1 Introduction

From the geomorphological perspective, the direct factor determining the change in the river system is the input and output of water and sand (Wright and Coleman 1973). However, human activities have been reported to have a more comprehensive impact on river systems in the past few decades (Yang et al. 2004; Caldwell et al. 2019; Nienhuis et al. 2020). Therefore, it is important to assess changes in river systems and potential hydrological responses under anthropogenic impacts, which can provide supportive information for maintaining river ecosystem balance and sustainable socioeconomic development (Zhang et al. 2018; Chen et al. 2023).

Previous studies have shown that the urbanization process changed morphological parameters of river networks (Bracken et al. 2013), and altered river connectivity (Zuo and Cui 2020). Chin (2006) surveyed more than 100 studies of river landscapes in a global context and found that urbanization has transformed river landscapes across the surface of the earth, including decreases in sinuosity, increases in bed material size and drainage density, along with other chemical, biological, and ecological effects. Urbanization has not only changed the external features of the river system but also to some extent weakened its storage and flow capacity (Vanacker, et al. 2005; Wang et al. 2016). For instance, Zhang et al. (2015) found that a decrease in the fractal dimension of a river network can cause an increase in flood frequency. Desmond and Patrick (2021) found that urbanization at the watershed scale led to river degradation and had comprehensive impacts on the frequency, magnitude, and volume of peak flows. At the same time, more attention is paid to the connectivity of rivers (Pringle 2003; Hooke 2003), which is defined as the connection between upstream and downstream (Jaeger et al. 2014). It has been proven that related changes in the physical connectivity of river systems can affect the transfer of matter, energy, and organisms, and the physical, chemical, and biological processes of river systems (Yang et al. 2004; Deng et al. 2016).

Although previous studies have well examined changes in the urban rivers, the results vary with natural properties (e.g., geographic characteristics) and social properties (e.g., urbanization level) (Zhao et al. 2011; Han and Wei 2022). In particular, it is easy for the river system on the delta plain, ranked among the most economically valuable environments on earth, to depart from the natural evolution process due to intense human activities (Zhou et al. 2013; Wu et al. 2017; Best 2019). For one thing, the delta plain generally has braided and crisscrossing streams, and the flat terrain and dense river networks with low stability make it vulnerable to inundation threats (Ma et al. 2012; Wang et al. 2019; Yamazaki et al. 2019). For another thing, rapid urbanization has resulted in significant changes in the river systems of the area, causing frequent floods and other hazards that pose a serious threat to the area's communities (Wu et al. 2018; Yuan et al. 2019). Since the role of urbanization in river changes is complex, it is of great significance to comprehensively investigate the evolution characteristics of the river system, including its structure, function, and subsequent hydrological effects.

The Yangtze River Delta (YRD) has become the strategic economic zone in China and one of the largest urban agglomerations in the world (Wang et al. 2016; Zhang et al. 2018). Since China officially launched reform and opening up in 1978, this area has experienced rapid urbanization, and urban expansion is the direct driving factor that has disturbed the natural evolution of the river. The development over the past few decades has also brought out some adverse influences on the local river systems, especially in plain areas with dense river networks. Urbanization has transformed many natural rivers into artificial landscapes. In addition, a large number of hydraulic engineering structures (e.g., sluices, pumps, and dams) have been built to reduce the loss of flood, and these constructions destroyed the connectivity of rivers and affected their natural regulation capacity (Deng et al. 2015; Han et al. 2016; Lu et al. 2020). Therefore, it is crucial to detect changes in the river systems of the YRD and then strike a balance between river evolution and urban development. The main aims of this study are as follows: (1) quantitatively investigate changes in various indicators of river system characteristic across the YRD under urbanization; (2) reveal responses of hydrological processes to these changes based on the long-term water level series in the Taihu Lake Plain; and (3) exploring potential impacts of river changes on regional floods and regulation storage. Our results will be useful for river protection and flood management in urbanized areas.

