Methane formation at Costa Rica continental margin - constraints for gas hydrate inventories and cross-decollement fluid flow

摘要

We present a numerical model study in order to quantify the effects of organic carbon (POC) degradation and fluid migration on methane and gas hydrate formation at ODP site 1040 (Costa Rica convergent margin). Various model runs show that POC-degradation in upper plate sediments yields a potential for methane hydrate formation between 0.8 and 2.5 vol.% of pore space. However, observed chlorinity anomalies cannot be explained by the amount and the distribution pattern of gas hydrates. Moreover, pore water profiles of ammonia do not match the observations. Setting up a moderate upward flow (0.03 cm yr(-1)) Of methane-enriched, low-chlorinity fluids (induced by dewatering of oceanic plate sediments) leads to a good approximation to measured pore water profiles, thus enabling a precise estimate of POC degradation kinetics. Fluid flow has a strong impact on the location of the upper limit of the modeled gas hydrate occurrence zone (GHOZ) and may increase the total amount of gas hydrate by more than 50%. Our best estimate of the amount of gas hydrate within the GHOZ is on average 1.65 vol.% of pore space, which corresponds to about 2.5 Tg of methane per km trench within the frontal prism of slope sediments.To comply with the fact that subducted pore waters are rich in sulfate and that there is striking evidence for fluid conduits at various depths we performed additional model runs, where we simulated fluid flow by using a Gauss-type rate law, allowing us to define distinct fluid sources. We can demonstrate that combined methane production in the upper plate sediments and sulfate reduction at the top of the down going slab is sufficient to prevent the upward movement of the zone of anaerobic oxidation of methane (AOM) to above the decollement at given upward advection rates. Steep pore water gradients along the plate boundary can be explained by lateral backflow within oceanic plate sediments. On a long term (in the order of at least some 100,000 years), fluid flow along conduits is likely to occur at low rates with temporarily increased pulses. All modeled runs are constrained by their compatibility to observed pore water profiles. (c) 2005 Elsevier B.V. All rights reserved.