ABSTRACT

The use of yield monitors, global positioning systems, remote sensing, and other attributes of site-specific crop management is increasing in California. California farmers who have adopted yield monitoring and mapping technology have frequently observed a high level of yield variability in their fields. In some cases growers have been able to interpret these yield maps based on their knowledge of the field and use this interpretation to improve their management and enhance profitability. However, the level of knowledge of this technology has not yet reached the state where growers can confidently adopt on a wide scale true site-specific management practices, that is, practices in which management is adjusted “on the go” to match the specific needs of each location in the field.

One of the most promising site-specific management practices is variable rate input application. In particular, variable rate application of fertilizers, especially fertilizer nitrogen, has been extensively studied in Midwestern cropping systems. Scientific investigations of the profitability of variable rate nitrogen application in the Midwest have produced equivocal results, with some investigations indicating a profit and others not. Much of the work in the upper Midwest has been motivated by regulatory concerns associated with potential contamination of ground and surface waters. This is also becoming an issue in California. Variable rate nitrogen application offers the potential for increasing profitability and reducing environmental effects of crop production if the increased costs associated with the practice can be offset by reduced input costs and/or reduced regulatory pressure.

In order to achieve a workable variable-rate fertilizer management program it is necessary to be able to estimate with sufficient accuracy the crop’s site-specific nitrogen demand prior to the time of fertilizer application. A number of researchers have found that under growing conditions of the semi-arid and arid West, Cotton removes approximately 50 to 60 lbs. N per bale of lint. It requires an additional 100 to 150 lbs./acre of N to support vegetative growth. Most of this latter N is returned to the soil when cotton stubble is disked in. Travis et al. studied over a five-year period the relationship between soil test nitrate levels and crop response to applied soil N. They found that at the field scale there is a general relation in which low soil test nitrate levels correspond to a higher yield response to applied N and high soil test nitrate levels correspond to a lower yield response to applied N. They also found that there was considerable variability in this relationship, which they attributed in part to within-field variability in soil and nitrogen conditions. We hypothesize that the precision of the relationship between soil test nitrate level and plant N response can be improved by a stratified sampling scheme taking into account within-field variation in soil conditions. Based on this working hypothesis, we developed a practical site-specific variable N rate application program for California cotton production.