Abstract

Previous research by the authors has established that southern root-knot nematode (SRKN, Meloidogyne incognita (Kofoid & White) Chitwood) and yellow and purple nutsedge (YNS, Cyperus esculentus L. and PNS, C. rotundus L.) form a pest-complex that adversely affects a wide variety of crops in the southern and western U.S. These pests appear to have co-evolved a mutually-beneficial relationship that promotes the survival of both nematodes and weeds to the detriment of crops. Traditional management has usually targeted one pest at a time, but managing this pest complex requires that all members of the complex be managed simultaneously. A series of experiments was performed to determine if this specific pest complex could be managed through crop-rotation using a non-dormant M. incognita-resistant variety of alfalfa (Medicago sativa) which can aggressively compete with, and hence decrease, occurrence of both species of nutsedges (NS), and subsequently decrease SRKN by decreasing the availability of root systems of host plants. A previous journal article discussed predicting counts of SRKN second-stage juveniles (SRKN-J2) as a function of YNS and PNS plant counts from a two-year alfalfa rotation experiment, using the Poisson distribution and a scale parameter to handle problems of overdispersion. In this paper, we examine three generalizations of the Poisson distribution that allow for the count variance being larger than the mean count: the Generalized Poisson, the Zero-Inflated Poisson (ZIP), and the Poisson Hurdle. The ZIP and Hurdle Poisson distributions both account for zero counts as a separate part of the distribution, while the Generalized Poisson incorporates a separate parameter that increases the variance relative to the mean. Different biological scenarios are presented for which each of these three general Poisson distributions might be logically appropriate. In addition, we use the alfalfa rotation data to present comparisons of fitted regression models of the three general Poisson distributions to the results from the previous analysis using the Poisson. For this data, there was no single probability distribution that worked best for all six sampling dates (three in each of the two years). This is not surprising in that over time the alfalfa rotation was, as planned, decreasing both nutsedge and nematode counts, thus presenting a "moving target" for the modeling process.

Keywords

Nematodes; Nutsedge; Poisson; Generalized Poisson; Zero-Inflated Poisson; Poisson Hurdle; Overdispersion

Creative Commons License


This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Share

COinS
 
May 1st, 11:00 AM

MODELING THE ROOT-KNOT NEMATODE/NUTSEDGE PEST COMPLEX: PERSPECTIVES FROM WEED SCIENCE, NEMATOLOGY AND STATISTICS

Previous research by the authors has established that southern root-knot nematode (SRKN, Meloidogyne incognita (Kofoid & White) Chitwood) and yellow and purple nutsedge (YNS, Cyperus esculentus L. and PNS, C. rotundus L.) form a pest-complex that adversely affects a wide variety of crops in the southern and western U.S. These pests appear to have co-evolved a mutually-beneficial relationship that promotes the survival of both nematodes and weeds to the detriment of crops. Traditional management has usually targeted one pest at a time, but managing this pest complex requires that all members of the complex be managed simultaneously. A series of experiments was performed to determine if this specific pest complex could be managed through crop-rotation using a non-dormant M. incognita-resistant variety of alfalfa (Medicago sativa) which can aggressively compete with, and hence decrease, occurrence of both species of nutsedges (NS), and subsequently decrease SRKN by decreasing the availability of root systems of host plants. A previous journal article discussed predicting counts of SRKN second-stage juveniles (SRKN-J2) as a function of YNS and PNS plant counts from a two-year alfalfa rotation experiment, using the Poisson distribution and a scale parameter to handle problems of overdispersion. In this paper, we examine three generalizations of the Poisson distribution that allow for the count variance being larger than the mean count: the Generalized Poisson, the Zero-Inflated Poisson (ZIP), and the Poisson Hurdle. The ZIP and Hurdle Poisson distributions both account for zero counts as a separate part of the distribution, while the Generalized Poisson incorporates a separate parameter that increases the variance relative to the mean. Different biological scenarios are presented for which each of these three general Poisson distributions might be logically appropriate. In addition, we use the alfalfa rotation data to present comparisons of fitted regression models of the three general Poisson distributions to the results from the previous analysis using the Poisson. For this data, there was no single probability distribution that worked best for all six sampling dates (three in each of the two years). This is not surprising in that over time the alfalfa rotation was, as planned, decreasing both nutsedge and nematode counts, thus presenting a "moving target" for the modeling process.