Lithologic column plotting for Expeditions 304/305
Donna Blackman

The starting point was the Igneous Petrology team's Unit Log. Dave Christie checked the Unit Log against the Barrel Sheets near the end of Expedition 304. I did the check on a daily basis on Exp. 305 and did a final pass at the end of the cruise. Differences between the Unit log and the Barrel Sheets were rare but when they occurred the Unit Log was made consistent with the Barrel Sheet. Oxide gabbro units in the lithologic column generally indicate any presence of oxide, in some cases this means that modal contents less than 2% are plotted with the oxide pattern shown.

Depths were included in the Unit Log by starting with the Janus database section-top depths. However, when curated lengths exceed the cored interval a shift was applied so that the locations of core pieces do not overlap. In almost all cases the shift was made in the upward sense (pushing into space below the previous cored interval). If several cores had high recovery, several sections in a row might require (variable) upward shifts.  In a very few cases a downward shift was applied if more space was available due to lower core recovery down section. The table of shifts applied to the depths (mbsf) in the Unit Log is called shiftz.xls. This table is archived with the Supplementary Material.

The top and bottom depths of each lithologic unit, grain size, and the rock type were extracted from the unit log and saved in a separate Excel file. Single-word abbreviation replacements were made to the rock type names. Medium-coarse grained and coarse-grained qualifiers (only) were retained. The Excel data were saved in CSV format and subsequently processing used UNIX scripts.

Unit rock type descriptors in *.csv and *lithnum files
Relevant grain size modifier precedes rock name; if absent fine- or medium-grained applies to the unit
mc: medium-coarse grained
c: coarse-grained

10 Basalt
11 Bbreccia: brecciated basalt
12 Bdike: basaltic dike
13 Diabase
20 Gabbro
20 GDike: gabbroic dike
21 MGabbro: microgabbro
22 OBGabbro: olivine-bearing gabbro
23 Gabbronorite
23 PGabbro: orthopyroxene-bearing gabbro
30 XGabbro: oxide
31 AGabbro: amphibole (possibly primary) bearing gabbro
40 OGabbro: olivine gabbro
41 TGabbro: troctolitic gabbro
50 Troctolite
60 Wehrlite 
61 Dunititc Troctolite
61 Dunite
62 Harzburgite
70 Fault
70 Cataclasite
71 Rubble
80 Sediment


Mac2unix was run to convert format of the *.csv file. Grain size qualifiers in front of the XGabbro and Rubble were deleted and some additional hand editing to clean up typos was done. The *.csv file was transferred to the SUN workstation and reformatted using an awk script mklis: 

mklis < d.csv >  dlith

Next a FORTAN program lithchk.f was run on the output of the awk script (dlith). This checked for remaining typos in the *.csv file and also assigned a numerical value to each unit/rocktype (as listed above). 

lithchk <<EOI
dlith
dlithnum
EOI

dlithnum, with numerical assignment & rock name vs top/bottom depths, was copied to the Strat Columns folder on Uservol

lithrave.f is a FORTRAN program that computes a running average of the fraction of each lithology recovered per cored interval.  Lack of recovery is included in the calculation, so the lithologic proportion is generally not the same as its proportion in the recovered rock. The length of the running average and the step at which the average window is shifted down core are input by the user. An example for a 20 m running average at 5-meter step is:

lithrave <<EOI
dlith
dlithout
20
5
EOI

The output is an ASCII listing of median depth for each averaging interval and columns with the fraction for each lithology (numbers correspond to rock types as on previous page)

depth (m) 	col1	col2	col3	col4	col5	col6
	10-13	20-23	50	30, 31	40, 41	60-62

A Matlab script (pltlithave.m) can be used to plot the output of lithrave (e.g. dlithout). Run a couple commands in matlab and then the script:

load dlithout
c=dlithout;
pltlithave

print -depsc lithave.eps

The last line produces an eps plot file in the directory that matlab is being running from.

To make lithologic column plots UNIX awk scripts and GMT (Generic Mapping Tools, Wessel and Smith) were used. 

The awk script doboxes runs another awk script mkunitbox which successively operates on the *.csv file and creates box files for each rock types. These will define the corners of boxes that GMT draws/fills when it makes the lithologic column plot.

doboxes is run by typing 'doboxes' in a UNIX command line. It is currently set up for filename 'd.csv' with output to a subdirectory 'box' in which several *.box files are written. Abbreviations for rocktype/filename are made and can be understood from the script itself.

To make core interval boxes and corresponding core number labels use Janus output of Hole/Core Summary pasted into an ascii file (e.g. d.txt) then run two awk scripts (mkcorebox, mkcoretxt):

mkcorebox < d.txt > d.cbox
mkcoretxt < d.txt > d.ctxt

The GMT plotting script pltunitbox must be edited to set output postscript plot filename and to modify the directory name where the *.box files exist. In addition the plot intervals must be specified. To make multiple columns plot on a single page the script is run repeatedly, with appropriate command lines commented or uncommented. Dont forget the GMT -O -K convention and make sure the uncommented lines employ these correctly for the results that you want (first (-K only), middle (-O -K), or last run of the script (-O only)). Pattern overlays for grain size are in the later lines of the script.

The postscript files were opened in Adobe Illustrator and a legend was made.

The scripts and programs discussed above are in the folder lithplot archived with the X305 material in the Misc folder.