Influence of soil geomorphic factors on vegetation patterns in a model white sands ecosystem complex

Abstract

Understanding the development of white sands ecosystems (WSEs) and their occurrence is critical for determining structural diversity in global biodiversity hotspots. WSE has fascinated ecologists for decades, with soil always opined to play a critical role in the development of complex ecosystem vegetation patterns. However, few to no studies have systematically investigated the effect of soil natural capital on the complex vegetation patterns in humid tropical forests that include the occurrence of open savannas. Instead, fires have been advanced as the key driver that sustains and expands open savannas in tropical forest biomes. In this study, we investigated the effects of soil geomorphic factors on vegetation patterns that range from tropical forests to open savannas in a model WSE complex known as the Aripo savannas scientific reserve. Soils were sampled from three white sands landscapes, each containing three vegetation types (open savanna, palm island, and marsh forest). Two transects of five soil sampling sites were randomly sampled in each vegetation type where stratified sampling of microtopography (hollows and flats) was applied. Samples were analyzed mainly for chemical properties, with some physical and mineralogical properties. Results showed patterns and relationships amongst vegetation types. Parent material was rich in quartz and zircon, poor in alkalis and sesquioxides, and resistant to weathering, forming white sands. Open savannas were significantly higher (P < 0.05) in coarse silt (32.8 %) and fine sand (34.1 %) and lower in fine silt and clay, suggesting a clear sorting associated with aeolian cover sands. In microtopographic areas, the upward gradient of water potential drove matter accumulation in hollows and upward translocation of elements under flat relief. The plant available nutrients and carbon were significantly lower (P < 0.05) in open savannas, demonstrating that plant nutrition was resource dependent. The Al concentration and pH matched with the vegetation patterns with a significantly decreasing Al concentration from forests (953 ppm) to open savannas (223 ppm) and significantly increasing pH from forests (3.7) to open savanna (4.1). Interestingly, the significantly lower Al concentration and higher pH in the open savanna suggest that these properties though more favorable in this locale, did not support tree growth and therefore were not predominant determinants of vegetation patterns. Quartz dominated all textural fractions, with clay fractions (which increased under forest) having higher levels of non-quartz minerals, e.g., silica. Silica, which is more degradable and initially provided more cations (nutrients and Al) to forest vegetation. Therefore, both texture (clay size) and minerals (clay minerals) together, degraded and delivered more cations. We conclude that WSEs in tropical forests developed on originally poor parent material. Since natural fire regimes are unlikely in this ecosystem, differences in vegetation distribution were not attributable to fires but are related to soil variability due to weathering and pre- weathering sorting. Most likely matter transport by lateral sheet water flow due to soil geomorphology is another factor controlling the development of vegetation patterns within WSE mainly found in humid tropical areas where wet season rainfall exceeds 2000 mm.