There are already some wats to disaggregate single-event rainfall to hourly data.
I'd like to know if there are any simple methods for disaggregating LONG-TERM rainfall data to hourly rainfall.
Because they say disaggregation of Single-event rainfall is completely different from Long-term (continuous) data of rainfall.
Can I use SMADA software to do that? (I mean for disaggregating long-term rainfall)?
Dear Negin
Disaggregating daily rainfall into hourly rainfall is often challenging. To do this you need to have some idea about the relationship between the one hour rainfall and the 24 hour rainfall. This relationship will depend on a number of factors (e.g. where you are in the world, type of rainfall etc.). It will also depend on what you are using the aggregated one hour rainfall for. For example, a rainfall disaggregation model developed intended to be used only for design flood purposes would probably focus on disaggregating larger daily rainfalls and may not be suitable for other purposes. You’ll need to have some rain gauges where sub-daily rainfall are available to develop a relationship between the daily and hourly data. If you don’t have evidence that is based on actual sub-daily rainfall measurements from a country or region with a similar climate to where you are I would say it is hard to disaggregate daily rainfall to hourly rainfall with any confidence. For example, a method or mathematical function to disaggregate daily to hourly rainfall for an arid area such as many countries in the Middle East would be very different to one that you would use in the United Kingdom.
You ask if the SMADA (Stormwater Management and Design Aid) can be used to disaggregate rainfall. I must confess that my knowledge of SMADA is limited; however, the documentation that I have read would indicate to me that you can only disaggregate daily rainfall using dimensionless rainfall curves developed for the USA. (Please note I stand to be corrected as to whether this is actually the case!).
It would be worth you looking at the following papers:
https://www.tandfonline.com/doi/pdf/10.1623/hysj.54.2.217 https://pdfs.semanticscholar.org/40e6/31dee9a0c83c44e9926789b0b0334c74603d.pdf https://andrewsforest.oregonstate.edu/sites/default/files/lter/pubs/pdf/pub3636.pdf
Best regards
Darren
As Darren mentioned, dis-aggregation is always challenging. I have used "Random cascade dis-aggregation model" to dis-aggregate Daily and 6- hourly precipitation into hourly precipitation data in my research, and it worked well.
This method first divides each precipitation amount into two equal sub-time intervals between two cascade levels. Then, the value of precipitation of each time intervals are determined by multiplying the total rainfall amount with weights W1 and W2 being calculated as follows;
📷
For further details please refer to attached "RANDOM CASCADE MODEL: documentation"
Hope this helps.
@Darren has given a good answer but doesnt quite go far enough. Basically the answer is that you cannot disaggregate daily total rainfall to hourly or sub-hourly data or link to rainfall intensity with any degree of accuracy without some other real sub-hourly data from a similar location to model with.
This is a painstaking exercise. However, as suggested by previous answers, you certainly need to compare your results with other external hourly dataset. This will ultimately assist you in calibrating whichever model you're using for your disaggregation. This is a proof of accuracy in its long term. Else, it becomes a routine exercise.
Thank you for the answers.
I continued this discussion in this page:
https://www.researchgate.net/post/Is_this_rainfall_disaggregation_method_reasonable
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series