Next Article in Journal
A New Method to Combine Coastal Sea Surface Height Estimates from Multiple Retrackers by Using the Dijkstra Algorithm
Next Article in Special Issue
Flood Extent and Volume Estimation Using Remote Sensing Data
Previous Article in Journal
First Open-Coast HF Radar Observations of a 2-Phase Volcanic Tsunami, Tonga 2022
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Remote Sensing
Volume 15
Issue 9
10.3390/rs15092324
remotesensing-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

River Bars and Vegetation Dynamics in Response to Upstream Damming: A Case Study of the Middle Yangtze River

by
Yong Hu
1,
Junxiong Zhou
2,
Jinyun Deng
1,*,
Yitian Li
1,
Chunrui Yang
1 and
Dongfeng Li
3,4
1
State Key Laboratory of Water Resources and Hydropower Engineering Science, Wuhan University, Wuhan 430072, China
2
Department of Bioproducts and Biosystems Engineering, University of Minnesota Twin Cities, St. Paul, MN 55108, USA
3
Key Laboratory for Water and Sediment Sciences, Ministry of Education, College of Environmental Sciences and Engineering, Peking University, Beijing 100871, China
4
Department of Geography, National University of Singapore, Kent Ridge, Singapore 117570, Singapore
*
Author to whom correspondence should be addressed.
Remote Sens. 2023, 15(9), 2324; https://doi.org/10.3390/rs15092324
Submission received: 28 March 2023 / Revised: 24 April 2023 / Accepted: 26 April 2023 / Published: 28 April 2023
(This article belongs to the Special Issue Flood Monitoring, Modelling, Forecasting and Analysis with Remote Sensing Tools)
Browse Figures
Versions Notes

Abstract

:
Investigating river bars and their vegetation dynamics in response to upstream damming is important for riverine flood management and ecological assessment. However, our mechanical understanding of the damming-induced changes in river bar and vegetation, such as bar area, morphology, and leaf area index (LAI), remains limited for large river systems. Leveraging satellite images and in situ observed hydrogeomorphic data from, we improve a machine learning-based LAI inversion model to quantify variations in river bar morphology, vegetation distribution, and LAI in the Middle Yangtze River (MYR) following the operation of the Three Gorges Dam (TGD). Then we analyze the mechanisms controlling the bar and vegetation dynamics based on high-resolution river cross-sectional profiles as well as daily discharge, water levels, and sediment in both the pre- and post-TGD periods. Our results indicate that the river bar area decreased by approximately 10% from 2003 to 2020, while the vegetation area and average LAI of these bars increased by >50% and >20%, respectively. Moreover, the plant community on most river bars tended to expand from the bar tail to the bar head and from the edge to the center. The main factor driving vegetation expansion in the MYR after the TGD’s operation was the reduction in bar submergence frequency (by 55%), along with a slight bar erosion. Further analysis revealed that the standard deviation of annual discharge decreased by approximately 37%, and the frequency of vegetation-erosive flow decreased by approximately 74%. Our data highlight the potential impact of large dams downstream flow regimes and vegetation encroachement. Such findings further the understanding of the biogeomorphological impacts of large dams on the river bar vegetation and have important implications for riverine plant flux estimatin, flood management and ecological restoration in dammed river systems.

1. Introduction

Rivers, as highly dynamic constituents of the Earth’s hydrological cycle, possess intrinsic economic and ecological value, playing a crucial role in the advancement of human societies [1], ecosystem sustainability [2], and regional climate [3]. Yet, their natural balance has been threatened by a wide range of anthropogenic stressors such as the construction of dams, which now number in the tens of thousands globally, and this number is growing [4]. Dams are often engineered to generate hydropower and alleviate flooding events [5,6]; however, they concurrently interrupt riverine continuity and engender considerable spatiotemporal modifications in fluvial flow and sedimentary patterns [7,8,9]. Such drastically changed flow-sediment regimes subsequent to dam development have precipitated significant downstream consequences, encompassing geomorphological, ecological, and societal domains [10,11]. The Three Gorges Dam (TGD; 30°44′18″N, 111°16′29″E), the world’s most expansive dam, is strategically constructed within the upper reaches of the Yangtze River, which holds the distinction of being Asia’s preeminent river (Figure 1). Commencing with the impoundment in 2003, the TGD has exerted considerable influence on the environmental dynamics of the middle and lower reaches of the river (e.g., a sharp reduction of ~90% in sediment loads from 1950–2002 to 2003–2020). The Middle Yangtze River (MYR), being the proximate reach downstream of the TGD, has consequently experienced substantial alterations to its riverine ecosystem, primarily attributable to the regulation of flow and sediment regimes. Recent studies suggest that the TGD could exacerbate floods in certain river systems by triggering enhanced vegetation resistance [12].
Changes in bar and vegetation characteristics reflect the dynamics and geometry of rivers and have consequently been studied for the purposes of flood control, water supply, waterway navigation, aquatic ecosystems, and biogeomorphology [13,14,15,16,17]. Variations in bar and vegetation downstream of dams have been documented in many rivers worldwide, and variation trends are different mainly due to runoff regimes, sediment transport, vegetation colonization, and so on [18,19]. For example, a marked degradation of bars in the lower Trinity River (USA) following the Livingston Dam construction was corroborated, a phenomenon precipitated by a considerable diminution in sediment load [20]. In addition, the vegetation growth in the Sauce Grande River (Argentina) subsequent to damming has been facilitated by the conservation of base flow levels and the diminution of flood-induced disruptions [21]. Furthermore, the downstream environmental perturbation resulting from the Balbina Dam (Brazil) spans in excess of 100 km, with vegetation experiencing mortality up to two decades following its establishment, attributable to the repercussions of hydropower plants on hydrological regimes [22]. Recently, the morphodynamic evolution of large bars in the Yangtze River and their responses to anthropogenic interferences have been investigated [23,24]. For example, the bar surface area in the Jianli reach exhibited greater dynamism under natural flow (1985–2003) as opposed to the regulated flow (2003–2012) [16]; the sandbars in the Yichang-Chenglingji reach were found to have primarily shrank due to the reduction in suspended sediment concentration following the TGD operation [25]; the river bar erosion and deposition patterns of the Shashi reach transitioned from “low bar erosion-high bar deposition” to “both low and high bar erosion” since the TGD damming [26,27]. Nevertheless, prior research is characterized by three principal limitations: (i) an emphasis on specific reaches or river bars rather than extensive spatiotemporal investigations; (ii) a focus on bar surface morphology without considering biological adjustments; (iii) a relative scarcity of information regarding the causality of river bar dynamics. However, the exploration of the variation principle concerning vegetation parameters, specifically the leaf area index (LAI), remains scant in extensive dammed river systems, despite its significance in hydrological, ecological, and geomorphic numerical simulations [28,29].
The advent of Earth observation satellites has enhanced the practicability of monitoring large-scale and protracted river channel transformations in remote regions where in situ measurements are scarce or entirely lacking [1]. LAI is a critical biophysical parameter for monitoring vegetation dynamics, which is defined as half of the total green leaf area per unit of ground area [30]. There are other indices or parameters that can be used for the objective of this study. For example, a previous study directly used the normalized difference vegetation index (NDVI) to monitor vegetation dynamics [9], where NDVI is just a mathematical combination of two reflectance bands. Although NDVI is highly related to vegetation growth status, it is rarely used for quantitative analyese because of lacking physical meaning. Another parameter is gross primary production (GPP), which is the total amount of carbon dioxide fixed by land plants per unit of time through photosynthesis. The GPP can also be used to monitor vegetation dynamics and it can be estimated from the satellite products [31,32], but GPP is relevant to vegetation biochemical processes, including photosynthesis and respiration.
In this study, we chose to use LAI because the physical interaction between water and vegetation was our focus. With the advance of Earth observation systems over the past decades, remote-sensing satellites provide a promising way to generate long-term LAI products for large-scale missions [33]. Currently, moderate-resolution LAI products have been applied for various studies, including global vegetation change and improving simulations of land surface models [34,35]. However, high-resolution LAI products are still required to resolve existing heterogeneity in hydrological and agricultural applications [36]. Current studies have used Landsat or Sentinel imagery to produce high-resolution LAI based on two types of methods: empirical and radiative transfer model (RTM)-based methods [37,38]. The empirical method requires in situ LAI samples, which are sometimes not available for long-term research, and the empirical relationships cannot be transferred over space and time [33]. Therefore, the RTM-based method is popular, especially when it is integrated with a powerful machine learning (ML) algorithm, which can build non-linear relationships between satellite reflectance and LAI well [39]. Another issue is that the inherent discontinuity of spatial and temporal coverage of high-resolution satellite input owing to revisit cycle and cloud contaminations limits the production of high-quality LAI products at a fine spatial resolution [40]. To address this issue, imagery compositing is proposed to combine several observations within a month or season to generate spatiotemporally seamless remote-sensing products [41,42,43]. Hence, we attempt to generate spatiotemporally seamless imagery and apply it to the RTM-based method with an ML algorithm so that the high-resolution LAI products can be gained and adopted for river bar and vegetation dynamics analysis.
Here, we aim to investigate river bar and vegetation dynamics in response to upstream damming using the case of the Middle Yangtze River (MYR) and the Three Gorges Dam (TGD). Specifically, we employ high-resolution river cross-sectional profiles, hydrological datasets, and remote-sensing images for the MYR covering both the pre- and post-TGD periods. The precise aims of this investigation include: (i) examining the spatiotemporal patterns in river bar and vegetation-related attributes, and (ii) exploring the underlying mechanisms governing vegetation alterations, with a particular emphasis on morphological dynamics and modifications in flow duration.

