I'd like to make sure if a GCM has enough accuracy in predicting Long-term Precipitation for a region, does it have the same accuracy in predicting the Extreme Rainfall (in the form of IDF curves, etc.) in that region?
Global Climate Models (GCMs) are indeed capable of predicting long-term precipitation trends for a region. However, their accuracy in predicting extreme rainfall events, such as those represented in Intensity-Duration-Frequency (IDF) curves, can vary12.
GCMs are designed to simulate large-scale climate patterns and may not capture local-scale phenomena such as extreme rainfall events accurately12. This is because extreme events are often influenced by local topography, land use, and other factors that may not be well-represented in the models12.
Furthermore, bias correction algorithms used to correct systematic distributional biases in precipitation outputs from climate models can artificially corrupt future model-projected trends2. This is particularly true for changes in long-period return values, which are crucial for IDF curves2.
Therefore, while GCMs can provide valuable insights into long-term precipitation trends, their predictions of extreme rainfall events should be interpreted with caution. It’s also important to note that the performance of GCMs can vary, and some models may perform better than others in simulating historical and projected future precipitation extremes1.
In conclusion, a GCM’s accuracy in predicting long-term precipitation does not necessarily guarantee the same level of accuracy in predicting extreme rainfall events. It’s always recommended to use multiple models and consider local factors when interpreting these predictions.
Thank you for the response. Did you get the information from a reliable source? If so, could you please provide it to me to have a look at it?
The spatial scale of GCMs (in the order of 100km) is too coarse for meaningful representation of extremes. The temporal scale (results stored at daily scale) is too coarse. Additionally, precipitation is a particularly difficult variable to model, first because precipitation-generating phenomena can have a variety of spatial and temporal scales, but also because of the chaotic nature of those processes. A regional climate model would get closer to what you want. What is often done is to assume that IDF curves change in the same way as daily precipitation, invoking a principle of self-similarity, though I haven't seen any good physical basis for that approach. Hopefully, someone else here knows more about that point. In any case, it is also true that precipitation intensity at sub-daily time scales can depend on things such as air quality.
Thank you all for the responses. But please notice that what I mean by GCM Accuracy is, of course, after being Downscaled (whether statistically or dynamically).
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling