|Gerard, P - MISS STATE UNIVERSITY|
|Mckibben, P - MCKIBBEN AG SERVICES|
|Hood, K - PERTHSHIRE FARMS|
|Samson, S - MISS STATE UNIVERSITY|
|Bethel, M - ITD SPECTRAL VISIONS|
Submitted to: Precision Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: September 13, 2006
Publication Date: December 1, 2005
Citation: Willers, J.L., Jenkins, J.N., Ladner, W.L, Gerard, P.D., Boykin, D.L., Hood, K.B., McKibben, P.L., Samson, S.A., Bethel, M.M. 2005. Site-specific approaches to cotton insect control. Sampling, and remote sensing analyses techniques. Precision Agriculture. 6:431-452. Interpretive Summary: Sampling for insects in cotton fields is a complicated process because several phenological stages of cotton plant growth are present in each field. These represent different ecological habitats for colonizing insects. This manuscript describes how to use ecological and sampling concepts in conjunction with remotely sensed multi-band imagery of cotton fields in a system that results in an improved sampling process to determine insect pest densities. We also describe how the choice of a single plant or a collection of plants as the sample unit impacts the validity of the assessment of pest density. In this system a remotely sensed image of the field is first used to identify the various phenological stages (insect habitats) and their GPS coordinates. This step guides the consultant in selecting representative GPS referenced areas in which to sample for insect pest densities which are used to answer the question, "Is there an insect problem that requires action or intervention in each of the phenological stages (insect habitats) in this field?" Using the properly classified image to determine where to sample and selecting the correct choice of sample unit can optimize sample size and sampling time for the consultant. When the system is used by an experienced consultant, a GPS referenced map of pest abundance over the entire field can be quickly and accurately developed. This map becomes the basis for a decision relative to the need for a variable rate application of a pesticide to optimize production in the field. This system greatly improves the consultant's pest management decisions, work time, and the grower's ability to optimize profits.
Technical Abstract: The accuracy of an estimate is in question when more than one population density actually exists in the field and the sampler is unaware of their existence. Multiple densities of the same species occur when the landscape is aportioned into different habitats and reproductive adults occupy each one at different rates of recruitment. Across the landscape, the various habitats differ temporally in their spatial interspersion pattern and vary in both shape and size. Therefore, when multiple population densities are sampled without consideration for the influence of habitat structure, the population mean that is estimated represents a summarization of haphazard processes and has less relevance for cotton insect pest management. Described in this paper are the primary concepts employed to link high resolution, multi-spectral imagery with several sampling concepts. This linkage when combined with consultant/producer experience and farm heuristics determines and maps the spatial abundance of several cotton insect pests. Using correctly sized sample units, delineating diverse habitats in fields by use of imagery, and detecting relevant mixtures of more than one statistical distribution (i.e., populations) in a collection of data are key elements of the methodology. The ability to minimize the observance of zeroes in samples and to segregate distributions representing different populations of insects (or pixel attributes) improves the sampling and management of cotton insect pests.