Update of results from the Cooperative Regional Model Evaluation (CReME) project

Results obtained from the Cooperative Regional Model Evaluation (CReME) project are summarized. First the authors recap briefly results from Phase 1, in which various photochemical models were evaluated on the Lake Michigan Ozone Study (LMOS) domain. Then new results are presented for New England, New York, and the combined Philadelphia-New York-New England UAM domains. These new results were obtained by using UAM-IV with various combinations of input files in order to test the sensitivity of model results to differing methods of providing boundary conditions, emissions, and meteorological conditions. The authors also plan to apply the SARMAP Air Quality Model (SAQM) to the LMOS domain using CReME inputs. Several conclusions were drawn from the model applications on the northeast and midwest domains: (1) UAM-IV output was very sensitive to both boundary conditions and mixing heights, (2) ROM-derived boundary conditions were generally inferior to those based on UAM-V output or observations; (3) a combined domain run gave improved simulations in the upwind regions of each domain relative to the single domain runs but did not show a general improvement over the entire domain; (4) there was little difference between UAM-IV and UAM-V performance; (5) the SAIMM prognostic meteorological model was difficult tomore » merge satisfactorily with UAM-IV; and (6) sensitivity runs with 50% emissions reductions in manmade NO{sub x} or VOC showed that, in general, VOC reductions always lead to ozone concentration reductions and NO{sub x} reductions usually lead to increases in ozone concentration close to large point sources or urban areas. However, the extent and amount of NO{sub x} disbenefits varied from day to day and with domain, and sometimes disbenefits were absent. NO{sub x} reductions always reduced ozone farther downwind.« less