SCERP Project Number: AQ94-2.1
Principal Investigator: Neil S. Berman
Arizona State University
Goal: The major objective is to develop procedures using synoptic classification, numerical modeling and physical modeling to identify sources of air toxics and small particulate matter in complex terrain and to determine exposures of populations to these air pollutants. A goal is to apply the method to "Ambos" Nogales on the border between Sonora, Mexico and Arizona, USA. The fundamental physics of the movement of pollutants during periods of highest concentrations in complex terrain is poorly understood and a goal of this project is to improve the understanding of this problem.
Rationale: One of the problems identified in the U.S.-Mexico border plan is the high concentration in the ambient air of small particulate matter (PM-10) in the sister city area of Nogales, Arizona/Nogales, Sonora ("Ambos" or both Nogales). Although the U.S. population adjacent to the border is small (20,000 in 1990), the population of Nogales, Sonora could be as high as 300,000. Automobile and truck traffic from central Mexico to the U.S. through Nogales is appreciable. Free trade agreements between the U.S. and Mexico will increase the population, the truck traffic, and the manufacturing plants on both sides of the border. There is a need to understand the present situation regarding air quality as well as a need to develop a means of predicting the future air quality.
Annual average concentrations of PM-10 measured in Nogales, Arizona have exceeded both 24 hour and yearly average standards and are similar to the concentrations measured in Phoenix. Long term historical data for the area are limited to the total suspended particulate (TSP) measurements and the PM-10 measurements in Arizona near the border crossing. Samples collected during a study by the Arizona Department of Environmental Quality (ADEQ) during the winter of 1988-89 showed similar composition to Phoenix samples except that the lead concentrations in Nogales were much larger. Surface wind measurements were measured at this location during the 1988-89 study and have been continued to the present. Additional monitoring has been instituted by ADEQ in Arizona and by the Mexican equivalent of the U.S. Environmental Protection Agency during the past two years. However, this area contains very complex terrain and a few measuring locations are not adequate to cover the movement of pollutants throughout the airshed. Most of the populated area of Nogales, Sonora, is in a narrow valley and the drainage flow at night funnels through roughly a one kilometer opening into Nogales, Arizona. Many side valleys are also present with different drainage directions and in some cases different entrance locations into Arizona.
The Arizona State University (ASU) group has been using physical and numerical models to develop detailed wind fields for the area on both sides of the border and synoptic climatology classifications to identify the meteorology for days with high and low PM-10 levels. In order to adequately understand the air quality of the Nogales area, detailed wind fields are being developed for each of the synoptic classes including times when the air flow is controlled by local topography as well as when outside air flows influence the local winds. Dispersion modeling based on these wind fields can identify "hot spots" and long term exposures to pollutants.
There are few if any studies of concentration distributions of pollutants in small mountain-valley areas with substantial populations. Ambos Nogales is an example where approximately 300,000 people reside and where the PM-10 ambient standards for 24 hours and annual averages are exceeded. Problems exist with current methods when slope and valley low speed winds are dominant and in modeling annual average pollutant concentrations. The vertical dispersion at night under very stable atmospheric conditions is poorly understood. Details of the three dimensional wind field in complex terrain and the method of averaging so that the true average flow is used in dispersion models are not available.
Approach: A grid point dispersion model based on the Urban Airshed Model is to be used with a grid size in the horizontal directions of 250 meters. Since the sources of the pollution are not precisely known, possible sources on both sides of the border will be tested to identify hot spot locations for field tests. This model simulates a 24 hour period and requires inputs of wind vectors at each grid point. A synoptic classification study is to be used to identify 15 types of daily weather and to associate the types with PM-10 measurements. The dispersion model will be run only for representative days of the types with high PM-10. Detailed wind fields are obtained for the high PM-10 types from the Diagnostic Wind Model with inputs from field measurements and physical model studies. In order to validate the model field sampling in conjunction with studies directed by the Arizona Department of Environmental Quality were anticipated. An eight stage cascade impactor will be used to collect particle size ranges below 10 microns and the samples will be analyzed by Particle Induced X-ray Emission (PIXE).
