Land use influences soil fungal community composition across central Victoria, south-eastern Australia

Current theory expects that fungi, on the one hand, are spatially ubiquitous but, on the other, are more susceptible than bacteria to disturbance such as land use change due to dispersal limitations. This study examined the relative importance of location and land use effects in determining soil fungal community composition in south-eastern Australia. We use terminal restriction fragment length polymorphism (T-RFLP; primer pair ITS1-F-ITS4) and multivariate statistical methods (NMDS ordinations, ANOSIM tests) to compare relative similarities of soil fungal communities from nine sites encompassing three locations (ca 50-200 km apart) and four land uses (native eucalypt forest, Pinus radiata plantation, Eucalyptus globulus plantation, and unimproved pasture). Location effects were generally weak (e.g. ANOSIM test statistic R <= 0.49) and were, in part, attributed to minor differences in soil texture. By contrast, we found clear and consistent evidence of land use effects on soil fungal community composition (R <= 0.95). That is, soils from sites of the same land use grouped together in NMDS ordinations of fungal composition despite geographic separations of up to ca 175 km (native eucalypt forests) and 215 km (P. radiata plantations). In addition, different land uses from the same location were clearly separate in NMDS ordinations, despite, in one case, being just 180 m apart and having similar land use histories (i.e. P. radiata versus E. globulus plantation both established on pasture in the previous decade). Given negligible management of all sites beyond the early establishment phase, we attribute much of the land use effects to changes in dominant plant species based on consistent evidence elsewhere of strong specificity in pine and eucalypt mycorrhizal associations. In addition, weak to moderate correlations between soil fungal community composition and soil chemical variables (e.g. Spearman rank correlation coefficients for individual variables of 0.08-0.32), indicated a minor contributing role of vegetation-mediated changes in litter and soil chemistry. Our data provide evidence of considerable plasticity in soil fungal community composition over time spans as short as 6-11 years. This suggests that - at least within geographic zones characterised by more-or-less contiguous forest cover - soil fungal community composition depends most on availability of suitable habitat because dispersal propagules are readily available for colonisation after land use change.