Towards a global high resolution water demand dataset: Effect of data quality and downscaling techniques - the case for Europe

Abstract

The development of water resource models requires reliable projections of gross and net water demand, and the ambition to bring these models to hyper-resolution makes this requirement even more important. The problem arises that high-resolution water demand data are universally missing resulting in absence of relevant studies that assess global water demand at high spatial resolution. As water scarcity threatens sustainable, economical and technological development, as well as worsen conditions for the urban poor there is a need to explore the underlying drivers, which are water availability and demand. This study focuses on two sectors that contribute to total water demand, households and industries. The main aim of this research is to set up a flexible framework to define high-resolution water demand for households and industries in Europe, using existing downscaling concepts and datasets with the final objective to develop and improve high-resolution global water demand estimates that may benefit from forthcoming data sources. A conceptual top-down approach on global water demand at 5 arc minute is used as a foundation, developed by Wada, which is changed to a high-resolution method by classifying possible improvements. These improvements are based on the methods of existing water demand approaches, with special focus on the pan-European empirical study of Bernhard to assess water use at high resolution. Possible improvements are conceptual changes, increase in resolution of used variables and changing the downscaling technique. Increasing the resolution of variables adds to the spatial distribution of domestic and industrial water demand in a country, which before was only done by the downscaling technique. GDP (main driver for industrial water demand) and population (main driver for domestic water demand) are increased and add to the spatial distribution within-country water demand. The downscaling techniques improve the performance of the method and are essential components in reaching a high-resolution dataset. A high-resolution water demand method for households and industries was constructed with a spatial resolution of 30 arcsec. This method performed significantly well for simulating both gross and net water demand in Europe for households, which was spatial distributed along population density. An industrial land cover fraction downscaling method was used as well as a population map for industrial water demand, and both had nearly the same performance.