Post-EGU Seminar in Hydrology

Dr. A. Paschalis, Imperial College London, England

Monday 15 Apr 2019 | Room D218 | 14:00 – 15:30  

Terrestrial ecosystems currently sequester annually ~1/3rd of the anthropogenic CO2 emissions on average, partially mitigating the rate of climate change. The fate of the terrestrial carbon sink is thus crucial, yet highly uncertain. Uncertainties arise from our limited understanding on how terrestrial vegetation responds to meteorological forcing and atmospheric CO2 at different time scale. This limits substantially our skill in projecting the fate of the terrestrial sink, subsequently increasing the uncertainties to climate change projections.
In this seminar, an extensive analysis of such uncertainties will be presented. Specifically:

  • How meteorological variability across scales affects the responses of ecosystems from the leaf to the ecosystem scale.
  • How elevated CO2 affects different ecosystems, and what is the importance of properly capturing stomata responses to changes in the concentration of CO2.
  • The key biotic and abiotic contributions and their interactions in shaping the total ecosystem responses at large scales and finally,
  • How different are the responses of ecosystems to changes is weather patterns as predicted by a set of ten widely used terrestrial biosphere models.

This presentation will build on data collected across multiple FLUXNET sites worldwide, nine rainfall manipulation experiments, three free air CO2 enrichment (FACE) experiments, and multiple modelling applications.


Time series modelling in hydroclimatic processes
Dr. S. M. Papalexiou, Global Institute for Water Security, Canada

Monday 15 Apr 2019 | Room D352 | 15:45 – 18:15

So, you've been given a time series, e.g, of hourly precipitation. That's great, but how can you generate as many as you like with exactly the same statistical properties? In this short course you'll find out.

You'll be introduced to a unified method of stochastic modelling and downscaling that makes feasible the generation of time series that preserve any desired marginal probability distribution and correlation structure including features like intermittency.

The workshop includes a rapid introduction in the stochastic properties of hydroclimatic processes like precipitation, flooding, wind, temperature, etc., and highlights features like stationarity, cyclostationarity, marginal distributions, correlations structures and intermittency.

We'll develop and apply on-the-spot and step-by-step: (a) the iconic AR(1) model, (b) higher order AR models as a method to approach arbitrary correlations structures; (c) the parent-Gaussian framework to simulate time series with any marginal distribution and correlation; and (d) intermittent time series modelling (like precipitation) at any time scale.