|LI, DAN - University Of Minnesota|
|MIAO, YUXIN - University Of Minnesota|
|BEAN, GREGORY - McCain Foods, Inc|
|FERNANDEZ, FABIAN - University Of Minnesota|
|SAWYER, JOHN - Iowa State University|
|CAMBERATO, JAMES - Purdue University|
|CARTER, PAUL - Retired Non ARS Employee|
|FERGUSON, RICHARD - University Of Nebraska|
|FRANZEN, DAVID - North Dakota State University|
|LABOSKI, CARRIE - University Of Wisconsin|
|NAFZIGER, EMERSON - University Of Illinois|
|SHANAHAN, JOHN - Agoro Carbon Alliance|
Submitted to: Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/13/2022
Publication Date: 1/15/2022
Citation: Li, D., Miao, Y., Ransom, C.J., Bean, G.M., Kitchen, N.R., Fernandez, F.G., Sawyer, J.E., Camberato, J.J., Carter, P., Ferguson, R.B., Franzen, D.W., Laboski, C.A., Nafziger, E.D., Shanahan, J. 2022. Corn nitrogen nutrition index prediction improved by integrating genetic, environmental, and management factors with active canopy sensing using machine learning. Remote Sensing. 14(2). Article 394. https://doi.org/10.3390/rs14020394.
Interpretive Summary: Current methods for determining when a corn crop requires more nitrogen (N) fertilizer to optimize grain yield are difficult as they rely on soil and plant tissue sampling. These methods are labor intensive, costly, and often not done at a high enough spatial resolution to determine the corn N status across large production fields. These issues can be resolved by measuring the corn color and biomass with optical sensors. However, current methods of using these sensors rarely consider genetic, environmental, or management factors that can affect the sensor’s accuracy. Therefore, this study was conducted to evaluate the best way to use sensor measurements to predict different corn growth characteristics associated with plant N status. Analysis of data obtained from 13 sites in 4 U.S. Midwest states found that when genetic, environmental and management information were integrated with optical sensing measurements four plant characteristics could be reliably predicted (R^2 >0.7). These characteristics included aboveground biomass, tissue N concentration, total N taken into the plant, and an index of corn N nutrition. Integrating multiple factors was accomplished using sophisticated computer machine learning techniques to generate new algorithms. Generally, the ability to predict these characteristics improved when more information was included in the algorithms, supporting the point that dozens of factors are known to impact inorganic N in agricultural soils. The result of these findings is that farmers using sensing technologies along with these new algorithms can adjust their N rates on-the-go to better match crop N need, which could improve their profitability and minimize environmental impacts of excessive N fertilization.
Technical Abstract: Accurate nitrogen (N) diagnosis early in the growing season across diverse soil, weather and management conditions is challenging. Strategies using multi-source data are hypothesized to perform significantly better than approaches using crop sensing information alone. The objective of this study was to evaluate across diverse environments the potential for integrating genetic (e.g., comparative relative maturity and growing degree days units to key developmental growth stages), environmental (e.g., soil and weather), and management (e.g., seeding rate, irrigation, previous crop, and preplant N) information with active canopy sensor data for improved corn N nutrition index (NNI) prediction using machine learning methods. Thirteen site-year corn (Zea mays L.) N rate experiments involving 8 N treatments conducted in four U.S. Midwest states in 2015 and 2016 were used for this study. A proximal RapidSCAN CS-45 active canopy sensor was used to collect corn canopy reflectance data around the V9 developmental growth stage. The performance of vegetation indices and ancillary data to predict corn aboveground biomass, plant N concentration, plant N uptake and NNI were evaluated using singular variable regression and machine learning methods. The results indicated that when the genetic, environmental and management data were used together with the active canopy sensor data, corn N status indicators could be more reliably predicted either using support vector regression (R2 = 0.74-0.90 for prediction) or random forest regression models (R2 = 0.84-0.93 for prediction), as compared with using best performing single vegetation index with regression based models (R2 = 0.22-0.29).. The N diagnostic accuracy based on NNI was 87% using the data fusion approach with random forest regression (Kappa statistic = 0.75), which was better than the result of a support vector regression model using the same inputs. The NDRE index was consistently ranked as the most important variable for predicting all the four corn N status indicators, followed by preplant N rate. It is concluded that incorporating genetic, environmental, and management information with canopy sensing data can significantly improve in-season corn N status prediction and diagnosis across diverse soil and weather conditions.