Status: The synoptic index for the Nogales region was originally developed from hourly surface observations collected at Tucson, Arizona (the nearest source of hourly weather measurements) for the winter seasons of 1973-1991, October through February. This period coincides with the record of particulate matter measurements in southern Arizona. The resulting calendar associating one of the 15 types with each day was compared with the particulate matter measurements and high particulate matter was found to be associated with three synoptic types. The high pollution days had low surface wind speeds, low cloud cover and a large diurnal temperature range suggesting that the surface wind flows were controlled by local slope and valley mechanisms.
The synoptic index has been updated through 1993. This was accomplished first by obtaining the necessary weather elements from the Western Regional Climate Center in Reno, Nevada and then merging the new weather data with the previous weather data. Using the six components and the 15 weather types derived for the earlier index, the new weather elements were combined with the old elements, resulting in a continuous index from 1973-1993. Average wind speeds and directions were determined for each synoptic type for the wind at the post office in Nogales, Arizona. The dominant wind direction is from the south to south-southwest. At about 0900 or 1000 LST the wind direction becomes variable with a general shift to the North. The wind becomes variable and shifts again at about 1600 LST, this time to the south-southwest.
In order to determine the correct parameters in the Diagnostic Wind Model, a synoptic index for Phoenix, Arizona was developed for the years 1948-1993. Twelve air mass classes were used in this index. Wind fields were determined for each class using data for a typical day in each class. The results are given in the Thesis of Mei-Hui Chen (1995). Most of the winter days had average wind speeds lower than the threshold measurement for many of the field instruments and only the airport location was used in the model. Attempts to determine some average wind speed to use in the model were not successful.
During the high pollution periods the low surface winds are determined by local surface cooling at night and there is little vertical mixing. This condition was simulated in a 0.79 m square box filled with water with the lower surface made to model a 12 km square region of the surface topography at Ambos Nogales. The aluminum base was cooled to produce the drainage flow and photographs of dye initially placed on the surface at many locations were used to obtain surface velocities. Vertical profiles of temperature and velocity were measured in the model at several different locations. Temperature measurements are given in the Thesis of Suvarna (1995).
Chen et al. (1993) and Berman et al. (1995) have discussed the similarity analysis for the modeling of physical systems. Although there are nine external dimensional parameters, only three dimensionless groups require matching in the model and the atmosphere. These are the dimensionless topography contour function, the dimensionless cooling time period and the dimensionless maximum elevation change. For the fixed location at Nogales the remaining two parameters suggest that an investigation of the vertical structure of the temperature and velocity in the model can be transferred to the atmosphere even though the flow in the model is laminar and the flow in the atmosphere is turbulent. We have investigated the variation in depth of the drainage flow and the maximum velocity near the surface. Although the experimental results seem to match measured wind velocities using the previous scaling analysis, it is clear that the results cannot be transferred to another valley. Therefore, more work needs to be done to develop the proper scaling for the three dimensional valley wind flow. This study is continuing with a PhD. student from Mexico supported by the Mexican government.
The physical model studies and the windfield study are the first steps in a dispersion analysis. Results of the dispersion calculations can be used in a risk analysis study when population density and emission models are coupled to the other parts of the model. With uniform emissions along the highway leading through the area, the highest population is always found in Mexico near the center of the City of Nogales, Sonora. The highest concentrations in the United States are at most 40% of those in Mexico. The complete analysis of the dispersion analysis for the 15 synoptic types will be presented in a thesis by Kirthi Kattige to be published in December, 1996.
An eight stage cascade impactor was setup at ASU to check the procedure for obtaining samples and analysis by PIXE. Initial results using glass fiber filters showed that the heavy metals in the filter were too high for use with PIXE. Additional trials with polycarbonate filters coated to allow particles to stick are in progress. Some PM-10 samples taken by the State of Arizona which had been previously analyzed by X- ray fluorescence were analyzed by PIXE. The relative results were similar but we have some work to do to obtain absolute numbers. Adequate procedures were not developed in time to do field measurements in Nogales.
Practical Use: Users include state and county air pollution professionals on both sides of the border. Discussions of the procedure and scientific analyses in meteorology and fluid dynamics will be available in the published literature.
References used in preparing this report
Last updated 7/1/99