Progress 07/01/03 to 06/30/08
Outputs
OUTPUTS: A survey was conducted by phone to 1,195 growers in six states (Illinois, Indiana, Iowa, Mississippi, Nebraska, and North Carolina) to measure producers' cropping history, perception of glyphosate-resistant weeds, past and present weed pressure, tillage practices, and herbicide use as affected by the adoption of glyphosate-resistant (GR) crops. When growers were asked if they were aware of documented cases of glyphosate weed resistance in their state, 42% of growers answered "yes", while 58% of growers answered "no". Responses to specific weed species in their state with documented glyphosate resistance include horseweed (39%), waterhemp (20%), and pigweed (14%). The majority of growers (76%) are not aware on glyphosate resistance in weeds. Of the percentage of growers aware of glyphosate resistance in weeds, only 62% ranked a "5" or higher for how serious a problem they believed resistant weeds to be. These results are surprising, considering the weed species that have developed resistance to glyphosate. Growers listed farm publications, dealers/retailers, and university/extension as their 3 primary sources of information concerning weed resistance issues. When a producer is making decisions about weed control programs one of the top deciding factors is the economic benefit of the program. This study was conducted to assess long-term viability of glyphosate-resistant technology as a foundation for cropping systems. The research team established 156 on-farm sites beginning in 2006. Production systems were divided into three groups for this assessment: continuous glyphosate-resistant (GR) crop, GR crop rotated with a different GR crop, and GR crop followed by a non-GR crop. The on-farm sites were divided into two treatments on halves of the field: the producer's normal glyphosate-based program and university recommendations based on weed resistance management principles. Across both treatments, costs were kept equal except for the application and herbicide cost. Across all production systems except for GR corn followed by a non-GR crop herbicide cost for the universities resistance weed management system was significantly higher than the producer's normal glyphosate-based program. Among all systems there was no statistically significant yield difference between the universities and producers herbicide programs. Although not significant, yield for the universities program was slightly higher than that of the producers program. Even though the university-recommended weed resistance management program required more intensive inputs and management, they resulted in similar yields and net returns. Thus, resistance management strategy can preserve glyphosate as an effective tool, without reduction in farm profits. PARTICIPANTS: Nothing significant to report during this reporting period. TARGET AUDIENCES: Nothing significant to report during this reporting period. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.
Impacts
Researchers and extension weed scientists and agronomists must be more diligent and proactive when it comes to providing information on weed resistance to glyphosate. Resistant weed management strategies have the potential to provide equivalent net returns, while improving long-term management of resistant weed occurrence.
Publications
- Shaw, D.R., W. Johnson, R. Wilson, M. Owen, B. Young, K. Gage, S. Weller, J. Wilcut, and D. Jordan. 2008. Assessing long-term viability of Roundup Ready technology as a foundation for cropping systems. Proc. International Weed Science Society 4:255.
- Owen, M.D., W.G. Johnson, S.C. Weller, D.R. Shaw, J.W. Wilcut, D.L. Jordan, B.G. Young, R.G. Wilson, D.J. Gibson, and K.L. Gage. 2009. How did weed species known to be confirmed glyphosate-resistant respond to various cropping systems Weed Science Society of America Abstracts 44:306.
- Weirich, J.W., D.R. Shaw, W.A. Givens, J.A. Huff, W.J. Everman, D.L. Jordan, W.G. Johnson, S.C. Weller, M.K. Owen, R.G. Wilson, and B.G. Young. 2009. Assessing long-term viability of glyphosate-resistant technology as a foundation for cropping systems-on-farm economic comparisons of management systems. Weed Science Society of America Abstracts 44:301.
