Introduction

Research has shown that yields can be improved by top-dressing wheat with N fertilizer, if the crop vigor (health status) can be estimated by some index and used to predict yields.

The NDVI (Normalized Difference Vegetation Index) is an index generated with remote sensors to assess crop vigor. It can be measured with the GreenSeeker sensor, for example, by collecting light reflected by the crop, and converting it to NDVI values. Higher values are indicative of a more vigorous crop, usually in response to N.

The objective of this research was to determine mid-season N need and rate to apply as top-dress using the NDVI data obtained with a handheld sensor and estimate the yield gained from top-dressing.

Research conducted in Carrington in 2019

Two wheat trials were established at Carrington under irrigation and dryland. Each trial received the same N treatments, which included five starter treatments (0, 40, 80, 120, 160 lbs N/ac) and two top-dress N rates based on the individual plot NDVI (0, 40 lbs N at planting, and additional N based on NDVI). The top-dress rates were determined from using several factors including NDVI, growing degree days, etc., to predict wheat yields.

Results

• Yields improved at higher N rates and began to decline above 120 lbs N. Protein data is not yet available.
• There was a strong relationship between NDVI and grain yields (Figure 1) at the irrigated site (R²= 81%), but not so good under dryland (R² = 42%). This means that NDVI was a good predictor of yields under irrigation since it explained about 81% of the differences in yields but could explain only 42% of the yield differences observed at the dryland site.

• Similarly, the relationship between N rates and NDVI was weak under dryland (23%), and fairly good under irrigation (52%).
• Yields, however, improved from top-dressing with N rates (Table 1.). The top-dress N rates, estimated using NDVI data, are shown on table 1.
• Under irrigation, for plots without starter N, yields improved by 6.7 bushels from top-dressing with 40 lbs N. With 40 lbs applied as a starter, top-dress estimate was 22 lbs N, which improved yields by 2.6 bushels.
• Under dryland, for plots without starter N, yields improved by 6.3 bushels from top-dressing with 16 lbs N. When 40 lbs was applied as a starter, and top-dressed with 20 lbs N, yields improved by 3 bushels.

Observations

The remote sensor was useful in assessing crop vigor, which was used to predict yield, and determine variable mid-season N rates to apply and enhance yields.

Jasper Teboh, Ph. D.
Jasper.Teboh@ndsu.edu
Research Soil Scientist