Appendix S1. Sampling methods for environmental variables; supplementary references

Table S1. Description of study system

Table S2. Results on sequential tests of relationships between predictor variables and α-diversity

1

Description of the study system

Table S1.Description of the study system with differences in biotic and abiotic factors (mean values ± 1 SE) among island size classes.

Response variable / Large island
(> 1.0 ha) / Medium island
(0.1-1.0 ha) / Small island (<0.1 ha) / F (P)
Disturbance regime
Time since last major fire (14C date) (year) / 585 (233) c / 2180 (385) b / 3250 (439) a / 13.4 (<0.001)
Number of fire scars caused in the past 250 years / 0.667 (0.256) a / 0.208 (0.085) b / 0.413 (0.016) b / 3.5 (0.037)
Humus properties
(0-10 cm depth)
pH / 3.51 (0.029) a / 3.42 (0.027) ab / 3.38 (0.039) b / 3.4 (0.034)
Total N (%) / 1.28 (0.06) b / 1.46 (0.04) a / 1.59 (0.07) a / 8.7 (<0.001)
Total P (%) / 0.087 (0.005) a / 0.097 (0.003) a / 0.091 (0.003) a / 1.5 (0.223)
Total C:N ratio / 41.0 (1.7) a / 35.3 (0.8) b / 32.8 (1.2) b / 10.8 (<0.001)
Total C:P ratio / 600 (31) a / 532 (16) a / 556 (20) a / 2.2 (0.127)
Total N:P ratio / 14.7 (0.5) a / 15.2 (0.5) a / 17.5 (0.8) b / 5.7 (0.009)
Mineral N (MIN) (µg g-1) / 38.2 (14.4) b / 58.1 (9.2) a / 25.3 (8.0) b / 13.9 (<0.001)
Dissolved organic N (DON)
(µg g-1) / 39.1 (7.2) b / 50.7 (5.5) a / 40.3 (4.6) b / 3.6 (0.028)
MIN/DON / 0.49 (0.04) a / 0.53 (0.05) a / 0.39 (0.03) / 5.7 (0.009)
Phosphate (µg g-1) / 43.6 (4.9) a / 37.7 (4.3) a / 24.7 (2.3) b / 5.9 (0.007)
Membrane-extractable P (µg g-1) / 87.4 (10.2) ab / 97.1 (6.5) a / 74.7 (5.3) b / 3.7 (0.039)
Carbon (C) storage (kg m-2)
Aboveground C / 3.5 (0.3) a / 3.1 (0.3) a / 1.8 (0.2) b / 13.3 (<0.001)
Belowground C / 6.4 (1.1) c / 16.2 (2.5) b / 27.2 (2.5) a / 24.2 (<0.001)
Total C / 9.9 (1.1) c / 19.2 (2.6) b / 29.0 (2.4) a / 20.1 (<0.001)
Vascular plant
biomass (g m-2)
P. sylvestris / 7409 (664) a / 3419 (1019) b / 586 (345) c / 22.1 (<0.001)
P. abies / 615 (343) a / 2443 (1078) a / 1838 (321) a / 1.9 (0.172)
B. pubescens / 987 (244) b / 2082 (379) a / 777 (207) b / 6.0 (0.007)
V. myrtillus / 108 (17) a / 59 (9) b / 33 (8) b / 10.4 (<0.001)
V. vitis-idaea / 198 (23) b / 271 (15) a / 142 (16) b / 12.2 (<0.001)
E. hermaphroditum / 32 (5) b / 29 (6) b / 92 (9) a / 25.3 (<0.001)
Net primary productivity (kg m-2 yr-1) / 169 (11) a / 167 (9) a / 101 (13) b / 12.1 (<0.001)
Vascular plant α-diversity / 6.6 (0.5) c / 8.7 (0.4) b / 10.6 (0.6) a / 17.2 (<0.001)

Data fromWardle et al. 1997, 2003, Wardle and Zackrisson 2005, Lagerström et al. 2009,Gundaleet al., 2011. Within each row, numbers followed by the same letter are not statistically significant at P = 0.05 (Tukey’s test following one-way ANOVA with 2,27df; n = 10 for each size class). Values of Pin bold from one-way ANOVAs are significant at P = 0.05. Island size in this table is used as a surrogate of fire history, as it integrates history of all fires, not just the most recent ones (Wardle et al. 1997, 2003).

Supplementary results

Table S2.Sequential tests of relationships between predictor variables and α-diversity (species richness and effective number of species).

