Wide river or narrow river: Future river training strategy for Lower Yellow River under global change

Abstract

The choice of a river training strategy is extremely important for the Lower Yellow River (LYR). Currently, the wide-river training strategy applies in the training of the LYR. However, remarkable changes in the hydrological processes in the Yellow River basin, as well as immediate pressure from socio-economic development in the Yellow River basin, make it necessary to consider if there is a possibility to change the river training strategy from wide-river training to narrow-river training. This research investigates the impacts of different river training strategies on the LYR through numerical simulations. A one-dimensional (1-D) model was used to simulate the fluvial processes for the future 50 years and a three-dimensional (3-D) model was applied to study typical floods. The study focused on river morphology, the results show that if the present decreasing trend in both water discharge and sediment load persists, the deposition rate in the LYR will further decrease no matter what strategy is applied. Especially, narrow-river training can achieve the aim to increase the sediment transport capacity in the LYR compared with wide-river training. However, if the incoming water and sediment load recovers to the mean level of the last century, main channel shrinkage due to sedimentation inevitably occurs for both wide-river and narrow-river training. Most importantly, this study shows that narrow-river training reduces the deposition amount over the whole LYR, but it provides little help in alleviating the development of the “suspended river”. Instead, narrow-river training can cause aggradation in the transitional reach where the river pattern changes from highly wandering to meandering, further worsening the “hump deposition” there. Because of uncertainty regarding future changes in hydrological processes in the Yellow River basin, and the lack of feasible engineering measures to mitigate “suspended river” and “hump deposition” problems in the LYR, caution should be exercised with respect to changes in the river training strategy for the LYR.

Introduction

The lower reach of the Yellow River, also referred to as the Lower Yellow River (LYR) has a length of 786 km measured from the Taohuayu Valley to the estuary of the Yellow River (Wang et al., 2005) (Fig. 1). The total elevation in this reach drops by 93.6 m, with an average slope of 0.12‰. Generally, based on its geomorphological features, the LYR is further divided into four sub-reaches along the river course (Chien & Wan, 1999). The first sub-reach, starting from Taohuayu Valley and ending at the Gaocun gauging station, is referred to as the highly wandering river with vast floodplains. The second sub-reach, located between the Gaocun gauging station and the Taochengpu, changes from highly wandering to meandering known as the transitional reach. The third sub-reach, from the Taochengpu to the Lijin gauging station, belongs to the category of a confined meandering river with a comparatively narrow plan configuration. The last sub-reach, from the Lijin gauging station to the Bohai Sea, is the estuarine area of the Yellow River.

Because of the serious soil erosion on the Loess Plateau, hyper-concentrated flows frequently occur in the middle and the lower reach of the Yellow River (Chien et al., 1987, Miao et al., 2011). Analysis of the hydrological records at the Sanmenxia (SMX) hydrological station shows that the annual average suspended sediment concentration (SSC) is as high as 35 kg/m³ over the last century, and an average of 1.6 billion t of silt and clay was delivered downstream annually (Wu et al., 2004); nearly 25% of the sediment load from upstream deposited in the LYR (van Maren et al., 2010, Wu et al., 2005). The continuous sedimentation in the LYR, which may lead to breaches of the levees, still imposes an immediate threat to the residents outside the Grand Levees that constrain the LYR to the narrow path shown in Fig. 1. Therefore, flood control by river training has long been one of the top priorities for both the local and central governments (Li, 2002, Ongley, 2000). Much effort has been devoted to training the river throughout the history of China.

A chronological review of the evolution in the river training strategies for the LYR shows that there are two completely different training strategies: wide-river training and narrow-river training (Li & Li, 2002). The fundamental idea of the wide-river training strategy is to keep vast floodplains between the main channel and the Grand Levees to retard floods and promote the deposition of over-loaded sediment. This strategy was put into practice two thousand years ago and can be traced back as early as to the Han Dynasty (206 BC–220 AD). In contrast, the aim of the narrow-river training strategy is to convey both floods and over-loaded sediment downstream through a comparatively narrow single river channel with the help from a variety of river training works to increase the sediment transport capacity. However, the narrow-river training strategy was applied only for a relatively short period because it had to face to the difficulties in maintaining the narrow channels and the increasing risk of frequent breaching of the levees (Chien, 1990). This situation of high risks finally led to gradual abandonment of narrow-river training, and later wide-river training was re-applied over the past two hundred years and is still currently applied.

