effect of different well shape factor determination methods on saturated hydraulic conductivity results in guelph permeameter analysis

saturated hydraulic conductivity (ks) is one of the most important physical features of soil that have wide usage in soil and water science. measuring ks carries on laboratory and field. guelph permeameter (gp) method is one of the field measuring method for determining field saturated hydraulic conductivity (kfs) above the ground water table. well-shape-factor (c) is one important parameter needed to calculate kfs, which related to pressure gradient of flow through the soil around the borehole. several approximated analytical solution are presented to determine c, such as porchet solution, glover solution, half-source (reynolds et. al.) solution. numerical solution is obtained using finite difference by reynolds et. al. (1985) and applied formulae fitted them by zhang et. al. (1998). developed theory of flow out a borehole corresponds the flow to hydraulic conductivity (kfs) and soil matrix potential (?m) which can be obtained by measuring outflow in two different ponding depth in single hole .one of the restrictions of this method is the negative results that may be seen in some cases. in this research, the gp experiments were down in 18 holes of 3 cm radius and 60 cm depth in a loamy soil with different ponding head (h), (maximum and minimum h was 5 and 20 cm, respectively). then the field saturated hydraulic conductivity (kfs) was determined by two head analysis of gp, richards, laplace and regression analysis with different solution methods for well shape factor (c). results showed that if the porchet method was used for c factor calculation, negative results have been minimized and other methods for the c calculation (glover, numerical and half-source) have been produced low to high negative results for (kfs), respectivity. according to this study, the increasing of ratio of h2 to h1 will be caused a large number of (kfs) values to be positive and also there were no significant differences between their means. in the richards, laplace and regression analyses, the glover, numerical and half-source solution of c had the least deviation of mean, respectively. analyses of regression model with numerical solution of c has the nearest results with two head analysis.