Predictor variables
Organism groups / Ecosystem age (yrs) / NPP (kg m-2 yr-1) / Total C storage (kg m-2) / Soil N:P ratio
Species richness
Spiders / 0.295** / 0.295 / 0.311 / 0.313
Nematodes / 0.001 / 0.059 / 0.076 / 0.076
Beetles / 0.239** / 0.241 / 0.311 / 0.332
Epiphythic lichens / 0.068 / 0.158 / 0.160 / 0.187
Vascular plants / 0.095 / 0.117 / 0.279* / 0.286
Litter fungi / 0.347** / 0.353 / 0.495** / 0.506
Root fungi / 0.229** / 0.241 / 0.318 / 0.362
Effective number of species
Spiders / 0.001 / 0.003 / 0.015 / 0.050
Nematodes / 0.086 / 0.213* / 0.246 / 0.262
Beetles / 0.002 / 0.011 / 0.069 / 0.073
Epiphythic lichens / 0.032 / 0.076 / 0.103 / 0.103
Vascular plants / 0.065 / 0.095 / 0.097 / 0.097
Litter fungi / 0.146* / 0.163 / 0.290* / 0.295
Root fungi / 0.168* / 0.180 / 0.240 / 0.331

Cumulative R2values are shown. Significant relationships are presented in bold, where , *, and ** represent significant relationships at P = 0.1, 0.05, and 0.01, respectively.

Sampling of environmental variables

Detailed descriptions of methods for measuring the four environmental variables used in this study are given in the original publications reporting the data that we used (i.e., Wardle et al. 1997, 2003, 2012b), and brief descriptions of these methods are as follows:

To determine ecosystem age (i.e., the time since the last catastrophic fire disturbance that reset the ecosystem) on each island, charcoal particles in the uppermost charcoal layer in the humus profile were sampled, and the time since the last fire was determined by 14C analyses of these particles at the Tandem laboratory, Uppsala, Sweden (Wardle et al. 1997, 2003).

To determine total ecosystem C storage on each island, C storage was measured separately both aboveground and belowground (see Wardle et al. 1997, 2003, 2004). For evaluation of aboveground C storage in tree biomass, one circular plot (10 m in radius) was established on each island. In each plot, the diameter of each tree was measured at a height of 1.3 m. Established relationships for converting diameter into biomass for each species were used for determining total tree biomass on an areal basis. Further, in the area used for these plots, four plots of 0.5 m by 0.5 m were also established, and all dwarf shrub and moss species were trimmed at ground level and sorted into component species for biomass determination. All plant biomass was converted to C by multiplying by 0.47. For belowground C, humus samples were collected to the entire humus depth from each of the four 0.5 m by 0.5 m plots used for ground layer vegetation determination and were combined for each island, and %C measured for this humus was used to determine C storage per unit area. Total %N and %P was also measured on this humus for determination of humus N:P ratio.

Total NPP was also measured aboveground for each island (Wardle et al. 2003, 2012b), as the sum of tree and understory vegetation productivity. To measure tree NPP, five individuals of each tree species were selected from each island. For each of these trees, a single core of wood was sampled at 1.3-m height, and its radial growth rate in the preceding three years was determined by measuring the width of the three most recent tree rings. Productivity of each tree in each plot was calculated by determining the mean radial growth of that species on that island over the previous three years, subtracting that value from the current radius of that tree, and using allometric relationships for estimating what its existing biomass would have been three years earlier. Total productivity of all trees alive in each plot over the previous three years is the sum of tree biomasses (see above) at the time of measurement minus the sum of biomasses of these trees three years earlier. To obtain understory vegetation productivity, the biomass and production of each of the three dominant ericaceous dwarf shrub species (collectively account for over 98% of vascular plant understory biomass) was measured. For each island, the point intercept method was used to determine the total number of times the vegetation of each species was intercepted by a total of 200 downwardly projected points. The total number of intercepts for each species was then converted to biomass per unit area through the use of established calibration equations. For each island, 20 stems of each of the three shrub species were then selected in the portion of the island used for point analysis; these were cut at ground level and the portion of the stem produced in that growing season was separated from the stem material produced in all preceding growing seasons. This was used to determine the proportion of biomass present in each growing season that has been produced earlier in that season for each species, and therefore its productivity per unit area that year.

Supplementary Literature Cited

Gundale, M. J., A. Fajardo, R. W. Lucas, M. -C. Nilsson, and D. A. Wardle. 2011. Resource heterogeneity does not explain the productivity-diversity relationship across a boreal island fertility gradient. Ecography34:887–896.

Lagerström, A., C. Esberg, D. A. Wardle, and R. Giesler. 2009. Soil phosphorus and microbial response to a long-term wildfire chronosequence in northern Sweden. Biogeochemistry95:199–213.

Wardle, D.A. and O. Zackrisson.2005. Effects of species and functional group loss on island ecosystem properties. Nature 435:806–810.

Wardle, D.A., G. Hörnberg, O. Zackrisson, M. Kalela-Brundin, and D. A. Coomes. 2003. Long-term effects of wildfire on ecosystem properties across an island area gradient. Science300:972–975.

Wardle, D.A., M. Jonsson, M. Kalela-Brundin, A. Lagerström, R. D. Bardgett, G. W. Yeates, and M. -C. Nilsson.2012b. Drivers of inter-year variability of plant production and decomposers across contrasting island ecosystems.Ecology 93:521–531.

Wardle, D.A., O. Zackrisson, G. Hörnberg, and C. Gallet.1997. The influence of island area on ecosystem properties. Science277:1296–1299.

1