Abstract
In this study, the fluctuation patterns of the amount of groundwater pumped over a 10 year period the period was investigated using Self-Organizing Maps (SOM), concentrating on 28 local cities in the Tama region of Tokyo Metropolis. The SOM classified the long-term fluctuation patterns of the groundwater pumped into six clusters showing clearly distinguishable patterns, and were then divided into four large groups. The results show that the fluctuation patterns of the decrease in groundwater pumped were classified into Group-A1 and A2. Most of the regular fluctuation patterns caused by a large amount of groundwater pumping were classified into Group-B1 and B2. Furthermore, the majority of the Group-B1 and B2 classifications were moved to Group-A1 and A2 due to a decrease in drinking water being pumped after being replaced by river (surface) water. However, some patterns remained in Group-B1 becaus of a major increase in water pumped for factories. These results show that the SOM analysis was successful in extracting the long-term fluctuations of groundwater pumped.