Flood Forecasting Model Using Multi-Source Rainfall Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional flood forecasting methods have limited lead times due to poor timeliness of rainfall data, low accuracy, and focus on hydrological simulation rather than forecasting, with evaluation criteria only suitable for single lead times, failing to effectively predict floods at multiple lead times.
Innovation Solution
A method that combines rainfall data from radars, gauges, and precipitation micro-physical characteristics sensors to obtain hydrological variables, estimates mean concentration time, and uses an optimized hydrological model with a specific objective function to predict streamflow at multiple lead times, enhancing forecasting accuracy and providing early warnings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If conventional methods use real-time observed rainfall data from gauges, then the data is available, but the lead time of forecast is relatively short due to poor timeliness
Solution Approach 1:
The patent applies preliminary action by using radar and satellite data to predict future rainfall before it occurs. The system generates predicted rainfall data for future time periods, allowing flood forecasts to be made in advance rather than waiting for actual rainfall measurements to be collected and transmitted from gauges.
2Loss of time
If conventional methods use weather-predicted rainfall data, then the lead time is extended, but the forecasting accuracy is relatively low due to uncertainty of weather-predicted rainfall
Solution Approach 1:
The patent applies feedback by using an objective function that compares predicted streamflow with observed streamflow data. The model parameters are iteratively adjusted to minimize the difference between predicted and observed values, thereby improving forecasting accuracy while maintaining extended lead times.
Solution Approach 2:
The patent applies parameter changes by optimizing hydrological model parameters through calibration against observed data. The objective function modifies model parameters to better match historical observations, reducing uncertainty in weather-predicted rainfall and improving overall forecasting accuracy.
3Reliability
If conventional methods focus on hydrological simulation rather than forecasting, then the rain-runoff relationship is modeled, but the forecasting ability is limited
Solution Approach 1:
The patent applies inversion by reversing the conventional approach: instead of using simulation to understand rain-runoff relationships and then attempting to forecast, the system directly optimizes the model for forecasting performance. The objective function is specifically designed to maximize forecasting accuracy by minimizing errors between predicted and observed streamflow, rather than merely simulating hydrological processes.
4Device complexity
If conventional evaluation criteria focus on forecasted result at a single lead time, then the evaluation is simple, but it is unable to simultaneously describe the forecasted results at multiple lead times
Solution Approach 1:
The patent applies multi-functionality by designing an objective function that simultaneously evaluates forecasting performance across multiple lead times. The function aggregates errors from different forecast horizons into a single optimization target, allowing the model to be calibrated for multiple forecast periods concurrently rather than requiring separate evaluations for each lead time.
Data Source
AI summary
A method for forecasting flood, the method including: calibrating the hydrological model by using an objective function that is a sum of squared difference between the observed streamflow and the corresponding forecasted streamflow at each lead time to obtain the optimized hydrological model; using the optimized hydrological model to forecast floods; and evaluating forecasting performance of the optimized hydrological model. The method improves the forecasting accuracy and provides forecasting results at various lead times.