Relationship of biodiversity to island size and isolation for a fire-prone oligotrophic ecosystem

Background/Question/Methods

The Theory of Island Biogeography (IBG) has been used in terrestrial systems to describe the consequences of habitat fragmentation to biodiversity. One aspect of IBG rarely considered is how biodiversity changes when disturbance is correlated with island (fragment) size. For instance, large lake islands in fire-prone boreal forests of Scandinavia burn more frequently and have lower plant diversity than small islands. This response in diversity is paradoxical to the theory of IBG. Little has been done to extend this work to ecosystems with higher fire frequencies or relate this to the Intermediate Disturbance Hypothesis (IDH). We used lake islands in Cree Lake, Saskatchewan to examine fire frequency across island size and isolation and to test IBG and IDH theories on biodiversity. Cree Lake contains 706 islands (up to 6,000 ha) and has one of the highest fire return intervals (FRI) for boreal forests of North America. We examined fire history for 706 islands over a 31 year period (1984-2012) using Landsat satellite data and related the occurrence of fires on islands to island size, isolation, and island characteristics to predict FRI. Plant diversity was sampled in 0.1 ha plots on 41 randomly stratified islands ranging in size from 0.1 to 6,000 ha. Structural equation models were used to relate biodiversity responses to both IBG and FRI factors.

Results/Conclusions

We predicted that FRI for any one particular island patch (pixel) followed an inverse power law relationship with island size (log₁₀FRI to log₁₀size). Specifically, FRI was predicted to range from 62 years on the largest (>1,000 ha) islands, 286 years on moderate-sized (100 ha) islands, and 1,600 years on small (0.1 ha) islands thus demonstrating a 25-fold difference in the disturbance regime. Vegetation differences on islands were apparent with small islands being spruce-bryophyte dominated and large islands typified as being lichen-woodlands. Plant richness was inversely related to island isolation and positively related to canopy cover, while canopy was positively related to FRI and inversely related to island size. In contrast to species richness, plant rarity was inversely related to canopy cover and stand age. We found support for IDH when considering beta diversity across island sizes.