Version 5.3.1, Oct 2016 S. Madronich
*** fix a few bugs in cross sections:
   - for acetone (absorption option 4) added T^3 term (previously had corrected
     only option 3)
   - for HCFC-123 (rxn32); HCFC-124 (rxn33); and HCFC-142b:
     corrected indexing problem, was
      t = MIN(295.,MAC(tlev(i),203.))-TBar
     changed to:
      t = MIN(295.,MAC(tlev(iz),203.))-TBar

________________________________________
Version 5.3, June 2016, S. Madronich

*** MOLECULAR DATA, CORRECTIONS AND ADDITIONS:

- In subroutine qyacet (acetone quantum yields) change 
      IF(w .GE. 302. .AND. w .LT. 327.) THEN
   to
      IF(w .GE. 302. .AND. w .LE. 327.) THEN
   to avoid possibility of undefined value at exactly w=327.
   (thanks to Peter Braeuer for pointing out error)
   Also added exponent overflow protection in qyacet.

- Corrected acetone cross section, was missing T^3 term in temperature
  correction equation.  Thanks to Glenn Wolfe (NASA) for pointing out this 
  error.

- For acetaledhyde qy1 pressure dependence, replace use of slope/intercept
  from Horowitz and Calvert, with direct Stern-Volmer assuming total yield = 1
  at zero pressure.  Makes some difference at altitudes about 30 km.

- Corrected read of methyl-ethyl-ketone absorption cross section from 
  Martinez et al. (was reading di-ethyl ketone by mistake). Thank you Peter Brauer 
  for pointing this out.
   (thanks to Peter Braeuer, U. of York, for pointing out error)

- Added photolysis of I2, IO, HOI

- Corrected typo in swbiol.f, Green's fit to Caldwell generalized
  plant action spectrum, preexonential changed to 2.618 from 2.628. Difference
  is about 0.4% is using this action spectrum (but not default).
  (thanks to Pedro Aphalo, U. Helsinki, for point out this problem)

- Changed CIE erythemal and UV Index spectrum to reflect CIE1998 definition rather
  than older (1987) definition.

- Re-normalize the DNA (Setlwo) action spectrum at 254 nm, in order to match with
  germicidal Hg lamps.

*** CODE BUG FIXES:

- Corrected line in addpnt.f:
  old:  IF(xnew. GT. x(i)) insert = i + 1
  new:  IF(xnew. GT. x(i-1)) insert = i

*** NEW FEATURES:

- Added absorption by surface waters, including the following subroutines,
  collected in file waters.f:
   WATERS(sza,kabs, ny, y, se_y, de_y) = driver for waters
       INTRFC(sza_a,se_a,sza_w,se_w, nang,dza_ai,df_a,dza_wi,de_wi) = 
           = computes interface angles, reflection for direct and diffuse
           SNEFRE(sza_a, sza_w, r) = Snell's refraction and Fresnel reflection
           ANGDIF(nang, dza_ai) = subdivides atmospheric diffuse ligth into 
                  NANG-1 equal areas (cosine-weighted), reports central angle.
  Results are reported as a function of detph, in meters.  Note that values just
  below the interface are slightly different than just above.
  For now, these routines are kept together in waters.f and are called from the 
  innermost loop (wavelength) of TUV.f.  However, many related parameters could be calculated
  externally saving computational time, e.g. refraction and reflection depend only on 
  solar zenith angle, and diffuse discretization depends only on chosen number of angles.

- Arbitrary constituent: Option to add any additional absorber or scatterer 
  (useful for sensitivity studies)
    dt_any(kz,kw) = optical depth for additional constituent
    om_any(kz,kw) = single scattering albedo for additional 
    g_any(kz,kw) 
  Can set manually or write a routine "setany" similar to other set routines

- Light pollution at night: Option to enable studying light pollution from sources at the surface, e.g.
  urban light sources.  Do this by turning off the incident solar beam, and adding diffuse emission from
  the surface.  For now, do only 2-stream, not with discrete ordinates.
  In subroutine ps2str, normally pifs = incident beam at top of atmospehre = 1.  But want to 
  keep pifs arbitrary, so carry into equations:  Replace
           fdr(lev) = EXP( -tausla(0) )
  with:      
	   fdr(lev) = pifs * EXP( -tausla(0) )
  and
         fdr(lev) = EXP(-tausla(lev-1))
  with
         fdr(lev) = pifs * EXP(-tausla(lev-1))
  So if night, pifs = 0.
  Also added surface (lower boundary) source in addition to albedo:
  so replaced
      ssfc = rsfc*mu*EXP(-tausla(nlayer))*pifs 
  with
      ssfc = rsfc*mu*EXP(-tausla(nlayer))*pifs + surfem
  where surfem = irradiance emission by the surface, e.g. normalized to 1.0

