Analyzing Inflow Trend of Indiana Reservoirs Using SOM

作者：Song Alfie Ningyu; Chandramouli V^*; Gupta Nimisha

来源：Journal of Hydrologic Engineering, 2012, 17(8): 880-887.

DOI：10.1061/(ASCE)HE.1943-5584.0000517

• Reservoirs
• Climate change
• Trend analysis
• Artificial intelligence-self-organizing maps
• Cluster analysis

Inflows to reservoir systems are affected by climatic changes. In the past, regional inflow trend analyses were conducted using statistical approaches. This research made use of an artificial intelligence technique called the self-organizing map (SOM) to perform trend and cluster analysis for the inflows into the flood-control reservoirs of Indiana. Along with SOM, this research also used the Mann-Kendall test and a revised Mann-Kendall test for regional analysis. Results indicate an increasing trend in the clusters that represent days with high inflows to the northern reservoirs of Indiana when inflows to the central and southern reservoirs were low or medium. A 7% increase was noticed in the annual daily counts belonging to this cluster during the past 20 years. Similar trends were observed concerning high-inflow days to the central reservoirs of Indiana. However, they are not statistically significant at a 95% confidence level. This study concludes that SOM is a useful tool for studying the trends at a regional level. DOI:10.1061/(ASCE)HE.1943-5584.0000517.

• 出版日期2012-8