Gå til innhold


Konferansebidrag og faglig presentasjon

Evaluating a forecast system for long-range atmospheric transport episodes of POPs.

Halse, A.K.; Eckhardt, S.; Schlabach, M.; Stohl, A.; Breivik, K.


Poster presented at SETAC Europe 23rd annual meeting 2013, Glasgow 12.05.13 - 16.05.13.
Fil: NILU PP 08/2013 (pdf)

Sammendrag: Background air measurements of persistent organic pollutants (POPs) within existing monitoring programs are typically conducted by use of active air samplers (AAS), but the high cost of AAS limits their spatial and temporal coverage. Sampling at many such sites furthermore occurs at fixed intervals (e.g. one day per week) without any a priori consideration of air mass transport (i.e., whether the air is likely to be polluted or not). While the current strategy is appropriate for the purpose of assessing long-term trends (years, decades), the fixed interval non-continuous sampling approach is at risk of missing out key long-range atmospheric transport (LRAT) episodes. The objectives of this study have been to (i) develop a forecast system using the Lagrangian transport model FLEXPART to predict long-range atmospheric transport episodes of POPs using PCB-28 as a model compound, (ii) to evaluate the capability of the forecast system to capture specific LRAT events at a background site in southern Norway (Birkenes) through targeted sampling (i.e. when LRAT events are predicted), (iii) to assess whether predicted LRAT events for PCB-28 coincide with elevated concentrations of additional PCBs and other POPs, and (iv) to identify source regions of POPs during individual episodes. The system has been initially evaluated by comparing targeted samples collected over 12 to 25 hours during individual LRAT episodes, with monitoring samples regularly collected over one day per week throughout 2011. The FLEXPART model was clearly successful in identifying LRAT episodes for both PCB-28 and other PCBs. The model fails to accurately reproduce the magnitude of PCB-28 concentrations during individual episodes, but this can be mainly attributed to uncertainties in the absolute emission rates of PCB-28 used to drive simulations. We conclude that forecasting of pollution episodes has the potential to add value to relevant monitoring efforts which are normally collecting active air samples at fixed intervals in a non-continuous manner. Observations targeted at strong pollution episodes (as in this study) or on transport from specific source regions with highly uncertain emissions (as could be done in a very similar forecasting framework) could significantly enhance our understanding of POP sources.