STARS - STatistical Analogue Resampling Scheme
What does the model do?
The STatistical Analogue Resampling Scheme (STARS) is based on the assumption that already observed weather situations will very likely recur in the same or similar way in the near future. Therefore, the model rearranges observed time series with respect to a given linear trend for a selected variable. In most cases the trend is prescribed for the mean temperature as this is a very robust signal in most regions of the world. However, it is also possible to choose another meteorological variable instead. The annual means of the simulated time series have to match the given linear trend for the simulation time period. The simulation is done by rearranging the input data (either station observations or gridded reanalysis data). Thus, the approach generates a mapping from dates of a observation period to dates of the simulation period. A set of heuristic rules makes sure that the resulting time series show a realistic behavior, such as annual cycle or weather persistence.
Stars-user-manual.pdf (460 KB)
Who maintains it?
What are the next steps in the development of the model architecture?
- Extension of the constraints to circulation pattern statistics in combination with flexible segment lengths for the date-to-date mapping
- Application for seasonal climate projections by prescribing the 3-monthly mean temperature trend
- Model comparison with dynamical downscaling methods using climate network analytics.
In what way is the model different from other models in the community?
STARS does not need a GCM as a driver. It delivers not only meteorological projections but any kind of interesting observations, e.g. river runoff or circulation indices, if they are provided for the observational period.
Hoffmann, P., Österle, H., Gerstengarbe, F.-W., Holldorb, C., Rumpel, F. (2014): The effects of climate change on winter service in Germany. In: Routes/Roads, 361, 30-37
Gerstengarbe, F.-W., Werner, P. C., Österle, H., Pardowitz, T., Burghoff, O. (2013): Winter storm- and summer thunderstorm-related loss events with regard to climate change in Germany. Theor. Appl. Climatol., Published online: 15 February 2013, DOI 10.1007/s00704-013-0843-y
Lutz, J., Volkholz, J., Gerstengarbe, F.-W. (2013): Climate projections for southern Africa using complementary methods. International Journal of Climate Change Strategies and Management, Special issue on climate change in Africa, 5, 2, 130-151, ISSN 1756-8692
Orlowsky, B., Bothe, O., Fraedrich, K., Gerstengarbe, F.‐W. & Zhu, X. (2010). Future climates from bootstrapped weather analogues: an application to the Yangtze river basin. Journal of Climate, 23: 3509‐3524
Orlowsky, B. & Fraedrich, K. (2009). Up‐scaling European surface temperatures to North‐Atlantic circulation pattern statistics. International Journal of Climatology, 29: 839‐849
Orlowsky, B., Gerstengarbe, F.‐W. & Werner, P. (2008). A resampling scheme for regional climate simulations and its performance compared to a dynamical RCM. Theoretical and Applied Climatology, 92: 209‐223
Werner, P. & Gerstengarbe, F.‐W. (1997). Proposal for the development of climate scenarios. Climate Research, 8: 171‐182
Some remarks on application of model STARS and its climate scenarios:
- The data calculated by STAR were used as input for 11 models in the project GLOWA-Elbe. From that, a considerable number of papers were published by project members which used the STAR simulations (see also www.glowa-elbe.de/german/publications_wiss.htm).
- Moreover, STAR was applied in the projects WAVES, KLIWA, KLARA and in different regions of Germany (BW, NRW, Brandenburg) and the world (China, NE-Brazil, Israel/Palestine, Tajikistan, Czech Republic).
- Actual, STAR is part of the projects GLOWA-Elbe, CLIMREG - Climate Impacts Register for Germany, “Auswirkungen des Klimawandels auf die Schadenssituation in der deutschen Versicherungswirtschaft“ and of about 15 project proposal.
- STAR-data were/are used at several research institutes and universities.