On the origin of rock fragment mulches on Vertisols: A case study from the Ethiopian highlands

Abstract

Many Vertisols in Tigray, Ethiopia, typically carry a discontinuous rock fragment (RF, size 0.5–> 40 · 10 − 2 m) cover with 10 to 100 RFs m − 2 . Such RF mulches are of agricultural and environmental significance because they influence the water balance in the underlying soils and the crop yield. Natural RF concentrations are mostly considered as eolian or hydraulic lag deposits, or as the result of lateral transport over the soil surface from a rock outcrop, upslope. In cultivated areas RF mulches can develop by tillage. This paper presents the case of a natural RF mulch whose lithology indicates that the RFs are up-squeezed by the local Vertisol. The study site is located in the pass of Enda Maryam, Tigray, Northern Ethiopia (39°8′ E and 13°36′ N). A circular area of 10 m diameter, about 200 m away from the water divide in the valley has been cleared annually between 01/1999 and 05/2003. During this period, 625 RFs, 17 being > 7.5 · 10 − 2 m in size, totalling a mass of nearly 62 kg, have been collected. After correction for measurement procedures, the rate of RF up-warping by the Vertisol at Enda Maryam is assessed at 5 RFs m − 2 in 3 years. At this rate of appearance, the formation of current RF concentrations on top of active valley Vertisols is only a matter of 10 1–2 years, provided the availability of RFs below the soil surface. Although important underground displacements were measured in the Vertisol between 01/1999 and 05/2002, the supposed link between up-squeezing of RFs and plastic deformations of ‘chimney’, ‘diapir’ or ‘intrusion’-like type in the Vertisol could not be evidenced. Instead, RFs are clearly concentrated on the soil surface as well as in depth, along the existing vertical desiccation cracks, often > 1 m deep which display polygonal configurations at the soil surface. Further, bundles of slickensides containing some RFs, have been mapped at the base of the Vertisol. The slickenside configuration suggests that the RF-bearing substrate is being scraped off. While the underground displacement of RFs along active slickensides seems normal, the process of RFs ascending in ‘upright’ position in the edge of desiccation cracks needs explanation. The closure of a desiccation crack is a peristaltic-like movement, following ascent or descent of the capillary fringe. It is hypothesized that this movement gradually pushes the RF to the surface or to another place or level in the soil profile where the crack closes in last instance. The apparent young age of the valley Vertisol mulches in Ethiopia might indicate the very recent formation of yearly recurrent desiccation cracks of Vertisols in the area. Available information confirms that most valleys in the study area used to be perennially marshy. Under these conditions no movements of RFs in the soil profile are expected to occur. Gullying, leading to pronounced seasonal desiccation of the Vertisols, started in several cases not more than 50 years ago.