CERRA single-levels: “invalid combination” for evaporation and total_column_water_vapour + requests stuck in “Accepted”

Datasets and Usage C3S - Datasets and Usage
,
1

Hello,

I am trying to download the evaporation, total_column_water_vapour and total_precipitation variables from the reanalysis-cerra-single-levels dataset to use as input for WAM2layers moisture tracking. I am consistently getting 400 Client Error: Bad Request — invalid combination of values for the first two, and CDS queue is extremely slow for the third (one forecast leadtime request taking ~6.5 hours since 2026-05-26).

The CERRA product user guide ( Copernicus European Regional ReAnalysis (CERRA): product user guide - Copernicus Knowledge Base - ECMWF Confluence Wiki ) lists these variables:

  • evaporation — forecast-only, accumulated, valid leadtimes 1–6 hourly.
  • total column integrated water vapour — analysis & forecast, instantaneous.
  • total precipitation — forecast-only, accumulated, valid leadtimes 1–6 hourly.

But I cannot find the exact API slug accepted by CDS for total column water vapour. I have tried (all return 400):

  • total_column_water_vapour
  • total_column_integrated_water_vapour
  • total_column_water
  • vertically_integrated_total_water_vapour

All requests with product_type: analysis, level_type: surface_or_atmosphere, and a single timestamp.

For evaporation with product_type: forecast, leadtime_hour: [3], level_type: surface_or_atmosphere I also get 400. With leadtime_hour: [1] the request was accepted but is still queued after 4+ hours.

Minimal reproducible example

import cdsapi
c = cdsapi.Client()

# Fails: 400 invalid combination
c.retrieve('reanalysis-cerra-single-levels', {
    'variable': ['total_column_water_vapour'],
    'level_type': 'surface_or_atmosphere',
    'data_type': 'reanalysis',
    'product_type': 'analysis',
    'year': '2014', 'month': '08', 'day': '20',
    'time': ['12:00'],
    'data_format': 'netcdf',
}, 'tcw_test.nc')

# Fails: 400 invalid combination
c.retrieve('reanalysis-cerra-single-levels', {
    'variable': ['evaporation'],
    'level_type': 'surface_or_atmosphere',
    'data_type': 'reanalysis',
    'product_type': 'forecast',
    'year': '2014', 'month': '08', 'day': '20',
    'time': ['09:00'],
    'leadtime_hour': ['3'],
    'data_format': 'netcdf',
}, 'evap_test.nc')

What does work

The same dataset with the following variables and parameters works fine in analysis:
2m_temperature, 2m_relative_humidity, surface_pressure, 10m_wind_speed, 10m_wind_direction
(also with level_type: surface_or_atmosphere).

What I’m asking

  1. What is the exact API slug for “total column integrated water vapour” in reanalysis-cerra-single-levels?
  2. What is the correct parameter combination for evaporation (accumulated, forecast)? Is leadtime_hour set correctly?
  3. Is there a known issue with CERRA queue priority since [date]? Requests appear to remain in “Accepted” indefinitely.

Environment:

Thank you!