2 Materials and methods

2.1 Overview of the Yangtze River Delta

The Yangtze River Delta (YRD), located in China's eastern coastal region, comprises Jiangsu Province Zhejiang Province, and Shanghai City. This region mostly belongs to the subtropical monsoon climate zone with an average annual temperature of 15 ℃. It receives heavy rainfall during the flood season, with an annual average of more than 1,000 mm, due to the influence of Meiyu rainfall and typhoon rainstorms (Han et al. 2016; Wang et al. 2019). The topography of the region consists primarily of plains and hills, with the southwestern portion of the region characterized by hilly and mountainous terrain (Han et al. 2016). Due to the delta landscape, the YRD features an extensive network of river systems, including the Huang-Pu River, the Tiao-Xi River, the Yong River, the Qin-Huai River, the Jing-Hang Grand Canal, and the rivers of the Taihu Lake Basin. The YRD is divided into different water conservancy regions for better water resource management, and HPJ, QHH, YCDM, YDN, HJH, and WCXY regions are typical regions among them (Fig. 1a). In particular, the WCXY, YCDM, and HJH regions compose the Taihu Lake Plain, with an average elevation of less than 5 m.

Fig. 1
figure 1

a Location of the YRD; b changes in the proportion of urban built area in main cities of the YRD

Full size image

In recent years, the YRD has rapidly urbanized with both social and economic development, and accounts for more than 20% of the gross domestic product (GDP) of China (Chen et al. 2023). The massive urbanization occurred after 1978, and Fig. 1b shows that the urban built land proportion in the main cities of the YRD has increased over the last few decades, indicating a significant growth in its urbanization level. At the same time, the cities in the YRD, especially Shanghai, Changzhou, Suzhou, and Wuxi, have been built on dense river networks (Wang et al. 2016; Deng et al. 2018). When the urban area continued to expand outward, some surface water bodies were shrinking. In addition, hydraulic engineering construction (e.g., sluices, water gates, pumps, and large-encirclement projects) has also led to complexity in river systems (Lu et al. 2020). Therefore, there is an urgent need to focus on the changes in river features, to provide support information for flood control and sustainable development in the YRD.

2.2 Data

The river data were derived from maps with a scale of 1:50000, including topographic maps in the 1960s (compiled during 1964 ~ 1969) and 1980s (compiled during 1980 ~ 1983), and digital linear maps in the 2010s (surveyed and drawn in 2015). The basic elements of river systems in the maps were extracted by ArcGIS, such as polygons (e.g., lakes, ponds, and main rivers) and polylines (e.g., tributaries). To improve accuracy and reliability, these data were checked and corrected by comparing them to Landsat high-resolution images, land-use maps, and regional conservancy survey data over the same period in the study area. Furthermore, in order to better understand how the hydrological process responds to river system changes, the daily mean water level data from 1960 to the present at 30 hydrological stations located in the Taihu Lake Plain, including the WCXY, YCDM and HJH regions (see Fig. 1a), from the Annual Hydrological Report of China. In addition, model simulations were carried out using a series of 1-h interval precipitation data of rainfall events from weather stations in the Yunbei-Large-Encirclement Project area.

2.3 Methods

2.3.1 Assessing changes in characteristics of river system

In order to quantitatively describe the changes in the river system in the study area, several indicators derived from the fluvial geomorphology, which reflects the geometric features of the river system, were established, including the density of rivers, the proportion of water surface, the evolutionary coefficient of rivers, the ratio of area to length of rivers, and the box dimension. Meanwhile, we used ArcGIS to modify the topological relationship between the river networks and further extracted the nodes and chains of the river networks. Then, the structural connectivity of the river network was calculated based on the evaluation theories in graph theory (Cook 2002; Cui et al. 2009). Three indicators were selected as follows: Network circuitry (α) is the ratio of the actual number of closed flow paths in a river network to the maximum possible number of closed flow paths. The ratio of edge to node (β) represents the average number of river chains connected to each node in the river network. Network connectivity (γ) means the ratio between the actual number of river chains in a river network and the possible maximum number. Details of all indicators can be found in Table 1.

