Hongjiadu Reservoir Research Articles

Elucidation of the dominant factors influencing N2O emission in water-level fluctuation zones in a karst canyon reservoir, southwest China

The water-level fluctuations zones (WLFZs) are crucial transitional interfaces within river-reservoir systems, serving as hotspots for N2O emission. However, the comprehension of response patterns and mechanisms governing N2O emission under hydrological fluctuation remains limited, especially in karstic canyon reservoirs, which introduces significant uncertainty to N2O flux assessments. Soil samples were collected from the WLFZs of the Hongjiadu (HJD) Reservoir along the water flow direction from transition zone (T1 and T2) to lacustrine zone (T3, T4 and T5) at three elevations for each site. These soil columns were used to conduct simulation experiments under various water-filled pore space gradients (WFPSs) to investigate the potential N2O flux pattern and elucidate the underlying mechanism. Our results showed that nutrient distribution and N2O flux pattern differed significantly between two zones, with the highest N2O fluxes in the transition zone sites and lacustrine zone sites were found at 75 % and 95 % WFPS, respectively. Soil nutrient loss in lower elevation areas is influenced by prolonged impoundment durations. The higher N2O fluxes in the lacustrine zone can be attributed to increased nutrient levels resulting from anthropogenic activities. Furthermore, correlation analysis revealed that soil bulk density significantly impacted N2O fluxes across all sites, while NO3−and SOC facilitated N2O emissions in T1-T2 and T4-T5, respectively. It was evident that N2O production primarily contributed to nitrification in the transition zone and was constrained by the mineralization process, whereas denitrification dominated in the lacustrine zone. Notably, the annual N2O efflux from WLFZs accounted for 27 % of that from the water-air interface in HJD Reservoir, indicating a considerably lower contribution than anticipated. Nevertheless, this study highlights the significance of WLFZs as a vital potential source of N2O emission, particularly under the influence of anthropogenic activities and high WFPS gradient.

Influences of hydrodynamics on dissolved inorganic carbon in deep subtropical reservoir: Insights from hydrodynamic model and carbon isotope analysis

Dam construction significantly impacts river hydrodynamics, subsequently influencing carbon biogeochemical processes. However, the influence of hydrodynamic conditions on the migration and transformation of Dissolved Inorganic Carbon (DIC) remains uncertain. To bridge this knowledge gap, we integrated hydrochemistry, isotopic composition (δ13CDIC), and a hydrodynamic model (CE-QUAL-W2) to examine the distinctions, control mechanisms, and environmental effects of DIC biogeochemical processes in a typical large and deep reservoir (Hongjiadu Reservoir) under different hydrodynamic conditions. We evaluated hydrodynamic alterations through the Schmidt stability index and relative water column stability. The analysis disclosed that during weak hydrodynamics periods, the energy necessary for complete mixing the surface and deep water was 34 times higher (3615.32 J/m2 vs.106.86 J/m2), and stability was 13 times greater (312.96 vs. 24.69) compared to periods of strong hydrodynamics. Additionally, the spatiotemporal heterogeneity of DIC concentrations (1.4 % to -9.1 %) and δ13CDIC (-1.7 % to -19.5 %) from the dry to wet seasons reflected disparities in DIC control mechanisms under varied hydrodynamic conditions. Based on model simulations, our calculations indicate that during weak hydrodynamics periods, the enhancement of the biological carbon pump effect resulted in substantial sequestration of DIC, reaching up to 379.6 t-DIC·d−1 in the water. Conversely, during strong hydrodynamics periods, DIC retention capacity decreased by 69.2 t·d−1, resulting in reservoir CO2 emissions of 22.7 × 104 t, which were more than 7 times higher than during weak hydrodynamics periods (3.2 × 104 t). Our findings emphasize the discernible impact of hydrodynamic conditions on reservoir biogeochemical processes related to DIC. Considering the increasing construction of reservoirs globally, understanding and controlling hydrodynamic conditions are crucial for mitigating CO2 emissions and optimizing reservoir management.

Unraveling the factors influencing CO2 emissions from hydroelectric reservoirs in karst and non-karst regions: A comparative analysis

Carbon migration, transformation, and emissions as CO2 in reservoir and lake systems have been extensively studied. However, uncertainties persist regarding carbon cycling variations in both karst and non-karst regions within large thermal stratified river-reservoir systems. To address this knowledge gap, we combined measurements of water chemistry, isotopic compositions (δ13C), partial pressure of CO2 (pCO2), and CO2 fluxes (FCO2) to elucidate the differences, control mechanisms, and environmental effects of various carbon biogeochemical processes in the Hongjiadu reservoir (HJD, karst reservoir) and Shangyoujiang reservoir (SYJ, non-karst reservoir), Yangtze River basin, China. Our results demonstrate that key biogeochemical processes are associated with CO2 production and emissions. The stronger biological carbon pump (BCP) effect in the lentic area of the HJD reservoir limited CO2 emission (0.9 ± 5.0 mmol m−2 d−1) compared to the SYJ reservoir (50.9 ± 33.4 mmol m−2 d−1), leading to low dissolved inorganic carbon (DIC) and high saturation state of carbonate minerals in the lentic epilimnion. Although the released water from both reservoirs acted as hotspots for CO2 emissions, the results of higher Revelle factor (RF) in the HJD reservoir than SYJ reservoir suggest the greater buffer capacity and lower sensitivity of OC degradation changes in DIC. Despite the shorter hydraulic retention time (HRT) of SYJ reservoir (120 d) compared to HJD reservoir (368 d), FCO2 was 3.5 times higher than that of the HJD reservoir (652.9 mmol m−2 d−1). Moreover, upon analyzing selected reservoirs worldwide, it becomes apparent that karst reservoirs demonstrate a reduced potential for CO2 emissions. Our results indicate that the impact of varying geologic settings should be considered to minimize errors in regional and global CO2 emission estimates.

Differential response of microbial diversity and abundance to hydrological residual time and age in cascade reservoirs

Freshwater sediments harbor taxonomically and metabolically diverse microbial communities, which play critical roles in aquatic ecosystems. However, little is known about the biogeographical patterns of assembly of such sedimentary communities, especially in cascade reservoirs of various age and hydrological characteristics. Using Illumina sequencing and quantitative real-time polymerase chain reaction targeting the 16S rRNA gene, we investigated spatial patterns of both bacteria and archaeal communities across eight reservoirs (including Hongfeng Reservoir (HFy), Hongjiadu Reservoir (HJDy), Wujiangdu Reservoir (WJDs), Dongfeng Reservoir (DFs), Silin Reservoir (SLd), Pengshui Reservoir (PSd), Suofengying Reservoir (SFYd), and Yinpan Reservoir (YPd)), forming the largest cascade reservoir system located on the Wujiang River, the tributary of the upper Yangtze River. Longer hydrological residual times promoted the proliferation of bacteria while reducing their diversity. In structures of different age, distinct archaeal communities were observed, with more abundant and diverse archaea in the earlier constructed reservoirs (HFy, WJDs). Both environmental factors (redundancy analysis) and spatial scale (non-metric multidimensional scaling analysis) determined community structures. The abundances of bacteria and archaea were both positively correlated with the presence of nutrients. In particular, a shift in the composition of the sedimentary archaeal community was observed along with increasing trophic complexity of these reservoirs, from Thaumarchaeota- towards Euryarchaeota-dominated communities. Also, the preference of several phyla of bacteria, including Latescibacteria and Bacteroidetes, for anaerobic and eutrophic habitats was significantly positively correlated with the presence of nutrients in these reservoirs. We demonstrated the existence of different distribution patterns of bacterial and archaeal communities in reservoir sediments, reflecting both deterministic and stochastic processes distinctly regulating the abundance of bacteria and archaea, in response to hydrological and biogeochemical changes after water impoundment in cascade reservoirs along a large river.

Carbon and phosphorus transformation during the deposition of particulate matter in the large deep reservoir

