Quantifying the influence of heterogeneity and preferential flow on the
scale and time dependence of weathering rates

Harihar Rajaram and Sachin Pandey


Numerous previous laboratory and field observations and inferences of weathering rates suggest significant scale
dependence (laboratory rates » field rates) and time dependence (rates appear to decrease progressively with time
of exposure). Preferential flow induced by heterogeneity, manifest as permeability variations, macropores, or dis-
crete fractures, has been suggested as one class of extrinsic mechanisms responsible for the observed scale and time
dependence. Additional intrinsic mechanisms proposed include the decrease in reactive surface area with weath-
ering. In this research, we present a quantitative evaluation of the influence of heterogeneity and preferential flow
on weathering rates using high-resolution reactive transport modeling. We employ PFLOTRAN, an open source
subsurface flow and reactive transport code that utilizes parallelization over multiple processors and provides a ro-
bust framework for simulating the complex weathering patterns resulting from preferential flow. Simulations were
performed in discrete fracture networks (DFNs) and correlated random permeability fields (CRPF). The behavior
in these simulations was compared to that in homogeneous permeability fields. The simulations reproduce to some
extent the scale and time dependence of weathering rates, although the modeled scale and time dependence are
less pronounced than indicated by observations. The simulations in DFNs indicated a systematic time-dependence
related to the formation of diffusion-controlled weathering fronts that propagate into matrix blocks. However, the
decline of system-averaged weathering rates does not follow a 1/sqrt(time) dependence characteristic of diffusion,
due to network scale effects and depletion of matrix blocks. The behavior in CRPF was akin to that in homoge-
neous permeability fields with enhanced dispersion, with a time-dependence that reflects the advective sweeping
of the weathering front from the simulation domain. Our results suggest that structured preferential flow, as re-
sulting from DFNs can readily explain an order-of-magnitude disparity between laboratory and field rates, and a
similar decrease with time over 1 million years. However, to explain the nearly four orders of magnitude de-
crease indicated by observations, preferential flow and diffusion limitations must act in consort with other intrinsic
mechanisms.