Date of Award


Document Type


Degree Name

Master of Science (MS)

Legacy Department

Biosystems Engineering

Committee Chair/Advisor

Khalilian, Ahmad

Committee Member

Kirk , Kendall R.


The main objective of this four year study was to develop, refine, and employ sensor-based algorithims to determine the mid season nitrogen requirements for production of irrigated and dryland cotton (Gossypium hirsutum L.) in Coastal Plains soils. The secondary objective of the project was to develop and test equipment for variable rate application of nitrogen to commerical cotton fields utilizing the developed algorithim. Two different production fields at Clemson's Edisto Research and Education Center near Blackville, SC were used. One field, equiped with an overhead irrigation system, was used during the 2007 and 2010 production seasons to develop the algorithm for irrigated cotton. The second field was used during the 2008 and 2009 seasons for developing the algorithm for dryland cotton nitrogen management. Each field was divided into three separate zones based on soil electrical-conductivity (EC) data. The algorithim was developed using 'Nitrogen Ramp Calibration Strips' (N-RCS) and varied prescription rate nitrogen plots. Three N-RCS were established in each production field, one per EC zone. The N-RCS was composed of 16 nitrogen rates (0 to 168.13 kgN*ha-1) on 5.0 meter intervals. For the varied prescription rate plots, five different rates of nitrogen fertilizer (0, 33, 67, 100, and 134 0kgN*ha-1) were replicated four times in plots of each zones using a Randomized Complete Block desgin arrangement.
Optical sensor readings were collected from the test plots to determine cotton plant Normalized Difference Vegetation Index (NDVI) at different growth stages. The sensor readings were used to develop two different algorithims to be used in the estimation of mid-season nitrogen need of the cotton plants. Sensor readings collected between 40 and 60 days after planting were highly correlated (average R2> 0.80) with the final yield and nitrogen requirement. The Response Index (RI), the extent to which the crop will respond to additional N, was calculated by dividing the highest NDVI reading from N-RCS and N-rich strips (established in each zone) by NDVI measurements of the adjacent area in each zone. In Season Estimated Yield (INSEY) was used along with the actual field yield to produce a yield potenial (YP0) for each growing season one for irrigated cotton and one for dry land cotton. The algorithm is N rate= (YP0*RI-YP0)*%N/NUE. Where the %N is the percentage of nitrogen in cotton seeds after harvest and NUE is the nitrogen use efficiency, typically 50%.
The algorithim developed from the 2008 growing season was used during the 2009 growing season to estimate the amount of mid-season side-dress nitrogen required for specific research plots in the production field. The algorithm reccommended a reduced rate of nitrogen (40% less) across the entire field compared to the normal grower practice (101 kgN*ha-1) with no reduction in cotton yield. Similar results were obtained when using the Oklahoma State University Algorithm.
Three different methods of nitrogen application were tested, one during each of the growing seasons of 2007-2009. During the 2007 production year a typical pull behind nitrogen side-dress applicator with a ground driven piston pump was used. This applicator was the most crude and innacurate method of fertilizer application used during the study. During the 2008 production year a custom built applicator was used. The applicator operated using a hydraulic pump in combination with an in-cab control system. The rates were adjusted using various orifices and solenoids. The final applicator, tested in 2009, was a typical three point hitch pull behind side-dress coulter rig controlled using a hydraullic Rawson controller for the piston pump. The three point hitch applicator has the potential to be the most accurate and versatile of any used during the project.
Various equipment was tested throughout the study to determine the best and most accurate way to apply the mid-season N algorithm fertilizer recommendation. The parameters of specific equipment such as the GreenSeekers¨ for measuring NDVI were tested to determine the true accuracy based on height above crop canopy and time of day, which is related to the sun angle and solar radiation. The results of this test proved that the sensor is height sensitive with an optimal height range of .8128 to 0.9144 meters. It was determined from the test that the sensors are not sun angle sensitive and return a non statistical difference in readings throughout the day between the hours of 10 a.m. and 8 p.m. (EST). The sensors returned a lower number once the sun had set but the main reason for the lower number is due to the physiological response of the plant. It was found due to the response of the plant that it is not possible to obtain an accurate sensor reading at night. Sensor readings taken from two different travel directions were found to not be statistically different, thus the sensors were found not to be travel direction specific. The data remained constant independent of the orientation of the field. This study confirmed that there is a significant possibility to accurately predict in-season expected yield (INSEY) in cotton using mid-season NDVI sensor readings in conjunction with an accurate prediction of a reduced nitrogen requirement without a significant reduction in yield.
Two different ultra-sonic height sensors were tested during the growing season of 2010 to determine the feasibility of determining plant height on-the-go. Both sensors gave promising results to accurately predict plant height with more testing and reprograming.



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.