Table 1 Details of indicators for assessing changes in characteristics of river system
Full size table

2.3.2 Hydrological connectivity

Hydrological connectivity is an important metric that represents the connection between the water transport of matter, energy, and organisms in the regional hydrologic processes (Freeman et al. 2007), and changes in river systems, especially for morphological and structural features, have led to an alteration in such process (Wu et al. 2018). In this study, a longitudinal functional connectivity index (LFCI) was selected to evaluate the degree of hydrological connectivity according to the functional connectivity evaluation method in the river network plain areas (Deng et al. 2018). This indicator can mainly reflect the synchronization degree of water level changes at adjacent stations on the same river. The greater the index, the better the hydrological connectivity of the river, and thus the better the synchronicity of changes in water flow.

$$LFCI=\frac{1}{10\times \left|\Delta {H}_{ab}\right|+1}=\frac{1}{10\times \left|\left({H}_{ai}-{H}_{aj}\right)-\left({H}_{bi}-{H}_{bj}\right)\right|+1}$$

where 0 < LFCI ≤ 1; ΔHab is the difference of water level between the two adjacent stations a and b. Hai and Haj are the water level of station a at times i and j, respectively; Hbi and Hbj are the water level of station b at times i and j, respectively. In particular, LFCI = 1 indicates the best longitudinal functional connectivity, while close to 0 indicates the worst.

2.3.3 Regulation and storage capacity

Flood regulation is a basic function of river systems in the delta plain and can be divided into two parts, namely, channel storage and regulation capacity (Xu et al. 2018). Here, we developed the specific storage capacity (SSC) and the specific regulation capacity (SRC) as indicators to characterize flood regulation changes corresponding to river system changes. The specific storage capacity (SSC) is defined as the total volume of water that a river holds under the ordinary water level, which is calculated as the product of the river area and water depth under the ordinary water level in a unit area. The specific regulation capacity (SRC) is defined as the spare volume of water that a river can hold before reaching its warning water level in a unit area, which is generally considered to be the product of the river area in a unit area and the water level difference between the ordinary water level and the warning water level. The unit of both indicators is m3/km2. Note that the ordinary water level is considered to be the mean observed daily water level, and the warning water level is considered to be the extremely high water level when the local water conservancy department should issue a flood warning.

2.3.4 Hydraulic model

To explore the impact of river changes on regional floods, we chose the Yunbei-Large-Encirclement Project region of Changzhou City as a typical study area for the case study (Fig. 1a), where urbanization has largely replaced natural surfaces and rivers have been profoundly modified by human activities. For this purpose, the MIKE11 model was used to construct simulations for different river network scenarios. MIKE11 is a professional hydraulic model software developed by DHI Water and Environment and is used for the simulation of one-dimensional flow in estuaries, rivers, irrigation systems, and channels (Rahman et al. 2011; Wang et al. 2022). The initial conditions, boundaries, and operation of the hydraulic engineering works constructed in the model scheme are based on field surveys, and then the range of river roughness is set between 0.020 and 0.045 according to the actual situation. We imported the rainfall and water level data as inputs in the model to simulate the changes in the rain-flood process. Several rainfall-flood events (from July 19 to September 9, 2017) were selected for calibration. The simulated results for each station were in general agreement with the observed processes. Two events ‘20150602’ and ‘20161018’ were selected for validation. Both the R2 and Nash coefficients in the calibration and validation process were above 0.90, and the simulated rain-flood processes at each station are in good agreement with the observed data. We then set up the following three scenarios of river systems: S1, tributaries and main rivers in the situation of the 1980s; S2, tributaries in the situation of the 1980s and main rivers in the situation of the 2010s; S3, tributaries and main rivers in the situation of the 2010s. The other parameters of the MIKE 11 model were kept constant during the simulation of these scenarios. The outputs of the simulations were flood hydrographs for different river network conditions, thus reflecting how floods respond to river changes.

