PhD (m/f/d) for uncertainty quantification

Biodiversity data spans a wide range both spatially and temporally (from regional to global scales, and from decadal to millennial time scales).
Observational
data can be used to understand past and predict future changes in biodiversity. Unfortunately, the sources of uncertainty for these data are numerous: process noise, observational error, and sampling fluctuations to mention a few. Similar statements also apply to

measured
data. In addition, the choice of model that relates response variables to explanatory variables is itself a source of uncertainty. We will construct probabilistic

models
via conditional probabilities estimated from real data. From these, we can generate artificial data sets that serve as inputs to an analysis pipeline, resulting in reconstructed conditional probabilities. Then, the overall uncertainty of the response (predicted community structure) will be quantified using information-theoretic measures (e.g., Kullback-Leibler divergence). The

analysis
pipeline can also be used to assess how the different sources of uncertainty contribute to the overall uncertainty.

Tasks

• The composition of species assemblages (biodiversity) is expected to reflect environmental variables that constrain establishment and success of species during community assembly. However, the assessment of both response variables (biodiversity) and explanatory variables (environment) are characterized by uncertainty, but also the association between them is itself a source of uncertainty.
• You will therefore assess multiple sources of uncertainty in the association between biodiversity and environmental data, e.g., process noise, observational error, and sampling fluctuations.
• Using probabilistic models based on real marine biodiversity data, we will generate artificial data sets that allow quantifying the overall uncertainty of the response (predicted community structure) using information-theoretic measures.

Requirements

• Master in environmental sciences or related fields (physics, applied mathematics, etc.)
• Strong background in statistics, mathematical modelling and data analysis
• Very good programming skills in R (or alternatively MatLab) are required
• Very good skills in English
• Knowledge of general ecology and ecological theory are a plus

Further Information

For further information please contact
Porf. Dr. Helmut Hillebrand
(helmut.hillebrand@hifmb.de; +49(471)4831-2542).
The position is limited to 3 years. The salary will be paid in accordance with the German Tarifvertrag des öffentlichen Dienstes (TVöD Bund), up to salary level

13 (66%).
The place of employment will be
Oldenburg.

This characterizes us

• our scientific success - excellent research.
• collaboration and cooperation - intra-institute, national and international, interdisciplinary.
• opportunities to develop – on the job, aiming at other positions and beyond AWI.
• a culture of reconciling work and family – audited, and even more than that.
• our outstanding research infrastructure – ships, stations, aircraft, laboratories and more.
• an international environment – everyday contacts with people from all over the world.
• having an influence – fundamental research with social and political relevance
• flat hierarchies – freedom and responsibility.
• exciting topics – also in technology, administration and infrastructure.

Equal opportunities for women and men are an integral part of our personnel policy. Therefore, we encourage women to apply.
As Ph.D. student at AWI you will be member of the Helmholtz Graduate School for Polar and Marine Research ‘POLMAR’ (
) or another graduate school.
Disabled applicants will be given preference when equal qualifications are present. The AWI fosters the compatibility of work and family through various means. Because of our engagement in the area of work-life compatibility we have been awarded the certificate “Career and Family”.

mehr...

zur Vakanz

Weitere Jobs von Alfred-Wegener