Critical evaluation of chemical mohometry 1: Calibration datasets and influences of intensive parameters on arc magma geochemistry

Chemical mohometry refers to the use of the geochemistry of arc magmas and/or magmatic minerals to infer depth to the Moho. This increasingly popular method in tectonics research relies on correlations between Moho depth and either trace element concentrations or their ratios, e.g., Hf, Ba, Sr/Y, La/Yb, Th/Yb, europium anomaly (Eu/Eu*), etc. These correlations are widely viewed to reflect depth-dependent changes to mineral stabilities, especially for feldspar and garnet. Here, we revisit the foundational whole-rock datasets to assess the influence of data scatter on calibrations. We also assess the effects of the following factors that influence mineral stabilities and melt geochemistry: oxygen fugacity (f_O2), water fugacity (f_H2O), temperature, and fractional crystallization of major and accessory minerals.

 

The following, depth-independent behaviors are evident:

• Data scatter is large – a single Moho depth is typically represented by 50-75% of the global range in a single trace element concentration or ratio. Calibrations are statistically robust only after averaging hundreds of data points.
• Different arcs have systematically different compositions, even at the same Moho depth, leading to systematic errors of ~10 km for some systems.
• Increasing f_H2O can depress plagioclase stability by 200 °C and increase garnet stability by hundreds of MPa. These changes to mineral stabilities strongly influence alkaline earth and rare-earth element concentrations and ratios (e.g., Sr/Y, La/Yb, etc.).
• Increasing f_O2 changes mineral assemblages and modes, and generally increases Eu/Eu* and apparent Moho depth.
• Increasing temperature changes melt fraction and mineralogy, and can strongly influence Sr/Y, La/Yb, Th/Yb, and (at low pressure) Eu/Eu*.
• Crystallization of accessory minerals, especially titanite, can strongly increase La/Yb. Sr/Y, and Eu/Eu* in separated liquids, leading to spuriously increasing calculated Moho depth.
• Realistic variations in f_O2, f_H2O, temperature, and fractionation can each shift calculated Moho depth by 10-20 km.

 

These behaviors lead to the following conclusions:

• Application of mohometry to the rock record should average hundreds of measurements per time slice. Such large datasets are rarely available, and may not be feasible to collect.
• Concurrent changes to temperature, f_O2, and f_H2O should be quantified, otherwise uncertainties in either calculated Moho depth or changes to Moho depth through time are tens of km.
• Crystallization sequences in genetically related magmas should be assessed to determine whether crystallization of trace phases affects trace-element concentrations and ratios used for mohometry.
• Systematic sampling errors are difficult to avoid and lead to systematic but unknown errors in estimated Moho depth.
• Extremely large datasets (hundreds of measurements per time slice) are rare, and simultaneous changes to f_O2, f_H2O, temperature, and melt fractionation are difficult to estimate. Consequently, future success in mohometry will require major streamlining of data collection and development of routine geochemical proxies for intensive parameters.
• Calculations to date of Moho depth using geochemical mohometry do not consider covariation in intensive parameters, melt fractionation, and/or sampling errors, so should be viewed as unreliable.