*** CODE ORGANIZATION

- Added new subroutine terint.f to TERminate and INTerpolate input data arrays.
  Takes the place of calling addpint 4 times and inter2 once.  Final two arguments
  specify if terminate with last constant (value = 1) or terminate with zero (value=0)

- Moved out of main TUV.f into subroutines:
    WRFLUT = look up table for WRF-Chem.  Code stops after table output.
    NEWLST = create new list of reactions. Code stops after list output.
  These are special applications that should not occupy much space in
  the driver TUV.f 
_____________________________________
Version 5.2, October 2014, S. Madronich

Moved correction for refraction to gridw routine, option 6

Create defin5 option, to generate molecular spectra x qy for use in wrf-chem.
For now, output to file ../sq_wrfc.txt

BrONO2 yields of NO2 and NO3 were reversed, now corrected.

Corrected error on CH2O qy molecular channel (was using pre-interpolation value).
_____________________________________
Version 5.1, Oct 2014, S. Madronich

August 2014:
added perfluoro 1-iodopropane photolysis 

July 2014:
Added several organic nitrate reactions from IUPAC 2006 (Atkinson et al. ACP).

July 2013:
updated most Ox, NOx, HOx, and HC reactions to jpl2011:
Reordered the reactions to follow JPL 2011.

Updated: HONO (thanks to Patryk Wojtal for pointing out need)
Updated: CH2O (note several errors in qy formulae in JPL 2011)

January 2012
Added photolysis reactions:
F 87 BrNO -> Br + NO                                   
F 88 BrNO2 -> Br + NO2                                 
F 89 BrONO -> Br + NO2                                 
F 90 BrONO -> BrO + NO                                 
F 91 HOCl -> HO + Cl                                   
F 92 NOCl -> NO + Cl                                   
F 93 OClO -> Products                                  
F 94 BrCl -> Br + Cl                                   
F 95 CH3(OONO2) -> CH3(OO) + NO2                       
F 96 C(CH3)3(ONO) -> C(CH3)3(O) + NO                   
F 97 ClONO -> Cl + NO2                                 
F 98 ClOOCl -> ClOO + Cl                               

Corrected error in calculation of qy for acetone at 327.5nm.
_________________________________________
Version 5.0, November 2010, S. Madronich

Added some data files on O3 column and profile climatology in directory 
DATAE1/ATM.  At this point, the code does not ingest the data but this can
be easily modified, and will be in the future.

Wrote individual read routines for O3 cross section data (WMO1985, Molina and Molina,
JPL2006, Reims group, Bass et al.).  This makes it easier to switch between the different
data sets.

Corrected O3 cross section wavelengths.  Original data is in standard air, 
now converted to vacuum wavelengths for use in TUV which is assumed in vacuum.
Previous versions of TUV did not make this conversion, so there was a slight
misalignment between wavelengths of the O3 cross sections (which were in air)
and the extraterrestrial flux (which was in vac)

Added O3 cross section from Reims group (Malicet et al., Brion et al.)
and also from JPL2006

Added ET flux from Chance and Kurucz (200-1000 nm by 0.01 nm steps)

Removed orphan references to ftrim (in savout.f) and to plkavg (in rtrans.f)

Added ClO photolysis per JPL2006

Corrected error in acetone quantum yield enforcing upper/lower bounds of 1./0.
(most likely inconsequential)  for Blitz qy (thanks to Barron Henderson, U. North Carolina,  for catching this)
      
Updated to JPL2006:
   NO2 absorption and quantum yield
Also moved NO2 absorption read/interpolate data into subroutines for different data sets,
so can call for (1) transmission, and (2) photolysis.

