|Stevens, William - UNIVERSITY OF MISSOURI|
|Wrather, Allen - UNIVERSITY OF MISSOURI|
|Rhine, Matthew - UNIVERSITY OF MISSOURI|
|Dunn, David - UNIVERSITY OF MISSOURI|
Submitted to: Agronomy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: September 12, 2007
Publication Date: March 1, 2008
Citation: Stevens, W., Wrather, A., Rhine, M., Dunn, D., Vories, E.D. 2008. Predicting rice yield response to midseason nitrogen with plant area measurments. Agronomy Journal. 100(2):387-392. Interpretive Summary: A simple method is needed to aid farmers with deciding when midseason nitrogen (N) is needed by the rice crop and this study was conducted to develop thresholds using visual and digital-photography measurements. Plant observations were made one or two days before the midseason N applications by floating a yardstick in the floodwater between two rows, with inch numbers counted that were totally unobstructed by overhanging rice leaves and digital images recorded from above the plant canopy. Computer software was used to analyze the images, and thresholds were determined for both the yardstick and digital photography methods. The results suggest that N recommendations can be developed based on easily obtained measurements. Guidelines for when midseason nitrogen is not required in rice that are easily followed will result in lower costs for the farmer and less excess nitrogen in the environment.
Technical Abstract: A simple method is needed to aid farmers with midseason N decisions in dry-seeded, delayed flood rice (Oryza sativa L.). This study was conducted to develop thresholds using visual and digital image measurements for predicting rice yield response to N topdressing. 'Francis' and 'Cheniere' (cv) rice were drill seeded from 2004 to 2006 on silt loam and clay soils at Glennonville and Portageville, Missouri. Pre-flood urea was applied 0, 39, 78, 118 and 157 kg N ha-1. Treatments were included with and without mid-season N applied 34 kg N ha-1 urea at panicle differentiation (R1) and the same rate one week later. Plant observations were made 1-2 d before midseason N applications. In each plot, a yardstick was placed in floodwater and floated parallel between two rows. Inch numbers were counted that were totally unobstructed by overhanging rice leaves. Plant height was also measured and digital images from a camera were recorded. Software was used to determine percentage of green pixels in images. Highest rice yields on both soils were most often achieved with 78 kg N ha-1 with midseason N or 118 kg N ha-1 without midseason N. Preflood N had highly significant effects on yardstick counts, plant height, and green pixels. Plant height was the least reliable indicator of rice N status. Using regression analysis, no rice yield increase from midseason N was predicted when fewer than 13 yardstick numbers were showing or more than 64 % of the pixels in an image were green.