Soil nutrient preferences between arbuscular and ectomycorrhizal trees

To monitor the recruitment dynamics of tree seedlings, we systematically established 1200 seedling quadrats within six 1-ha plots in early 2008. Within each 1-ha tree plot (100 m × 100 m), 100 seedling quadrats (1 m × 1 m) were set up in groups of four and placed within every alternate 10 m × 10 m sub-plot.
We collected a total of 600 soil samples from the six 1-ha plots in April 2017 and 1288 soil samples across the 50-ha plot in September 2023, and measured soil inorganic and organic fractions of N and P. Within each of the six 1-ha plots, soil samples were collected from alternating 10 m × 10 m sub-plot, obtaining a total of 50 regularly distributed soil samples (Figure S1). To characterize fine-scale directional variations in soil properties, we randomly selected one of the five regular sampling locations from each column of subplots. At each selected location, five soil samples were collected at distances of 0.1, 0.3, 0.8, 2, and 8 m along a randomly determined compass direction, generating 50 cluster samples per 1-ha plot. For the 50-ha plot, a total of 1288 soil sample were collected using a similar sampling design, with 588 fixed sampling points on the 20 m × 20 m grid, and 100 cluster sampling points were randomly selected with seven soil samples at distances of 2.66 m, 3.69 m, 4.02 m, 4.17 m, 4.78 m, 12.45 m and 18.67 m along a random azimuth from each fixed point.