Making a Corn Canopy Sensor Algorithm Better for Nitrogen Recommendations

Something went wrong. Please try again later...

Rationale


  • Canopy sensors use light reflectance from a corn canopy as an indicator of the plant’s N health status.
  • Studies have shown canopy sensor algorithms used in partnership with light reflectance for corn N fertilizer recommendations are not consistently accurate (Bean et al., 2018).

Objectives


  • Compare red and red-edge waveband sensitivity to soil and crop measurements.
  • Evaluate across the US Midwest region the performance of the University of Missouri Corn Canopy Sensor Algorithm (ALGMU).
  • Improve ALGMU N fertilizer recommendations with site-specific soil and weather information.

Study Description

 
  • A public-industry partnership between Pioneer and eight US Midwest land-grant universities (IA, IL, IN, MN, MO, ND, NE, and WI; Figure 1).
  • A total of 49 site locations over three growing seasons (2014-2016).
Map showing research locations across eight states included in a 3-year split corn nitrogen application study (2014-2016).

Figure 1. Research locations across eight states included in the 3-year split N application study (2014-2016).

  • Used a RapidSCAN (Holland Scientific Inc., Lincoln, NE) corn canopy sensor with measurements at V9±one corn growth development stage.
  • The ALGMU, developed using the red reflectance waveband to calculate in-season N fertilizer recommendations.
  • The economically optimal N fertilizer rate (EONR) was calculated for each site using $4.00/bu corn and $0.40/lb of N prices.
  • Performance of ALGMU N fertilizer recommendations was measured by comparing to end-of-season calculated EONR.
  • Weather information used for ALGMU adjustment was from the time of planting to the time of side-dress. Soil properties used for ALGMU adjustment were from pre-plant soil samples. Soil nitrate samples were collected at V5. The relative yield (RY) was calculated by dividing the individual plot yield by the averaged site-level EONR.

Results - Objective 1


  • Sufficiency indices (SI; N deficient corn reflectance / N reference corn reflectance) calculated with either red or red-edge wavebands were significantly related to V5 soil nitrate (Figure 2; ɑ = 0.05).
  • As V5 soil nitrate increased, the SI approached “1”, a point at which V5 soil nitrate was ̴ 16 ppm. This suggests 16 ppm soil nitrate at V5 leads to equivalent canopy reflectance between V9 unfertilized corn and N reference corn.
  • The red-edge waveband (R2 = 0.51) was better related to V5 soil nitrate than the red waveband (R2 = 0.32).
Chart showing the percent of sites where corn nitrogen uptake at maturity was affected by N application timing at total N application rates of 160 and 240 lbs/acre.

Figure 2. The relationships between red and red-edge sufficiency indices (calculated from V9 collected canopy sensor reflectance measurements) and V5 soil nitrate.

  • Sufficiency indices were also related to relative yield using both the red and red-edge wavebands (Figure 3; ɑ = 0.05).
  • As RY approached “1”, meaning there was no additional yield increase with added N fertilizer, the SI also reached “1”. This was expected. Corn that received additional N fertilizer with no yield increase “looked” the same as the N reference corn. Similar to V5 soil nitrate, the red-edge waveband better related to changes in RY than the red waveband.
Chart showing the percent of sites where corn nitrogen uptake at maturity was affected by N application timing at total N application rates of 160 and 240 lbs/acre.

Figure 3. The relationships between relative yield and the red and red-edge sufficiency indices (calculated from canopy sensor reflectance measurements collected at V9).

Results - Objective 2


  • The ALGMU N recommendations did not relate well to EONR (Figure 4). For most sites needing > 100 lbs N/acre, the ALGMU underestimated crop N need.
  • The range in ALGMU N fertilizer recommendations was between 50 and 160 lbs N/acre while the range in EONR was from 0 to 240 lbs N/acre, showing a lack of sensitivity to site-specific N need.
  • A state-by-state analysis revealed the ALGMU performed better outside of MO than it did inside MO. Interestingly, sites where the ALGMU performed best were those with > 3.4% organic matter.
Chart showing the performance of unadjusted University of Missouri Corn Canopy Sensor Algorithm compared to the end-of-season calculated economically optimal N fertilizer rate (EONR).

Figure 4. The performance of the unadjusted ALGMU compared to the end-of-season calculated EONR. Data points on or near the 1:1 diagonal line were sites that the ALGMU performed reasonably well for making an in-season N fertilizer recommendation. Sites below and above the 1:1 line represent recommendations that under- and over-estimated N need, respectively. Sites that fell within the yellow shaded region are those within 30 lbs N/acre of EONR (the percent of sites in the white box in the top right-hand corner). The dashed line shows the linear relationship between the ALGMU and EONR.

Results - Objective 3


  • Adjusting the ALGMU N fertilizer recommendations with site-specific soil and weather information resulted in improved EONR estimation (Figure 5). This was helpful since early-season precipitation and soil properties greatly influence corn N response, especially over a geographically diverse area.
  • Following adjustment, 51% percent of the sites (11 additional sites over the unadjusted ALGMU) fell within 30 lbs N/acre of EONR.
  • The range of adjusted ALGMU recommendations more accurately mirrors the range in EONR values.
Chart showing the performance of adjusted University of Missouri Corn Canopy Sensor Algorithm using soil and weather data compared to the calculated economically optimal N fertilizer rate (EONR).

Figure 5. The performance of the adjusted ALGMU using soil and weather information compared to calculated EONR (see Figure 4 caption for details).

Conclusions


  • Canopy sensor measurements, especially the red-edge waveband, was related to V5 soil nitrate and yield response to added N fertilizer.
  • ALGMU recommendations may have improved if the algorithm employed the red-edge waveband instead of the red waveband.
  • The unadjusted ALGMU was poor in estimating EONR values across a large geographical region.
  • Using site-specific soil and weather information improved the ALGMU N fertilizer recommendations.
  • Use of the ALGMU algorithm regionally only would be advised when including site-specific soil and weather adjustments.

 

Authors: G. Mac Bean, University of Missouri and Dr. Newell R. Kitchen, USDA-ARS

Bean, G.M., N.R. Kitchen, J.J. Camberato, R.B. Ferguson, F.G. Fernández, D.W. Franzen, C.A.M. Laboski,  E.D. Nafziger, J.E. Sawyer, P.C. Scharf, J. Schepers, and J.F. Shanahan. 2018. Improving an Active-Optical Reflectance Sensor Algorithm Using Soil and Weather Information. Agron. J. 110:1-11. doi:10.2134/agronj2017.12.0733

Research was conducted by G. Mac Bean, University of Missouri, Dr. Newell R. Kitchen, University of Missouri, and the others involved in this regional project was a part of the Pioneer Crop Management Research Awards (CMRA) Program. This program provides funds for agronomic and precision farming studies by university and USDA cooperators throughout North America. The awards extend for up to four years and address crop management information needs of Pioneer agronomists, sales professionals, and customers.

The foregoing is provided for informational use only. Please contact your Pioneer sales professional for information and suggestions specific to your operation. 2014-2016 data are based on average of all comparisons made in over 49 locations through December 1, 2016. Multi-year and multi-location is a better predictor of future performance. Do not use these or any other data from a limited number of trials as a significant factor in product selection. Product responses are variable and subject to a variety of environmental, disease, and pest pressures. Individual results may vary.