(ICM) on Small New Jersey Farms
M. G. Hughes 1,2 , P. Tocco1, C. McGarrity1, D. Stanker1, J. Singer1, and D. L. Lee1
Rutgers Cooperative Extension1
Grant F. Walton Center for Remote Sensing and Spatial Analysis2
New Brunswick, NJ, USA
For large fields and farms in the Midwest and Western United States, precision agriculture tools based on Global Positioning Systems (GPS), Geographic Information Systems (GIS) and Remote Sensing Systems (RS) are used successfully to guide and manage large farming operations. GPS-led site-specific field management coupled with color and color-infrared aerial photography and satellite imagery, can optimize chemical inputs to fields, improve yields, identify crop stress and weed infestations, and improve environmental quality by decreasing nutrient run-off into watersheds (NRC, 1997). Two years ago, Rutgers Cooperative Extension (RCE) and the Grant F. Walton Center for Remote Sensing and Spatial Analysis (CRSSA) at Rutgers University began a unique program aimed at utilizing Global Positioning Systems (GPS) and its related technology-based tools (Geographic Information Systems and Remote Sensing) for agriculturally related projects. Our objective was to test whether site-specific precision agriculture tools work accurately and economically on the smaller fields and farms common to New Jersey grain, forage, fruit, and vegetable producers.
GPS and Precision Agriculture
The ability to determine an earth location with sub-meter accuracy creates
new opportunities for New Jersey agriculture, where field acreage varies
from small plots to hundreds of acres. A farmer can now 1) accurately map
field boundaries, 2) analyze specific sites within fields, 3) map and analyze
variability on a sub-field scale, 4) integrate generated field maps with
GPS controlled equipment such as seeders, sprayers, irrigation systems,
and yield monitors to fully adapt site-specific management practices and
5) utilize new digital camera and video technology to produce timely remotely
sensed, high resolution imagery for identification of crop stress, weed
infestations, and nutrient run-off into watersheds.
Over the past two years, a number of demonstration projects have been initiated through RCE and CRSSA using geospatial tools in agriculture. Approximately $12,000 was spent to purchase the following field equipment: 1) A Trimble AgGPSä Model 122 differential GPS receiver used for collecting positional information. Output from the receiver is captured in a Fujitsu pen-top computer running an integrated GPS/GIS site-specific agricultural software package FarmGPS (Red Hen Systems, Inc., purchased through the Farmer?s Software Association, Fort Collins, CO). 2) A DYCAM Agricultural Digital Camera (ADC), a portable rugged digital camera that senses in the near-infrared and the visible red portions of the electromagnetic spectrum. The camera has a resolution of 496x365 pixels and connects to a host computer through an RS-232 serial port. The software package that comes with the camera allows for calibration of ambient conditions and the computation of three different vegetative indices for identification of crop stress and weeds. 3) A VMS 200 video mapping system (Red Hen Systems, Inc., Fort Collins, CO ) used for aerial mapping of farmland and watersheds over large areas. The unit integrates with the AgGPSä 122 GPS unit to provide spatially referenced video information. The 8mm camcorder records video, GPS locations, and audio commentary. 4) A 35mm SLR camera equipped with a #12 yellow filter and Kodak Ektachrome Professional infrared EIR 135-36 film.
Nutrient and Pest Mapping
To date, over 100 fields on 16 farms in Burlington, Gloucester, and Salem counties have been mapped using the integrated GPS/GIS unit. An outreach package was created for each participating grower consisting of accurate field acreage and soil delineation maps. For some farmers who rent the land they farm, this provided the first accurate determination of arable land and resulted in immediate economic savings. In addition to boundary mapping, point sampling grids for selected fields were constructed for site-specific data collection throughout the growing season. Factors important to a number of growers such as pH, soil nitrate and phosphate levels, soil compaction, and presence/absence of various insects (potato leafhopper) and diseases (phytophthora) were measured. Spatial analysis techniques such as surface interpolation and statistical correlation were used to map and analyze such diverse factors as 1) the distribution and intensity of diseases, 2) insect outbreaks, and 3) the distribution of variables such as pH, soil type, soil nutrients, and organic matter. As New Jersey?s Department of Environmental Protection (NJDEP) is now taking a watershed management approach to solving environmental problems (NJDEP, 1998), many farmers are being implicated for non-point sources of nutrient run-off into watersheds (NJDA & USDA, 1998); however, they have little knowledge of the distribution of these nutrients on their fields. Preliminary results of these mapping projects show high within-field variability of pH, soil nutrient (nitrate and phosphate) levels on many fields. In addition, the integration of these site-specific data sets with existing 1995/1997 United States Geological Survey (USGS) 1 meter resolution digital ortho-photography enables us to analyze the proximity of these fields to watersheds. In the future, these measurements will be used in conjunction with digital elevation models to target areas within fields that contribute to nutrient runoff.
