If you’ve ever looked at the rocks in the channel at different sites in the Buffalo River, you may have noticed that the size of rocks in the channel at upstream sites like at Boxley and Steel Creek are larger than at sites farther downstream like South Maumee and Buffalo Point. In most rivers, the average size of clasts (loose rocks) decreases downstream. Geomorphologists call this downstream fining. It happens for 2 main reasons. Smaller grains get winnowed or washed away by fairly frequent flows that have enough energy to carry small grains but not larger ones. Clasts also break apart during larger flows because of smashing into bedrock channel beds and against each other. This also causes rocks to become rounded. It’s like the river is a big rock tumbler. In some really steep rivers, with the right set of instruments, you can even hear the rocks smashing together. Some scientists are now using seismic acoustic technology to measure discharge of large floods from the impacts of rocks in the river rock tumbler.
When the flow is low and it is safe to enter the river, we measure average surface grain size of bars or of the channel using the ‘Wolman Pebble Count’ method, published by ‘Reds’ Wolman, one of the forefathers of fluvial geomorphology, in 1954. His summary of the method was to ‘walk like a drunk and point with a stick’. Taking a wandering walk, the person making the measurements stops randomly and points a stick at the gravel surface. Pointing with a stick reduces the bias we humans inherently have to pick up rocks that are easy to pick up and easily fit in the hand. We measure across the ‘b-axis’. If your cell phone was a clast, the b-axis would be the width across the face if held upright. The a-axis would be the longest width from earpiece to the speaker at the bottom; the c-axis would be the thickness.
We make at least 100 measurements to get a good statistical population, but of course, larger bars require more measurements. Some studies indicate that at least 400 measurements are needed to get accurate results. Instead of using the average, scientists typically report the median grain size. When 100 measurements are ordered from smallest to largest, the measurement of the 50th grain is the median or D50. The term 'D' is typically used to refer to the diameter of a particular clast size in sediment transport equations. Other measurements are also used in some assessments of grain size, for example, when considering the larger portion of the grain size distribution, the D84, or size of the 84th grain, is often used.
In the Buffalo River, the median grain size decreases from 35 mm at Boxley to 10 mm at South Maumee. Of course, the downstream trend isn’t exactly linear. Inputs of sediment from tributaries can increase the grain size locally, particularly if the tributary is steep. Rivers in watersheds that have lots of tributaries (what we call a high level of dissection) have lots of inputs of sediment along their length and may not have downstream fining at all. It may be easiest to conceptualize downstream fining on a continental scale. Rivers in steep high mountain streams often have huge boulders the size of cars while at the coast, rivers supply sand to beaches. So, when you’re sitting on the beach on the Gulf of Mexico, you can think about downstream fining and imagine that you are enjoying sand that may have started its journey within the large sandstone blocks in the Buffalo River watershed.