In 2001 I began working towards a Master’s degree from Central Washington University under the guidance of Dr. Wendy Bohrson and her collaborator Dr. Michael Clynne, a volcanologist with the United States Geological Survey (USGS). Clynne had published a really cool study of the 1915 Lassen Peak eruption and hypothesized a complex magma mixing origin for the four distinct magma compositions that spewed from the volcano. I got involved to collect crystal size, texture, and compositional data of the abundant plagioclase crystals within the 1915 lavas and to relate these data to magmatic models of the eruption.
Rock collection at Lassen Peak in 2002.
With the help of Dr. Martin Streck at Portland State University, we etched thin sections of the lavas with acid to create compositional-based topography on the plagioclase crystals. The crystals were then photographed (with a film camera attached to the microscope) and a selection were chosen for analysis with the electron microprobe. I made my first visit to Corvallis and Oregon State University where Dr. Roger Neilson was running the probe lab and helped set up the analyses.
You can see the textural discontinuities on the plagioclase crystal on the left. The dots are where each microprobe analysis were made. The corresponding chemistry (An = anorthite, the Ca-rich endmember of plagioclase)) along the traverse is shown on the right.
The magmas showed disequilibrium textures indicative of magma mixing.
Plagioclase dissolution textures surrounded by calcic (high An) growth rims (left) are often thought to result from magma mixing, as is quartz dissolution with pyroxene halos (right).
Mixing is also indicated by the many mafic inclusions (left), and heterogenous glass (right).
The crystal size distributions (CSDs) involved getting Al-concentration maps of the thin sections (using the microprobe at OSU) and counting the crystals with image software. Below is a .gif summary of the process.
The CSDs clearly showed three crystal populations with distinct crystallization histories. The plagioclase compositional data matched up very well with each size range characterized by distinct elemental differences. The larger phenocrysts seem to be from a silicic magma with long-lived (1000s of years) crystallization, whereas the smaller microphenocrysts are from a more mafic magma with much shorter (days to years) crystallization times that we interpreted as forming during magma mixing. The smallest population, the microlites, likely formed in the hours or days during decompression and eruption.
This image from our 2008 paper in Journal of Petrology shows the strong correlation between crystal size and core chemistry. We interpreted each of the three size classes as representing a unique phase of crystallization.
Frank Ramos and the lab at Washington State University also helped us to collect isotopic (87Sr/86Sr) data on select plagioclase phenocrysts. We found that there was significant heterogeneity in a few of the plagioclase demonstrating that these crystals had some pretty complex, open system evolutions.
Here are isotope data for two points in the same crystal.
Overall, we were able to support much of Clynne’s magma mixing story and to quantify the process in more detail. In addition to our interpretations, we were also able to provide quality data for a major debate in the petrologic community concerning the timescales, crystallization processes, and ascent histories of magmas. Here is a summary of our interpretation of the mixing process.
1. Addition of mafic magma to phenocryst-rich silicic magma.
2. Mixing between magmas forming a hybridized andesite with rapid plagioclase growth
3. Vesiculation and disaggregation
4. The four rock types prior to eruption
5. Further mixing induced by the eruption itself.
see the complete study in Salisbury et al. 2008.