Algorithmische Optimale Steuerung - CO2-Aufnahme des Meeres

Dr. rer. nat. Anna Heinle

wiss. Mitarbeiterin, Mathematikerin

Exzellenzcluster Future Ocean

zur Person

Thema der Doktorarbeit:

The parameters used in numerical models in climate research are often poorly known. To improve the predictive power of those models, optimization techniques can be applied to the parameters. 
The model I'm currently working on is a vertically resolved nitrogen-based marine ecosystem model, considering nutrients (N), phytoplankton (P), zooplankton (Z) and detritus (D). The parameters include growth rates, mortality rates, sinking rates and others. Due to the high complexity of the parameter set and the model itself, parameter optimization is difficult and computationally expensive. My attempt to perform nevertheless an optimization is to reduce the dimensions of the model by eliminating the vertical dependency and get optimal parameter estimates for the new, less complex model using automatic differentiation. Afterwards, the thus obtained parameter values shall be used to initialize an optimization process of the original model.


seit 06/13: PostDoc im Future Ocean-Projekt Surrogate-based Optimization of Marine Ecosystem Models, Institut für Informatik, CAU Kiel

03/13: Promotion zum Dr. rer. nat.

07/09--03/13: Doktorandin, Institut für Informatik, CAU Kiel, Junior Research Group A3 (Future Ocean): Algorithmische Optimale Steuerung - CO2-Aufnahme des Ozeans, Gruppenleiter: Prof. Dr. Thomas Slawig

09/08–04/09: Diplomarbeit - in Kooperation mit IFM-GEOMAR „Automatic cloud classification of whole sky images“

08/07: Summer school COPS, Baden-Baden

10/03–04/09: Mathematikstudium, CAU Kiel, Nebenfach: Meteorologie


Heinle, A.,Slawig, T.: Theoretical analysis and optimization of nonlinear ODE systems for marine ecosystem models, System Modelling and Optimization, 501-510 (2013) 

Heinle, A., Slawig, T.: Internal dynamics of NPZD type ecosystem models, in: Ecological Modelling, 254, 33-42 (2013) 

Heinle A., Macke A. , Srivastav A.: Automatic cloud classification of whole sky images. Atmospheric Measurement Techniques 3, 557–567 (2010)