3 Results

3.1 Changes in the river system of the Yangtze River Delta

3.1.1 Changes in the geometric features

The scale of the analysis has a practical impact on quantifying changes in regional river systems, and here we demonstrate the spatial heterogeneity of river system evolution in terms of water conservancy areas and urbanized areas.

Water conservancy region scale

To visually demonstrate the changes that have occurred in the river system, we used the HJH region as an example (Fig. 2). It is obvious that there is an increase in the main rivers but a decrease in the tributaries. However, the losses of tributary were much more severe, particularly in the southeastern part of the region. Various indicators for different periods in typical regions of the Yangtze River Delta are shown in Table 2. The maximum and minimum density of rivers in the 1960s were 3.80 and 1.25 km/km2, respectively. The water conservancy regions located in the Taihu Lake Plain (i.e., the WCXY, YCDM, and HJH regions) have the highest density of rivers (Dd) with more than 3.5 km/km2 in the 1960s, while the QHH region has the smallest Dd of 1.4 km/km2. Moreover, the average Dd of the YRD decreased from 3.41 km/km2 in the 1960s to 1.14 km/km2 in the 2010s, with WCXY, HJH, and YDN regions showing a significant reduction of nearly 20%. However, the YCDM and QHH regions show a smaller decrease of 3.8% and 9.1%, respectively. Apart from the QHH region, the proportion of water surface (Wp) has also significantly decreased in these regions, and the decreasing trend has intensified. Over the past few decades, the Wp of the YDN region has decreased most (29.5%), and the Wp of the WCXY, YCDM and HJH regions has decreased by 23.5%, 19.4%, and 17.5%, respectively. However, the water surface in the QHH region has been increasing since the 1960s and grew to 7.52% in the 2010s, which is closely related to the large increase in artificial ponds since the 1980s (Han et al. 2016).

Fig. 2
figure 2

River system in the HJH region at different periods: a main rivers; b tributaries

Full size image
Table 2 Changes in the geometric features of river system in the typical water conservancy regions
Full size table

From the perspective of the ratio of area to length of rivers (RAL), there was much less change in all regions except a significant growth in the QHH region, which increased from 40.0 in the 1980s to 73.2 in the 2010s. This is mainly due to the widening of main rivers and the excavation of new drainage channels during the process of urban development. Meanwhile, the evolutionary coefficient of rivers (K) varied among different regions. The WCXY region has the highest value of K, indicating a highly developed tributary network. On the contrary, the YCDM region with many lakes and the QHH region with sparse river network have lower K values. The results indicate that the tributary development in the WCXY, HJH, and YDN regions has been declining over the past fifty years, with the most significant reduction in the HJH region, where the K value decreased by 46.8% from the 1960s to the 2010s. This trend indicates the variability in the stability of the river structure, with many tributaries being converted into main rivers.

There is a significant spatial heterogeneity in the box dimension (D) that the central Taihu Lake Plain (i.e., the WCXY, YCDM, and HJH regions) have relatively high values of D, while the QHH region has the lowest. This difference is similar to Dd, indicating that there is a positive correlation exists between the box dimension and the density of the rivers. In addition, the value of D has decreased in all regions, ranging from 1.32 to 1.71 in the 1960s, and from 1.27 to 1.65 in the 2010s. In particular, the WCXY and HJH regions have experienced the most significant reduction, reaching a ratio of 7.8% and 6.5%, respectively, while the YCDM region with dense lakes showed a decrease of only 2.0%.

Urbanized area scale

The number of rivers in major urbanized areas of the YRD has declined significantly, as shown in Fig. 3a. Suzhou, which has the most developed rivers, saw its Wp drop from 12.21% to 9.6%, and its Dd from 3.17 km/km2 to 2.63 km/km2. In contrast, the Dd and Wp of Nanjing both increased by 34.50% and 47.81%, respectively. After the 1980s, the degeneration of river systems in each city accelerated, with variations ranging from -35.25% to 1.84% for the Dd and Wp. An increase was found in the RAL from the 1980s to the 2010s in all cities, and Nanjing increased most by a ratio of 71.46% (Fig. 3b). These results indicate that small streams in the urbanized areas have gradually disappeared and the status of main rivers has been improved, leading to a simplification in the structure of river systems. Moreover, Fig. 3c shows that Suzhou has the most developed river system with a mean D value of 1.51 and is relatively constant, while the trends of the box dimension in the other cities vary considerably, indicating that anthropogenic modification can lead to differences in regional river structures.

