Skip to main content
ARS Home » Plains Area » Lincoln, Nebraska » Agroecosystem Management Research » Research » Publications at this Location » Publication #251234
Title: An Active Sensor Algorithm for Corn Nitrogen Recommendations Based on a Chlorophyll Meter Algorithm

Author
item SOLARI, FERNANDO - Monsanto Biotechnology
item Shanahan, John
item FERGUSON, RICHARD - University Of Nebraska
item ADAMCHUK, V - University Of Nebraska

Submitted to: Agronomy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/1/2010
Publication Date: 7/1/2010
Citation: Solari, F., Shanahan, J.F., Ferguson, R.B., Adamchuk, V.I. 2010. An Active Sensor Algorithm for Corn Nitrogen Recommendations Based on a Chlorophyll Meter Algorithm. Agronomy Journal. 102:1090-1098.

Interpretive Summary: The development of alternative nitrogen (N) fertilizer management strategies is crucial for sustaining future corn production in the U.S. and around the world, because current N management approaches have led to low N use efficiency (NUE), economic losses, and environmental contamination issues. In previous work we proposed use of active sensor reflectance measurements of the corn canopy to guide spatially variable in-season N applications as a means to improve corn NUE and decrease environmental contamination. The active canopy sensor used in this work was the Crop Circle model ACS-210 (Holland Scientific, Inc., Lincoln, NE, USA), which measures canopy reflectance in the amber and near infrared (NIR) light wave bands. We found that sensor readings (canopy reflectance in amber and NIR bands) acquired during vegetative growth (11 to 15 leaf growth stage) and converted to chlorophyll index (CI) values were highly correlated with crop N status as determined with a Minolta SPAD chlorophyll meter ( a previously proven method for measuring crop N status). Because active sensor readings obtained during this growth stage are much less laborious than SPAD chlorophyll meter readings for assessing crop N status, it was concluded that active sensors have much greater potential for directing spatially variable in-season N applications for improving crop NUE. However, it was noted that it would first be necessary to develop an algorithm for converting active sensor measurements into appropriate N application (Napp) rates. Hence, the objectives of our work were to: 1) develop an active canopy sensor algorithm based on a chlorophyll meter algorithm and 2) validate the active canopy sensor algorithm using data collected from the companion study. In this paper, an algorithm for converting active canopy sensor measurements into site-specific N applications is proposed. Development of the algorithm incorporates results from our previous work showing a linear relationship between normalized active canopy sensors (SIsensor) and normalized SPAD chlorophyll meter SISPAD readings and was based on a SPAD algorithm previously published that outlines procedures for translating SPAD readings into corn Napp rates. The resulting equation developed for this work was: Napp = (317 * square root of(0.97 – SIsensor)) which represents the function for translating SIsensor to Napp. To validate the sensor algorithm, SIsensor values were collected at two growth stages in a field study involving small plots receiving varying N fertilizer amounts and were converted into Napp using the algorithm. Then Napp was converted into crop N balance (Nbalance) estimates, where Nbalance= applied N - Napp. Negative Nbalance values indicate sensor-based estimates of crop N deficiency while positive values indicate excess N. The Nbalance values were compared with relative yields and a quadratic-plateau model fit to the data set for both growth stages (V11 and V15), producing and R2 of 0.66. Relative yields reached a plateau for Nbalance = 11 kg N ha-1, implying the algorithm provided reasonable estimates of Napp for maximizing yields. Results presented in this paper indicate that using in-season canopy reflectance measurements via an active canopy sensor can be used to assess the amount of N needed to maximize corn yield. This method requires areas where N is not limiting (N reference) to determine SIsensor values. Once the N reference area has been established, the model developed indicates that sensor data collected anytime during the vegetative growth period from V11 through V15 can be used to determine SIsensor values, which can in turn be put into the generalized algorithm shown in Fig. 3 and solved for N rate. This N rate is the amount of N fertilizer required to maximize yield. Additionally, for this method to be successful, a nominal amount (45-90 kg ha-1) of

Technical Abstract: In previous work we found active canopy sensor reflectance assessments of corn (Zea mays L.) N status acquired at two growth stages (V11 and V15) have the greatest potential for directing in-season N applications, but emphasized an algorithm was needed to translate sensor readings into appropriate N application (Napp) rates. Hence, the objectives of this work were to: 1) develop an active canopy sensor algorithm based on a SPAD chlorophyll meter algorithm and 2) validate the active canopy sensor algorithm using data collected from the companion study. We derived the active canopy sensor algorithm using a linear relationship between sensor (sufficiency index, SIsensor) and SISPAD readings in previous research and a published SPAD algorithm employing a quadratic equation to calculate Napp as a function of SISPAD. The resulting equation: Napp = (317 * square root of(0.97 – SIsensor))represents the function for translating SIsensor to Napp. To validate the algorithm, SIsensor values collected at two growth stages in the preceding study from small plots receiving varying N amounts were converted into Napp using the algorithm. Then Napp was converted into crop N balance (Nbalance) estimates, where Nbalance= applied N - Napp. Negative Nbalance values indicate sensor-based estimates of N deficiency while positive values indicate excess N. The Nbalance values were compared with relative yields and a quadratic-plateau model fit to the data set for both growth stages (V11 and V15), producing and R2 of 0.66. Relative yields reached a plateau for Nbalance = 11 kg N ha-1, implying the algorithm provided reasonable estimates of Napp for maximizing yields.