Non-stationary Regional Weather Generator (nsRWG) is a multivariate autoregressive (MAR-1) stochastic weather model that can generate arbitrarily long synthetic time series of weather variables including precipitation and temperature at several locations. The spatial and temporal dependence of precipitation and temperature at different locations is taken into account, so that consistent spatio-temporal patterns of weather variables are generated that correspond to the observations in terms of their statistical properties.

The non-stationary model nsRWG (Ngueyen et al., in review) is based on the stationary model of Hundecha et al. (2009), which was extended and evaluated by Nguyen et al., (2021). In contrast to the stationary versions (Nguyen et al., 2021), precipitation generation in nsRWG is now conditioned on circulation patterns as a latent variable that characterizes changes in atmospheric dynamics. In addition, the regional average of mean daily temperature, which characterizes thermodynamic changes, is used as a covariate for the non-stationary probability distributions at individual locations. nsRWG can be set up on both a station and raster basis. The model works on a daily resolution. However, precipitation time series can be disaggregated to hourly values if required using a disaggregation model that maintains spatiotemporal consistency (Guan et al., 2023).

Publications:

Guan, X., Nissen, K., Nguyen, V. D., Merz, B., Winter, B., & Vorogushyn, S. (2023). Multisite temporal rainfall disaggregation using methods of fragments conditioned on circulation patterns. Journal of Hydrology, 621, 129640. doi.org/10.1016/j.jhydrol.2023.129640

Nguyen, V. D., Merz, B., Hundecha, Y., Haberlandt, U., & Vorogushyn, S. (2021). Comprehensive evaluation of an improved large-scale multi-site weather generator for Germany. International Journal of Climatology, 41(10). https://doi.org/10.1002/joc.7107

Nguyen, V. D., Vorogushyn, S., Nissen, K., Brunner, L., & Merz, B. (2024). A non-stationary climate-informed weather generator for assessing future flood risks. Advances in Statistical Climatology, Meteorology and Oceanography, 10(2), 195–216. https://doi.org/10.5194/ascmo-10-195-2024

Ullrich, S. L., Hegnauer, M., Nguyen, D. V., Merz, B., Kwadijk, J., & Vorogushyn, S. (2021). Comparative evaluation of two types of stochastic weather generators for synthetic precipitation in the Rhine basin. Journal of Hydrology, 601, 126544. doi.org/10.1016/j.jhydrol.2021.126544