Factors driving the relationships between vegetation and soil properties in the Yellow River Delta, China

Fig. 4 shows that cover percentage was determined by OM, K+, TN and SO42− at regional scale, while at sub-regional scale, it was positively determined by SW, TN, TP and OM. Fig. 4 also shows that plant height was positively correlated with OM, Na+ and K+, and negatively correlated with SO42− at regional scale; plant height were mainly determined by OM, TP, TN and SW at sub-regional scale and landscape scale, plant height were positively correlated with K+, TP and SO42−, while negatively with NH4+. For the aboveground biomass, were positively correlated with TP and NO3− at regional scale, while negatively correlated at sub-regional and landscape scales. On the basis of CCA results, we selected the lowest Akaike Information Criteria value using the stepAIC function to quantify the effect of soil factors on vegetation properties at different scales.154