Chaining calculations#

A derived quantity is an input like any other.

Sometimes you’ll have a pipeline that is two or three calculations deep. Because every calculation returns the same form you passed in, the result of one is a valid argument to the next: in Polars you add it as a column and refer to it by name, in Pandas you assign it and pass it straight back.

For example, estimate evapotranspiration from an eddy-covariance energy balance: first latent heat flux from net radiation, soil heat flux and sensible heat flux, then evapotranspiration from that.

Polars

df = df.with_columns(
    le=evapotranspiration.latent_heat_flux("rn", "g", "h"),
).with_columns(
    et=evapotranspiration.evapotranspiration_from_latent_heat_flux("le", "ta"),
)

shape: (2, 6)
┌───────┬─────┬──────┬──────┬──────┬──────────┐
│ rn    ┆ g   ┆ h    ┆ ta   ┆ le   ┆ et       │
│ ---   ┆ --- ┆ ---  ┆ ---  ┆ ---  ┆ ---      │
│ f64   ┆ f64 ┆ f64  ┆ f64  ┆ f64  ┆ f64      │
╞═══════╪═════╪══════╪══════╪══════╪══════════╡
│ 80.0  ┆ 5.0 ┆ 30.0 ┆ 15.0 ┆ 45.0 ┆ 0.065704 │
│ 120.0 ┆ 8.0 ┆ 40.0 ┆ 18.0 ┆ 72.0 ┆ 0.10543  │
└───────┴─────┴──────┴──────┴──────┴──────────┘

Pandas

df["le"] = evapotranspiration.latent_heat_flux(
    df["rn"], df["g"], df["h"]
)

df["et"] = evapotranspiration.evapotranspiration_from_latent_heat_flux(
    df["le"], df["ta"]
)

      rn    g     h    ta    le        et
0   80.0  5.0  30.0  15.0  45.0  0.065704
1  120.0  8.0  40.0  18.0  72.0  0.105430

Some calculations exist mainly to feed others. The FAO-56 intermediates in evapotranspiration - saturation vapour pressure, the psychrometric constant, the slope of the vapour pressure curve - are available individually if you want to inspect them, but potential_evapotranspiration_30min() computes them internally, so you do not have to chain them by hand.