Carbon Lost vs. carbon gained: A study of carbon tradeoffs among land uses in Phoenix, AZ, reveals the inadequacy of statistical spatial scaling techniques and the need for new methodologies for understanding carbon dynamics across cities

Background/Question/Methods

Research has shown that with human modification and management urban landscapes can have considerable carbon storage pools associated with both vegetation and soils, as well as experience increased rates of net primary productivity. However, as there are now a number of studies that indicate the social and ecological drivers of urban vegetation cover are complex and variable across cities, there is an urgent need to understand the heterogeneity of carbon dynamics across multiple land use and land cover types. We estimated carbon storage and net primary productivity associated with trees and shrubs across the Central-Arizona Phoenix Long-Term Ecological Research site (CAP LTER).   Plot-level data were scaled up to regional values for a diversity of land uses using traditional scaling and hierarchical Bayesian and non-parametric statistical modeling approaches. 

Results/Conclusions

Statistical differences among carbon storage and NPP for the different land-use types indicated that vegetation cover was a function of human decision-making, with the urban land uses having higher tree carbon storage/NPP than the surrounding desert and agricultural landscapes.  Our results also show that shrubs actually make a substantial contribution to carbon storage/NPP in the CAP ecosystem, and much of the gains in carbon storage/NPP associated with increased tree cover in the urbanization process were likely partially offset by losses of shrub cover from desert landscapes.  Although all of the analyses showed a change in carbon storage/NPP with urbanization, the results varied considerably among the scaling methodologies.  Unlike previous research that has shown potential for Bayesian scaling techniques applied to soil carbon and nitrogen pools, our results indicate that the statistical models do not do well at predicting the spatial distribution of vegetation-related carbon dynamics.  This study supports evidence that there is a need to develop spatially explicit and standardized methods for analyzing carbon dynamics associated with vegetation in urbanizing areas.  Furthermore, our analysis indicates studies on carbon change with urbanization should consider including data for shrubs. Until now, many studies have excluded this vegetation type and have assumed it has very little to contribute in terms of carbon storage/NPP.