Technical Abstract: Measuring gas emissions from treatment lagoons and storage ponds poses challenging conditions for existing micrometeorological techniques because of non-ideal wind conditions. These include those induced by trees and crops surrounding the lagoons, and lagoons with dimensions too small to establish equilibrated microclimate conditions within the water boundary. Using a synthetic floating emission source with known emission rates from an irrigation pond, this study evaluated the accuracy of an emerging backward Lagrangian stochastic (bLS) inverse-dispersion technique to measure lagoon emissions. The measured parameters were wind statistics and path-integrated concentrations (PICs) from multiple locations. Anemometers were located on the upwind, downwind, side berm parallel to wind, or directly above pond water surface. PICs were monitored within the pond and on the downwind berm. Additionally, the berm surface was deliberately roughened during the summer by placing pine straw bales along the berms to simulate vegetation growth. The accuracy of the inverse-dispersion technique was significantly affected by the location of the three-dimensional sonic anemometers. Generally using an anemometer located on the berm produced the more accurate results than using an anemometer located directly above water surface. The total average accuracy of all combinations of anemometer location and PICs for both smooth and rough berm surface conditions was 0.77 ± 0.23 (N = 398). This lagoon study showed an accuracy level not very different from environments that meet the ideal assumptions of the inverse-dispersion model, thus, demonstrating the robustness of the inverse-dispersion technique even in non-ideal settings.