Designing a network for monitoring groundwater level using the Principal Component Analysis technique

Well designed monitoring networks are essential for the effective management of groundwater resources but the costs of monitoring well installations and sampling can prove prohibitive. Principal Component Analysis (PCA) is one of the data reduction techniques used to extract the important components that explain the variance of a system. In this paper, the PCA was used to identify the effective wells and remove the less important ones. For this purpose, 160 wells were constructed in the Salman Farsi agro-industry, which was measured twice in a month during 10 months. In this technique, variation factors called principle components are identified with considering data structures. Using the PCA, the relative importance of each well was calculated for groundwater depth estimation. In the present study, the acceptable threshold was taken to be 0.8, according to which the number of wells in determining groundwater depth was reduced to 33 wells. By identifying important wells, important points for sampling are identified, and groundwater depth monitoring is performed only in these wells. As a result, it can save a great deal of time and cost of studies.