Actually I don't have much of an idea how some weather generators work. Many of them have no input for monthly weather data. If there any R package for that kind of work it is much better.
On Research Gate, we uploaded a Matlab code for stochastic rainfall generation through the Neyman Scott Rectangular Pulse model:
https://www.researchgate.net/publication/259310279_Neyman-Scott_Rectangular_Pulse_model_(Matlab_code)
It could help...
Luca: Can you tell me whether the output is close to a lognormal distribution? Thanks.
Dear Roger,
I am not sure that I really understood your question. Anyhow, I'm sorry but we never tested if the outputs of the Neyman Scott Rectangular Pulse model follow a lognormal distribution. You can do by yourself using our code, if you like.
.
Sorry guys, but I would say "grosso modo" it is impossible to generate daily from monthly data. Had some daily series of at least one year could create some proportionality, however, I'm sorry, but generate daily data from monthly data find an actual physical nonsense without any representability, because as the rainfall in the region these can be concentrated or dispersed throughout the month. The result of concentrated rains and scattered in any simulation of the effect of rain on another parameter is completely different.
.
I do not understand how the proposed by Luca can serve to the problem, as the applications presented in the two articles is done exactly the opposite, hourly rainfall are created monthly hydrographs!
.
Remember if there was a way to turn monthly rainfall in daily rainfall there would be only too Intensity-Duration-Frequency Curve, ie the distribution of rainfall over time would be universal!
.
Dear Luca, sorry for the late reply. This paper may help explain why I asked
On the lognormality of rain rate
http://www.pnas.org/content/84/4/901.full.pdf
I believe lognormality is observed in real observational data too.
- Badges
- Science topic
- Similar topics
- Geoscience
- Hydrology