Geostatistics: Roadblocks and Challenges

Misconceptions about the role of random function models may have distracted geostatistical research into unrewarding debates about model properties and led some practitioners into using more complex algorithms when simpler ones would have sufficed.Typical of such roadblocks are concerns about stationarity, ergodicity, model consistency when random function models have no intrinsic siginificance. They are but data analysis tools and, as such, can be bent, retooled or simply discarded.The challenge proposed to geostatistical theory is to data-charge its algorithms to account for an ever increasing variety of information, of both structural nature (multiple-event dependence beyond traditional two-point covariances) and local nature (hard and soft data). Last, like recent developments in simulated annealing, these algorithms need not be probabilistic.