“Backward” Lagrangian stochastic models calculate an ensemble of fluid element (particle) trajectories that are distinguished by each passing through an observation point. As shown, they can be faster and more flexible in calculating short-range turbulent dispersion from surface area sources than “forward” models, which simulate trajectories emanating from a source. Using a backward model, one may catalog a set of “touchdown” points (where trajectories reflect off the ground) and vertical touchdown velocities w0 of particles “on their way to” a sensor location. It is then trivial to deduce the average concentration resulting from a surface source using the touchdown catalog: by summing the reciprocal of w0 for touchdowns occurring within the source boundary. An advantage of this methodology is that while forward model trajectories are linked to a specific source, backward trajectories have no such dependence. In horizontally homogeneous flow, a “library” of touchdown catalogs (for representative ... Abstract “Backward” Lagrangian stochastic models calculate an ensemble of fluid element (particle) trajectories that are distinguished by each passing through an observation point. As shown, they can be faster and more flexible in calculating short-range turbulent dispersion from surface area sources than “forward” models, which simulate trajectories emanating from a source. Using a backward model, one may catalog a set of “touchdown” points (where trajectories reflect off the ground) and vertical touchdown velocities w0 of particles “on their way to” a sensor location. It is then trivial to deduce the average concentration resulting from a surface source using the touchdown catalog: by summing the reciprocal of w0 for touchdowns occurring within the source boundary. An advantage of this methodology is that while forward model trajectories are linked to a specific source, backward trajectories have no such dependence. In horizontally homogeneous flow, a “library” of touchdown catalogs (for representative ...

The gas emission rate Q from an artificial 36-m2 surface area source was inferred from line-average concentration CL measured by an open-path laser situated up to 100 m downwind. Using a backward Lagrangian stochastic (bLS) model, a theoretical CL/Q relationship was established for each experimental trial by simulating an ensemble of fluid-element paths arriving in the laser beam under the prevailing micrometeorological conditions. The diagnosed emission rates (QbLS) were satisfactory for trials done when Monin–Obukhov similarity theory gave a good description of the surface layer, but were poor during periods of extreme atmospheric stability (|L| ≤ 2 m) and transition periods in stratification. With such periods eliminated, the average value of the 15-min ratios QbLS/Q over n = 77 fifteen-minute trials spanning 6 days was 1.02. Individual 15-min estimates, however, exhibited sizable variability about the true rate, with the standard deviation in QbLS/Q being σQ/Q = 0.36. This variability is less... Abstract The gas emission rate Q from an artificial 36-m2 surface area source was inferred from line-average concentration CL measured by an open-path laser situated up to 100 m downwind. Using a backward Lagrangian stochastic (bLS) model, a theoretical CL/Q relationship was established for each experimental trial by simulating an ensemble of fluid-element paths arriving in the laser beam under the prevailing micrometeorological conditions. The diagnosed emission rates (QbLS) were satisfactory for trials done when Monin–Obukhov similarity theory gave a good description of the surface layer, but were poor during periods of extreme atmospheric stability (|L| ≤ 2 m) and transition periods in stratification. With such periods eliminated, the average value of the 15-min ratios QbLS/Q over n = 77 fifteen-minute trials spanning 6 days was 1.02. Individual 15-min estimates, however, exhibited sizable variability about the true rate, with the standard deviation in QbLS/Q being σQ/Q = 0.36. This variability is less...