Kiyoshi Hoshi

Flood forecasting by the filter separation AR method and comparison with modeling efficiencies by some rainfall-runoff models

Abstract In this paper, the authors firstly propose a flood forecasting system with and without rainfall data, applying the filter separation AR method. The runoff time series are separated sequentially into two runoff components. Since each hydrologic subsystem is expressible by linear input-output relationships, the rainfall components are inversely estimated from the ARX model or from the response function type model. Future runoff is computed by the ARX or the response function type model utilizing the inversely estimated effective past rainfalls and the extrapolated future rainfalls as input, and, as for the shorter-period runoff, using also the past rainfall data. Secondly, we calculate and compare modeling efficiencies for four rainfall-runoff models. The four models are (1) the filter separation AR model, (2) and (3) two l inds of the generalized storage function model (Prasad model and Hoshi model) to which Karman filtering theory is applied, and (4) the tank model. The generalized storage function method contains the nonlinearity in the runoff equation itself and the single component nonlinear model. The filter separation AR method is composed of linear subsystems, and the nonlinearity of the total system is explained by the nonlinearity of the rainfall separation process into subsystems and the multicomponent model.

The impact of seasonal flow characteristics and demand patterns on required reservoir storage

Abstract Annual streamflow volumes, modeled by AR(1) and ARMA(1,1) processes, were generated and disaggregated to seasonal flow volumes and routed through a reservoir algorithm to determine the within-year capacity of a single reservoir needed to satisfy a specified demand sequence. Six seasons, each of two months duration, were used to investigate the impact of seasonal demand—seasonal flow variability coupling. Annual synthetic flows were disaggregated by a modified version of the disaggregation scheme which has been developed by Valencia and Schaake. Annual demand levels of 0.3, 0.5, and 0.7 of the mean annual flow were used. Three seasonal patterns representing relatively uniform demand, demand in phase with streamflow, and demand out of phase with streamflow were used. Modeling the skewed distributions of seasonal flows was shown to be important at all demand levels investigated.

Practical Issues and Their Resolutions in Flood Runoff Prediction

The present paper describes practical issues and their resolutions for increasing predictive capabilities of real-time flood forecasting methods. Special attention is directed to explicit incorporations of the water level-discharge (H-Q) relationship and rainfall measurements by radar into the filter prediction formulation.It is well recognized from hydrologic practices that measurement errors involved in the H-Q relationship are quite significant in flow and water stage forecastiong problems. The current approach proves extremely helpful in assessing the relative importance of modeling and measurement errors, and inaccuracy in rainfall forecasts for flood prediction applications.For the behavior of watershed response, the present study uses a nonlinear storage routing model converted from the kinematic wave equations where accurate linkages between the parameters for the two approaches are maintained. The proposed algorithms. were applied to the river stage forecasts of the 1988 Flood data in the Uryu River, which is a tributary of the Ishikari River.

Fuzzy Support System for Gate Operations in the Barato River Basin

The Barato River basin, located in the northern part of Sapporo City, is characterized by low-lying areas. The prevention and mitigation of flood damages in this basin are mainly controlled by the two gates; the gate of Shibi canal plays an important role in shutting off the backwater effect of the Ishikari River, while the gate of Ishikari Floodway draining out the inflow volumes from three upstream tributaries of the Barato River as fast as possible. The current gate operations in the Barato River system are heavily relied on intuition of the experienced engineers, mainly because decision-makings should be done on complicated predictions of inflows to the Barato River, water levels of Ishikari and Barato Rivers and the tide in a river mouth. The present study describes the basic concept of a Fuzzy support system applied to gate operations of the Barato River basin and some simulation results.