2. Study Area

The Yangtze River, ranked as the world’s third-longest river with an approximate length of 6300 km, is commonly segregated into upper, middle, and lower reaches, delineated according to distinct hydrological attributes and geographical settings [44]. The MYR encompasses a ~955 km stretch between the Yichang and Hukou reaches, situated immediately downstream of the TGD, and incorporates floodplain-type lakes, such as Dongting and Poyang [45] (Figure 1). The regional climate of the MYR is characterized by a subtropical monsoon climate, with annual precipitation levels ranging between 1000 and 1600 mm and an average annual temperature of 16–18 °C [46]. The topography in the MYR is predominantly composed of mountains, hills, and plains. The highest elevation is situated in the mountains of western Hubei, while the lowest is found within the central plains, with mountains and hills constituting over 50% of the landscape. The terrain exhibits a west-to-east gradient, transitioning from higher to lower elevations, with an average altitude of approximately 1497 m. The observed flora encompasses 36 species of annual herbs (e.g., Polypogon fugax), 34 varieties of perennial herbs (e.g., Phragmites communis), 4 types of trees (e.g., Pterocarya stenoptera), and 3 classifications of vines (e.g., Cayratia japonica) [47]. The MYR predominantly exhibits a meandering channel configuration, characterized by numerous bars and a riverbed composition that progressively transitions to finer grain sizes in the downstream direction [48].
Based on local geographical contexts and channel patterns, nine sub-reaches were demarcated, including Yichang-Zhicheng, Zhicheng-Majiadian, Majiadian-Shashi, Shashi-Shishou, Shishou-Jianli, Jianli-Chenglingji, Chenglingji-Hankou, Hankou-Jiujiang, and Jiujiang-Hukou (Figure 1). Sediment and flow fluxes within each sub-reach were assessed at 10 hydrometric stations. For the analysis, 16 relatively large bars (with an area exceeding 1 km2) were selected (Figure 1, Table 1 and Table 2). We used ArcGIS Pro to generate maps in this study (Esri Inc., Redlands, CA, USA, (2022). ArcGIS Pro (Version 3.0.2)).

3. Materials and Methods

3.1. Data Sources

All original raw data utilized in this study (excluding Landsat images) were procured from the Changjiang Water Resources Commission (CWRC; http://www.cjw.gov.cn/, accessed on 1 March 2023), encompassing hydrological and sediment data (e.g., daily mean discharge, water levels, and sediment concentrations at multiple hydrometric stations; Figure 1), as well as topographical data (i.e., annual one-dimensional cross-sectional profiles at 450 specific locations, measured using a multi-beam echo-sounder). Prior to public release, these datasets undergo stringent verification and uncertainty analysis in accordance with governmental protocols, such as the protocol for liquid flow measurement in open channels, the protocol for measurements of suspended sediment in open channels, and the protocol for port and waterway engineering survey, as published by the Ministry of Water Resources, China [9] (Table 3).
The satellite data employed in this research were Landsat 5, 7, and 8 surface reflectance products, which were obtained from Google Earth Engine (GEE, https://developers.google.com/earth-engine/datasets/catalog/landsat, accessed on 1 March 2023), with the aim of monitoring vegetation dynamics for 16 bars spanning the period of 2003 to 2020 (Table 3). Additionally, GEE is a cloud computing platform that provides access to an enormous amount of geospatial data and computing resources. It allows users to process and analyze petabytes of satellite imagery and other geospatial data, making it possible to perform large-scale analyses that were previously impossible due to computational constraints. In this study, we downloaded satellite images from GEE and performed some processes on them (please refer to Section 3.2.1).

3.2. Machine-Learning-Based Leaf Area Index Inversion Model

3.2.1. Preprocessing of Satellite Imagery

Before inversing LAI from remote-sensing reflectance imagery, there are several steps need to be implemented. The first step is using the maximum value compositing to generate annual satellite imagery over the dry season since Landsat satellites have a 16-day revisit cycle, which means vegetation can only be observed every 16 days and the condition could be worse when cloud contaminates the Landsat imagery, such as thick clouds blocking vegetation information [40]. Specifically, we computed the normalized difference vegetation index (NDVI):
N D V I = N I R − R N I R + R
where NIR and R are near-infrared and red reflectance bands, respectively. The NDVI is a good indicator for monitoring vegetation dynamics [49,50], and large NDVI values reflect strong vegetation signals. We selected the pixel along with the maximum NDVI value as the composited value from several candidate pixels from October to April every year. Note that images acquired from 2011–2013 were not used in this study because these images suffered many errors, including geo-registration errors and cloud contamination issues. Then, water masks were generated by Landsat water index (WI2015) [51]:
W I 2015 = 1.7204 + 171 G + 3 R − 70 N I R − 45 S W I R 1 + 71 S W I R 2
where G, SWIR1, and SWIR2 are the green band and shortwave infrared bands 1 and 2, respectively. Normally, a remote-sensing index is designed to divide remote-sensing imagery into two categories. In this study, we used the NDVI and WI2015 to identify vegetation and non-vegetation and water and non-water, respectively. Large WI2015 values reflect strong water signals. Therefore, the threshold of WI2015 is set to 0.5 to mask out water for monitoring bar vegetation dynamics. Subsequently, the threshold of the NDVI was set to 0.3 to identify vegetation. In detail, a pixel with a WI2015 value greater than 0.5 will be classified as water; otherwise, it is non-water. Note that a pixel with an NDVI value greater than 0.3 will be classified as vegetation; otherwise, it is non-vegetation.

