Spatializing Agronomic Modeling

Agricultural production is wholly related to a crop's interaction with the environment in which it is situated. Such an obvious statement seems redundant until it is understood that the majority of current agronomic models do not contain a spatial component. Attempting to understand and quantify the effect that variations in climate, substrate, water availability, and management have on a harvest has occupied scientific research for many years. The successful result of any agronomic model is the ability to predict how a specific crop will behave under varying conditions and then extrapolate this information throughout a broader landscape. The spatialization of agronomic models can be achieved by combining physiological information related to the crop, environmental conditions, and management practices within a geographic information system.

Collecting enough information at a suitable scale for the implementation of a spatialized agronomic model is challenging. Decisions need to be made defining the scales at which results from the model can be extrapolated. Indicating a scale too large can cause excessive detail in the final analysis, rendering the model ineffective, while including information of a smaller scale can cause problems associated with data generalization.

GIS provides not only a means to incorporate data layers from many scientific disciplines and display them but a method to include within traditional agronomic studies—a spatial component at the heart of any agronomic model. The Geostatistical Analyst extension for ArcGIS provides the tools by which a researcher can quantify the degree to which a specific parameter affects the overall production of a crop. The ModelBuilder extension allows scientists to interactively create new agronomic models using a simple interface that can include both vectorized and rasterized data sets. Spatial regression analysis is key to extrapolating results from existing or future agronomic models and will continue to grow in importance as the relationship between crop productivity and location is reexamined.