - Add: support for Mac OS.
    - Add: now de-spiking method after Mauder et al. (2013) is available as an
      alternative to the default method of Vickers and Mahrt (1997) in the
      Statistical Analysis page.
    - Add: now ensemble averaged spectra output also includes de-noised spectra.
      If noise elimination is not requested by user, de-noised spectra are
      equal to original spectra.
    - Refine: now "Missing Sample Allowance" operates not only on full records,
      but also on individual variables, so that any given variables (and related
      flux results) is set to error if for that variable more data than the
      missing allowance is missing, regardless of the records available for
      other variables. This is most likely the behavior of "Missing samples
      allowance" expected by most users.
    - Refine: introduced control on linear detrending time constant,
      automatically set equal to the flux averaging interval if user set it
      larger than that.
    - Refine: now GHG-Europe output complies with new specifications of
      AmeriFlux and European Carbon databases. EddyPro creates two files
      - one for EC-results and one for biomet results - that can be directly
      submitted to the databases.
    - Refine: now RP writes path of essentials file on stdout so that GUI can
      safely retrieve it and write it on the .eddypro file. RP keeps writing it
      on the .eddypro file too, as before.
    - Add: binned (co)spectra filtering based on results of micro-meteorological
      tests. User can select whether to discard (co)spectra when the
      corresponding period is flagged as moderate and/or low quality.
    - Refine: more explicit subperiod start/end indication when importing
      binned cospectra files, to clarify overlap between available (co)spectra
      files and the selected subperiod.

      GUI
    - Add: built and packaged the Mac OS X version.
    - Add: graphic design review.
    - Add: implement drag and drop capabilities in the file system browsing
      widgets.
    - Add: implement warning level guided mode messages, for example to notify
      when IRGA separations and flow rate are not greater than zero.
    - Add: new checkbox 'Ensemble averaged spectra' in the output settings page.
    - Add: introduce handling of file name format in case of GHG files.
    - Add: new setting for (co)spectra filtering according to Mauder and
      Foken, 2004.
    - Add: new despiking method after Mauder et al. 2013.
    - Refinement: change label of units in Turbulent Fluctuations time constant
      from [m] to [min].
    - Refinement: verify that is possible to run multiple instances of the
      program at the same time.
    - Refinement: set default time lag test to the nominal values in the
      Raw file description.
    - Refinement: remove 7zip dependency to read/write zip files.
    - Refinement: extract 'spectral' settings from spectral corrections, make
      it more general and review spectral analysis page.
    - Refinement: clean the metadata file path opening a project with an
      inexistent metadata file.
    - Refinement: review WPL correction interface.
    - Refinement: improve detection of start/end date of dataset with
      'Detect dataset dates'.
    - Refinement: remove setting automatically the Angle of Attack correction.
    - Refinement: set a maximum limit to 'Missing samples allowance' of 40%, as
      already implemented in the Engine.
    - Refinement: improve update notification message reminding the
      SmartFlux update too.
    - Refinement: review policies for start/end dates settings.

5.2.1 (2015-01-29)

    Engine
    - Fix: bug causing the engine to hang before run completion,
      if no averaging period could be processed completely. The bug
      was unlikely to affect EddyPro-desktop users, but could at
      times show up in SmartFlux.

    GUI
    - Add: Save and Continue buttons to Spectral, Planar fit and
      Time lag assessment files testing dialog.
    - Fix: Time lag assessment file testing crash in case of missing
      RH classes.
    - Fix: Planar fit assessment file testing too restrictive.
    - Fix: 4th gas molecular weight and diffusivity automatic management.
    - Fix: magnetic declination retrieval from NOAA website was broken
      due to changes in the website response.

