Seismic-pattern recognition applied to an ultradeepwater oil field

Quantitative seismic interpretation is key in the integration of geoscience data on scales of basinwide, reservoir, and field-development studies. Quantitative modeling can be deterministic or probabilistic. Probabilistic modeling can be divided into two groups: multivariate statistics and geostatistics approaches. Close to probabilistic modeling is the neural network method. Neural network modeling was applied to seismic-pattern recognition (seismic-facies analysis) in an ultradeep-water turbidite oilfield reservoir in Campos basin, offshore Brazil.