Abstract In the 1960s, anticipating the award of concessions in the Gulf of Thailand, BP carried out reconnaissance geological mapping over all of Southern Thailand including its Gulf islands. The resulting eight geological maps at 1:250 000 scale provided a database that underpinned discussions of Thailand's place in global plate-tectonic reconstructions that were emerging at that time. In particular, new light was thrown on the Paleozoic succession, and important differences were found between the Upper Thai Peninsula and the Lower Peninsula (respectively, NW and SE of the prominent bend in the peninsula's outline). The Phuket Group of the Upper Peninsula is a very thick diamictite-bearing succession of possible Devonian to Early Permian age, apparently sourced in the west, a region now occupied by the Indian Ocean. That in turn suggested a possible Gondwana origin for SE Asia, a contentious concept at the time, but one that is now widely accepted. The emphasis of field mapping these days is on obtaining detailed data that necessarily means over areas of limited extent. But a very broad regional mapping project such as BP's in Thailand revealed stratigraphic and tectonic insights that could be revealed only from a wide-ranging regional study of this kind.

Abstract We propose and validate methods for risk analysis of fault-bounded hydrocarbon traps in exploration. We concentrate on cross-fault leakage and consider lateral seals due to (1) juxtaposition and (2) high capillary-entry-pressure fault rock (membrane seal). We conclude that stochastic methods for fault seal analysis are essential, due to the large number of structural and stratigraphic parameters and the uncertainties. Central to the methods proposed is a Monte Carlo simulation which models geometrical and stratigraphic uncertainty. Multiple Allan maps (fault-parallel cross-sections) are produced and analysed for juxtaposition and shale gouge ratio (SGR). For validation, known discoveries with independently observed hydrocarbon–water contacts (IHWC) have been back-analysed. We present two case studies in this paper, and an additional 40 case studies are summarized (four public domain and 36 confidential case studies). The model outputs were compared with the IHWC. Juxtaposition analysis with no SGR contribution gives the smallest error. The inclusion of any fault rock seal mechanisms (such as SGR) matches or increases predicted hydrocarbon column heights compared to juxtaposition and gives larger errors. We conclude there is no reason to include fault rock membrane seals in exploration prospect risking.