Interpolating daily rainfall to generate hourly rainfall is difficult without some assumption on the distribution of the events within the day. Weather generators assume a distribution (mostly gamma) to generate rainfall series. If synthetic weather generator is not an option, other hourly observation can be used to develop some rainfall distribution pattern within a day. But this could be unreliable as rainfall intensity and incidence can vary with a day and can't be generalized. If you plan to use weather generator, try to simulate multiple series similar to stochastic modeling.  

Hi M.L. Gaur,

You can do a temporal disaggregation of the time series to transform daily to hourly rainfall. In R you can use the package 'tempdisagg'. More information on the techniques available for this task can be found here: http://journal.r-project.org/archive/2013-2/sax-steiner.pdf.

Hope it helps.

Regards,

Augusto

It is not an easy task, however, depends on what you're looking for. I recommend you first do an analysis considering all the rain falls in an hour and another considering that the rain falls evenly throughout the day. You can observe significant changes in what you want to model? If you want to model the hourly flow in a basin Mediterranean climate, I do not think is a problem, however, if you want to model the structure of recharge in an aquifer in an arid or semi arid area itself will be important rain level schedule. If you realize that the time is really important rain, I recommend looking for some instrumental record in a nearby basin or similar, albeit small extent. With this, you can replicate the temporal structure of the hourly measured data in your data on a daily basis, not forgetting adjusting for scale.

Good Luck 

Best Regards, 

Gino 

Thanks Timothy for suggesting distribution (gamma)

Thanks L. Augusto Sanabria for a nice advise for opting a temporal disaggregation of the time series to transform daily to hourly rainfall. In R and giving an important document

Thanks Gino Sturla Zerene for caring the question asked by me

A good option would be to use satellite-based rainfall estimate, such as GSMaP product as it covers hourly and global rainfall intensity. These satellite products are often biased (over- or underestimation), but temporal fluctuation at sub-daily scale is well captured. Thus, using your daily record, you can adjust the intensity bias and capture the hourly variability of rainfall. 

Thanks Masahiro

yes it could be a smart way to opt reverse way to reach destination. Can you help me to have some such stuff for indian continent or the source and ways to retrieve it 

Hi Gaur,

You may want to first register to get an access to the data: http://sharaku.eorc.jaxa.jp/GSMaP/index.htm

Then you can download hourly rainfall estimates at global scale and clip the data to narrow down to your target region.

Thanks a lot Ryo

its excellent help . I tried right now. I have registered but facing 2 problems. first the language shown on screen is other than English and second the exact procedural steps to retrieve the desired hourly data in numerical form or digits say in excel..

I am sorry i am bothering you a lot,,but it looks that i am moving towards one good solution or answer to mine intended question..

thanks

Gaur

You may find the language other than English slightly, but is not it mostly written in English? I think most of this kind of global data requires you to convert from a specific format (e.g., gz or TIFF) to a sort of excel data sheet by yourself using GIS and an automating script... If you are unfamiliar with this kind of data processing, maybe you need to find another option. 

thanks a lot Masahiro

RMC (Random Multiplicative Cascade) model is also an option to disaggregated daily rainfall into hourly rainfall. You can google about it e.g.

https://www.researchgate.net/publication/223276406_Rainfall_stochastic_disaggregation_models_Calibration_and_validation_of_a_multiplicative_cascade_model

I have tested this approach in some of the stations in South India. This approach work reasonable to get the reasonable estimate of distribution at hourly scale.

Article Rainfall stochastic disaggregation models: Calibration and v...

So nice of you Sat

I am extremely thankful

an excellent paper ..

thanks again

let me read ..me not so smart statistician..lol

but great help from you dear

Hello Gaur,

Regarding your question, I am of a view that it may not be possible to dis aggregate your daily data to hourly data.

In that case, how were you able to solve the problem?

Philip you are right but stilli have a hypothesis in my mind that certainly some kind of distribution shape function could be evolved by seeing the generalized trends at nearby or other mass observed data at varied locations if made available ..

Yes Gaur. I think that is indeed true.