Exhaustive statistical processes to estimate pest densities


Exhaustive Statistical Processes to Estimate Pest Densities

The efficiency of the control of soybean pests is subject to the precision of the sampling process to estimate pest densities. Inadequate information about pest populations can result in crop loss or unnecessary chemical inputs, as well as higher costs of production and insect resistance. The objective of this project is developing effective sampling plans aiming to determine the densities of soybean insect pests such as Aphis glycines (Hemiptera: Aphididae). Statistical analyses will be exhaustive to take into consideration variability of the pest densities. Intensive resampling of field-collected data sets on pest densities via bootstrap sampling methods will allow use of the empirical distribution of the observed data rather than preassumed distributions. Bootstrap methods are necessary, because it is  not feasible to collect such large datasets from the field. These techniques will be useful to increase the power and scope of the inferences for pest management. The code will perform a simple random sampling with replacement in each data set starting with a sample size of one unit. It will keep increasing the sample size by one sample until the desired precision (d=0.10) is attained. For each sample size, mean and variance of the number of insects will be obtained. Each process is typically performed about a thousand times until the desired precision is achieved. This process is repeated 999 times for each dataset. The first analysis has fifteen datasets, which will result in around 14,985,000 processes to obtain results.


Group name: