Practice note: Selecting flow data

Practice note: Selecting flow data

This practice note is one of a set developed to provide consistency and transparency of river system models being used within the Murray-Darling Basin. The notes cover modelling practices, such as naming conventions for folder structures, to model methods, such as for flow routing and residual inflow estimation, and have been developed through a collaboration between the MDBA, Basin States and CSIRO.

Produced in collaboration with:

This practice note, Selecting flow data, describes some general principles to be followed when assessing the availability and quality of observed flow data. It summaries flow data sources and the quality assurance checks that are used by different jurisdictions in the Basin and provides examples of also provides examples of implementing the principles.

Background

Flow data is essential for building any river model. The way in which data are used within the model depends on its quality and modelling need. Having an understanding of different data sources, the quality of the data, and how it is used in the model is essential for others to understand the underlying data quality.

General principles

• Streamflow data should be obtained from the most appropriate source
• Streamflow data should be extracted from 9am to 9am to align with the climate data used in the model. The data stamp should be at the end of the time period (i.e. the total to 9am)
• Appropriate quality assurance checks should be undertaken to aid in data classification
• Streamflow data should be classified based on the quality of the data and length of record
• Based on the data classification, how the data will be used in model development and calibration should be decided upon and documented.

An example of how general principles may be implemented

Obtaining streamflow data

• Most gauged flow data is held by the controlling water authority and stored in a HYDSTRA data base. Flow data should be extracted from HYDSTRA using the system most appropriate for the jurisdiction
• Daily data should be extracted to maintain a consistent time base between rainfall, evaporation and streamflow. Flow data should align with climate data which is typically measured from 9am to 9am.
• Where high flow data are missing from a site, investigation should be undertaken to see if missing data can be infilled using data maintained by the Bureau of Meteorology (BoM) for their flood alert stations.

Quality assurance and classification of gauging stations

• When assessing the quality of the flow data, it is important to:
  • understand the quality codes and the required time base for the data
  • distinguish which gauges cover the full range of flows vs gauges that may be fine until bank full is reached, after which their data become only vague estimates (because the floodplain is so wide or the channel cross section is generally ill defined.)
• Quality assurance checks should include
  • A visual check of the data – this should include looking at the data as a daily time series to look for obvious issues
  • A review of individual stations including:
    1. The site summary report including the stability of ratings, period of record, quality of record, periods of missing data, flow range covered by ratings
    2. Any additional information gained from hydrographers
    3. Quality of the rating curve – the site should be checked for, for example, looped ratings or unstable ratings indicated by large changes between successive rating curves
    4. Quantitative analysis of uncertainty may be warranted for important gauges
  • A comparison of catchment gauges. This should include:
    1. Upstream / downstream analysis to ensure reasonable gaining /losing patterns:
       • Visual inspections of hydrographs (and upstream vs downstream where available)
       • Monthly totals subtracted upstream from downstream to check for negatives or significant differences to patterns.
    2. Check for shifts in gauging, including Double Mass Curves to check the consistency of data collection between nearby stations over time.
    3. Check for outliers in terms of flow regime – for example, daily flow duration plots to compare general shape to surrounding catchments or with upstream and downstream gauges.

• Following quality assurance, gauges should be classified based on the quality of the available data and how they will be used in the model (see example in Table 1).

Table 1: Example classification of flow data and typical model use

References

Vaze J, P Jordan, R Beecham, A Frost, G Summerell, (eWater CRC, 2011) Guidelines for rainfall-runoff modelling: Towards best practice model application.

Links to relevant sections of the eWater wiki site

• Data preparation
• General requirements for data