**2. Methodology**

#### **2.1 Sampling**

During three two-week sampling campaigns (November 2003 and 2004, and March 2004), PM2.5 samples were collected at two sites with different land use. **Figure 1** shows the two sampling sites, along with the orographic characteristics of the Mexico City Metropolitan Area. La Merced (MER) is an important commercial site with a major food market, located approximately 0.2 km west of downtown; the site is near to moderately traveled paved streets. Xalostoc (XAL) is located in the northeast of the city, adjacent to heavily traveled paved and unpaved roads where old and new gasoline and diesel vehicles transit in an industrialized area; this site is 5 km west of the dry Lake Texcoco, thus resuspended dust from the dry lake bed may affect this site.

Particle samples were collected with battery-powered MiniVol portable PM2.5 samplers (Airmetrics) operating with a volume intake of air of 5 L/min for 24 h. Before sampling, the filtration volume of the samplers was previously calibrated at standard conditions of temperature and pressure. Teflon membranes were used to measure

mass by gravimetry. Details of the equipment and sampling procedures can be found elsewhere [9].

### **2.2 PM2.5 concentration and composition measurements**

PM2.5 mass was determined by duplicate samples collected on 47 mm Teflon filters, which were weighted in a microbalance prior to and after sampling.

Ions were collected on quartz filters and extracted in deionized water by sonication. Ions were analyzed by high-performance liquid chromatography. Nitrate, sulfate, and ammonium were analyzed by ion exchange chromatography. Organic carbon and elemental carbon were analyzed by thermal–optical reflectance.

#### **2.3 Stable nitrogen and carbon isotopes measurements**

Stable carbon and nitrogen isotopes of PM2.5 were collected in pre-combusted GF/F filters, which were placed in the MiniVol samplers. After sampling, filters were freeze-dried to remove any water or humidity in the sample, placed in aluminum vials, and inserted into a combustion column. Generated gases (CO2 and N2) were analyzed in a MAT-Finnigan 252 stable isotope mass spectrometer to quantify the 13C/12C and 15N/14N isotopes in each sample. Standards used for instrument calibration were acetanilide, methionine, and urea, all with a known isotopic value. An acetanilide substandard was run for every 10 samples. Carbon and nitrogen isotopes are referred to as VPDB and Vienna air nitrogen standards, respectively.

#### **2.4 Meteorological and ambient air data**

Ancillary ambient data were obtained from the RAMA and REDMET data sets. RAMA (Spanish acronym for automatic ambient monitoring network; [17]) measures criteria atmospheric pollutants (PM10, CO, CO2, NOx, NO2, SO2, and O3) at 15 locations within Mexico City on an hourly basis. At MER, PM2.5 hourly data are also measured by RAMA. Wind velocity and direction, relative humidity, and temperature measured on an hourly basis were obtained from the REDMET (meteorological network) database [17]. For each sampling period, the hourly data were averaged, except for wind direction where the hourly mode was used.
