Carbon and nitrogen dynamics associated with post-wildfire stand development of boreal mixed conifer ecosystems : application of data in model validation and refinement

Abstract

Study site : Northwestern Ontario

To make accurate and meaningful predictions about boreal forest ecosystem development, a strong understanding of the processes that drive change within the ecosystem s in required. Once this knowledge is gained, it can be incorporated into process-based ecosystem models that can be valuable tools for both making management decisions and also guiding further research. In northwestern Ontario, various calibration and testing exercises for the CENTURY Soil Organic Matter model have been performed. However, it was identified that a validation exercise was required in order to properly evaluate the ability of the model to simulate C and N dynamics through stand development. In response to this need, this project involved the set up of a 110 year old upland mixed conifer chronosequence of fire origin from which C and N dynamics were measured (i.e. C and N pool levels and fluxes). By dividing the chronosequence into four stand development stages (regeneration, crown closure, self-thinning, and mature), it was found that pool levels were often not significantly different between stand development stages; however, N mineralization and soil respiration rates were linked to soil temperature, and substrate quality seems to decrease with time since fire disturbance. The validation exercise employed eight different tests to evaluate the ability of the CENTURY model to simulate thirteen C and N pools, fluxes, and soil temperature calculations. The exercise was much less successful than expected. In most cases the pool levels were different and in other cases, the over time trends were very different. For example, simulated N mineralization rates increase gradually through stand development whereas real-world measured rates decreased into immobilization. It was found that growing season soil temperatures (May to September) calculated by the model were too high and thought that lowering these temperatures would result in some of the simulated pools and fluxes being m ore similar to the real-world data. Validation of the new low er soil temperatures was very successful; however, this success did not translate into better validation results for the other parameters. Further work with CENTURY is required before it can be relied upon to accurately simulate C and N dynamics through stand development; however, this type of work is valuable in itself because it generates new information and furthers out understanding of these forest ecosystems.