_________________________________________
Version 4.6 March 2009, Sasha Madronich
Corrected indexing problem for PAN and PPN, two channels:
        j = j+1
        jlabel(j) = 'CH3CO(OONO2) -> CH3CO(OO) + NO2'
        j = j+1
        jlabel(j) = 'CH3CO(OONO2) -> CH3CO(O) + NO3'
    The sq assignment was to j and j+1 rather than to j-1 and j.  Therefore the 
    NO2 channel was not assigned (zeros) and the NO3 channel contained the NO2 channel
    instead.  Same problem for PPN.
Deleted old input files in directory INPUTS (mex, j01, t1).  The order of reactions was 
    old and different than new order. Their potential use would lead to mis-assignment of reactions.
Changed interopolation schemes in vpair.  Now interpolating log concentrations onto z-grid, 
    then doing log integration to get column increments.  Also changed notation.
Changed interpolation schemes in vpo3. Linear interpolations of concentrations.  Also extended the 
    ussa O3 profile to 121 km (from 74 km) using exponential scale height of 4.5 km.
___________________________________________
Version 4.5 July-September 2007, Sasha Madronich
Added biological weighting functions for:
	vitamin-D production (CIE, 2006a)
	non-melanoma skin cancer spectrum (CIE, 2006b)
Added photolysis reactions: 
	CH3COCH2CH3 (MEK)
	CH3CH2COO2NO2 (PPN) 
	CH2=CHCHO (acrolein)
	CH3CO(OOH) (peracetic acid)
Modified:
	CH2=C(CH3)CHO (methacrolein) updated to JPL2007
	Put in quantum yields for PAN photolysis to NO2 and NO3
Created z-grid for mexico city. Cleaned up some gridz options.
Created option to overwrite PBL pollutants O3, SO2, NO2, aerosol
   This is a preliminary add-on, with some cautions requred. See TUV.f
Changed deltax to 1e-5 (in file params). This uncovered minor data 
   termination error in file rxn.f, subroutine r43, for Halon-2402 (CF2BrCF2Br)
Changed order of photolysis reactions (in swchem.f), to group together 
   related reactions.
Moved all input files (including default inputs, and user-defined inputs) 
   to new directory INPUTS
Regenerated SAMPLE files.  Most values are identical to before.  Some 
   minor changes, mostly due to new/modifed photolysis reaction.

_____
15 November 2005, J. Lee-Taylor
Version 4.4 includes calculation of snowpack radiative properties:
* Lee-Taylor, J., and S. Madronich (2002), Calculation of actinic fluxes 
* with a coupled atmosphere-snow radiative transfer model, J. Geophys. 
* Res., 107(D24) 4796 (2002) doi:10.1029/2002JD002084
It includes absorption by ice:
*Grenfell and Perovich, J. Geophys. Res., 86, 7447-7450 (1981),
*Perovich and Govoni, Geophys. Res. Lett.,18(7),1233-1235 (1991)
and photodissociation of NO3-(aq):
*Cross section: Burley & Johnston, Geophys. Res. Lett., 19, 1359-1362, (1992)
*Quantum yield: Warneck & Wurzinger, J. Phys. Chem., 92, 6278-6283, (1988)   
Snowpack properties are set manually in subroutine setsnw.f
There are no other changes since V4.3
______
10 October 2005, S. Madronich
Version 4.3 uses the new acetone (CH3COCH3) quantum yields from:
* 4:  Blitz, M. A., D. E. Heard, M. J. Pilling, S. R. Arnold, and M. P. Chipperfield 
*       (2004), Pressure and temperature-dependent quantum yields for the 
*       photodissociation of acetone between 279 and 327.5 nm, Geophys. 
*       Res. Lett., 31, L06111, doi:10.1029/2003GL018793.
This was already implemented in version 4.2 on May 4 2004.  However, earlier versions of tuv4.2 may
use the older data from McKeen et al.
______
10 October 2005, S. Madronich
Version 4.3 corrects an error existing in version 4.2.  The error concerns
the option to shift between in-vacuum and in-air wavelengths.  This option 
is enabled by setting the logical variable lrefr=.TRUE. in the driver 
program, TUV.f.  The error was in subroutine wshift.f, and applied the 
shift in the wrong direction.  The correct eqations
(used in version 4.3) are:

      IF(mrefr .EQ. 1) THEN
         DO i = 1, n
            w(i) = w(i) * refrac(w(i),airden)
         ENDDO
      ELSEIF(mrefr .EQ. -1) THEN
         DO i = 1, n
            w(i) = w(i) / refrac(w(i),airden)
         ENDDO
      ENDIF

and replace the incorrect equations (in version 4.2)

      IF(mrefr .EQ. 1) THEN
         DO i = 1, n
            w(i) = w(i) * refrac(w(i),airden)  !! INCORRECT
         ENDDO
      ELSEIF(mrefr .EQ. -1) THEN
         DO i = 1, n
            w(i) = w(i) / refrac(w(i),airden)  !! INCORRECT
         ENDDO
      ENDIF

* Thanks to Joe Michalsky (NOAA) for catching this error.