3.2.2. ML-Based LAI Inversion Model

A famous radiative transfer model, the PROSPECT + SAIL (PROSAIL) model, is used to simulate samples with both LAI and reflectance [52,53]. The parameters of PROSAIL are shown in Table 4. Other parameters were set to the default values for all simulations: leaf structure parameter (1.75), equivalent water thickness (0.015 cm), and leaf mass per area (0.0075 g/cm2). After setting all the parameters, the PROSAIL model was used to generate 100,000 samples.
We chose the random forest regressor (RFR) as the ML algorithm, which is an ensemble method and can model non-linear relationships between predictors and responses [54]. Random forest (RF) has been used as a classifier in remote sensing, but it is also capable of estimating vegetation parameters [55]. Compared to other ML methods, training a RF model is relatively less time-consuming and this model is robust to outliers [56]. N_estimator, max_depth, and min_samples were set to 100, 10, and 2, respectively. The flowchart of the ML-based LAI inversion model is shown in Figure 2. Generally, RTM is used to produce synthetic reflectance and NDVI with different LAIs and other vegetation parameters as input and the RF is trained based on this synthetic dataset. Then the trained RF predicts satellite LAI products with real satellite reflectance and the NDVI as input. All the processes were implemented based on Python with numpy and Scikit-learn packages. The PROSAIL source code is publicly available (http://teledetection.ipgp.jussieu.fr/prosail/, accessed on 1 March 2023).

3.3. Sediment Balance Method

The cross-sectional topography method is widely used to investigate geomorphic changes in channels and can, therefore, be applied to calculate the amount of erosion and deposition [26]. The calculation equation (Equation (2)) is expressed as follows:
V = ∑ i = 1 n − 1 1 3 ∆ A i + ∆ A i + 1 + ∆ A i ∆ A i + 1 × ∆ L i
where V is the erosion (+) or deposition (−) volume, m3; A i is the i th cross-sectional area, m2; ∆ A i is the area change at the i th cross-section, m2; ∆ L i is the distance between the i th and ( i + 1 ) th cross-sections, m2; n is the number of cross-sections (51, 21, 33, 52, 34, 36, 92, 124, and 15 in each sub-reach).

4. Results

Subsequent to the TGD operation, a multitude of bars within the MYR have experienced varying degrees of transformation. These alterations predominantly manifest in the modification of the spatial extent and geometric characteristics of the bars, shifts in vegetation coverage, and fluctuations in LAI pertaining to the bars. In the MYR, the channel distances are relatively extensive, resulting in temporal and spatial variations in the bar and vegetation areas across distinct sub-reaches. A systematic analysis of the nine sub-reaches in the MYR from 2003 to 2020 is conducted individually.

4.1. Variations in River Bar and Vegetation Areas

Figure 3 shows the variations in bar and vegetation areas in the MYR after the operation of the TGD. These results indicate that the bar area in the MYR experienced an overall modest decline (approximately 10%), while the vegetated area witnessed a substantial increase (roughly 66%), and the vegetation coverage on the bar notably surged (by around 98%). Owing to the diverse geological conditions present within the sub-reaches, the dynamics in bar characteristics and vegetation exhibit considerable disparities across the varying sub-reaches. Specifically, with respect to the bar area, vegetation area, and vegetation coverage (i.e., the ratio of vegetation and bar area), sub-reaches 1–9 have exhibited significant changes post-TGD operation (Table 5).

4.2. Variations in Vegetation Distribution and Leaf Area Index

The alterations in vegetation distribution characteristics and LAI provide a more detailed representation of the impacts that bar and vegetation dynamics have on the river channel. Utilizing the ML-based LAI inversion model, variations in vegetation distribution and LAI in the MYR after the TGD operation are depicted in Figure 4 and Figure 5.
Figure 4 shows the fluctuations in LAI within the MYR subsequent to the commencement of the TGD operation. The results reveal an overall 24% augmentation in LAI. More specifically, the LAI in sub-reaches 1–9 have undergone changes of +79%, +15%, +66%, +18%, +7%, +7%, −3%, +34%, and −3%, respectively. Despite sub-reach 1 having the least dense bar vegetation (LAI = 0.60), its LAI growth rate is the most substantial among all the reaches (+79%). Simultaneously, sub-reach 6 possesses the densest bar vegetation (LAI = 1.40), exhibiting only a slight LAI growth rate following the TGD operation (+7%). Furthermore, while the bar vegetation in most sub-reaches typically displays a trend towards density, the bar vegetation in sub-reaches 7 and 9 exhibits a relatively sparse trend, with LAI growth rates both equal to −3%.
Figure 5 displays the alterations in LAI distribution for four representative bars within the MYR after the TGD operation. These four representative bars are selected due to two main reasons. First, the four bars chosen are situated hundreds of kilometers apart, as the downstream influence of water and sediment conditions stemming from the dam exhibits a discernible spatial impact on the riverine geomorphology. Second, since channel type can influence the bar evolution process, the four bars are located in different channel types covering straight, braided, and meandering channels. Obviously, most vegetations have exhibited a tendency to expand and become denser, irrespective of various river and bar types. In terms of vegetation zone morphology, a vegetation expansion transpired from the tail to the head, as well as from the edge to the center, while maintaining a slight shrinkage of the bar area.

5. Discussion

The driving factors behind the dynamics of bars and vegetation in the MYR are multifaceted; however, the primary causes can be attributed to adjustments in hydrology and deposition processes within the fluvial channel. These can be further delineated into specific factors, such as the alteration of the annual flow process, the frequency of bar submergence, adjustments in channel and bar deposition, modifications in the transverse and longitudinal sections of the channel, and shifts in the river’s hydraulic geometry. Despite the advancements made in prior research, studies have predominantly concentrated on the analysis of singular factors. For instance, variations in the river bar area in the Shashi reach were elucidated by [26] through changes in erosion amplitude at distinct discharges, while [18] demonstrated that flow conditions governed river bar area fluctuations in the Jingjiang reach. In contrast, our work offers a holistic causal analysis, examining the interplay between hydrodynamic and geomorphic transformations in the MYR following the TGD operation.

5.1. Hydrodynamic Changes

5.1.1. Alteration in Annual Flow Process

Here, years featuring flood discharge within distinct time series (1998, 2007, and 2012) are chosen to construct the annual discharge curve. Figure 6 depicts the annual discharge curves for hydrometric stations situated along the MYR. The results indicate that the standard deviation of the average annual discharge in the MYR is 12,704 m3/s. Following the impingement of the TGD, the annual discharge curve became flatter (the standard deviation of the annual discharge decreased by approximately 37%), and the number of days with overbank flows progressively diminished (from 83 days in 1998 to 40 days in 2012, a reduction of 52%). It is noteworthy that the differences in discharge standard deviation reduction among various hydrometric stations are relatively small. For instance, the most modest reduction is observed at Station 6 (−36%), while the most substantial reduction takes place at Station 4 (−39%). Consequently, after the TGD operation, the less fluctuating flow process in the MYR may diminish the impact on the bar to some degree, thereby offering opportunities for the development and growth of the bar and its vegetation.

