Using Drones to Measure Cover Crop Biomass as a Predictor of Soil Nitrogen And Corn Emergence Issues

Progress report for GNE20-241

Project Type: Graduate Student
Funds awarded in 2020: $14,832.00
Projected End Date: 07/31/2021
Grant Recipient: University of Delaware
Region: Northeast
State: Delaware
Graduate Student:
Faculty Advisor:
Dr. Jarrod Miller
University of Delaware
Expand All

Project Information


Cover crops are a common soil health management practice adopted by Delaware farmers who seek to capitalize on ecosystem services like N fixation, nutrient scavenging, and soil cover. Cover crop termination timing can play a huge role in the level of ecosystem services provided to the following cash crop. Later terminated cover crops are able to establish more above and below ground biomass and, in return, can produce additional root systems that encourage microbial diversity or, in the case of legumes, fix more N. However, later terminated cover crops can inhibit seedling germination by reducing seed to soil contact at planting. We will investigate different seeding rates of cover crop mixes to identify the seeding rate that provides the most opportunity for ecosystem services from cover crops yet does not inhibit germination of the subsequent cash crop. We will use drone imagery to observe and compare cover crop biomass with results from soil N tests. Drone technology could prove useful in predicting potential N from its biomass readings and stand counts, which is useful to improve decision making for timing cover crop termination. We will share our findings with farmers, crop consultants, and other researchers through interactive presentations, like ArcGIS Story Map, at Cooperative Extension events such as field days, Delaware Ag Week, or the Mid-Atlantic Crop Management School.

Project Objectives:

The goal of this project is to use consumer drone technologies to quickly perform stand counts, assess cover crop biomass N potential, and identify fields where cover crop residue density may cause cash crop planting issues.

Objective 1: Use drone technology to capture imagery of the cover crop mixes to approximate biomass and ground cover to help determine potential cover crop N.

Research plots will be planted with rye, clover, and a rye clover mix at five different rates. We will scout the plots with a drone at various times throughout the season (i.e., up to two early season flights to observe fall cover crop stands, one flight during the winter, and bi-weekly flights in the spring until the cover crops are terminated).  Specialized cameras attached to the drone, will capture biomass imagery that will be analyzed on software that is easily accessible to farmers and crop consultants. We will determine the relationship cover crop biomass, soil N, and drone imagery. From this imagery a rate curve of biomass and N content will be developed.

Objective 2: Verify the mathematical relationship between cover crops biomass and biomass available N from the research plots using data collected from farmer-planted cover crops.

The drone derived biomass to N relationship determined in the research field trial (Objective 1) will be evaluated in farmer-planted cover crop stands. We will fly each field using the consumer drone and develop a map of the cover crop biomass. Samples collected from each field will be used to test the accuracy of the drone derived rate curve.

Objective 3: Improve farmer productivity using rapid drone assessments to determine potential planting issues related to cover crop biomass.

Thick cover crop residue can lead to cash crop planting issues and act as a mechanical barrier to seedling emergence (Loydi et al., 2012). We will compare stand counts taken in our research plots following corn planting to in-field biomass measurements and drone imagery to determine if cover crop biomass effects on final stands can be quickly measured through aerial imagery. Establishment of biomass residue thresholds will allow for better decision making on timing cover crop termination or adjusting planter settings, thereby reducing potential reseeding costs and increasing cash crop yields.


The purpose of this project is to determine if drones can provide rapid information on field-wide cover crop biomass to identify potential planting issues due to cover crop residues and to guide N management by predicting N availability from cover crop biomass. The N content of cover crop biomass is directly related to various ecosystem services, yet it is hardly ever measured by farmers or researchers due to sampling and analysis expenses (White et al., 2019). Expanding our understanding of N content in cover crop biomass would provide for improved predictions of N needs for the following cash crop, allowing for farmers to use less N fertilizer, which is essential in minimizing environmental impacts (Bundy et al., 1999).

