A numerical model based on the convective-diffusive transport of reacting and adsorbing proteolytic enzymes within erodible fibrous biopolymers was used to predict lysis fronts moving across biogels such as fibrin or collagen. The fiber structure and the transport properties of solutes in fibrin (or collagen) were related to the local extent of dissolution within the dissolving structure. An accounting for solubilization of adsorbed species into solution from the eroding fiber phase provided for complete conservation of mass in reacting systems containing over 10 species. At conditions of fibrinolysis typical of clinical situations, the model accurately predicted the dynamic rate of lysis front movement for plasmin, urokinase, and tissue plasminogen activator (tPA)-mediated lysis of fibrin gels measured in vitro. However, under conditions of extremely fast fibrinolysis using high enzyme concentrations, fibrinolytic fronts moved very rapidly (>0.1 mm/mm)-faster than predicted for diffusion-limited reactions-at nearly constant velocity for over 2 h, indicating non-Fickian behavior. This was due to proteolysis-mediated retraction of dissolving fibrin fibers that resulted in fiber convection and front-sharpening within 3 mu m of the reaction front, as observed by digitally enhanced microscopy. In comparing the model to fibrinolysis measurements using human lys(77)-plasmin, the average first order rate constant for non-crosslinked fibrin bond cleavage by fibrin-bound plasmin was calculated to be 5 s(-1), assuming that 10 cleavages per fibrin monomer were required to solubilize each monomer. The model accurately predicted lysis front movement using pressure-driven permeation of plasmin or urokinase into fibrin as well as literature data obtained under well-mixed conditions for tPA-mediated fibrinolysis. This numerical formulation provides predictive capability for optimization of proteolytic systems which include thrombolytic therapy, wound healing, controlled drug release, and tissue engineering applications.