On the effect of short temporal scale climate variability on long-term water, energy and carbon fluxes: Insights from a modeling approach

Abstract:

While the impact of rapid (i.e. sub-daily) fluctuations in climatic forcing on short-term fluctuations in water and carbon fluxes is rarely disputed, their aggregate effect on long-term fluxes and stores of carbon and water continues to be the subject of active research.

A process based ecohydrological model was used to unfold the effect of short time scale variability in precipitation, temperature and radiation on the water and carbon cycles across different climatic regimes and biomes worldwide. Specifically, synthetic climate inputs are first generated with prescribed statistical properties, modifying the small scale structure of each of the investigated forcing variable to evaluate its impact on ecosystem fluxes and stores at all temporal scales. The key statistical properties investigated here include: precipitation extremes, correlation structure, intermittency and its inter- and intra-storm structure; temperature distribution and correlation; and radiation distribution and their correlation structure. A statistical analysis and mechanistic explanations of how climate variability at short temporal scales affects evapotranspiration (ET) and its two components, net and gross primary productivity and vegetation dynamics at a wide range of temporal scales spanning from hourly to inter-annual is provided.

The main result is that short-lived excursions in climate forcings play a major role in controlling water, energy and carbon fluxes at all temporal scales. This finding is supported by the fact that nonlinearities describing the pathways by which climate variables impact carbon and water fluxes at short time scale do not necessarily ‘weaken’ with increasing temporal scales due to feedbacks and ‘soft’ thresholds and switches. The feedbacks, thresholds and switches depend on the limiting factor for each of the climatic regimes and are not expected to be universal, but site specific. Furthermore, the responses are mostly dependent on climate, rather than vegetation type. A future goal is include such effects in dynamic vegetation models using parameterization schemes that are aware of such cross-scale information flow.