Upload 40 years of ERA5 forcing data to sampo

[juliasloan25]

Is your feature request related to a problem? Please describe.
We need multiple decades of forcing data available for longruns of the full land model. We intend to do these runs on clima, so we can upload the data to sampo. We want data in the years 1979-present.

We need these variables at hourly resolution & at 1 degree resolution. Please note that some of the variables we had before were accumulations (e.g. J/m^2 or kg/m^2) and the quantities below are all mean fluxes (accumulation divided by accumulation time period). So our era5 forcing function may change in the preprocessing function.

• Total downwelling SW radiation Mean surface downward short-wave radiation flux
• Direct downwelling SW radiation - not currently used, but we should use this instead of estimating the diffuse fraction internally: Mean surface direct short-wave radiation flux
• downwelling LW radiation at the surface Mean surface downward long-wave radiation flux
• Total precip at the surface: Mean total precipitation rate
• precipitation in snow at the surface: Mean convective snowfall rate or Mean large-scale snowfall rate? @braghiere or @szy21 could you advise? (Zhaoyi edit: We should use Mean snowfall rate, which should be the sum of convective and large-scale snowfall.)
• wind speed at the surface (or 10m): 10 metre U wind component AND 10 metre V wind component
• dew point temperature at the surface: 2 metre dewpoint temperature
• temperature at the surface: 2 metre temperature
• air pressure at the surface Surface pressure

Separate data file:
These can be monthly

• LAI (high vegetation and low vegetation): Leaf area index, low vegetation AND Leaf area index, high vegetation @braghiere please confirm. (Renato edit: We should use MODIS LAI instead, which is already available in sampo.)
  Kat comment: let's get the ERA5 data just for completeness, but plan to use MODIS when available

This function shows exactly what we need and how we use it:

ClimaLand.jl/src/shared_utilities/drivers.jl

Line 1020 in 295ea68

function prescribed_forcing_era5(

Part of #859

[ph-kev]

Pull request for only 2013 in ClimaArtifacts: CliMA/ClimaArtifacts#63

The current variables in the 2021 ERA5 dataset being used right now are these:

• Latitude ("lat")
• Longitude ("lon")
• Time ("time")
• Temperature at 2 meters (lon, lat, time; "t2m")
• Downwelling SW flux accumulation at surface (lon, lat, time; "ssrd")
• Downwelling LW flux accumulation at surface (lon, lat, time; "strd")
• Neutral wind at 10m, u-component (lon, lat, time; "u10n")
• Neutral wind at 10m, v-component (lon, lat, time; "v10n")
• Surface air pressure (lon, lat, time; "sp")
• Dew point temperature at 2 meters (lon, lat, time; "d2m")
• Total accumulated precipitation (lon, lat, time; "tp")
• Accumulated snowfall (lon, lat, time; "sf")
• Leaf area index, high vegetation (lon, lat, time; "lai_hv")
• Leaf area index, low vegetation (lon, lat, time; "lai_lv")

[braghiere]

That is right. Have a look also at /groups/esm/ClimaArtifacts/artifacts/era5_land_forcing_complete/download_era5_2021.py

I used that script to get the 2021 data. Can we already replace LAI (low and high) by the MODIS one in this file? In the ILAMB intercomparison, we've noticed ERA5 LAI is too high.

[ph-kev]

That is right. Have a look also at /groups/esm/ClimaArtifacts/artifacts/era5_land_forcing_complete/download_era5_2021.py

I used that script to get the 2021 data. Can we already replace LAI (low and high) by the MODIS one in this file? In the ILAMB intercomparison, we've noticed ERA5 LAI is too high.

Is the MODIS dataset the same as the MODIS dataset on ILAMB for Leaf Area Index? I checked that one and the dates only go up to 2005 and the dates are monthly instead of hourly.

[kmdeck]

Pull request for only 2013 in ClimaArtifacts: CliMA/ClimaArtifacts#63

The current variables in the 2021 ERA5 dataset being used right now are these:

* Downwelling SW flux accumulation at surface (lon, lat, time; "ssrd")

* Downwelling LW flux accumulation at surface (lon, lat, time; "strd")



* Total accumulated precipitation (lon, lat, time; "tp")

* Accumulated snowfall (lon, lat, time; "sf")

We can switch from accumulations to means (and alter the preprocessing functions used). But they should be equivalent, so whatever is easier

[kmdeck]

That is right. Have a look also at /groups/esm/ClimaArtifacts/artifacts/era5_land_forcing_complete/download_era5_2021.py

I used that script to get the 2021 data. Can we already replace LAI (low and high) by the MODIS one in this file? In the ILAMB intercomparison, we've noticed ERA5 LAI is too high.

Let's still grab the ERA5 LAI data. I updated the issue to note that this can be monthly resolution (and in a different file). We can use MODIS when possible, though

[Sbozzolo]

We have the ERA5 data in /net/sampo/data1/era5, is anything missing?

[kmdeck]

as far as ERA5 goes, yes, we can close this.

Ill open another issue on Monday to add MODIS data to ClimaArtifacts and to switch ClimaLand to that