6.6   Variable Rate Nitrogen Management

Investigators:

Thomas G. Mueller, Agronomy, mueller@uky.edu

Richard I. Barnhisel, Agronomy, rbarnhis@uky.edu

Haluk Cetin, Murray State University, haluk.cetin@murraystate.edu

Carl Dillon, Agricultural Economics, cdillon@uky.edu

Lloyd Murdock, Agronomy, lmurdock@uky.edu

Scott Shearer, Biosystems and Agricultural Engineering, shearer@bae.uky.edu

Timothy Stombaugh, Biosystems and Agricultural Engineering, tss@uky.edu

Cooperators

Rick Murdock (Ponderosa Farms)

Mike Ellis (Worth and Dee Ellis Farms)

Introduction:

            Many of the factors that govern crop response to inputs vary spatially within agricultural fields such as soil fertility status, soil water storage, drainage, organic matter, landscape position, etc. Variable rate technologies allow farmers to manage site specific variability site-specifically. The premise of variable rate management is that the optimal rate can be adequately predicted across an agricultural field. The predictions must also be timely and at a spatial scale suitable for management. The economic potential of variable rate technologies will be great for Kentucky; however, a great deal of work is still needed to develop site-specific recommendations. The main purpose of this proposal is to develop variable rate nitrogen recommendations for corn and wheat in Kentucky.

 Objectives

            The specific objectives of this proposal are to 1) develop a method for detection of nutrient stress in wheat; 2) develop site-specific nitrogen recommendations for corn and wheat in Kentucky.

Background

A crop or plant stress is a factor that prevents normal function and results in a reduction in growth or reproduction (Biswal and Biswal, 1999).  Plants have many direct and indirect responses to stresses including changes in pigmentation.  Pigments are important indicators of stress because they interact with light and therefore can be detected with remote sensing.

            Chlorophyll is perhaps the most important indicator of plant stress for remote sensing applications because it is responsive to many crop stresses (e.g. nutrient, drought, disease), and it plays a critical role in photosynthesis, a fundamental plant process.  Furthermore, differences in chlorophyll can be detected very accurately (Carter et al., 2001, reflectance at 638 nm; Blackburn et al., 1998, r2 = 0.96; [reflectance at 800 nm]/[ reflectance at 680 or 636], and Daughtry et al., r2 = 0.95, slope of the relationship between NIR/Green vs NIR/red]).  The chlorophyll in a normally functioning plant will absorb red and blue light while reflecting green light.  Often, leaf chlorophyll content decreases when plants are affected by some environmental stressor (Murtha, 1982).  As a result, absorption of visible light decreases, reflectance is increased, and chlorosis, or a general yellowing of leaves, may be observed. Remote sensing systems can often provide earlier and more comprehensive detection of such stress symptoms than visual observation in the field.  Since nitrogen molecules are an important constituent of chlorophyll, nitrogen stress in plants is often associated with chlorosis (reduction in chlorophyll content) therefore it can readily be detected (Blackmer et al., 1996; Schepers, 1998).  Further, some have found a linear relationship between the intensity of remotely sensed imagery and optimal side dressed nitrogen rates (Scharf, 2000; r2 = 0.51) by the intensity of green reflectance. 

We have found that remote sensed imagery can be used to detect nitrogen stress in corn in Kentucky with simple indices (Cetin et al., 2002; Fig. 1). But stress is very difficult to detect early in the growing season because of incomplete canopy coverage.  Tractor based “remote sensing” systems can be used to get around this problem. Unfortunately, even when it is detected, other areas in the field may become nitrogen stressed as the growing season progresses. An alternative approach to remote sensing for corn is to predict where nitrogen deficiencies are likely to occur. In areas where water stress is likely to occur, the potential for a nitrogen response is less likely.  Sensors that can be used to assess topography (Fig. 2), electrical conductivity, or soil permeability have may be useful in identifying areas where water stress is likely and therefore a nitrogen response is unlikely.  Elevation models (Fig. 3) and relationships between EC and depth of top soil are shown in Fig. 3

 

 

 

 

 

 

Fig. 1. Relationship between NDVI calculated from RDACS airborne imagery and IKONOS imagery in Calloway Co. Kentucky (Cetin et al., 2002).

