Artificial intelligence (AI) is a state-of-the-art technology and has nowadays become highly popu... more Artificial intelligence (AI) is a state-of-the-art technology and has nowadays become highly popular in scientific and technological fields. AI possesses great capability in handling mass information and formulating intelligent algorithms with human-like logical inference through learning messages and storing knowledge from input information. AI has been applied with great success to water resources management in Taiwan. This study aims to systematically present the development and achievements of AI techniques on integrated water sources management and hydro-informatics with respect to diversified domains including hydrology, engineering, environment, eco-hydrology and hydro-meteorology in Taiwan. The continual integration of AI techniques (neural networks, fuzzy inference, genetic algorithms) with domain-driven technologies from hydrological, water resources, eco-environmental and informatics engineering fields will be our future mission, which is dedicated to the development of advanced intelligent techniques on hydro-related systems/ platforms. The creation of a new era on water resources management and sustainable eco-environment with AI is an everlasting goal for us all to pursue.
An Investigation of the Impact from Different Rainfall Sources on Flash Flood Prediction using Artificial Neural Networks
ABSTRACT An artificial neural network (ANN) based rainfall-runoff model was developed to forecast... more ABSTRACT An artificial neural network (ANN) based rainfall-runoff model was developed to forecast flash floods by using satellite-derived forcing data during typhoon periods over Keelung watershed, located on northern Taiwan. The satellite rainfall estimates over Taiwan are generated through PERSIANN CCS at grid of 4 km and temporal resolution of half hour. Validation of satellite estimates with gauge measurements shows the PERSIANN CCS captures the heavy rainfall in terms of trend and peak volume but slightly underestimates the light rainfall, in particular the initial stage of the storm events. The major goal of this study is to investigate the impact of rainfall forcing data from different sources on flood forecasting. First, an ANN flood prediction model was calibrated by using datasets of 13 historical gauge-streamflow events. Second, 6 latest flood events were used to investigate the flood prediction results driven by satellite-derived rainfall and gauge observations, respectively. Finally, the realization of uncertainty quantification was examined by propagating satellite-derived precipitation ensemble into the flood prediction model. Our results exemplify the need for a better representation of satellite-derived precipitation error structure and a detailed investigation of the techniques that propagate the input error into hydrological models.
AI techniques for optimizing multi-objective reservoir operation upon human and riverine ecosystem demands
Journal of Hydrology, 2015
Summary Flow regime is the key driver of the riverine ecology. This study proposes a novel hybrid... more Summary Flow regime is the key driver of the riverine ecology. This study proposes a novel hybrid methodology based on artificial intelligence (AI) techniques for quantifying riverine ecosystems requirements and delivering suitable flow regimes that sustain river and floodplain ecology through optimizing reservoir operation. This approach addresses issues to better fit riverine ecosystem requirements with existing human demands. We first explored and characterized the relationship between flow regimes and fish communities through a hybrid artificial neural network (ANN). Then the non-dominated sorting genetic algorithm II (NSGA-II) was established for river flow management over the Shihmen Reservoir in northern Taiwan. The ecosystem requirement took the form of maximizing fish diversity, which could be estimated by the hybrid ANN. The human requirement was to provide a higher satisfaction degree of water supply. The results demonstrated that the proposed methodology could offer a number of diversified alternative strategies for reservoir operation and improve reservoir operational strategies producing downstream flows that could meet both human and ecosystem needs. Applications that make this methodology attractive to water resources managers benefit from the wide spread of Pareto-front (optimal) solutions allowing decision makers to easily determine the best compromise through the trade-off between reservoir operational strategies for human and ecosystem needs.
