We are attempting to download data at hourly cadence using the Python API for reanalysis-era5-single-levels (type=reanalysis) by specifying year, month, day, and hour alongside the requested variables. The format is netcdf, and because we request variables from different streams and steps, we expect to get back a .zip file. As part of our process, we read the metadata from all of the archived .nc files to ensure that all of the variables we requested are present.
For year=2025, month=3, day=1, and hour in (1, 2, 3, 4, 5, 6), we find that some of the requested variables are missing. In the past, we’d run into issues where the parameter short name changed, but now we’re looking at parameter ids instead. Sure enough, we request 262 variables but only get back 243.
For this combination of datetime arguments, the missing params are: 260121 (k_index) 260123 (total_totals_index), 228001 (convective_inhibition), 228003 (friction_velocity), 228015 (minimum_vertical_gradient_of_refractivity_inside_trapping_layer), 228016 (mean_vertical_gradient_of_refractivity_inside_trapping_layer), 228017 (duct_base_height), 228018 (trapping_layer_base_height), 228019 (trapping_layer_top_height), 260015 (precipitation_type), 228023 (cloud_base_height), 228024 (zero_degree_level), 228029 (instantaneous_10m_wind_gust), 228088 (total_column_supercooled_liquid_water), 228217 (instantaneous_large_scale_surface_precipitation_fraction), 228218 (convective_rain_rate), 228219 (large_scale_rain_rate), 228220 (convective_snowfall_rate_water_equivalent), and
228221 (large_scale_snowfall_rate_water_equivalent).
The hours immediately surrounding this range downloaded and validated fine, so we were wondering if this is a bug or if some variables are occasionally just not present for a given datetime.
Thanks for any insight.