Review Stage 2: Synthesis and Assessment Country report

TIME SERIES

Party: RS

Submission year: 2011

The aim of this test is to identify instances of dips, jumps, and sudden trends in time series data reported by countries.  Only data in NFR02 and NFR09 reporting format were analyzed, and data for which at least three years were reported. We converted all years where only NFR02 sectors are available to NFR09 and show only the converted sectors in the time series analysis.

The table below shows data, where outliers in the time series were identified based on the following methodology:
Reported time series data were log 10-transformed prior to analysis to reduce intra-series variability and improve general time series linearity. A linear regression was subsequently applied to the log-transformed values for each time series.

An individual value within the time series was identified as a dip/jump if the respective residual value (regression forecast value - reported value) was greater than 2 standard deviations from the mean of all residuals within the time series. Only time series where the flagged data value contributed a significant fraction (>3%) of the national total for the given year are included in this dataset for expert review.

Time series where sigma grow largely (i.e. larger than 20% of the time series mean value) because of extreme variation in the emission data are also flagged. Zero values indicate small numbers rounded to zero.

 

The test was performed for the following time series (if data was available):

Main pollutants:

1990 – 2009

HMs and POPs:

1990 – 2009

PMs:

2000 – 2009

 

Color Key

 

 Indicates a dip in the time series data

 

 Indicates a jump in the time series data

 

 Indicates time series data with large sigma


MAIN (Unit = Mg)
  19901991199219931994199519961997199819992000200120022003200420052006200720082009
NOx1A2fi               2410239102340