Skip to content

Monthly-to-Daily Pipeline

Many applications need daily synthetic flows, but the most robust generation methods operate at a monthly scale. SynHydro's pipelines chain a monthly generator with a temporal disaggregator in a single interface.

KirschNowakPipeline combines the Kirsch monthly bootstrap with the Nowak KNN disaggregator. You provide daily observed data; it handles the internal monthly aggregation and disaggregation automatically.

Generate daily synthetic flows

import synhydro

Q_daily = synhydro.load_example_data()

pipeline = synhydro.KirschNowakPipeline()
pipeline.fit(Q_daily)
daily_ensemble = pipeline.generate(n_realizations=10, n_years=30, seed=42)

Q_syn_daily = daily_ensemble.data_by_realization[0]
print(Q_syn_daily.shape)  # (~10957 days x n_sites)

Single-site alternative

pipeline = synhydro.ThomasFieringNowakPipeline()
pipeline.fit(Q_daily.iloc[:, [0]])

Visualize

Plot a one-year window to inspect daily variability:

from synhydro.plotting import plot_timeseries

site = Q_daily.columns[0]

fig, ax = plot_timeseries(
    daily_ensemble,
    observed=Q_daily[site],
    start_date="2000-01-01",
    end_date="2000-12-31",
    show_members=3,
)

Next steps