One of the fundamental questions faced in ecological restoration is that of trying to find a “natural” baseline to work towards. This is especially true in Europe and eastern North America – these are the areas where so much fundamental ecological work was done, but by the time this work was done, they were modified to the point where it was often impossible to distinguish “natural” from anthropogenic.
I’ve always been a little confused by streams. As a terrestrial ecologist, steams often seem to be these alien things that intrude into the landscape. It isn’t that bodies of water are themselves incomprehensible, but rather, the way that they interact with the terrestrial landscape. Too often, they just don’t seem to add up – it always feels like current processes couldn’t produce what I see. Turns out that, at least in eastern North America, I’m at least partly correct to feel that way.
In a paper published in the January 18th issue of Science Robert Walter and Dorothy Merrits of Franklin and Marshall College examined the established ideas about the characteristic form of rivers in the Mid-Atlantic and Western parts of the United States. The norm is seen to be a single meandering channel with a gravel bottom and alternating riffles and pools, often with a channel that is deeply incised into the surrounding streambanks. These streams flow through “valley flats” – broad deposits that fill the valley. The stream cuts into these relatively coarse deposits. These deposits were interpreted to have been deposited when the river floods, while the valley was thought to have been produced by the stream as it meandered across the valley. (That’s what I was taught in high school geography). However, many observations seemed inconsistent with these explanations. Modern flood patterns don’t deposit the type or amount of sediments that are observed in these valleys. In addition, Walter and Merritts observed that the tops of historic milldams (which were typically built 2.5 to 3.7 m tall) were at the level of the valley flats which suggested that the valley flats had been deposited in recent times, after the dams were built.
Dam construction began with European settlement. Until they were replaced by steam engines in the twentieth century forges, furnaces, mining operations and mills were driven by water from millponds. Walter and Merritts estimated that there are 16,000 – 18,000 millponds in Pennsylvania, covering a period of over 200 years. Since millponds were built early in the settlement period, they rapidly silted up as the surrounding landscape was deforested. In the late 1800s and early 1900s newer dams were built along streams that had breached older dams. Many of these dams have also breached, leading to an even more complex pattern.
When the dug into the valley flats, Walter and Merritts found several metres of sediment on top of soils characteristic of wetlands. These soils are rich in organic material and sometimes include tree stumps and the remains of corduroy roads – roads made of logs or planks on swampy ground. By digging trenches across the entire valley, they were able to get a sense of the entire river, and were able to conclude that:
The characteristics of the presettlement sediments and organic material suggest that valley bottoms were broad, forested wetlands (alder shrub-scrub) with small, shallow (<1-m) anabranching and chain-of-pool streams that experienced frequent overbank flow, which is consistent with accounts by early explorers of ubiquitous swampy meadows and marshes fed by springs at the base of valley side-slopes
Walter and Merritts conclude by saying
These conclusions change the interpretation of hydraulic geometry in eastern U.S. streams that is based on the archetype of an “ideal meandering river form” and imply the need to reconsider current procedures for stream restoration that rely on reference reach conditions and the assumption that eroding channel banks are natural and replenishable. The current condition of single gravel-bedded channels with high, fine-grained banks and relatively dry valley-flat surfaces disconnected from groundwater is in stark contrast to the presettlement condition of swampy meadows (shrub-scrub) and shallow anabranching streams described here.
These observations also have important implications for European rivers, were millpond construction dates back to the 1100s. It also makes me rethink Trinidadian streams and rivers – but that needs to be a separate blog post.
Walter, R.C., Merritts, D.J. (2008). Natural Streams and the Legacy of Water-Powered Mills. Science, 319(5861), 299-304. DOI: 10.1126/science.1151716