Probabilistic Analysis of Optimal Management of Storage Areas 

Introduction

Flood is one of the main and most frequent natural disasters in the world (Jiang et al. , 2006). As revealed by World Meteorological Organization (WMO, 2014), storms and floods accounted for 79 percent of total number of disasters during 1970 to 2012, causing 54 percent of deaths and 84 percent of economic losses. It is found by many researchers (IPCC, 2014; Saghafian et al. , 2010; Jiang et al. , 2006; Li, 2011) that the flood events have been increasingly frequent and show a trend of greater damage in last decades. The global issues, such as climate change, population growth and deterioration of ecosystem services, may further contribute to higher flood risk (WMO, 2018). China is abundant in water resources, having considerable rivers with 42000 km length in total. Meanwhile, China is also one of the countries that are most affected by floods in the world in terms of both number of deaths and economic losses.

The map (figure 1) provided by WMO (2014) illustrates that China lost more than 100 billion dollar by disasters in the period of 1970 to 2012, which is mainly due to floods. With increasing impact of human intervention and climate change, diversity of flood hazards, such as higher risk of dam break, is defined as new flood risk features (Wan and Wang, 2011). Its length is 1000 km and basin area is 270000 km2 with 170 million population and 12 million hectare cultivated land in total, ranking third among all Chinese basins. The annual runoff and precipitation in Huai river is recorded as 62. 2 billion m3 (Gebeyehu et al. , 2018) and 875 mm respectively. The precipitation also shows an uneven distribution which concentrates in the period of July to September (70%) and decrease from south (1400-1600 mm) to northern areas (600-700 mm) (Qian et al. , 2017). Due to these characteristics, Huai River constantly suffers from floods. For example, from 1994 to 2007 floods occurred ten times (Gebeyehu et al. , 2018). What is more, Jun et al. (2012) found that water resources in Huai River are rather sensitive to climate change. Therefore, the water management in Huai River is facing not only problems with the current situation but also future challenges.

Problem Definition

As mentioned above, many natural disasters, especially flood events, often occur in Huai River basin. To deal with flood together with drought problems Chinese government has constructed a large amount of water projects along Huai River, such as flood gates and dams, as well as projects with non-structural measures. Among the structural measures, the operation of storage areas can effectively alleviate flood situation by storing flood water inside them so that important downstream cities are protected. However, the interests of residents living in storage areas would then be harmed, as they utilise the areas for housing and agricultural activities. Apparently, the protection of downstream areas/cities and minimizing flood in storage areas are two conflicting objectives.

Therefore, there are significant and urgent needs to develop an effective and reliable improvement of storage area operations to reduce flood damage in flood-prone areas. Some optimal strategies for middle part of Huai River have been proposed in previous research (Sun, 2016; He, 2017; Mu, 2018) by simulating the water movement in the river channel and in the storage areas. Sun (2016) developed a HEC-RAS 1D river model coupled with NSGA II multi-objective optimization algorithm, in which storage areas are modelled as conceptual reservoirs. He (2017) compared the results of three different multi-objective optimized algorithms with same model as Sun’s (2016). Mu (2018) updated the damage estimation function and further developed a terrain based 1D-SA model which allocates water with spatial distribution, and a 1D-2D model. No matter which kind of model or algorithm were previously used, what is common to them is that same data was used as boundary conditions.

These are the flow hydrograph and rating curve starting from 28th June to 4th September, 2007. It means that the model results and proposed strategies are valid for this flow condition. In other words, it is unknown whether those strategies work well or not with different inflow hydrograph. Testing the performance of strategies under different inflow conditions can help solve this problem and some updated suggestions may be generated. Currently, the operation of storage area gates are determined by the hydrological data (normally water level) at nearby station referring to the rules (The Huaihe River Commission of the Ministry of Water Resources, P. R. C. , 2004) obtained by analysing historical events. Some shortcomings are found about these rules mainly because they neglect the technology of forecasting. To be more specific, they give much focus on small possibility of flood, resulting in unnecessary damage in storage areas. However, in case an extreme rainfall happens, the operation of storage areas is unable to ensure the security of downstream cities. Given these problems, an adaptive measure such as real time operational scheme for storage area operation should be developed by considering the upstream inflow and forecasting.

Objective and Research Question

To reduce flood damage on economy, society and environment is usually the objective of flood management. In this case study, conflict exists between protecting downstream city Bengbu and preserving the livelihood in storage areas where local settlements and agricultural activities develop. So, how to balance the interest between Bengbu and storage areas should be answered in this study. Furthermore, based on analysis of previous studies (see section 2), their disadvantages (single inflow condition and lack of real-time control strategy) should be overcome. In conclusion, this study will try to improve optimal management for storage areas in order to reduce flood risk by considering possible variation of upstream inflow and it will explore possibilities for real time operation. The sub-objectives are defined as:

• To determine the performance of existing optimal operation strategies under conditions of different flood hydrographs coming from upstream;
• To determine possible improvements of the optimal operation strategies, given the uncertainty of the flood hydrographs coming from upstream; and
• To propose an initial version of operating the storage areas in real time, given the uncertainty of the flood hydrograph coming from upstream.

The objectives can be achieved by answering those research questions:

• What are the possible variation of flood hydrograph?
• What is the difference and similarity of the flood damages under different flood hydrographs using existing optimal operation strategies? What is the reason for these differences?
• What are the optimal operation strategies under different flood hydrograph? Which are the most robust ones?
• Which location of upstream flow information is suitable for real-time control? How the forecast can be used in real-time operation? How can the lead time be extended in a real-time scheme?