Ovesen, N.B., Kronvang, B., Larsen, S.E. & Andersen, P.M. 2023. Betydning af skift i instrument-typer til vandføringsmåling ved hydrometristationer i NOVANA. Aarhus Universitet, DCE – Nationalt Center for Miljø og Energi, 63 s. - Teknisk rapport nr. 258 http://dce2.au.dk/pub/TR258.pdf
Measurement campaigns carried out at nine hydrometric stations are used to elucidate whether there are any systematic differences in water flow measured by different types of instruments and, thus, whether there is a risk of a homogeneity breach of runoff time series. The study includes propeller instrument OTT C31-Universalflügel (propeller-instrument), ADCP-Acoustic Doppler Current profiler from RD Instruments (ADCP-StreamPro) and electromagnetic velocimeter from OTT (OTT MFPro).
The statistical analyses show that there are clear and significant differences in the results of the measurement instruments. The differences between the measurements at the individual stations vary between -13.1 and +11.1%. The overall tests, where all measurement results are included, show that the ADCP-StreamPro instrument measures 2.6% more than the propeller instrument and that the OTT MFPro instrument measures 5.7% less than the propeller. The results also show that the differences between OTT-MFPro and propeller are caused by different registration of the velocity and not the cross-sectional area. The same partly applies to the relationship between ADCP-StreamPro and propeller, but not quite as clearly. At the same time, there is a tendency for the percentage differences between the instruments to be greatest for the smallest water flow measurements, and, correspondingly, for the relationship between OTT-MFPro and propeller, there is a tendency for the difference in the measured velocity to be highest at the lowest velocities.
For the ADCP-StreamPro measurements, the significance of the choice of extrapolation routines for the calculation of the unmeasured part of the water flow at the bottom and surface has been studied. Consistent use of the "power-power" routine, a regular routine up to 2019, leads to water flow values that are, on average, 1.9% higher than those using the current procedure with individual assessment of the most appropriate calculation routine. There is a tendency for the differences in measured water flow between the instruments to be lower in the largest streams where velocity is measured in a larger proportion of the profile.
In the international literature, with information on comparable measurements with ADCP and propeller instrument as reference, no significant differences between the water flow measured with the two instruments have been identified. The vast majority of the studies are from the USA – and from very large rivers. This study appears to be the first to conduct a systematic control of ADCP against propeller instruments in small streams.
Introduction of ADCP-StreamPro for the measurement of water flow in large streams and OTT MFPro in the smaller streams instead of the previously applied propeller instruments has resulted in a homogeneity breach of the water flow time series. The OTT-MFPro instrument undoubtedly measures a smaller water velocity than the ‘true’ one in the measurement profile, but the specific reason for this is not determined. Turbulence in the measurement profile is a possible explanation, which should be investigated further. It cannot be decided whether the ADCP-StreamPro or the propeller instrument measures ‘true’ or actually correct water flow.
The observed homogeneity breach in precipitation-runoff relations around 2010/11 at a number of NOVANA measurement stations may, to a certain extent, be linked to changes in measurement instruments for water flow determination.
It is recommended, if possible, to recalculate results from stations with ADCP-StreamPro measurements from before 2019 using new software and new computational routines. In addition, it is recommended to establish a new measurement project including many small-sized streams using both propeller (Klein-Flügel) and OTT-MFPro. The possibility of correcting measurements with OTT-MFPro should then be examined.
