Feature request - deriving precipitation events from precipitation time series

[rchlumsk]

A function to basically isolate rainfall events from precipitation data would be helpful, and has been requested for usage in groundwater event modelling in particular.

First cut of this could use some first order derivative criteria to detect a rate of change in precip to capture rise and fall in precip intensity (?). Would likely require hourly or higher resolution data

[KevinShook]

Is this to separate rainfall from total precipitation, i.e. to separate rain and snow? If so, some good work has been done on this by Philip Harder at the Centre for Hydrology, who has produced code which can estimate the precipitation phase from the air temperature and humidity, without the need for a site-specific calibrated parameter. It's in my CRHMr package.

[rchlumsk]

I would be interested to take a look at that as well. But no, this is to simply identify events in a time series. I imagine this code exists in multiple places but not sure if there is a good bit of R code to do that specifically for rainfall/total precip.

[PaulWhitfield]

We have code that does this for POT using a threshold. What the real issue is that there is no consistent idea of what an event consists of. I suspect we would need to accommodate different thoughts on this as there all types of nuances. Are we talking about independent events for extreme value work, or combining non-independent events?

[rchlumsk]

Good questions, this was a request from some time ago so not sure what the best approach would be. I image some kind of parameterized method that allows users to isolate events in their data. POT is certainly one approach, I'm sure we could include others.

There is not much more guidance on this question beyond what we wish to do with it, we can also close this is we are looking to clear issues since the POT functionality already exists.