Interpreting first results

This guide explains the exact, fully known results of the first evaluation. Other analyses can contain unknown outcomes; see the unknown-outcomes guide for their evaluation and reporting behavior.

Read the timestep CSV

Open first-run_output/flow_sustained_flow.csv in a text editor or spreadsheet. It contains these columns:

Column	Meaning
time	The observation date.
flow	The original scenario's observation, in the input units.
dowy	Day of water year; this example starts the water year on January 1.
magnitude_gt1	The magnitude characteristic outcome: 1 where flow exceeds 1, otherwise 0.
duration_ge2	The duration characteristic outcome: 1 throughout a qualifying run of at least two timesteps, otherwise 0.
sustained_flow	The final component outcome. Here it agrees with duration.

In this example, 1 means configured success and 0 means configured failure. These are software classifications, not proof of ecological benefit or harm. All eight outcomes are known.

Date	Flow	Magnitude outcome	Duration outcome	Component outcome
2020-01-01	0	0	0	0
2020-01-02	2	1	1	1
2020-01-03	3	1	1	1
2020-01-04	0	0	0	0
2020-01-05	4	1	0	0
2020-01-06	0	0	0	0
2020-01-07	0	0	0	0
2020-01-08	5	1	0	0

January 2 and 3 form a qualifying run of two timesteps, so both are marked successful by duration. The isolated qualifying timesteps on January 5 and 8 fail the duration condition. This whole-run assessment is retrospective: January 2's duration outcome uses the later observation on January 3.

Read the summary workbook

Open sustained_flow_summary.xlsx. Each sheet has a flow column for the scenario and rows labelled total and 2020. The total row summarizes the whole observed record. The 2020 row summarizes the observations in that water year; it does not mean a full year was observed.

Sheet	total and 2020 portions in this example
magnitude_gt1	0.50
duration_ge2	0.25
sustained_flow	0.25

The component has two successful timesteps among eight fully known timesteps, giving 2/8 = 0.25, or 25%. Magnitude qualifies on four timesteps, but two of those fail the later duration assessment. Read characteristic summaries separately from the final component summary.

This is a deliberately fully known example. Do not extrapolate its denominator to intervals with unknown outcomes: summaries use known outcomes only, and all-unknown intervals have no defined portion. For broader configuration rules, continue to the reference.