GeoScPlot is a Python "library" created to provide nice, high-quality (high-resolution) templates for common discriminant plots used in Earth Sciences. Basically, all plots are those I often used while studying Geosciences at Universidad de Los Andes, Bogotá, Colombia.

Templates are built over Matplotlib and Ternary so you should install them both beforehand.

The package is also capable of plotting a bunch of geochemical discriminant plots from (geochemical) data, for this, I used McDonough & Sun's data on chondrites and MORB. To test it, I used data that was provided to me in the courses of Geochemistry (2015-1) and Petrology (2016-2). In principle, any set of geochemical data can be processed to be used with the library, here a few important steps.

The library relies strongly on Pandas, so install that, too.

• Import raw dataset as pandas DataFrame.
• Make sure that the sample name is the DataFrame's index and that each oxide/element is a column, avoid repetition.
• I believe anyone would be able to workaround an unexpected issue. For example, the code is designed to plot 6 samples of different colors, I believe there would be trouble with more samples than that but a look at the source code should give the answer (which is: put more colors in loscolores list.)

There are a handful of quite useful templates, divided in two "modules"

To use it, save the GeoScPlot folder in your working directory and import it as

from GeoScPlot.TwoDPlots import *

This module contains:

• convertToAnhBase() given a LOI column
• buildTas() Total Alkali vs. feldespar
• plotTAS(data)
• plotHarker(data) Harker plots (plenty)
• buildZrTi_NbY() Igneous rock classification
• plotZrTi_NbY(data)
• plotREE(data) Rare Earth Elements plot
• plotSpider(data) Spider plot
• buildShervais() Shervais discriminant for igneous provenance
• plotShervais(data)
• buildThTa() Igneous provenance classification/ series classification
• plotThTa(data)
• buildPierce() Igneous provenance
• plotPierce(data)

All build* functions return a figure, ax, so they are to be used as fig, ax = buildTas(). This will return the template for the TAS discriminant.

All plot* functions have keyword arguments save and nombre (sorry for the spanglish) so they can be called as plotTAS(data, save=True, nombre = "proyectoPetrologia"). Default save resolution is 300dpi (which I think is more than enough).

A Shervais diagram would look like

Some examples may be found here.

To use it, save the GeoScPlot folder in your working directory and import it as

from GeoScPlot.TernaryPlots import *

These were usually a pain to plot in a nice way, until I found Ternary library. Nevertheless, the templates were still missing, so I made them. This module does not include plotting functions, but I'll show how to use the templates since I find them easier to use on their own.

This module contains:

• buildProvenancePlot() Tectonic provenance classification
• buildSSTClassification() Sandstone classification
• buildPettijohnSST() Pettijohn sandstone classification
• buildPettijohnWacke() Pettijohn wacke classification.
• buildCong() conglomerate/gravel classification.
• buildQAP() the upper part of the QAPF diagram
• buildAPF() the other part of the QAPF diagram.

These functions return figure, tax. You should read Ternary's docs for accurate information on tax (TernaryAxesSubplot) objects, but they work quite similarly to AxesSubplot from Matplotlib.

from GeoScPlot.TernaryPlots import *

Get the template that you need, for example let's say you are to classify an igneous rock using the QAP diagram.

fig, tax = buildQAP() #Plots the template

Your data (sample composition) must be in a python tuple of the form rock = (P, Q, A), although you can have rock = (P,Q) and Ternary will figure out A since they all must add 100.

tax.scatter([rock], s = 100, marker = 's', c='r') #Plots the dot with size 100 and a red square marker.

In general, data should be passed as a list of tuples or lists, one per sample. These samples must be organized as (Bottom, Right, Left) = (Bottom right, Top, Bottom left), the latter in terms of triangle vertices.

Now, let's use other template:

f, tax = buildCong()
rock = [50., 20.] #50% Sand, 20% Gravel, then 30% Mud
tax.scatter([rock], s =100, marker ='s', c='r')
plt.gcf()
plt.tight_layout()
plt.show()

For more details on the use of the ternary library, visit their GitHub.

• W.F. McDonough and S. -s. Sun, Chemical Geology 120, 223 (1995).
• J.A. Pearce, T. Alabaster, A.W. Shelton, and M.P. Searle, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 300, 299 (1981).
• J.W. Shervais, Earth and Planetary Science Letters 59, 101 (1982).
• S. -s. Sun and W.F. McDonough, Geological Society, London, Special Publications 42, 313 (1989).
• I.F. Blanco-Quintero, A. García-Casco, L.M. Toro, M. Moreno, E.C. Ruiz, C.J. Vinasco, A. Cardona, C. Lázaro, and D. Morata, International Geology Review 56, 1852 (2014).