Variation in soil microbial communities across a boreal spruce forest landscape Lisbet Holm Bach, A ˚ sa Frostega ˚ rd, and Mikael Ohlson Abstract: We investigated soil microbial community structure by phospholipid fatty acid (PLFA) analysis in a mature boreal spruce forest landscape in southern Norway, with low diversity of vascular plants. We investigated the spatial variation in PLFAs and the importance of environmental variables in 10 plots (each 13 samples) in a study area of 1 km  1 km. The scales investigated were 15 cm to 10 m within study plots and 100 m to 1 km between study plots. Soil microbial biomass var- ied 10-fold and we found a large variation in microbial community structure, even at distances of 15 cm. Samples aggregated into plots when PLFAs were subjected to a principal components analysis. Plot identity explained 36.3% of the variation in the PLFAs and geostatistical analysis showed that the microbial community structure displayed spatial dependence at within- plot distances. Environmental variables differed significantly between all plots but explained only minor parts of the variation in the overall PLFA pattern. The vegetation variables were, however, the best at explaining the PLFA pattern, and up to 60% of within-plot variation in individual plots, respectively, could be explained by vegetation variables, pH, and soil depth. Re ´sume ´: Nous avons e ´tudie ´ la structure de la communaute ´ microbienne au moyen de l’analyse des acides gras phospholi- pidiques (AGPL) dans un paysage de pessie `re bore ´ale mature, situe ´ dans le sud de la Norve `ge, ou ` la diversite ´ des plantes vasculaires est faible. Nous avons e ´tudie ´ la variation spatiale des AGPL et l’importance des variables environnementales dans 10 parcelles (13 e ´chantillons chacune) a ` l’inte ´rieur d’une zone d’e ´tude de 1 km  1 km. L’e ´tude a e ´te ´ re ´alise ´e a ` des e ´chelles de 15 cm a ` 10 m dans les parcelles et de 10 m a ` 1 km entre les parcelles. La biomasse microbienne variait par un facteur de 10 et nous avons observe ´ une forte variation dans la structure de la communaute ´ microbienne, me ˆme a ` des distances de 15 cm. Les e ´chantillons se regroupaient par parcelle lorsque les AGPL e ´taient soumis a ` une analyse en com- posantes principales. L’identite ´ des parcelles expliquait 36,3 % de la variation dans les AGPL et l’analyse ge ´ostatistique a montre ´ que la structure de la communaute ´ microbienne pre ´sentait une de ´pendance spatiale a ` des distances infe ´rieures a ` la dimension des parcelles. Les variables environnementales e ´taient significativement diffe ´rentes d’une parcelle a ` l’autre mais elles expliquaient une faible partie seulement de la variation dans le profil d’ensemble des AGPL. Cependant, les variables de la ve ´ge ´tation expliquaient le mieux le profil des AGPL et jusqu’a ` 60 % de la variation interne dans chaque parcelle pouvait e ˆtre explique ´e respectivement par les variables de la ve ´ge ´tation, le pH et l’e ´paisseur du sol. [Traduit par la Re ´daction] Introduction Soil biota are central to forest ecosystem functioning, as organic matter turnover and nutrient mineralization are gov- erned by the belowground community. The soil organisms live within a very heterogeneous matrix (Young and Ritz 1998), i.e., the soil. Here, physical and chemical characteris- tics are spatially aggregated at scales ranging from milli- metres or less to several kilometres (Cambardella et al. 1994; Brodsky ´ et al. 2001), leading to distinct spatial patterns often occurring at multiple scales simultaneously (Robertson and Gross 1994). These patterns in turn give rise to a corre- sponding variability in the soil microbial community (Morris and Boerner 1999; Green et al. 2004). The high abundance of soil microorganisms, in numbers of both species and individuals, is believed to be caused and maintained by resource heterogeneity and spatial isolation (Siira-Pietika ¨inen et al. 2001; Zhou et al. 2002). This in turn leads to great variation in the structure, size, and activity of the soil microbial community on both vertical (Fritze et al. 2000; Fierer et al. 2003; Garcia-Pichel et al. 2003; Agnelli et al. 2004) and horizontal (Saetre and Ba ˚a ˚th 2000) scales. Although the structure of the soil microbial community may change fundamentally over a few metres’ distance along nat- ural nutrient gradients in boreal forests (Nilsson et al. 2005), little is known about the scale of variation across boreal for- est landscapes. This is in contrast with most other groups of boreal forest organisms for which we have fairly detailed knowledge (e.g., see Berg et al. 1994; Ohlson et al. 1997). Describing such spatial patterns also for soil microbial com- munities is the first step in understanding the spatial distri- bution of the functional properties of these communities (Morris 1999). Soil microbial communities are largely structured by the vegetation, with plant species, plant diversity, and plant bio- mass being important factors (Pennanen et al. 1999; Saetre 1999; Hooper et al. 2000). Phospholipid fatty acid (PLFA) analysis is a sensitive method for analyzing microbial com- munity structure of the living microbial biomass (Tunlid and Received 26 April 2007. Accepted 3 December 2007. Published on the NRC Research Press Web site at cjfr.nrc.ca on 2 May 2008. L.H. Bach 1 and M. Ohlson. Department of Ecology and Natural Resource Management, Norwegian University of Life Sciences, 1432 A ˚ s, Norway. A ˚ . Frostega ˚rd. Department of Chemistry, Biotechnology and Food Science, Norwegian University of Life Sciences, 1432 A ˚ s, Norway. 1 Corresponding author (e-mail: lisbet.bach@umb.no). 1504 Can. J. For. Res. 38: 1504–1516 (2008) doi:10.1139/X07-232 # 2008 NRC Canada