Date of Graduation

5-2016

Document Type

Thesis

Degree Name

Master of Science in Crop, Soil & Environmental Sciences (MS)

Degree Level

Graduate

Department

Crop, Soil & Environmental Sciences

Advisor

Larry Purcell

Committee Member

Trenton L. Roberts

Second Committee Member

Douglas Karcher

Third Committee Member

Jason Kelly

Abstract

Nitrogen (N) is one of the most limiting factors for maize (Zea mays L.) production worldwide. Over-fertilization of N may decrease yields and increase NO3- contamination of water. The Dark Green Color Index (DGCI) is a quantitative measure of greenness that is closely related to leaf N concentration. Previous research determined that DGCI values from maize at V6 to V10 could be used to predict the amount of N fertilizer to recover 90 or 95% of the potential yield. These DGCI algorithms were used by Spectrum Technologies Company to develop a smartphone application (app-method) for predicting maize N needs. Our objectives were to (1) evaluate the N recommendations based upon DGCI values made at V6 to V10 using the app-method and the previously developed method using a digital camera (camera-method) with standard N recommendation from the University of Arkansas Cooperative Extension Service (246 kg ha-1); (2) determine if DGCI values made by the app-method agreed with values determined by the camera-method; and (3) identify potential sources of error leading to discrepancies between the camera- and the app-methods for determining DGCI. Field and laboratory experiments were conducted to answer these objectives. When residual soil N was high and the crop was unresponsive to additional N fertilizer, the app still recommended additional N to be applied. The evaluation of the camera-method at low- and medium-residual N soils showed that DGCI predicted significantly less N than the standard recommendation without affecting yield potential. There was a relatively poor agreement between DGCI values made on the same leaves, resulting in large differences in the recommended amounts of N to apply using the camera- and the app-methods (r2=0.52, 0.67). Two major sources were identified that were responsible for the discrepancies between the camera and the app-methods. First, the DGCI values of the internal standards used by the two methods were different from their reported nominal values. Second, the camera-method uses the entire leaf blade in calculating DGCI whereas the app calculates DGCI from only a small portion of the leaf. Dissection of leaf blades indicated that N concentration increased from the basal to the apical portion of leaves and that DGCI values were different across the leaf blade. Future research should improve the relationship between the camera- and the app-methods in predicting DGCI values.

Share

COinS