Argo float profiles — quickstart¶

Pull a set of Argo float profiles (temperature, salinity, pressure) for a small box in the North Atlantic and plot a temperature–depth profile. Argo is open data — no credentials. This backend needs the argopy SDK: pip install earthlens[argo].

Setup¶

earthlens provides the unified EarthLens entry point; matplotlib draws the profile. Argo is a tabular backend — download() returns a long-format DataFrame (one row per measured level) and also writes it to disk.

In [ ]:

import matplotlib.pyplot as plt

from earthlens.earthlens import EarthLens

import matplotlib.pyplot as plt from earthlens.earthlens import EarthLens

1 — Fetch a region of profiles¶

Name the family parameters you're interested in (TEMP, PSAL); the bounding box, time window and depth range build the Argo region query. We use source="gdac" (the Global Data Assembly Centre) as the data backend. A region fetch returns the whole family (here PRES/TEMP/PSAL plus QC/error columns) — the names validate intent, they don't subset the columns.

Build the request¶

Describe what we want — the core physical parameters, a two-week January 2020 window, and a small box south of Nova Scotia. Building the request first keeps it easy to read and re-run.

In [ ]:

floats = EarthLens(
    data_source="argo",
    variables=["TEMP", "PSAL"],
    start="2020-01-01",
    end="2020-01-15",
    aoi=[-60.0, 40.0, -55.0, 45.0],
    path="_nb_out",
    source="gdac",
)

floats = EarthLens( data_source="argo", variables=["TEMP", "PSAL"], start="2020-01-01", end="2020-01-15", aoi=[-60.0, 40.0, -55.0, 45.0], path="_nb_out", source="gdac", )

Download into a DataFrame¶

download() realises the profiles into memory (and writes the backing CSV). head() previews the first few measured levels.

In [ ]:

df = floats.download(progress_bar=False)
df.head()

df = floats.download(progress_bar=False) df.head()

2 — Plot a temperature–depth profile¶

Each float reports temperature at many pressure levels. Plot TEMP against PRES (pressure in dbar ≈ depth in metres) with the y-axis inverted, so the surface is at the top — the classic Argo profile view.

In [ ]:

ax = df.plot.scatter(x="TEMP", y="PRES", s=4, figsize=(4, 6))
ax.invert_yaxis()
ax.set_xlabel("Temperature (°C)")
ax.set_ylabel("Pressure (dbar)")
ax.set_title("Argo temperature profiles — N. Atlantic, Jan 2020")
plt.tight_layout()

ax = df.plot.scatter(x="TEMP", y="PRES", s=4, figsize=(4, 6)) ax.invert_yaxis() ax.set_xlabel("Temperature (°C)") ax.set_ylabel("Pressure (dbar)") ax.set_title("Argo temperature profiles — N. Atlantic, Jan 2020") plt.tight_layout()

3 — A single float by WMO id¶

Instead of a region, a float: selector targets one float by its WMO id (the bbox is then ignored). A profile:<WMO>/<cycle> selector narrows to one cycle.

In [ ]:

one_float = EarthLens(
    data_source="argo",
    variables=["float:6902746"],
    start="2020-01-01",
    end="2020-12-31",
    path="_nb_out",
    source="gdac",
).download(progress_bar=False)

one_float[["PLATFORM_NUMBER", "CYCLE_NUMBER", "TIME", "PRES", "TEMP", "PSAL"]].head()

one_float = EarthLens( data_source="argo", variables=["float:6902746"], start="2020-01-01", end="2020-12-31", path="_nb_out", source="gdac", ).download(progress_bar=False) one_float[["PLATFORM_NUMBER", "CYCLE_NUMBER", "TIME", "PRES", "TEMP", "PSAL"]].head()