Incorporating sink and source dynamics to improve the performance of a forest ecosystem model

Background/Question/Methods

It remains unclear to what extent tree growth is limited by either the availability of carbon (source) or the ability to use the carbon (sink).  Despite the uncertainty, most models of ecosystem biogeochemistry implicitly assume source limitation, and consequently have difficulties predicting known responses such as the patterns of root respiration under changing conditions.  Experiments show that the increase in atmospheric CO₂ leads to an increase in root respiration, whereas N fertilization leads to a decrease.  These two conditions represent the extremes of sink or source limited scenarios, as elevated CO₂ increases photosynthesis and N fertilization increases growth.  Would including sink limitation improve model performance?  We examined this question by incorporating sink limitation into an ecosystem model, Century, specifically the forest version with a daily time step, DayCent-Photosyn. We compared the model predictions to the published data from the Duke FACE experiment. 

Results/Conclusions

We incorporated sink and source dynamics into the model by creating a carbohydrate pool.  The resulting model first estimates photosynthesis and partitions the fixed carbon into a carbohydrate pool.  It then partitions the carbohydrate pool into maintenance respiration and biomass production (and associated growth respiration).  This structure allowed for source and sink dynamics to be controlled separately.  We further incorporated sink limitation and feedback inhibition through known mechanisms, by allowing growth to be more sensitive to environmental conditions than photosynthesis, and by allowing the carbohydrate pool to regulate photosynthesis. The model successfully simulated the dynamics of soil respiration under both N fertilization and elevated CO₂.  Fine root growth in the model responded most rapidly to changes in carbohydrate levels, resulting in lowered root growth under N fertilization and associated growth respiration.  Lower root growth also led to lower root biomass and thus decreased maintenance respiration.  Elevated CO₂ increased the carbohydrate pool, and it in turn stimulated fine root growth.  This resulted in increased growth respiration, root biomass, and maintenance respiration.  The phenology of growth was a major factor in adequately simulating the seasonal changes in the carbohydrate pool.  Our results indicate that incorporating sink and source dynamics improves model performance.