The construction of the flow duration curve and the regionalization parameters analysis in the northwest of China

Daily runoff is the data to estimate the water resources in a river. In many catchments, the daily discharge is not well observed. Flow duration curve is an important characteristic of daily runoff, and important for the design of water conservancy projects. In the ungauged catchments, the evaluation of distribution functions and the parameters of flow duration curve is a helpful method to understand the characteristics of the flow. This study uses data from 19 hydrological stations to evaluate the applicability of 11 distribution functions to simulate flow duration curves in the northwest of China. The fitted flow duration curves are evaluated by Nash-Sutcliffe efficiency, the root mean square relative error and the coefficient of determination. The evaluation shows that, among the 11 distribution functions, the log normal model is the most suitable model to construct flow duration curves of 19 hydrological stations. Based on a multivariate linear regression model, a regional model of distribution parameters is constructed, including functions of watershed geomorphologic and climatic characteristics. The analysis of Baijiachuan hydrological station shows that the parameters a and b showed a decreasing trend. This study presents an innovative approach to evaluate regionalized parameters of flow duration curves considering the impacts of geomorphologic and climatic characteristics.

Advantages of calibrating a daily rainfall-runoff model to monthly streamflow data

<p>Development of robust approaches for calibrating daily rainfall-runoff models to monthly streamflow data enable modelling platforms that operate at daily time step to be applied in practical situations. Here precipitation is available at the daily scale, but observed streamflow is available only at the monthly scale (e.g. predicting inflows into large dams). This study compares the performance of the daily GR4J hydrological model when calibrated against (1) daily and (2) monthly streamflow data. The performance comparison relies on a wide range of metrics and is undertaken for 508 Australian catchments. Two evaluation periods (1975–1992 and 1992–2015) and four objective functions (including sum-of-squared-errors of Box-Cox transformed streamflow and the Kling-Gupta efficiency) were tested.</p><p>Monthly calibration performs similar to or better than daily calibration in most sites and both periods in terms of bias and fit of the flow duration curve. This result remains the same when the flow duration curve is computed at the daily time step, which constitutes a significant finding of this study.</p><p>However, the performance of monthly calibration is worse than daily calibration for daily pattern metrics such as Nash-Sutcliffe efficiency in most sites and both periods. Significant improvement can be achieved if the flow-timing parameter of GR4J is regionalised, effectively reducing the number of calibrated parameters. Similar results are obtained for other pattern metrics and all objective functions.</p><p>These findings suggest that monthly calibration of rainfall-runoff models using daily-rainfall and monthly-streamflow data is a viable alternative to daily calibration when no daily streamflow data are available.</p>

Environmental Flows Assessment in Nepal: The Case of Kaligandaki River

Environmental flow assessments (e-flows) are relatively new practices, especially in developing countries such as Nepal. This study presents a comprehensive analysis of the influence of hydrologically based e-flow methods in the natural flow regime. The study used different hydrological-based methods, namely, the Global Environmental Flow Calculator, the Tennant method, the flow duration curve method, the dynamic method, the mean annual flow method, and the annual distribution method to allocate e-flows in the Kaligandaki River. The most common practice for setting e-flows consists of allocating a specific percentage of mean annual flow or portion of flow derived from specific percentiles of the flow duration curve. However, e-flow releases should mimic the river’s intra-annual variability to meet the specific ecological function at different river trophic levels and in different periods over a year covering biotas life stages. The suitability of the methods was analyzed using the Indicators of Hydrological Alterations and e-flows components. The annual distribution method and the 30%Q-D (30% of daily discharge) methods showed a low alteration at the five global indexes for each group of Indicators of Hydrological Alterations and e-flows components, which allowed us to conclude that these methods are superior to the other methods. Hence, the study results concluded that 30%Q-D and annual distribution methods are more suitable for the e-flows implementation to meet the riverine ecosystem’s annual dynamic demand to maintain the river’s health. This case study can be used as a guideline to allocate e-flows in the Kaligandaki River, particularly for small hydropower plants.

Monthly Flow Duration Curve Model for Ungauged River Basins

Flow duration curve (FDC) is widely used in hydrology to assess streamflow in a river basin. In this study, a simple FDC model is developed for monthly streamflow data. The model consists of several steps including the nondimensionalization and then normalization in case the monthly streamflow data do not fit the normal probability distribution function. The normalized quantiles are calculated after which a back transformation is applied to the normalized quantiles to return back to the original dimensional streamflow data. In order to calculate annual streamflow of the river basin, an empirical regression equation is proposed using the drainage area and the annual total precipitation only as the input. As the final step of the model, dimensional quantiles of FDC are calculated. Ceyhan River basin in southern Turkey is chosen for the case study. Forty-two streamflow gauging stations are considered; two thirds of the gauging stations are used for the model calibration, and one third for validation. The modeled FDCs are compared to the observation and assessed with a number of performance metrics. They are found similar to the observed ones with a relatively good performance; they are good in the mid and high flow parts particularly while the low flow part of FDCs might require further detailed analysis.