Abstract
Abstract. Streamflow drought forecasting is a key element of contemporary drought early warning systems (DEWS). The term streamflow drought forecasting (not streamflow forecasting), however, has created confusion within the scientific hydrometeorological community as well as in operational weather and water management services. Streamflow drought forecasting requires an additional step, which is the application of a drought identification method to the forecasted streamflow time series. The way streamflow drought is identified is the main reason for this misperception. The purpose of this study, therefore, is to provide a comprehensive overview of the differences between different drought identification approaches to identify droughts in European rivers, including an analysis of both historical drought and implications for forecasting. Streamflow data were obtained from the LISFLOOD hydrological model forced with gridded meteorological observations (known as LISFLOOD-Simulation Forced with Observed, SFO). The same model fed with seasonal meteorological forecasts of the European Centre for Medium-Range Weather Forecasts system 5 (ECMWF SEAS 5) was used to obtain the forecasted streamflow. Streamflow droughts were analyzed using the daily and monthly variable threshold methods (VTD and VTM, respectively), the daily and monthly fixed threshold methods (FTD and FTM, respectively), and the Standardized Streamflow Index (SSI). Our results clearly show that streamflow droughts derived from different approaches deviate from each other in their characteristics, which also vary in different climate regions across Europe. The daily threshold methods (FTD and VTD) identify 25 %–50 % more drought events than the monthly threshold methods (FTM and VTM), and accordingly the average drought duration is longer for the monthly than for the daily threshold methods. The FTD and FTM, in general, identify drought occurrences earlier in the year than the VTD and VTM. In addition, the droughts obtained with the VTM and FTM approaches also have higher drought deficit volumes (about 25 %–30 %) than the VTD and FTD approaches. Overall, the characteristics of SSI-1 drought are close to what is being identified by the VTM. The different outcome obtained with the drought identification methods illustrated with the historical analysis is also found in drought forecasting, as documented for the 2003 drought across Europe and for the Rhine River specifically. In the end, there is no unique hydrological drought definition (identification method) that fits all purposes, and hence developers of DEWS and end-users should clearly agree in the co-design phase upon a sharp definition of which type of streamflow drought is required to be forecasted for a specific application.
Highlights
Drought is a creeping natural disaster that has major socioeconomic and environmental impacts across the world (e.g., Tallaksen and Van Lanen, 2004; Wilhite et al, 2007; Ding et al, 2011; Van Dijk et al, 2013; Stahl et al, 2016; Haile et al, 2019)
We present the differences of streamflow droughts identified using different drought identification approaches in two parts
For a better overview of forecasted drought characteristics in the Rhine River than in Fig. 8, we summarize all 7month forecast results done from January 2003 to December 2003 in Table 3 for daily drought approaches (VTD and fixed threshold approaches are applied (FTD)) and in Table 4 for monthly drought approaches (VTM, FTM, and Standardized Streamflow Index (SSI)-1) using the median of the ensemble members
Summary
Drought is a creeping natural disaster that has major socioeconomic and environmental impacts across the world (e.g., Tallaksen and Van Lanen, 2004; Wilhite et al, 2007; Ding et al, 2011; Van Dijk et al, 2013; Stahl et al, 2016; Haile et al, 2019). Hydrological drought deviates from meteorological drought (e.g., Changnon, 1987; Peters et al, 2003; Mishra and Singh, 2010; Van Loon and Van Lanen, 2012; Barker et al, 2016; Sutanto et al, 2020b), which means that the latter cannot straightforwardly be used to predict drought in groundwater or river flow Because of all these differences, an explicit delineation of what is being forecasted is a prerequisite. The purpose of this study, is firstly to provide a clear overview of the differences between streamflow drought using different definitions (i.e., identification methods) and temporal resolutions, i.e., daily and monthly This is done through a historic analysis using streamflow data from 1990 to 2018.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
