Wildfire effects on lipid composition and hydrophobicity in bulk soil and soil size fractions

Abstract

Low soil-water affinity and soil water repellency (SWR, hydrophobicity) prevents water from wetting or infiltrating soils in burnt and unburnt ecosystems, causing various changes on their hydrology, geomorphology, geochemistry, and biochemistry. Wildfire may destroy, develop or enhance SWR in previously wettable or water-repellent soils (e.g., Doerr et al., 2009; Jordán et al., 2013 and references therein). SWR is at least in part attributed to a lipid-like cover, rich in fatty acids (FAs). Recently, it was shown that FAs had a major role in increasing the water repellency of unburnt sandy soils in Doñana National Park (DNP, SW-Spain), with Mediterranean climate and developed under trees (Quercus suber, Pinus pinea) and shrubs (Pteridium aquilinum, Halimium halimifolium) dominated vegetation (Jiménez-Morillo et al., 2016). To get further insight into how fire affect the distribution of soil lipids and their role in the SWR, a study was performed on different size fractions of a DNP sandy soil under Quercus suber canopy cover. Two soil samples were taken, one in a burnt site and another in an adjacent unburnt (control) one, both having the same physiographic characteristics. SWR was determined using water-drop-penetration-time test in the <2 mm sieved (bulk) soils and in six size fractions: 1-2 mm, 0.5-1 mm, 0.25-0.5 mm, 0.1-0.25 mm, 0.05-0.1 mm and <0.05 mm. Lipids were extracted from all samples (n = 14), and the FAs and neutral lipids were identified and quantified by GC/MS and GC/FID. The carbon isotope ratios (¿13C values) for the individual fatty acids were determined by GC/C/IRMS.

The SWR values of soil samples and fractions were statistically different (p < 0.01), for both, the fire affected and unaffected soils, and different grain-size fractions. SWR values in burnt bulk soil and 0.05-0.1 mm fraction were higher than in unburnt homologues. The coarsest and finest soil fractions (1-2 mm and <0.05 mm, respectively) of the unburnt soil were the most hydrophobic; in contrast, the finer fractions (0.05-0.1 mm and <0.05 mm) were the most hydrophobic in burnt soils. The total amount of lipids and total FAs were higher in burnt bulk sample and all the size fractions, except the coarser one, which had twice the amount of lipids, compared to the burnt one. All samples showed similar distribution of saponifiable lipids, characterized by straight chain saturated acids in the C14-C32 range and only differing in their relative abundance. In bulk soil and <0.5 mm fractions the concentrations (g FA/g soil) of the FAs were higher in burnt compared to unburnt soil (this difference was small or absent in C22). For the coarser fractions, the opposite trend was observed in most FAs, except C18, and for C<20 acids in the 0.5-1 mm fraction. All the samples showed generally very similar distribution of non-saponifiable lipids, dominated by C27, C28 and C29 sterols and triterpenols, and a less abundant homologous series of n-alkan-1-ols in the C20 to C28 range and small amount of n-alkanes in the C23 to C28 range. Bulk soil and the size-fractions, except the coarser one, had higher concentrations in most neutral lipids. Principal component analysis (PCA) performed on lipid concentration, concentration ratios and SWR (Fig. 1A) indicated that hydrophobicity of soils were positively correlated to total amount of lipids, normal C>24 FAs and branched C>24 FAs, and negatively correlated with the even/odd FAs ratio. The scatterplot of the PC1 vs PC2 scores showed that all the burnt samples formed a cluster at values around 0, the unburnt coarse (>0.5 mm) and finest (<0.05 mm) fractions were positively correlated with SWR and total lipids, the bulk and intermediate fractions (0.5-0.05 mm) negatively (Fig. 1B). The biosynthetic origin of these lipids, and their transformation pathways during fire will be discussed with the results of the ongoing measurements of the 13CFA values.