Model coupling and atmospheric dispersion of malodorous compounds over coastal regions

BERRI, GUILLERMO JORGE; ORCELLET EMILIANA ELISABET; AGUIRRE CESAR A; GABRIELA V. MÜLLER

Salt Lake City

Simposio; 22nd Symposium on Boundary Layers and Turbulence.; 2016

American Meteorological Society

We conduct an atmospheric dispersion study of malodorouscompounds emitted from a pulp mill plant over coastal regions of the UruguayRiver in South America. A total of 11 events are simulated in which theauthorities received complaints from the local population about the presence ofmalodorous substances in the area, with indication of date, hour and locationof the event. The simulations are performed with an atmospheric dispersionmodel coupled to a boundary layer forecast model, and the modeled plumes andwind forecasts are validated with local information. The transport anddiffusion of malodorous compounds is simulated with the Advanced RegionalPrediction System (ARPS, University of Oklahoma), coupled to the operative BoundaryLayer Forecast Model (BLM, National Meteorological Service of Argentina). ARPS performsthe transport and diffusion calculations and BLM provides the 1-km horizontalresolution 3D meteorological fields required by ARPS as initial and boundaryconditions. BLM has been specifically developed for modeling low-levelatmospheric circulations over coastal regions where the daily cycle ofwater-land thermal contrasts at the surface is the major driving mechanism forthe atmosphere. The validation of BLM winds with the observations from a 42 m meteorologicaltower indicates reasonably accurate wind forecasts. The spatial layout ofmodeled plumes is compared with the geographical distribution of points in thearea where authorities received the complaints of the local population aboutthe presence of malodor. Nine of the 11 studied events are considered successfulmodeling cases since a positive (negative) in situ verification of malodorpresence matches with a plume position over (far from) the site. In one of thetwo unsuccessful modeling cases, although the plume is marginally distant fromthe site, the average wind direction error is the largest one of all theevents. In the other case the modeled plume is in fact over the site, but thesituation was negatively verified. The reason for the disagreement could be thewind direction changes during the event. This was the longest modeled casethat  lasted for 7 hours and the plumewas meandering during that time; first from SSW to the S, then back the SSW,and finally to the S and SSE. The conclusion of the study is that, despite theinherent uncertainty of numerical simulations, the implemented modeling systemproves to be a useful tool not only for diagnostic studies but also forpreventing conflictive situations since it can produce reasonably accurateforecast of plume position and its potential impact.