Skip to content

Importing monitoring data

A model is a hypothesis until you compare it with a measurement.

Time-series measurements from permanent instruments or a temporary flow survey: level, flow, velocity, rainfall. Anything with a timestamp and a value.

CSV. The importer expects a timestamp column and one or more value columns, and it will show you what it found before anything is applied. Common export formats from telemetry systems work without reshaping.

Each series is attached to a model element - the manhole or pipe where the instrument sits - so the comparison is like for like.

Once imported, the observed series is drawn as a dashed line on that element’s results chart, over the modelled solid line. This one view answers most calibration questions before any automation gets involved:

  • Modelled peak much higher than observed - the catchment or its imperviousness is too big.
  • Modelled peak much lower - missing inflow, or infiltration you have not represented.
  • Right magnitude, wrong timing - the routing or the time of concentration is off.
  • Dry-weather baseline wrong - the dry-weather flow is wrong, and nothing about the storm response can be judged until that is fixed.

Comparing dry days and wet days separately is the single most useful thing you can do with a flow survey. Dry-weather flow is sanitary and infiltration only, so it calibrates population, per-capita flow and steady infiltration. What is left over in wet weather is rainfall-derived, and that is what your RDII parameters have to reproduce.

See I&I and RDII and Calibration.