As the running time of reservoirs is increasing, a large number of reservoirs are becoming eutrophicated. Organic phosphorus (OP) is a key factor in eutrophication. However, the mechanism and extent to which organic matter degradation affects P recycling in water column of large deep reservoirs are unclear, especially for the newly-built ones. In this study, different forms of carbon (C) and P in the water column of Hongjiadu Reservoir were investigated. The contents of particulate organic carbon (POC) and particulate organic phosphorus (POP) both decreased with depth in summer, indicating that organic matter was degraded during the deposition of particulates. In contrast, the contents of POC and POP varied slightly with depth in winter. This difference may result from the double thermal stratification and the corresponding double oxygen stratification in summer. The POC/POP ratios were lower in the epilimnion and increased with depth, suggesting that P was preferentially regenerated relative to C during organic matter degradation. The contents of particulate inorganic phosphorus (PIP) and POP were significantly negatively correlated, indicating that POP transformed into PIP in deeper water. The double thermoclines and oxyclines in Hongjiadu Reservoir lead to very low dissolved oxygen (DO) concentrations in the hypolimnion, which should receive sufficient attention. If water becomes hypoxic, enhanced P release during organic matter degradation will promote phytoplankton growth, leading to higher phytoplankton biomass and more severe DO depletion. Thus, a positive feedback loop may form among hypoxia, enhanced P release, higher primary productivity, and more severe hypoxia, accelerating P recycling in large deep reservoirs. Once if eutrophication occurs in these reservoirs, it will be very difficult to restore the water ecosystem. Thus, it is particularly important to prevent the occurrence of eutrophication and the formation of positive feedback loop as early as possible. This highlights the importance of both reducing external loading and improving DO level in large deep reservoirs.

Monthly runoff forecasting via an improved extreme learning machine

Generally, monthly runoff prediction is of great importance for effective water resource planning and management. Extreme learning machine (ELM) is a novel training tool for the famous single layer feed-forward neural network. Due to the satisfying performance, ELM is chosen for monthly runoff forecasting in this research. Nevertheless, it is unfortunately found that ELM easily falls into local optima in practice because the randomly-determined computational parameters remain unchanged during the learning process. Thus, this paper tries to develop an improved extreme learning algorithm (IELM) where the evolutionary algorithm is used to search for satisfying computational parameters while the Moore-Penrose generalized inverse method is used to determine the output weights. The IELM method is applied to forecast the monthly runoff of Hongjiadu reservoir in southwest China. The results show that the proposed method outperforms several traditional algorithms with respect to the performance indicators. Thus, this paper provides a new and effective artificial intelligence approach for the monthly runoff forecasting.

The influence of cascade reservoir construction on water chemistry distribution in Karst area

为深入了解河流梯级筑坝对喀斯特地区河流水化学分布的影响，于2017年1、4、7和10月别对乌江干流洪家渡水库（多年调节）、乌江渡水库（季调节）和索风营水库（日调节）3个具有不同滞留时间的水库进行水样采集，分析入库水、坝前剖面水和下泄水的水化学特征，探讨河流梯级筑坝对水化学分布及风化速率估算的影响.研究结果表明：3个水库深层水比表层水HCO₃^-浓度分别高12.9%、5.5%和8.0%，Ca²⁺浓度分别高15.9%、2.4%和8.5%.河流梯级筑坝一定程度上改变了水体水化学组成，从而影响碳酸盐岩风化速率估算.整体上，洪家渡水库、索风营水库和乌江渡水库的全年风化速率变化范围分别为：-1.7%~15.4%、-5.6%~1.1%和-0.3%~3.4%.河流筑坝作用对风化速率估算及主量离子浓度的影响：HCO₃^-与Ca²⁺浓度分布均为：洪家水库 > 乌江渡水库 > 索风营水库，这与水体滞留时间长短规律一致，表明水体滞留时间影响着水化学的组成分布.同时水体离子浓度表现出明显的季节性差异，丰水期各水库变化率明显大于枯水期.上述结论表明喀斯特地区河流连续筑坝后水化学组成及分布特征发生了一定程度改变，影响流域化学风化速率的估算，且影响程度随水体滞留时间延长而增大，并受气温影响.因此，今后在估算流域风化速率及探究水化学空间变化时应对筑坝作用加以考虑，以便更加准确地评估喀斯特流域岩石风化在全球碳循环中的作用.;In order to better understand the effect of cascade damming on river chemistry distribution in the karst region, we collected water samples from Hongjiadu Reservoir (multi-yearly regulation), Wujiangdu Reservoir (seasonal regulation) and Suofengying Reservoir (daily regulation) in January, April, July and October of 2017. We analyzed the water chemistry characteristics of the inflow, outflow and stratified water samples of these three reservoirs and investigated the cascade damming effect on water chemistry. The results show that water chemistry distribution has been highly affected by the damming effect. The HCO₃^- concentration of the hypolimnion is 12.9%, 5.5% and 8.0%, and the Ca²⁺ concentration is 15.9%, 2.4% and 8.5% higher than that of the surface water, respectively, for the Hongjiadu Reservoir, Suofengying Reservoir and Wujiangdu Reservoir. Consequently, the chemical weathering rate estimation of carbonate rock has also been affected. The annual weathering rate variation of Hongjiadu Reservoir, Suofengying Reservoir and Wujiangdu Reservoir is -1.7%-15.4%, -5.6%-1.1% and -0.3%-3.4%. The calculated weathering rate and HCO₃^- and Ca²⁺ concentrations in decreasing order is Hongjiadu Reservoir > Wujiangdu Reservoir > Suofengying Reservoir, which is consistent with the water retention time, indicating that water retention time affects the redistribution of water chemistry. At the same time, the ions concentration showed obvious seasonal variations, and the change rates in the three reservoirs were significantly greater during the flood season than that of the dry season. The above conclusions indicate that the water chemistry of rivers in karst area is highly affected by damming, which affects the estimation of chemical weathering rate. Water residence time and temperature are the main controlling factors on water chemistry redistribution. Therefore, we proposed that the damming effects should be taken into account to calculate the weathering rate accurately.

