A personal philosophy of precision agriculture:

Precision agriculture took off with the introduction of yield monitors, GPS and variable rate spreadeers. Take your eyes off the fancy technology! Precision agriculture is really about the changes this technologybrings by enabling farmers and their partners to manage uncertainty better. The technology helps manage spatial variation. This is a major source of uncertainty in agriculture because it has never been measured before in detail over farms. But it is only ONE of many sources of uncertainty in agriculture.

I have seen precision agriculture since the early 1990's. Together with Dr Rob Bramley, Dr Rob Corner and Dr Matt Adams in Australia, I developed a different philosophy of precision agriculture as technology to help better process control in agriculture.

Agriculture, like all industries, faces continual demands for change. These demands may be derived internally - for example the need to protect a scarce land resource - or externally, such as the need to maintain profitability in the face of increasing international competition. Change is impeded by uncertainty of new outcomes, in response to which decision makers acquire strategic knowledge; that is, knowledge about what to do (Gruber, 1989).

Rowe (1994) provides a useful classification of uncertainty and explains how uncertainty extends through many parts of the decision problem.

Metric uncertainty occurs when a farmer does not know whether adding fertilizer will increase crop yield to (say) 1.5t/ha or 3.0 t/ha.

Temporal uncertainty occurs if yield increase depends on timing, or on the season.

Structural uncertainty occurs if the farmer doesnt know whether fertilizer is the main determinant, or if other variables are more important, such as soil type, weed infestation or rainfall.

Translational uncertainty is perhaps the most difficult to describe, but can best be characterized as: 'How much do I care?'. Mintzberg (1994) points to the fallacies of defining a complex, unstable and vague system by formal representation. This is very relevant to the provider of spatial information because careful judgements have to be made about how information can be formalised to improve its accuracy without constraining its meaning.

28 Oct 08

Directory