Hello,

First, as the hydrology depends on the region, it is not the same to estimate the runoff in arid zones than in Mediterranean areas, for example. Also important is the hydrological regime in the basin.

If we assume that the basin has a pluvial regime and also that the monthly rainfall is significantly correlated with the monthly runoff and you want to explain the monthly runoff, you hasve to use the monthly rainfall, which corresponds to the sum of the daily rainfall of  precipitation. In general, almost always, the resolution of the input is always the same as the output.

Another case is, for example, if you know that the basin does not accumulate snow or groundwater from one hydrological year to another, and also precipitation correlates with runoff, a great way to estimate annual runoff is using annual rainfall, The sum of 365 days. It could also occur in watersheds with a snow melt regime.

It is very difficult to estimate the dialyzed runoff, much information is needed from the basin, almost always the model will be more specified: a large number of parameters.

Now, if you want to estimate runoff by using a continuous time series for rainfall, you can use continuous transfer functions. That is, you have an input function, precipitation, and an output function, runoff. It is not usual in hydrology, but if its transfer function represents the process in a good way, and does not have many parameters, its model would be adequate. You can also use a discrete time process, in Matlab there are good tools to manage the transfer functions.

Definitely, in discreet time, you should always use the same resolution for the precipitation and runnof, maybe your model will have delays, but all with the same length of time; And we always have to add the precipitation "requires an input volume and an output volume and, in continuous time, the resolution is not a problem, because we will work with integrals, perhaps also with lags, depending on the climatic region and The physical characteristics of the basin.

Please, I do not know your case, there are a lot of other variables that could be important for your odel selection challenge. However, i can say you, that always, in discrete time you have to use the same resolution for both variables and if you work in continuous time this is not a problem.

I hope, my words could help you. If you need more specific recomendations, you can write me to [email protected], and tell me more about the hydrological process that you want to modelate; sure i could help you in a better way.

Best regards,

Gino

helo,

its my perception, if you have rainfall for say 1 hour duration and if there will be a gap of more than 1 hour between next rainfall,

than it can be treated as single rainfall event otherwise continuous

Your question is related to event-based and continuous simulation of runoff simulation. For event-based modelling, your storm event depends on the problem you are studying. Say you are interested in quantifying monthly runoff from daily precipitations- in this case, you can accumulate the daily rainfall to get a monthly total and use it to calculate monthly runoff. If you want to simulate annual runoff from annual precipitation, you can accumulate daily precipitation to get annual total and use it in annual runoff simulation. In most continuous simulations, you can use daily precipitation in a runoff equation to generate daily runoff series. In this case, when you have no precipitation, the runoff will also be zero. A popular method you could explore is SCS Curve Number methods.

If rainfall occurs continuously for 2 or 3 days then it would be consider as continuous rainfall event and else it would be discrete events.

Regarding runoff same has to be follow based on hydrograph pattern.

Please go through the following article hope you may get something.

http://www.hydrol-earth-syst-sci-discuss.net/hess-2016-129/hess-2016-129.pdf

http://www.tandfonline.com/doi/pdf/10.1080/02626668809491261

http://and.lternet.edu/lter/pubs/pdf/pub3919.pdf