Ecosystem service mapping based on land-cover classification has proven to be a valuable tool for studying the spatial distribution of services in order to inform management decisions. However, the lack of emphasis on the underlying mechanisms makes identification of thresholds of response to environmental change that will alter service provision problematic. A potential solution is to apply trait-based approaches to predict how the response of plant communities to environmental changes will impact ecosystem processes and the services they provide. While this approach has been employed in several studies in temperate and alpine grasslands, its applicability to different ecosystems remains unclear. Here, we implement trait-based approaches on the San Pedro River, a semi-arid riparian corridor in the Sonoran Desert, in order to predict the role of plants in mediating the impacts of decreased water availability on soil properties which serve as proxies for supporting ecosystem services. We selected sites along a gradient of water availability at which we measured soil and leaf litter C:N rations, soil organic matter, and specific leaf area (SLA), height, stem size, leaf C:N ratio and C isotopes for woody plants on transects beginning at bank-top and moving out to a distance of 50m.
Results/Conclusions
Initial results showed that both soil organic material and specific leaf area exhibit high within-site variability as is consistent with the spatial heterogeneity of semi-arid and riparian environments. Specific leaf area decreased by up to 40% along a gradient moving away from the river which corresponded with species turnover from cottonwood/willow gallery forest to mesquite bosque. We also saw high intraspecific variation of SLA in plants closest to the main channel which decreased by up to 60% with distance from the river. Despite these patters with landscape position, plant functional traits appeared to be weak predictors for soil properties with location-based characteristics showing greater importance. This suggests that while plant functional traits may be predictable to a degree based on environmental gradients, in highly variable environments such as arid riparian corridors they are weaker predictors of soil properties related to ecosystem processes and services than in other systems. Thus prediction of ecosystem service provision in such areas should take into account this high spatial variability. Further study should aim to identify typical levels of variation within these systems in contrast to thresholds representing substantial changes in ecosystem functioning.