Skip to main content
ARS Home » Midwest Area » Columbia, Missouri » Cropping Systems and Water Quality Research » Research » Publications at this Location » Publication #66589


item Sudduth, Kenneth - Ken

Submitted to: Canadian Symposium on Remote Sensing Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 3/29/1996
Publication Date: N/A
Citation: N/A

Interpretive Summary: Precision farming is a management system which attempts to improve production efficiency by adjusting crop treatments, especially fertilizer and chemical applications, to varying local conditions within the field. Input application may be based on any factor or combination of factors affecting crop growth and yields, such as nutrient status, weed pressure, soil moisture status, landscape position, soil organic matter, soil acidit or topsoil depth. One focus of our research has been the collection and interpretation of the spatial data which is the foundation of precision farming. In this paper, we examine the process of collecting and analyzing these data to make management decisions. Grain yield and topsoil depth data were collected with electronic sensors, while soil fertility data were obtained from laboratory analysis. Images obtained by aerial photography were useful to help us understand patterns of yield variation. Several standard statistical methods were applied to the data but did not provide reasonable results. A nonparametric regression method was applied to the data and was able to estimate yields based on soil property data. Use of this approach may lead to improved methods of management planning for precision farming implementation by producers and agribusinesses.

Technical Abstract: The process of data acquisition and interpretation for precision farming was examined using a research field in central Missouri as a case study. Grain yield data were collected with a commercial combine-mounted sensor. Topsoil depth was estimated from on-the-go soil conductivity sensing. Soil fertility data were obtained from laboratory analysis of collected samples. .Visual interpretation of maps of these data provided useful insight into spatial relationships. In a normal growing season, variations in crop yield were primarily influenced by topsoil depth and topography. Aerial images were found to show within-season differences in crop growth which correlated well with yield variations. Statistical evaluation of the data was difficult with standard correlation and regression analysis. A nonparametric regression method, projection pursuit regression, successfully estimated spatial yield patterns based on soil property data, with r^2 values ranging from 0.57 to 0.72. Estimated yield maps based on this technique compared well with actual yield maps, reproducing all important features of the original data.