Standard years for large-scale hydrological scenario simulations

Scenario analyses are regularly characterised by a large number of degrees of freedom. It is usually unfeasible to perform sophisticated hydrological simulations with continuous long-term meteorological time series for all combinations of scenarios and adaptation strategies. To reduce computation time while retaining sufficient degrees of freedom, a so-called “standard year” has been developed. Average hydrological conditions according to simulations driven by this standard year should optimally resemble results from full 30-year simulations. Therefore, the standard year should optimally represent intra-annual variability. The objective of this paper is to explore how the errors, introduced by using standard years compare to natural variability. In addition, the standard year was also tested for future climate scenario simulations.The standard year was constructed by perturbing meteorological quantities recorded during a selected meteorological year. The selection of this year was based on precipitation (P) and reference crop evapotranspiration (Eref) in 1967, due to the good resemblance of the potential cumulative precipitation deficit in the Netherlands (Eref − P) during April–September in this year with the mean climatology. Subsequently, the 1967 precipitation and reference crop evapotranspiration was modified such that the standard year optimally captured intra-annual modes of variability. This is done by the application of spatially varying correction factors, which set the 2-monthly precipitation and reference crop evapotranspiration sums equal to climatology.As a standard year only contains one year of meteorological data, it has no interannual variability by construct. As a consequence, extreme daily events are biased and small but systematic errors are simulated in the major hydrological budget terms. Surface runoff appeared oversensitive to the underestimated number of heavy rainfall events and was considerably underestimated by standard year simulations. The simulated climate change response of the major hydrological terms was reproduced very well and the error introduced by the standard year methodology appeared much smaller than the climate change signal.

Research highlights
► A meteorological standard year is developed for current and future climate.
► Hydrological simulations with standard years compare well to full 30-year simulations.
► A standard year has no interannual variability and extreme daily events are biased.
► Small, systematic errors are simulated in the major hydrological budget terms.
► Errors due to the proposed methodology are much smaller than simulated climate change.