GTOL Version 6

GTOL Version 6.4b [December 3,2003]

Author: Thomas W. Burrows

Dept. of Energy Sciences and Technology

National Nuclear Data Center Bldg. 197-D

Brookhaven National Laboratory

P.O. Box 5000 Upton, NY 11973-5000

Phone: 631-344-5084 FAX: 631-344-2806

(Original authors: W.B. Ewbank, Nuclear Data Project, Oak Ridge National Laboratory; B.J. Barton, National Nuclear Data Center, Brookhaven National Laboratory; and L.P. Ekstrom and P. Andersson, Department of Nuclear Physics, Lund University)

In this program, gamma-ray energies are used to derive a set of least-squares adjusted level energies. The net feeding at each level is calculated from the input gamma intensities and conversion coefficients. Unplaced or questionable gammas or gammas whose final level is ambiguous or unknown are ignored. The program parses the DSID of each data set and, if there is no indication of possible gamma records within the data set, skips it. In addition, the program will not calculate the intensity balancing for adopted data sets.

Input file: An ENSDF formatted file with the following optional information: An option record with 'OPTION' in col. 1-6 may precede any data set and contain any of the following options in free format: OPTION MEANING NOREC No recoil correction, i.e. recoil correction has already been applied to E(GAMMA) RECOIL Perform recoil correction (DEFAULT) MARKED Process only data sets preceded by a card with '*GTOL' in col. 1-5 ALL Process all data sets (DEFAULT) DEG= For the current data set, override default assumption of 1 keV where no uncertainty on the gamma energy is given. Following the equal sign may be either a number or a number followed by a percent sign. A number alone indicates the uncertainty on EG in keV while a number followed by a percent sign indicates the fractional percent uncertainty to be assigned.

GTOL Version 6.4b [December 3, 2003] DRI= For the current data set, assume a default uncertainty for the relative photon intensity when none given. A number alone indicates the uncertainty on RI in the current relative units while a number followed by a percent sign indicates the fractional percent uncertainty to be assigned. DTI= For the current data set, assume a default uncertainty for the total transition intensity when none given. A number alone indicates the uncertainty on TI in the current relative units while a number followed by a percent sign indicates the fractional percent uncertainty to be assigned. Note that an option card resets the defaults. A level energy can be held fixed by adding the letter 'F' somewhere in the energy field (columns 10-21). If the output option to create a new file containing the adjusted level energies is chosen, the 'F' will be removed and a level documentation record will be added (LEVEL ENERGY HELD FIXED IN LEAST-SQUARES ADJUSTMENT). If DRI or DTI are specified on an OPTION record, the assumed uncertainty may be overridden for an individual intensity, by adding an "E" separated from the intensity in either the RI or TI fields. If DEG, DRI, or DTI are specified on an OPTION record and a new file is created, FOOTNOTE COMMENTS will be generated and inserted as necessary.

Sample input file: DATA.TST

Output files: 1. Report file.

The report file will contain a summary of the data input and actions taken by the program (e.g., unplaced or questionable gammas ignored) and the following optional outputs for each data set:

(a) Comparison of input gamma energies to those calculated based on the

Adjusted level energies.

(b) Comparison of calculated net feedings to each level with values input on

B, E, or A records.

Note: If the calculated net feeding overlaps zero within three standard deviations, the program will calculate estimated upper limits (90% confidence level) using two methods suggested by Louis Lyons in Statistics for nuclear and particle physicists (Cambridge University Press) and report these

GTOL Version 6.4b [December 3, 2003] estimates if they differ by more than 0.01. The two methods are:

1. (Integral of gdB from 0 to Bl)/(Integral of gdB from 0 to infinity)=0.9 where g is the normal (Gaussian) distribution.

