Model Change Bulletin

                               MCB#4                       8/19/93

                       CTSCREEN (dated 93228)


     This Model Change Bulletin is to notify users of changes
made to correct problems in CTSCREEN (dated 93076).  The specific
changes, the problems addressed, and the consequences of the
changes are summarized below.

     These revisions constitute CTSCREEN (dated 93228) and
completely replace CTSCREEN (dated 93076).  With this bulletin,
the source codes and executables for CTSCREEN (dated 93228) are
being made available for downloading from the Air Quality Models
area of the File Transfer Section of SCRAM.  CTSCREEN (dated
93076) is obsolete and, with this bulletin, will no longer be
available on SCRAM.  All copies of such should be discarded.


1.  Non-convergence in WPDF

PROBLEM:

      The WPDF subroutine determines the vertical velocity that connects
the source to the receptor through the terrain-disturbed flow.  Thirty
iterations are allowed to converge on a value for the vertical velocity.  The
resulting vertical velocity (wprime) is used in the probability density
function calculation and the y-coordinate of the position of the streamline is
used in the calculation of sigma-y.  In CTSCREEN (dated 93076), when
convergence was not obtained, the last value of wprime was retained
correctly.  However, the position was incorrectly reset to the receptor
location.

SOLUTION:

      In CTSCREEN (dated 93228), the position that is returned is the
one corresponding to the last guess for wprime.  For most cases, this coding
change will have no effect on predicted concentrations.  For cases in which
the non-convergence occurred, we can not predict the magnitude or the
direction of changes that could be expected.

SUBROUTINES MODIFIED:

      WPDF.FOR


2.  Wind speed used in WPDF subroutine

PROBLEM:

      In CTSCREEN (dated 93076), the wind speed used in determining the
vertical velocity needed to connect the source to the receptor through the
terrain-disturbed flow was set to the wind speed at stack height.  This is
inconsistant since the wind speed at half the plume height is used in the
concentration equation.

SOLUTION:

      In CTSCREEN (dated 93228), the wind speed at half the plume height is
used for all calculations.  This coding change will have no effect on results
from CTSCREEN since the wind speed profile is constant with height.  The code
was modified to maintain consistency with CTDMPLUS.

SUBROUTINES MODIFIED:
      WPDF.FOR, FLOWSP.FOR, DAYSCR.FOR


3.  Problem with circular ellipses

PROBLEM:

      In CTSCREEN (dated 93076), the code for handling the special case of a
circular ellipse did not account for a value of the variable ARSQ less than
1.0.

SOLUTION:

      A check for a value of ARSQ greater than 1.0 has been added to
CTSCREEN (dated 93228).  For most cases, this change will have no effect.  A
runtime error would have been reported if this problem had been encountered in
a particular run.

SUBROUTINES MODIFIED:
      SEQSCR.FOR


4.  Error in DO loop

PROBLEM:

      In CTSCREEN (dated 93076), the "DO 110" loop contained a GO TO 300
statement that caused the loop to be exited prematurely.


SOLUTION:

      The GO TO 300 statement has been moved outside of the loop in
CTSCREEN (dated 93228).  The source contribution array is zeroed in the "DO
110" loop.  There may be cases where the reported source contribution may
differ from previous versions, although we have not observed any differences
to date.

SUBROUTINES MODIFIED:
      CONCALC.FOR


5.  Problems with non-buoyant releases

PROBLEM:

      In CTSCREEN (dated 93076) an error occurred when the stack exit
temperature was less than the ambient temperature.

SOLUTION:

      A check was added to detect the case where the stack temperature is
less than the ambient temperature and the stack temperature is set equal to
the ambient temperature.  For most cases, this coding change will have no
effect.  For non-buoyant releases, a runtime error will be eliminated.

SUBROUTINES MODIFIED:
      SEQSCR.FOR

6.  Problems associated with the SUMRE file

PROBLEM:

      There were three problems with the SUMRE file produced by CTSCREEN
(dated 93076) that were reported by users.  First, the source contribution
reported in the receptor summary in the SUMRE file was incorrect.  Second, the
format for writing concentrations did not allow for large values.  Third, a
runtime error occurred if all concentrations were exactly 0.0.

SOLUTION:

      These problems have all been remedied in CTSCREEN (dated 93228).

SUBROUTINES MODIFIED:
      WRITSCR.FOR


7.  New value for IUNSTA switch

        A new value for the IUNSTA switch has been added that allows the
user to model only unstable hours.  The possible values for this switch are
now:
         0 = model only stable hours
         1 = model both stable and unstable hours
         2 = model only unstable hours

This change does not affect CTSCREEN runs since the IUNSTA switch is not
used by CTSCREEN.  The code was modified to maintain consistency with
CTDMPLUS.


8.  Changes made to the menu driver

      a.  DRIVEIT now deletes any existing output files (e.g. STCONC, SUMRE)
          before a new model run begins.

      b.  The error handling associated with running CTSCREEN has been
          improved.  In addition to the detection of runtime errors that
          was a part of the last version, DRIVEIT now indicates when
          non-runtime errors (e.g.  lowest contour above base elevation) occur
          and refers you to the CTDM.OUT file for more information.

9.  Changes to RECGEN

      In some cases (especially multiple contours), RECGEN did not correctly
repaint the screen when the F2 key was pressed.  This has been corrected.


10.  Contacts for Questions (NO CHANGE)

     Questions related to the use of CTSCREEN in specific
regulatory applications should be directed to the appropriate
regulatory agency or Regional Modeling Contact (see State/Local
Modeling Contacts in the Agency Communications Section of SCRAM).

     Generic questions related to the use and regulatory
applicability of the model should be directed to Desmond Bailey
via SCRAM EMAIL or phone (919) 541-5248.

     Questions or problems related to program execution should be
reported to Donna Schwede via SCRAM EMAIL or phone (919) 541-3255
or to Steve Perry phone (919) 541-1341.

     Users who would like to share comments, suggestions, or
helpful hints are encouraged to use the Model Conferencing area
on the SCRAM Bulletin Board (see Model Conferencing area in the
Public Communications Section of SCRAM).


     Technical correspondence should be directed to:

     Ms. Donna Schwede or
     Dr. Steven Perry
     Applied Modeling Research Branch
     ACMD/AREAL/USEPA
     MD-80
     Research Triangle Park, NC  27711

     Fax: (919) 541-1379