Submitted to: Weed Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/1/1998
Publication Date: N/A
Interpretive Summary: In the United States, it has been proposed that by the year 2000,Integrated Pest Management (IPM) will be required on 75% of our cropland and agchemicals will be reduced by two-thirds. One method of reducing herbicide use and improving environmental quality is to use computer decision-aid tools for purposes of weed management in crop production systems. Several years ago a computer weed/crop bioeconomic model called PALWEED:WHEAT was developed for the winter wheat producing areas of the Pacific Northwest (PNW). History has shown that very few first generation bioeconomic models developed have actually been field-tested for sensitivity, accuracy, and revision. PALWEED:WHEAT was therefore field tested for 2-yrs. The new model, PALWEED:WHEAT II differed from the first model in several ways including using separate indices for broadleaf and grass weeds rather than an aggregate competition index for the yield function and using separate models for wheat depending on crop rotation. Herbicide recommendations were considerably more balanced and realistic with the revised model. Models transferred to growers without field testing may be very risky. Based on our field tests and grower interactions, we are confident that PALWEED:WHEAT II is better suited to the PNW than the first generation model.
Technical Abstract: PALWEED:WHEAT is a bioeconomic model for determining profit-maximizing postemergence herbicide treatments for winter wheat in the Washington-Idaho Palouse region. PALWEED:WHEAT performed relatively well economically in 2 yr of on-farm field tests. However, the model was less sensitive than desired in prescribing postemergence broadleaved herbicides in the presence eof high densities of broadleaved weed seedlings. Therefore, PALWEED:WHEAT was revised in response to the field testing. This paper compares the revised model's agronomic and economic performance to the original model in computer simulations. The revised model, PALWEED:WHEAT II, differs from the original model in several respects: (1) exponential functions replace linear functions in predicting weed survival, (2) preplant application of a nonselective herbicide is entered as an exogenous binary variable, (3) separate indices of broadleaved and grass competition are substituted for an aggregate weed competition index in the wheat yield function, (4) hyperbolic replaces logistic functional representation of weed damage to wheat yield, and (5) separate models are estimated for winter wheat after spring dry pea and for winter wheat in all examined crop rotations positions. In simulations including a variety of agronomic and economic conditions, PALWEED:WHEAT II recommended postemergence herbicide types and rates that consistently complied with agronomic and economic theory. Furthermore, the revised model, especially when estimated from the relevant wheat after pea data set, was markedly more balanced in recommending both broadleaved and grass herbicides in response to observed densities of both weed groups.