Abstracts – Session C1
Impact of extremes on land-atmosphere biogeochemical cycling
C101 - ORAL-0400: The impacts of the 2015/2016 El Nino on the carbon cycle of tropical forests: insights from a global forest monitiring network
1University of Oxford, Oxford, United Kingdom
The response of the terrestrial tropics to climate extreme events is known to dominate the interannual variability of the global carbon cycle, and is a major factor in determining the carbon sink capacity and biodiversity of the biosphere under climate warming. Atmospheric observations tell us that during El Niño events the terrestrial tropics become a net source of carbon to the atmosphere, resulting in an anomalous rise in global CO2 concentrations. However, which processes drive this carbon source (e.g. fire, reduced productivity and carbon uptake, increased plant or soil respiration, tree death) remains unclear. Understanding this carbon source may provide vital clues as to how stable the tropical biosphere may be under future climate change.
Terrestrial biosphere models try to capture the key processes in the biosphere carbon cycle under current and future climates, and have been increasing in sophistication in the past decade. However, they remain untested against direct observations of the tropical biosphere under an extreme climate event. The El Niño of 2015/2016 was one of the strongest tropical climate events for over a century, and tipped the atmosphere carbon dioxide concentration permanently above the symbolic 400 pm threshold for the indefinite future. Moreover, it was the first strong El Niño to be observed in detail by terrestrial observation networks,. Notably, we have deceloped a Global Ecosystems Monitoring (GEM) network which has tracked in detailed the monthly variation in the carbon cycle in over 50 forest plots across the tropics, over several years before, during and after the El Niño. The regions monitored span Amazonia, Africa and Borneo. These observations enable direct observation of how productivity, respiration and net carbon budget have responded to El Nino extremes, and what comment elements there are across tropical regions. We present initial analysis of results from contrasting sites across the tropics.
C102 - ORAL-0185: Dependence of drivers affects risks associated with compound events
Jakob Zscheischler, Sonia Seneviratne
Compound climate extremes are receiving increasing attention because of their disproportionate impacts on humans and ecosystems. Risks assessments, however, generally focus on univariate statistics. We analyze the co-occurrence of hot and dry summers and show that these are correlated, inducing a much higher frequency of concurrent hot and dry summers than what would be assumed from the independent combination of the univariate statistics. We further show that in a strong greenhouse-gas forcing scenario, the frequency of concurrent extremely hot and dry summers increases in many regions of the world in addition to long-term trends in temperature and precipitation. Extremely hot and dry conditions heavily affect terrestrial carbon fluxes, and analyzing the frequency of such events can help to better understand carbon dynamics. In summary, our results demonstrate how the dependence structure between variables affects the occurrence frequency of multivariate extremes. Assessments based on univariate statistics can thus strongly underestimate risks associated with given extremes, if impacts depend on multiple (dependent) variables. We conclude that a multivariate perspective is necessary in order to appropriately assess changes in climate extremes and their impacts.
C103 - ORAL-0124: Evaluating CMIP5 model agreement for multiple drought metrics
Anna Ukkola, Andy Pitman, Martin de Kauwe, Gab Abramowitz, Nadja Herger, Jason Evans, Mark Decker
Global climate models play an important role in quantifying past and projecting future changes in drought. Previous studies have pointed to shortcomings in these models for simulating droughts but systematic evaluation of their level of agreement has been limited. Here, historical simulations (1950-2004) for 20 models from the latest Coupled Model Intercomparison project (CMIP5) were analysed for a variety of drought metrics and thresholds. Model agreement was investigated for different types of drought (precipitation, runoff and soil moisture) and how this varied with drought severity and duration. The models were shown to agree well on rainfall drought metrics at the global scale but systematically underestimated rainfall drought intensity compared to observations. Conversely, model estimates for runoff and soil moisture droughts varied significantly, particularly for intensity. Differences in precipitation simulations were found to explain discrepancies in runoff and soil moisture drought metrics only over some regions, predominantly with respect to drought intensity, pointing to a limited control by rainfall as the source of model differences in runoff and soil moisture droughts. This study shows large, but metric-dependent discrepancies in CMIP5 for modelling different types of droughts. It points to a need for further model development and evaluation to better understand the capability of CMIP5 models in capturing droughts and to increase confidence in future projections of drought.
C104 - ORAL-0246: Soil-moisture drought affects runoff more than evapotranspiration
Rene Orth1, Georgia Destouni1
1Stockholm University, Stockholm, Sweden
Drought comprehensively affects different interlinked aspects of the terrestrial water cycle, which have so far been mostly investigated separately and without direct comparison. With runoff and evapotranspiration anomalies impacting societal sectors and ecosystems in different ways, a consistent understanding of their relative roles is vital for drought understanding, prevention and response. By use of comprehensive multi-decadal data from >400 near-natural catchments along a steep climate gradient across Europe we here analyze these roles. We show that, across different climates, the precipitation deficits that induce soil moisture droughts are mainly counteracted by decreased runoff. Only in dry climate does evapotranspiration have a relevant, yet still relatively small role in drought evolution. In severe droughts, the role of runoff is even more dominant. These results highlight a key drought-regulation role of runoff, calling for relevant understanding and modelling account of local-to-global freshwater changes in the Earth system and in water resource management.
