Introduction
Landslide early warning systems in arid regions typically rely on probabilistic physical models that focus on internal soil properties (moisture, temperature, mechanical stress). However, external environmental processes may also significantly influence slope stability. This study investigates six field-observable factors: (1) surface clay crust from wind-blown dust; (2) shallow groundwater associated with alluvial aquifers beneath gully beds; (3) grazing-related surface disturbances; (4) manure-induced changes in soil geochemical properties; (5) cascade-type rockfall dynamics; and (6) wind-assisted erosion of dry gully sidewalls through particle detachment.
Methods
Field observations were conducted at two Algerian sites: Meziraa (Biskra) and the Babar Mountains. Methods included visual monitoring of gully sidewall failures, manual hand-excavation to compare surface versus subsurface moisture, characterization of surface crust, mapping of grazing intensity and manure distribution, observation of rockfall initiation, and assessment of wind effects on dry sidewalls. Comparative qualitative analysis identified recurring instability patterns across both sites.
Results
In Meziraa, slope failures initiated predominantly from sidewalls rather than channel beds. Although surface layers appeared dry and cracked, hand-excavation revealed moist near-surface conditions, indicating shallow groundwater beneath the gully bed. Areas with intensive grazing and manure accumulation showed altered soil geochemical properties and reduced apparent cohesion. Under strong wind conditions, dry sidewalls exhibited increased particle detachment and minor collapses. In the Babar Mountains, small rock movements frequently triggered cascading downslope displacement of larger rocks, indicating metastable slope conditions governed by chain-reaction-like granular behavior.
Conclusions
Surface clay crust, shallow groundwater beneath gully beds, grazing disturbances, manure-induced geochemical changes, cascade-type rockfalls, and wind-assisted sidewall erosion collectively contribute to landslide susceptibility in arid environments. Integrating these six field-observable indicators into existing probabilistic early warning models could improve landslide assessment in data-scarce regions.