Stonelines are ubiquitous on the hills around Bananal (here above Lavapés). They are essentially thin (usually less than 20 cm) stringers of resistant clasts which follow the slope at depth of 0.5 to c. 2 m; often there is more than one stoneline at a site. There exist many theories as to how stonelines develop and what is their significance, although as Donald Johnson frequently pointed out during our trip, soil fauna perhaps plays a crucial role in the development of the superficial sediment layer, including the stonelines.
At Bom Retiro it is possible to demonstrate that the process of stripping and 'rampa' formation was intermittent. Two colluvial bodies, a few metres thick each, are separated by a distinct palaeosol horizon (hammer) dated for 9700 BP. The subsequent massive deposition of colluvium, recorded throughout the region, is called the 'Manso event' and taken as the evidence of significant environmental change in the early to middle Holocene.
The area around Campinho provides a splendid example of what the Brazilian colleagues call the 'drowned landscape'. Hilly compartments, usually weathered throughout, are being progressively 'drowned' in an alluvial fill of first- and second-order valley. Stripping of weathering mantle from the slopes is so fast that the capacity of rivers to transport this material further away is insufficient. Therefore the colluvial-alluvial fill accumulates, valleys become choked up with the sediments, and valley floors aggrade. At Campinho the thickness of Holocene alluvial infill, the top part of which is seen in the road cut, is about 12 m. However, at this site at present erosion seems to prevail over aggradation.
At Bela Vista we were shown a marvellous example of gullying and 'rampa' dissection. Here the gully is developing within the first-order valley filled by colluvial deposits washed down from the surrounding slopes. The gully is 5-10 m deep and has several branches that develop independently from each other. Headward erosion accomplished mainly by pipe collapse and earth falls is very fast; in the last 15 years the rear scarp has retreated by a few hundreds of meters. Note the horse rider for scale.
Deforested interfluves near Porto Alegre allowed us to see the hilly landsurface of Bananal in its wider geomorphic context. Mountain ridge of Serra da Bocaina forms the skyline. Within the hilly landscape numerous degraded landslide scars and depositional 'rampas' can be identified.
Best places to appreciate the depth of weathering in the piedmont of Serra do Mar are road cuts such as this one located west of Barra Mansa. Precambrian gneiss is weathered down to at least 20 m though the original rock structure is still preserved despite weathering. From this area depths of 'immature' sandy weathering in excess of 50 m have been reported and are indicators of very rapid advance of weathering front the rate of which is not matched by mineralogical evolution of the saprolites. This is perhaps the way the weathering system responds to uplift and rapid erosional differentiation of relief in the humid sub-tropical environment. Weathering is also important for assessing the potential of superficial processes. Deeply weathered slopes, especially if stripped of woodland vegetation, are prone to rapid gully erosion and many deep-seated landslides occur within the saprolite.
Reduction of strength of rock due to deeply penetrating weathering enhances susceptibility of slopes to landsliding. This photograph was taken on the road going from Bananal westwards and shows two landslides of different ages. The one on the right is a good example of slump and, judging from the very fresh appearance of the rear scarp, is very recent. To the left is the scar of an older landslide, partly vegetated but also subjected to shallow translational sliding and gullying. The landslide mass has now a smooth evened surface and the form approaches the depositional 'rampa' stage. Low earth mound in the foreground testifies to the important role played by soil organisms in forming the structure of the superficial layer.
A typical 'rampa' near Bananal. Two segments of different origin can be recognised. The upper one is a hollow within the convex slope, moulded by landsliding, gully erosion and sheet wash. The lower one is a gently concave surface sloping down towards the nearby valley floor and underlain by colluvial deposits. Absence of clear signs of recent slope instability suggests that this particular feature has already been fossilised.