5.1.2. Alteration in Bar Submergence Frequency

The variation pattern of the annual discharge process in typical years is relatively consistent, with the phenomenon of the discharge process flattening occurring in all instances. Nevertheless, to assess the influence of flow on the interaction between channel and bars, we select frequency changes of overbank flows and flood discharge at stations along the MYR in different time series. This approach helps to minimize any cognitive bias that may arise due to the typical selection of years in the previous section.
Figure 7 illustrates the alterations in bar submergence frequency and flood occurrence probability for six hydrometric stations along the MYR during pre- and post-TGD periods. The findings reveal that both the bar submergence frequency and flood occurrence probability at the stations in the MYR have declined following the TGD operation.
Overall, from period 1 to period 3 (Table 1), the bar submergence frequency at hydrometric stations in the MYR declined by 49–62%, while the flood occurrence probability dropped by 69–76%. Bar submergence frequency and flood occurrence probability decreased as follows: at Station 1, by 51% and 69%; at Station 2, by 55% and 74%; at Station 4, by 54% and 74%; at Station 6, by 49% and 73%; at Station 8, by 61% and 76%; at Station 9, by 62% and 74%, respectively.
Generally, when the water depth reaches approximately 1.5 m, it is unfavorable for the growth of bar vegetation [57,58,59]. Here, based on the relationship between water level and discharge, the corresponding water depth of flood discharge chosen by each hydrometric station is roughly 2 m higher than that of floodplain discharge (approximately 0.5 m higher than the water depth threshold suitable for vegetation growth). As illustrated in [9] (with Yichang Station and Luoshan Station as examples), we believe that the occurrence of flood flow is detrimental to vegetation growth. Following the TGD operation, the frequency of floods decreases at all stations along the MYR (with a concurrent reduction in bar submergence frequency), which will provide more favorable conditions for the growth of vegetation in elevated bar areas.

5.2. Geomorphic Changes

5.2.1. Adjustment of Bar and Channel Erosion Distribution

The entire MYR is partitioned into three sub-reaches, namely the Yichang-Chenglingji reach (sub-reaches 1–6; 408 km), the Chenglingji-Hankou reach (sub-reach 7; 251 km), and the Hankou-Hukou reach (sub-reaches 8–9; 295 km), based on the river length and the relationship between the main stream and the lake. This division is employed to calculate the distribution of erosion and deposition within these sub-reaches. Figure 8 shows the annual bar and channel deposition in the MYR from 2003 to 2020. The results indicate that, in the post-TGD period, a more intense channel scouring was observed along the MYR, accompanied by slight yet relatively stable bar scouring (channel scouring being approximately 9–14 times more intense than bar scouring). The strongest scouring occurred in the sub-reach proximate to the dam, where the scouring intensity above the Chenglingji section was about three times greater than that in the section below Chenglingji.
Overall, the erosion of bars and channels in the MYR displays relatively distinct spatial and temporal variability. Specifically, the annual channel erosion in Yichang-Chenglingji reach (7328 × 104 m3/y), Chenglingji-Hankou reach (2864 × 104 m3/y), and Hankou-Hukou reach (3993 × 104 m3/y) are significantly greater than the corresponding bar erosion (784 × 104 m3/y; 299 × 104 m3/y; 286 × 104 m3/y). However, the relatively pronounced channel erosion downstream of the Chenglingji reach occurred with a lag of approximately 10 years compared to the upstream Chenglingji reach.

5.2.2. Adjustment of Cross-Sectional and Longitudinal Profiles

Previous studies have comprehensively examined the adjustments of cross-sectional and longitudinal profiles in the MYR during the post-TGD period, drawing upon in situ data [60]. For instance, Ref. [9] has shown the longitudinal profiles in sub-reaches 1, 2, 3, 7, and 8, along with the cross-sectional profiles encompassing bars 2, 3, 5, 6, 7, 8, 12, and 15 from 2004 to 2015. The findings reveal that the intense scouring induced by unsaturated flow following the TGD operation has primarily influenced the longitudinal profile of the channel, rendering it more undulating. However, the impact on bars has been more restrained, with the elevation of most bars remaining stable.

5.2.3. Adjustment of Hydraulic Geometry

Figure 9 presents statistics illustrating the changes in water depth, river width, cross-section area, and river regime coefficient after the TGD operation for each sub-reach in the MYR under a given bank-full discharge at Station 1. The results indicate that there is an increase in cross-sectional area (by ~19%; from 20,678 m2 to 24,608 m2), water depth (by ~11%; from 13.5 m to 15.0 m), and river width (by ~7%; from 1516 m to 1629 m). Additionally, the channel becomes deepened and narrower (with a ~7% reduction in the river regime coefficient; from 2.91 m−1/2 to 2.70 m−1/2).
Viewed along the MYR, the reduction in the river regime coefficient is greater in the near-TGD reach (~15% in sub-reaches 2–6) than in the far dam reach (~5–8% in sub-reaches 8–9). Concurrently, the increase in water depth in sub-reaches 2–6 is relatively obvious (~20%), while the increase in the downstream reach near Hukou station is more modest (~10%), with the increase in sub-reach 1 being only 8%. The primary reason for the minor depth increase in sub-reach 1 may be attributed to its riverbed being composed of gravel and cobble, which is more resistant to scouring compared to the downstream sandy reach. Consequently, the riverbed can easily form a protective layer due to roughening, resulting in a smaller erosion magnitude.
In general, although the MYR exhibits channel undercutting and bank widening throughout its course, there are subtle variations along the way. These include relatively minor changes in channel morphology in sub-reach 1, an increase in water depth in sub-reaches 2–6, and an increase in river width in sub-reaches 7–9.

5.3. Broader Implications

River bars have always been at the forefront of biogeomorphological research [17,59]. Our study contributes to the enhancement of quantitative insights into biogeomorphic interactions, fostering interdisciplinary engagement in the process [61] as the MYR, a non-equilibrium river undergoing strong anthropogenic disturbance, provides an ideal case for disentangling the complex bidirectional biogeomorphic interactions in fluvial and riparian systems, owing to the detailed in situ observations [13,62,63,64]. Simultaneously, our enhanced ML-based LAI inversion model is employed to generate spatiotemporally continuous 30 m resolution LAI products, thereby addressing the deficiency of biophysical vegetation parameters in riverine ecosystems.
Compared to current moderate-resolution (MODIS) LAI products [65], Landsat LAI products generated in this study can accurately capture bar LAI heterogeneity in the MYR (Figure 10). At the same time, these current moderate-resolution LAI products sometimes will misclassify bars into water owing to the mismatch between coarse resolution and small bar areas, resulting in no LAI values in these small bars and thus limiting further analyses. Generally, the proposed framework that combines maximum-value composites, RTM simulations, and ML can easily be transferred to predict long-term LAIs at other bars owing to the high generalization of RTM and ML as well as no requirement of in situ LAI measurements.
There are several advantages of the LAI estimation method mentioned above; however, there is still room to improve LAI estimations for monitoring bar vegetation dynamics. Firstly, maximum-value composites trade temporal resolution for spatial continuity. When we want to observe bar vegetation activity at both high and temporal resolutions, some advanced sensors or algorithms can be utilized, such as Sentinel-2 and PlanetScope imagery and the spatiotemporal fusion methods [66,67]. Moreover, the proposed framework does not require in situ LAI measurements, but these measurements are still necessary if it is possible to be collected, which can be used as a calibration for producing accurate LAI products [33].

6. Conclusions

Examining river bar and vegetation dynamics in response to upstream damming is crucial for riverine flood management and ecological assessment. In this study, we evaluate the hypothesis that intensified erosion and attenuated discharge pulses downstream of dams contribute to the expansion of river bars and vegetation. Our investigation employs a case study from the Middle Yangtze River (MYR) in China to substantiate this hypothesis. We find that the bar area diminished by ~10% due to minor bar erosion, whereas the vegetation area expanded by ~66% in a from-tail-to-head and from-edge-to-center trend, which can be attributed to the reduced frequency of bar submergence. Further analysis revealed that the diminished bar submergence frequency is resulting from the flatter discharge pulses and larger bar-channel elevation gap, which is caused by the approximately tenfold discrepancy in erosion volume between the channel and bar. In addition, we have improved the machine learning-based leaf area index (LAI) inversion model, identifying a ~24% increase in LAI after the TGD operation.
Our preliminary investigation, situated at the intersection of hydrology, geomorphology, biology, and remote sensing, was effectively executed in the MYR to examine river bar and vegetation dynamics in response to the construction of the TGD. This approach furnishes an accessible methodology (ML-based LAI inversion model) and conceptual framework for causal analysis when addressing the biogeomorphological impacts of large dams. Our findings carry significant implications for flood management and ecological restoration in dammed river systems.

