Diffusion is the spread of a phenomenon, such as an idea, a technological innovation, or a disease, over space and time. The origins of interest in diffusion in geography can be traced back to the work of the German geographer Frederick Ratzel (1844–1904). In the second volume of his highly influential Anthropogeographie, Ratzel described the diffusion of cultural traits. This work laid the foundation for the studies by geographers on cultural history, which came to be most influential in the work of Carl Sauer (1889–1975) and the geography department at the University of California, Berkeley, from the 1930s through the 1950s, which is referred to as the Berkeley School.
Sauer (1952) argued that issues of cultural diffusion should be 1 of the main concerns of geography. Sauer believed that the diffusion of ideas, such as agriculture, from "cultural hearths” (cultural centers) has been 1 of the main driving forces in human history. The research undertaken by the Berkeley School also influenced the study of the origin and spread of culture in the 1960s and, more recently, the study of environmentalism and the processes of cultural globalization.
Another key geographer in ideas of diffusion was the Swede Torsten Hägarstrand (1916–2004). His 1952 doctoral dissertation on innovation diffusion as a spatial process had limited impact until its publication in English in 1967. Hägarstrand observed that the diffusion process can be likened to a wave pattern that loses its strength as it moves away from its source of origin.
Hägarstrand provided a mathematical basis for simulating innovation diffusion on the basis of probability although it is rare for diffusion to be completely random. Each location is regarded as having a different set of probabilities for diffusion based on a mean information field that structures the way in which diffusion flows through a region. Four different types of diffusion are usually recognized.
In expansion diffusion (also referred to as contagious diffusion), a phenomenon, such as knowledge of an innovation or a disease, is spread by direct contact or word of mouth. This type of diffusion exhibits the frictional effects of distance by which those furthest away are less likely to receive this information than those closer to the initial source (distance decay).
A more developed model of the simple epidemic model of contagious diffusion in which all members of a population are regarded as susceptible to a disease is the General Epidemic Model. In this approach a threefold division of the population is made into susceptibles, infectives and removals (infectives who after a period of time either cease to pass a disease on to others or communicate information regarding an innovation). The General Epidemic Model has been utilized to study a number of different types of diffusion with respect to disease and the dispersal of plants and animals and has also been refined to introduce greater complexity into studying transmission.
The third type of diffusion is termed hierarchical. In this case diffusion can leap over intervening people and places. Hierarchical diffusion helps explain diffusion within large bureaucratic systems such as multinational firms. It is also a useful way to explain diffusion in the fashion industry, where innovations may originate in fashion centers such as , , or Melbourne, then diffuse to chain stores in larger cities and from there to retail stores in smaller towns.
The fourth type of diffusion is relocation. In relocation diffusion, information (or plants and animals) moves along with the people who know it. Once relocated, migrants will often then spread innovations via an expansion diffusion process. Indeed, the different types of diffusion are usually occurring almost simultaneously.
Absolute or absorbing barriers are a feature or condition that completely prevent diffusion, for example, mountain ranges that prevent population dispersal. However, over time barriers may become permeable because of technological advances or changes in perception, such as the Appalachians with respect to European settlement in the United States. Finally, there are reflective barriers that deflect an innovation wave back on itself, such as in the case of human settlement on a coastline.
Substantial research continues to be done on diffusion in geography with respect to innovation particularly with respect to technology and products. In this there is a strong relationship to research on time geography as well as trying to understand the social processes behind individual adoption and resistance to adoption.