Preparing time-series data

Provide a CSV or Excel file with one date column followed by one or more numeric flow columns. Each flow column is evaluated independently as a scenario.

Arrange the input columns

For CSV, put the date in the first column and use time as its heading:

time,flow
2020-01-01,0
2020-01-02,2
2020-01-03,3

For several scenarios, add one numeric column for each:

time,baseline,alternative
2020-01-01,0,1
2020-01-02,2,0
2020-01-03,3,2

Each observation date should correspond to one row. The dates must be readable by the date parser; set date_format in the time-series configuration when automatic parsing is unsuitable. Excel input uses the first column for dates and the selected worksheet.

Place files and run the command

Relative data paths resolve from the terminal's current working folder, not from the folder containing the TOML configuration. Keep the data file in the working folder named by the relative path, or use an appropriate path in the configuration. The first evaluation demonstrates this explicitly.

Choose characteristics for the question

Use the scientific foundations and characteristic reference to choose conditions that match the analysis. Examples demonstrate syntax and software behavior only; their thresholds are not universal ecological criteria.

For seasonal comparisons, use the timing characteristic's calendar-day convention. For duration and frequency, identify whether a timestep means a day, month, or another sampling interval in your data; their configured lengths are counts of observations, not automatically calendar lengths.

Daily calendars and leap day

For daily records, hydropattern accepts an otherwise complete water year whether or not February 29 is present. A missing February 29 is treated as an optional omission, whether it was removed from a synthetic calendar or was a missing real observation; the software cannot tell these apart. Other missing days are not accepted. Do not use omitted leap days as a general gap-filling rule.

Daily completeness and exposure use a denominator of 366 when the water year's recorded dates contain February 29 and 365 when they do not. A partial year that ends before February 29 uses 365, even when a leap day would occur later. February 28 and 29 share a normalized day-of-water-year label, but remain separate recorded observations. Do not infer the exposure denominator from those labels.

Unsupported cadence, duplicate dates, other missing dates, or insufficient timestamp information make annual completeness undetermined. Observed-outcome counts and fractions are still available, but nested annual evaluation and time-based event rates cannot be calculated reliably. Use supported complete daily or monthly timestamps for those analyses; see unknown outcomes and water-year reporting.