5.2.0 (2014-11-20)

    Engine
    - Added: Analytic spectra correction after Massman (2000, 2001).
    - Added: Now biomet outputs are provided also for periods when
      corresponding high-frequency data are not available or are corrupted.
    - Added: eddypro_rp now edits .eddypro file, so that eddypro_fcc can be
      launched when eddypro_rp completes, without further interventions.
      This improves portability and facilitate usages from command line and
      automation.
    - Added: option to run engine with command-line specificied .eddypro
      file (full path). See help for usage (run "$eddypro_rp -h")
    - Fixed: bug causing off-season uptake correction to switch to
      'simple linear' and stick there, if radiation measurements were
      missing for just one half-hour.
    - Fixed: management of external biomet files (1 file, files in folder,
      1 minute time-step).
    - Fixed: bug causing TOB1 files without header (rare case) to be deleted
      from the data folder, if they happened to contain less data than expected.
    - Fixed: bug causing soil temperature not to be duly treated as a
      biomet variable.
    - Fixed: call to CorrectionFactorsIbrom07 in BPCF_Fratini12. The wrong call
      prevented use of backup solution (Ibrom model) in periods of too-low
      fluxes. So, the correction was always using the direct method of
      Fratini et al. 2012, also in low-flux conditions.
    - Fixed: functioning of metadata retriever, that got broken after
      refactoring for 5.0.
    - Fixed: bug in footprint model of Kormann and Meixner 2001, as implemented
      in 'eddypro_fcc' (no corresponding bug in 'eddypro_rp')
    - Fixed: missing exception could cause timelag optimization to associate a
      timelag instances to first RH class, which were erroneous. The problem
      would occur if, for a given period, not enough H2O data were available.
    - Fixed: bug causing cross-wind correction to be applied to
      'fast temperature' measurements, but only if those were used in
      conjunction with an R2 anemometer.
    - Fixed: exit condition for subs importing ASCII data. Was crashing in some
      conditions, if a file contained more records than expected.
    - Fixed: module call to m_common_global_var in RenameTmpFilesCommon.
      The bug was not causing problems because the same module is called by
      the module that was mistakenly called.
    - Fixed: duplicated biomet outputs when raw biomet data are missing.
    - Refinement: eliminated squared root operating on the transfer function H,
      in Fratini et al. (2012) spectral correction method. This refinement
      shall be made publicly available in the form of a corrigendum to the
      paper. Thanks a lot to Johannes Laubach for suggesting the change,
      demonstrating its foundation and verifying the new implementation in
      EddyPro.
    - Refinement: minimum timelag searching window for H2O in timelag optimizer
      reduced from +/-0.7s to +/-0.3s, which fits better to short-tube set-ups.
    - Refinement: refined format of GHG-Europe output file after
      specifications from European Database managers:
        * Eliminated date/time columns: now only ISOdate remains as a timestamp
        * Eliminated WS and WD
        * Eliminated raw file name from output strings
        * Eliminated Pa, Ta, RH from file to avoid conflicts with meteo files
        * Added ISO basic format timestamp
    - Refinement: in timelag optimization session and in main raw data
      processing session, moved call to retrieve_sensor_params before time-lag
      estimations to have instruments properties available for refined
      "starting time-lags" definition.
    - Refinement: moved 'normal exit instruction' in all subs importing raw data,
      in such a way that the exception of reaching the end of the file is caught
      in most (possibly all) occasions.
    - Refinement: implementation of random error after Finkelstein and
      Sims (2001):
        * Improved code quality and readability.
        * Improved speed in calculating the ITS.
    - Refinement: revised retrieval of files in list, to safely skip files that
      do not comply the raw file name format.
    - Refinement: eliminated N = N - 1 at the end of sub ReadSLTEddySoft.
    - Refinement: completely rewritten subroutine to import full cospectra and
      modified approach in FCC to get the size of cospectra from the first file
      instead of allocating in advance.
    - Refinement: improved control over missing lines, with an additional control
      after statistical screening, so that period is skipped if not enough data
      remain after the screening, also including native missing data such as -9999.
    - Refinement: improved format and function calls related to exception
      handling and corresponding output messages.
    - Refinement: Simplified ShowDailyAdvancement and introduced advancement
      marks on output, to allow for GUI estimates of processing time.
    - Refinement: revised the text displayed when running engines with option -h.
    - Refinement: in "filelist_by_ext", anticipated exception handling if no
      files are found in folder.
    - Refinement: proper setting of FileEndReached in ReadSLTEddySoft and
      ReadSLTEdisol.
    - Refinement: adjusted algorithms to calculated default timelags, which now
      include "tube time", "cell time" and a "safety margin".
    - Refinement: eliminated useless "+1" on both sides of a "greater than" test,
      in subs importing all binary files (TOB1, generic binary).
    - Refinement: eliminated useless "+1" on both sides of a "greater than" test,
      in subs importing SLT files.
    - Refinement: Removed the usage of auxiliary folder ..\mod in both RPP and
      FCC projects, to ease and streamline the port to other OS.
    - Refinement: in case of space separation in raw file with non-numeric field,
      defaults to considering multiple consecutive separations as only one.
    - Refinement: calculation of cutoff frequencies and dependence cutoff/RH
      when data for very few RH classes are available.
    - Refinement: Moved definition of 'null' spectra from fx to common module.
    - Refinement: moved function NameMatchesTemplate into
      src_common/dir_sub.f90 file.
    - Refinement: eliminated useless output 'Verifying time series integrity'
      during planar fit data import.
    - Refinement: corrected the spelling of subroutine ShrinkString.

