Guidance on the use of meteorological time series constructed to match the KNMI’23 climate scenarios

In the previously issued KNMI climate scenarios Klein Tank et al. (2014), time-series of weather observations were transformed to time-series representative of scenarios for future climates, using climate change signals (such as differences in monthly percentiles of temperature) extracted from climate model projections. These data were used to produce various data products linked to the scenarios, and the timeseries themselves were made available for the assessment of the impacts of the climate scenarios. In addition, external users were enabled to perform this transformation on their own time-series, either off-line (by a stand-alone program) or online through
the KNMI Climate Explorer.

For the KNMI ’23 climate scenarios van Dorland et al. (2023), a different approach has been adopted: in addition to time-series of transformed observations, also bias-corrected climate model projections are provided; see van Dorland et al. (2023). The bias-corrected climate model projections are available for all 12 km ⇥ 12 km grid cells of the RACMO regional climate model. They have also been used to compute the climate change signal in the scenario tables; see van Dorland et al. (2023), Section 2.1.11.

Bias-corrected climate model projections are suitable for a range of applications, but not for all. Therefore, in addition, observed time-series transformed to match the climate change signals corresponding to the four scenarios considered are provided for the sites with daily precipitation measurements, with data of other variables taken from the nearest automated weather station.

The two datasets (bias-corrected model projections and transformed time-series of observations) each have their strengths and weaknesses. For each application, these need to be considered carefully before deciding which dataset is to be used.

This document provides guidance for this decision, as well as a little background on the methods used to provide the data and how they differ from the methods used previously for transforming observed time-series (see KNMI (2015). Further explanation of the bias-correction of model projections and the transformation of observed time-series can be found in the Appendix of this document, which reproduces relevant sections from van Dorland et al. (2023). General guidance on the use of the KNMI ’23 climate scenarios can be found in Chapter 11 of van Dorland et al. (2023).