As you are aware from your studies, rainfall is not always well distributed, so having a network of raingauges is helpful in evaluating differences relative to time and space of individual storms, and effective totals to your watershed.  Due to detail, it is often necessary to effectively stratify and combine the data into monthly or annual amounts if not studying specific hydrograph reponses.  Topography can have a big influence on rainfall totals as well as normal storm movement pattern, as compared to abnormal instances such a cyclonic or thunderstorm events of rare intensity and complexity.  

The answer to your question may be in how you intend to use data, such as annual water balance using monthly rainfall, streamflow, ET data or generated estimates.  Or your intent may be evaluating how the watershed responded to a rare rainfall event where more detail of event specifics become important.  There are various methods that can spatially assimilate rainfall from varying stations such as krieging that can help automate spatial mapping of individual events to annual totals from a network from rain gauges, and applying that information to the watershed and topography of interest, and this can be done in a GIS format.  Be thankful, in the old days, we had to hand plot these data and spatially estimate the rainfall pattern across the landscape or watershed.  

Sometimes you may not know the amount of data you need until it is analyzed or have a specific reason for using a model other than making it work.  Generally the data within the watershed will be more important than locations at substantial distances, but the presence of highly variable event totals can help signal the presence of extreme conditions.  If the goal is to enter data and get a model output, that is going to be different from using the data to support some analysis to validate the use and application of the model for your dissertation or technical article.  

In the model description, you may find direction or hints on how it was developed, assumptions, limitations, and uses.  HEC-HMS was developed with a variety of uses, whether individual storms and unit hydrograph to longer term analyses such as water balance components.

thank u for your reply,it's very helpful.. My purpose is to realize he hydrograph of my watershed so i 'd like to ask u if my methodology is right or not in deallig with rainfull data.

first i choose the best staions by knowig the correlation between each others:Second i realize the hyetogramm for each station(in this step i have to choose daily or monthly data?) .Then i achieve the average hyetogramme using the Thiessen polygon's method. Then  i have to study the spatial variability of rainfull with interpolation (IDW,krigeage..) but how can i use it or enter it in hec hms on WMS?

Dear,

The choice of a time interval of yours precipitation data depends on the aims of research or study. If it is about rainfall reservoir storage, you can use monthly data. If you need rainfall to analyze drainage issues in watersheds with a small time of concentration, you should use daily rainfall data.

Regards

i agree with Hansen

The time interval of rainfall data will depend on your modeling objective. If you are trying to develop a water resource model, better to go for daily data and if it is a flood model, you can select hourly or even smaller time interval.

Je travaille sur le phénomène d'inondation et j'ai trouvé une difficulté dans la saisie des années pluvio pour le modèle, si quelqu'un connait ce modèle,pourriez vous m'aider ?