Applied Statistical Methods on Price Data and Analyzing the Effects of Environmental and Chemical Factors For Aphanizomenon Flos-aquae Density Thesis uri icon



  • Thesis (M.S., Statistical Sciences)--University of Idaho, June 2014 | This work is divided into two research projects: (i) Comparison of the price reports for lumber products between Random Lengths and Crows' in North American softwood market; (ii) Analyzing the effects of environmental and chemical factors for Aphanizomenon flos-aquae density in Willow Creek, Oregon. When compared the price reports for lumber products, the parametric tests for different means were not proper methods. This was due to the time dependent of the data which the observations were conducted weekly over 20 years. The paired comparison of price differences changing over time showed how much difference between Random Lengths and Crows' price reports. Based on the results of autocorrelation function and binomial test, it showed that the prices between Random Lengths and Crows' had a mean difference of less than $1 and standard error around 0.15 for the 12 different products of which most southern yellow pine and some other products; these differences were stationary time series. Therefore, this was the average difference between Random Lengths and Crows'. For Douglas Fir Green 8' 2X4 Std&Btr-US and Douglas Fir Green 8' 2X6 #2&Btr-US, the difference means were $26 and $9 and the standard error were 0.027 and 0.026, respectively, and they were strong non-stationary time series. Thus, the difference between the price reports and the real market prices should be mentioned when study the lumber product prices in North America by using third-party price data. Analysis of Aphanizomenon flos-aquae density in the Willow Creek, Oregon area was different from other studies targeting of blue-green algae. Modeling a single species rather than the whole biomass was more difficult because the dominant species strongly influenced the model. If Aphanizomenon flos-aquae was not the dominant species, the accuracy of the model would be reduced. The reactive phosphorus may give a negative effect on Aphanizomenon flos-aquae density, but positive effect on whole biomass density. When modelling the Aphanizomenon flos-aquae density, it was necessary to select the observations which Aphanizomenon flos-aquae density higher than 200 counts per liter. If it was lower than 200 counts per liter, it turned to poor model-fit. Moreover, lower density of Aphanizomenon flos-aquae will not cause toxin problems of public health. Model with Expectation Maximization (EM) algorithm showed that nitrite, reactive phosphorus, pH, and temperature had highly significant effects on Aphanizomenon flos-aquae density, DO however, was less significant.

publication date

  • June 1, 2014