Agriculture is extremely vulnerable to climate change. Greenhouse farming is recognized as a prom... more Agriculture is extremely vulnerable to climate change. Greenhouse farming is recognized as a promising measure against climate change. Nevertheless, greenhouse farming frequently encounters environmental adversity, especially greenhouses built to protect against typhoons. Short-term microclimate prediction is challenging because meteorological variables are strongly interconnected and change rapidly. Therefore, this study proposes a water-centric smart microclimate-control system (SMCS) that fuses system dynamics and machine-learning techniques in consideration of the internal hydro-meteorological process to regulate the greenhouse micro-environment within the canopy for environmental cooling with improved resource-use efficiency. SMCS was assessed by in situ data collected from a tomato greenhouse in Taiwan. The results demonstrate that the proposed SMCS could save 66.8% of water and energy (electricity) used for early spraying during the entire cultivation period compared to the t...
Taiwan is located in themonsoon zone of the North Pacific Ocean and experiences an average of 4-5... more Taiwan is located in themonsoon zone of the North Pacific Ocean and experiences an average of 4-5 typhoons annually. The particular topography of Taiwan makes rivers short and steep, and thus rivers flow rapidly from catchments to reservoirs within a few hours during typhoon events. This study aims to construct realtime multi-step-ahead reservoir inflow forecast models by using Artificial Neural Networks (ANNs) based on radar rainfall data and reservoir inflow data. The Back PropagationNeural Network (BPNN) and the Recurrent Neural Network (RNN) are adopted for forecasting. Results indicate that the correlation coefficients in the testing phases of both models exceed 0.86 for one- to three-hour-ahead forecasts and exceed 0.69 for six-hour-head forecasts. The RNN model outperforms the BPNN model, which indicates the recurrent property of the RNN can effectively improve forecast accuracy when making several step-ahead forecasts. Results demonstrates that the constructed multi-step-ahe...
In the face of increasingly flood disasters, on-line regional flood inundation forecasting in urb... more In the face of increasingly flood disasters, on-line regional flood inundation forecasting in urban areas is vital for city flood management, while it remains a significant challenge because of the complex interactions and disruptions associated with highly uncertain hydro-meteorological variables and the lack of high-resolution hydro-geomorphological data. Effective on-line flood forecasting models through the rapid dissemination of inundation information regarding threatened areas deserve to develop appropriate technologies for early warning and disaster prevention. Artificial Intelligence (AI) becomes one of the popular techniques in the study of flood forecasts in the last decades. We apply the AI techniques with the newly implemented IoT-based real-time monitoring flood depth data to build an urban AI flood warning system. The AI system integrates the self-organizing feature mapping networks (SOM) with the recurrent nonlinear autoregressive with exogenous inputs network (R-NARX) for modelling the regional flooding prediction. The proposed AI model with the IoT-based real-time monitoring flood depth datasets can increase the value-added application of diversified disaster prevention information and improve the accuracy of flood forecasting. We develop an on-line correction algorithm for continuously learning and correcting model's parameters, automatic operation modules, forecast results output modules, and web page display interface. The proposed AI system can provide the smart early flooding warnings in the urban area and help the Water Resources Agency to promote the intelligent water disaster prevention services.
Rapid urbanization in metropolitan areas easily triggers flashy floods. Urban drainage systems co... more Rapid urbanization in metropolitan areas easily triggers flashy floods. Urban drainage systems conveying stormwater out of cities are key infrastructure elements for flood mitigation. This study develops an intelligent urban flood drainage system accounting for carryover storage through optimizing the multi-objective operation rules of pumping stations for effectual flood management in Taipei City. The Yu-Cheng pumping station constitutes the study case, and a large number of datasets collected from 17 typhoon/storm events are adopted for model construction and validation. Three objective functions are designed to minimize: (1) the sum of water level fluctuations in the flood storage pond (FSP); (2) the sum of peak FSP water levels; and (3) the mean absolute difference of pump switches between two consecutive times along operation sequence. The non-dominated sorting genetic 2 algorithm II (NSGA-II) is applied to searching the Pareto-optimal solutions that optimize the trade-off between the objectives. We next formulate the optimal operation rules through a two-tier sorting process based on a compromised Paretooptimal solution. The comparison of the simulated results obtained from both the optimal operation rules and current operation rules indicate that the optimal operation rules outperform current operation rules for all three objectives, with improvement rates reaching 43% (OBJ1), 3% (OBJ2) and 71% (OBJ3), respectively. We demonstrate that the derived intelligent urban flood drainage system can serve as reliable and efficient operational strategies for urban flood management and flood risk mitigation.
