ERA5 Precipitation Type vs Precipitation Rate


I'm trying to get hourly liquid precipitation and solid precipitation from the ERA5 hourly data on single levels dataset.  I'm doing this by downloading the "Mean total precipitation rate" and "Precipitation type" variables, and using the precipitation type variable to determine if the precipitation should be classified as liquid or solid.  The problem is that there are many cases where the precipitation type for a given timestamp and location is 0 when the precipitation rate for that same timestamp and location is non-zero.  The description for the precipitation type variable (here) states that a value of 0 means no precipitation, so it seems like there may be some inconsistencies between precipitation type and mean total precipitation rate.

Does anyone have a suggestion for how I should handle this case?  If not, are there different variables I should be using instead to get hourly liquid precipitation and solid precipitation?

Thank you!

- Jordan

Hello Jordan,

This discrepancy is arising because the "Mean total precipitation rate" is an average over one hour, see ERA5: data documentation#Meanratesandaccumulations, whereas "Precipitation type"  is an instantaneous parameter, see Parameters valid at the specified time. The corresponding instantaneous precipitation parameters are: "Convective rain rate", "Large scale rain rate", "Convective snowfall rate water equivalent" and "Large scale snowfall rate water equivalent" (see Table 2 in the ERA5: data documentation). I agree that the description for "Precipitation type" is misleading, so I will change that.

Unfortunately, this means that you can only diagnose the precipitation in this way with the hourly instantaneous data. You cannot do so with the mean parameters over each hour. In this latter case, you can only split by rain and snow (and convective and large-scale) using parameters: "Mean total precipitation rate", "Mean snowfall rate", "Mean large-scale precipitation rate", "Mean large-scale snowfall rate", "Mean convective precipitation rate" and "Mean convective snowfall rate" (see Table 4 in the ERA5: data documentation).

Note, the ERA5 "reanalysis" (HRES) data is produced with a model that uses a 12 minute time step, so the mean parameters over each hour use 5 time steps of information per hour, whereas the instantaneous parameters only make use of the information in 1 time step per hour.

Also note, that the mean rate parameters in Table 4 provide the same information as the accumulated parameters in Table 3 (of the ERA5: data documentation) , except the units are different and the latter parameters are expressed as an accumulation and the former as an average.

Let us know if you have further questions.



Hi Paul,

Thank you for the detailed response, this is very helpful.  I will let you know if I have any further questions.

- Jordan