The summer of 1999, from June through mid-August was the driest growing season on record in New Jersey according to the National Oceanic and Atmospheric Administration (NOAA). Rainfall totals were well below normal (1961-1990 averages) with precipitation deficits ranging from approximately 22.9 cm (9 in.) in northern part of the state to 8.9 cm (3.5 in.) in the south. This past summer, a site-specific crop management program was established with two New Jersey corn growers. The boundary of a 12 ha (30 acre) irrigated cornfield and a 40.5 ha (100 acre) non-irrigated cornfield were mapped and site-specific sampling grids were laid out in each field. The location of irrigation risers for the 12 ha (30 acre) field were mapped and irrigation buffers created using ArcViewâ GIS (ESRI, Redlands, CA). At each sampling point a gypsum block moisture sensor (Delmhorst Inc.) was placed within the corn row. Electrical conductivity (EC) and chlorophyll content (Minolta-Chlorophyll Meter SPAD-502) readings were obtained twice weekly (irrigated) and weekly (non-irrigated) at each point throughout the growing season. Critical moisture thresholds were determined and converted to a color-coded irrigation recommendation map for the grower. Results of this project show that the irrigation strategy saved the grower both money and water throughout the summer, and resulted in a healthy high yielding corn crop.
Mapping crop stress
Detection and monitoring of the impact of insect, disease, and soil drainage properties on a wide range of vegetation and crop types has been successfully undertaken using remote sensing techniques (Moran et al., 1997; see Frazier et al., 1997 for review). Remote sensors are made up of detectors that record specific wavelengths of the electromagnetic spectrum (ERDAS, 1997). Indicators of plant stress can be developed using a variety of techniques based upon the spectral reflectance properties of the crop of interest and the radiometric information available from the remote sensing instrument in use. In general, chlorophyll in plants strongly absorbs energy in the blue and red portion of the electromagnetic spectrum, and reflects in the green. At slightly longer wavelengths, in the near-infrared part of the spectrum, the internal structure of plant leaves strongly reflects incoming solar radiation. The ratio of the near-IR to red reflectance known as a vegetative index (NDVI) is often used as a measure of plant health (or stress) and for plant species discrimination (Lillesand & Kieffer, 1987). This past summer three of our demonstration projects utilized a 35mm SLR camera with color and color-infrared film, an agricultural digital camera with a red and near-IR sensor, and a GPS-compatible video camera. A local pilot was contracted throughout the summer to obtain the images/photographs. The plane was not altered in any way, so images were obtained simply through an open window. Problems were encountered holding the camera steady and vertical at altitudes between 152 m (500 ft.) and 305 m (1000 ft.), and speeds of approximately 160 kph (100 mph). The process of incorporating imagery into a GIS requires ground control points, locations with known earth coordinates. Verification of exact field placement for integration with in-situ data was difficult when working on large fields. Nonetheless, obtaining any photographs/imagery remotely proved to be valuable. Figure 1 (shown in black and white) shows an image taken with the DYCAM ADC of a forage grass study at the Snyder Research Farm in Pittstown, NJ. The field is divided into three varieties of grasses and four within species treatments for each grass variety (fig. 1a). The image was imported into ERDASâ Imagine (Atlanta, GA) an image processing package and georeferenced. An NDVI was calculated and used to identify differences in plot treatments between varieties (fig. 1b). The final map was combined with previously GPS?ed boundaries in ArcViewâ GIS.