|
Progress 01/01/07 to 12/31/07
Outputs
OUTPUTS: A survey was conducted by phone to 1,195 growers in six states (Illinois, Indiana, Iowa, Mississippi, Nebraska, and North Carolina). The survey measured producers' cropping history, perception of glyphosate-resistant weeds, past and present weed pressure, tillage practices, and herbicide use as affected by the adoption of glyphosate-resistant (GR) crops. GR cropping systems substantially increased the percentage of growers using no-till and reduced-till systems. Twenty five percent of growers in reduced-till systems changed to no-till systems after the adoption of a GR crop. Of the growers using conventional tillage, 24 and 31 percent changed to no-till and reduced-till systems, respectively, after the adoption of a GR crop. The adoption of a GR cropping system also had a significant effect on growers' tillage system. Analysis of tillage shifts by crop rotation indicated that in continuous GR cotton and continuous GR soybean, a significant impact was noted with respect to the change in tillage. No difference was found among the other crop rotations. Among continuous GR cotton producers, there was a 26 percent increase in the number of growers using a no-till system after adoption of GR cotton. Among continuous GR soybean producers, there was a 21 and 24 percent increase in the number of growers using no-till and reduced-till systems, respectively, after adoption of GR soybean. Analysis of tillage shifts by state indicated a significant effect of state with respect to tillage shifts after the implementation of a GR cropping system. The largest and most significant shift was from Nebraska with 22 percent of growers shifting from conventional and reduced-till systems to no-till systems. The lowest shifts toward more conservative tillage systems came from the states of Illinois and Iowa, with 11 percent of growers shifting from conventional and reduced-till systems to no-till systems. One of the factors a producer considers when selecting a production practice is the economic benefit of the system. Data from 2006 and 2007 were analyzed by herbicide cost, yield and net return. Among the four cropping systems herbicide cost was only different for corn in the 2007 growing season. In 2006 and 2007, weed management systems were not different in yield or net returns. In continuous GR cotton herbicide costs for grower vs. university systems were 43.54 dollars and 60.17 dollars, respectively; net returns were 417.19 dollars and 409.79 dollars, respectively. In continuous GR soybean herbicide costs for grower vs. university were 22.50 dollars and 30.27 dollars, respectively; net returns were 220.52 dollars and 223.91 dollars, respectively. When GR soybean was followed by a non GR crop, the GR soybean herbicide costs for grower vs. university were 24.76 dollars and 33.15 dollars, respectively; net returns were 257.59 dollars and 265.70 dollars. Thus, even though the university-recommended weed resistance management program required more intensive inputs and management, they resulted in similar yields and net returns. PARTICIPANTS: No Participant information reported. TARGET AUDIENCES: No Target Audiences information reported. PROJECT MODIFICATIONS: No Project Modifications information reported.
Impacts
Adoption of glyphosate-resistant cropping systems has substantially increased the use of conservation tillage, thereby enhancing soil conservation and environmental impact. Herbicide resistance management programs resulted in no reduction in net income; thus resistance management does not cause any profit penalty, while at the same time preserves the value of glyphosate-resistant crop technology.
Publications
- Gray, C. J., D. R. Shaw, J. A. Bond, D. O. Stephenson, IV, and L. R. Oliver. 2007. Assessing the reflective characteristics of Palmer amaranth (Amaranthus palmeri) and pitted morningglory (Ipomoea lacunosa) accessions. Weed Science 55:293-298.
- Shaw, D. R., W. A. Givens, P. D. Gerard, L. A. Farno, J. W. Wilcut, B. G. Young, R. G. Wilson, M. D. K. Owen, S. C. Weller, and W. G. Johnson. 2007. Assessing long-term viability of Roundup Ready technology as a foundation for cropping systems. Weed Sci. Soc. Am Abstr. 43:318.
- Prince, J. M., D. R. Shaw, L. A. Farno, W. A. Givens, P. D. Gerard, J. W. Wilcut, B. G. Young, R. G. Wilson, M. D. K. Owen, S. C. Weller and W. Johnson. 2007. Shifts in weed problems following adoption of Roundup Ready technology in continuous soybean cropping systems. Proc. South Weed Sci. Soc. 60:38.
|
Progress 01/01/06 to 12/31/06
Outputs
Remotely sensed images offer the potential for rapidly developing weed maps for precision agriculture. To effectively detect weed patches in remotely sensed multispectral or hyperspectral images, one can use spatial and/or spectral information. In low spatial resolution imagery, spectral information can be used to discriminate between broad classes, such as vegetation and non-vegetation, whereas spatial information can be used to discriminate between homogeneous areas versus non-homogeneous areas in the image using techniques such as texture analysis. The objective of this research was to test the ability of multiresolutional texture analysis methods for the detection of aggregated weed patches in a soybean. Weed species, density of each species, and overall percent ground cover was recorded in each sq. m of a 70 x 70 m area of the field. The remote sensing data collected was multispectral imagery acquired the same day as the ground truth data with data acquisition
dates of June 24, 2005 and August 9, 2005. Discrete wavelet transforms were used for feature detection; these features were compared to ground truth data collected the day of image acquisition to determine their ability to (a) discriminate between areas free of weeds and areas where weeds were present, and (b) discriminate between areas requiring a herbicide application and areas not requiring a herbicide application. The method was tested on a Normalized Difference Vegetative Index (NDVI), Soil Adjusted Vegetative Index (SAVI), and Transformed Difference Vegetative Index (TDVI) derived from the multispectral imagery. The near infrared (NIR) band was also used individually in the analysis. The analysis was performed at three different spatial resolutions: 0.14, 0.5, and 1.0 m. Weed presence or absence data produced classification accuracies 22% higher than the herbicide application data. Furthermore, classification accuracies using the weed presence or absence data using the August 9,
2005 dataset were higher than those from the June 24, 2005 dataset. The framework for a site-specific herbicide application decision support system (DSS) is under development. This DSS will use the concepts developed by North Carolina's HADDS (Herbicide Application Decision Support System). This DSS with utilize interpolated weed maps developed by the user to generate optimized herbicide recommendations for either single product, variable rate applicators or multiple product, variable applicators.
