Spatial patterns of stand characteristics and soil nutrient availability in reclaimed boreal forests

Background/Question/Methods

Oil sands surface mining has currently disturbed roughly 650 km² of boreal ecosystems in northern Alberta and this area is projected to increase to approximately 1500 km² in the next 10-15 years.  Forest reclamation after the complete removal and replacement of soil materials represents a major ongoing challenge that is partially contingent on re-establishing plant connections to soil biogeochemical cycles. Site characteristics including forest floor mass, tree location, canopy area, pH, and seasonal nutrient availability were quantified with a spatially explicit sampling protocol in three reclaimed stands (Populus tremuloides, Pinus banksiana, and Picea abies).  Spatially explicit sampling allowed us to remove the effect of unmeasured variables and determine mechanistic relationships. A random walk design was used with a minimum lag distance was 0.5 m and at least 40 pairwise comparisons per lag distance. Stand characteristics included tree location, canopy cover, and forest floor mass. We assessed seasonal nutrient availability using three sets of plant root simulator probes (PRS^TM, Western Ag Innovations) placed in August 2008, November 2008, and May 2009.  Spatial patterns were accessed with semi-variograms and spatial ANOVA in R.

Results/Conclusions

Spatial patterns in soil nutrient availability were generally patchy at scales of 3 to 20 m and nutrients varied among seasons both in patch size and quantity measured.  After 25 years, fine scale patterns for some features (for example aspen nitrogen (N) availability, 3 m patch) were similar to natural forest structure (aspen canopy cover, 3 m patch). However, large scale patterns (for example aspen phosphorus (P) availability, >20 m patch) must represent relics of soil placement during reclamation.  Spatial ANOVA indicated that canopy cover was the dominant factor leading to increased N, sulphur, and calcium in the aspen stand, but was only related to potassium in the spruce stand.  Forest floor mass, a partial indicator of soil development, was only related to phosphorus and potassium availability in the aspen stand.  The spatial residuals were significantly related to N cycling in the conifer stand types, and P in the aspen stand, indicating that a mechanistic relationship had been missed, which was most likely related to microbial turnover. These results provide evidence that soil-plant relations are re-establishing differently in aspen and conifer stands, with aspen nutrient availability more dominated by canopy dynamics (top down) and conifer nutrient availability more dominated by spatial residuals (bottom up?).