In the majority of Danish streams, weed growth affects the effective stream width and bed roughness thus imposes temporal variations on the stage-discharge relationship. Small stream-gradients and firm ecology based restrictions prevent that hydraulic structures are made at the discharge stations and thus remove or limit such influences. Hence, estimation of the hydrograph is based on continuous stream gauging combined with monthly control measurements of discharge and assuming linear variation of bed roughness between the monthly measurements. As a result, any non-linear drift in weed density or structure which affect the frictional characteristics of the stream during both normal and peak flows are ignored. The present investigation studies if such temporal variation in the conveyance may be detected and eventually compensated for when estimating the hydrograph. Therefore acoustic Dopplers have been placed at the main discharge station in one of the largest Danish catchments (the Skjern). The instruments were set out in early February 2010 during the winter season and have been running since then. The long term average discharge at the station is near 14 m3/s and the cross sectional profile is roughly trapezoidal having width about 15 m, but slightly skew so that the stream is about 0.5 m deeper off the right than off the left bank. During winter, the depths are typically near 2 m while during summer they are about 1.5 m. During peak flows, when the discharge exceeds 35 m3/s, the depth increases to more than 3 m. The Doppler instruments (Nortek) are placed on a vertical pole about 2 m off the right bank at three fixed elevations above the streambed (0.3, 0.6, and 1.3 m); the beams point horizontally towards the left bank perpendicularly to the average flow direction. At each depth, the Doppler sensor records 10 minute average stream velocities in the central 10 m section of the stream. During summer periods with low flow, stream velocity has only been recorded at two depths since the water table drops below the uppermost sensor. A pressure transducer is also placed at the pole where it records the stage. The monthly control measurements made by the water authority office are made with an ADCP-instrument (Streampro, acoustic Doppler current profiler) which is pulled across the stream next to the in situ Doppler instruments. As part of the present project supplementary ADCP-data have been collected for different stage-values. For each stage many-pull traverses (>20) have been combined in order to produce reliable high-density stream velocity profiles across the stream. Currently, the different sets of velocity data are being analyzed in order to investigate how well the discharge may be predicted from the in situ Dopplers. The poster will present the field-set up and the results of the data analysis.