Mapping terrestrial evaporation regimes : a data-driven analysis of land-atmosphere interactions under climate change

Terrestrial ecosystems are essential for human life, as they provide several key services, such as food and water security, sequestration of anthropogenic carbon dioxide (CO2) and evaporative cooling. Healthy ecosystems perform photosynthesis, during which the plants’ stomata, tiny apertures on the leaf surface, are open to take up CO2 and evaporate water into the atmosphere, thereby providing a cooling effect. At the Earth’s surface, all the energy that is available from solar radiation that is not used to evaporate water, is used to heat near-surface air. Thereby, ecosystems play a key role in connecting the energy and the water balance.In Chapter 1, I introduce the energy and the water balance, which are linked through the evaporation of terrestrial water by ecosystems. How much of the available energy is used to evaporate water from the land surface, the evaporative fraction (EF), depends on water supply (soil moisture); When the soil is too dry for plants to extract water from the soil, plants close their stomata to prevent water loss. As the soil moistens, plants can access this water and increase their transpiration with further soil moisture increases until the maximum EF is reached. At this point, any additional soil moisture increases will no longer affect the EF, rendering the magnitude of the evaporation of water, the terrestrial evaporation (ET), solely dependent on available energy. The soil moisture and the corresponding terrestrial evaporation regime can dampen or amplify heat waves and other extreme events. Therefore, distinguishing energy versus water-limited conditions and accurately representing in Earth system models is essential.The aim of this thesis is to study the relevance of water versus energy limited conditions and associated regime transitions for weather, climate and related extremes under climate change. In this data-driven analysis, a multitude of data streams is used, amongst which satellite observations, in-situ observations, such as weather balloons, gridded machine-learning based data sets and heavily constrained and unconstrained model output from Earth system models. Using such a combination of data products provides a comprehensive perspective that allows to infer the role of limitations in individual data products by comparing results between them.In addition to research focusing solely on either energy or water, I introduce a novel index in Chapter 2 that considers the relevance of both water and energy availability for terrestrial evaporation. Using satellite observations over Europe, energy limitation is found in Northern Europe and water limitation in the Mediterranean, separated by a transitional zone across Central Europe. A soil moisture threshold, the critical soil moisture (CSM), which separates energy and water-limited conditions, is determined at the continental and the grid cell scale. This transition matters, as it separates water-limited conditions, where soil moisture can impact the near-surface temperature through evaporation, from energy-limited conditions, where atmospheric dynamics and related temperature and solar radiation impact the evaporation of water. The CSM is sensitive to local climate, soil and vegetation characteristics, but is robust when determined using proxies from other data products for energy availability and terrestrial evaporation.As land-atmosphere interactions work both ways, land surface conditions can also be inferred from an atmosphere perspective: In Chapter 3, atmospheric temperature and humidity measurements from weather balloons are translated into landscape-scale surface fluxes using a mixed-layer model. Juxtaposing these surface fluxes against satellite-observed surface soil moisture, the CSM is estimated, which is then used to separate water versus energy-limited conditions. Further, I find that afternoon atmospheric boundary layers (ABLs) are warmer, drier and deeper over water-limited as compared to energy-limited surfaces.In Chapter 4, future trends ecosystem water limitation trends are studied with model output from a suite of state-of-the-art Earth system models, which (i) can complement data lacking from observations (e.g. root-zone soil moisture) and (ii) allows to infer future climate conditions. Ecosystem Limitation Index (ELI) trends based on historical simulations and future projections reveal a widespread increase in ecosystem water limitation under climate change. Although globally this is driven mostly by increasing energy availability, several hot spot regions are identified where ecosystem water limitation increases particularly rapid owing to simultaneous decreases in water availability. This shift towards ecosystem water limitation occurs in space, as transitional zones shift northwards in Canada and northern Eurasia and northeastwards in the Amazon, and in time through widespread lengthening of the water-limited season. Although net surface radiation is the most dominant driver for ELI trends, a full range of eco-climate variables is needed to understand local differences.In previous chapters, I mainly focused on mean climate conditions, whereas the role of land- atmosphere interactions for climate extremes is even more pronounced: In Chapter 5, I find that in about half of the land area with active vegetation, vegetation productivity extremes jointly occur with hydrometeorological hazards. The drivers of such vegetation extremes vary in space: In the Northern Hemisphere, temperature-related hazards are found important above 50°N, whereas water-related hazards most strongly impact vegetation productivity below 50°N. Further, several key regions are identified where vegetation productivity maxima are energy- driven and minima are governed by water, which is characteristic for transitional regimes and strong land-atmosphere coupling.In Chapter 6, an ensemble of Earth system models is used to find that in 77% of the vegetated land surface, maximum temperature is increasing more rapidly than mean temperatures under climate change. In several hot spot regions, such temperature divergence increases occur alongside shifts towards ecosystem water limitation and related reductions in evaporative cooling. Whereas the magnitude of temperature divergence trends is largest in initially energy-limited and transitional regions, where ELI trends are strongest, the sensitivity is highest in initially water- limited regions, where evaporative cooling is most strongly affected by increasing ecosystem water limitation.In Chapter 7, specific contributions, limitations and future research opportunities are synthesized. In conclusion, a novel index was introduced that jointly considers the relevance of energy and water availability for terrestrial evaporation. Typically, water-limited conditions are characterized by dry soils and correspondingly low terrestrial evaporation and high sensible heating of near- surface air. As such, ABLs are warmer and drier over water versus energy limitation, thereby amplifying heat extremes. Hot spot regions are found where heat extremes are amplified by reductions in evaporative cooling. Identifying such regions is essential to tailor adaptation strategies, such as irrigation or tillage practices, city greening and afforestation, to mitigate impact on human health and infrastructure. Further, these studies allow the comparison of modelled and observed land-atmosphere coupling, which could inform model development and lead to better weather and climate forecasts and accurate determination of land-atmosphere coupling hot spots and associated extreme events