How can i model rainfall and flood frequency?

You could start by looking at the daily rainfall models developed by Zucchini and Adamson for South Africa. They are based on a first order Markov chain approach seasonally adjusted for particular sites. Using these it is possible to calculate rainfall exceedence probabilities. Then you can go on to hydrological models to get estimates of flood exceedence probabilities. https://journals.co.za/doi/pdf/10.10520/AJA00382353_9448 I note that these models have some difficulty with specific weather events such a extreme cut-off lows which can produce floods such as those experienced in Spain in 2024.

the approach you take will depend on the scale of your study. Flood flows from Large basins are best modeled using historic flow data from gauged rivers in the region. Smaller basins can be modeled with gauged data from nearby basins, or by using rainfall data with a watershed tool. Small watersheds (under 1000 ha) are best modeled with rainfall data and runoff modeling tools. For rainfall data, if there are no nearby weather stations, I tend to get monthly rainfall depth and number of wet days from online sites, and then use an online climate generator I developed to generate a long period of record. From that record, I can determine return period rainfall events. This may not be the best method, but it works if more local data are not available. I can give you more details on my methods if you don’t get better suggestions.

Use models like HEC-RAS, SWAT, or MIKE FLOOD to simulate how rainfall translates into runoff and flooding.