Skip to main content
Fig. 2 | BMC Ecology

Fig. 2

From: Using multivariate cross correlations, Granger causality and graphical models to quantify spatiotemporal synchronization and causality between pest populations

Fig. 2

Cross correlation (CRCO) and partial cross correlation (PACO) statistical similarity measure matrices and the respective network configurations; a similarity-values matrices (E) and related full connected weighted networks: computed using multivariate cross correlations and partial cross correlations. b Significant p- values matrices (E’) and related full connected weighted networks: computed using multivariate tests of correlation using the α = 0.05 significance level for thresholding and c binary-adjacency matrices (E’’) and related binary networks: computed using false discovery rate (FDR) to correct and adjust false p values. Notice that the binary-adjacency matrix is composed of an upper and lower triangular part which are identical in which the white cells indicate significant interaction between the time-series and are used to draw the links. Matrix corresponds to K = 15 ecological time series of length n = 285 (intervals of 3 Julian day during the growth season; W1: temperature (oC), W2: relative humidity (%), X1–X8: A. orana, Y1–Y3: A. lineatella and Z1–Z2: G. molesta moth population (individuals). Deeper (hot) colors, in either matrices or networks, indicate higher correlations and implications for each pair of variables tested, while light colors and thinner vertices are connected to lower correlations and implications

Back to article page