Preliminary scaling model for fracture lengths in a strike-slip tectonic setting: the Liquiñe-Ofqui Fault System and the Andean Transverse Faults, Southern Andes (39-40ºS).

Abstract

Geological fracture networks govern fluid flow in the upper crust. This implies that theirgeometry, size and spatial distribution directly influence the nature and development ofshallow magmatic and hydrothermal systems, especially when fractures occur in impervious,crystalline rocks. However, fractures occur as complex meshes in a wide range of sizes andorientations, in which the scaling nature of their geometrical features (e.g. length, aperture)may follow different statistical distributions. One of the basic distributions defining themulti-scale properties of flow is the power-law length density distribution, expressed asN(L,l)=αLDl−a. This expression relates the number of fractures N(L,l)dl with lengthsbetween l and l+dl in a system of given size L. The distribution is defined by thefractal density term α, the mass dimension D, and the exponent a representingthe balance between small and large fractures of the system. We chose a portionof the Southern Volcanic Zone of the Andes (SVZ) (39-40ºS) as a case study totest the consistency of the scaling model. In the SVZ, recent studies suggest thatcrustal flow is controlled by two groups of faults: the Liquiñe-Ofqui Fault System(LOFS) and the Andean Transverse Faults (ATF). Regionally, the LOFS is an active,intra-arc fault system composed of NNE-striking dextral and dextral-reverse masterfaults and NE to ENE-striking subsidiary faults with dextral and dextral-normalkinematics. The ATF are apparently older than the LOFS, and include a set of NW toWNW-striking faults and morphotectonic lineaments that show sinistral and sinistral-reversekinematics. To define a preliminary scaling model for flow, we produced and measureddifferent fracture trace maps at the regional-, outcrop- and thin section-scales. A totalof 3347 fractures mainly hosted in grandioritic to tonalitic rocks were measured,with lengths ranging in eight orders of magnitude (L ∼104-10−3 m). Assuming arepresentative mass dimension D = 1.8, the obtained best fit is a normalized power-lawlength density distribution n(l)=N(L,l)/LD, as given by the expression n(l)=3.5l−2.8.These preliminary results are in agreement with distributions reported in fracturedsystems, and indicate that fracture connectivity is dominated by both microscopic andmacroscopic fractures, which excludes permeability within individual fractures or classicalpercolation theory as applicable models for the multi-scale flow structure of the studyarea.