Background on Quantitative Network Design

The task of quantitative network design is the evaluation of a set of (candidate) networks in terms of how well they constrain a given target quantity. In our context, an example of a target quantity is the uptake of carbon by the terrestrial biosphere over Europe. Quantitative network design is based on an assimilation or inverse modelling system that fulfils two requirements. First, it must be capable of assimilating the observations provided by the candidate networks, together with estimated uncertainties. Second, it must be capable of simulating the target quantity, together with an estimated uncertainty. The candidate networks can then be ranked according their respective uncertainty in the target quantity. Details on the methodological background are provided in an overview paper by Kaminski and Rayner, who also describe applications to the carbon cycle.

The first application of quantitative network design to the carbon cycle was built around an inverse model of the atmopsheric transport by Rayner et al. (1996) and was restriced to flask sampling networks. This project integrates a number of additional data types, in particular continuous atmospheric samples and local eddy-flux measurements. It uses an assimilation system built around the terrestrial biosphere model BETHY, which is called the Carbon Cycle Data Assimilation System (CCDAS).