Towards better representation of water, CO2, and temperature exchanges at the land-atmosphere interface

Abstract

The movement of water between the surface and the atmosphere is often linked to exchanges of other quantities such as CO2 and temperature. Therefore, understanding how to parameterize these fluxes is essential in many applications. To this end, this dissertation starts by investigating the Monin-Obukhov Similarity theory (MOST), which is widely adopted to parameterize fluxes in Earth observations and models. One of the key findings is that MOST misrepresents fluxes of low magnitude, such as CO2 fluxes from lakes and water bodies. An alternative to MOST --- the Relaxed Eddy Accumulation (REA) --- was examined and found to be more robust under non-ideal conditions. Broader use of REA is thus recommended in field experiments, while improvements or alternatives to MOST in numerical models should be pursued in future research. This dissertation also studies the partitioning of water and CO2 fluxes into plant (photosynthesis and transpiration) and ground (respiration and evaporation) components. On one hand, reliable partitioning methods or measurement techniques capable of distinguishing the four components are still lacking. On the other hand, better quantification of these fluxes is essential in many applications such as optimization of irrigation techniques. This problem is tackled in this dissertation from the perspective of scalar turbulence behavior in the atmosphere, where four methods relying on high frequency measurements are explored. Their robustness and performance are first evaluated using experimental data, followed by a more detailed investigation of their physical assumptions with large eddy simulations. Overall, by combining experimental and numerical methods, this dissertation discusses when and where such methods are expected to perform well across various ecosystems. It also underlines a set of guidelines to be followed in new field experiments aiming at maximizing the accuracy of these partitioning models. Moving from plant canopies to cities, this dissertation concludes by investigating the interactions between rainfall and urban environments. More specifically, the heat exchange between runoff and hot pavements during summer storms. A study of 100 urban streams in the US shows that stream water temperature can dramatically increase by up to 10 C a few minutes after the beginning of the storm in very developed areas. In addition to these short-lived temperature jumps, this study identified that urban streams are, on average, hotter than their less disturbed "rural'' counterparts. This phenomenon is here defined as a Hydrologycal Urban Heat Island, which is a first step towards better characterizing the different impacts that urbanization have on urban hydrology and atmosphere.