Reduction of bedrock blocks as corestones in the weathering profile: Observations and model

Blocks of metavolcaniclastic rock exposed at the surface in upland catchments of the Espiritu Santo and Mameyes rivers within the Luquillo Experimental Forest (Puerto Rico) are interpreted as corestones, reduced from initial joint-bounded bedrock blocks by subsurface weathering. Maximum corestone size, expressed as the geometric mean of the three dimensions, S = (3)root abc, shows a smooth envelope when plotted against elevation. We postulate that, for each catchment, they represent in situ corestones within a stratified weathering profile, many tens of meters in thickness, that has been subsequently exhumed by younger erosion. We formulate a simplified one-dimensional model for reduction in corestone size within a steady-state weathering profile that incorporates: (i) vertical fluid transport of the reactant and the soluble products of chemical weathering; (ii) linear kinetics of corestone reduction; and, subsequently, (iii) erosion. The rate of advance of a steady-state weathering profile is a statement of the mass balance between entering reactants and weathering components, here idealized as H+ and albite. The mathematical relations, 1/2 k nu lambda/f(AR) (gamma c(R))(N) = K(gamma c(R))(N) = V(S-0/6L), tie the laboratory-determined rate constant for dissolution of albite (k) to a generalized kinetic constant for the rate of decrease (K) in corestone diameter to the advance rate of the weathering profile (V). The last parentheses contain an effective roughness at the scale of the weathering profile, where So is the maximum size of initial bedrock blocks, inferred to be set by initial bedrock fracture spacing, and 3L* is the profile thickness. The laboratory scale roughness value, X, is the ratio of the surface area accessed by BET analysis to that of the corestone grain scale. In the model, erosion is not coupled with weathering, although the presence of corestones of finite size, S-E > 0, exiting at the erosional surface may be postulated to affect the erosional flux. The thickness of the corestone weathering profile derived for the model for the distance between bedrock and a corestone-free saprolite cap is approximately 3L* = (beta qc(R))(Mab/f(ab)rho(ab))S-0/2[K(gamma c(R))(N)]. This expression is the product of the effective pH buffering-adjusted input reactant flux per unit area times a stoichiometeric factor linking this to net albite dissolution, divided by the rate of corestone size reduction at the input concentration of protons. Further, the profile thickness scales with the input "particle" size, S-0. The model fit, which yields the ratio S-0/3L* approximate to 0.02, is consistent with a rate constant for albite dissolution that lies between laboratory-measured and field-estimated values. Sensitivity to the reaction order of albite dissolution with respect to H+, N, is small, except near the base of the profile. This model yields insights into the relationship between fracture spacing and the evolution of particle size and chemistry in weathering profiles.