Improve assignment of lakes to regine catchments
Closed this issue · 2 comments
NVE's lakes database assigns each lake to a unique regine catchment. This is useful, because large lakes may cover several regines, but for the accumulation calculations we need to assign each lake to a single catchment.
At present, TEOTIL3 uses the "official" regine assignments from NVE for each lake. In general these assignments seem sensible, but there are some cases where a lake is assigned to a seemingly random catchment intersecting the lake. I am not sure what criteria NVE have used but, for example, Tyrifjorden is assigned to regine 012.D71, which is part of a side catchment, rather than the "main stem". As a consequence, Tyrifjorden has a smaller effect on accumulated retention than might be expected.
Although using the official NVE designations seems reasonable, it might be worth exploring alternatives. The most sensible option is probably to assign each lake to the downstream-most regine that intersects it. This should not be too difficult, but we'd need to carefully explore the implications for accumulated fluxes.
I have been in contact with NVE about this, and it seems there are some errors in the lakes database, mostly caused because the regine dataset has been updated but the lakes database not adjusted to match. See e-mail from Stian Lundanes at NVE for details (received 07.12.2023).
Stian says that NVE will try to fix the errors as part of their annual update. I have also added a file here to TEOTIL3, which makes it possible to manually assign lakes to regines. Updating this file and then re-running the notebooks for Tasks 2.4 to 2.7 should modify the assignment of lakes.
This is unlikely to be resolved by NVE in the near future and the workaround is OK. Closing.