Using mineral forensics to determine the nature of gabbro: primitive mantle melt or crystal cumulate?

Distinguishing between cumulate and non-cumulate gabbros is important when interpreting petrogenesis. The composition of gabbro bulk rock, e.g. high Mg#, is often used to infer a mantle source; however, quite often gabbros represent crystal cumulates skewing this interpretation. Mineral accumulation may or may not be apparent texturally or from bulk composition alone. We used petrography, bulk chemistry, and electron microprobe and LA-ICP-MS mineral chemistry to distinguish cumulate vs non-cumulate gabbros on six gabbro samples, with primitive Sri values, from three Cretaceous arc segments: the Peninsular Ranges Batholith (PRB) of California, the Peruvian Coastal Batholith (PCB), and the Chilean Coastal Batholith (CCB). Plagioclase petrochemistry was integrated to reconstruct the melt in equilibrium with plagioclase crystals. Three types of gabbros were found: A type I gabbro has non-cumulate texture, Mg# 0.8, unzoned plagioclase with core compositions An95-87, and almost constant core-to-rim Sr values (860-630 ppm). These plagioclases were in equilibrium with the bulk rock and crystallized in a closed magmatic system. The type II gabbro has cumulate textures, Mg# from 0.7 to 0.5, plagioclase with zoning patterns and resorbed core and mantle zones, and widely varying core-to-rim Sr (1130-350 ppm) and An90-30 values. The type III gabbro has cumulate textures, Mg# from 0.7 to 0.5, very few plagioclase cores and mantles (less than 10%) with largely constant Sr and An values that are in equilibrium with the bulk rock, followed by an abrupt decrease in both in the plagioclase rim, representing the residual felsic melt before complete crystallization. We found that the Green Acres-PRB is a type I gabbro, PCB and San Marcos-PRB are type II gabbros, and the CCB is a type III (cumulate) gabbro. Only the PCB and PRB gabbros contain plagioclase in equilibrium with their bulk rock compositions. We found that cumulate gabbros naturally exhibit disagreements between bulk rock and plagioclase chemistry derived melt compositions (e.g. Sr values) and can have high or low Mg# dependent on the cumulate phase. This challenges the use of Mg#s and high concentrations of compatible elements to interpret the petrogenetic nature of gabbros. We recommend the use of plagioclase chemistry in comparison with bulk rock compositions instead.