Filters

2018 - 20204

More filters

2021 - 20234
Reset filters

Settings

Author: Kretschmer, Marlene × Has files: Yes ×

Search Results

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Tropical and mid-latitude teleconnections interacting with the Indian summer monsoon rainfall: a theory-guided causal effect network approach
    (Göttingen : Copernicus Publ., 2020) Di Capua, Giorgia; Kretschmer, Marlene; Donner, Reik V.; van den Hurk, Bart; Vellore, Ramesh; Krishnan, Raghavan; Coumou, Dim
    The alternation of active and break phases in Indian summer monsoon (ISM) rainfall at intraseasonal timescales characterizes each ISM season. Both tropical and mid-latitude drivers influence this intraseasonal ISM variability. The circumglobal teleconnection observed in boreal summer drives intraseasonal variability across the mid-latitudes, and a two-way interaction between the ISM and the circumglobal teleconnection pattern has been hypothesized. We use causal discovery algorithms to test the ISM circumglobal teleconnection hypothesis in a causal framework. A robust causal link from the circumglobal teleconnection pattern and the North Atlantic region to ISM rainfall is identified, and we estimate the normalized causal effect (CE) of this link to be about 0.2 (a 1 standard deviation shift in the circumglobal teleconnection causes a 0.2 standard deviation shift in the ISM rainfall 1 week later). The ISM rainfall feeds back on the circumglobal teleconnection pattern, however weakly. Moreover, we identify a negative feedback between strong updraft located over India and the Bay of Bengal and the ISM rainfall acting at a biweekly timescale, with enhanced ISM rainfall following strong updraft by 1 week. This mechanism is possibly related to the boreal summer intraseasonal oscillation. The updraft has the strongest CE of 0.5, while the Madden–Julian oscillation variability has a CE of 0.2–0.3. Our results show that most of the ISM variability on weekly timescales comes from these tropical drivers, though the mid-latitude teleconnection also exerts a substantial influence. Identifying these local and remote drivers paves the way for improved subseasonal forecasts.
  • Thumbnail Image
    Item
    Inferring causation from time series in Earth system sciences
    ([London] : Nature Publishing Group UK, 2019) Runge, Jakob; Bathiany, Sebastian; Bollt, Erik; Camps-Valls, Gustau; Coumou, Dim; Deyle, Ethan; Glymour, Clark; Kretschmer, Marlene; Mahecha, Miguel D.; Muñoz-Marí, Jordi; van Nes, Egbert H.; Peters, Jonas; Quax, Rick; Reichstein, Markus; Scheffer, Marten; Schölkopf, Bernhard; Spirtes, Peter; Sugihara, George; Sun, Jie; Zhang, Kun; Zscheischler, Jakob
    The heart of the scientific enterprise is a rational effort to understand the causes behind the phenomena we observe. In large-scale complex dynamical systems such as the Earth system, real experiments are rarely feasible. However, a rapidly increasing amount of observational and simulated data opens up the use of novel data-driven causal methods beyond the commonly adopted correlation techniques. Here, we give an overview of causal inference frameworks and identify promising generic application cases common in Earth system sciences and beyond. We discuss challenges and initiate the benchmark platform causeme.net to close the gap between method users and developers. © 2019, The Author(s).
  • Thumbnail Image
    Item
    The different stratospheric influence on cold-extremes in Eurasia and North America
    (London : Springer Nature, 2018) Kretschmer, Marlene; Cohen, Judah; Matthias, Vivien; Runge, Jakob; Coumou, Dim
    The stratospheric polar vortex can influence the tropospheric circulation and thereby winter weather in the mid-latitudes. Weak vortex states, often associated with sudden stratospheric warmings (SSW), have been shown to increase the risk of cold-spells especially over Eurasia, but its role for North American winters is less clear. Using cluster analysis, we show that there are two dominant patterns of increased polar cap heights in the lower stratosphere. Both patterns represent a weak polar vortex but they are associated with different wave mechanisms and different regional tropospheric impacts. The first pattern is zonally symmetric and associated with absorbed upward-propagating wave activity, leading to a negative phase of the North Atlantic Oscillation (NAO) and cold-air outbreaks over northern Eurasia. This coupling mechanism is well-documented in the literature and is consistent with the downward migration of the northern annular mode (NAM). The second pattern is zonally asymmetric and linked to downward reflected planetary waves over Canada followed by a negative phase of the Western Pacific Oscillation (WPO) and cold-spells in Central Canada and the Great Lakes region. Causal effect network (CEN) analyses confirm the atmospheric pathways associated with this asymmetric pattern. Moreover, our findings suggest the reflective mechanism to be sensitive to the exact region of upward wave-activity fluxes and to be state-dependent on the strength of the vortex. Identifying the causal pathways that operate on weekly to monthly timescales can pave the way for improved sub-seasonal to seasonal forecasting of cold spells in the mid-latitudes.
  • Thumbnail Image
    Item
    Detecting and quantifying causal associations in large nonlinear time series datasets
    (Washington, DC [u.a.] : Assoc., 2019) Runge, Jakob; Nowack, Peer; Kretschmer, Marlene; Flaxman, Seth; Sejdinovic, Dino
    Identifying causal relationships and quantifying their strength from observational time series data are key problems in disciplines dealing with complex dynamical systems such as the Earth system or the human body. Data-driven causal inference in such systems is challenging since datasets are often high dimensional and nonlinear with limited sample sizes. Here, we introduce a novel method that flexibly combines linear or nonlinear conditional independence tests with a causal discovery algorithm to estimate causal networks from large-scale time series datasets. We validate the method on time series of well-understood physical mechanisms in the climate system and the human heart and using large-scale synthetic datasets mimicking the typical properties of real-world data. The experiments demonstrate that our method outperforms state-of-the-art techniques in detection power, which opens up entirely new possibilities to discover and quantify causal networks from time series across a range of research fields. Copyright © 2019 The Authors,