Validation of predictive empirical weed emergence models of Abutilon theophrasti Medik based on intercontinental data

Authors: EGEA-COBRERO, V - Institute For Sustainable Agriculture; BRADLEY, K - University Of Missouri; CALHA, I - Instituto Nacional De Investigação Agrária E Veterinária; Davis, Adam; DORADO, J - Institute For Sustainable Agriculture; Forcella, Frank; LINDQUIST, J - University Of Nebraska; SPRAGUE, C - Michigan State University; GONZALEZ-ANDUJAR, J - Institute For Sustainable Agriculture

Submitted to: Weed Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/24/2020
Publication Date: 5/19/2020
Citation: Egea-Cobrero, V., Bradley, K., Calha, I., Davis, A.S., Dorado, J., Forcella, F., Lindquist, J.L., Sprague, C.L., Gonzalez-Andujar, J.L. 2020. Validation of predictive empirical weed emergence models of Abutilon theophrasti Medik based on intercontinental data. Weed Research. 60(4):297-302. https://doi.org/10.1111/wre.12428.
DOI: https://doi.org/10.1111/wre.12428

Interpretive Summary: Successful weed management relies on the proper timing of weed control practices in relation to the timing of weed seedling emergence. Therefore, the development of models that predict the weed emergence timing may help provide growers with tools to make better weed management decisions. The aim of this study was to compare the performance of two previously published models of seedling emergence in relation to accumulated thermal time for the weed velvetleaf (Abutilon theophrasti) in corn with data sets from the USA and Europe. We tested the hypothesis that a robust and general weed emergence model can be developed for this species. Our results indicated that predictions made with a more flexible model (Weibull function) were more accurate than those made with a less flexible model (Logistic function). However, both models showed appreciable bias, demonstrating the need for a more comprehensive experimental understanding of the environmental factors contributing to variability in weed seedling emergence across different locations.

Technical Abstract: Good weed management relies on the proper timing of weed control practices in relation to weed emergence dynamics. Therefore, the development of models that predict the timing of emergence may help provide growers with tools to make better weed management decisions. The aim of this study was to validate and compare two previously published predictive empirical thermal time models of the emergence of Abutilon theophrasti growing in maize with data sets from the USA and Europe, and test the hypothesis that a robust and general weed emergence model can be developed for this species. Previously developed Weibull and Logistic models were validated against new data sets collected from 11 site-years, using four measures of validation. Our results indicated that predictions made with the Weibull model were more reliable than those made with the Logistic model. However, Weibull model results still contained appreciable biases that prevent its use as a general model of A. theophrasti emergence. Our findings highlight the need to develop more accurate models if the ultimate goal is to make more precise predictions of weed seedling emergence globally to provide growers with universally consistent tools to make better weed management decisions.