Comparison of flume and towing methods for verifying the calibration of a suspended-sediment sampler