Abstract: Hybrid data assimilation (DA) is a new method used in recent hydrology and water resources research. In this paper, a DA method coupled with the support vector machines (SVM) and the ensemble kalman filter (EnKF) technology is used for the prediction of soil moisture at different soil layers: 0 cm, 30 cm, 50 cm, 100 cm, 150 cm and 200 cm. SVM method is a statistically sound and robust approach for solving the inverse problem by building statistical models. So far, SVM has a great use in such problems to classify or predict data which often contain some useful information. The principle strength of this machine lies in the use of Structural Risk Minimization (SRM) rather than Empirical Risk Minimization (ERM). EnKF is an extension of the kalman filter, a well-known method for updating information. It is one of the mostly used sequential DA methods in recently land data assimilation research. Herein, the SVM methodology is firstly used to train the ground measurements of soil moisture and meteorological parameters from Meilin study area to construct the soil moisture statistical predictor models. Then the subsequent observations and their statistics were used for the future predictions by using two approaches: SVM predictor and SVM-EnKF model by coupling SVM model with EnKF technique using DA method. Validation results showed that the proposed SVM model coupled with EnKF technology can effectively improve the predictions of soil moisture in different layers, from surface to root zone.
Abstract: In order to assess the effects of calibration data length on the performance and optimal parameter values of hydrological model in ungauged or data limited catchments (actually, data are non-continuous and fragmental in some catchments), we choose to use non-continuous calibration periods to have more independent streamflow data for SIMHYD model calibration. Nash-Sutcliffe efficiency (NSE) and percentage water balance error (WBE) are used as performance measures. The Particle Swarm Optimization (PSO) method is used to calibrate the rainfall-runoff models. Different length of data range from 1 year to 10 years randomly sampled used for study on impact of calibration data length. 55 relatively unimpaired catchments all over Australia with daily precipitation, potential evapotranspiration (PET), and streamflow data are tested to obtain more general conclusions. The results show that, longer calibration data does not necessarily result in better model performance. In general, 8 years data are sufficient to obtain steady estimates of model performance and parameters for SIMHYD model. It is also show that most humid catchments require fewer calibration data to get good performance and stable parameter values. The model performs better in humid and semi-humid catchments than arid catchments. Our results may have useful and interesting implications in the efficiency of limited observation data used for hydrological model calibration in different climatic catchments.
Abstract: Isotopic fractionation is the foundation of tracing water cycle using hydrogen and oxygen isotopes. Isotopic fractionation factors in evaporation from free water body are mainly affected by temperature and relative humidity, and greatly vary with these atmospheric factors in a day. Evaporation rate can properly reveal the effects of atmospheric factors. Therefore, there should be a certain function relationship existing in isotopic fractionation factors and evaporation rate. An average isotopic fractionation factor was defined to describe isotopic differences between vapor and liquid phases in evaporation with time interval of hours or days. The relationship of average isotopic fractionation factor and evaporation based on isotopic mass balance was investigated through an evaporation pan experiment with no inflow. The experimental results showed that the isotopic compositions of residual water became more enrichment with time; the average isotopic fractionation factor was affected by air temperature, relative humidity and other atmospheric factors, and had a good functional relation with evaporation rate. The values of average isotopic fractionation factor could be easily calculated with the known of evaporation rate, the initial volume of water in pan and isotopic compositions of residual water.
Abstract: Three high-resolution satellite precipitation products, Tropical Rainfall Measuring Mission (TRMM) Standard precipitation products (3B42V6 and 3B42RT) and NOAA Climate Precipitation Center (CPC) morphing technique precipitation products (CMORPH), were evaluated against dense surface rain gauge measurements over Laohahe Basin in northern China. Widely used validation statistical indices and categorical statistics were adopted in the study. The evaluations were performed at multiple time scales, ranging from daily to annual, for the years from 2003 to 2008. The results show that all the three satellite precipitation products perform very well in detecting the occurrence of the rainfall events, while there are some different biases in rainfall amount, the 3B42V6 having a bias of 19.6% fits the best with the gauge observations both at daily and monthly scales, while, the biases of 3B42RT and CMORPH (with the values of 78% and 65.7%, respectively) are much higher than a normal receivable threshold. The quality of the satellite precipitation products also have a monthly and yearly variation: 3B42RT has a big positive bias in cold months (Sep. to Mar.), while CMORPH has a big positive bias in warm month (May to Aug.), and they all get the best values in 2006 (with 6%, 41% and -6% biases for the 3B42V6, 3B42RT and CMORPH, respectively). Our evaluation shows that, over the Laohahe Basin, 3B42V6 has the best corresponding with the ground observations , and CMORPH reveals a litter better than 3B42RT. The high errors of the real-time satellite precipitation products (i.e., 3B42RT and CMORPH) reminds us that some new quests for improving the precision of the satellite precipitation products must be proceeded.
