To enhance disaster preparation for incidents that involve rainfall, we developed a system to support evacuation decision-making by people in charge of disaster prevention and by the general public. A reproduction calculation result for the 2009 Sayo event is presented herein. The risk of a flood disaster and a sediment disaster is estimated using calculations incorporating single rainfall data and ensemble calculations using multiple rainfall data.

For flood disasters, an inundation depth map, a risk map, and an optimal evacuation path are calculated. Inundation records and other methods used to validate the accuracy of the calculation results show that the inundation depth distribution accuracy is enhanced using 5 mDEM instead of 10 mDEM. For a sediment disaster, a warning level map, risk map, and risk index are calculated. Collapse records and other methods used to validate the calculation result accuracy demonstrate that the correlation between the calculated risk index and collapse record is enhanced by consideration of multiple risk factors, which include not only rainfall and the soil precipitation index but also the slope form, geology, and others. For a complex disaster, a combination of a flood disaster and a sediment disaster, a risk map, and optimal evacuation path are calculated to show evacuation guidelines for people in charge of disaster prevention and for the general public.

 Industrial water demand is increasing rapidly because of the growth of economically developing countries. Predictions of water demand are therefore necessary for effective management. One model proposed by Takahashi et al. (2000) predicts industrial water demand based on GDP data and a coefficient related to water use efficiency (η). This study specifically examines a method of determining η to improve predictive accuracy. A cluster analysis using the gross value added of industry super-sectors was completed to divide 61 countries into four clusters (“oil producer,” “developed or middle developed,” “middle developed or developing,” and “developing” countries). For each cluster, multi-regression analyses were conducted to elucidate the relation between η and some typical η determinants (e.g., GDP change ratio, inventory of water resources). Relations between past changes in η (Δη) and GDP_PPP-per-capita or the initial η were also studied. Results of multi-regression analyses show that determinants of η were significantly different among groups. Typical determinants explained η for countries in two groups: “oil producer” and “developed or middle developed” countries. Furthermore, countries with a GDPppp-per-capita higher than 14,000 $ present almost stable Δη/year (=0.00345), whereas countries presenting a lower GDP_PPP-per-capita have no tendency of Δη/year. Results show the possibility of predicting industrial water consumption more accurately by the prediction of η using the results described above.

 From experience with challenges and solutions during a technical cooperation project for capacity development in the field of flood management in the Philippines, some means of improving future technical cooperation projects in economically developing countries can be inferred. During the construction stage of boulder type spur dykes as a pilot project, spur dyke boulder diameters were indicated as 45 cm, on average, in the design specifications. Japanese experts expected that boulders of almost 45 cm diameter would be used for those spur dykes. However, both engineers in the government which ordered the construction work and engineers in the construction company in the Philippines recognized that from small 10-cm-diameter stones to large rocks with 1 m diameter could be used, if the average diameter were equal to or greater than 45 cm. Consequently, boulders were installed in the center of the spur dykes. Then smaller stones were arranged around those boulders to shape the spur dykes. Means of improving future technical cooperation projects related to capacity development in the field of flood management were suggested to solve such differences of recognition and raise the probability of alleviating such differences.