Relative abundance, spatial distribution and sequential sampling for four main predators of cereal aphids in winter wheat fields of Gorgan, northern Iran

In order to investigate seasonal abundance and spatial distribution of the main predators of cereal aphids and develop a fixed-precision sequential sampling plan, a weekly sweepnet sampling was conducted at five winter wheat fields in Gorgan region in northern Iran, during two growing seasons, 2006 and 2007. Spatial distribution of the predators was described by fitting data to Poisson (random) distribution, as well as by calculating three dispersion indices. A sequential sampling plan was also developed using the fixed-precision method of Green for estimating the mean abundance of predators' population. Sitobion avenae Fabricius and Schizaphis graminum Rondani were the dominant aphid species and four species of Coccinella septempunctata L., Propylea quatuordecimpunctata L., Eupeodes corollae (Fabricius) and Episyrphus balteatus (De Geer) comprised 18.75, 14.25, 41.54 and 6.02 percent of the predators' community, respectively. For all predators, with the exception of E. corollae, the parameters b of Taylor’s power law did not differ significantly from one, indicating that populations of them exhibited random spatial distribution. Fitting population’s frequency data to distribution models also showed that Poisson (random) distribution provided a good fit to the population frequencies during most of the wheat growing season. Sequential sampling results showed that the number of sample units required to stop sampling was depended upon mean population abundance and desired level of precision. So that, for C. septempunctata, at abundance of 0.033-0.6 adult/10 sweepnet and precision level of 0.25, the required sample size ranged from 400 to 41 sweepnets. At the same time, for P. quatuordecimpunctata, at abundance of 0.02-1.8 adult/10 sweepnets, the required sample size ranged from 775 to 10 sweepnets. For E. corollae and E. balteatus, required sample size ranged from 4 to 400 and 57 to 175 sweepnets, respectively. Regarding number of sample units, at high abundance of predators, sweepnet sampling was a cost-effective method to estimate predators' population, whereas at low population levels, sweepnet sampling was a time-consuming method and a quite large sample was required to achieve the desired precision of 0.25. Therefore, we recommend the comparison of precision and efficiency of sweepnet with other sampling procedures to determine the best sampling method for estimating abundance of predators in wheat fields.