Scientific Visualization

fig, (ax_rain, ax_flow) = plt.subplots(
    2, 1,
    sharex=True,
    figsize=(10, 6),
    gridspec_kw={"height_ratios": [1, 2]},
    layout="constrained",
)

• 2, 1: two rows, one column; returns two axes you can name
• sharex=True: one time axis, so dates line up straight down
• height_ratios: the effect (river) gets twice the height of the cause (rain)
• layout="constrained": matplotlib spaces labels so panels do not overlap
• figsize is in inches; 10 × 6 fits a slide or a full-width page

fig, ax = plt.subplots(figsize=(3, 2))
ax.plot(stream["date"], stream["discharge_cfs"])
ax.plot(weather["date"],
        weather["precipitation_mm"])
plt.show()

Same 2019 data as the rest of this section. Work in pairs for two minutes.

w = weather[weather["date"].dt.year == 2019]
s = stream[stream["date"].dt.year == 2019]

fig, (ax_rain, ax_flow) = plt.subplots(
    2, 1, sharex=True, figsize=(10, 6),
    gridspec_kw={"height_ratios": [1, 2]})

ax_rain.bar(w["date"], w["precipitation_mm"],
            width=1)
ax_flow.plot(s["date"], s["discharge_cfs"])

Fixes problem 2. Rain is daily totals, so it is drawn as bars; flow is continuous, so it is a line.

ax_rain.invert_yaxis()
ax_rain.set_ylabel("Rain (mm/day)")
ax_flow.set_ylabel("Discharge (cfs)")
ax_flow.set_xlabel("Date, 2019")

ax_flow.xaxis.set_major_formatter(
    mdates.DateFormatter("%b"))
ax_flow.yaxis.set_major_formatter(
    StrMethodFormatter("{x:,.0f}"))

Fixes problems 1 and 4. Month names replace ISO dates; thousands separators make 12,000 readable at a glance.

flood = (pd.Timestamp("2019-05-17"),
         pd.Timestamp("2019-07-04"))
for ax in (ax_rain, ax_flow):
    ax.axvspan(*flood, color="0.92", zorder=0)

peak = s.loc[s["discharge_cfs"].idxmax()]
q = peak["discharge_cfs"]
ax_flow.annotate(f"Peak {q:,.0f} cfs, 29 May",
    xy=(peak["date"], q), xytext=(20, 0),
    textcoords="offset points",
    arrowprops={"arrowstyle": "-"})

Fixes problem 5. The band marks 17 May to 4 July, when flow rose and stayed above 9,000 cfs.

from earthcourse.hydro import flow_duration_curve

fdc = flow_duration_curve(stream)  # 4,018 days

fig, ax = plt.subplots(figsize=(8, 5))
ax.semilogy(fdc["exceedance_probability"] * 100,
            fdc["discharge_cfs"])
ax.set_xlabel("Percent of days flow is exceeded")
ax.set_ylabel("Discharge (cfs)")
ax.grid(True, which="both", alpha=0.25)

The 2019 peak, 12,600 cfs, sits above Q10 by a factor of two.

grid = temp.pivot(index="y_km", columns="x_km",
                  values="temperature_c")
anom = grid - grid.values.mean()
lim = np.abs(anom.values).max()

fig, ax = plt.subplots(figsize=(7, 5))
im = ax.pcolormesh(
    grid.columns, grid.index, anom,
    cmap="RdBu_r", vmin=-lim, vmax=lim)
fig.colorbar(im, ax=ax, label="Anomaly (°C)")
ax.set_aspect("equal")
ax.set_xlabel("x (km)")
ax.set_ylabel("y (km)")

from earthcourse.plotting import save_figure

saved_path = save_figure(
    fig,
    PROJECT_ROOT / "figures"
    / "grouplab01_rain_river_2019.png",
)
print(saved_path)

• Creates figures/ if it does not exist
• Saves at 150 dpi with bbox_inches="tight", so labels are never clipped
• The saved file is the deliverable; name it like you mean it: lab, content, year
• Reference the filename in your caption and README