Rainfall estimation in the Chikugo River Basin by Atmospheric downscaling using artificial networks

Izumi Ishikawa, Jonas Olsson, Kenji Jinno, Akira Kawamura, Koji Nishiyama, Ronny Berndtsson

Research output: Contribution to journalArticlepeer-review


For the proper water resources management of the Chikugo River basin, the prediction of both drought and heavy rainfall needs to be carried out by the conventional and engineering method which can be useful to for the practitioners who work on the water resources management and flood control. A relatively simple and efficient way to estimate local and regional rainfall, as well as other hydrometeorological variables, is now intensively discussed. This method utilizes the grid data point value (GPV) to predict the regional rainfall based on the so called atmospheric downscaling. In this paper, artificial neural networks (ANNs) are employed. As the input variables, three large-scale meteorological variables, precipitable water, and zonal and meridional wind speeds, are used. Output is the mean rainfall intensity in the Chikugo River basin during a 12-hour period. In the model, the serially combined ANNs were employed to predict the rainfall amount exactly. The result from the serially combined ANNs is slightly better than the result from the neumerical weather prediction model of the Japan Meteorological Agency by comparing the values of CC and RMSE.

Original languageEnglish
Pages (from-to)85-96
Number of pages12
JournalMemoirs of the Faculty of Engineering, Kyushu University
Issue number2
Publication statusPublished - Jun 2002
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • General Energy
  • Atmospheric Science
  • General Earth and Planetary Sciences
  • Management of Technology and Innovation


Dive into the research topics of 'Rainfall estimation in the Chikugo River Basin by Atmospheric downscaling using artificial networks'. Together they form a unique fingerprint.

Cite this