Knowledge-Based Information Systems in Practice, 2015
Taiwan is located in the monsoon zone of the North Pacific Ocean and experiences an average of 4-... more Taiwan is located in the monsoon zone of the North Pacific Ocean and experiences an average of 4-5 typhoons annually. The particular topography of Taiwan makes rivers short and steep, and thus rivers flow rapidly from catchments to reservoirs within a few hours during typhoon events. This study aims to construct realtime multi-step-ahead reservoir inflow forecast models by using Artificial Neural Networks (ANNs) based on radar rainfall data and reservoir inflow data. The Back Propagation Neural Network (BPNN) and the Recurrent Neural Network (RNN) are adopted for forecasting. Results indicate that the correlation coefficients in the testing phases of both models exceed 0.86 for one-to three-hour-ahead forecasts and exceed 0.69 for six-hour-head forecasts. The RNN model outperforms the BPNN model, which indicates the recurrent property of the RNN can effectively improve forecast accuracy when making several step-ahead forecasts. Results demonstrates that the constructed multi-step-ahead rainfall-runoff models can provide valuable instantaneous inflow forecasts for the coming six hours so that decision makers can implement more suitable reservoir operations in consideration of inflow forecasts, rather than just depend on historical scenarios.
Hydrology and Earth System Sciences Discussions, 2013
The accurate stream flow composition simulated by different models is rarely discussed, and few s... more The accurate stream flow composition simulated by different models is rarely discussed, and few studies addressed the model behaviors affected by the model structures. This study compared the simulated stream flow composition derived from two models, namely HBV and TOPMODEL. A total of 23 storms with a wide rainfall spectrum were utilized and independent geochemical data (to derive the stream composition using end-member mixing analysis, EMMA) were introduced. Results showed that both hydrological models generally perform stream discharge satisfactory in terms of the Nash efficiency coefficient, correlation coefficient, and discharge volume. However, the three simulated flows (surface flow, interflow, and base flow) derived from the two models were different with the change of storm intensity and duration. Both simulated surface flows showed the same patterns. The HBV simulated base flow dramatically increased with the increase of storm duration. However, the TOP-derived base flow remained stable. Meanwhile, the two models showed contrasting behaviors in the interflow. HBV prefers to generate less interflow but percolates more to the base flow to match the stream flow, which implies that this model might be suited for thin soil layer. The use of the models should consider more environmental background data into account. Compared with the EMMA-derived flows, both models showed a significant 2 to 4 h time lag, indicating that the base-flow responses were faster than the models represented. Our study suggested that model intercomparison under a wide spectrum of rainstorms and with independent validation data (geochemical data) is a good means of studying the model behaviors. Rethinking the characterization of the model structure and the watershed characteristics is necessary in selecting the more appropriate hydrological model.
The sustainable management of water cycles is crucial in the context of climate change and global... more The sustainable management of water cycles is crucial in the context of climate change and global warming. It involves managing global, regional, and local water cycles—as well as urban, agricultural, and industrial water cycles—to conserve water resources and their relationships with energy, food, microclimates, biodiversity, ecosystem functioning, and anthropogenic activities. Hydrological modeling is indispensable for achieving this goal, as it is essential for water resources management and mitigation of natural disasters. In recent decades, the application of artificial intelligence (AI) techniques in hydrology and water resources management has made notable advances. In the face of hydro-geo-meteorological uncertainty, AI approaches have proven to be powerful tools for accurately modeling complex, non-linear hydrological processes and effectively utilizing various digital and imaging data sources, such as ground gauges, remote sensing tools, and in situ Internet of Things (IoT...