    GUI
    - Improve information available in Run page (for example, show warnings and
      errors, computate time estimates, etc).
    - Improve progress bars in Run page.
    - Add separation of Standard variables and Custom variables in the Metadata
      editor Raw file description. Provide a Clear button for the Custom
      variables.
    - Integrate Massman 2000/2001 correction.
    - Implement testing of ancillary files (Spectral correction assessment,
      Planar fit and Time lag optimization) at loading time.
    - Add a dialog to inform about the Angle of attack automatic selection.
    - Add an 'Open output directory' button in Run page.
    - Add keyboard shortcuts to navigate between pages (Alt+1...Alt+5).
    - Fix minor bugs.

5.1.1 (2014-03-04)

    Engine
    - Fixed: bug causing ignoring next raw file when current file is shorter
        than expected.
    - Fixed: bug causing 'full co-spectra' files not to be read in 'eddypro_fcc'.

5.1.0 (2014-01-31)

    Minor update mostly for bug fixing and refinement.

    Engine
    - Fixed: import of TOB1, no longer copies the TOB1 file multiple times and
      speeds up processing of long TOB1 files (typical use case).
    - Fixed: bug that was causing storage fluxes to be identically zero in most
        use cases.
    - Fixed: file in src_common was importing module rp-related (did not cause
        computation issues).
    - Refinement: modified absolute limits statistical test defaults as follows:
        * Minimum sonic temperature set to -40 C (was -20 C).
        * Maximum CO2 set to 900 ppm (was 600 ppm).
    - Refinement: fixes to ranges of accepted biomet values:
        * Changed minimum ambient pressure to 40 kPa (was 80 kPa).
        * Made all interval min/max closed instead of open,
          most important for RH, to include values 0 and 100.
    - Refinement: optimized the output time stamp in case
        of shorter-than-expected raw files.

    GUI
    - Fixed: behavior of the progress bar.
    - Fixed: decimal digits of the 4th gas molecular values were lost in the
        Basic settings page when saved.
    - Refinement: improved information available in Run page.
    - Refinement: added default button selection to SMARTFlux create package
        dialog.
    - Refinement: modified absolute limits statistical test as listed above

5.0.0 (2013-12-09)
For a complete list of new features see:
http://www.licor.com/env/support/EddyPro/topics_eddypro/Whats_New.html

    Introduced support to the package creation for SMARTFlux systems.

    Engine
    - Fixed: bug causing crash on short raw files.
    - Fixed: bug causing the software not to read strings longer than 200 characters
        in .eddypro file.
    - Fixed: potential problems with timestamps in non-ISO format.
    - Fixed: bug causing crash if last raw file of the dataset is not a valid one
        (too short, not valid header, etc.).
    - Fixed: bug causing the software not to process correctly averaging intervals
        shorter than file length in certain circumstances.
    - Fixed: use dynamic metadata files. Same variables as usual, but now full
        independence of processed time period from definition of dynamic metadata
        dates.
    - Fixed: bug causing the spectral correction of Fratini et al. (2012) to virtually
        always use the fallback solution (model) rather than the direct method
        (thanks Olli Peltola and Ivan Mammarella!).
    - Fixed: bug causing the spectral assessment to fail if a large number of spectra
        files were to be used.
    - Refinement: changed units of ET fluxes in the full output file. New units is
        mm+1hour-1.
    - Refinement: changed Express settings. Cross-wind correction is no longer
        applied by default.
    - Refinement: exception handling with SLT-EdiSol files if header reports
        implausible record length.
    - Refinement: merged and simplified spectral correction code.
    - Refinement: period to be processed extended to include the very last
        averaging interval, which was excluded from the automatic selection of
        start&end date performed in the GUI.
    - Refinement: initialization of stats and their values when variables are missing.
    - Added: Creation of unique temp folder for allowing parallel runs.
    - Added: possibility to process indefinitely long raw files. Possibility to process
        datasets for multiple years in one session.
    - Added: smart understanding of initial timestamp in the dataset.

