Development of On-the-Go Nitrogen Application Algorithms Based on NDVI

Precision agriculture technologies can be categorized as either information-intensive or embodied-knowledge.  Information-intensive technologies such as yield monitors and variable rate applications based on grid soil sampling take additional management ability and effort to make use of the information.  Embodied-knowledge technologies such as automated guidance require no additional management abilities to utilize the technology because it is purchased in the form of an input.  Although commercialized later than information-intensive technologies, embodied-knowledge technologies have been more readily adopted due to the ease of use and faster payback.  A hybridization of these two categories that are inherently ‘embodied’ is on-the-go applications based on real-time sensing. The development of on-the-go nitrogen application algorithms based on active NDVI measurements are discussed for cotton production. The statistical issues involving the combination of seemingly incompatible data is presented as well as the regression and response surface methodologies employed to estimate algorithms for utilization at the farm-level.  Once data has been combined into a rectangular dataset, a range of candidate functional forms were tested for goodness-of-fit for the given data based on agro-economic theory.