Ground affected by periglacial and glacial processes can be among the most variable formed by nature. Previous chapters have graphically illustrated this variability and explained the topographic and sedimentary associations to be expected within former and present-day cold regions. This chapter shows how that background is needed to design and execute an investigation for predicting either the ground response to engineering change or the volumes of material the ground contains. Such an investigation of the ground is also needed to explain its current and former state of stability on slopes and its natural groundwater flow.The starting point of any such investigation is a conceptual model of the ground which subsequent investigation tests and refines; investigations conducted without such a model can easily become sterile and expensive exercises in collecting data. Such a model starts with knowledge of landscape, cold climate processes and their products, initially refined with the aid of a desk study. This then develops with each phase of the investigation, starting with what is known via desk studies, and progressing through what can be readily seen by walkover surveys and shallow investigations, including surface geophysics and remote sensing, all leading towards a model that can be tested directly by various intrusive investigations. Techniques appropriate for such investigations, including sampling, in glaciated and frost-disturbed ground both onshore and offshore are reviewed.Great care must be taken with the description of coarse materials, glaciotectonic structures and the materials within them; a unique feature of this chapter is the correlation it presents between the engineering descriptions of glacial sediments, as used in ground engineering, and the descriptions used by glacial sedimentologists for the same materials. Water levels are also obtained during these investigations, and in these types of ground they are often misinterpreted by applying thinking more appropriate to aquifer hydrogeology. A surprising feature of glaciated ground is its low permeability overall, and the correct interpretation of heads measured in such environments is often that for aquitards rather than aquifers.The initial conceptual model starts with little more than an idea and a broad outline, and evolves as the investigation progresses. It should continue to evolve throughout construction as more and more of the ground is exposed and its behaviour is better known; in this way, the ground model can be thought of as a living document, especially appropriate in such variable ground. The chapter concludes with a review of how this information can be brought together as three-dimensional models that effectively communicate the knowns and unknowns of a volume of ground and their associated risks, in both deterministic and probabilistic ways.