 

 

 

 Figure 2. Topographic maps for various fields in Kentucky (T.G. Mueller, unpublished research data).

 

Fig. 3. Relationships in Kentucky between EC and topsoil depth in the surface 50 cm (20 in; Hartsock et al., in submission).

 

            Winter wheat has full canopy closure early in the spring prior to nitrogen application; therefore detecting nitrogen stress is less problematic than for corn. In a variable rate wheat study conducted in Owensboro, Kentucky (Spectra Visions, 2001), there was no significant increase in yield with variable rate N-management over stand practices, but there was some reduction in the amount of fertilizer applied (6 lbs acre-1). It is very unlikely that the reduced costs of N inputs offset the overhead costs associated with the technology in this study. Another study in North Carolina showed more promise (Flowers, 2001). As with corn, nitrogen deficiencies may occur well into the growing season for example in areas that are prone to denitrification losses. Combining ground based sensors with remote sensing may improve the economics of variable rate nitrogen for wheat in Kentucky.

            Our interest is in describing the spatial variability of crop response across Kentucky fields. Then, we will attempt to understand the factors that affect this variability and try to develop proxies to predict where responses are likely to occur.

            There is currently a great deal of ongoing work in Kentucky in the development of variable rate nitrogen rates. For example here are some examples of funded projects.

1) Shearer, S.A. S.F. Higgins, T.G. Mueller, T.S. Stombaugh, C.R. Dillon, J.P. Fulton. 2001. Equipment enhancements to support variable rate response surface development. Phase III funding; 2) T.G. Mueller, T.S. Stombaugh, S.A. Shearer, R.L. Barnhisel. C.R. Dillon, L. Murdock, H, Cetin, M. Bitzer, M. Collins, G. Thomas, J. Grove. 2001. Sensors and variable rate management. Phase III funding; 3) Murdock, L. S.A. Shearer, C. Dillon, T. Mueller, P. Needham, S. Fischer, and J. Ebersbacher, and T. Gress. The use of remotely sensed imagery to make nitrogen recommendations for winter wheat in Western Kentucky. NASA Ag. 20/20 project; 4) Cetin, H., W. Spencer, C.J. Liu, T.G. Mueller, S.A. Shearer. G.A. Carter, R.O. Green, T. Gress. Establishment of a research cluster: commercialization of remote sensing in Kentucky. NASA-EPSCoR. We will use these funds to continue our work in this area.

Procedures

            We will relate corn and wheat response to nitrogen to terrain attributes, EC and other sensor data, and remote sensed imagery. The experimental design for the nitrogen experiments will follow the same model that was proposed in the phase III proposal (Fig. 4)

Fig.4. Experimental design and layout of nitrogen experiment for corn which will be replicated multiple times throughout each field. From Mueller et al., 2001, phase III USDA funding proposal.

 

            For the Phase III funding, we requested funding for a post doc. This post doc will be responsible for day to day project organization, ground based sensor acquisition, airborne imagery acquisition, experimental design, and data analysis. We are requesting funding in the following areas: 1) a part time engineering associate who will be responsible for planting, nitrogen application, and harvesting of the experiment, 2) support from the Mid America Center for Remote Sensing for aerial image analysis, 3) funds for a multispectral camera, 4) support in year 3 for economic analysis of the spatial response data and variable rate nitrogen recommendations.

            Dr. Murdock is providing expertise with wheat management and soil nitrogen management. Dr. Barnhisel is providing his expertise in variable rate nitrogen management. Dr. Cetin will provide expertise in the handling of remote sensed data. Dr. Mueller, Stombaugh, and Shearer will provide support in with sensors and RTK GPS data collection. Dr. Mueller will provide support in statistical and spatial analysis.  Dr. Mueller and Dr. Cetin will be involved with ground truthing.  All PI’s will have some involvement in outreach activities but Dr. Murdock and Stombaugh will have leading roles

Expected Benefits

            We expect to develop variable rate fertilizer recommendations for corn and soybeans. Our hope is that these will make crop production more profitable in Kentucky.

Deliverables

            Recommendations for variable rate nitrogen management will be made available to Kentucky farmers through UK cooperative extension publications.