Soil water repellency as a vegetation-driven strategy in arid ecosystems: Banksia woodlands (Western Australia) case study.

Abstract

Soil water repellency (SWR) is caused by hydrophobic organic substances released by plant residues, roots or soil microorganisms that inhibit or delay rainwater infiltration. Banksia woodlands (BW) are iconic ecosystems of semi-arid environments of Western Australia (WA) comprised by an overstorey dominated by Proteaceae, e.g. Banksia menziesii and Banksia attenuata, in combination with other species, such as Eucalyptus spp. Although located in poor dune soils, BW provide numerous ecosystem services and sustain a high biodiversity. A characteristic feature of BW is their dimorphic root architecture, formed by a proteoid (cluster) system that radiate from the parent root and spreads to form thick mats below the soil surface. These clusters secrete large amounts of organic acids and phenolics to increase the uptake of P and other minerals that can be related to SWR. In contrast, the parent root penetrates soil deeply, reaching the water table. This study, conducted in natural BW of WA, aimed to assess SWR and its impact on water economy in relation with soil functioning and plant strategies for water uptake in pristine BW. Soil samples were collected at different depths (0-1, 1-10, 20-30 and 40- 50 cm) based on the different SWR severities found in each layer under field conditions. SWR was assessed under laboratory conditions in oven-dry samples (48 h, 105 ºC) and the chemical organic assemblage of bulked soil subsamples from each layer was analysed by direct analytical pyrolysis (Py-GC/MS). SWR distributed discontinuously through the soil profile. The first thin layer (0-1 cm) composed of coarse sand and litter, located immediately above Banksia root clusters, showed wettable conditions. In contrast, the well aggregated soil layer where the cluster root system is located (1-10 cm) was severely water-repellent. The 20-30 and 40-50 cm deep layers rendered wettable or subcritically water-repellent. After Py-GC/MS analysis, major compounds were identified and grouped according to their probable biogenic origin. Among other soil organic compounds, well resolved bimodal alkane/akene (C8-C31, maxima at C13 and C26) and fatty acids series (short-chained, C5-C9, and long-chained even-numbered C12-C18) were associated to the root cluster soil layer (1-10 cm). Also, a relatively high contribution of fire-derived polycyclic aromatic hydrocarbons (PAHs) was observed (7%). These results point to possible indirect links between organic substances released by roots and soil wettability involving soil microorganisms. Further discussion should shed light on possible ecological plant strategies and specific adaptations for water uptake in such arid ecosystems of WA.