Fig. 3
figure 3

Changes in the geometric features of river system in the typical urbanized areas: a the density of rivers and the proportion of water surface, b the evolutionary coefficient of rivers and the ratio of area to length of rivers, and c the box dimension. Note that the data from Shanghai was derived from Cheng (2019)

Full size image

3.1.2 Changes in the structural connectivity

Structural connectivity is described as the spatial pattern of a river network landscape, which affects water transport and flow paths, and can reflect how river systems change in a pattern of morphological composition. We quantified the structural connectivity of the main rivers in the Taihu Lake Plain, as shown in Table 3. River elements (i.e., nodes and chains) all show increasing trends, with more significant increases in the 1960s-1980s than in the 1980s-2010s. During the former period, the number of nodes and chains increased by 45.27% and 52.82%, respectively. In terms of structural connectivity indicators, affected by the changes of nodes and chains, the increase rates of the network circuitry (α), the ratio of edge to node (β), and network connectivity (γ) were much higher in the former period than in the latter. Among them, α increased from 0.1640 to 0.2022 by an increase of 23.29%, β increased from 1.3234 to 1.4018 by an increase of 5.92%, and γ increased from 0.4433 to 0.4685 by an increase of 5.68%. However, the most ideal values of α, β, γ are 1, 3, and 1, respectively (Cook 2002), so the structural connectivity of the area was not at a good level. Combined with the results in changes in the river feature indicators, it can be known that the structural connectivity has been altered by the changes in the composition of nodes and chains, and the reconstruction of the main rivers is not enough to counteract the negative effects caused by the degradation of the small streams.

Table 3 Changes in the structural connectivity of the main rivers
Full size table

3.2 Responses of hydrological processes to changes in river systems

3.2.1 Changes in the long-term water level series

A direct response of hydrological processes to changes in rivers is the dynamics of normal water flowing (Cheng 2019). Therefore, we analyzed the trends of extreme water level (i.e., annual mean maximum daily water level) and rise of water level (i.e., mean rise of water level during rainfall events) in the Taihu Lake Plain since 1970. The former indicator can represent the upper limit of the change in the hydrological process in a river during an extreme rainfall event, and the latter indicator can describe the change in the mean volume of water in a river during a rainfall event. As Fig. 4a shows, the results suggest that the extreme water level of the river systems in this area was detected to have a continuously increasing trend, which increased from 3.56 m in the 1970s to 4.22 m in the 2010s, with an increase of 18.54% (slope = 0.13 m/10a, P < 0.01), especially with high values exceeding 4.5 m in 1991, 1999, and 2016. As shown in Fig. 4b, the mean rise of water level in the past fifty years is 0.27 m. However, its trend was relatively fluctuating. The average value was 0.26 m from 1970 to 1990, with a higher phase in the 1990s when the average value of 0.30 m, before falling back to its previous level after 2000. The results indicate that changes in the river systems have led to significant impacts on the increase in the extreme water levels and uncertainty in the rise of water level, making the delta plain area vulnerable to flood disasters.

Fig. 4
figure 4

Trends of water level change in the Taihu Lake Plain from 1970 to 2020: a extreme water level and b rise of water level during rainfall events

