Quantifying feedbacks in the plant-atmosphere-cloud continuum

Abstract

The main objective of this thesis was to gain insights on the impact of diurnal and local interactions between the vegetation, atmosphere and boundary layer clouds in current and future atmospheres. Special focus was placed on the consequences on moist convection, as it is one of the main uncertainties in the global climate and weather models. Moist convection is strongly influenced by the vegetated surface characteristics, which has consequences on the sub-weekly atmospheric state. In this thesis, a balanced approach is taken, which takes into account local (meters) and short (minutes) dynamic vegetation responses to atmospheric and cloud perturbations, that subsequently influence the atmospheric boundary layer and cloud development. To deepen our understanding of the processes that act on the smaller scales, a Large-Eddy Simulation (LES) model was employed and coupled with a mechanistic land-surface submodel. The LES model explicitly resolved the various dynamical processes at a scale of 50 m, which has an advantage over the coarser atmospheric models as minimal parametrizations are required. The investigations were based on a combined approach of advanced measurements and numerical experiments. The numerical experiments are based on observations over Western Europe, while for the future atmospheres the numerical experiments were inspired on results from a Free-Air CO2 Enrichment (FACE) experiment in Japan and combined with findings from literature. By following a systematic approach, our results highlighted the regional effects of a strong plant-atmosphere-cloud coupling (Chapter 2). In low wind and convective situations, a lowering in surface energy fluxes resulted in stabilized cloud development, although there was a distinct response based on the cloud optical properties. With increasing background wind, atmospheric roll vortices forced the cloud population into streets (i.e., parallel strips of clouds alternated by clear sky). As a consequence of an asymmetric stomatal plant response, vegetation streaks arose due to cloud shading that negatively affected the surface energy balance. This result shed light to a new coupling mechanism that constrained cloud development and reduced the in-cloud moisture flux (Chapter 3). To determine whether the plant-atmosphere-cloud coupling could be captured by homogeneous surface responses (i.e., related to a response in a NWP or GCM grid box), we performed simulations that were similar in the domain averaged surface energy, but differed in their response: interactive versus prescribed (Chapter 3). Our findings showed that large misrepresentation of up to 56% occurred in the regional moisture flux when the locality of the dynamic plant responses to atmospheric perturbations was not taken into account. This highlighted the need that these atmospheric flow dependent plant-atmosphere-cloud interactions need to be included in the parameterizations of the coarser NWPs and GCMs. By analyzing a comprehensive observational FACE dataset of two distinct rice varieties in ambient and elevated CO2 environments (+200 ppm) in Chapter 4, we identified a strong interplay between influencers on the plant-atmosphere interaction. In elevated CO2 environments, the physiological response to this factor became apparent, with warmer and drier in-canopy levels in a more closed and less photosynthetic active canopy, while the opposite was found in a more open and photosynthetic active canopy. Inspired by our findings of Chapter 4, we simulated and investigated the sensitivity of plant responses to elevations in both air temperature (+2 K) as [CO2] (+200 ppm) in Chapter 5. Our findings showed contrasting responses to elevations in temperature and [CO2] on the surface energy balance and momentum transfer. Elevations in temperature yields enhanced plant transpiration, thus latent heat flux, and reduced the sensible heat flux. As a consequence, the turbulent kinetic energy and buoyancy rates reduced, which caused reductions in cloud cover and mid-tropospheric moisture transport. With elevations in [CO2], a distinct response occurred, leading to higher sensible heat fluxes and lower plant transpiration and latent heat fluxes. With more momentum in the atmospheric boundary layer, clouds were able to become deeper and transport more moisture into the troposphere. When simulating a future atmosphere with both elevations in temperature and [CO2] in Chapter 5, we found an offset in the surface energy balance with nearly identical energy fluxes as compared to current situations. However, the plant physiological state was affected, with reductions in plant transpiration and increased CO2 assimilation. In conclusion, our results highlight the necessity of small scales and interactions, which require a bottom-up approach to be able to accurately capture the nonlinear plant-atmosphere interactions. Neglecting these interactions cause the coarser global climate and numerical weather prediction models to be liable to misrepresentations when modelling current and future atmospheres.