Time reference from ERA5 reanalysis

Hi all, I only starting to use ERA5 data and I have a simple question regarding the time zone of the data. I am downloading hourly data on a single level of temperature. I am interested in information from Brazil (UTC-3).

To investigate hourly temperature along a single day, say 2012-09-31, how should I search in ERA5? Should I look for the exact date and time I want, and the application adjusts the time of each point in the grid for the time I asked, or it gives information for every point referenced to the UTC?

In other words, if I download global data for 07:00 of the day 2012-09-31, will I have: (i) a morning temperature for a grid point in Brazil and also for a grid point in Beijing, or (ii) a nighttime temperature for Brazil (07:00-3) and an afternoon temperature for Beijing (07:00+8)?

All times are UTC based. So a 12UTC is then 12-3 in Brazil. Local time conventions (daylight savings etc) are usually handled in analysis software packages.

Both python and R have tools for that.


After spend hours surfing the Internet (forums) looking for some Python code to convert UTC time to local time (e.g. Lima, Peru), I was wondering if someone knows how to do it.

I'm working with ERA 5 land hourly data downloaded in netCDF format.

Any help will be appreciated!


I am working with era5 hourly data in Brazil and couldn't understand one thing.

I downloaded precipitation and temperature data from 1979 to 2019; Precipitation data (also evaporation, potential evaporation, i.e. forecast variables) starts on "1979-01-01 07:00:00", temperature data (also soil water content, i.e., analysis variables with instantaneous values) starts on "1979-01-01 00:00:00".

Why the first starts on 7h? I would expect 3h, since this is the difference (Longitude = -40). Is the second local time?

All the best


for accumulated variables, the values for 0000-0600 come from the forecast run on the previous day, so these timesteps are missing for 1/1/1979.



That makes sense!

Thank you a lot Michela!