Open AccessDetection of forest stand‐level spatial structure in ectomycorrhizal fungal communities
Author(s) -
Lilleskov Erik A.,
Bruns Thomas D.,
Horton Thomas R.,
Taylor D.,
Grogan Paul
Publication year - 2004
Publication title -
fems microbiology ecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.377
H-Index - 155
eISSN - 1574-6941
pISSN - 0168-6496
DOI - 10.1016/j.femsec.2004.04.004
Subject(s) - mantel test , biology , species evenness , spatial analysis , sampling (signal processing) , spatial ecology , ecology , community structure , spatial distribution , common spatial pattern , statistics , similarity (geometry) , biomass (ecology) , range (aeronautics) , spatial variability , species richness , mathematics , genetic variation , biochemistry , materials science , filter (signal processing) , composite material , artificial intelligence , gene , computer science , image (mathematics) , computer vision
Ectomycorrhizal fungal (EMF) communities are highly diverse at the stand level. To begin to understand what might lead to such diversity, and to improve sampling designs, we investigated the spatial structure of these communities. We used EMF community data from a number of studies carried out in seven mature and one recently fire‐initiated forest stand. We applied various measures of spatial pattern to characterize distributions at EMF community and species levels: Mantel tests, Mantel correlograms, variance/mean and standardized variograms. Mantel tests indicated that in four of eight sites community similarity decreased with distance, whereas Mantel correlograms also found spatial autocorrelation in those four plus two additional sites. In all but one of these sites elevated similarity was evident only at relatively small spatial scales (< 2.6 m), whereas one exhibited a larger scale pattern (∼25 m). Evenness of biomass distribution among cores varied widely among taxa. Standardized variograms indicated that most of the dominant taxa showed patchiness at a scale of less than 3 m, with a range from 0 to 17 m. These results have implications for both sampling scale and intensity to achieve maximum efficiency of community sampling. In the systems we examined, cores should be at least 3 m apart to achieve the greatest sampling efficiency for stand‐level community analysis. In some cases even this spacing may result in reduced sampling efficiency arising from patterns of spatial autocorrelation. Interpretation of the causes and significance of these patterns requires information on the genetic identity of individuals in the communities.