    GUI
    - Implemented SMARTFlux bar and file package creation button.
    - Removed 'Number of files to merge' control.
    - Removed gas analyzer height in the Metadata editor.
    - Introduced 'Open Sans' font.

4.2.1 (2013-10-04)
    Minor update that fixes a GUI bug related to the retrieval of declination correction
    for magnetic north from the U.S. National Oceanic and Atmospheric Administration
    (NOAA) website. The update corrects an error that occurs when the website is
    unavailable.

4.2.0 (2013-08-22)
    For a complete list of new features see:
    http://envsupport.licor.com/help/EddyPro4/Default.htm#Whats_New.htm

    Engine
    - Fixed: use of "Flags" (set in the "Basic Settings" page) for filtering out individual
        raw data records.
    - Fixed: import of SLT-EddySoft raw files featuring low-resolution data.
    - Fixed: formulation of tube transfer function for laminar regimes and - only in FCC -
        for turbulent regimes.
    - Fixed: problem with Fratini et al. 2012 spectral correction giving NaN in case of high
        fluxes.
    - Fixed: bug in the running mean and exponential running mean detrending
        algorithms.
    - Fixed: bug in planar fit with no velocity bias that caused rotations not to be
        performed.
    - Fixed: bug in time lag optimizer causing a crash.
    - Fixed: initialization of footprint results for Kormann and Meixner (2001) and
        Hsieh (2000) models.
    - Fixed: labels of custom variables in full output, when created by FCC program
        (added "_mean" padding to each variable name).
    - Fixed: Management of missing variables when passing from RP to FCC.
    - Fixed: Using flow rate from raw data files when available, with both GHG and
        non-GHG files, if time lags are not explicitly set by user.
    - Fixed: bug that caused mismatch in full output headers when advanced processing
        settings were used.
    - Fixed: missing LI-7700 CH4 flux output in full output file when data is screened
        using LI-7700 Diagnostic Value.
    - Fixed: wind coordinates rotation when AXIS configuration (GIll sonics) or R2
        anemometer is selected, for angle-of-attack correction.
    - Addition: added ET (evapotranspiration flux, in mm) to full output.
    - Addition: optional cross-wind correction of sonic temperature for Gill WindMaster
        and WindMaster-Pro anemometers (needed for type 1352, not for types
        1561 and 1590).
    - Refinement: running in EXP mode, the default version of the angle of attack
        correction depends on the anemometer model (2006 for R3 and R2).
    - Refinement: increased resolution of footprint results, from 5 to 1 meter, for
        Kormann and Meixner (2001) and Hsieh (2000) models.
    - Refinement: support for TOB1 files with no header lines (explicit user specification
        of variable types 'IEEE4' or 'FP2' is needed).
    - Refinement: stationarity test evaluated after time lag compensation and
        tilt correction.
    - Refinement: minor modifications to QC flags to better match flag definitions
        according to TK3 approach (M. Mauder, personal communications).
    - Refinement: resolving Integral Turbulence Time Scale to the simple definition
        from Billesbach (2012) if direct calculation fails.
    - Refinement: support of flux averaging intervals equal to 1 minute.
    - Refinement: Raw data filtering for "absolute limits" postponed, in order to
        account for actual T and P rather than "normal values".
    - Refinement: Extended automatic definition of time lag plausibility windows (min/max)
        for both open and closed path instruments.
    - Refinement: extended range of accepted ambient pressures, now the minimum
        value accepted by EddyPro is 40 kPa, which replaces the former value of 80 kPa.

