Characterizing nitrogen deficiency of maize at early growth stages using fluorescence measurements
Date
2016
Authors
Siqueira, Rafael Telles Tenorio de, author
Khosla, Raj, advisor
Longchamps, Louis, committee member
Reich, Robin, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
Among all nutrients that are important for crop production, nitrogen (N) is one of the least efficiently utilized, mainly due to its high mobility in soil. The possibility of using crop sensing in real-time to detect variability in N deficiency within a field has the potential to enhance N efficiency, increase crop yield, and reduce potential environmental risks and crop production costs. Potassium (K), another important crop nutrient, can also lead to higher yield when applied in the right amount and manner. Real-time fluoro-sensing is a new technology for crop sensing and studies have shown that it could enable variable rate nutrient management for precision agriculture practices. The objective of this study was (1) to evaluate if fluorescence sensing can detect variability of N and K in crop canopy at early growth stages of maize (prior to V6 crop growth stage) under controlled condition (greenhouse), (2) to evaluate the effect of different fertilization dosages of N over the plant growth, and (3) to verify if induced fluorescence can detect in situ N variability at early growth stages of maize. Research was conducted in two stages, first in a greenhouse condition and later in field spread over three site-years. The greenhouse research was conduct in year 2011 and plants were grown in plant-pots with silica sand and supplied with modified Hoagland solution with different rates of N and K. Field trials were conducted in year 2012 and 2013 in northern Colorado. For the greenhouse study, data collected via fluorescence sensor (Multiplex®3) were analyzed using ANOVA and Tukey’s HSD to test significant differences among treatments in each experiment. For the N experiment, regression analysis between the seven fluorescence indices and N uptake was performed for the 12 days of data acquisition at five different growth stages (i.e. 2-leaf to 6-leaf growth stages) and coefficient of determination was used to identify the best fluorescence indices to detect N status. Also, root mean square error (RMSE) was used to test the precision of the estimates for each index. Results of this study indicated that all fluorescence indices were able to detect N variability in maize canopy prior to V2 growth stage. However, the fluorescence indices failed to identify K deficiency as the maize plants with K treatments showed small variability at early crop growth stages. For the field study, two site-years had 5 N rate treatments applied as UAN 32% (urea ammonium nitrate; 32-0-0), while one site-year had 6 N treatments applied pre-planting. Sensors used in this study were the Multiplex®3 for fluorescence sensing and the GreenSeeker® for reflectance sensing (NDVI). Sensor measurements were correlated with aboveground biomass, N content, and N uptake measured at two growth stages (V6 and V9 maize growth stage). The aboveground biomass, N content, N uptake, yield, and sensors readings were analyzed using ANOVA and Tukey’s HSD to test significant differences among the N treatments. Also, a regression tree between N uptake and the fluorescence indices was fitted along with the coefficient of determination (R2). The N rates had no effect on aboveground biomass, N content and N uptake (for both sampled growth stages). Under field conditions, fluorescence indices failed to detect N variability in maize at early growth stages for all three site-years. This finding may require further investigation, as for most of the N treatment plots, maize plants had sufficient N levels and another biotic or abiotic stress may be responsible for unexplained differences in N variability as measured by fluorescence sensor. Contrasting findings under greenhouse conditions versus field conditions limit the application of fluorosensing sensor. Further field studies are needed to evaluate the potential of this sensor for detecting N variability in situ.