Author Contributions

Conceptualization, Y.H., D.L., J.D. and Y.L.; methodology, Y.H. and J.Z.; validation, Y.H. and J.Z.; formal analysis, Y.H.; investigation, Y.H. and C.Y.; data curation, Y.H. and J.Z.; writing—original draft preparation, Y.H.; writing—review and editing, D.L., J.Z., J.D., C.Y. and Y.L.; visualization, Y.H. and J.Z.; supervision, J.D.; project administration, J.D.; funding acquisition, J.D. All authors have read and agreed to the published version of the manuscript.

Funding

This work was funded by the National Natural Science Foundation of China (Grant No. 51779185), and the National Key R&D Program of China (Grant No. 2018YFC0407201).

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Wu, Q.; Ke, L.; Wang, J.; Pavelsky, T.M.; Allen, G.H.; Sheng, Y.; Duan, X.; Zhu, Y.; Wu, J.; Wang, L.; et al. Satellites reveal hotspots of global river extent change. Nat. Commun. 2023, 14, 1587. [Google Scholar] [CrossRef]
  2. Marcarelli, A.M.; Kirk, R.W.V.; Baxter, C.V. Predicting effects of hydrologic alteration and climate change on ecosystem metabolism in a western US river. Ecol. Appl. 2010, 20, 2081–2088. [Google Scholar] [CrossRef] [PubMed]
  3. Siam, M.S.; Eltahir, E.A. Climate change enhances interannual variability of the Nile river flow. Nat. Clim. Chang. 2017, 7, 350–354. [Google Scholar] [CrossRef]
  4. Chen, W.; Olden, J.D. Designing flows to resolve human and environmental water needs in a dam-regulated river. Nat. Commun. 2017, 8, 2158. [Google Scholar] [CrossRef] [PubMed]
  5. Li, D.; Lu, X.; Walling, D.E.; Zhang, T.; Steiner, J.F.; Wasson, R.J.; Harrison, S.; Nepal, S.; Nie, Y.; Immerzeel, W.W.; et al. High Mountain Asia hydropower systems threatened by climate-driven landscape instability. Nat. Geosci. 2022, 15, 520–530. [Google Scholar] [CrossRef]
  6. Ma, H.; Nittrouer, J.A.; Fu, X.; Parker, G.; Zhang, Y.; Wang, Y.; Wang, Y.; Lamb, M.P.; Cisneros, J.; Best, J.; et al. Amplification of downstream flood stage due to damming of fine-grained rivers. Nat. Commun. 2022, 13, 3054. [Google Scholar] [CrossRef] [PubMed]
  7. Li, D.; Lu, X.X.; Yang, X.; Chen, L.; Lin, L. Sediment load responses to climate variation and cascade reservoirs in the Yangtze River: A case study of the Jinsha River. Geomorphology 2018, 322, 41–52. [Google Scholar] [CrossRef]
  8. Zhou, Y.; Li, D.; Lu, J.; Yao, S.; Yan, X.; Jin, Z.; Liu, L.; Lu, X.X. Distinguishing the multiple controls on the decreased sediment flux in the Jialing River basin of the Yangtze River, Southwestern China. Catena 2020, 193, 104593. [Google Scholar] [CrossRef]
  9. Hu, Y.; Li, D.; Deng, J.; Yue, Y.; Zhou, J.; Chai, Y.; Li, Y. Mechanisms Controlling Water-Level Variations in the Middle Yangtze River Following the Operation of the Three Gorges Dam. Water Resour. Res. 2022, 58, e2022WR032338. [Google Scholar] [CrossRef]
  10. Best, J. Anthropogenic stresses on the world’s big rivers. Nat. Geosci. 2019, 12, 7–21. [Google Scholar] [CrossRef]
  11. Syvitski, J.; Ángel, J.R.; Saito, Y.; Overeem, I.; Vörösmarty, C.J.; Wang, H.; Olago, D. Earth’s sediment cycle during the Anthropocene. Nat. Rev. Earth Environ. 2022, 3, 179–196. [Google Scholar] [CrossRef]
  12. Hu, Y.; Deng, J.; Li, Y.; Liu, C.; He, Z. Flow resistance adjustments of channel and bars in the middle reaches of the Yangtze River in response to the operation of the Three Gorges Dam. J. Geogr. Sci. 2022, 32, 2013–2035. [Google Scholar] [CrossRef]
  13. Viles, H.A.; Naylor, L.A.; Carter, N.E.A.; Chaput, D. Biogeomorphological disturbance regimes: Progress in linking ecological and geomorphological systems. Earth Sur. Process. Landf. 2008, 33, 1419–1435. [Google Scholar] [CrossRef]
  14. Chai, Y.; Yang, Y.; Deng, J.; Sun, Z.; Li, Y.; Zhu, L. Evolution characteristics and drivers of the water level at an identical discharge in the Jingjiang reaches of the Yangtze River. J. Geogr. Sci. 2020, 30, 1633–1648. [Google Scholar] [CrossRef]
  15. Gierszewski, P.J.; Habel, M.; Szmańda, J.; Luc, M. Evaluating effects of dam operation on flow regimes and riverbed adaptation to those changes. Sci. Total Environ. 2020, 710, 136202. [Google Scholar] [CrossRef]
  16. Long, J.; Li, H.; Wang, Z.; Wang, B.; Xu, Y. Three decadal morphodynamic evolution of a large channel bar in the middle Yangtze River: Influence of natural and anthropogenic interferences. Catena 2021, 199, 105128. [Google Scholar] [CrossRef]
  17. Wohl, E.; Marshall, A.E.; Scamardo, J.; White, D.; Morrison, R.R. Biogeomorphic influences on river corridor resilience to wildfire disturbances in a mountain stream of the Southern Rockies, USA. Sci. Total Environ. 2022, 820, 153321. [Google Scholar] [CrossRef]
  18. Zhu, G.; Li, Y.; Sun, Z.; Kanae, S. Response of vegetation to submergence along Jingjiang Reach of the Yangtze River. PLoS ONE 2021, 16, e0251015. [Google Scholar] [CrossRef]
  19. Lou, Y.; Dai, Z.; Lu, X.; Li, D. Anthropogenic pressures induced hydromorphodynamic changes of riverine islands in the Upper Jingjiang reach along the Changjiang (Yangtze) River. Catena 2022, 217, 106488. [Google Scholar] [CrossRef]
  20. Phillips, J.D.; Slattery, M.C.; Musselman, Z.A. Channel adjustments of the lower Trinity River, Texas, downstream of Livingston Dam. Earth Sur. Process. Landf. 2005, 30, 1419–1439. [Google Scholar] [CrossRef]
  21. Casado, A.; Peiry, J.L.; Campo, A.M. Geomorphic and vegetation changes in a meandering dryland river regulated by a large dam, Sauce Grande River, Argentina. Geomorphology 2016, 268, 21–34. [Google Scholar] [CrossRef]
  22. Assahira, C.; Piedade, M.T.F.; Trumbore, S.E.; Wittmann, F.; Cintra, B.B.L.; Batista, E.S.; Resende, A.F.; Schöngart, J. Tree mortality of a flood-adapted species in response of hydrographic changes caused by an Amazonian river dam. For. Ecol. Manag. 2017, 396, 113–123. [Google Scholar] [CrossRef]
  23. Lou, Y.; Mei, X.; Dai, Z.; Wang, J.; Wei, W. Evolution of the mid-channel bars in the middle and lower reaches of the Changjiang (Yangtze) River from 1989 to 2014 based on the Landsat satellite images: Impact of the Three Gorges Dam. Environ. Earth Sci. 2018, 77, 1–18. [Google Scholar] [CrossRef]
  24. Xia, J.; Wang, Y.; Zhou, M.; Deng, S.; Li, Z.; Wang, Z. Variations in channel centerline migration rate and intensity of a braided reach in the Lower Yellow River. Remote Sens. 2021, 13, 1680. [Google Scholar] [CrossRef]
