RISC standards and resources

Without knowing the stability of the reference gauge, it is hard to determine the stability of the stage sensor and apply corrections if necessary. Stage data will likely be graded U-unknown. To obtain RISC grade A or B data, full level surveys of the benchmark network and reference gauge must be conducted bi-annually, and at least annually for RISC grade C.

No, regular level surveys to the entire benchmark network help confirm and assess benchmark stability, including the reference gauge. If a benchmark is unstable, consider installing additional benchmarks and/or increasing the frequency of level surveys.

Errors may be detected using a stable reference gauge to allow correction of the water level data to the local datum. Errors vary according to the type of stage sensor used. It is recommended that sensors be verified annually and, if necessary, calibrated according to the manufacturer’s specifications. Error corrections to stage data typically include:

• Offsets – for correcting unstable reference gauges to local datums
• Offsets – for adjusting unstable stage sensors
• Deletions and/or trims – for removing erroneous data
• Drifts – for correcting unstable stage sensors or accounting for sensor calibration drift

Rating curves may include multiple segments that can be described by the rating equation. In general, 6 discharge measurements are recommended per segment. Discharge measurements may be limited by available field resources or the timing of field visits. Discharge measurements should be selected to cover the full range of flows observed at the station and apply to stable control conditions.

High water discharge measurements are challenging due to elevated hazard and risk. Do not undertake any field activities without the appropriate safety training, experience, monitoring equipment, and PPE. 

Rating curves may be extended above the highest measured flow to a point where transitions in the rating curve control features are expected to occur such as changes in channel shape, or bank full elevation. If there are no expected transitions, the curve may be extended up to twice the highest measured discharge.

Curve extensions are often graded as Estimated to account for uncertainty associated with extrapolation above or below the highest measured discharge or changes in channel shape.

No. Data grades are generally intended to be applied to continuous datasets. Data grades of discrete measurements, although informative, do not represent all the sources of uncertainty and how data is graded and qualified. Professional judgement is recommended when reviewing salt dilution discharge measurements to ensure that best practices were followed and the discharge calculated is accurate within an acceptable estimate of uncertainty. Meta-data is vital when reviewing salt dilution discharge measurements.

The use of RISC BP – Best Practice grade is recommended for ADCP measurements, assuming best practices have been followed. Although the BC Hydrometric RISC Manual does not define such practices yet, resources may be consulted from the Water Survey of Canada (WSC), USGS, or World Meteorological Organization (WMO). Some key best practices include, valid compass calibration, moving bed test, appropriate track reference selected, reciprocally paired transects, sampling time >720 s, >50% sampled area, appropriate extrapolation methods used, and post-processing with software such as QRev.

BC Hydrometric RISC Grades are intended for continuous hydrometric datasets as a qualitative indicator of uncertainty. Alternate discharge measurement methods not covered by RISC can still be used to develop and verify rating curves and produce discharge datasets. Best practices should be followed and documented whenever possible. The World Meteorological Organization (WMO) and Water Survey of Canada (WSC) publish publicly available documents on such best practices.

First, review the outlier discharge measurement for errors, this may include assessing relevant meta-data. Second, review the mean gauge height used to calculate the rating error or shift for errors including the application of any gauge corrections to the reference gauge. Errors should be corrected, or if unable to be corrected, the measurement can be omitted for rating curve verification and discharge production.

Assuming no errors, consider the quantitative uncertainty (2 sigma, 95% confidence) of the measurement in relation to the rating error. Quantitative uncertainty that is larger than the rating error may indicate uncertainty due to challenging measurement conditions more so than changing or unstable rating curve control conditions.

If the rating error is assessed to represent a valid change to the rating curve, consider applying a stage or time-based shift correction to the rating curve or developing a new rating curve. Meta-data analysis and comprehensive field notes and pictures will make this assessment easier. Consider altering the hydrometric RISC grade of the derived discharge dataset to best represent quality.

Data should be reviewed and graded by a Qualified Hydrometric Data Reviewer as outlined in the Hydrometric RISC manual Chapter 6. Departures from the published hydrometric RISC standards should be documented and justified where applicable.

No, Grade A is not necessarily better than Grade C and does not indicate a lack of skill or attention to detail by the data collector or reviewer. Hydrometric RISC data grades are qualitative indicators of uncertainty in hydrometric data.  Station meta-data such as rating curve documents and station analyses are vital for end users understanding how data was graded and whether it is suitable for its intended use.

Overall discharge rating accuracy for a rating curve may be calculated as the Root Mean Square Error of the concurrent stage and discharge observations used to define the curve. RMSE is an indication of the average error between observed and rated discharge. RMSE may be used as an indicator of rating curve uncertainty and the derived discharge dataset. Other statistical measures of fit, such as mean average percent error, may be used to characterize rating accuracy. 

Hydrometric RISC data grades are applied according to the rating accuracy threshold criteria outlined in Table 1-1 of the BC Hydrometric RISC Manual.

Rating curves must be verified periodically due to seasonal or morphological changes in their control features that are common at many hydrometric gauging stations. Rating curves are often verified using streamflow measurements to calculate the rating error of the observed discharge compared to the rated discharge.  The rated discharge is the discharge predicted by the rating curve at the same mean gauge height as the measured discharge. Rating error for a given discrete streamflow measurement is calculated as the percent difference between the rated and measured discharge. 

If a rating error is larger than the target rating curve shift deviation threshold, the input stage data to the rating curve may require shifting for the output discharge data to match the observed discharge. Shift applications should be supported by sufficient evidence and not be used to account for random uncertainty associated with the measurement of stage, discharge, or the rating curve itself. 

In some instances, excessive shift applications may result in grading the output discharge data as RISC E – estimated. These may include instances where ice or unstable control conditions occur.

Rating errors larger than target rating curve shift deviation thresholds may also be accounted for my downgrading the output discharge data to the appropriate hydrometric RISC grade.

RISC data grades should generally be applied from one field visit to the next since data accuracy, and therefore the data grade, is verified during field visits. However, data grades may also be applied during periods that do not include a field visit. Some examples include:

• applying RISC E - estimated grade to ice affected periods of discharge data,
• applying RISC E – estimated grade to estimated data
• applying RISC E – estimated grade to periods of discharge data that are extrapolated beyond maximum or minimum thresholds
• Altering applied RISC grade to correspond with known stage transition points in rating curve segments, for example, floodplain activation
• Altering applied RISC grade during periods of known channel instability or control alteration due to the presence of beavers or human interventions

Hydrometric RISC grades should be documented and explained in station analysis summarising the computational period.