Masters Theses

Date of Award

5-1998

Degree Type

Thesis

Degree Name

Master of Science

Major

Biosystems Engineering Technology

Major Professor

William E. Hart

Committee Members

Michael Buschermohle, John Wilkerson

Abstract

Precision farming technologies were used to measure potential yield variability due to three postemergence herbicide treatments and the soils inherent physical and chemical variability. The two-year study was conducted on two production size fields having different soil types and landscape features. One field was bottom land using conventional tillage practices. The second field was an upland using no-till cropping practices. Production fields were treated as case studies, rather than comparisons of cropping systems. Herbicide treatments included Dupont's recommended program with Staple, a widely accepted program that includes broadcasting DSMA, and a program developed by The University of Tennessee's Milan Experiment Station that involves post-directing MSMA and Cotoran. Fields were divided into blocks 24 rows in width. Eight rows of each block were treated with one of the three postemergence broadleaf herbicide programs.

Several physiological growth factors of individual cotton plants were measured throughout the season. Random plants were mapped to measure terminal height, number of nodes, and number of squares for a representative sample from each treatment. Plants were mapped to determine if physiological differences in plants were due to herbicide treatments, also these factors were correlated to yield and soil nutrient data.

Yield data from the fields in this study was used to try to explain variation due to measured environmental effects and the different management practices. Yield variation was shown by calculating standard deviation and graphically illustrated by generating yield maps.

Additional data collected throughout the study included gin samples, plant populations, soil nutrients, and soil types. Grab samples were collected at harvest to correlate gin turnout differences with herbicide treatments. Plant population counts were taken to quantify plant densities between herbicide treatments. The soil nutrients measured included phosphorus, potassium, and soil pH. Also, soil types were mapped within each field. All of these spatial factors were correlated with yields within a field.

A basket weigh system mounted inside the basket of a Case 2155 4-row picker was used to measure yield on production scale in real time. The system was originally designed, built, and field tested during the 1995 cropping season. This system was refined and improved for the first year of this study (1996) and was very successful with some limitations. The system was further refined for the second year harvest to include a system to measure both static and dynamic air pressure within the weigh basket during harvest. Air pressure measurements were collected to determine offsets to correct weigh data biased during harvest from picker fans. Although the system would not be feasible for production scale harvesting (manual unloading), it was an excellent method for researchers to collect yield data in real time on a field scale.

Plant populations differed between years only. Results from the plant height and the yield showed that the Staple treatment performed slightly better. Gin sampling results for field A-5 indicated that gin turnout in the Staple treatment was significantly higher than for MSMA/Cotoran. The DSMA treatment was between these two treatments and not significantly different from either. Correlations between soil nutrients and yield showed phosphorus and potassium both had good correlation with yield, while the correlation between pH and yield was weak. Soil type data showed that the soil type Henry in field A-5 had significantly lower yields. The combined data showed that phosphorus had the highest correlation to yield. Other variable pairs that were correlated included height and nodes; height and squares; nodes and squares; phosphorus and potassium; phosphorus and pH; and potassium and nodes.

Results from the pressure sensing units were poor because they were not sensitive enough for this application.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS