The growth of forecasting models has resulted in the development of an excellent model known as the support vector machine (SVM). SVMs can find a global optimal solution equipped with kernel functions. This research trains and tests the SVM network and constructs the support vector regression prediction model by using hydrologic data. Six hydrologic time series were calculated by different kernel functions (namely, linear, polynomial, radial basis function (RBF)), to determine which kernel is the more suitable hydrologic time series in practice. A new solution is presented to identify the good parameter (C; g) by using grid-search and cross-validation. Results prove that linear SVM is a superior model to polynomial and RBF and produced the most accurate results for modeling hydrologic time series behavior as complex hydrologic phenomena. The case study also shows that the calculation errors were correlated with data characteristics. More stable raw data will result in a more accurate result, whereas more random data will result in a more inaccurate result. Model performance could also be dependent on base data nonlinearity.

This study presents an integrated probabilistic framework by combining Monte Carlo Simulation with a gas transport model of a horizontal well with multi-fracturing stages to assess shale gas resources in the Wangyinpu Formation of the Xiuwu Basin, China. Modeling results suggest that the 30-year cumulative production of a single horizontal well is predicted at a likely value of 3.50×10 8 m 3 with a maximum of 6.78×10 9 m 3. Potential shale gas production from a "sweet spot" area is estimated at a range of 1.13×10 10 m 3 to 1.76×10 13 m 3 with a likely value of 8.24×10 11 m 3. Sensitivity analysis indicates that the gas production rate and cumulative gas production of a single horizontal well are most sensitive to the relative volume occupied by kerogen in the bulk volume of the shale, gas desorption rate, number of fracturing stages, and permeability of the stimulated zone. Assessment of water demand for horizontal well drilling and hydraulic fracturing suggests that shale gas development at the Xiuwu Basin will not likely cause regional water-supply stress because of abundant water resources in the region. The probabilistic approach presented in this study can provide valuable information for planning shale gas development and can also be applied to other shale gas reservoirs.

The Yangtze River is the main water source in Jiujiang City. Thus, the city faces a crisis in domestic water supply whenever the river suffers from any unexpected contamination. The availability of groundwater and the risk of pollution caused by extraction from the riverside aquifers should be determined before the groundwater can be used for emergency water supply. A numerical model of flow and transport based on the regional hydrogeological condition and 52 borehole data is developed in this study for evaluating groundwater availability and pollution risk. Results showed that the riverside aquifers can be used for emergency water supply despite the river contamination. However, the contamination had environmental effects on the water source. Thus, temporal and spatial variations of contaminants in groundwater were analyzed using the transport model. Some measures, such as pumping and injecting schemes, were suggested to protect groundwater from pollutants. An acceptable measure was also suggested to protect the water source by calculation and comparison with different methods. This research provides an important basis for proper exploitation and scientific protection and management of groundwater resource in Jiujiang City. In addition, the study has reference value to other riverside aquifers.

Abstract. The process of dam design up to the management of dam operation involves many uncertain factors, such as hydrologic, hydraulic and flood control factors, which cause risks for the flood control safety of dams. This study presents an integrated probabilistic framework that combines Monte Carlo Simulation and a flood control risk model. Results show that the highest flood level of 1000-year return periods of the Zhelin Reservoir exceeds the designed flood level. However, the overtopping risk probability is small because super safe elevation is considered in the crest elevation design of the earth dam. Sensitivity analysis indicates that flood peak flow, line type hydrological factors and flood control level are more sensitive than hydraulic factors. When many factors are considered, the comprehensive risk rate is small because of the positive and negative effects of these factors. Numerical experiments indicate that hydrology and flood control level influence the estimated maximum water level more than hydraulics does. Because of these uncertain factors, it is necessary to consider super safe elevation in dam planning and design. And pay attention to the sensitive factor of flood control level in reservoir management and operation.