Abstract: High-quality rainfall information is critical for accurate simulation of runoff and the water cycle process on the land surface. In situ monitoring of precipitation has a very limited usefulness on the regional and global scale because of the high spatial and temporal variability of precipitation. It is unlikely to improve rainfall station network and rainfall radar network in short term in all regions. The inaccuracy of rainfall in some areas with sparse rainfall stations is bound to bring great uncertainty into the runoff and water cycle process simulation. On the way to overcome these problems, microwave remote sensing observations are appropriate for retrieving the spatial and temporal precipitation coverage because of the global and frequent availability. With the development of the satellite radar technology, TRMM precipitation data get more and more attention by meteorologists and hydrologists because of its high temporal (3 hours) and spatial (0.25°×0.25°) resolution. For distributed hydrological model, the distributed input data is the basic condition for hydrological simulation. This paper addresses the question of whether remotely sensed precipitation estimates over a catchment can be used to improve distributed hydrological model water balance computations. The TRMM (Tropical Rainfall Measuring Mission) precipitation product was introduced into the hydrological cycle simulation for Yangtze basin, in south China. A tool was developed to interpolate the observations into the same spatial and temporal resolution with TRMM data and then evaluate the precision of TRMM data from 1998 to 2006. It shows that TRMM rainfall in the moderate-intensity with high accuracy, but low-intensity and heavy-intensity rainfall with low fidelity. So the integrated application of TRMM data and ground observations was proposed to the watershed hydrology process simulation and achieved good results.
Abstract: Abstract: When subway tunnels are routed underneath rivers, riverbed scour may expose the structure with potentially severe consequences. Thus it is significant to research the maximum scouring depth to ensure the designed buried depth is adequate. There are a range of methods that may be applied to this problem, including fluvial process analysis, geological structure analysis, scouring formula, scouring model experiment and numerical simulation. However, the applicable range and forecasting precision of these methods vary considerably. In order to quantitatively analyze the characteristics of the different methods, a subway passing underneath a river is selected with the aforementioned five methods used to forecast the maximum scouring depth. The research results show that the fluvial process analysis method was used to characterise the river regime and evolution trend, which were the baseline for researching the scouring depth of the riverbed. The results obtained from the scouring model experiment and the numerical simulation method were reliable, these two methods were suitable to be applied to subways or tunnel projects passing under rivers. The scouring formula method was less accurate precision than the souring model experiment, so it was suitable to be applied to the “lower risk” projects as pipelines. The result of the geological structure analysis was of low precision, and it was suitable to be an assistant research method. To forecast the maximum scouring depth of the riverbed above the subway tunnel, a combination of methods was suggested, and the appropriate analysis method was chosen with respect to the local conditions.
Abstract: Toward solving the actual operation problems of cascade hydropower stations under hydrologic uncertainty, this paper presents the process of extraction of statistical characteristics from long-term optimal cascade operation, and proposes a monthly operation function algorithm for the actual operation of cascade hydropower stations through the identification, processing, and screening of available information during long-term optimal operation. Applying the operation function to the cascade hydropower stations on the Jinshajiang-Yangtze River system, the modeled long-term electric generation is shown to have high precision and provide benefits. Through comparison with optimal operation, the simulation results show that the operation function proposed retains the characteristics of optimal operation. Also, the inadequacies and attribution of the algorithm are discussed based on case study, providing decision support and reference information for research on large-scale cascade operation work.
Abstract: In recent years, rural drinking water safety has become a growing concern in China. The present study aims to the core problems in rural drinking water safety in Mingshan County, Ya’an, China where groundwater is used as drinking water. The objective of this study was to determine groundwater quality and to assess the health risk of pollution in groundwater in Mingshan County, Ya’an, China. Over the 19-year period from 1991 to 2010, the total number of groundwater samples was 47.This study inspected source water quality, applied the health risk assessment model recommended by U.S.EPA, and calculated carcinogenic risk, non-carcinogenic risk and total risk of factors exceeding standards. Health risk thematic maps of carcinogenic risk and non-carcinogenic risk caused by single factor and their combination can be obtained with the method of inverse-distance weighting interpolation (IDW) and the function of geo-statistical analysis in ArcGIS (Geographic Information System, GIS). The results showed, carcinogenic risk caused by chemicals in drinking groundwater is low, which is in the acceptable interval in risk management both at home and abroad for years. However, non-carcinogenic risk is high and number of sampling sites with risk values exceeding the standards amounted to 30, in which, non-carcinogenic risk of Cr (6+), nitrate, fluoride and Fe of number 43, 46, 74, 64, 67, 50 were the sources that cause total health risk high. This study revealed the risk level of drinking groundwater quality, orders of treatment on pollutants and provided a scientific basis for groundwater management in this area.
Abstract: Wetlands are almost the exclusive source of natural resources upon which rural economic depend. In fact wetlands are generally highly productive eco-systems, providing many important benefits. These benefits are mainly flood control and ground water recharging and pollution reduction and act as filtering aquatic system. Wetland, however, contains numerous goods and services that have an economic value not only to the local populations, but also to people living outside the periphery of the wetland. Stakeholders’ participation is very essential to protect and preserve wetland because it plays a very important role economically as well as ecologically in the wetland system. The objective of the present paper is to analyze whether there is any dependency of the gender, educational status, mouzas (Mouzas are constituents of a block according to land reform of West Bengal Government in India) and wetland functions on annual income of the local community. Considering a flood-plain wetland in rural India, the focus is extended to recognize the pattern of wetland functions according to the nature of involvement of people through cluster analysis for the male and the female population. Using the statistical software, R-2.8.1, ANOVA table is constructed. Since the p-value is lower than 0.5 for each case, it can be remarked that there is a significance dependency of gender, educational status, mouzas, and wetland functions on annual income. However, S-plus-2000 is applied to get a complete scenario of pattern of the wetland functions, in terms of involvement of males and females, through the cluster analysis. The main conclusion is that, the significance effects of gender, educational status, mouzas, and wetland function on annual income have been found to be satisfactory while the pattern of occupation of the local community based on wetland functions is completely different for the male and the female population.
