La simulazione annealing per la pianificazione del campionamento: applicazione in tre areali

Sampling scheme is the major factor influencing the efficiency and costs of a survey. Moreover, in designing sampling scheme, earlier observations and knowledge on the area can provide valuable information. The aim of this study was to describe a method to optimize the spatial sampling scheme, taking physics constraints and preliminary information into account. The method was based upon a spatial simulated annealing algorithm. Spatial sampling schemes can be optimised for minimal Kriging variance or distance between observations. Three case studies are presented, in different pedoclimatic conditions (mountain, plain, hill), to locate 100 samples. In Val Chiavenna (mountain, Sondrio, Lombardy) minimal distance between observations criterion was used. Lithology, aerial photographs and two pedological profiles were the preliminary information. In Lodi (plain, Lombardy) preliminary information was soil use and 158 soil samples. The optimization process was carried out in two steps: 50 samples were located with the minimal distance between observations criterion, while additional 50 samples were located with the variance Kriging criterion. In Mustigarufi (hill, Caltanissetta, Sicily), 50 samples were located with the minimal distance between observations criterion without taking preliminary information into account. Additional 50 samples were located with the minimal distance between observations criterion and soil electrical conductivity variability, 6 pedological profiles and the earlier 50 observations as preliminary information.