Accumulation of cover crop residue on the soil surface can provide many beneficial services, such as retaining moisture and regulating temperatures (among other services); however, cover crop residue can act as a mechanical barrier to seedling emergence, thereby interfering with development and growth of the cash crop (Kaspar and Erbach, 1998). Residue may also reduce light quality and quantity, soil temperature, and moisture regime (Loydi et al., 2012), all of which are all fundamental in seed establishment. If a cash crop is planted into a dense residue and seedlings are unable to thrive, poor stand and potential need to replant decreases the farmer productivity, increases production costs, and can result in negative environmental impact. Understanding the threshold of cover crop residue density that will not interfere with seed development can help farmers make important early season management decisions that are time and money conducive.

The goal of this project is to integrate consumer drone technology into cover crop scouting to improve productivity while maximizing cover crop effects on soil health. Drones can provide quick observations of vegetation conditions and stand densities to provide insight on potential N availability and identify conditions that might deter seedling establishment.


Materials and methods:

The experiment is currently being conducted at the University of Delaware Carvel Research and Education Center in Georgetown, DE. The experiment includes three (3) common cover crops (rye, clover, rye/clover mix) and a control (no cover). Each cover crop has been seeded into 10 ft × 40 ft plots at four (4) rates (full DE NRCS recommended rate, 0.75 full rate, 0.5 full rate, and 0.25 full rate) and replicated three (3) times, with the control also being replicated three times to align with each cover crop. We have planted a total of 39 plots in a randomized complete block design which was planted October 6th, 2020.

Objective 1: Use drone technology to capture imagery of the cover crop mixes to approximate biomass and ground cover to help determine potential cover crop N.

Using the DJI Mavic Air drone, equipped with an RGB camera, and the Matrice 200 drone, equipped with a multispectral camera, we have conducted flights:

Sept 23
Oct 7
Oct 14
Oct 21
Oct 27
Nov 9
Nov 24
Dec 27
Jan 13

By using the Matrice 200 drone equipped with a multispectral camera, we are able to create normalized difference vegetation index (NDVI) images of the research plots. Adding the multispectral camera allows for us to measure light energy reflected off of vegetation and estimate physical and chemical properties that we would otherwise not see with RGB imagery. All flights have been and will continue to be conducted at 400 ft above ground level (AGL) with a 75% image overlap. Plots have been flown multiple times throughout the fall to observe initial cover crop stands, a few times throughout the winter, and will be flown at least bi-weekly during the spring until cover crop termination. Drone images will be stitched together using Pix4D photogrammetric software (Prilly, Switzerland) and then uploaded onto ArcGIS, where plot level data will be extracted and analyzed. Imagery from the Mavic Air will also be processed by Drone Deploy, a commercially available software more likely to be used by consultants and farmers. Visual imagery will be analyzed by Visible Atmospherically Resistant Index (VARI – (GREEN − RED) / (GREEN + RED − BLUE) and multispectral imagery will be processed by NDVI. Imagery from the Mavic Air (RGB) will also be processed by Canopeo, a free app that can estimate ground cover. Pix4D will also be used to create point clouds to determine biomass through structure from motion (Bedell et al., 2017). This method is not likely to be used by crop consultants but would help determine the accuracy of two dimensional biomass analyses.

Prior to cover crop termination, biomass will be collected from each plot by cutting all aboveground biomass at the soil surface from three 10.8 ft2 locations within each plot. The three biomass samples collected in one plot will be composited and oven-dried to a constant mass. Total carbon (C) and N content will be determined in each sample by combustion using an Elementar Vario Max CN Analyzer (elementar Americas, Mt. Holly, NJ) at the University of Delaware (UD) Soil Testing lab. Composite soil samples (consisting of 12 soil cores) will be collected from each plot at a depth of 8 in. Soil composites will be bagged and air-dried prior to being sent to the UD Soil Testing Lab where they will be oven-dried and sieved to 2 mm prior to analysis. Total N will be determined by combustion as described for biomass samples. Soil samples will also be analyzed for ammonium-N and nitrate-N colorimetrically using a Bran & Luebbe AutoAnalyzer 3 (Buffalo Grove, IL; Bran+Luebbe, 1998) following extraction with 2 M KCl solution (1:10 weight:volume) following methods of Mulvaney (1996).