This study intends to model the shoreline change by investigating monthly shoreline position data... more This study intends to model the shoreline change by investigating monthly shoreline position data collected from seven sandy beaches located at the Yilan County in Taiwan during 2004-2011. The harmonic analysis results indicate shorelines appear significantly periodic with great variation. The adaptive neuro-fuzzy inference system network (ANFIS) is configured with two scenarios, namely lumped and site-specific ones, to extract significant features of shoreline changes for making shoreline position predictions in the next year. The lumped models for all stations are first investigated based on a number of possible input information, such as month, location, and the maximum and mean wave heights. The results, however, are not as favorable as expected, and wave heights do not contribute to modeling due to their high variability. Consequently, a site-specific model is constructed for each station, with its current position and nearby stations' positions as model inputs, to predict its shoreline position in the next year. Compared with the harmonic analysis and the autoregressive exogenous (ARX) model, the ANFIS model produces more accurate prediction results. The results indicate that the constructed ANFIS models can accurately predict shoreline changes and can serve as a valuable tool for future coastline erosion warning and management.
Optimal reservoir operation strategy for balancing ecosystem and human needs
The frequency of extreme hydrological events varies highly in Taiwan, and increasing attention ha... more The frequency of extreme hydrological events varies highly in Taiwan, and increasing attention has been paid to the optimal reservoir operations. This study establishes an optimization model for watershed management through reservoir operations subject to human and ecosystem needs. The Shihmen Reservoir in Taiwan is used as a case study. This study adopts the Taiwan Eco-hydrological Indicator System (TEIS) to classify river flow patterns. We combine the non-dominated sorting genetic algorithm II (NSGA-II) with the self-organizing radial basis network (SORBN) to develop the optimal model of reservoir operation. The results indicate that it is possible to simultaneously satisfy human and ecosystem needs, where ecosystem diversity can be retained in high SI values (1.7-1.9) while human demands can also be highly satisfied (α higher than 0.85). The proposed approach allows decision makers to easily determine the best compromise in water allocation through the trade-off between human and...
This study proposes a multi-objective optimization model of two cascade reservoirs in the Upper Y... more This study proposes a multi-objective optimization model of two cascade reservoirs in the Upper Yellow River basin for increasing social well-beings in general while simultaneously mitigating ice/flood threats. We first develop a strategy of dimensionality reduction and constraint transformation to largely diminish the complexity of the optimization system and next propose a novel search method that fuses a Feasible Search Space (FSS) into the Particle Swarm Optimization (PSO) algorithm, i.e. FSS-PSO, to effectively solve the optimization problem. To investigate the applicability and effectiveness of the proposed method, this study compares the FSS-PSO model with historical operation. The results indicate that the proposed model produces much better performances in all the objectives than historical operation. To assess the superiority and efficiency of the proposed FSS-PSO, the classical PSO and the Chaos Particle Swarm Optimization (CPSO) are also implemented to compare their comp...
Application of Self-Organizing Radial Basis Neural Networks for estimating riverine biodiversity
Investigation on environmental flow for conservation of river ecosystem has been focused on ecolo... more Investigation on environmental flow for conservation of river ecosystem has been focused on ecological flow regime approach which is more comprehensive than the traditional minimum flow management schemes that merely consider single flow value. Therefore, the pivotal difficulty in developing ecological flow regime is how to take into account the interaction and relation between flow regime and river ecosystem. In this study we first present an idea of considering the relation between ecological flow regime and fish communities and then applying the gradient analysis technique in quantitative ecology theory to the ecological response model. The model is built by using the fish abundance (diversity) and the Taiwan Ecohydrology Indicator System (TEIS). Moreover, the introduction of dummy variables represent synthetic environment gradient to identify the niche in each fish species on ecohydrological gradient axis. The main advantages of this technique are: (1)Approximate the natural flo...
Uploads
Papers by Fi-john Chang