|A definition of the consistency of the carbon budget of an ecosystem, and its application to the Oosterschelde estuary, S.W. Netherlands|
Klepper, O.; van de Kamer, J.P.G. (1988). A definition of the consistency of the carbon budget of an ecosystem, and its application to the Oosterschelde estuary, S.W. Netherlands. Ecol. Model. 42: 217-232
In: Ecological Modelling. Elsevier: Amsterdam; Lausanne; New York; Oxford; Shannon; Tokyo. ISSN 0304-3800, more
|Authors|| || Top |
- Klepper, O.
- van de Kamer, J.P.G.
Previous studies have produced estimates of the carbon flows that occurred in the Oosterschelde estuary before a storm-surge barrier was constructed in its mouth. This paper describes research done to ascertain whether these estimates are consistent, i.e. whether balanced carbon budgets can be found for all the ecological groups investigated when each carbon flow is within the range of value obtained from laboratory or field data.In general, published carbon budgets either ignore errors or only adjust certain flows in an ad hoc fashion to obtain balanced budgets. This paper presents a well-defined method of obtaining a balanced budget within the pre-specified ranges for the flows. Starting from flow estimates with uncertainty ranges, an optimal balanced budget that minimizes the maximum (weighted) deviation between estimated and balanced flows is sought. It is shown that linear programming is required to achieve this. The set of flows and uncertainty estimates is defined as being consistent if the maximal deviation of the optimal solution lies within the uncertainty range of that estimate.The results obtained by following the above-mentioned procedure on a set of data collected in the Oosterschelde show that the maximum deviation of a carbon flow from its estimated value is 75% of its uncertainty range. The requirement that the carbon flows form balanced budgets and also fall within their uncertainty range results in a marked reduction in the initial uncertainty range of primary production and detritus mineralization. Two possible extensions of the method are discussed that could make it applicable to data sets that are more detailed in space and/or time.