The Input File

All analyses in meltPT require an input file containing the compositions of samples to be analaysed in the form of a simple csv. In the source code you will find many examples – here we briefly describe what the file must contain, what it may contain, and what will be done with different types of information.

Major elements: Required

An important step in meltPT analyses is backtracking fractional crystallisation by olivine to estimate sample primary compositions (i.e., melt compositions when they were last in equilibrium with the mantle). We assume that crystallising olivine consists of SiO2, MgO and FeO. We need to estimate the composition of olivine that would be in equilibrium with our melt, which is a function of Mg and Fe content. For the these reasons, we require input files to include columns with the headings: “SiO2”, “MgO” and “FeO”; if these columns are not present, the program will crash.

Major elements: Optional

Each of the thermobarometers we inlude in meltPT parameterise pressure and temperature as functions of different major elements. As such, we will look for the following input columns:

  • Al2O3

  • Fe2O3 & FeO_tot (i.e., combined FeO and Fe2O3)

  • CaO

  • Na2O

  • K2O

  • TiO2

  • MnO

  • Cr2O3

  • P2O5

  • NiO

  • CoO

  • H2O

  • CO2

Though none of these are required.

If any of these columns are not present, they will generally be assigned zero values. This procedure is necessary for the smooth running of the code. However, it will lead to inaccurate results if you apply a thermobarometer that relies on an element for which you do not provide a concentration. We therefore urge users to familiarise themselves with the assumptions behind any thermobarometer they use, and carefully consider whether it is appropriate to apply a specific thermobarometer to their data set.

Trace elements and major-element parameterisation

Trace-element concentrations are not generally required for analyses in meltPT but will be used in some specific circumstances.

Concentrations of volatile phases such as H2O and CO2 are notoriously difficult to measure and often not available. An alternative is to use a calibrated proxy to estimate their concentrations. In meltPT, if H2O concentrations are set to zero, we search for a Ce column, expected in ppm, and estimate water content using a specified H2O/Ce value. For CO2 we optionally use a parameterisation based on SiO2 content (Sun and Dasgupta, 2020, EPSL).

Finally, since NiO, CoO, and Cr2O3 are not always treated as major elements, if they are not present, we will search for columns with headings Ni, Co and Cr. Assuming these columns contain concentrations in ppm, they are then used to calculate oxide concentrations in percent.

Primary compositions

We acknowledge that alternative methods exist to correct sample compositions for the effects of fractionaly crystallisation of olivine (and other phases). We therefore include an option to read the contents of the input file as primary compositions. Column headings should be as above, but they will be treated by meltPT as primary compositions. See Tutorial 4: Benchmarking for examples.

Other information

Any other information (e.g. sample metadata, other elemental or isotopic concentrations), will be read in and stored, maintaining the given column headings. These data will not be used by meltPT’s functions but will be readily available to facilitate easy comparision with thermobarometric results.