uni-potsdam.de

You are using an old browser with security vulnerabilities and can not use the features of this website.

Here you will see how you can easily upgrade your browser.

PhD-Project Q6: Investigation of interrelationships between floods and climate variability using multi-layer complex networks

Timescale: Oct.2018 – Sept.2021

Supervisors:

Prof. Dr. Bruno Merz, GFZ Potsdam,

Prof. Dr. Jürgen Kurths, PIK Potsdam

Dr. Norbert Marwan, PIK Potsdam

Background

Unusually large floods in drainage basins of all sizes can be related to large-scale atmospheric circulation anomalies. However, their impact depends on further factors and preconditions that can even change over time, thus resulting in more or less strong consequences of floods. Such factors can be, e.g., catchment wetness or timing of the snow melt, factors that can be clearly linked to climate change. The project will apply a multi-layer complex network approach to investigate the interrelationship between potential influencing factors, floods, and climate variability.

Objectives and Methods

In this project we aim to identify structural changes in climate and flood-influencing variables (e.g. catchment wetness) that affect the occurrence of extreme floods. A promising approach for studying space-time dynamics of strong natural events are complex networks based on event synchronization. Event synchronization has been combined with a multiscale approach and successfully applied in Q2 (Agarwal et al. 2017). This technique will be extended by a multivariate approach allowing to include more physical parameters, resulting in a multilayer network or a network of networks. A special challenge will be to combine synchronization event based networks with networks obtained from other climate and flood-influencing variables, such as temperature or catchment wetness. This includes the development of appropriate measures of causality to quantify possible interactions between the networks, including recurrence based approaches as developed in Q1. The methods will be adapted for spatio-temporal climate and flood data, with the ultimate goal of better understanding how extreme river floods in Central Europe and in the Mekong River Basin are linked to climate.

Up to 10 river gauges in the Mekong River Basin and a large number of catchments in Europe will be selected, comprising a variety of flood regimes in Europe and a monsoonal region. Flood times series of different magnitude (e.g. peak-over-threshold with on average 3 events per year) and associated meteorological and catchment variables (sea surface temperature, atmospheric pressure, rainfall, snow cover, soil moisture, etc.) will be derived. From the data a multilayer network or a network of networks is constructed where the interactions within each network and, more challenging, between the networks should provide new insights in the underlying mechanisms. Moreover, time-dependent analysis techniques for non-stationary data and statistical methods for testing the significance of transient structural properties in this network of networks will be further developed.