The observational method describes a risk-based approach to geoengineering that employs adaptive management, including advanced monitoring and measurement techniques, to substantially reduce costs while protecting capital investment, human health, and the environment. Development of the observational method in geoengineering is generally attributed to Terzaghi (Casagrande, 1965; Peck, 1969). The method consists of the following steps (Peck, 1969):
The observational method has several caveats. One must be able to define an action plan for every conceivable adverse condition. The method cannot be used if you cannot develop a predictive model for the behavior (i.e., you must have a model that can calculate the parameters you will subsequently observe). You must be able to monitor the parameters you can predict. This is not a trivial problem as often we can measure what we cannot calculate and vice versa. This means that the monitoring plan must be chosen very carefully with a good understanding of the significance to the problem. Mistaken preconceptions about the dominant phenomena that control system behavior can lead to choosing irrelevant observational parameters and cause the method to fail.
Casagrande (1965) described limitations to the use of the observational method in his classic geotechnical paper on “The Role of the Calculated Risk in Earthwork Engineering.” Casagrande postulated that risks inherent to geotechnical practice include engineering risks and human risks, calculated risks and unknown risks, and voluntary risks and involuntary risks. Calculated risks are risks based on uncertainties associated with engineering analyses of known phenomena. Casagrande called for the use of the observational approach (i.e., an adaptive management method employing instrumentation and monitoring) to manage calculated risks.
The observational approach is also embodied in what is sometimes referred to as “adaptive management,” or “staging,” approaches to complex engineering problems. Like the observational method, adaptive management is designed to be used on problems where it is not possible to definitively predict the outcome of engineering choices because the system is too complex, the processes are not well enough designed, or the systems cannot be characterized adequately. The use of adaptive management to deal with a seemingly intractable geoenvironmental problem is discussed with respect to the Yucca Mountain Project in the NRC report One Step at a Time (NRC, 2003c). The method is most applicable when the project is one of a kind, the methods and the outcomes are controversial, and the consequences of the project will take a long time to evolve.