2. Bl<Bm+1.28sigma. Sample output file: GTOL.RPT 2. New file containing the adjusted level energies (Optional). Sample output file: None Terminal dialog: The program will request the input and report (default: GTOL.RPT) file names and ask if you wish a new file created (default: no new file) and for the new file name, to suppress the gamma-energy comparison (default: no suppression), and to suppress the intensity comparison (default: no suppression). The progress of the program will be noted on the terminal as well as possible problems. Compilation and loading instructions: This program requires subroutines from the NSDFLIB.FOR package.

Additional documentation:

B.J. Barton and J.K. Tuli. PHYSICS ANALYSIS PROGRAMS FOR NUCLEAR STRUCTURE EVALUATION. Brookhaven National Laboratory Informal Report BNL-NCS-23375/R (1977). L.P. Ekstrom and P. Andersson. FORTRAN 77 VERSIONS OF STRING HANDLING SUBPROGRAMS AND THE PROGRAMS GTOL AND MEDLIST. Nuclear Physics Report LUNFD/(NFFR-3049)/1-27 (Lund University, Lund Sweden. 1983).

Version History

VERSION 5(0): FORTRAN77 Version August, 1983, LPE, Lund, Sweden.

Based on VERSION 4(15) as of 6-Jan-82.

MAIN CHANGES:

1) FORTRAN 77 using STR77 Library.

2) Number of levels, gammas, fixed levels as

parameters.

3) Some unused options removed.

4) Some options added.

VERSION 5(1): February 1986

MAIN CHANGES:

1) Localized program to improve paging rate on

TOPS-10.

2) Changed terminal dialog.

3) Size of output reduced.

4) Option to create a new file with calculated

level energies included.

5) Check of GAMMA continuation record for final

level (FL=) formalism and other quantities

which may affect the calculations.

6) Ignores data sets whose DSID's indicate that

there are no gammas present.

KNOWN PROBLEMS (Common to all versions):

1) Field width overflows when there are very

precise gamma energies.

2) Matrix inversion sometimes unstable when there

is only one transition to or one transition

from a level.

VERSION 5(2): June 1986

MAJOR CHANGES:

1) Converted from Swedish STR77 library to NNDC

to NNDC F77STR and NSDCNV libraries

VERSION 5(3): 8-Aug-1986 Add VAX MDC

VERSION 5(4): 11-Dec-1986 Add IbmPC MDC

VERSION 5(5): 4-Sep-1987

1) Corrected logic causing floating-point overflow

in subroutine MINV.

2) Corrected minor parsing problems in subroutine

GA2DEC. Also rewrote this subroutine to reduce

redundant coding using modules from RADLST and

the new subroutines CHNGS1 and CHNGS2.

3) Corrected field-width overflow problems by

increasing associated string lengths and

format statements from 8 to 10 (Maximum size

of energy fields in ENSDF). Since this

increased the size of the output, added checks

to not list blank lines when there are no old

gamma energies to be compared to.

4) Removed output of "F" for level energies.

5) Brought modules associated with these

corrections more up to current F77 standards

and philosophy.

VERSION 5(6): 2-Nov-87

VAX mdc OPEN READONLY added for input dataset

file

VERSION 5(7): 30-Aug-89

1) Added checks for

a) Number of levels exceeding number of

gammas plus fixed levels

b) Matrix being singular

In both of these cases it will list connected

but not fixed levels which either have

a) No gammas feeding them

b) No gammas deexciting them

since this seems to be the most common cause

of the problem

2) General cleanup of code

3) Added logic to reduce extraneous calculations a

output

4) Restored lost coding to remove output of "F" for

level energies and to output a "D L" record for

these levels

5) Automatic suppression of intensity comparison

for ADOPTED LEVELS, GAMMAS dataset

VERSION 5(8): 12-SEP-89

1) Corrected logic error which caused (G,G') data

sets to be rejected

2) Changed string lengths and output formats for

I/O files to reduce changes of truncation

VERSION 5(9): 12-JUN-90

1) Changed statement order(CHKALF),F format spec(I

etc for PC fortran

VERSION 5(10): 24-JUL-90

1) Corrected error in INTOUT which gave log of a

negative number

VERSION 5(11): 13-DEC-90

1) Corrected error in INTOUT which gave log of

zero

2) Delinted using F-LINT 2.71

VERSION 5(12): 15-Oct-92 Added Machine coding for ANS

VERSION 6.0: 7-Apr-93 Merged IBM PC test version of 28-Feb-91

with distributed version 5(12) [15-Oct-92]. IBM PC

changes were:

