Extended Data Table 5 Coefficients for the analysis of the relative abundance of EcM trees among and within 16 lowland tropical forests at a fine scale

From: Tropical tree ectomycorrhiza are distributed independently of soil nutrients

  1. Parameter is the variable included in the model, and its estimate, LCI, and UCI (95% credible intervals) are calculated from the posterior predictive distribution. The table is partitioned both horizontally and vertically. Horizontally, the left section shows fixed effect coefficients, whereas the right section shows the random effect coefficients. Vertically, the table is split into three sections. The top and middle sections show coefficients for the discrete and continuous components of the ZABE model, respectively. The bottom section lists coefficients for the spatial random terms (BYM2 model) for both components. Fixed effects coefficients in bold exclude zero and are considered influential, whereas fixed effects coefficients with credible intervals that contain zero (not in bold) are considered negligible. Random effects are restricted to be positive. Precision is indicated by τ, with τ = 1/√σ, where σ is the standard deviation. For the spatial random terms in the discrete component, the parameter ‘τ ID.z’ indicates the posterior of the precision of the spatial field and the parameter ‘ɸ ID.z’ indicates the posterior of the precision of the mixing parameter (see104 for details). The parameter ‘β ID.o’ in the spatial random terms for the continuous component represents the estimated scale parameter. All continuous predictors in the model were standardized using their z-score, and the total basal area was transformed using the natural logarithm before standardization ((log) BA).
Back to article page