Projecting future climate impacts on soil biogeochemical processes in spatially heterogeneous drylands

Background/Question/Methods

Ecological processes are often under different mechanistic controls in drylands than better-characterized mesic systems, challenging efforts to project dryland responses to climate change. This challenge is compounded by the outsized role played in drylands by spatial heterogeneity at multiple hierarchical levels of organization. Characterizing biogeochemical processes in drylands is critical, however, as these systems cover approximately 45% of the global land surface and are predicted to expand and become drier in the future. These systems contribute roughly one third of terrestrial net carbon (C) uptake through photosynthesis, while accounting for roughly one third of the global soil organic C pool and a substantial portion of the interannual variability in atmospheric C. Understanding the controls over C release from drylands is therefore critical for accurate prediction of atmospheric C pools, both now and in the future climate scenarios. Our understanding of the controls over C cycle processes is particularly lacking in hyper-arid and arid systems that account for 20% of the terrestrial area but less than 1% of soil respiration observations. We explored variability in dryland soil respiration response to rainfall pulses across a hyper-arid to arid rainfall gradient in the Namib Desert, allowing us to assess how future climate might influence soil respiration. The rainfall gradient is superimposed across multiple levels of spatial heterogeneity, allowing us to assess responses across two soil surfaces that differ in stability and composition and among different geomorphic and vegetation patch types.

Results/Conclusions

Soil respiration was highly responsive to rainfall pulses, although soil surfaces and patch types often exerted more control on soil respiration than did rainfall alone. Soils at the wetter (arid) end of the rainfall gradient were generally more responsive to rainfall pulses than were soils in the hyper-arid zone. Greater soil respiration occurred on the relatively stable gravel plains soil surfaces than on the shifting, unstable sandy soil surfaces. However, soil respiration was also greater in patches characterized by depositional inputs of organic matter rather than in erosional patch types characterized by organic matter losses. We discuss these results in light of the challenges to data interpretation given extreme spatial heterogeneity of drylands. We suggest approaches for considering multiple scales of spatial heterogeneity when interpreting and projecting dryland biogeochemical fluxes in response to climate change.