• J. Alex Thomasson (PI) - Agricultural & Biological Engineering, Mississippi State University
• Scott A. Shearer (Co-PI) - Biosystems & Agricultural Engineering, University of Kentucky

• Jian Chen (Postdoc) - Agricultural & Biological Engineering, Mississippi State University
• James Wooten (Graduate Student) - Agricultural & Biological Engineering, Mississippi State University
• John Fulton (Research Engineer) - Biosystems & Agricultural Engineering, University of Kentucky

The objectives of this project are to answer the following questions:

1. Can yield of cotton in Mississippi, corn in Kentucky, and wheat in Idaho, be accurately estimated from remotely sensed multispectral satellite data?
2. Can crop moisture stress be assessed from remotely sensed multispectral and thermal data?
3. Can biomass production of corn in Kentucky be estimated from remotely sensed multispectral data?
4. Can a fertilizer decision-support system developed for potatoes be used for corn in Kentucky and cotton in Mississippi with the aid of remote sensing data?

Yield: Multispectral satellite data will be collected on several dates over fields in the Delta region of Mississippi and in Kentucky. Historical grid soil sample data will be available for each field in the study. Data on crop canopy temperature will be collected in Mississippi. Geo-referenced yield data will be collected on corn in Kentucky and on cotton in Mississippi with an experimental harvester-mounted yield monitor, in addition to collecting handpicked yield information. All data will be incorporated into a GIS database, and the data will be examined in an effort to answer the questions in the objectives. In the proposed work herein, the new commercial earth observation satellite (Space Imaging Ikonos, 4-m resolution multispectral) is proposed as a primary data source along with Landsat 7 and aerial data from RSTC projects.

The data to be collected for this portion of the project are:

1. multispectral satellite data
2. ground-based infrared thermometry data and weather data for the Mississippi location
3. visual assessment of crop conditions at each location.

1. multispectral satellite data
2. spectrophotometric data collected by the PI on various plant parts, soils, and residues for the Mississippi location
3. geo-referenced yield monitoring data

Crop Stress: Two methods of assessing crop water stress will be utilized. The first involves determining the expression of crop water stress in the spectral bands offered by Ikonos (450-520 nm, 520-600 nm, 630-690 nm, and 760-900 nm), as in the work reported by Bausch (1995). At the Mississippi fields, canopy temperature will be measured with in situ infrared thermometers, and it will thus be used as the standard by which water stress is measured there. At all locations in this work, fields will be scouted manually, and subjective assessments of crop conditions, including water stress, will be made in light of the latest satellite imagery. The second method of examining crop water stress involves using the spatial variability of the satellite spectral data. It is known that certain parts of a field arrive at the point of water stress ahead of other parts of a field. In fact, the variability in water stress in a field increases as the field dries out. It is theorized that this fact will be expressed in increasing variability of spectral responses within the field.

Biomass: To effect variable-rate planting in no-till situations, field maps of variations in residue levels would be required. It is proposed that remote sensing data can be used to predict residue levels based on estimates of biomass production in the previous growing season. Again from Wiegand and Richardson (1990) and Wiegand et al. (1991, 1992), vegetation indices calculated from remote spectral observations measure the amount of photosynthetic tissue in crop canopies. It is thus believed that the remote sensing data proposed will provide excellent information regarding spatial biomass variations. The data to be collected for this portion of the project include multispectral satellite data for the Kentucky location.

Decision-Support System (DSS): Soil fertility data were collected immediately after fertilization in 1999. These data, along with 1999 yield data, will be incorporated into the DSS for its use in the 2000 growing season. Soil samples will be required in 2000 just prior to the time of fertilization. These samples will be collected on a one-acre grid from two cotton fields near Vance, Mississippi, totaling approximately 700 acres, and from two corn fields in Shelby County, Kentucky, totaling approximately 300 acres. These data and remote-sensing will be added to the DSS to develop prescriptions for variable-rate fertilization in 2000. A test will be developed to compare the economics of the DSS with the blanket-application method and the "conventional" variable-rate fertilization method.

Expected results include the ability to accurately estimate crop yield with multitemporal, multispectral remote sensing data. Success in this area will extend the current state of the art and provide farmers a well-defined method for using remote sensing data to estimate crop yield on a spatially varying basis. Expected results also include the ability to accurately relate estimated crop stress (primarily water-related) from common satellite remote sensing data with crop stress as determined by canopy temperature and visual crop-condition assessments. Success in this area will provide provide farmers a method for using remote-sensing data to determine proper irrigation timing. Further, it is expected that residue maps in no-till corn will be achievable. Finally, a sophisticated variable-rate fertilizer-application DSS will be adapted for corn in Kentucky and cotton in Mississippi.