Impacts
Computer software to develop site-specific herbicide application spray maps based on weed populations enables reductions in herbicide usage and an increase in effectiveness, thus providing both economic and environmental benefits. Intensive ground sampling is required to accurately determine weed distributions in fields. Other means such as remote sensing techniques will be essential to make site-specific weed management affordable.
Publications
- Hutto, K. C., D. R. Shaw, J. D. Byrd, Jr., and R. L. King. 2006. Differerentiation of turfgrass and weed species using hyperspectral radiometry. Weed Sci. 54:335-339.
- Shaw, D. R., and J. L. Willers. 2006. Improving pest management with remote sensing. Outlooks in Pest Management 17:197-201.
|
Progress 01/01/05 to 12/31/05
Outputs
The ability to manage different parts of a field according to need at that particular site in the field is the basic concept of precision agriculture. One practice that site-specific applications could be justified is weed management since it has been established that weeds often occur in patches across agricultural landscapes. However, to take advantage of these new technologies, there must be a way of accurately producing weed distribution maps that can be used to develop site-specific herbicide recommendations. The objective of this study was to evaluate two traditional sampling techniques for their ability to create accurate weed maps. The two techniques evaluated were the W- and Z-shaped sampling technique. Both interpolation methods tend to overestimate the densities for all 3 weed species in both sampling schemes. Weed maps interpolated using the W-shaped sampling method in general tended to overestimate weed population values less than those interpolated using
the Z-shaped sampling method. The best results for each species were obtained using the: Z-shaped sampling method for horsenettle, and the W-shaped sampling method for yellow nutsedge and broadleaf signalgrass. Several decision support systems (DSSs) for weed management have been developed to aid producers in this decision making process. These DSSs give herbicide recommendations for the entire field based on the competitive index for each weed, crop growth stage, herbicide cost, and application cost. One of the pitfalls of these DSSs is their whole field approach to herbicide recommendations. A site-specific herbicide application DSS would be a more logical weed management tool. This DSS utilizes interpolated weed maps developed by the user to generate optimized herbicide recommendations for either single product, variable rate applicators (VRA) or multiple product, variable applicators. Results give users ten herbicide recommendations for each map cell ranked according to net
return. Another product generated is a table of the top ten unique herbicide recommendations for all map cells. Information for each herbicide recommendation includes the net return if applied to the entire field (whole field recommendation) and net return if applied site-specifically. Field experiments were conducted to evaluate the effects of residual grass herbicides applied mid-POST in tank-mixes with glyphosate on early-planted glyphosate-resistant soybean at two row spacings for annual grass control. It was thought that tank mixtures of glyphosate might provide added residual control for late-season weed flushes. At two weeks, the only different treatment was a 0.69 kg/ha early-POST application of glyphosate with no tank-mix or M-POST treatment, which was not as effective. No differences in row spacing with regard to control were noted. At six weeks, all treatments containing a tank mixture were more effective than glyphosate alone, but were not different from one other. A 0.69
kg/ha early-POST application of glyphosate was not different than the untreated check. At this evaluation timing weed control in the 38-cm row spacing was better than control in the 76-cm row spacing.
Impacts
Computer software to develop site-specific herbicide application spray maps based on weed populations enables reductions in herbicide usage and an increase in effectiveness, thus providing both economic and environmental benefits. Intensive ground sampling is required to accurately determine weed distributions in fields. Other means such as remote sensing techniques will be essential to make site-specific weed management affordable. Tank mixing a residual herbicide with glyphosate postemergence in soybean provides more effective full-season weed control than glyphosate used alone.
Publications
- Flint, S.G., D.R. Shaw, F.S. Kelley, and J.C. Holloway. 2005. Effect of herbicide systems on weed shifts in soybean and cotton. Weed Technology19:266-273.
- Shaw, D R, and F. S. Kelley. 2005. Remote sensing for determining and classifying soybean anomalies. Precision Agriculture 6:162-170.
- LaMastus, F.E., and D.R. Shaw. 2005. Comparison of different sampling scales to estimate weed populations in three soybean fields. Precision Agriculture 6:271-280.
