Water repellent soils: From molecule to ecosystem

Abstract

Soil water repellency (SWR) interrupts water infiltration, diminish plant growth and potentially promote soil erosion. Hydrophobic organic compounds coating soil mineral surfaces that originate from plants (leaves/roots) and microorganisms can induce SWR. These SWR-causing compounds are defined as ‘SWR-markers’. We aim to investigate the relations between SWR-markers and SWR to predict the persistence of SWR, linking SWR-markers to their plant origins to interpret the influence of vegetation cover on the SWR and SWR-markers of the underlying soil and, conversely, to explore the effect of SWR on vegetation dynamics in ecosystems. Dutch coastal dune sandy soils are collected under various plant species at different depths. The degree of SWR of the soils increases with increasing content of total organic carbon. A sequential extraction method is applied to these soils to obtain individual fractions and are analysed by gas chromatograph-mass spectrometry: free lipids (the so-called D fraction) are removed by dichloromethane/methanol (DCM/MeOH); the residual soils are extracted by iso-propanol/ammonia solution (IPA/NH3) and the extracts separated into a DCM/MeOH soluble part (AS fraction) and an insoluble part (AI fraction). The predominant groups in the D fraction originate from plant leaf waxes, while the AI fraction contains mainly root-derived suberins. The majority of compounds in the AS fraction are derived from suberins hydrolysed by microbes and a minor part is from leaf waxes. Therefore, plants are the primary sources of these SWR-markers, while microorganisms contribute a minor part. The soils become more hydrophobic after the D fraction is removed. The SWR of the residual soils dramatically decreases and the majority of them render wettable after the AI and AS fractions are extracted. Roots induce stronger SWR than leaf waxes. Suberin-derived ω-hydroxy fatty acids and α,ω-dicarboxylic acids both with C22 and C24 can well predict SWR. Due to a less mixed origin, more single long-chain (>20) SWR-markers are correlated significantly to SWR in a simple ecosystem with only a few plant species than in a complex ecosystem with various vegetation species. Plant leaves and roots mainly contribute SWR-markers to the top- and subsoils, respectively and consequently influence the SWR of corresponding soil layers. However, roots can also affect the SWR of the topsoils resulting from their distribution in the soil upper layer. Consequently, depending on plant species and plant tissues, vegetation cover has different influences on the abundance and compositions of SWR-markers and, therefore, on the distribution of SWR. After upscaling SWR from the molecular to the ecosystem scale, theoretical modelling suggest that SWR may affect ecosystems by triggering cyclic vegetation dynamics. High water competitive species (e.g. woody plants) with a low SWR-marker content show stable temporal dynamics. By contrast, low water competitive species (e.g. grasses) with a high SWR-marker content in tissues are more likely to experience cyclic dynamics.