![]() ![]() We are grateful to the anonymous reviewers for their valuable comments and suggestions to improve the paper. This study was funded by the Luxembourg Institute of Science and Technology (LIST). The software showed stable Acknowledgements More than half of the synthetic gaps were reconstructed with NS > 0.7. The tool was used for infilling ∼5000 synthetic gaps, of different lengths and positions, randomly created along the entire records of all stations. It was tested in the gauging network of Luxembourg to perform gap infilling on daily values, leading to satisfactory results on synthetic gaps. In this paper, we presented gapIt, a tool for infilling gaps in hydrological discharge time series. Before infilling real gaps, a first analysis was performed on synthetic gaps, in order to test the tool's capabilities and, at the same time, to build a knowledge database with synthetic gaps appropriately created and infilled. The gapIt software was applied to infill gaps in discharge times series measured at 24 gauging stations of the Luxembourgish gauge network. In the following, we designate as target station (respectively, donor station(s)) the station characterized by a gap to be infilled Results This loosens dependency on other types of variables, for instance catchment rainfall, which may not always be available (Harvey et al., 2010). It has to be noted that this approach is based on a single variable, discharge, provided as input to the software. In this section, algorithm implementation and input data requirements for gapIt will be described. ![]() High discharge values are Methods and tools Precipitation is relatively uniform throughout the year, although strong seasonality in low flow exists due to higher evapotranspiration from July to September. The region has a temperate, semi-oceanic climate. ![]() The gauge network considered here is composed of 24 stations, displayed in Fig. 1, including both very responsive and groundwater-fed rivers. The dense river network of hydrometric stations in Luxembourg offers an excellent opportunity to test the proposed tool. The same holds true for the selection of Case study Among all possible infilling methods, the choice of the most appropriate one is not a trivial task. Most of the methods proposed in the literature are based on data transfer from one or more donor stations (gauges) to a target station. infilling missing values.ĭata infilling is a challenging task that has been addressed by previous research work.įor infilling gaps in hydrological time series, classical methods of data analysis have long been applied (Salas, 1980) and recent studies have proposed more efficient techniques (Harvey et al., 2010, Mwale et al., 2012). As a result, hydrologists have access to a collection of usable tools, but they still need to deal with several technical issues (like data wrangling, tuning predictive algorithms) before solving their initial problem, i.e. But most of these tools require some data mining and machine learning expertise, as well as fine-tuning in order to meet user needs and be properly exploitable by end-users (Serban et al., 2013). For instance, specific user friendly software tools are already available or can be developed in platforms like R 1 or Matlab 2 to interpolate missing data and/or address hydrological problems. A consequence of these issues is the need of data infilling methods to reconstruct missing data, when appropriate and before hydrological time series can be used in a number of applications.įrom a technical point of view, a wide choice of data analysis tools is nowadays offered to hydrologists. It also limits the use of such data for hydrological or hydrodynamic model calibration/validation purposes. Specifically, the presence of discontinuities precludes the computation of hydrological statistics and physiographic indices. The existence of gaps results in difficulties in data interpretation and is a large source of uncertainty in data analysis. Missing data in river flow records represent a loss of information and a serious drawback in water management. They are an inevitable consequence of factors such as station maintenance, equipment malfunctioning, human errors, changes in instrumentation and data processing issues (Harvey et al., 2010). Rather, time series of hydrological data are often affected by data gaps, which are discontinuities in the record of data. Long uninterrupted hydrological time series are often not available for many of the stream gauges in the world. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |