The terrestrial water, energy and carbon cycles are tightly coupled through land-atmosphere (L-A) interactions, not only regulating local plant physiological activities and also modulating regional and global climate. With ongoing anthropogenic greenhouse gas emissions, many of these interactions can be modified and complicated. To better anticipate and adapt to future climate, it is of great importance and necessity to deepen and refine our understanding of the complex L-A interactions. In this dissertation, three topics are investigated across the ecosystem, regional and global scales respectively, throughout which, the critical role of dryness or drying in the context of global warming is highlighted.
ππ‘ππ©πππ« π: Evapotranspiration (ET) is a key component that connects the continental water, carbon and energy cycles and a proxy that measures the coupling strength between the biosphere and atmosphere. A wide range of biophysical factors, which usually exhibit nonlinearity and strong covariation, collectively modulate ET and complicate the overall understanding of ET dynamics. In the first study, the causal discovery frameworks PCMCI+ and Latent PCMCI are utilized with integrated priori physical knowledge to identify the dominant drivers and constraints of ET in the growing seasons across sites, with a particular focus on the role of site dryness degree. The Dryness Index (DI), defined as the ratio of annual mean net radiation to precipitation, has been introduced to assess the water availability relative to energy supply at different locations. By analyzing the daily observations from 115 flux tower sites and satellite remote sensing, it has been discovered that the feedbacks around ET are mediated by the degree of dryness: at sites with adequate water supply (using PCMCI+, the DI value averaged from such sites is 1.33), the atmospheric conditions, including incoming solar radiation and atmospheric demand for water (indicated by vapor pressure deficit, VPD), prevail in driving ET; in contrast, in semi-arid and arid areas where the water stress is high (using PCMCI+, the DI value averaged from such sites is 3.32), soil water content is the primary factor to constrain ET due to the plant regulation of stomatal conductance as part of the water conservation strategy. Additionally, as DI increases across sites, the sign of the contemporaneous causal relationship between VPD and ET can reverse from positiveβindicating that atmospheric demand for water drives ETβto negativeβreflecting that plant stomatal closure limits ET in response to the dryer atmosphere.
ππ‘ππ©πππ« π: As summer heatwaves and droughts are becoming more frequent and intense, such as in Western Europe, there is a growing interest in unraveling the physical mechanisms behind their occurrences and their changes. Soil desiccation is critical for the intensification and propagation of heatwaves, but its relative importance compared to other well-known large-scale atmospheric mechanisms, such as persistent atmospheric blocking systems and horizontal warm advection, remains elusive, especially in the context of a changing climate. In the second study, we utilize machine learning along with intervention experiments to estimate the respective contributions of soil water content πΆ_π π€π and atmospheric circulation πΆ_ππ‘π to daily maximum temperature in Western Europe, with a particular focus on the 2022 summer events. Our results reveal that during the two unparalleled heatwave events that occurred in June and July of 2022, the impact πΆ_π π€π on the heatwave intensity was on average approximately 40% of πΆ_ππ‘π, and was comparable to πΆ_ππ‘π in continental dry-to-wet transition regions. Reviewing heatwaves in recent three decades, the percentage of heatwave areas that are significantly influenced by soil moisture-air temperature coupling has increased by 11.4% per decade. Additionally, for regions that have experienced heatwaves in at least 5 out of the past 33 years, about 21.7% areas, mostly in the transition zones, witness a significant increase in πΆ_π π€π; while only 2.5% exhibit a substantial increase in πΆ_ππ‘π. Furthermore, we find within the transitional climates, the intensification of heat extremes is mainly resulted from soil moisture depletion rather than atmospheric anomalies; while in (dry) Spain and the (wet) northern areas of central Europe, it is the variations in atmospheric circulation and soil desiccation that jointly fuel the persistent heatwaves. Our study emphasizes the observation-based large and increasing importance of soil moisture coupling in intensifying summer heatwaves and provides insights into future climates in extra-tropical regions like Western Europe, where a warmer and drier future is projected.
ππ‘ππ©πππ« π: Earth system models (ESMs) and climate simulations are extensively employed to study the dynamics of climate and project long-term changes in the climate system. Despite their widespread use, large uncertainties persist among these models regarding the estimation of the continental gross primary productivity (GPP) and land carbon sink, which compromise the reliability of projections concerning future atmospheric carbon dioxide (πΆπβ) concentrations and the assessment of how terrestrial ecosystems respond to and might mitigate some of global warming. In ESMs, convection and clouds are one major source of such uncertaintiesβthey are not only the most uncertain factors in the modeling of ``physical'' climate and also significantly affect the land carbon cycle through complex interactions involving radiation, moisture, and thermal pathways. In the third study, to isolate the role of clouds on the terrestrial carbon cycle, two modelsβthe Community Earth System Model (CESM) and its super-parameterized counterpart (SPCESM, abbreviated as SP), which only differ in their representation of convection and clouds, are analyzed under present-day climatology to assess the impact of cloud representations on GPP. Compared with CESM, SP shows a 12.8% decrease in total cloud fraction within the 60Β°π βΌ 60Β°π range, which results in a notable GPP decline of 5.6 πππΆ π¦πβ»ΒΉ. This divergence, equivalent to 4.4% of terrestrial GPP in CESM, is comparable to the inter-annual variability in GPP and the uncertainty of GPP observed across climate models with diverse representations, extending beyond just cloud-related processes.
Further analysis decomposes the GPP divergence between CESM and SP into two additive components and demonstrates that three-quarters of the difference is attributed to the negative impact from reduced cloud cover on light use efficiency (LUE) from CESM to SP, while the remaining one quarter is due to the positive impact from enhanced photosynthetically active radiation (PAR). An explainable machine learning model equipped with SHAP values further identifies two primary mechanisms underlying the lower LUE estimation in SP. Firstly, diminished clouds lead to higher air temperatures and reduced precipitation, creating a drier environment that prompts plants to regulate stomatal conductance to minimize water loss through transpiration, thereby suppressing the exchange rate of πΆπβ between biosphere and atmosphere. Secondly, the reduction in diffused radiation restricts the photosynthesis of shaded leaves. Combined, these two mechanisms reduce plant LUE, outweigh the beneficial impacts of increased PAR on photosynthesis, and ultimately lead to the declined terrestrial biosphere productivity in SP. Overall, we identify the representation of clouds as a key process for the terrestrial carbon cycle.
| Identifer | oai:union.ndltd.org:columbia.edu/oai:academiccommons.columbia.edu:10.7916/y27n-wa29 |
| Date | January 2024 |
| Creators | Huang, Yu |
| Source Sets | Columbia University |
| Language | English |
| Detected Language | English |
| Type | Theses |
Page generated in 0.0031 seconds