Calculating ERA-Interim monthly total precipitation on single levels

Datasets and Usage Other - Datasets and Usage
, ,
1

Good day

I have a question about ERA-Interim's "Monthly Means of Daily Forecast Accumulations". I am looking for the total precipitation for each month, based on a 24 hour forecast. My question is: which "time step" do I need to download? Is it the 0-12 or is it the 12-24? I thought I would need both and then sum them for each month, but those totals seem too high.

I am aware I need to multiply by 1000 and then by the number of days per month to get it in mm/month, but would appreciate guidance on the correct data to download for a monthly total. I am new to working with ERA data.

Thank you in advance.

2

Hi,

please have a look at the last row of the table at the following link:

Conversion table for accumulated variables (total precipitation/fluxes) - Hydrological parameter table


Thanks

Michela

3

Hi Michela Giusti

I did look at that table and it was based off it that I decided to choose both the 0-12 and 12-24 steps and then sum them for each month. However, as mentioned, the values seem too high. So I am wondering if my interpretation to sum the two values for each month is incorrect? Should they be averaged or do I only need one of the time steps?

Many thanks

Jessica

4

Hi,

as shown in the table, the data from time step labelled 12 has to be selected from reference time 00UTC and added to the data from time step labelled 12 selected from reference time 12UTC ; the sum of these will then contain the accumulated total precipitation over the 24 hours for that day. By repeating this process for all days in a month and then adding all the values of each day of the month, you will get the monthly total in metres. To get the monthly total in mm, you just need to multiply this value by 1000.

Thanks

Michela

5

Hi Michela Giusti

I understand and fully agree with your explanation. However, that description is based on daily data. As indicated in my original question, I am downloading monthly data and am therefore unsure how to use the time steps available (0-12, 12-24 or 24-36) in order to ensure I get a monthly total that is based on a 24 hour forecast period. I have tried downloading the monthly estimates for 0-12 and 12-24 and then summing these, but the values are too high (i.e. each month had two estimates, one for 0-12 and one for 12-14, which I then added together for each month). Please can you explain which interval is best to use as a total for each month?

I am downloading the data here: https://apps.ecmwf.int/datasets/data/interim-full-moda/levtype=sfc/

Many thanks

Jessica

6

Hi  Jessica,

According to the documentation in the Era Interim paper (section 3.3, on page 17) you can see that:

Monthly means of daily forecast accumulations are produced for the accumulated surface and single level fields by averaging the twice daily forecasts (from 00 and 12 UTC) over the month, for the forecast ranges of 0-12 hours, 12-24 hours and 24-36 hours, and then scaling the results to have units “per day”.

So you can choose which period 0-12 or 12-24 to use, because both are scaled for the 24 hours period.

I hope this helps.

Best regards,
Milana

7

Hi Milana

Thank you very much, that definitely helps.

Best regards

Jessica

8

Hi Jessica,

The ERA-Interim, monthly means of daily forecast accumulations have 3 step ranges: 0-12, 12-24 and 24-36 hours. They are all produced using the twice daily forecasts from 00 and 12 UTC, but using the 3 different step ranges ie the first 12 hours, second 12 hours and third 12 hours of each forecast within the month. Each of these step ranges provides an estimate of the precipitation for the complete month, with units m/day. The reason these 3 choices are available is because of spinup/spindown, which can occur to precipitation as the forecast progresses i.e. the area average, precipitation can become more or less intense during the forecast. (Of course, precipitation will still vary locally, due to the passage of weather systems and convective storms etc.) With these 3 forecast step ranges, you can investigate the best compromise between the spinup/spindown and the length of the forecast (longer forecasts are prone to more error) for your parameter (precipitation in this case) and region of interest.

Best wishes,

Paul

9

Hi Paul

Thanks very much for providing this detailed explanation. I will investigate the best compromise between spinup/spindown over Africa.

Kind regards

Jessica