Page Banner

United States Department of Agriculture

Agricultural Research Service

Title: Evaluating Field-Scale Sampling Methods for the Estimation of Mean Plant Densities of Weeds

Authors
item Colbach, N - INRA, FRANCE
item Dessaint, F - INRA, FRANCE
item Forcella, Frank

Submitted to: Weed Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: October 25, 2000
Publication Date: N/A

Interpretive Summary: Crop advisors often scout fields to determine levels of weed infestations. Many fields may be scouted in a single day, so scouting methods must be fast and efficient. Traditionally, scouts only make cursory observations, however as computerized weed management decision aids become more realistic, scouting may have to become more quantitative. Accordingly, methods of weed assessment will be more time-consuming. Our objective, therefore, was to compare different scouting schemes, from simple to complex, for reliability in terms of assessing weed vegetation. In general, simple systematic sampling patterns resulted in the lowest error (e.g., observations made along a diagonal or "zigzag" across a field). Use of maps based upon last year's weed populations resulted in the highest error. Tedious random sampling had intermediate error. Specification of a minimum distance between sampling points also lessened sampling error. These results indicate that quantitative estimates of weed populations can be obtained using very simple surveying techniques. This is a very important guideline for crop consultants who must scout for weed problems in a timely and accurate manner.

Technical Abstract: Weed vegetation (Amaranthus retroflexus L.; Asclepias syriaca L.; Chenopodium album L.; Cirsium arvense (L.); SCOP.; Elytrigia repens (L.) Nevski; Setaria viridis (L.) P. Beauv.; Sinapis arvensis L.) of fields continuously grown with soybean (Glycine max (L.) Merr.) was simulated for four years, using semivariograms established from previous field observations. Various sampling methods were applied and compared for accurately estimating mean plant densities, differing weed species and years. The tested methods were based on: a) random selection, where samples were chosen either entirely randomly, randomly with at least 10 or 20 m between samples, or randomly after stratifying the field; b) systematic selection, where samples were placed along diagonals or zigzag lines across the field c) predicted S. viridis seedling maps, which were used to divide the field into low- and high-density areas and to choose the largest sample eproportion in the high-density area. For each method, sampling was performed with 5 to 40 samples. A ranking of methods, independent of the sampled species, was established. Systematic methods generally resulted in the lowest estimation error, followed by the random methods, and finally by the predicted-map methods. In cases of species over- or under-represented along the diagonals or the zigzag sampling line, the systematic methods performed badly, especially with low sample numbers. In those instances, random methods were best, especially those imposing a minimal distance between samples. Even for S. viridis, the methods based on predicted S. viridis maps were not satisfactory, except with low sample numbers. The relationships between sampling error and species characteristics (mean density, variability, spatial structures) also were studied.

Last Modified: 4/24/2014
Footer Content Back to Top of Page