The use of bootstrapping to estimate conditional probability fields for source locations of airborne pollutants

摘要: Abstract A receptor model has been developed in which meteorological information the form of air parcel back trajectories are combined with on atmospheric constituent concentration data to produce conditional probability fields pointing areas that likely have made significant contributions samples higher than average concentrations. This approach, potential source contribution function (PSCF) analysis, proven quite successful producing maps a good correspondence known high emissions variety spatial scales from large urban scale problems basin includes Los Angeles, CA regional transport pollutants southern Ontario semi-global several sites Arctic. However, there cells having limited numbers endpoints because region low probabilities and is estimate uncertainties PSCF values. Thus, we examined use bootstrapping provide better estimates values their uncertainties. approach tested locations at differing levels geographical for varying selected trials made. The results studies Arctic Ny Alesund Spitsbergen (78°55′ N, 11°57′ E, 5 m above mean sea level) presented. these suggest many cases bootstrapped clearer more easily interpreted terms sources.