Non-Technical Progress Report--2001

Reporting Period: 9/15/00 - 9/14/01

Specific Aims: Ground- and surface water nitrogen (N) contamination from southeastern Coastal Plain agriculture is a regulatory and social issue threatening regional crop production. Research has shown that we can use remote sensing techniques such as aerial photography to determine crop N status and guide management of N fertilizer. Site-specific management (a.k.a. "precision agriculture") seeks to determine how soils and crops vary within fields and alter management accordingly from place to place to accommodate that variability. Our goal is to evaluate remote sensing-informed, site-specific, variable-rate N management in winter wheat and corn for its potential to optimize N-use efficiency (NUE) and reduce water N contamination in Coastal Plain cropping systems. Our specific objectives are to:

1. compare conventional, uniform N fertilizer management, in which a constant rate of fertilizer is applied everywhere in a field, to intensive strategies that use remote sensing of crop status and other crop and soil information to optimize N fertilizer rate and timing on both a whole-field and site-specific basis;
2. evaluate management effects on NUE and on soluble soil and groundwater N;
3. continue development of N management techniques and decision support systems based on remote sensing and crop and soil information; and
4. demonstrate N management strategies and consequences to producers, extension agents, crop advisors, students, the media, and general public through field expos, training programs, course modules, and the internet.

Results: In Fall, 2000, we established a large wheat-soybean-corn rotation experiment in a 33-acre field on a research station in Kinston, NC. This field has four main soil types typical of the Coastal Plain. To characterize soil spatial variability, we took soil samples every fifty feet in an equilateral triangle grid and analyzed them for plant-available nutrients and other properties. Over the four-year course of the experiment, we will impose three N fertilizer management treatments on wheat and corn in rotation:Treatment 1: whole-field N management based on a realistic yield expectation previously estimated for the main soil type in the field, which is a current management strategy; Treatment 2: whole-field N management based on in-season estimates of optimal N timing and rates derived from remote sensing and ground sampling of crop status and soil conditions averaged across the field, representing intensive management; and Treatment 3: site-specific N management based on the same information as Treatment 2, but with N applied at different rates in various parts of the field according to varying crop and soil conditions. We planted winter wheat and divided the field into30 large (0.9 acre) plots, providing 10 replications of the treatments, which will help us distinguish differences among them. To monitor the environmental fate of fertilizer N, we installed devices called suction lysimeters to extract water from soil; we also installed wells to monitor N leaching from the crop rooting zone into the groundwater. Samples were collected every two weeks and analyzed for N.

To achieve optimal yield, wheat must establish a sufficient number of stems or "tillers" early in the season. If wheat tiller density is below a threshold of about 500 tillers per square meter, an early-season N fertilizer application is needed to promote tiller formation and optimize yield potential. In previous research (Flowers et al., 2001), we showed that we can estimate early-season wheat tiller density from aerial color-infrared (CIR) photographs. Using aerial CIR photographs taken in February 2001, and tiller densities determined on the ground, we developed a mathematical relationship to estimate tiller density at any point in the field. We improved our estimates by developing one relationship for one soil type and a second for the rest of the field. Tiller densities indicated that no N was needed for the uniform Treatment 1 plots, nor for the uniform, crop-informed Treatment 2 plot. However, some areas within the crop-informed, site-specific management Treatment 3 plots had tiller densities below the N application threshold. We developed and used fertilizer application maps to apply N only to these areas. We also created calibration strips by applying various rates of N fertilizer to wheat growing just outside the experimental area.

Previous research showed that information in CIR photographs could be combined into an index (Green Normalized Difference Vegetation Index: GNDVI) which correlates with wheat N content, which in turn can be used to determine optimal fertilizer N rates somewhat later in the season. Four weeks after the first N decision, we used a new aerial CIR photograph and information from the fertilizer calibration strips to estimate wheat N content everywhere in the field and determine the appropriate amount of N to apply to the crop-informed treatments. We applied fertilizer N either uniformly (Treatments 1 & 2), or on a site-specific basis at one of seven discrete N rates to the Treatment 3 subplots. Wheat grew well until April 2001, when a late freeze aborted wheat flowers, preventing grain from developing. We had planned to harvest wheat grain with a combine equipped with a yield monitor and differential global positioning system (DGPS). Instead, we used a forage plot harvester to harvest strips from each plot. We analyzed wheat forage samples for N content. We georeferenced the harvest strips using DGPS, which will enable us to develop maps showing how wheat yield and wheat N content varied across the field in response to the treatments. Following wheat harvest, we planted soybeans to continue the planned rotation. We procured and installed profiling time domain reflectometry (TDR) probes that allow us to measure soil moisture periodically over multiple depths in the crop rooting zone. Moisture availability is an important factor affecting crop growth, yield, and NUE. We will use the information from the TDR probes to better understand crop-soil-moisture relations and improve our N management strategies. Collaborating researchers from Virginia Tech. are using aerial photography to study relationships between soybean development, pests and diseases, and soil moisture.

We established a web site (http://www.precisionag.ncsu.edu/projects/ifafs2000/) to publicize the project. We recruited and engaged a postdoctoral research associate and two Ph.D. students. We have offered a third graduate student an assistantship for Spring 2002, and we are recruiting a fourth Ph.D. student.

Plans for the Coming Year: We will harvest soybeans using a combine equipped with a yield monitor and DGPS, enabling us to determine how soybean yield varied across the field in response to soil and moisture conditions. We will continue to monitor soil solution and ground water N, and to analyze this year=s data to determine the effects of the experimental treatments. Outside the main experiment, we will continue corn and wheat research that is being done to refine techniques for determining crop status and appropriate rates, timing, and placement of N fertilizer. In Summer 2002, we will plant corn and make in-season N application decisions based on techniques similar to those described above for wheat. The ultimate goal is to provide growers with tools enabling them to farm more efficiently and profitably, while improving environmental stewardship.

Publications: None to date.

Reference: Flowers, M., R. Weisz, and R. Heiniger. 2001. Remote sensing of winter wheat tiller density for early nitrogen application decisions. Agron. J. 93:783-789.