fixed it

[kareina]

Back in January I attempted to do garnet fractionation calculations for one of my samples. This means that I take a guess as to what changes to temperature might have happened, define it in terms of change in temperature with respect to pressure, and tell the computer to calculate how the growth of garnet will change the composition of the rock if the temperature and pressure changes along the path I've selected, and as the garnet grows the ingredients it uses up are "locked" into the garnet core, and so unavailable for further chemical reactions, but *everything* else in the rock is available for reactions. Once it has done that, I open up the file, copy over the data and create a graph which shows the changes to the garnet composition over time & has horizontal lines which correspond to the composition of the outermost garnet rim that was actually measured in the sample. A path is considered to have "worked" if the iron, magnesium, calcium, and manganese for the calculated garnet intersect those horizontal lines all at one pressure (and temperature), and doesn't work if A) one or more of those elements never reaches the horizontal line on the graph showing the target composition, or B)they all reach their target goals, but a quite different pressures. Needless to say, it often takes lots of path attempts before finding the one which will work for any given sample.

Today's project was writing up the results for the above mentioned sample. When I got to the part about the garnet fractionation calculations, I looked at what I'd done, and typed something about how 10 different fractionation paths had been attempted, and none of them worked, and given the pattern displayed, one wasn't possible for this sample. Then I thought about it again, looked to see what numbers I'd used for the garnet rim "target" goal (above mentioned horizontal lines) for the paths and realized that the "rim" I'd chosen doesn't actually represent the composition of most of the garnet rims. So I went through and changed all 10 graphs from January's work to show the new, improved, target goal horizontal lines. Suddenly, instead of the combination of graphs showing that there is no path possible for the theoretical garnet to match the rims, it looked like there just might be one, about half way between this path and that path. So I ran another path, and then one more for good measure, and lo and behold, on the newest path, with the improved target it is predicted that the garnet would grow at the same composition as measured in my sample.

That process took most of the day, and once done and the results typed up, I've only 568 new words for my thesis, and another 7 tables/figures referred to. Not as many words as I'd hoped to write today, but at least I didn't have to leave the "tried ten paths, none worked" version of the paragraph in the thesis, but instead can say "tried 12, this one works". (only more eloquently, and with all of the details of what and why, of course!)

Time for yoga and bed!