    GUI
    - Fixed: behavior of North alignment for Generic anemometer in the Metadata editor.
    - Fixed: default angle-of-attack overrides user selection when re-opening a saved
        project.
    - Fixed: flags threshold, rounding and policy persistence re-opening a saved project.
    - Fixed: prevented overriding "Lowest noise frequency" settings.
    - Fixed: prevented overwrite of the anemometer user selection on basic settings page
        in case of multiple anemometers.
    - Fixed: prevented undue field persistency changing units of the input in the
        Metadata editor.
    - Fixed: restored Magnetic declination fetching not working properly due to NOAA
        website changes.
    - Fixed: improved reset behavior on metadata file deletion
    - Fixed: magnetic declination not included in final wind direction calculations
    - Refinement: added dialog when clearing the raw data directory.
    - Refinement: modified raw file name format displayed in the corresponding dialog.
    - Refinement: permitted negative altitudes in Metadata editor.
    - Refinement: improved Metadata editor stability.
    - Refinement: added Gain-Offset automatic selection in the Metadata editor.
    - Refinement: added colors to variables ignored or not numeric in the Metadata editor.
    - Refinement: disabled not necessary fields and set Ignore to yes in case of
        Not numeric variables in the Metadata editor.
    - Refinement: prevented that column selection in Metadata tables triggers first row
        editing.
    - Refinement: Ctrl+mouse-wheel resizes the GUI main window.
    - Refinement: disabled Ctrl+F12 shortcut to launch the program.
    - Refinement: cross-wind correction checkbox is always enabled, though checked
        as suggestion when needed
    - Refinement: changed angle-of-attack correction policy with Gill anemometers
        (see help for details)
    - Refinement: increased absolute limits ranges for statistical test on gases
    - Refinement: changed policies to update software, project and metadata files
        versions at saving time
    - Refinement: custom variables created in the file description table of the metadata
        file editor will be permanently available in the local computer for future uses.

4.1.0 (2012-11-01)
      For a complete list of new features see:
      http://envsupport.licor.com/help/EddyPro4/Default.htm#Whats_New.htm

    - New features
      - A new spectral correction scheme based on Fratini et al. (2012),
        specifically designed for closed-path systems, but applicable to any EC setup.
      - A new angle-of-attack correction algorithm, implemented after Nakai and
        Shimoyama (2012).

    - Improvements
      - Usage of "previous results" to dramatically reduce program execution time.
      - Binary SLT files containing more than 6 variables can now be processed with
        EddyPro.
      - Usage of external biomet data files, not anymore limited to 18,000 data records.
        In version 4.1, biomet files can contain any number of data records.
      - Express processing uses the new angle-of-attack correction algorithm from
        Nakai and Shimoyama (2012) rather than the correction from Nakai et al. (2006)
        used in previous versions.

    - Major bug fixes
      - Automatic time lag optimization
        - Bug description
          The bug is found in the "Automatic time lag optimization" procedure
          (Advanced Settings >> Processing Options >> Time lag compensation"). The bug
          is such that time lags assessed for H2O (if this gas is treated in flux
          computation) are erroneously used also for CO2 or CH4. Selection of an
          erroneous time lag results in flux underestimates: the more the used time
          lag deviates from the real one, the more fluxes are underestimated.
        - Who is affected
          Anyone who used the "Automatic time lag optimization" option for flux
          computation is likely to be affected. However, the severity of the effects
          changes according to the gas analyzer in use. For open-path analyzers (e.g.,
          LI-7500A), the bias if most likely negligible, if present at all. For
          enclosed-path analyzers (e.g., LI-7200), the bias is probably detectable but
          still negligible, because time lags of H2O do not deviate dramatically from
          those of CO2, especially if conditions of low relative humidity (<50-60%,
          typical of daytime). In these analyzers, effects are further minimized
          if a short (< 1 m) and/or heated or insulated sampling line was used. For
          closed-path analyzers with sampling lines longer than 2 m (e.g., with an
          LI-7000 or an LI-6262), the effects are likely to be relevant and we thus
          recommend to recalculate fluxes using version 4.1.
        - Who is not affected
          All users who ran EddyPro in "Express mode", or those who did not use the
          "Automatic timelag optimization" option are not affected by this bug.
        - Note
          We recommend using the "Automatic timelag optimization" procedure especially for
          closed-path setups featuring medium and long (> 3-4 m) sampling lines. The bug
          notified here was fixed and the correct procedure is available in this release
          (EddyPro 4.1).

      - Calculation of average CH4 mole fractions and mixing ratios from LI-7700
        - Bug description
          Fixed a bug that caused the band-broadening correction to not get applied in the
          calculation of average CH4 concentrations. Note that this did not affect fluxes
          at any extent, as the band-broadening correction was applied on fluxes
          separately in 4.0. The effects of this bug were visible especially in conditions
          of very low ambient pressure with respect to normal values.

