There are two kinds of precipitation data on ERA5 datasets naming ‘total column rain water’ and ‘total precipitation’. I want to compare ERA5 precipitation data with observed data so which data sets will be better for this?
a main difference in ERA5 to ERA-Interim is that accumulated (or mean rate) fields are accumulated from the end of the previous time step or in other words: cover the last hour. That is, correct me if I'm wrong, you have to download all hourly fields of a day and sum them up afterwards by yourself (with e.g. CDO). I use this script for hourly total precipitation:
#!/usr/bin/env python
from ecmwfapi import ECMWFDataServer
server = ECMWFDataServer()
server.retrieve({
"class": "ea",
"dataset": "era5",
"expver": "1",
"stream": "oper",
"type": "fc",
"levtype": "sfc",
"param": "228.128",
"date": "2016-01-01/to/2016-01-02",
"time": "00:06:00/18:00:00",
"step": "0/1/2/3/4/5/6/7/8/9/10/11",
"grid": "0.25/0.25",
"format": "netcdf",
"target": "test.nc"
})
'time' contains the two forecasts at 6UTC and 18UTC, and 'step' contains 1 to 11, because every step contains the sum over the last hour (and not from the start of the forecast as in ERA-Interim). For example, 6UTC + step 2 means the total precipitation from 7UTC to 8 UTC.
With that you cover:
6 UTC + step 1 +...+ step 11 --> 6 UTC to17 UTC and from
18 UTC + step 1 +...+ step 11 --> 18UTC to 5 UTC (next day).
Step 0 is import because it covers 5UTC to 6UTC and 17UTC to 18UTC.
Once you have downloaded the files you can do a 'cdo daysum ifile ofile' to get daily sums.