Full size image

3.2.2 Changes in the hydrological connectivity

The variations of hydrological longitudinal connectivity are influenced by the structure of the river system and hydrological features in the delta plain (Deng et al. 2018; Lu et al. 2020). Based on the long-term daily water level data, a longitudinal functional connectivity index (LFCI) was utilized to demonstrate the characteristics of hydrological connectivity under changes in river hydrological processes. Figure 5a shows that the average annual hydrological connectivity index in the past few decades increased at the rate of 0.4%/10a, with an average value of 0.84. Based on the hydrological connectivity index for each decade shown in Fig. 5b, it is easy to see that the hydrological connectivity of the Taihu Lake Plain gradually increased from 0.83 in the 1960s to 0.86 in the 2010s. Furthermore, the average hydrological connectivity index (LFCI = 0.816) during the flood season (i.e., from May to September) is lower than the index during the non-flood season (LFCI = 0.852). However, both the flood and non-flood seasons showed a slightly increasing trend, as observed in Fig. 5c and d. Flood season increases were higher than non-flood season increases, at 0.6%/10a and 0.3%/10a, respectively.

Fig. 5
figure 5

Changes in the longitudinal functional connectivity index (LFCI). a annual, b decadal, c flood season and d non-flood season

Full size image

3.3 Impacts on regulation storage and regional floods

3.3.1 Impacts on regulation and storage capacity

River networks play a crucial role in regulating streamflow when storms occur, and the capacity of river regulation can be affected by changes in the water surface. We quantified the specific storage capacity (SSC) and the specific regulation capacity (SRC) to reveal the changes in the regulation storage function of the Taihu Lake Plain. The results show that SSC of the area has decreased from 34.16 × 104m3/km2 in the 1960s to 27.70 × 104m3/km2 in the 2010s, and the SRC decreased from 10.09 × 104m3/km2 in the 1960s to 8.07 × 104m3/km2 in the 2010s (Fig. 6a). The degradation ratio of SSC was higher in the recent decades (1980s ~ 2010s) than in the 1960s-1980s. This indicates that changes in the delta plain rivers have affected the regional regulation and storage function over the past fifty years. We further analyzed the variations in the regulation and storage function of the main rivers and tributaries, as can be seen in Fig. 6b. The results show that the main rivers in the Taihu Lake Plain have maintained better storage capacity than the tributaries. In addition, both SSC and SRC of the tributaries exhibited a significant decrease by a ratio of over 15% in each period (Fig. 6b). As a result, such reductions contribute the most to the overall reduction in the regulation and storage capacity. This indicates that the loss of tributaries closely affects flood control capacity, which could threaten regional water security.

Fig. 6
figure 6

Change in the specific storage capacity (SSC) and specific regulation capacity (SRC): a value and b change ratio in Period I (1960s ~ 1980s) and II (1980s ~ 2010s)

Full size image

3.3.2 Impacts on regional floods

The results of flood process simulation under three scenarios (S1, S2, and S3) of the river network are shown in Table 4. We used the Yunbei-Large-Encirclement Project of Changzhou City as a typical study area for the case study, where the Large-Encirclement Project is a special water conservation zone, with a series of sluices and pumps built at the outlet of rivers to prevent flooding in the inner urban areas. By comparing S1 with S2, the impact of changes in the main rivers on the peak water level showed spatial differences. From the 1980s to the 2010s, the peak water level decreased by 0.07 m within the Large-Encirclement area but increased outside the area. Comparison of S2 and S3 shows that the decrease in the tributaries consequently caused an increase in peak water level at various stations (0.03 ~ 0.04 m) from the 1980s to 2010s, with the peak occurring one hour earlier. Therefore, it can be concluded that the impact of changes in main rivers on the flood process depends on the connectivity of the rivers controlled by hydraulic engineering constructions, while the tributaries mainly affect the flood intensity through their regulation and storage function. Furthermore, a comparison of S1 and S3 shows that the peak water level decreased by 0.03 m and the duration of the high water level inside the Large-Encirclement area decreased by 4 h, but the peak water level outside the area increased (0.07 ~ 0.12 m), indicating that the flood magnitude is tightly bound to the regulation and storage capacity of tributaries. River systems have been severely altered by human activities, leading to a decrease in drainage functions and a degradation in the static regulation and storage functions of regional river systems, increasing the risk of floods. Therefore, river system protection and flood control are important tasks in the Yangtze River Delta.

