1. - - - - - - - - - - - - - - NDVI - - - - - - - - - - - - - -
Active Sensors in Sugarbeet Production for In-Season and Whole Rotation Nitrogen Management
T.J. Boring and R.J. Gehl
Dept. of Crop and Soil Sciences, Michigan State Univ., East Lansing, MI 48824
Introduction
Sugarbeet processing efficiency depends on both root quantity and sugar quality – factors that are largely influenced by N fertilization. While inadequate N supply limits total yield, excess N uptake affects the processing quality through decreased sucrose levels and increased impurities.
Recently, measurement of the reflectance characteristics of crop foliage has been adopted for N status assessment in various crops. Research efforts have focused on the use of active sensors to estimate N use efficiency, N requirement, and yield potential for crops including corn and wheat. Similar in-season assessments for sugarbeets would be a valuable tool to assist in N management, as well as harvest scheduling and prioritization.
Previous research has indicated that the use of active sensors during the sugarbeet growing season has potential as a means to predict root yield and quality and as an indicator of N return to the cropping system (from foliage). Accounting for residual plant N and applying that information for subsequent crops can improve crop yield and quality, and increasing grower profitability.
Methods
Sugarbeet canopy NDVI was measured in mid-June, mid-July, mid-August and at harvest using a Greenseeker (Ntech Industries, Inc., Ukiah, CA) optical sensor. Above-ground biomass samples were obtained immediately following Greenseeker scanning on day of harvest by removing all foliage from 4.5 m of each of the center two rows of each plot. Total biomass was measured as fresh weight and dry matter following drying, and subsamples analyzed for total N content.
Root yield was determined by harvesting 2 row lengths of about 21 m each. Subsamples of roots were analyzed for percent sucrose and clear juice purity (CJP). Recoverable white sucrose as Mg ha-1 (RWSA), an economic parameter used for grower payments, was calculated from the following formula:
Results
In-Season Yield Prediction
NDVI measurements collected during the growing season were strongly related to root yield in both July (R2=0.88, Fig. 1) and August (R2=0.88, data not shown).
Results
End-of-Season Total N
NDVI readings at harvest were strongly related to total N remaining in the above-ground biomass (R2=0.90, Fig. 3), which would be subsequently returned back to the soil in production field systems.
Figure 3. Day of harvest NDVI versus sugarbeet top total N determined in foliage biomass on a dry weight basis for 3 sites in 2006 and 2 sites in 2007.
Figure 1. July NDVI versus sugarbeet yield for 3 sites in 2006 and 4 sites in 2007.
Results
In-Season NDVI
Differences in relative greenness of the sugarbeet canopy between N rate treatments became more pronounced as the season progressed (Table 1). Only control N treatments were consistently differentiated from higher N rates at June and July sensing events in both 2006 and NDVI tended to increase throughout the growing season as the sugarbeet canopy developed, along with the ability to differentiate N rate treatments.
Summary
In-season NDVI was as an effective indicator of both sugarbeet yield and RWSA. NDVI readings collected during the season as a means for yield prediction may prove to be an effective tool for growers to determine harvest order of multiple fields and for processors to improve harvest scheduling and storage efficiency.
End-of-season NDVI was strongly related to sugarbeet top total N. Our results indicate that delineation of N zones for subsequent crop N utilization using NDVI may be possible for Michigan sugarbeet cropping systems. This approach is similar to previous work using, among other tools, satellite imagery and aerial photography to construct N management zones. Determination of total N from crop residues, and subsequent compensation of that N for following crops, has potential to improve N management for the sugarbeet cropping system as a whole.
In-Season RWSA Prediction
Growing season NDVI readings were strongly related to RWSA (Fig. 2). For measurements recorded in July, a relationship of R2=0.87 was found for data from 3 sites in 2006, and R2=0.81 for 2 sites in 2007 (data from 2 additional sites in 2007 not yet available). When combining the 5 sites over both years, the relationship between NDVI and RWSA remained strong (R2=0.89). Similar results were found for the August NDVI readings (R2=0.84, 0.67, and 0.86 for 2006, 2007, and combined, respectively).
Table 1. Mean NDVI measurements for 3 sites in 2006 and 4 sites in 2007, in relation to N fertilizer rate and sampling time. Means labeled with same letter on a given date are not different as determined by LSD at a=0.05.
2006
N rate
June
July
August
Harvest
kg N ha-1
NDVI
0.2049b
0.5545b
0.5099b
0.6787c 45
0.2567ab
0.7221a
0.6092a
0.6921c 90
0.2741a
0.7432a
0.6781a
0.7270bc
135
0.2628a
0.7242a
0.6828a
0.7348bc
180
0.2584a
0.7581a
0.6836a
0.7705ab
225
0.2572ab
0.7255a
0.6900a
0.7935a
2007
0.4219b
0.6572b
0.6985c
0.6775e
0.4939a
0.6855b
0.7134c
0.6898de
0.4993a
0.7123a
0.7423b
0.7172cd
0.4990a
0.7188a
0.7518ab
0.7406bc
0.4896a
0.7226a
0.7691a
0.7601ab
0.5000a
0.7652a
0.7784a
Objectives
Evaluate optical sensor technology for in-season prediction of sugarbeet harvest yield and quality.
Evaluate optical sensor technology for end-of-season measurement of sugarbeet above-ground biomass total N content.
Figure 2. July NDVI versus sugarbeet recoverable white sucrose yield for 3 sites in 2006 and 2 sites in 2007.
Methods
Field experiments were established in Michigan at 3 locations in 2006 and 4 locations in Treatments included 6 N rates ranging from 0 to 224 kg N ha-1 in 45 kg N ha-1 increments. Granular urea N fertilizer at 45 kg N ha-1 was banded 5 cm x 5 cm at planting and sidedress N as 28% UAN was injected between rows in early June. Plots measured 4.6 by 12.2 m and were arranged in a RCBD with 4 replications.