    - Other bug fixes
      - Fixed bug that caused the program to crash when the number of files selected for
        planar fit calculations were more than the maximum allowed (3,000). The bug was
        fixed and this maximum value was increased to 18,000.
      - Fixed bug that caused the calculation of maximal wind speed to fail in case of
        existence of any raw data records with at least one wind components set to -9999
        (EddyPro's internal error code). Most often this bug resulted in a maximal wind
        speed of 17318.7 m/s, that was the result of wind speed calculated from 3 wind
        components set to -9999. This bug had no affect on flux results.
      - Fixed bug that caused the cross-wind correction of sonic temperature to be
        calculated erroneously for individual data records that have any wind component
        set to -9999. This bug resulted in implausible values for sonic temperature and
        thus either in fluxes set to -9999, or to extremely spiky fluxes. Note that for
        most anemometers the cross-wind correction is applied in the firmware, so there
        is no need to apply it at processing time with EddyPro.
      - Fixed bug that caused the header of the "full output" file to be erroneous when
        using the option "Use standard output format".
      - Fixed bug that caused the night-time/daytime indication in the "full output" file
        to fail in some circumstances.
      - Fixed a bug in the Random uncertainty estimate that caused the software to
        crash with an "Out of memory" error message.

    - GUI changes
      - Added a general "Restore Default Values" button to restore all the Advanced
        Settings to the Default (express) Settings.
      - Added many constraints between fields to help fill out the Metadata File Editor
        tables.
      - Improved management of previous versions of Project and Metadata files for
        backward compatibility.
      - Added a Software Version field in the Metadata File Editor IRGA table.
      - Added an automatic "Detect Dataset Dates" button for the raw data files.
      - Prevented wheel mouse scrolling on setting controls.
      - Improved management at low resolution displays.
      - Added automatic saving on sub-dialogs exit.
      - Fixed opening associated .eddypro project files when double clicking the file.
      - Added button to clear the output console.
      - Improved guided mode information messages.
      - Fixed many minor GUI bugs.

4.0.0 (2012-04-25)
    - New features
      - Support for biomet data (biological and meteorological data) collected
        from ancillary sensors.
      - Metadata Retriever tool to compile all site metadata from GHG files
        into a single metadata file that can be edited or viewed.
      - Planar Fit tool customization with graphical controls.
      - Time Lag Optimization tool now supports automatic time lag optimization
        to account for time lags.
      - Random Uncertainty estimation tool.

3.0.1 (2012-03-12, not released)
    - Major bug fixes
      - Kljun et al. 2004 footprint estimation: Previous version used
        measurement height (z). Update uses the difference between the measurement
        height and displacement height (z-d).
      - Initialization of latitude, longitude and altitude when using
        dynamic metadata file.
      - Sampling line flow-rate calculation when average flow rate from
        high-frequency data is used to initialize nominal, minimum and maximum
        time lags.
      - Spectral assessment output file when selected correction method is
        Horst (1997).
      - A bug prevented runs without any spectral corrections applied.
      - Refined handling of the "Select a subperiod" option.
      - GUI dialog windows. Users in the previous version could potentially
        loose settings in the dialog windows. When a user clicks on a button
        to enter data, such as "Settings..." next to Generic binary, they can
        enter their data, save the project, close the project, re-open it,
        re-open the widget and their data will not be visible in the GUI. The
        data is still there but not appearing. If the user saves the project, the
        information will be overwritten. If they select cancel, it will remain as
        it was before.
      - Improved project and data selection pages by adding page splitters.
      - Fixed behavior with the cross-wind correction checkbox in relation to
        being properly deactivated with the Master Anemometer in data selection
        page.
      - Improved multiple items in the raw file description. The measurement type
        "Molar Density" was changed to "Molar/Mass Density." The option "Raw File
        Settings" is now deactivated when SLT file types are selected. The algorithm
        for detection of valid minimum variable settings was corrected for "Speed
        of Sound." The saving and reloading of the correct field separator values
        was fixed.
      - Fixed and improved multiple items in the advanced settings. The [%] unit
        was removed from the "Steadiness of Horizontal Wind" feature. Under the
        processing options, the behavior of the widgets for "compensate density
        fluctuations" and "add instrument sensible heat components" was corrected
        to gray and disabled when appropriate.
      - Fixed the error label default number to correctly display in the project
        file as -9999. (Advanced settings/output files).
      - Fixed the default setting under Advanced Settings/Output Files so "Build
        continuous dataset" is unchecked.
      - Match the default unchecked state of Metadata output checkbox with
        project file setting and remove the incorrect association with Full Output.
      - Fixed incomplete offline help system.
      - Fixed the Guided mode for the Metadata File Editor algorithm to properly
        assess the fixed unit of measurement for both input and output
        relationships.