Comparison of Multiple Linear Regression, Artificial Neural Network, Extreme Learning Machine, and Support Vector Machine in Deriving Operation Rule of Hydropower Reservoir

Operation rule plays an important role in the scientific management of hydropower reservoirs, because a scientifically sound operating rule can help operators make an approximately optimal decision with limited runoff prediction information. In past decades, various effective methods have been developed by researchers all the over world, but there are few publications evaluating the performances of different methods in deriving the hydropower reservoir operation rule. To achieve satisfying scheduling process triggered by limited streamflow data, four methods are used to derive the operation rule of hydropower reservoirs, including multiple linear regression (MLR), artificial neural network (ANN), extreme learning machine (ELM), and support vector machine (SVM). Then, the data from 1952 to 2015 in Hongjiadu reservoir of China are chosen as the survey case, and several quantitative statistical indexes are adopted to evaluate the performances of different models. The radial basis function is chosen as the kernel function of SVM, while the sigmoid function is used in the hidden layer of ELM and ANN. The simulations show that three artificial intelligence algorithms (ANN, SVM, and ELM) are able to provide better performances than the conventional MLR and scheduling graph method. Hence, for scholars in the hydropower operation field, the applications of artificial intelligence algorithms in deriving the operation rule of hydropower reservoir might be a challenge, but represents valuable research work for the future.

Model and Algorithm of Sequence‐Based Quantum‐Inspired Neural Networks

To enhance the approximation ability of traditional Artificial neural network (ANN), by introducing the quantum rotation gates and the multi-qubits controlled-NOT gates to ANN, we proposed a Sequence input-based quantum-inspired neural network (SIQNN). In our model, the hidden nodes are composed of some multi-qubits controlled-NOT gates, the inputs are described by the multi-dimensional discrete qubits sequences, the output nodes are the traditional neurons. The model parameters include the rotation angles of quantum rotation gates in hide layer and the weights in output layer. The learning algorithms were derived by employing the Levenberg-Marquardt algorithm. Simulation results of predicting the runoff of the Hongjiadu Reservoir show that, the SIQNN is obviously superior to the ANN.

Temporal and spatial characteristics of dissolved organic carbon in the Wujiang River, Southwest China

River systems play an important role in the global carbon cycle. Rivers transport carbon to the ocean and also affect the carbon cycle in the coastal ocean. The flux from land to the ocean is thought to be a very important part of the land carbon budget. To investigate the effect of dam-building on dissolved organic carbon (DOC) in rivers, three reservoirs of different trophic states in the Wujiang basin, Guizhou Province, were sampled twice per month between May 2011 and May 2012. Temporal and spatial distributions of DOC in the reservoirs and their released waters were studied. It was found that different factors controlled DOC in river water, reservoir water, and released water. DOC in the rivers tended to be affected by primary production. For reservoirs, the main controlling factors of DOC concentration varied by trophic state. For the mesotrophic Hongjiadu Reservoir, the effect of primary production on DOC concentration was obvious. For the eutrophic Dongfengdu Reservoir and the hypereutrophic Wujiangdu Reservoir, primary production was not significant and DOC came instead from soil and plant litter.