The VARI, NDVI, and Canopeo measurements will be compared to the weight of cover crop biomass, as well as cover crop biomass N and soil N contents to determine the empirical relationships between the drone imagery and field tests; these relationships will be used to determine if potential N contributions  of cover crop residue to the cash crop can be predicted through drone imagery.

Objective 2: Verify the mathematical relationship between cover crops biomass and biomass available N from the research plots using data collected from farmer-planted cover crops.

The drone-derived biomass N availability rate curve will be tested in a minimum of three farmer-planted cover crop stands. Farmers will be recruited through a network of previous partnerships. Participating farmers will have planted cover crops in the fall but have not yet terminated them at the time of sampling in spring (about two weeks before planting corn). We will conduct a drone flight using the DJI Mavic Air (RGB camera) drone at each field. Based on the RGB imagery, we will map out 10 zones based on high and low biomass areas. Within each zone, we will conduct the same tissue and soil sampling and analysis as described for the research plots. Each zone that we are extracting samples from will be marked by a GPS unit that will allow for us to split them along a gradient using ArcGIS software to determine how useful drones are for choosing soil sampling locations. We will determine this by analyzing the results given back by the UD Soil Testing Lab and comparing them to the intensity of biomass depicted by the RGB images from the drone and look for any correlation.

Objective 3: Improve farmer productivity using rapid drone assessments to determine potential planting issues related to cover crop biomass.

Two weeks prior to cover crop termination of our research plots, we will plant corn as our cash crop in each (June of 2021). Once corn emerges, a stand count will be collected by hand in each plot to determine a cover crop residue’s potential to deter seedling germination. Stand count will be measured by counting the number of alive corn plants in the middle 17.5 feet of each plot. We will then take the stand count and multiply by 1000 to represent a whole acre. We will then compare stand successes to the dry weight of biomass density data collected from objective 1.

Participation Summary

Education & Outreach Activities and Participation Summary

1 Curricula, factsheets or educational tools
1 Webinars / talks / presentations

Participation Summary

Education/outreach description:

Findings from this study will be shared with farmers, Extension personnel, and crop consultants using a variety of educational methods. Upon the completion of the project, we will create a succinct report highlighting the results and outcomes of the study. Elements of this report will be placed into an ArcGIS Story Map, to outline the positive aspects of cover crops in Delaware, the use of drones in agriculture, and the success in mapping cover crops to improve soil health and crop production. This report and story map will be shared with farmers, certified crop consultants, researchers and Extension personnel via a publicly accessible website. RGB images will be a more economical option for farmers and consultants, so outreach will focus on providing the best value with consumer drones. We will also present results of this study at various University of Delaware Cooperative Extension events, Delaware Agriculture Week (January 2021/2022), Delaware Nutrient Management Certification Sessions, and the Mid-Atlantic Crop Management School (November 2021/2022). A field day on cover crops and drone scouting will also be held at the University of Delaware Carvel Research and Education Center in the spring of 2021.

Another source of outreach will include newsletter articles (e.g., University of Delaware Weekly Crop Update; and fact sheets for publication on the University of Delaware Cooperative Extension website. We will also utilize the Delaware Agronomy Blog to share results from this project in short articles addressing cover crops and the basics of drone flights. These resources are frequently accessed and freely available to farmers, crop consultants, researchers, the general public and Extension agents. With the data we compile, we will share results with professionals and agencies, like the Delaware Department of Agriculture (DDA) or DNREC, to support their endeavors.

Current efforts of outreach include a working ArcGIS Story Map sharing the need and reason for the study. The StoryMap will be shared through a presentation at Delaware Agriculture-Week on January 20th, 2021. Once all data is collected, the StoryMap will be updated and published for the public.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.