Table 4 Simulations of changes in flood characteristics under different scenarios of the river network
Full size table

4 Discussion

4.1 Understanding the causes and effects of changes in the river system

The changes in river systems at a long-term scale are the consequence of the comprehensive effects of basin geology, topography, climate change, and human activities, but can be mainly influenced by human activities in a short period (Chin 2006; Belletti et al. 2020). When high-intensity human activities break the original stable evolution of rivers, the erosion and sedimentation processes of rivers are also altered. The Yangtze River Delta region, featuring rapid economic development and urbanization, is one of the largest urban agglomerations in the world (Hou et al. 2021).

With the growth of the built area, the dominance of the natural landscape in the Yangtze River Delta has been gradually weakened (Zhou and Wang 2022). Generally, tributaries and small water bodies were the first to be buried to develop industry and buildings (Deng et al. 2015). As a result, a large number of rivers in the Yangtze River Delta have disappeared, leading to varying degrees of changes in the number, morphology, and structure of rivers. Table 2 shows that the mean density of rivers (Dd) decreased from 3.15 km/km2 in the 1960s to 2.63 km/km2 in the 2010s, the mean proportion of water surface (Wp) decreased from 10.12% in the 1960s to 8.57% in the 2010s. In addition, the main rivers were usually dredged and reconstructed to improve the capacities of storage for flood control, and some tributaries were also widened into the main rivers. Thus, whether at the water-conservancy-region scale or urbanized-area scale, the evolutionary coefficient of rivers (K) and the ratio of area to length of rivers (RAL) exhibited obvious changes due to regional substantial hydraulic engineering practices. Moreover, our findings suggest that the degradation of the tributaries led to a significant reduction in the structural connectivity of the total river network (Table 3), indicating that the structure of river networks has been simplified.

The degeneration of the river systems will cause a decline in regulation function (Steele and Heffernan 2014). In Fig. 6, the SSC and SRC of main rivers were found to be reduced, which is due to the changes in rivers over the past few decades. Thus, the flood regulation function of river channels has been relatively weakened. In addition, sluices or dams are constructed on rivers and such barriers cut off the original hydrological connectivity of river systems (Nilsson et al. 2005; Grill et al. 2014). This kind of phenomenon is particularly common in the plains of the Yangtze River Delta, leaving the river channels largely under artificial control. The hydraulic connections between areas inside and outside of sluices are limited extensively in delta plains (Deng et al. 2018). However, these constructions are often closed during the flood season for flood control and opened during the non-flood season to exchange water between rivers. As a result, the longitudinal functional connectivity of the non-flood season can be better than that of the flood season.

The process of runoff generation and confluence in urban areas is relatively complex (Kaspersen et al. 2015; Sun et al. 2017). However, an important manifestation of the impact of urbanization on hydrological processes in plain river network areas, such as the Taihu Lake basin, is that it leads to changes in water level. Figure 5 shows a significant growth in the average annual water level and the rise in water level during rainfall events. The trends of these observed data series have proven that human activities do affect regional hydrological processes to a large extent. For instance, previous research revealed that human activities contributed more than 80% to the annual water level change after the 1980s (Xu et al. 2020). Furthermore, it is widely reported that urbanization led to a decrease in the underlying roughness and water permeability, resulting in an increase in surface runoff generation and a shorter confluence time (Hammond et al. 2015; Guan et al. 2016). As shown in the simulation results in Table 4, the flood peak and occurrence time change in different river network scenarios, with opposite trends in the inner and outer regions of the Large-Encirclement Project area. With rapid urbanization, a large number of rivers in the Yangtze River Delta have disappeared, and the structure and connectivity of rivers have changed to varying degrees (Deng et al. 2015). Overall, the length of river channels in the region has decreased from 1,330.7 km in the 1980s to 899.6 km in the 2010s with a decrease of 32.4%. While the length of the main rivers has remained essentially constant, the structure of the river network has changed, especially with the relocation project of the Changzhou urban section of the Jing-Hang Grand Canal, which has optimized the structure of the main rivers (Wu et al. 2018). The decline of the river system and the decrease in its storage capacity have led to an increase in the risk of floods in the region. In addition, the increase of pipelines in urbanized areas and hydraulic engineering constructions over rivers not only altered the path of confluence but also led to a reduction in peak time lag (Kaspersen et al. 2015). Ultimately, this may lead to earlier flood peaks and larger peak volumes, an effect that is particularly evident in small and medium-sized flood events.