1) Reworked internal storage to fit into

limitations of real mode of MS/DOS.

a) Most argument passing to subprograms replace

by COMMONs.

b) Replaced WAA(NLE,NLE) by WAA(NSTORE) with

NSTORE=(NLE*NLE+NLE+1)/2 and added

bookkeeping function STRLOC

2) Replaced general purpose matrix inversion

routine MATINV and MINV by a new MATINV which

uses a specific algorithm for symmetric

matrices.

3) Added PCTIME and PCDATE routines for IBM PC

which access the MS FORTRAN routines GETTIM and

GETDATE

*** As a result of 1b) and 2) the code runs slower

however, for NLE=300 memory requirements were

reduced by about 658k.

Other changes:

1) Added overlay module indicators for use in

separating source code for compilation and

linking on IBM PC

2) Added check on diagonal matrix elements after

inversion - If any negative values, no

least-squares adjustment

3) Added check on E(level) after matrix

multiplication - If any negative values, level

processing terminated

4) Removed redundant output of fixed levels list

5) Replaced numeric comparison for FL= with

string comparison and corrected minor logic

errors

6) Added bookkeeping on how the level was fixed

7) Always assume that the first level is fixed

8) Added logic for non-numeric levels

9) Update FL= when new file option specified

10) Warn about levels being out of order in new

output

11) Increased NFIX=NLE/2 to NFIX=NLE

12) Reduced verbosity of report by only

putting out relevant input data (should also

reduce elapsed time due to I/O

13) Corrected output field width overflow in

intensities

14) Kept non-numeric uncertainties on feeding

radiations

15) Hold levels of the form SP+x or SN+x fixed

and ignore deexciting gammas

16) Correction of minor logic errors in calculating

total intensities

17) If NB not given, assume 1.0/BR in agreement wit

other codes

18) Accounted for 3% uncertainty in CC theory when

calculating TI

VERSION 6.1: 12-Jul-93 Corrected error in calculating net g.s.

feeding (100*BR+TNRF to 100+ TNRF)

VERSION 6.2: 26-Nov-93

1) Removed error introduced in version 6.1.

2) Ignore gammas with non-numeric EG.

3) Implemented Asilomar F and P recommendations to

allow specifications of DEG, DRI, and DTI when

not given.

4) For DEG, DRI, or DTI of "AP", uncertainty set

to three times that for field.

5) If "&", in column 77, RI=DRI=(RI+DRI)/2 and

TI=DTI=(TI+DTI)/2 assumed.

6) If "LT" or "LE" in DRI or DTI field,

RI=DRI=RI/2 or TI=DTI=TI/2 assumed.

7) Automatically remove previous GTOL-generated

"DL E" records.

Version 6.2a: 09-Apr-1999

1) Y2K compliance

2) Increased ANSI FORTRAN 77 compliance

3) Check for and skip Ionized Atom

4) Properly recognize H record

Version 6.3: 23-May-2000

1) Implemented logic for FL=?

2) Added estimates of upper limits of the

calculated net feeding using the methods of Lyon

3) Corrected bug in subroutine CHGCRD

4) Removed overlay logic

Version 6.3a: 08-Jan-2001

Corrected logic errors in replacing "FL="'s

Version 6.4 01-Mar-2001

Added MDC code for f77 on Linux

Version 6.4a 11-Jul-2001

Corrected output overflows when adding new Comment

record for DEG, etc.

Version 6.4b 03-Dec-2003

1) Increased NLE from 300 to 1000 and NGA from

3*NLE to 4*NLE.

2) Left justify new level energies in output file.

3) Corrected outout overflow problem in RADDEC.

4) Corrected string range problem in IDDEC.