Daily Reservoir Runoff Forecasting Method Using Artificial Neural Network Based on Quantum-behaved Particle Swarm Optimization

Accurate daily runoff forecasting is of great significance for the operation control of hydropower station and power grid. Conventional methods including rainfall-runoff models and statistical techniques usually rely on a number of assumptions, leading to some deviation from the exact results. Artificial neural network (ANN) has the advantages of high fault-tolerance, strong nonlinear mapping and learning ability, which provides an effective method for the daily runoff forecasting. However, its training has certain drawbacks such as time-consuming, slow learning speed and easily falling into local optimum, which cannot be ignored in the real world application. In order to overcome the disadvantages of ANN model, the artificial neural network model based on quantum-behaved particle swarm optimization (QPSO), ANN-QPSO for short, is presented for the daily runoff forecasting in this paper, where QPSO was employed to select the synaptic weights and thresholds of ANN, while ANN was used for the prediction. The proposed model can combine the advantages of both QPSO and ANN to enhance the generalization performance of the forecasting model. The methodology is assessed by using the daily runoff data of Hongjiadu reservoir in southeast Guizhou province of China from 2006 to 2014. The results demonstrate that the proposed approach achieves much better forecast accuracy than the basic ANN model, and the QPSO algorithm is an alternative training technique for the ANN parameters selection.

Optimized operation of cascade reservoirs on Wujiang River during 2009–2010 drought in southwest China

The effects of optimized operation principles implemented at reservoirs on the Wujiang River in southwest China between September 2009 and April 2010 under drought conditions were analyzed based on operational data collected from the Guizhou Wujiang Hydropower Development Co., Ltd. A set of linear regression equations was developed to identify the key factors impacting the electric power generation at reservoirs. A 59% reduction in the inflow discharge at the Hongjiadu Reservoir led to a decrease of only 38% in the total electric power generation at the Hongjiadu, Dongfeng, Suofengying, and Wujiangdu reservoirs on the Wujiang River, indicating that optimized operation can play an important role in drought management. The water level and the amount of other water inputs at the Hongjiadu Reservoir and the outflow discharge at all of the reservoirs except the Wujiangdu Reservoir were key factors affecting the total electric power generation at reservoirs on the Wujiang River under optimized operation.

Impacts of river impoundment on the riverine water chemistry composition and their response to chemical weathering rate

Currently, most rivers worldwide have been intensively impounded. River damming becomes a big problem, not only in inducing the physical obstruction between upstream and downstream, but also in destroying the natural continuity of river. But the discontinuity of water quality was often neglected, which presents a challenge to traditional river geochemistry research. To understand the changes in basic chemistry of water upstream and downstream of the dam, we investigated the Miaotiao River reservoirs in series in the Wujiang River Basin, and the Hongjiadu, Dongfeng Reservoir on the upper reaches of the Wujiang River. Chemical weathering rates were calculated using the water chemistry data of the reservoir surface and downstream of the dam, in each reservoir, respectively. The results showed that the difference between the chemical weathering rates calculated from reservoir surface water and water downstream of the dam was greater in reservoirs with a longer water retention time. In Hongjiadu Reservoir with the longest water retention time among the studied reservoirs, this difference reaches 9%. As a result, the influence of river damming, especially the influence of reservoirs in series, should be taken into account when calculating the chemical weathering rate of a river basin.

Diurnal variations of pCO2 in relation to environmental factors in the cascade reservoirs along the Wujiang River, China

We have investigated the diurnal variations of the pCO2 and related environmental factors in the cascade reservoirs with different trophic levels along Wujiang River. In surface water the pCO2 was 357±11 μatm in Hongjiadu Reservoir, 338±48 μatm in Dongfeng Reservoir, 682±303 μatm in Wujiangdu Reservoir, and 1677±429 μatm in Liuguang, respectively. The results indicated that these cascade reservoirs had much lower pCO2 values in surface water than river did, and hypereutrophic reservoir showed larger diurnal variations of pCO2 than meso-eutrophic reservoir. In water column, pCO2 tended to increase with the depth. Phytoplankton and the environmental factors such as temperature and pH had different influences on pCO2 diurnal variations due to different trophic levels, and the effect of phytoplankton on pCO2 variation increased with the increase of trophic level in these reservoirs.

