This paper describes a practical method for developing predictive soil models for large offshore areas. The models are developed by synthesizing geologic background information, high-resolution seismic reflection data, and soil boring results. The geologic information provides reasonable limits for the predicted range in soil conditions. Seismic data define the three-dimensional relationships among soil units and provide a framework for relating soil borings to one another. The soil borings 'calibrate' the model by defining soil stratigraphy and geotechnical properties at key locations. Components of a soil model are: a soil province map showing areas where soil conditions are similar; soil cross sections to show the vertical relationships among soil units; and a representative soil profile for each soil province. Soil models can be used to predict general soil conditions throughout a study area, but not exact soil conditions at specific sites. Reliability of predictive models varies according to geologic complexity; the amount, type, and quality of data used; and the experience of those who develop the model. The method described has been used successfully to develop models for several Alaskan OCS areas and other frontier areas.
The purpose of this paper is to describe a practical method for predicting general soil conditions offshore, and to discuss the limitations and factors affecting the reliability of this method. The method is most applicable for predicting soil conditions of large frontier areas tens of miles on a side, but ranging downwards to areas about the size of an individual lease block. Knowledge of general soil conditions is needed for conceptual and preliminary design of foundations, for various pre-lease economic analyses, for feasibility studies, for preliminary drilling and construction planning, and for planning geotechnical site investigations. Predicting soils at individual sites is not specifically addressed here, but aspects of the general method and its limitations as described here would also apply to site investigations.
The method consists of first developing a geologic model and then transforming the geologic model into a predictive soil model. The models are developed by synthesizing geologic background information, high-resolution seismic reflection data, and soil boring data. The method involves considerable interpretation and judgment, but is the most rational method readily available today for predicting general soil conditions for large areas.
Geophysical techniques are being developed that will complement the seismic data now used. These techniques are anticipated to provide direct identification of materials and to significantly increase the reliability of regional soil models. However, none of these techniques are expected to be perfected for geotechnical applications and be routinely available on an operational basis for at least several years and probably much longer.
The method described here involves three basic steps;
development of a geologic model;
transforming the geologic model into a preliminary soil model; and
calibrating and refining the soil model.