4.2 Implications for water management and flood control

The terrain in the Yangtze River delta plain is extremely flat, with fluctuating flowing directions, which determines that the basin needs to face complex relationships between rivers, lakes, and seas for flood control (Wang et al. 2021). Our results show that the river has changed dramatically over the past few decades. The physical mechanisms of hydrological processes have changed as a result of human activities and the effects of modifications to the underlying surface of the watershed. Therefore, reasonable land-use policies that maintain a balance between urban development and river conservation are crucial in the early years of urbanization. In addition, attention should be paid to the increasing water level and it is important to establish a monitoring and early warning system. At the same time, it would be unwise to go to all the lengths of hydraulic regulation, including placing numerous hydraulic structures along rivers or draining water from the region. Instead, actions should be taken primarily to protect water surfaces during urbanization, thus maintaining natural watershed-scale storage and relieving pressure on regional flood control. In order to harmonize the development of social progress and river systems under urbanization, it is necessary to protect the natural features of rivers so that they can better serve human society in the long term.

5 Conclusion

This study mainly investigated changes in the river systems across the Yangtze River Delta region and explored responses of hydrological processes in a typical plain river network area. Several characteristic indicators were used to reveal the evolution of river systems from the 1960s to the 2010s at the scale of water conservancy regions and urbanized areas. Furthermore, focusing on the Taihu Lake Plain, we quantified trends in water level and hydrological connectivity, as well as the impact of river changes on regulation and storage capacity and regional floods. The main conclusions are as follows:

  1. (1)

    The quantitative features of river systems in the YRD showed a significant reduction, the average density of rivers (Dd) decreased from 3.41 km/km2 in the 1960s to 1.14 km/km2 in the 2010s, while the proportion of water surface (Wp) exhibited a decrease in most of the subregions. Trends of the morphological features and structural features of river systems varied spatially, but an increase was found in the ratio of area to length of main rivers (RAL) in all cities. Overall, the total number of rivers in the YRD has decreased and the structure of the river system has been simplified, with many tributaries disappearing or being transformed into main rivers.

  2. (2)

    Relevant hydrological responses to river system changes have been found in the Taihu Lake Plain. The extreme water level of the rivers in the area showed a significant trend of continuous increase, which increased from 3.56 m in the 1970s to 4.22 m in the 2010s, presenting potential challenges in flood control. In addition, the longitudinal functional connectivity index (LFCI) of the area slightly increased, and the average hydrological connectivity index (LFCI = 0.816) during the flood season was lower than the index during the non-flood season, indicating that hydraulic engineering structures have governed the regional river system.

  3. (3)

    Both the specific storage capacity (SSC) and the specific regulation capacity (SRC) of the tributaries exhibited a significant decrease, indicating that the loss of multiple tributaries caused a decline in the regulation and storage capacity of the regional river system. The MIKE 11 model simulation results showed that the changes in the tributaries resulted in an increase in peak water level and an advance in peak timing. Still, the trends of peak water level due to changes in the main rivers are opposite in and out of the Large-Encirclement area, suggesting that anthropogenic activities have altered floods by influencing the connectivity of the main rivers and the storage capacity of the tributaries.

There are many environmental factors affecting the structural and connectivity changes of river systems in the study area, but human activities featuring rapid urbanization have played a dominant role, especially the impact of social economy and hydraulic engineering constructions. When decision-makers recognize that urbanization is the main cause of changes in the river system, they can control urban expansion and improve river system resilience. In this context, further research should be conducted on the impact of urbanization on river systems based on multi-dimensional analysis, exploring the mutual coordination and coupling relationship between urbanization and river systems.