Yield loss prediction models based on early estimation of weed pressure

Ali Asif; Jens Carl Streibig; Christian Andreasen

doi:10.1016/j.cropro.2013.06.010

Yield loss prediction models based on early estimation of weed pressure

Ali Asif, Jens Carl Streibig, Christian Andreasen

Afgrødevidenskab

18 Citationer (Scopus)

Abstract

Weed control thresholds have been used to reduce costs and avoid unacceptable yield loss. Estimation of weed infestation has often been based on counts of weed plants per unit area or measurement of their relative leaf area index. Various linear, hyperbolic, and sigmoidal regression models have been proposed to predict yield loss, relative to yield in weed free environment from early measurements of weed infestation. The models are integrated in some weed management advisory systems. Generally, the recommendations from the advisory systems are applied to the whole field, but weed control thresholds are more relevant for site-specific weed management, because weeds are unevenly distributed in fields. Precision of prediction of yield loss is influenced by various factors such as locations, yield potential at the site, variation in competitive ability of mix stands of weed species and emergence time of weeds relative to crop. The aim of the review is to analyze various approaches to estimate infestation of weeds and the literature about yield loss prediction for multispecies. We discuss limitations of regression models and possible modifications to include the influential factors related to locations and species composition in context of their implementation in real time patch spraying.

Originalsprog	Engelsk
Tidsskrift	Crop Protection
Vol/bind	53
Sider (fra-til)	125-131
Antal sider	7
ISSN	0261-2194
DOI	https://doi.org/10.1016/j.cropro.2013.06.010
Status	Udgivet - nov. 2013

Adgang til dokumentet

10.1016/j.cropro.2013.06.010

Citationsformater

@article{fbb572b865b64e79a663c8213c9e16de,

title = "Yield loss prediction models based on early estimation of weed pressure",

abstract = "Weed control thresholds have been used to reduce costs and avoid unacceptable yield loss. Estimation of weed infestation has often been based on counts of weed plants per unit area or measurement of their relative leaf area index. Various linear, hyperbolic, and sigmoidal regression models have been proposed to predict yield loss, relative to yield in weed free environment from early measurements of weed infestation. The models are integrated in some weed management advisory systems. Generally, the recommendations from the advisory systems are applied to the whole field, but weed control thresholds are more relevant for site-specific weed management, because weeds are unevenly distributed in fields. Precision of prediction of yield loss is influenced by various factors such as locations, yield potential at the site, variation in competitive ability of mix stands of weed species and emergence time of weeds relative to crop. The aim of the review is to analyze various approaches to estimate infestation of weeds and the literature about yield loss prediction for multispecies. We discuss limitations of regression models and possible modifications to include the influential factors related to locations and species composition in context of their implementation in real time patch spraying.",

author = "Ali Asif and Streibig, {Jens Carl} and Christian Andreasen",

year = "2013",

month = nov,

doi = "10.1016/j.cropro.2013.06.010",

language = "English",

volume = "53",

pages = "125--131",

journal = "Crop Protection",

issn = "0261-2194",

publisher = "Elsevier",

}

TY - JOUR

T1 - Yield loss prediction models based on early estimation of weed pressure

AU - Asif, Ali

AU - Streibig, Jens Carl

AU - Andreasen, Christian

PY - 2013/11

Y1 - 2013/11

N2 - Weed control thresholds have been used to reduce costs and avoid unacceptable yield loss. Estimation of weed infestation has often been based on counts of weed plants per unit area or measurement of their relative leaf area index. Various linear, hyperbolic, and sigmoidal regression models have been proposed to predict yield loss, relative to yield in weed free environment from early measurements of weed infestation. The models are integrated in some weed management advisory systems. Generally, the recommendations from the advisory systems are applied to the whole field, but weed control thresholds are more relevant for site-specific weed management, because weeds are unevenly distributed in fields. Precision of prediction of yield loss is influenced by various factors such as locations, yield potential at the site, variation in competitive ability of mix stands of weed species and emergence time of weeds relative to crop. The aim of the review is to analyze various approaches to estimate infestation of weeds and the literature about yield loss prediction for multispecies. We discuss limitations of regression models and possible modifications to include the influential factors related to locations and species composition in context of their implementation in real time patch spraying.

AB - Weed control thresholds have been used to reduce costs and avoid unacceptable yield loss. Estimation of weed infestation has often been based on counts of weed plants per unit area or measurement of their relative leaf area index. Various linear, hyperbolic, and sigmoidal regression models have been proposed to predict yield loss, relative to yield in weed free environment from early measurements of weed infestation. The models are integrated in some weed management advisory systems. Generally, the recommendations from the advisory systems are applied to the whole field, but weed control thresholds are more relevant for site-specific weed management, because weeds are unevenly distributed in fields. Precision of prediction of yield loss is influenced by various factors such as locations, yield potential at the site, variation in competitive ability of mix stands of weed species and emergence time of weeds relative to crop. The aim of the review is to analyze various approaches to estimate infestation of weeds and the literature about yield loss prediction for multispecies. We discuss limitations of regression models and possible modifications to include the influential factors related to locations and species composition in context of their implementation in real time patch spraying.

U2 - 10.1016/j.cropro.2013.06.010

DO - 10.1016/j.cropro.2013.06.010

M3 - Journal article

SN - 0261-2194

VL - 53

SP - 125

EP - 131

JO - Crop Protection

JF - Crop Protection

ER -