Bayesian thermal history modelling of detrital thermochronometric data
Résumé
We have developed a Bayesian inversion
approach to modeling both detrital and bedrock
thermochronometric data. Following the approach
presented in Gallagher (2012), we use Markov
chain Monte Carlo to sample many candidate
thermal histories models. We use the present day
hypsometric curve in a drainage basin as a
starting point to sample age-elevation profiles
predicted for each candidate thermal history. From
these we can then predict the detrital age
distribution for a detrital sample representative of
the catchment. In principle, discrepancies between
the predictions and the observed data may allow
us to refine the sampling of the age-elevation
profile and infer a detrital sampling distribution
different to that implied from the hypsometric
curve. The methodology can be applied to the
profile data alone, combined profile-detrital data,
or just the detrital data alone. The results do not
differ too much, implying that detrital
thermochronological data can be used directly to
reconstruct thermal histories of a catchment. We
give examples of the method applied to apatite
fission track data (AFT) data from small (< 900
km2) river catchments from the Santa Marta
Sierra Nevada in northern Colombia with
elevations up to 5.8 km. The resultsreveal spatially
variable, episodic exhumation with a major peak in
middle to late-Miocene (30-15).
Gallagher, K. (2012), doi:10.1029/2011JB00882