Learning in national inventory reporting: a bivariate approach

The paper addresses the problem of uncertainty assessment in National Inventory Reports (NIRs) on greenhouse gas (GHG) emission, provided by cosignatories to the UNFCCC and its Kyoto Protocol. The data on GHG emission in a given year, along with revisions for past data are reported anually, resulting in the data matrix with rows consisting of the data series, formed by consecutive revisions, and columns, corresponding to emissions for a given year, estimated in consecutive revisions. The idea is to consider each row of that matrix a realization of a discrete-time non-stationary stochastic process, being a sum of two other processes, and to analyze the uncertainty related to both of them. The uncertainty is interpreted in terms of errors and expressed by the mean values of the component processes, as functions of time. The existence and uniqueness of these estimates are investigated, and the iterative estimation algorithm is proposed. Some methods to estimate standard deviations are also discussed and their interpretation in terms of uncertainty is provided. The estimation results are presented for a few selected EU-15 countries.

Bibliografia s. 21-22

23 stron ; 21 cm

Bibliography p. 21-22

23 pages ; 21 cm