Biogeochemistry of granitic weathering

Open Access
Moore, Joel
Graduate Program:
Doctor of Philosophy
Document Type:
Date of Defense:
June 06, 2008
Committee Members:
  • Susan Louise Brantley, Committee Chair
  • Jennifer Macalady, Committee Member
  • Maryann Victoria Bruns, Committee Member
  • Lee Kump, Committee Member
  • granite
  • chronosequence
  • mineral
  • weathering
  • soil
Dissolution of minerals in soils and saprolite plays many important roles in the earth system including the release of nutrients into soils, the neutralization of acid deposition, and control of atmospheric CO2 concentrations on the million-year time scale. In this dissertation, aspects of mineral dissolution and weathering were studied in granitic chronosequences. A chronosequence is a set of soils where four of the five soil-forming factors–climate, lithology (or similar parent material), organisms, and relief are similar. Chronosequence soils are differentiated by their age, or by time, and are useful natural laboratories for studying shifts in geochemistry and ecology through time. Two granitic chronosequences were studied as part of this dissertation: Santa Cruz and Merced. The Santa Cruz chronosequence is located on the coast of California near the city of Santa Cruz. The Merced chronosequence is located along the Merced River in the western foothills of the Sierra Nevada, Ca. To provide kinetic data for modeling chronosequence weathering, dissolution rate data for gibbsite, &#948;-Al2O3, and the sheet silicate minerals (SSM) chlorite, kaolinite, vermiculite, illite, and smectite were compiled and synthesized (chapter 2). The synthesized data sets were used to propose dissolution rate versus pH rate laws for each of the minerals. Rate laws produced from several data sets are more reliable than data sets from any single investigator, and comparison of kinetic data sets from multiple minerals may point to similar dissolution processes. For example, the differences among the dissolution rates of the common soil minerals illite, smectite, and kaolinite were <0.5 log units. Some of the rate laws proposed were used as model input for the reactive transport modeling conducted in chapter 4. To understand how subsurface biota varies across chronosequences, the size and composition of soil microbial communities were investigated at Santa Cruz (chapter 2). Soil microbial community size and composition changed as a function of soil age with more significant changes in the subsurface. The shifts in size and composition were primarily correlated to C within a soil (with C decreasing as a function of depth) and with chemical changes (driven by mineral dissolution) between soils. In the subsurface communities, the proportion of fungi increased with soil age, consistent with the fungal proportion increasing as nutrient concentrations decreased. Microbial community activity may feedback into weathering processes by promoting kaolinite and Fe oxide precipitation. Precipitation is promoted as microorganisms consume organic ligands complexed with Al and Fe, releasing the Al and Fe back into solution where precipitation into the solid phase is likely. A reactive transport code, FLOTRAN, was successfully used to model mineral dissolution in granitic soil and saprolite as old as 3000 ka. FLOTRAN was used to explore a number of factors proposed to be responsible for the three to five order of magnitude difference in laboratory and field rates. The most important factors controlling the depth and thickness of plagioclase reaction fronts were the pore fluid flow rate, the reactive surface area, and to a lesser extent the form of the rate law (i.e., transition-state theory or sigmoidal). Successful model predictions were achieved by adjusting the pore fluid flow rate to match the depth of plagioclase reaction fronts and by adjusting the reactive surface area to match the thickness of plagioclase reaction fronts. The best models, which included a sigmoidal rate law, predicted the difference between laboratory and field rates to be less than one order of magnitude for the Davis Run, Virginia (USA) saprolite (800 ka) and approximately two orders of magnitude for the Merced, California (USA) chronosequence (40 - 3000 ka). The smaller difference for Davis Run where >3 times as much precipitation falls as Merced indicated that heterogeneous hydrology, such as dead end pores, may slow rates by at least an order of magnitude in dry soils. A model that successfully predicted soil mineralogy after 250 ka of weathering also predicted aqueous Na+ concentrations and pH reasonably well.