Spatiotemporal variations of nitrous oxide (N2O) emissions from two reservoirs in SW China

Abstract Greenhouse gas emissions from hydroelectric dams have recently given rise to controversies about whether hydropower still provides clean energy. China has a large number of dams used for energy supply and irrigation, but few studies have been carried out on aquatic nitrous oxide (N 2 O) variation and its emissions in Chinese river-reservoir systems. In this study, N 2 O spatiotemporal variations were investigated monthly in two reservoirs along the Wujiang River, Southwest China, and the emission fluxes of N 2 O were estimated. N 2 O production in the reservoirs tended to be dominated by nitrification, according to the correlation between N 2 O and other parameters. N 2 O saturation in the surface water of the Wujiangdu reservoir ranged from 214% to 662%, with an average fluctuation of 388%, while in the Hongjiadu reservoir, it ranged from 201% to 484%, with an average fluctuation of 312%. The dissolved N 2 O in both reservoirs was over-saturated with respect to atmospheric equilibrium levels, suggesting that the reservoirs were net sources of N 2 O emissions to the atmosphere. The averaged N 2 O emission flux in the Wujiangdu reservoir was 0.64 μmol m −2 h −1 , while it was 0.45 μmol m −2 h −1 in the Hongjiadu reservoir, indicating that these two reservoirs had moderate N 2 O emission fluxes as compared to other lakes in the world. Downstream water of the dams had quite high levels of N 2 O saturation, and the estimated annual N 2 O emissions from hydropower generation were 3.60 × 10 5 and 2.15 × 10 5 mol N 2 O for the Wujiangdu and the Hongjiadu reservoir, respectively. These fluxes were similar to the total N 2 O emissions from the reservoir surfaces, suggesting that water released from reservoirs would be another important way for N 2 O to diffuse into the atmosphere. It can be concluded that dam construction significantly changes the water environment, especially in terms of nutrient status and physicochemical conditions, which have obvious influences on the N 2 O spatiotemporal variations and emissions.

Intercomparison and applicability of some dynamic and equilibrium approaches to determine methylated mercury species in pore water

To assess adequately the impact of methylmercury (MeHg) on sensitive wetland ecosystems, accurate measurements of MeHg in pore water are required. In the present study, the feasibility of three methods for porewater sampling was investigated with respect to MeHg in sediments and rice paddy fields. The performance of dialysis samplers (peepers), sediment core sectioning followed by porewater separation by centrifugation (core), and the thin film diffusive samplers (DGT) were evaluated. These methods were intercompared in field experiments at two sites in Guizhou province, SW China disparately impacted by mercury pollution. All the methods report that the concentrations of MeHg in the soils of the Gouxi (GX) rice paddy near Wanshan were much higher than that in the sediment of the Hongjiadu Reservoir (HR), which is located within the Wujiang River basin. The three methods also report different MeHg profiles at the same site. However, these methods exhibit different temporal and spatial resolution scales, due to the differing operations involved with the sampling methods, may also reflect different states of MeHg in pore water. This corresponds to MeHg derived from diffusive flux, equilibrium concentration, and bulk concentration in pore water detected by DGT, peeper, and sediment core, respectively. The advantages and limits of the three methods are also presented.

Dissolved silicate retention and transport in cascade reservoirs in Karst area, Southwest China

Presently, the fluvial fluxes of Dissolved silicate (DSi) by many famous rivers in the world have been reported significantly decreased in the past several decades. DSi retention by river damming hence becomes a great concern. In this study, the impact on DSi retention and transport by cascade reservoirs in the Karst area were investigated. Based on the monitoring data for four seasons, budget of DSi in these reservoirs were calculated. Results showed that, only the downstream eutrophic Wujiangdu reservoir (WJD) was the sink for DSi, with a value of ca. 3.5 kT DSi trapped annually. The new Hongjiadu reservoir (HJD) in the headwater area revealed to be an important source for DSi. Both Dongfeng (DF) and Suofengyin (SFY) reservoirs also released more DSi to the downstream than entering them. The great discrepancy of DSi source and sink effects among these reservoirs indicates that, (i) DSi was obviously taken up in summer and spring, and the concentration of DSi could be dropped down to below 30 μΜ; (ii) during the period of thermal stratification, the concentration stratification of DSi is also developed along the water column. Possibly DSi releasing from inundation soil and bottom sediment was the important compensation for the silicon depletion in the epilimnion. In the newly constructed reservoirs ( e.g. HJD and SFY), the releasing flux of DSi can significantly exceed the assimilating flux by diatom; and (iii) due to the hypolimnion introducing for hydropower operation, water leaving the dam generally had high DSi content, and then masked the DSi taking up in the epilimnion. This process also decreased the ratio of DIN to DSi, mitigating the silicon limiting situation in the downstream.