  25. Wang, J.; Dai, Z.; Mei, X.; Lou, Y.; Wei, W.; Ge, Z. Immediately downstream effects of Three Gorges Dam on channel sandbars morphodynamics between Yichang-Chenglingji Reach of the Changjiang River, China. J. Geogr. Sci. 2018, 28, 629–646. [Google Scholar] [CrossRef]
  26. Yang, Y.; Zheng, J.; Zhang, M.; Zhu, L.; Zhu, Y.; Wang, J.; Zhao, W. Sandy riverbed shoal under anthropogenic activities: The sandy reach of the Yangtze River, China. J. Hydrol. 2021, 603, 126861. [Google Scholar] [CrossRef]
  27. Yang, Y.; Zheng, J.; Zhang, H.; Chai, Y.; Zhu, Y.; Wang, C. Impact of the Three Gorges Dam on riverbed scour and siltation of the middle reaches of the Yangtze River. Earth Sur. Process. Landf. 2022, 47, 1514–1531. [Google Scholar] [CrossRef]
  28. Zhang, Y.; Wang, P.; Cheng, J.; Wang, W.J.; Zeng, L.; Wang, B. Drag coefficient of emergent flexible vegetation in steady nonuniform flow. Water Resour. Res. 2020, 56, e2020WR027613. [Google Scholar] [CrossRef]
  29. Box, W.; Järvelä, J.; Västilä, K. Flow resistance of floodplain vegetation mixtures for modelling river flows. J. Hydrol. 2021, 601, 126593. [Google Scholar] [CrossRef]
  30. Chen, J.M.; Black, T.A. Defining leaf area index for non-flat leaves. Plant Cell Environ. 1992, 15, 421–429. [Google Scholar] [CrossRef]
  31. Jiang, C.; Guan, K.; Wu, G.; Peng, B.; Wang, S. A daily, 250 m and real-time gross primary productivity product (2000–present) covering the contiguous United States. Earth Syst. Sci. Data. 2021, 13, 281–298. [Google Scholar] [CrossRef]
  32. Yang, P.; Prikaziuk, E.; Verhoef, W.; van Der Tol, C. SCOPE 2.0: A model to simulate vegetated land surface fluxes and satellite signals. Geosci. Model Dev. 2021, 14, 4697–4712. [Google Scholar] [CrossRef]
  33. Fang, H.; Baret, F.; Plummer, S.; Schaepman-Strub, G. An overview of global leaf area index (LAI): Methods, products, validation, and applications. Rev. Geophys. 2019, 57, 739–799. [Google Scholar] [CrossRef]
  34. Piao, S.; Yin, G.; Tan, J.; Cheng, L.; Huang, M.; Li, Y.; Liu, R.; Mao, J.; Myneni, R.B.; Peng, S.; et al. Detection and attribution of vegetation greening trend in China over the last 30 years. Glob. Chang. Biol. 2015, 21, 1601–1609. [Google Scholar] [CrossRef] [PubMed]
  35. Albergel, C.; Munier, S.; Leroux, D.J.; Dewaele, H.; Fairbairn, D.; Barbu, A.L.; Gelati, E.; Dorigo, W.; Faroux, S.; Meurey, C.; et al. Sequential assimilation of satellite-derived vegetation and soil moisture products using SURFEX_v8. 0: LDAS-Monde assessment over the Euro-Mediterranean area. Geosci. Model Dev. 2017, 10, 3889–3912. [Google Scholar] [CrossRef]
  36. Kang, Y.; Ozdogan, M.; Gao, F.; Anderson, M.C.; White, W.A.; Yang, Y.; Yang, Y.; Erickson, T.A. A data-driven approach to estimate leaf area index for Landsat images over the contiguous US. Remote Sens. Environ. 2021, 258, 112383. [Google Scholar] [CrossRef]
  37. Delegido, J.; Verrelst, J.; Alonso, L.; Moreno, J. Evaluation of sentinel-2 red-edge bands for empirical estimation of green LAI and chlorophyll content. Sensors 2011, 11, 7063–7081. [Google Scholar] [CrossRef]
  38. Liu, J.; Pattey, E.; Jégo, G. Assessment of vegetation indices for regional crop green LAI estimation from Landsat images over multiple growing seasons. Remote Sens. Environ. 2012, 123, 347–358. [Google Scholar] [CrossRef]
  39. Verrelst, J.; Rivera, J.P.; Veroustraete, F.; Muñoz-Marí, J.; Clevers, J.G.; Camps-Valls, G.; Moreno, J. Experimental Sentinel-2 LAI estimation using parametric, non-parametric and physical retrieval methods–A comparison. ISPRS J. Photogramm. Remote Sens. 2015, 108, 260–272. [Google Scholar] [CrossRef]
  40. Ju, J.; Roy, D.P. The availability of cloud-free Landsat ETM+ data over the conterminous United States and globally. Remote Sens. Environ. 2008, 112, 1196–1211. [Google Scholar] [CrossRef]
  41. Holben, B.N. Characteristics of maximum-value composite images from temporal AVHRR data. Int. J. Remote Sens. 1986, 7, 1417–1434. [Google Scholar] [CrossRef]
  42. Weng, Q.; Fu, P. Modeling annual parameters of clear-sky land surface temperature variations and evaluating the impact of cloud cover using time series of Landsat TIR data. Remote Sens. Environ. 2014, 140, 267–278. [Google Scholar] [CrossRef]
  43. Jin, Z.; Azzari, G.; You, C.; Di Tommaso, S.; Aston, S.; Burke, M.; Lobell, D.B. Smallholder maize area and yield mapping at national scales with Google Earth Engine. Remote Sens. Environ. 2019, 228, 115–128. [Google Scholar] [CrossRef]
  44. Cao, G.J.; Wang, J. Measurements and Studies of Hydrological and Sediment Data in the Three Gorges Project; Science Press: Beijing, China, 2015. [Google Scholar]
  45. Yang, Y.; Li, M.; Liu, W.; Chai, Y.; Zhang, J.; Yu, W. Relationship between potential waterway depth improvement and evolution of the Jingjiang Reach of the Yangtze River in China. J. Geogr. Sci. 2023, 33, 547–575. [Google Scholar] [CrossRef]
  46. Yi, Y.; Hu, X.; Shi, M.; Kang, H.; Wang, B.; Zhang, C.; Liu, C. Vegetation dynamics and its relationship with climate factors in the middle reaches of the Yangtze River based on MODIS NDVI. Sheng Tai Xue Bao 2021, 41, 7796–7807. [Google Scholar]
  47. Zhang, Z.; Li, M.; Yang, Z.; Liu, Q.; Zou, M.; Yuan, Y. Vegetation distribution on the ecological protection slopes of the Middle and Lower Yangtze River. Shui Sheng Tai Xue Za Zhi 2021, 42, 7–15. [Google Scholar]
  48. He, Z.; Sun, Z.; Li, Y.; Zhao, Q.; Hu, Y.; Chen, Z. Response of the gravel–sand transition in the Yangtze River to hydrological and sediment regime changes after upstream damming. Earth Sur. Process. Landf. 2022, 47, 383–398. [Google Scholar] [CrossRef]
  49. Huang, S.; Tang, L.; Hupy, J.P.; Wang, Y.; Shao, G. A commentary review on the use of normalized difference vegetation index (NDVI) in the era of popular remote sensing. J. For. Res. 2021, 32, 1–6. [Google Scholar] [CrossRef]
  50. Zhou, J.; Chen, J.; Chen, X.; Zhu, X.; Qiu, Y.; Song, H.; Rao, Y.; Zhang, C.; Cao, X.; Cui, X. Sensitivity of six typical spatiotemporal fusion methods to different influential factors: A comparative study for a normalized difference vegetation index time series reconstruction. Remote Sens. Environ. 2021, 252, 112130. [Google Scholar] [CrossRef]
