Information Technology Reference
In-Depth Information
Marine Ecosystem Model Calibration through
Enhanced Surrogate-Based Optimization
Malte Prieß 1 , Slawomir Koziel 2 , and Thomas Slawig 1
1 Institute for Computer Science, Cluster The Future Ocean
Christian-Albrechts Universitat zu Kiel, 24098 Kiel, Germany
2 Engineering Optimization & Modeling Center, School of Science and Engineering
Reykjavik University, Menntavegur 1, 101 Reykjavik, Iceland
{ mpr,ts } @informatik.uni-kiel.de, koziel@ru.is
Abstract. Mathematical optimization of models based on simulations usually
requires a substantial number of computationally expensive model evaluations
and it is therefore often impractical. An improved surrogate-based optimization
methodology, which addresses these issues, is developed for the optimization of
a representative of the class of one-dimensional marine ecosystem models. Our
technique is based upon a multiplicative response correction technique to cre-
ate a computationally cheap but yet reasonably accurate surrogate from a tem-
porarily coarser discretized physics-based coarse model. The original version of
this methodology was capable of yielding about 84% computational cost sav-
ings when compared to the fine ecosystem model optimization. Here, we demon-
strate that by employing relatively simple modifications, the surrogate model
accuracy and the efficiency of the optimization process can be further improved.
More specifically, for the considered test case, the optimization cost is reduced
three times, i.e., from about 15% to only 5% of the cost of the direct fine model
optimization.
Keywords: Marine
Ecosystem
Models,
Surrogate-based
Optimization,
Parameter Optimization, Response Correction, Data Assimilation.
1
Introduction
Numerical simulations nowadays play an important role to simulate the earth's climate
system and to forecast its future behavior. The processes to be modeled and simulated
are ranging from fluid mechanics (in atmosphere and oceans) to bio- and biochemical
interactions, e.g., in marine or other type of ecosystems. The underlying models are
typically given as time-dependent partial differential or differential algebraic equations
[7,10,12].
Among them, marine ecosystem models describe photosynthesis and other biogeo-
chemical processes in the marine ecosystem that are important, e.g., to compute and
predict the oceanic uptake of carbon dioxide ( CO 2 ) as part of the global carbon cycle
[17]. They are typically coupled to ocean circulation models. Since many important
processes are non-linear, the numerical effort to simulate the whole or parts of such a
coupled system with a satisfying accuracy and resolution is quite high.
 
Search WWH ::




Custom Search