Search

River water level prediction (WLP) plays an important role in flood control, navigation, and water supply. In this study, a WaveNet-based convolutional neural network (WCNN) with a lightweight structure and good parallelism was developed to improve the prediction accuracy and time effectiveness of WLP. It was applied to predict 1/2/3 days ahead the water levels at the Waizhou gauging station of the Ganjiang River (GR) in China, and it was compared with two recurrent neural networks (long short-term memory (LSTM) and gated recurrent unit (GRU)). The results showed that the WCNN model achieved the best prediction performance with the fewest training parameters and time. Compared with the LSTM and GRU models in the 1-day ahead prediction, the training parameters were reduced from 73,851 and 55,851 to 32,937, respectively. The root mean square error (RMSE) was reduced from 0.071 and 0.076 to 0.057, respectively. The mean absolute error (MAE) was reduced from 0.052 and 0.059 to 0.038, respectively. The Nash–Sutcliffe efficiency (NSE) and coefficient of determination (R2) both increased to 0.998. This result indicated that the improved model was more efficient for WLP. HIGHLIGHTS A WaveNet-based convolutional neural network was proposed for water level prediction.; WCNN with a lightweight structure and good parallelism achieved better prediction performance.; WCNN obtained higher accuracy results with the fewest parameters and training time than RNN.; Influence of different inputs and hyperparameters of models on prediction results was revealed.;

The response of blue and green water to climate and land-use change in the Ganjiang River Basin (GRB) is evaluated via the SWAT model that combines three scenarios (the land-use/land-cover (LULC), climate change, and integrated climate and LULC change scenarios) in the 2040s (2031–2050) and 2060s (2051–2070). The results indicate that, for the GRB, cropland, woodland, and grassland show a decreasing trend, while build-up and water areas show an increasing trend in terms of future land-use change. The climatic conditions projected using NORESM1-M model data under the RCP4.5 and RCP8.5 scenarios suggest, respectively, increases in precipitation (31.17 and 27.24 mm), maximum temperature (2.25 and 2.69 °C), and minimum temperature (1.96 and 2.58 °C). Under climate change conditions, blue water is estimated to decrease by up to 16.89 and 21.4 mm under RCP4.5 and RCP8.5, while green water is estimated to increase up to 19.14 and 20.22 mm, respectively. Under the LULC changes, blue water is projected to increase by up to 5.50 and 7.57 mm, while green water shows decreases of 4.05 and 7.80 mm for the LULC2035 and LULC2055 scenarios, respectively. Under the four combined LULC and climate change conditions (RCP4.5_2040s, RCP4.5_2060s, RCP8.5_2040s, and RCP8.5_2060s), blue water tends to decrease by 0.67, 7.47, 7.28, and 9.99 mm, while green water increases by 19.24, 20.8, 13.87, and 22.30 mm. The influence of climate variation on blue and green water resources is comparatively higher than that of the integrated impacts of climate and land-use changes. The results of this study offer a scientific reference for the water resources management and planning department responsible for scheduling water resource management plan in the GRB. HIGHLIGHTS This study investigates the effects of climate changes and land-use changes on blue/green water.; The NORESM1-M model data from the CMIP5 are combined with the SWAT model to analyse the influence of future climate changes on blue/green water.; The CA–Markov model is applied to generate future land-use scenarios.; Multiple climate and land-use scenarios are set to quantitatively analyse the changes of blue/green water.;

There are abundant resources like climate, water, biological and tourism in the middle and upper reaches of the Ganjiang River (MU-GJR), which is an important ecological protection area in hilly areas of South China. The four periods of land use and land cover (LULC) data from 1980 to 2018 and DEM data are the main data support. The LULC change were analyzed with the area proportion, transfers chord diagram and change characteristics map, the response of ecological environment from LULC change was analyzed by using the landscape ecological environment quality (LEEQ) and contribution index. Then the terrain gradient was used to analyze the impact of LULC transfer on ecological environment under different terrain conditions. (1) The LULC characteristics in the study area show that was dominated by forest land and cultivated land (more than 90 %), various types of land transfer significantly from 2010 to 2018, especially the cultivated land, forest land and grassland. The stable type of LULC change was the most widely distributed, followed by the later change type (2010–2018) and mainly distributed in the flat valley, with the largest change range of cultivated land, forest land and grassland. (2) The LEEQ was generally well, but gradually deteriorates. The obvious change of LEEQ was found after 2010. The deterioration type of quality mostly occurs around the city, while the getting better type occurs in the transition area between the city and the mountains. (3) There are more deterioration categories than improvement categories of LULC transfer from 1980 to 2018. The improvement categories were mainly contributed by the transfer to other land to forest land and build-up land to other land. The deterioration categories were mainly contributed by the transfer of forest land to other land. (4) The impact of LULC change on the ecological environment decreases with the increase of elevation, slope and terrain position. There are differences in the impacts of various types of transfer on the ecological environment under the three gradients. In the elevation gradient, the improvement effect were dominant at −105–100 m and 200–500 m, and the deterioration effect were dominant at 100–200 m, 500–1000 m and 1000–2128 m. In the slope gradient, the improvement effect were dominant at 0–2°, 15–25° and 25–69.7°, and the deterioration effect were dominant at 2–6° and 6–15°. In terrain position gradient, improvement effect were dominant at 0–0.32, 0.32–0.51, 0.7–0.9 and 0.9–1.58, and deterioration effect were dominant at 0.51–0.7. Therefore, the specific conditions of different terrain gradients should be considered for land resource development in the middle and upper reaches of Ganjiang River region.

Application of the Soil and Water Assessment Tool (SWAT) model is limited in plain polders, where hydrological and nutrient processes are confined by water conservancy facilities such as dikes and pumping stations. Watershed delineation techniques are proposed to enable SWAT to simulate these processes in a plain polder in Jiangxiang Town (Nanchang County, Jiangxi Province, China). Drainage unit division and multiple-outlet modelling approaches are introduced, and the main river network and land features are incorporated into a digital elevation model (DEM), so that drainage unit delineation can agree well with real flow direction and concentration. The impoundments of pumping stations are regarded as functioning like reservoirs, which are set up in the SWAT model to simulate confined hydrological processes. The results show that confined hydrological and nutrient processes in plain polders are simulated well; the retention rates of total nitrogen and total phosphorus are estimated as, respectively, 39% and 29% in both streams and impoundments. [ABSTRACT FROM AUTHOR]

Rainfall information is a prerequisite to and plays a vital role in driving hydrological models. However, limited by the observation methods, the obtained precipitation data, at present, are still too coarse. In this study, a new downscaling method was proposed to obtain high spatial resolution (~1 km/hourly) precipitation estimates based on Integrated Multi-satellitE Retrievals for GPM (IMERG) data at hourly scale. Compared with original IMERG data, the downscaled precipitation results showed the similar spatial patterns with those of original IMERG data, but with finer spatial resolution. In addition, the downscaled precipitation estimates were further analyzed to quantify their improvements using the Coupled Routing and Excess STorage (CREST) model across Ganjiang River basin. Compared with the observed streamflow, the downscaled precipitation results showed satisfying hydrological performance, with Nash-Sutcliffe Coefficient of Efficiency (NSCE), Root Mean Square Error (RMSE), Relative Bias (BIAS), and Correlation Coefficient (CC). The improvement in terms of four statistic metrics in terms of streamflow simulation also indicated great potential of hydrological utility for the downscaled precipitation results.

• Study on land use and land cover change in the hilly watershed in South China. • Evaluation of landscape ecological environment based on LULC change. • Terrain gradient response of ecological environment to LULC change be analyzed. There are abundant resources like climate, water, biological and tourism in the middle and upper reaches of the Ganjiang River (MU-GJR), which is an important ecological protection area in hilly areas of South China. The four periods of land use and land cover (LULC) data from 1980 to 2018 and DEM data are the main data support. The LULC change were analyzed with the area proportion, transfers chord diagram and change characteristics map, the response of ecological environment from LULC change was analyzed by using the landscape ecological environment quality (LEEQ) and contribution index. Then the terrain gradient was used to analyze the impact of LULC transfer on ecological environment under different terrain conditions. (1) The LULC characteristics in the study area show that was dominated by forest land and cultivated land (more than 90 %), various types of land transfer significantly from 2010 to 2018, especially the cultivated land, forest land and grassland. The stable type of LULC change was the most widely distributed, followed by the later change type (2010–2018) and mainly distributed in the flat valley, with the largest change range of cultivated land, forest land and grassland. (2) The LEEQ was generally well, but gradually deteriorates. The obvious change of LEEQ was found after 2010. The deterioration type of quality mostly occurs around the city, while the getting better type occurs in the transition area between the city and the mountains. (3) There are more deterioration categories than improvement categories of LULC transfer from 1980 to 2018. The improvement categories were mainly contributed by the transfer to other land to forest land and build-up land to other land. The deterioration categories were mainly contributed by the transfer of forest land to other land. (4) The impact of LULC change on the ecological environment decreases with the increase of elevation, slope and terrain position. There are differences in the impacts of various types of transfer on the ecological environment under the three gradients. In the elevation gradient, the improvement effect were dominant at −105–100 m and 200–500 m, and the deterioration effect were dominant at 100–200 m, 500–1000 m and 1000–2128 m. In the slope gradient, the improvement effect were dominant at 0–2°, 15–25° and 25–69.7°, and the deterioration effect were dominant at 2–6° and 6–15°. In terrain position gradient, improvement effect were dominant at 0–0.32, 0.32–0.51, 0.7–0.9 and 0.9–1.58, and deterioration effect were dominant at 0.51–0.7. Therefore, the specific conditions of different terrain gradients should be considered for land resource development in the middle and upper reaches of Ganjiang River region. [ABSTRACT FROM AUTHOR]

Ecological flow is an important indicator for reflecting the stability of a watershed ecosystem. The calculation of ecological discharge under hydrological variation has become a research hot-spot. The Ganjiang River south of Poyang Lake in China was taken as an example in this study. Hydrological Alteration Diagnosis System methods were used to detect the change-points. The Distributed Time Variation Gain Model (DTVGM) was used to carry out runoff restoration. The Probability-weighted Flow Duration Curve was applied to calculate the ecological flow. The results showed that: (1) The hydrological alteration of the Waizhou Station occurred in 1991, the annual runoff increased by 10%, and the Gini coefficient (GI) increased by 0.07 after the change-point. The change in precipitation was the main driving factors. (2) The R value and NSE of the DTVGM were greater than 0.84, which represents the feasibility of the model used to restore runoff. (3) Compared to the traditional hydrological method, the proposed method can better reflect the inter-annual difference of ecological flow, flow ranges for high, normal, and low flow years are 398-3771 m 3 /s, 352-2160 m 3 /s, and 277-1657 m 3 /s, respectively. The calculation method of ecological flow in rivers considering hydrological variation can more scientifically reflect the impact of hydrological variation on ecological flow process, ecological flow under different human activities that can be calculated, such as dam control, water intake and water transfer, furthermore, it also provides a scientific basis for water resources planning and allocation under changing environment.

River flow reconstruction under the background of long-term climate change is of great significance for understanding the regional response to future drought and flood disasters, and the sustainable development of water resources. Investigating the basic characteristics and changing trends of the streamflow of the Ganjiang River is scientifically important to mitigate drought and flood disasters in the future. This study reconstructed drought and flood grade series of five regional stations of the Ganjiang River based on spatially explicit and well-dated local chronicle materials and used a linear regression model of modern drought/flood grades and precipitation to reconstruct historical precipitation for the past 515 years. The relationships between the modern precipitation of five regional stations and streamflow of Waizhou Station, which is the last hydrological station of the Ganjiang River were analyzed through principal component regression. The adjusted R2 is 0.909, with a low relative bias of −1.82%. The variation of streamflow from AD 1500 to AD 2014 was reconstructed using the proposed model. Result shows that high flows occur in nine periods and low flows occur in 11 periods. Extremely low stream flow in 515 years appears during the middle and late 17th century. Cumulative anomaly and Mann-Kendall mutation test results reveal that a transition point from predominantly low to high flows occur in AD 1720. Redfit power spectrum analysis result shows that the variation periods of streamflow are 2–5, 7–8 years, and approximately 32 years, where the most significant period is 2–3 years. Continuous wavelet transform indicates that the corresponding relation occurs between streamflow and El Niño/Southern Oscillation for eight years. Streamflow is affected by temperature and East Asian monsoon that is controlled by solar activities. The flood may be related to strong solar activity, monsoon failure, and vice versa. Hydrological frequency curve analysis shows that the streamflow of the Ganjiang River once in a hundred years may reach up to 1031 × 108 m3 for flood or 485 × 108 m3 for drought and the standard of once in a millennium runoff may reach up to 1188 × 108 m3 for flood or 450 × 108 m3 for drought. These results may provide basic hydrological data for the sustainable development of society and serve as a reference for mitigating the impact of drought and flood disasters in the future. [ABSTRACT FROM AUTHOR]

Rainfall information is a prerequisite to and plays a vital role in driving hydrological models. However, limited by the observation methods, the obtained precipitation data, at present, are still too coarse. In this study, a new downscaling method was proposed to obtain high spatial resolution (~1 km/hourly) precipitation estimates based on Integrated Multi-satellitE Retrievals for GPM (IMERG) data at hourly scale. Compared with original IMERG data, the downscaled precipitation results showed the similar spatial patterns with those of original IMERG data, but with finer spatial resolution. In addition, the downscaled precipitation estimates were further analyzed to quantify their improvements using the Coupled Routing and Excess STorage (CREST) model across Ganjiang River basin. Compared with the observed streamflow, the downscaled precipitation results showed satisfying hydrological performance, with Nash-Sutcliffe Coefficient of Efficiency (NSCE), Root Mean Square Error (RMSE), Relative Bias (BIAS), and Correlation Coefficient (CC). The improvement in terms of four statistic metrics in terms of streamflow simulation also indicated great potential of hydrological utility for the downscaled precipitation results. [ABSTRACT FROM AUTHOR]