  51. Fisher, A.; Flood, N.; Danaher, T. Comparing Landsat water index methods for automated water classification in eastern Australia. Remote Sens. Environ. 2016, 175, 167–182. [Google Scholar] [CrossRef]
  52. Feret, J.B.; François, C.; Asner, G.P.; Gitelson, A.A.; Martin, R.E.; Bidel, L.P.; Ustin, S.L.; Maire, G.; Jacquemoud, S. PROSPECT-4 and 5: Advances in the leaf optical properties model separating photosynthetic pigments. Remote Sens. Environ. 2008, 112, 3030–3043. [Google Scholar] [CrossRef]
  53. Verhoef, W.; Jia, L.; Xiao, Q.; Su, Z. Unified optical-thermal four-stream radiative transfer theory for homogeneous vegetation canopies. IEEE Trans. Geosci. Remote Sens. 2007, 45, 1808–1822. [Google Scholar] [CrossRef]
  54. Breiman, L. Random forests. Mach. Learn. 2001, 45, 5–32. [Google Scholar] [CrossRef]
  55. Belgiu, M.; Drăguţ, L. Random forest in remote sensing: A review of applications and future directions. ISPRS J. Photogramm. Remote Sens. 2016, 114, 24–31. [Google Scholar] [CrossRef]
  56. Wolanin, A.; Camps-Valls, G.; Gómez-Chova, L.; Mateo-García, G.; van der Tol, C.; Zhang, Y.; Guanter, L. Estimating crop primary productivity with Sentinel-2 and Landsat 8 using machine learning methods trained with radiative transfer simulations. Remote Sens. Environ. 2019, 225, 441–457. [Google Scholar] [CrossRef]
  57. Weisner, S.E.; Strand, J.A. Rhizome architecture in Phragmites australis in relation to water depth: Implications for within-plant oxygen transport distances. Folia Geobot. 1996, 31, 91–97. [Google Scholar] [CrossRef]
  58. Xu, X.; Zhang, Q.; Li, Y.; Tan, Z.; Wang, X. Reed population features and relationships between feature indicators and flooding depth, groundwater depth in an islet in Poyang Lake. Wetl. Sci. 2014, 12, 714–722. [Google Scholar]
  59. Xiong, H.; Sun, Z.; Chen, L.; Yang, X.; Zhou, X.; Li, Z.; Zhou, W. Influence of hydrological and morphological changes on the habitats of Miscanthus lutarioriparius and Phragmites australis in a river flood plain: A case study of Wuhan reach of the Yangtze River. J. Lake Sci. 2022, 34, 1250–1261. [Google Scholar]
  60. Yang, X.; Xiong, H.; Li, D.; Li, Y.; Hu, Y. Disproportional erosion of the middle-lower Yangtze River following the operation of the Three Gorges Dam. Sci. Total Environ. 2023, 859, 160264. [Google Scholar] [CrossRef]
  61. Wohl, E.; Knox, R.L. A first-order approximation of floodplain soil organic carbon stocks in a river network: The South Platte River, Colorado, USA as a case study. Sci. Total Environ. 2022, 852, 158507. [Google Scholar] [CrossRef]
  62. Knox, R.L.; Wohl, E.E.; Morrison, R.R. Levees don’t protect, they disconnect: A critical review of how artificial levees impact floodplain functions. Sci. Total Environ. 2022, 837, 155773. [Google Scholar] [CrossRef] [PubMed]
  63. Majhi, I.; Yang, D. Streamflow characteristics and changes in Kolyma Basin in Siberia. J. Hydrometeorol. 2008, 9, 267–279. [Google Scholar] [CrossRef]
  64. Suzuki, K.; Park, H.; Makarieva, O.; Kanamori, H.; Hori, M.; Matsuo, K.; Matsumura, S.; Nesterova, N.; Hiyama, T. Effect of permafrost thawing on discharge of the Kolyma River, northeastern Siberia. Remote Sens. 2021, 13, 4389. [Google Scholar] [CrossRef]
  65. Yan, K.; Park, T.; Yan, G.; Chen, C.; Yang, B.; Liu, Z.; Nemani, R.R.; Knyazikhin, Y.; Myneni, R.B. Evaluation of MODIS LAI/FPAR product collection 6. Part 1: Consistency and improvements. Remote Sens. 2016, 8, 359. [Google Scholar] [CrossRef]
  66. Kimm, H.; Guan, K.; Jiang, C.; Peng, B.; Gentry, L.F.; Wilkin, S.C.; Wang, S.; Bernacchi, C.J.; Peng, J.; Luo, Y. Deriving high-spatiotemporal-resolution leaf area index for agroecosystems in the US Corn Belt using Planet Labs CubeSat and STAIR fusion data. Remote Sens. Environ. 2020, 239, 111615. [Google Scholar] [CrossRef]
  67. Estévez, J.; Vicent, J.; Rivera-Caicedo, J.P.; Morcillo-Pallarés, P.; Vuolo, F.; Sabater, N.; Camps-Valls, G.; Moreno, J.; Verrelst, J. Gaussian processes retrieval of LAI from Sentinel-2 top-of-atmosphere radiance data. ISPRS J. Photogramm. Remote Sens. 2020, 167, 289–304. [Google Scholar] [CrossRef]
Remotesensing 15 02324 g001 550
Figure 1. Sketch map of the study area: the Yangtze River Basin and the Middle Yangtze River with locations of sub-reaches, hydrometric stations, bars, and diversion inlets.
Figure 1. Sketch map of the study area: the Yangtze River Basin and the Middle Yangtze River with locations of sub-reaches, hydrometric stations, bars, and diversion inlets.
Remotesensing 15 02324 g001
Remotesensing 15 02324 g002 550
Figure 2. Flowchart of the machine learning-based LAI inversion model, including three main steps: (1) RTM simulation; (2) ML training; (3) LAI prediction.
Figure 2. Flowchart of the machine learning-based LAI inversion model, including three main steps: (1) RTM simulation; (2) ML training; (3) LAI prediction.
Remotesensing 15 02324 g002
Remotesensing 15 02324 g003 550
Figure 3. Variations in bar and vegetation areas in the MYR after the operation of the TGD. The blue and red lines represent the temporal trends of the bar area and vegetation area, respectively, and the shades delineate the corresponding 95% confidential intervals of the trends.
Figure 3. Variations in bar and vegetation areas in the MYR after the operation of the TGD. The blue and red lines represent the temporal trends of the bar area and vegetation area, respectively, and the shades delineate the corresponding 95% confidential intervals of the trends.
Remotesensing 15 02324 g003
Remotesensing 15 02324 g004 550
Figure 4. Fluctuations in LAI within the MYR subsequent to the commencement of the TGD operation. The red line is the linear fitting of LAI over time and the red/blue shades represent the corresponding 95% confidence/prediction bands.
Figure 4. Fluctuations in LAI within the MYR subsequent to the commencement of the TGD operation. The red line is the linear fitting of LAI over time and the red/blue shades represent the corresponding 95% confidence/prediction bands.
Remotesensing 15 02324 g004
Remotesensing 15 02324 g005 550
Figure 5. Alterations in LAI distribution for bars 1, 5, 8, and 16.
Figure 5. Alterations in LAI distribution for bars 1, 5, 8, and 16.
Remotesensing 15 02324 g005
Remotesensing 15 02324 g006 550
Figure 6. Daily discharge pluses for different hydrometric stations along the Middle Yangtze River from 1998 to 2012.