C105 - ORAL-0340: Attribution of a changing evaporation regime: Identifying three distinct responses within the Horn of Africa bimodal rainfall region
Toby Marthews, Friederike Otto, Daniel Mitchell, Simon Dadson, Richard Jones, Myles Allen
A severe drought hit the Greater Horn of Africa (GHoA) in 2014. The long rains (March-May) in this region have been in long-term decline and in 2014 they failed completely, causing drought in some areas of Ethiopia and Kenya, however it remains unclear whether this was attributable to anthropogenic climate change. Precipitation patterns are known to be changing across the GHoA, but trajectories in land surface variables such as evaporation and runoff and variations in the land-atmospheric coupling strength are less well-known and do not always correlate with changes in precipitation. Here we demonstrate a new approach to assessing these effects by applying novel probabilistic event attribution techniques to components of the land surface energy balance and water balance, combining publicly-volunteered distributed computing (through the weather@home project) with land surface simulations using the JULES land environment simulator and perspective from the Max Planck Institute's Earth System Data Cube. We show that in the bimodal rainfall zone of the GHoA during 2014, identifiable areas of the region followed state space trajectories of drought risk that differed from the regional mean, e.g. the Lake Victoria region moved towards receiving less precipitation, but the drought-affected Kenya-Ethiopia-Somalia cross-border area simultaneously became somewhat less dry. Our results are in line with expected land surface responses across the complex GHoA landscape and climate and provide evidence for these in a data-poor region where climate models have previously returned conflicting predictions of both increasing and decreasing drought risk.
C106 - ORAL-0224: Towards constraining extreme temperature projections of the CMIP5 ensemble for Central Europe
Martha Marie Vogel1, Sonia Seneviratne1
1Institute for Atmospheric and Climate Science, Eidgenössische Technische Hochschule (ETH) Zürich, Zürich, Switzerland
Under rising atmospheric CO2 emissions temperature extremes are projected to increase in intensity and frequency. However, large uncertainties exist for future changes of temperature extremes in regions like Central Europe. This is likely associated with uncertainties land physical processes as they considerably contribute to a regional amplification of extreme temperatures. Recent progress was made by constraining extreme temperature projections but uncertainties remain large. In this study we link projections of changes in extreme temperatures to changes in radiation and land-atmosphere interactions in Central Europe.
We focus on changes in latent heat flux as this is a variable which combines the effect of radiation and soil moisture changes and for which observations are available. We find that an increase in net radiation raises the available energy and can therefore enhance both (sensible and latent heat) surface fluxes particularly in the end of the 20th century. The soil moisture-temperature feedback also contributes to the development of extreme temperatures due to the partitioning of surface fluxes. A drying of soil moisture is associated with a decrease in latent heat flux and increase the sensible heat flux and finally an increase of temperatures, which can then again influence soil moisture. We find that the latter process is becoming more important in Central Europe after the first decades of the 21st century. However, models and observations show large differences in present summer trends of latent heat flux. A large part of cmip5 models tend to show an increase in latent heat flux whereas there is no trend in the observations. To reduce the uncertainties of hot extremes we hence employ observational evapotranspiration data sets to potentially constrain the CMIP5 model ensemble in the current climate.
C107 - ORAL-0047: Reduced terrestrial gross primary productivity in North America associated with anomalous Arctic warming
Jin-Soo Kim1, Jong-Seong Kug1, Su-Jong Jeong2
1Pohang University of Science and Technology (POSTECH), Pohang, The Republic Of Korea 2Southern University of Science and Technology (SUSTech), Shenzhen, China
Because of cold-season warming, the productivity of terrestrial vegetation over the Northern Hemisphere has increased. However, despite the ongoing climatic warming, North America has recently experienced more frequent and stronger cold weather events during winters and springs throughout the continent, which may affect terrestrial processes. Herein, we demonstrate the significant impacts of Arctic temperature variation on terrestrial processes in North America using different types of observations and process-based models. It is shown that springtime anomalous Arctic warming leads to negative anomalies in gross primary productivity over the majority of North America by approximately 0.31 PgC yr−1. This negative terrestrial response is mainly explained by two climatic factors: severe cold conditions in the northern North America and lower precipitation in the South Central United States, which resulted from a remote teleconnection in the downstream regions of the anomalous Arctic warming. Decreases in springtime terrestrial gross primary productivity, resulting from anomalous Arctic warming can be sustained until following season because of the aftereffects of plant physiological damage. In addition, United States crop-yield data reveal that annual yields of corn, soybeans, and wheat for Arctic warming years declined by approximately 1.74, 3.96, and 3.62%, respectively. This suggests that the anomalous Arctic warming remotely influences not only natural terrestrial productivity but also on cropland productivity over North America