Photographs taken with a 35mm camera using both color and color-infrared film proved to be very useful for identification of crop stress. Figure 2 (shown in black and white) shows two corn fields in southern New Jersey one irrigated (fig. 2a), one non-irrigated (fig. 2b). The photograph of the irrigated cornfield appears uniform and dark red using the color-infrared film (representing healthy vegetation). Areas of water stress are minimal and concentrated in the high (dry) elevations of the field. The non-irrigated field appears to be highly stressed as evidenced by the light patches (shown in black and white) occurring throughout the photograph. Similar results were found using the DYCAM ADC although the resolution of the camera (496x365 pixels) was not as good as using film. The images from the cameras were supplemented by GPS-video imagery. The video camera provided growers with perhaps the best tool for identification of crop stress over large areas. The video was shown to growers and governmental officials at local meetings.
The GPS-video camera was used this past summer to determine its potential use as a management tool for watershed analysis. Weekly flights to evaluate the Salem Watershed, Salem County, NJ were made during the month of August. One of the biggest benefits initially realized from the use of aerial imagery, was the educational value it provided to producers who could visually identify the location of their fields with respect to the watershed.
Video images were analyzed to identify aquatic growth (algal blooms), sedimentation, and proximity to existing agricultural fields and urban/industrial centers (Figures 3a and 3b). Photographs taken with the 35mm SLR camera using color and color-infrared images were also collected. Future studies will deal with the correlation between site-specific water sample analyses and aerial imagery to determine potential for nutrient overload and sedimentation relative to livestock and cropping operations.
Growers in New Jersey are facing major obstacles in maintaining profitability.
New state and federal guidelines are placing more restrictions on the way
growers can farm. Crop prices continue to fall, and land values and development
pressure continue to rise. Those that stay in farming look toward Rutgers
Cooperative Extension to implement new technologies to enhance their farm
operations. Results from these demonstration projects suggest that GPS
and its related technologies (GIS and RS) will continue to play a significant
role in improving farm management practices in New Jersey. The demonstration
programs in Integrated Crop Management (ICM) and watershed management are
utilizing new technologies to 1) reduce fertilizer and pesticide costs,
2) improve crop management, 3) improve spray recommendations for pest control,
4) monitor irrigation systems, 5) monitor nutrient pollution to watersheds,
and perhaps most importantly, 6) provide value-added information to growers
to improve farm management and farm profits even on the finer spatial and
temporal scales applicable to New Jersey agriculture.
ERDAS, Inc., "Remote Sensing." In ERDAS Imagine Field Guide, Fourth Ed. Atlanta, GA, pp. 5-10, 1997.
Frazier, B. E., C. S. Walters, and E. M. Perry, "Role of Remote Sensing in Site-Specific Management," In: The Site-Specific Management for Agricultural Systems, ASA-CSSA-SSSA, Madison, WI. pp. 149-160, 1997.
Lillesand, T. M. and R. W. Kieffer, "Concepts and Foundations of Remote Sensing" In Remote Sensing and Image Interpretation." John Wiley & Sons, New York, pp. 15-19, 1987.
Moran, M. S., Y. Inoue, and E. M. Barnes, "Opportunities and limitations for image-based remote sensing in precision crop management." Remote Sens. Environ, Vol. 61, pp. 319-346, 1997.
NJDA and USDA, "On-farm strategies to Protect Water Quality." Prepared by the NJ Assoc. of Conserv. Dist. In cooperation with the NJ Dept. of Ag. and US Dept. of Ag. 95 p., 1998.
NJDEP, "New Jersey?s Environment 1998." NJ Dept. of Env. Prot., Division of Science and Research, 28 p., 1998.
NRC, "Executive Summary." Precision Agriculture in the 21st Century. Geospatial and Information Technologies in Crop Management, Eds. J. Dixon and M. McCann, National Research Council, National Academy Press, Washington, D. C., pp. 1-16, 1997.