Scientists have been looking at sediments formed by the Yellow River in China and it seems the river eroded and incised the Tibetan plateau over millions of years and in the process created the wind blown dust that is thought to form the Chinese loess formations – see http://phys.org/print363580304.html
I see to remember this story several months ago so perhaps it has just been published (Nature Communications, October 2015) which may indicate they had to alter the text as it was not acceptable to peer reviewers. That is speculation but as the research has reversed some mainstream opinions on the loess formation one if left wondering. A joint Swedish and Chinese research team explored modern and ancient sediments with a focus on zircon – a very hard silicate mineral that is resistant to weathering – which is useful as far as the loess is concerned as the texture is fine grained. In some parts of the UK loess is known as brick earth – and is also very fertile. For instance, in the Thames Valley brickearth was sought out by market gardeners in the 18th and 19th centuries. At the same time it was useful in the making of bricks.
The idea that loess is windblown with an origin in cold Ice Age northern lands, whipping across the tundra and a tree-less landscape, is part of text book geology. When it came to the huge loess formations in China (see above) it was thought these had formed during Ice Ages and the Yellow River carried loess downstream (hence the name of the river). The research established, instead, that the loess formation was in fact a sink. Material washed out of the Tibetan Plateau by the river (and presumably by rain and flooding) accumulated in the loess formation. The research demonstrates large scale sediment storage on land. However, it is still held that the sediment was worn away by the wind as well as the river, so the old theory has not been completely abandoned.
Allan and Delair, in their book or an article in SIS Review, claimed loess had an origin in water dispersal. To form the loess of China an awful lot of water would be necessary, in a region not renowned for its rain fall in the modern world. Were heavy rains and flooding a feature of the Tibetan Plateau in the past?
The loess was not laid down in one lump but in a succession of layers – so we are talking about a succession of climate changes. Each layer has been defined by fossils within it. Elsewhere in the world loess is very often found along river valleys (and flood plains) which supports the idea it is dispersed by water (rivers in flood). It is also found on former terraces of rivers. For example, the Thames has cut down in a succession of channels which are assigned to various ice ages (or glacial episodes), and loess can be found on terraces now far above the course of the modern river. This research tends to suggest loess is not entirely wind blown sediment and is primarily dispersed by water. Whether that will be acceptable to mainstream or not is something altogether different.