You are here: Home PIK Members Goswami Bedartha

Dr. Bedartha Goswami :: Research Profile

Scientist at Transdisciplinary Concepts and Methods: RD IV, PIK Potsdam
Picture

Research Domain IV

Transdisciplinary Concepts and Methods 

Contact:

Room 3.51, Building A56
14473 Potsdam


+49 331 288 2467
email mage courier 10pt

         

Research Interests

Uncertainties in paleoclimate data

Fundamentally, all of paleoclimate research involves the following paradigm: the measurement of paleoclimate proxies along the depth of a paleoclimate archive, and the measurement of the age of the depth levels of the archive with techniques such as radiocarbon dating, U/Th dating, layer counting, etc. The goal is then to combine these two sets of measurements to obtain a proxy versus time data set called the proxy record. Typically, the chronological uncertainty involved in the dating is not appropriately carried forward to the proxy record in a quantitative sense.

I focus on estimating the proxy records using a Bayesian approach that can provide posterior proxy estimates such as the mean or the median along with estimates of the uncertainty such as the variance or inter-quantile ranges. I am also involved in developing a toolbox that seeks to provide proxy records with uncertainty estimates using Mote Carlo procedures.

 

Recurrences of dynamical systems

Recurrences are defined as the (approximate) return of the trajectory of a dynamical system to an earlier state. Patterns of recurrences contain information about the fundamental dynamical nature of the system itself. Over the past decades, recurrences have been shown to a be a powerful time series analysis tool that can help quantify the dynamics, behavior and even similarities of complex systems — systems that are otherwise extremely intractable.

I attempt to extend existing recurrence-based methods for the quantification of synchronization between pairs of time series to apply them to complex datasets and estimate the similarities between them. 

 

Surrogate-based significance testing of measures estimated from time series

Normally, time series obtained from real-world and natural systems are the sole realization of the dynamical black boxes that give rise to them. Thus, when we estimate measures such as cross-correlation, variance, power spectra, etc. from such time series, we have no way of comparing the obtained values of these measures to a benchmark test case in order to be able to interpret them properly. For instance, if we get a correlation value of 0.61 between two time series obtained from, let's say, a stalagmite in China and another stalagmite in India, we have no way of knowing whether this value of 0.61 could actually have been caused by a random chance between the two time series given their inherent dynamical natures. Neither can we say anything about whether 0.61 should be interpreted as a case of a strong or moderate connectivity because of the lack of a baseline to compare it to. This is why the measured values have to placed against a background distribution of values obtained using surrogate data sets — data sets generated by randomization of the observed time series.

I seek to find a proper formal framework of describing the statistical significance of measures that have been estimated from time series. This also involves an in-depth understanding on the various types of surrogate creation and their corresponding null hypotheses.

 

Publications

  • D. Traxl, N. Boers, A. Rheinwalt, B. Goswami, J. Kurths, The size distribution of spatiotemporal extreme rainfall clusters around the globe, Geophysical Research Letters 43 (2016) doi:10.1002/2016GL070692

  • B. Goswami, S. M. Shekatkar, A. Rheinwalt, G. Ambika, J. Kurths, A random interacting network model for complex networks, Scientific Reports 5 (2015) 18183 doi:10.1038/srep18183
  • S. Dey, B. Goswami, A. Joshi, A possible mechanism for the attainment of out-of-phase periodic dynamics in two chaotic subpopulations coupled at low dispersal rate, Journal of Theoretical Biology 367 (2015) 100–110 doi:10.1016/j.jtbi.2014.11.028
  • B. Goswami, J. Heitzig, K. Rehfeld, N. Marwan, A. Ambili, S. Prasad, J. Kurths, Estimation of sedimentary proxy records together with associated uncertainty,  Nonlinear Processes in Geophysics 21 (2014) 10931111doi:10.5194/npg-21-1093-2014
  • S. Dey, B. Goswami, A. Joshi, Effects of symmetric and asymmetric dispersal on the dynamics of heterogeneous metapopulations: Two-patch systems revisited, Journal of Theoretical Biology 345 (2014) 52–60 doi:10.1016/j.jtbi.2013.12.005
  • B. Goswami, N. Marwan, G. Feulner, J. Kurths, How do global temperature drivers influence each other? A network perspective using recurrences, European Physical Journal Special Topics 222 (2013) 861–873 doi:10.1140/epjst/e2013-01889-8
    • S. F. M. Breitenbach, K. Rehfeld, B. Goswami, J. U. L. Baldini, H. E. Ridley, D. J. Kennett, K. M. Prufer, V. V. Aquino, Y. Asmerom, V. J. Polyak, H. Cheng, J. Kurths, N. Marwan, COnstructing Proxy Records from Age models (COPRA), Climate of the Past 8 (2012) 1765–1779 doi:10.5194/cp-8-1765-2012
      • B. Goswami, G. Ambika, N. Marwan, J. Kurths, On interrelations of recurrences and connectivity trends between stock indices, Physica A 391 (2012) 4364–4376 doi:10.1016/j.physa.2012.04.018
       
       

      Book Chapter

      • A. Rheinwalt, B. Goswami, N. Boers, J. Heitzig, N. Marwan, R. Krishnan and J. Kurths, Teleconnections in Climate Networks: A Network of Networks Approach to Investigate the Influence of Sea Surface Temperature Variability on Monsoon Systems, in: Machine Learning and Data Mining Approaches to Climate Science (2015) Springer International Publishing. pp 23–33
       
       
       copra graphic minimal ultra small    titter icon ultra small