- Shaw, D. R. 2005. Introduction to the symposium on site-specific weed management. Weed Science 53:220.
- Shaw, D. R. 2005. Remote sensing and site-specific weed management. Frontiers in Ecology 3:526-532.
- Shaw, D. R. 2005. Translation of remote sensing data into weed management decisions. Weed Science 53:264-273.
|
Progress 01/01/04 to 12/30/04
Outputs
Recent advances in remote sensing technology have improved weed detection capabilities. Experiments evaluated the ability to separate reflectance characteristics from biotypes of Palmer amaranth and pitted morningglory accessions originating from across the species' indigenous range. There were no predictable trends in accession collection origin for either species, which would indicate that remote sensing can be used to separate species without regard to local conditions. To effectively detect weed patches in a remotely sensed multispectral or hyperspectral image, one can use spatial and/or spectral information. In low spatial resolution imagery, spectral information can be used to discriminate between broad classes, such as vegetation and non-vegetation; whereas, spatial information can be used to discriminate between homogeneous areas versus non-homogeneous areas in the image. The objective of this research was to test the ability of multi-resolutional texture
analysis methods for the detection of aggregated weed patches in a soybean production setting. Significant correlations occurred in the vertical decompositions for both grass and broadleaf weed categories, suggesting that discrete wavelet transforms show some promise as an alternative method for weed detection with spectral remote sensing. Weeds are typically absent in many portions of fields; however, until recently entire fields required treatment when any portion was infested. This study combined all previously collected data (>30 species) into a large dataset and examine classification accuracies with newly developed feature extraction techniques. When species were analyzed as a whole (no ancillary data included) overall classification accuracy was 42%. However, Hyperspec was successful at classifying hemp sesbania 95% correctly, kudzu 75% correctly, and Dubbs soil 86% of the time. However, classification accuracies were generally inadequate. Species were then aggregated into
groups commonly observed together in the natural environment. Classification accuracies for Dubbs, Marietta, Dundee, and Leeper soils were greater than 99%. Also classification accuracies for field crop residue were more than 95%. Classification for selected turfgrass and turfgrass weed species were more than 93%. Results indicate that hyperspectral remote sensing may be a viable option for detection of selected species and soil types. However, more research is needed to determine the level of ancillary data needed to correctly classify all species and soil types.
Impacts
Site-specific herbicide applications can reduce the amount of herbicide applied by over 50%. This can lead to economic and environmental benefits. Remote sensing can be used to delineate weedy from weed-free portions of fields, thus enabling these precision applications.
Publications
- Henry, W.B., D.R. Shaw, and L.M. Bruce. 2004. Spectral reflectance curves to distinguish soybean from common cocklebur (Xanthium strumarium) and sicklepod (Cassia obtusifolia) grown from varying soil moisture. Weed Science 52:788-796.
- Henry, W.B., D.R. Shaw, K.R. Reddy, L.M. Bruce, and H.D. Tamhankar. 2004. Remote sensing to detect herbicide drift on crops. Weed Technology 18:358-368.
- Henry, W.B., D.R. Shaw, K.R. Reddy, L.M. Bruce, and H.D. Tamhankar. 2004. Remote sensing to distinguish soybean from weeds after herbicide application. Weed Technology 18:594-604.
- Koger, C.H., D.R. Shaw, K.N. Reddy, and L.M. Bruce. 2004. Detection of pitted morningglory (Ipomoea lacunosa) by hyperspectral remote sensing. I. Effects of tillage and cover crop residue. Weed Science 52:222-229.
- LaMastus-Stanford, F.E., and D.R. Shaw. 2004. Evaluation of site-specific weed management implementing the herbicide application decision support system (HADSS). Precision Agriculture 5:411-426.
