Hello all

I know how to download hourly minimum and maximum air temperature from API, now my question is: is there any way to download monthly values of min and max air temperature from API (cdapi)?

Hello all

I know how to download hourly minimum and maximum air temperature from API, now my question is: is there any way to download monthly values of min and max air temperature from API (cdapi)?

Dear Behzad,

I presume you are concerned with the minimum and maximum air temperature (Minimum/Maximum 2m temperature since previous post-processing) in ERA5, is that correct? If so, we only have those parameters in the sub-daily form, which is hourly in the ERA5 reanalysis data. We don't have daily or monthly values, which you would need to calculate from the hourly values.

Regards,

Paul

Dear Paul,

Thank you for your reply.

Best Regards

Dear Behzad,

You could easily compute the monthly max and min with the Toolbox.

Here is an example of workflow that does what you want: https://cds.climate.copernicus.eu/toolbox-editor/168/forum_monthly_max_min

Let me know if you have any questions.

Regards.

Vivien

Dear Vivien,

In my study, I also need the monthly max and min temperature. Thanks to the Toolbox you provided above, I was able to retrieve these types of data (I used ERA5-Land data).

However, when I calculated the average values over some study regions, I found that the min/max temperature seems to be shifted about 3 months compared to the mean temperature. What do you think about this problem?

Thanks in advance!

Best regards,

Lan Anh

Dear Lan Anh,

Thanks for your question. Do you have an example of workflow or application where I could see the shift you are talking about? I don't see any reason why there should be a shift in time dure the max and min calculation. Did you use 'MS' as frequency when calculating the mean?

Here is a workflow where I:

- retrieve ERA5 Land data
- calculate min, max and mean
- extract the values for Hanoi
- plot the three time series.

https://cds.climate.copernicus.eu/toolbox-editor/168/forum_monthly_max_min_mean_plot

I hope this helps. Let me know if you have any question.

Vivien

Dear Vivien,

Thanks very much for your reply.

I tested your workflow applied for Hanoi. I agreed there is no problem (e.g., the shift) here.

In my case, I used the same workflow as you mentioned previously for 'reanalysis-era5-land-monthly-means' to calculate the max/min temperature.

https://cds.climate.copernicus.eu/toolbox-editor/32518/forum_monthly_max_min_la

However, for the mean temperature, I considered directly the 2m temperature from https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-land-monthly-means?tab=overview . Do you think the shift comes from this process?

In my procedure, I projected and averaged the data (Tmean,min,max) over the whole region instead of considering a point (lat,lon). So I checked again this procedure, but it seems to work very well for precipitation.

Thanks and best regards,

Lan Anh

Dear Lan Anh,

In this example I take the mean monthly values from ERA5-land instead of calculating them and I don't get any shift.

It is hard for me to tell what is happening if you do any processing outside the Toolbox.

Have you tried selecting a bounding box which is very close to the area you are interested in in the Toolbox and make a spatial_average?

Regards.

Vivien

Dear Vivien,

Yes. In this case, it is difficult because I also make some calculations outside the Toolbox. But it is just kind of strange when I apply **the same processing** for both Tmean and Tmin/max and then the shift appeared...

Thanks for your suggestion, I will test the spatial average over a bounding box with the Toolbox to see.

Thanks again!

Cheers,

Lan Anh

Hi Lan Anh,

I would also suggest printing the time coordinates of your data locally, before and after the processing.

Regards.

Vivien

Dear Vivien,

Thanks again for your suggestion.

I finally found the problem: I downloaded the Tmean a long time ago, so the data is only available until Sept 2019. Meanwhile, Tmax/min are available the whole year 2019. But I did not realize that and supposed they are the same time period.

Sorry for bothering you.

Best regards,

Lan Anh

Dear Lan Anh,

No problem at all, I am glad you solved your problem and hope you got some more interest in the Toolbox.

There is a recently released documentation about the Toolbox if you are interested in it: https://cds.climate.copernicus.eu/toolbox/doc/index.html

Regards

Vivien

Dear Vivien,

Thanks a lot. I will check it out.

Have a nice day.

Best regards,

Lan Anh

hello. How to download minimum and maximum air temperature MONTHLY (reanalysis era5 single levels) grid 0.25 * 0.25 ?

Dear Vivien Mavel,

Thank you for the workflow. It's very helpful.

However, I would like to ask you if that workflow compute the monthly mean daily maximum temperature and the monthly mean daily minimum temperature. I am not sure what exactly do you refer exactly by monthly max and min, it's just pick the max and min temperature in that month? If so, is it possible to compute the monthly mean daily maximum temperature?

Best regards,

Diego.

Dear Behzad,

Note that the ERA5 forecast model tends to be biased cold at low levels. This means that in the CDS the 2m temperature (which comes from the analyses) often lies outside the range of the Minimum/Maximum 2m temperature (which comes from the forecasts). So you need to decide which is best for your application, ie whether to use the 2m temperature or the Minimum/Maximum 2m temperature to calculate the monthly max and min.

In ERA5, the time step is 12 minutes, so for each hour the Minimum/Maximum 2m temperature is calculated from 5 time steps of the forecast model whereas the 2m temperature is the analysis at the specified time (see Parameters valid at the specified time and Model grid box and time step), so only uses information from 1 time step every hour.

Regards,

Paul

Dear Paul.

thank you very much for your comment. May I also ask a question related to this issue? We are aiming at using ERA5 for driving some hydrological and ecosystem models, which all require daily minimum, maximum and average temperature. According to your comment, it is not a good idea to mix variables from the analysis and the forecast as this might lead to average temperatures outside of the tmin-tmax range. As we also need precipitation, wind and humidity, does it make sense to focus on the forecast variables as they are "consistent"? Then, we would compute the daily average temperatures by first calculating the hourly average from tmin and tmax and then averaging over one day. The other option would be to use the t2 from the analysis to calculate daily minima and maxima. I assume that there is no "unique" and "best" solution but maybe you can give some suggestions from the viewpoint of an ERA5-expert about how to approach this issue.

Best regards and thank you very much in advance,

Christof

Dear Christof,

Whether you use the analyses or forecasts depends on what you are trying to do. Usually, if the parameter is available from the analyses, it would be best to use those. However, analyses can have problems, eg the discontinuities in 10m winds between the assimilation windows, see the ERA5 diurnal cycle for winds problem in:

ERA5: data documentation#Knownissues

Some parameters, such as the total precipitation, are only available from the forecasts.

Clearly, it wouldn't make sense to use the analysed 2m temperature as well as the minimum and maximum 2m temperature from the forecasts.

If you did want to calculate the average temperature from the forecasts, it might be better to use the hourly output (of 2m temperature) from the forecasts, rather than using tmin and tmax.

Best wishes,

Paul

Dear Paul,

thank you very much for your detailed and informative answer!!! This indeed makes the usage of these variables much more clear. For the future, I'll simply use the analysed 2m temperature.

Best regards and, again, thanks a lot,

Christof