The Kansas Geological Survey has actively pursued a long-term interest in petroleum reserve estimation; the earliest Kansas reserve appraisals were published in 1896. Periodically updated reserve assessments have been issued routinely and are currently monitored by a computer file system. The major decline in Kansas crude oil production since the mid-1950s has placed a key emphasis on estimation of potential undiscovered reserves as a decision guide to future petroleum exploration in the state. Initial work in this area was undertaken by Griffiths, who applied the unit-value concept to the regional evaluation of potential mineral wealth across Kansas.

Techniques of regional resource evaluation may be related to two basic analytical models. In a causal predictive model, estimation of unknown reserves is derived as the effective outcome of causative basin parameters such as geometry and magnitude of source beds and potential reservoir formations. In contrast, an empirical predictive model entails the extrapolation of regional production history into the future, utilizing statistical time-series methods. The Survey is currently developing the Kansas Oil Exploration (KOX) Decision System incorporating features of both these models in a capability for estimation of unknown reserves in either local or regional areal contexts. Information sources of drilling histories and diagnostic geologic variables are analyzed interactively in a condi ional probability framework geared to likelihood of undiscovered production.

The varying degree of oil exploitation across the state permits the application of statistical "experience" from more mature areas to the analysis of less-developed regions. Because the perception of basin characteristics undergoes a progressive evolution with the exploration history of a basin, the consequent dynamic character of key variables is selectively analyzed in a structured "re-experience" model of basin development. The production outcomes of wildcat areas are expressed as conditional probabilities drawn from contingency tables relating known production to perceived subsurface geology. These probabilities, which may be mapped areally, provide basic data for exploration decisions or economic analysis.