The conflict between natural and societal functions of the LYR has long been one of the most intractable problems under the wide-river training strategy (Li et al., 2002, Niu et al., 2013, Wang and Zhang, 2013). Along both sides of the main channel of the LYR, the floodplains, which extend over an area about 4000 km², are divided into hundreds of patches enclosed by levees constructed at low standards by local residents to protect their farmland and cottages. These low standard levees block the lateral exchange of water and sediment between floodplains and the main channel during inundation (Zhang et al., 2011), which gradually leads to “secondary suspended rivers”. Additionally, the low standard levees are not well-designed and constructed under the guidance of the principles of river engineering. The uncertainty with respect to safety of these structures is, therefore, a potential threat to the residents living on the floodplains when unexpected large floods occur (Li et al., 2002). As a result, there has been a controversy over many years as to whether these low standard hydraulic structures should be completely abandoned and the local residents relocated, or the local government should use their power to enforce the construction of permanent levees with sufficiently high standards to provide adequate protection for farmland and cottages.

Because of human activities and global climate change, there now seems to be an opportunity to resolve this dispute persisting for centuries. One of the most significant impacts on the LYR comes from the continuous damming in the Yellow River in the past fifty years (Chen et al., 2017, Hou and Wang, 2017, Zhou and Zhang, 2012), especially the construction of the Xiaolangdi (XLD) Reservoir which became fully functional in 1999. Up to 2014, XLD Reservoir accumulated to 32.5×10⁸ m³ of sediment, indicating that nearly 80% of the total sediment load transported from the upstream catchment was trapped in the reservoir (Liu, 2016). At the same time, the hydrological features of the LYR have changed dramatically over the past decades (Huang et al., 2017, Liu et al., 2012, Liu et al., 2013, Miao et al., 2010, Si et al., 2017, Wang et al., 2006), showing a remarkable decrease of both runoff and sediment load (Shi et al., 2017). The varied runoff and sediment load under the operation of the XLD Reservoir have led the main channel to change from experiencing aggradation to experiencing degradation since 1999 (Fig. 2a) (Hu & Zhang, 2010). In turn, the cross-sectional bankfull discharge, an index that represents the conveyance capacity of main channels, has recently recovered to the levels of 1980s (Fig. 2b). Both the degradation and the increasing bank-full discharge in the LYR reduce the risk of overbank floods, and, thus, local residents believe that it is a luxury to maintain vast floodplains solely for flood detention and sediment deposition. Consequently, a growing opinion believes that narrow-river training should be re-examined, so that floodplains can be protected by permanent levees to promote local socio-economic development (Niu et al., 2013, Wang and Zhang, 2013, Zhang, 2004).

Therefore, several measures for narrow-river training have been proposed (An et al., 2013a, Hu, 2015, Niu et al., 2013, Zhang et al., 2011). The one that is preferred by the majority of the engineers is the one proposed by Zhang et al. (2011), in which a couple of longitudinal embankments along both sides of the main channel are to be built. Meanwhile, floodplains are separated into sub-floodplains by the lateral embankments. The purpose of the construction of longitudinal embankments is to increase flow velocity to obtain a high transport capacity of sediment load and at the same time, to prevent the lands inside the embankments from inundation when floods are below a certain return period. In addition, each of the sub-floodplains is equipped with two weirs on the longitudinal embankment to allow water to flow into/out of the sub-floodplains during floods larger than a certain return period. This arrangement is expected to help to alleviate the development of a “suspended river” (Fig. 3a).

However, whether it is possible to shift from wide-river training to narrow-river training has not been fully discussed under the new hydrological circumstances. The proposed measure to implement narrow-river training in the LYR requires detailed study concerning both long-term and short-term reactions to the changes in the river training strategy. Of particular importance, the long-term river erosion and deposition, and short-period flood events are two key problems that need detailed study. To realize narrow-river training in this study (Fig. 3b), the alignment of the embankments, which was refined by An et al. (2013a) on the basis of the proposal suggested by Zhang et al. (2011), was selected.

In this study, one-dimensional (1-D) and three-dimensional (3-D) numerical models are applied to tackle these issues and to assess the responses of the LYR to the two training measures. The aim of using a 1-D model was to study the long-term evolution of the LYR under different river training measures, while a 3-D model was used to simulate design floods with different return periods, focusing mainly on river morphology.