Steady state of ecosystem flow networks: a comparison between balancing procedures

When modelling real ecosystems, a number of techniques need steady state condition to proceed in their analysis. In ecosystem network models this means that energy entering the system exactly balances the output. Steady state, however, is not a straightforward outcome of network construction and, to have this condition satisfied, network analysis uses balancing procedure. This operation leads to restructuring the weighted network, changing the values of some network flows; this can affect drastically the results of the analysis. Presently, two algorithms are used for balancing ecosystem networks: input-based approach and output-based approach. In the former input flows are kept constant while outputs and transfer coefficients are manipulated; the latter requires that inputs and intercompartmental flows are modified. This paper discusses the effects of these algorithms on some products of network ana lysis, in particular system level indices such as total system throughput (TST) and ascendency. Also it suggests four new procedures that, while balancing the networks, can minimise changes on measured flows and distortion on results of the analysis. (C) 2003 Elsevier Science B.V. All rights reserved.