Estructura de bandas

The many apparent contradictions in the literature can probably be accounted for in part by clarifying the definition of the entities described, specifying the precise methods employed, adopting a standardised protocol in terms of the measurements taken. Bresson and Valentin (1994), in their review of 50 studies of crusting referred to earlier, were unable to make a statistical comparison because of the lack of common categories of data provided by the authors (Bresson 1995). They had hoped to systematise our knowledge o f crusting mechanims by comparing these studies on the basis o f the soil and environmental conditions of formation; namely, the soil texture, initial structure, initial microrelief, antecedent moisture, intensity and

kinetic energy of the rainfall, and state of development of the crust (Bresson 1995). The last point is particularly interesting, as it is not apparent from many of the papers reviewed for this present study whether the authors conceptualise crusts as temporary, dynamic states; during formation, yes, but the implicit assumption seems to be made that once formed (taken to be by the end of the experiment), they are in some final state.

Thus, in addition to the issue of common experimental protocol, some of the apparent

contradictions in the literature could potentially be resolved by recognising that any crust is the net product o f various forces operating at a particular point at the soil - atmosphere interface, and which vary in time, and which are affected in turn by conditions in neighbouring areas of time and space. Thus a particular crust observed by researcher A under conditions B may in fact be the 'same' crust as that observed by researcher X under conditions Y. Hence, any crust must be described with respect to its position in time and space, for example its location along a micro or macro catena; its management history, if any; and its place in any cycle of crusting observed from longitudinal studies and /or from general models of crust development.

A generic within-event multiple path temporal sequence is proposed by West et al (1992) for a cultivated soil. In the study area the crusted soils examined are used primarily for grazing, however the widespread trampling (apparent in the aerial photography) breaks up the crust and hence the initial stage in West et aVs (1992) system of a cloddy aggregated surface with large variability in surface morphology is often a valid initial condition.

Stage 0

Freshly tilled soil before rainfall

Stage 1

Aggregate breakdown and particle rearrangement due to slaking (wetting effect) and raindrop impact (energy effect), resulting in a 'disruptional' layer, which is the same as structural crust. Soil and rainfall characteristics determine the thickness, porosity and continuity o f this layer

Further crust development is now controlled by soil dispersibility (chemistry effect); there are two possible pathways, effectively controlled by the chemistry of the soil-water system. In soils with low dispersibilty, the dismptional layer continues to develop and aggregate coalescence may occur below the disruption zone as part of this process. Raindrop impact may cause particle disjunction, with the released fines removed in runoff. The lateral water movement results in washed-out layer, but no washed-in layer, as these fines are removed from the system.

In dispersible soils, on the other hand, extensive particle disjunction occurs early in the rain event, resulting m the clogging of what porosity remains in the disruptional layer. Thus a washed-in layer forms, which may be accompanied by a washed-out layer above it.

Stage 3

Because of the low permeability of the disruptional and/or washed-in layer, runoff is maximum at this stage. This is the stage o f particular interest for water harvesting studies; the sustained runoff yield plateau. The washed-out layer, if present, will be eroded by this runoff water, exposing the underlying layer to further raindrop impact and hence crust formation, by way of crust

coalescence. In some cases a thin, secondary washed-out layer will develop, with loss of fines. These crusts are very flat, with a depositional crust covering most of the surface, together with a thin, oriented clay seal after the end of the rain event if there is sufficient clay in suspension.

Farres (1978) introduced the term 'equihbrium crust', which has not been taken up by other authors, and is not probably valid in its original context (crusts 'filling up' from the bottom of the crusting zone until the protection it offers prevents further aggregate breakdown and hence the release o f material for further infilling; but see seal erosion issues below), and yet it could be used to describe a crust as a function of various forces, as proposed above, as long as the term

equihbrium is acknowledged to be a temporary equilibrium. If this perspective on crusting is valid, can we then move on to a systématisation of our knowledge of crust types and mechanisms of their formation by explicitly classifying them with respect to their position in time and space?

(3.13)