|
Progress 01/01/03 to 12/31/03
Outputs
Weeds are known to occur in patches; however, production fields typically receive herbicide applications on the entire field. Research was conducted on multiple soybean fields in eastern Mississippi, in which multiple scouting techniques were used to estimate weed populations. The first method involved establishing grids and sampling. Another method involved tracing the perimeter of infested areas using a global positioning system. The final method used multispectral aerial imagery. Grids may be used to produce accurate weed maps, but may not be feasible due to the intensity of labor required. The tracing method appears to have promise in fields with weeds distributed in patches; however, when weeds are evenly distributed across fields the tracing method appears to have reduced utility. Aerial imagery classified fields as weedy or weed-free correctly 88 to 91% of the time. An experiment was designed to determine if ground-truthed weed maps could be correlated to
multispectral aerial imagery. Weed species and the total weed populations were interpolated by inverse distance weighting to develop weed maps. Multispectral aerial imagery at 0.5-m resolution containing four spectral bands was also collected. These bands were then used to create six vegetation indices. These six indices, the four spectral bands, and elevation were then correlated to the eight weed species and total weed populations. Correlation between variables was either non-significant or relatively low. Broadleaf signalgrass and total weed population correlation was greater then any other weed species. Broadleaf signalgrass correlation to the vegetation indices ranged from 0.11 to 0.29. Total weed population correlation to the vegetation indices ranged from 0.07 to 0.23. No significant correlation was determined for knotroot foxtail and entireleaf morningglory. Low correlations can be expected due to the size of the weeds at the time of data collection and the presence of the
soybean crop. The small weed and crop sizes also allowed for greater background soil reflectance, which ultimately gave lower spectral band values when the multispectral image was acquired. These results also suggest only the major weed species present in a production field will be detected using early-season multispectral imagery. Research was conducted to evaluate multispectral imagery for the development of herbicide application maps. Aerial imagery of each field was obtained within 5 days of sampling. Supervised and unsupervised classifications, as well as NDVIs, were used to build site-specific herbicide application maps. These prescription maps were then compared to the grid-sampled data, and actual herbicide recommendations to determine the prescription maps accuracy. Aerial imagery was successfully used to detect weeds and build site-specific prescription maps. Classification accuracies for distinguishing weeds from soybean ranged from 63 to 90% over the two years.
Additionally, prescription spray maps generated for the fields showed areas equal to 53 and 52% of the fields required no herbicide spray in 2002 and 2003, respectively.
Impacts
Site-specific herbicide applications can reduce the amount of herbicide applied by over 50%. This can lead to economic and environmental benefits. Remote sensing can be used to delineate weedy from weed-free portions of fields, thus enabling these precision applications.
Publications
- Hutto, K. C., D. R. Shaw, G. E. Coats, and J. Vickery. 2003. Differentiation of turfgrass stresses with hyperspectral radiometry. Proc. South. Weed Sci. 56:230.
- Kelley, F. S., D. R. Shaw, and J. W. Easley. 2003. Weed detection and classification in soybean using hyperspectral remote sensing. Proc. South. Weed Sci. 56:22.
- Koger, C. H., D. R. Shaw, C. E. Watson, and K. N. Reddy. 2003. Detecting late-season weed infestations in soybean (Glycine max). Weed Technology 17:696-704.
- Koger, C. H., L. M. Bruce, D. R. Shaw, and K. N. Reddy. 2003. Wavelet analysis of hyperspectral reflectance data for detecting pitted morningglory (Ipomoea lacunosa) in soybean (Glycine max). Remote Sensing of Environment 86:108-119.
- Leon, C. T. Leon, D. R. Shaw, L. M. Bruce, and C. Watson. 2003. Effect of purple (Cyperus rotundus) and yellow nutsedge (C. esculentus) on the growth and reflectance characteristics of cotton and soybean. Weed Science 51:557-564.
- Leon, C. T., D. R. Shaw, M. S. Cox, C. Watson, M. J. Abshire, B. Ward, and M. C. Wardlaw, III. 2003. Utility of remote sensing in predicting crop and soil characteristics. Precision Agriculture 4:359-384.
- Easley, J. W., D. R. Shaw, and M. L. Tagert. 2003. Assessment of sampling techniques for generation of site-specific spray maps in soybean. Proc. South. Weed Sci. 56:312.
- Easley, J. W., W. B. Henry, and D. R. Shaw. 2003. Hyperspectral reflectance as a tool for estimating herbicide injury in corn (Zea mays), soybean (Glycine max), and four weed species. Proc. Am. Soc. Photogramm. and Remote Sensing 108:194.
- Gao, W., and D. R. Shaw (editors). 2003. Ecosystems dynamics, agricultural remote sensing and modeling, and site-specific agriculture. SPIE the International Society for Optical Engineering Conference, 7 August, 2003, San Diego, CA.
- Gray, C. J., and D. R. Shaw. 2003. Reflectance characteristics of various weed and crop species. Proc. Am. Soc. Photogramm. and Remote Sensing 108:172.
- Gray, C. J., and D. R. Shaw. 2003. Remote sensing for differentiating soybean and six weed species. Weed Sci. Soc. Am. Abstr. 43:218.
- Henry, W. B., D. R. Shaw, J. W. Easley, and W. P. Williams. 2003. The use of GPS, GIS, and remotely sensed hyperspectral imagery to detect paraquat injury to corn. Weed Sci. Soc. Am. Abstr. 43:246.
|
|