3.0.0 (2012-02-06)
    - Final official release of introducing the Advanced Settings.

3.0.0 Beta (2011-12-14)
    - First public release of 3.0 Beta. This version introduces the Advanced Settings.

2.3.0 (2011-07-20)
    - Major bug fixes
      - An error occurred in the calculation of corrected fluxes if the
        N2O analyzer was a closed-path one, while the CH4 analyzer was open-path
        or CH4 was not present at all. The bug resulted in corrected fluxes of
        N2O equal to -9999 (that is, if your N2O fluxes are not -9999, they are
        not affected by this bug).
      - Definition of "relative separations". No impact on calculations,
        only on the output metadata.
      - Control on use of cell temperature (fixed re-initialization of column
        information)
      - In statistical tests (Spike count/removal and Amplitude Resolution/Drop
        Out) length of implied windows modified so as to scale with the length
        of the averaging period.
    - Improvements
      - Handling of situations where H2O readings are not available
        (air density, momentum flux).
      - Control over peculiar situations in Fluxes23 (WPL section).
      - Introduced support for TOB1 format "FP2"
      - Introduced detection of very implausible values in the raw
        files, and their substitution with error codes.
      - Error codes in raw files (-9999) are no longer modified
        during conversions of raw units into EddyPro units
      - Better handling of situation when all GHG files have invalid
        metadata or are corrupted.
      - Better handling of the computations exit codes

2.2.0 (2011-06-20)
    - Major bug fixes
      - Fixed bug concerning fluxes calculated from CH4 and N2O mole fraction
        measurements from closed path systems. The bug affected also fluxes
        calculated with CO2 mole fractions, if the paired H2O was measured as
        either molar density or mixing ratio. In all cases, the bug resulted
        in flux values =-9999 (that is, if your fluxes are not -9999, they
        are not affected by this bug).
      - Fixed Altitude field zeroing in the Metadata editor.
      - Changed the docking policies of the Tips dock: now it's movable and
        floatable top and bottom and it could also be tabbed with the Console.
      - Changed the dimension policy of the Tips dock: now it's expandable.
      - Fixed misspellings.
      - Used native dialog for the 'Save metadata as' button.
      - GUI code cleaned up.

2.1.4 (2011-06-02)
      - Fixed incorrect filtering of CH4 vars in the Processing page, in case of
        Generic gas analyzer.
      - Added the changelog file viewer in the About dialog.

2.1.3 (2011-05-31)
      - Fixed the SoS to temperature conversion.
      - More detailed diagnostic on the metadata file validation at console.
      - Quality flags recalculated according to Mauder and Foken (2004), which
        should eliminate most (possibly not all) the -9999 in the quality flags
      - Added CHANGELOG.txt to source and installer packages.

2.1.2 (2011-05-18, not released)
      - Fixed problem of Access denied authorization during the processing of the
        data.

2.1.1 (2011-04-26)
    - Fixed typo about the unit of measure of the Raw Data Description table
      (umol/mol^3, instead of umol/m^3).

2.1.0 (2011-04-21)
    - Optimized GUI for support low resolution screens (up to 800x600 px).
    - Rearranged widgets between Project page and Processing page (ID moved to
      Processing).
    - Added input dialog for the ID field.
    - Improved look&feel of Combo box and Spin box.
    - Changed policies for scrollbar and scrolling behavior into the tables.
    - Changed default value in some table fields.
    - Changed from QByteArrays to QString where there is the possibilities of
      weak conversions.
    - Fixed selection of None variables into Processing page combo's when there
      are already some variable.
    - Updated the manual.

2.0.0 (2011-04-05)
    - First public release of EddyPro (Express version only).