Figure 6. Daily discharge pluses for different hydrometric stations along the Middle Yangtze River from 1998 to 2012.
Remotesensing 15 02324 g006
Remotesensing 15 02324 g007 550
Figure 7. Alterations in bar submergence frequency and flood discharge frequency for different hydrometric stations in the pre- and post-TGD periods. The blue (red) bar is the bar submerged frequency (flood discharge frequency) of each hydrometric station and the color gradient from lighter to darker signifies the progression of time, ranging from far to near (period 1 through period 3).
Figure 7. Alterations in bar submergence frequency and flood discharge frequency for different hydrometric stations in the pre- and post-TGD periods. The blue (red) bar is the bar submerged frequency (flood discharge frequency) of each hydrometric station and the color gradient from lighter to darker signifies the progression of time, ranging from far to near (period 1 through period 3).
Remotesensing 15 02324 g007
Remotesensing 15 02324 g008 550
Figure 8. Annual bar and channel deposition volume in different sub-reaches after the operation of the Three Gorges Dam.
Figure 8. Annual bar and channel deposition volume in different sub-reaches after the operation of the Three Gorges Dam.
Remotesensing 15 02324 g008
Remotesensing 15 02324 g009 550
Figure 9. Adjustments of hydraulic geometry at different sub-reaches along the Middle Yangtze River from 2004 to 2015. The blue (red) bars represent river depth (river width) and cross-section area (river regime coefficient) in the left and right figures, respectively. The color gradient from lighter to darker signifies the progression of time, ranging from 2004 to 2015.
Figure 9. Adjustments of hydraulic geometry at different sub-reaches along the Middle Yangtze River from 2004 to 2015. The blue (red) bars represent river depth (river width) and cross-section area (river regime coefficient) in the left and right figures, respectively. The color gradient from lighter to darker signifies the progression of time, ranging from 2004 to 2015.
Remotesensing 15 02324 g009
Remotesensing 15 02324 g010 550
Figure 10. Comparison of different LAI products in 2020: (a,b) are LAI products at a spatial resolution of 30 m for bars 10 and 16; (c,d) are LAI products at a spatial resolution of 500 m for bars 10 and 16.
Figure 10. Comparison of different LAI products in 2020: (a,b) are LAI products at a spatial resolution of 30 m for bars 10 and 16; (c,d) are LAI products at a spatial resolution of 500 m for bars 10 and 16.
Remotesensing 15 02324 g010
Table
Table 1. Essential information on 9 sub-reaches, 10 hydrometric stations, and 16 bars in the Middle Yangtze River (MYR).
Table 1. Essential information on 9 sub-reaches, 10 hydrometric stations, and 16 bars in the Middle Yangtze River (MYR).
Sub-reach number123456789
Included reachesYichang-ZhichengZhicheng-MajiadianMajiadian-ShashiShashi-ShishouShishou-JianliJianli-ChenglingjiChenglingji-HankouHankou-JiujiangJiujiang-Hukou
Hydrometric station number12345678910
Hydrometric station nameYichangZhichengMajiadianShashiShishouJianliChenglingjiHankouJiujiangHukou
Bar number12345678910111213141516
Bar nameYanzhiGuanLiutiaoLalinTuqiOuchikouWuguiDamaXiongjiaZhongFuxingHankouTianxingDongcaoDaijiaZhangjia
Table
Table 2. Essential information on 3 operation periods of the Three Gorges Dam (TGD).
Table 2. Essential information on 3 operation periods of the Three Gorges Dam (TGD).
Period NumberImpounded Water LevelIncluded Years
1--2002
2135.0–139.0 m
144.0–156.0 m
145.0–172.8 m		2003–2005
2006–2007
2008
3145.0–171.4 m
145.0–175.0 m		2009
2010–present
Table
Table 3. Sources and resolution of measurements.
Table 3. Sources and resolution of measurements.
Data TypeSpatial ResolutionTemporal ResolutionPeriodSource
Imagery30 m16 days2003–2012Landsat-5, USGS
2014–2021Landsat-8, USGS
Discharge-1 day1991–2020CWRC
Water level-1 day1991–2020CWRC
Cross-section profile2 km1 year2003–2020CWRC
Table
Table 4. Parameter settings in the PROSAIL model. The first four parameters were randomly sampled based on the range between min and max values in the simulation. Solar zenith angle and soil brightness were also randomly sampled based on values as shown in the table.
Table 4. Parameter settings in the PROSAIL model. The first four parameters were randomly sampled based on the range between min and max values in the simulation. Solar zenith angle and soil brightness were also randomly sampled based on values as shown in the table.
ParameterUnitMinMax
Chlorophyll contentug·cm−2060
Carotenoid contentug·cm−2040
Leaf area indexm2·m−206
Leaf inclination -−11
Solar zenith angle°25, 30, 35, 40, 45, 50
Soil brightness-0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0
Table
Table 5. The increase rate of the bar area, vegetation area, and vegetation coverage in each sub-reach in 2019 relative to 2004.
Table 5. The increase rate of the bar area, vegetation area, and vegetation coverage in each sub-reach in 2019 relative to 2004.
Sub-Reach NumberBar AreaVegetation AreaVegetation Coverage
1−14%309%376%
2−37%194%364%
3−6%2%9%
4−3%49%55%
5−1%16%17%
6−5%2%8%
7−13%2%17%
8−6%22%30%
9−4%−3%2%
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Hu, Y.; Zhou, J.; Deng, J.; Li, Y.; Yang, C.; Li, D. River Bars and Vegetation Dynamics in Response to Upstream Damming: A Case Study of the Middle Yangtze River. Remote Sens. 2023, 15, 2324. https://doi.org/10.3390/rs15092324

AMA Style

Hu Y, Zhou J, Deng J, Li Y, Yang C, Li D. River Bars and Vegetation Dynamics in Response to Upstream Damming: A Case Study of the Middle Yangtze River. Remote Sensing. 2023; 15(9):2324. https://doi.org/10.3390/rs15092324

Chicago/Turabian Style

Hu, Yong, Junxiong Zhou, Jinyun Deng, Yitian Li, Chunrui Yang, and Dongfeng Li. 2023. "River Bars and Vegetation Dynamics in Response to Upstream Damming: A Case Study of the Middle Yangtze River" Remote Sensing 15, no. 9: 2324. https://doi.org/10.3390/rs15092324

APA Style

Hu, Y., Zhou, J., Deng, J., Li, Y., Yang, C., & Li, D. (2023). River Bars and Vegetation Dynamics in Response to Upstream Damming: A Case Study of the Middle Yangtze River. Remote Sensing, 15(9), 2324. https://doi.org/10.3390/rs15092324

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop