Dry deposition of aeolian (windblown) dust and sand represents an important physical and chemical flux of matter and nutrients in and downwind of deserts (Prakash et al., 2016). Falling dust and sand has numerous environmental effects, affecting the composition and structure of dryland soils (Munroe et al., 2024;Rea et al., 2020;Reheis et al., 1995), impacting plant physiology (Jiao et al., 2018;Soheili et al., 2023), and altering snowpack composition and hydrology (Reynolds et al., 2014). Falling aeolian sediment impacts human infrastructure, soiling buildings and machines (Duniway et al., 2019), contaminating drinking water (Sánchez et al., 2015), reducing traction on paved roads (Pan et al., 2021), inhibiting the performance of photovoltaic systems (Abuzaid et al., 2022), and degrading overhead electric power lines (Maliszewski et al., 2012). Dust is a major component of airborne particulate matter in drylands and regions downwind of them, resulting in potentially hazardous concentrations of air pollution (Rivera Rivera et al., 2010).
An effective, low-tech, widely-used technique to collect and quantify windblown sediment is the deployment of passive collectors (dust traps).
Marble dust collectors (MDCOs) (Goossens and Offer, 2000) have been frequently utilized to capture falling dust and sand and determine dust deposition rates (Sow et al., 2006). Advantages of MDCOs are that collected dust particles are protected against deflation by wind and rain, and help diminish bias in the particle size distribution of collected dust (Sow et al., 2006).
The largest metropolitan area in the Chihuahuan Desert of North America is El Paso, Texas, USA/Ciudad Juarez, Chihuahua, Mexico (the Paso del Norte region), with a combined population of over 2 million. The city of El Paso, at the westernmost point of Texas, is separated from Ciudad Juarez by the Rio Grande/Rio Bravo Del Norte river. There is a large residential and industrial component of El Paso along with open desert and mountain topography contained within city limits. Air quality issues in the metropolitan area include excessive ozone (Karle et al., 2021), gaseous pollutants (McCoy et al., 2010), traffic-related emissions (Raysoni et al., 2011), industrial emissions (Grineski and Collins, 2010), and fugitive dust resuspended within the metropolitan area (Garcia et al., 2003(Garcia et al., , 2006)). However, windblown desert dust represents the most frequent and most extreme threat to air quality in El Paso (Novlan et al., 2007;Rivera Rivera et al., 2010) (Fig. 1).
Encompassing a climatic boundary between the subtropics and the middle latitudes (Castiglia and Fawcett, 2006), the northern Chihuahuan Desert experiences strong winds causing frequent dust events, forming one of the main dust sources in North America (Prospero et al., 2002). El Paso averages approximately 22 dust events per year (Robinson and Ardon-Dryer, 2024). Dust and sand transport around El Paso causes exceedances of air quality standards (Lee et al., 2009;Rivera Rivera et al., 2010), deflates soil and leads to sand dune formation (Gillette and Monger, 2006;Langford, 2000), transports aquatic microinvertebrates to new habitats (Rivas et al., 2018(Rivas et al., , 2019)), reduces visibility presenting a significant driving hazard (Li et al., 2018;Tong et al., 2023), and impacts human health (Grineski et al., 2011;Herrera-Molina et al., 2021;Tong et al., 2017). Dust generated in the Chihuahuan Desert near El Paso can cause transient spikes in particulate matter concentrations hundreds of kilometers downwind (Lee et al., 2009) and has been detected in wet deposition thousands of kilometers downwind, in southeastern Canada (Park et al., 2007).
Windblown dust and sand in the Chihuahuan Desert tends to occur as distinct types of events in different seasons. Synoptic events occur with west to southwest winds during the drier, cooler part of the year (October-May) (Novlan et al., 2007), typically associated with Pacific frontal passages and low-pressure areas traversing to the north (Rivera Rivera et al., 2009). Mesoscale events, typically convective outflows (haboobs) associated with the North American monsoon, primarily occur in June through August (Novlan et al., 2007). Source areas upwind of El Paso include playas (dry lake beds), dry riverbeds, and alluvial lowlands in the Chihuahuan Desert (Baddock et al., 2011(Baddock et al., , 2016(Baddock et al., , 2021)), as well as agricultural lands in Chihuahua, Mexico (Lee et al., 2009;Rivera Rivera et al., 2010). Desert sand sheets surrounding El Paso (Langford, 2000) emit dust which impacts the city (Baddock et al., 2011), while gypsum dust from White Sands, New Mexico (White et al., 2015) occasionally advects into El Paso from the north-northeast. In the air quality management context, El Paso's dust events and the strong influence of regional transport pathways into the city represent "exceptional events" from natural sources and/or from beyond international borders, which are not reasonably controllable or preventable (Dayalu et al., 2024). Within the metropolitan area, locally-generated fugitive dust emanates from unpaved roads, bare lots, construction sites, and industrial facilities (Garcıa et al., 2003).
Aeolian dust/sand deposition has been measured elsewhere in west Texas, including Lubbock (Crabtree, 2004;Warn and Cox, 1951), Big Spring (Crabtree, 2004), Salt Flat Basin (Perez and Gill, 2009), and rural El Paso County (Ortiz and Jin, 2021;Rivas, 2019), but not within urban El Paso. Given that knowledge gap, the aim of this project was to determine deposition rates, particle size, elemental, and mineralogical composition of aeolian dust and sand being transported to the city of El Paso, Texas during synoptic-scale events, and place these properties in context to other dust-prone cities in North America and elsewhere. The samples used in this work were collected using MDCO passive deposition traps initially to investigate wind dispersal of aquatic microinvertebrates (Rivas et al., 2018(Rivas et al., , 2019)). Rivas et al. (2018Rivas et al. ( , 2019) ) include some event-specific data on amounts of sediment deposited from individual wind events along with selected HYSPLIT trajectory results showing regional dust source areas. This study aggregates more than five years of deposition data, alongside particulate matter concentration data from a nearby continuous air monitoring station, to quantify and characterize dry dust/sand deposition in El Paso.
Samples of the wind-deposited dust and sand from which microinvertebrate propagules were rehydrated by Rivas et al. (2018Rivas et al. ( , 2019) ) were analyzed to determine dust deposition rates, particle sizes, and chemical and elemental composition. Wind trajectories from the events and measurements from an adjacent air monitoring station were used to contextualize transport pathways and particulate concentrations associated with these dry deposition event characteristics. Some experimental methods pertinent to this study were summarized in Rivas et al. (2018). A full description of the methods is provided and expanded on here.
Bulk falling aeolian sediment was collected with MDCOs on the flat rooftop of the Biology Building on the campus of the University of Texas at El Paso (UTEP) (31.768 N, -106.504 W; elevation 1170 m, height of 21 m above the ground), in an urbanized southwest part of the metropolitan area (~7 km downwind of the edge of open desert, but in the upwind part of the Paso del Norte metropolitan area for most events). Collections were made from March 2011-May 2016 and stopped in fall 2016 when major land clearance and construction activities commenced <150m away, exposing bare soil and causing fugitive dust to be emitted near the sampling site.
A total of 68 synoptic wind events were sampled on an event-byevent basis. Thirteen events were briefly described in Rivas et al. (2018) and meteorological, event-specific deposition, and PM data on 67 of the events were included in Supplemental Document 1 in Rivas et al. (2019). A collection season was defined as from October through May, as dust/sand events from June through September are almost exclusively mesoscale events associated with convective storms of the North American Monsoon (Novlan et al., 2007;Robinson and Ardon-Dryer, 2024), and only synoptic events with only dry deposition (no precipitation while traps were deployed) were sampled. Traps were not deployed for dust events when precipitation was also forecast due to the need for samples to remain dry for biological analyses (Rivas et al., 2018(Rivas et al., , 2019)). Any unforecasted synoptic-scale blowing dust/sand events, all of short duration and intensity, were not sampled. When a forecasted event did not materialize, no detectable amounts of sample were deposited in the MDCOs, and it was not counted as an event. Six samples from Spring 2011 are not included in collection year deposition rate data. PM 2.5 and PM 10 data from the adjacent air quality monitoring site were not available during three events due to monitor malfunction or maintenance happening at those times. Only seven samples were collected during 2014-2015, primarily due to transport capacity limitation (lack of wind). For this reason, several samples representing long-term collection periods or background conditions were undertaken during this year, including stretches of multiple days (12/13/2014-2/15/2015) in which dust was not reported in airport weather observations. MDCOs (traps) (n = 7) with a total surface area of 0.9 m 2 were placed on the rooftop in scattered orientations the day before each forecasted dust event and were deployed until meteorological observations of dust at El Paso had concluded and continued dusty conditions were no longer forecasted. Events were forecasted by monitoring the United States National Weather Service-El Paso's twice-daily Area Forecast Discussions for mention of possible blowing dust or dust storm during the next day. Hourly or more frequent weather observations (including visibility, wind speed, wind direction, wind gust, event start time, and event end time) at El Paso International Airport, 10.6 km east of the dust sampling site, were collected, with airport weather observations of dust events (Blowing Dust, Dust, Dust Storm, or Haze with wind speed >20 mph (8.9 m/s) and without precipitation; as reported by and with the criteria used by the USA National Weather Service (Ardon-Dryer et al., 2023)) noted.
Traps were also set out during a non-dusty wind event from 1/10/13 to 1/13/13 in which wind speeds were light to breezy and dust was not being forecast or suspended, to characterize local "background" deposition rates and characteristics outside of identified events. Average wind speed during this collection period was 2.6 m/s. During the October 2014-May 2015 synoptic dust season El Paso experienced very few wind events, so traps were left out for longer periods of time, from weeks to months, and data from the 2014-2015 season includes a larger proportion of background conditions.
Sample collection procedures were adapted from Goossens and Rajot (2008). Using fine angled clean paintbrushes, one for each trap, the dust was brushed into a corner, placed into small clean containers and weighed with an analytical balance. Each sample was divided into up to four subsamples; the first for rehydration of biota (see Rivas et al., 2018Rivas et al., , 2019)), the second for particle size analysis, the third for mineral composition, and if sufficient sample remained, the fourth for elemental analysis. A minimum of 1.5 g of deposited sediment per event aggregated from all MDCOs was sufficient to run all four analyses. Deposition rate was calculated by dividing the total amount of material collected per event by the total surface area of all traps over the time MDCOs were deployed. Deposition rate over the synoptic dust season (October-May) was expressed as g/m 2 /year, a convenient and widely-used metric for such studies (Lawrence and Neff, 2009).
A Texas Commission on Environmental Quality (TCEQ) continuous air monitoring station (CAMS-12) was located 217m east-southeast of the MDCO collection site. The CAMS continuously recorded hourly data on weather and air quality, including PM 2.5 and PM 10 concentrations measured by tapered element oscillating microbalance (TEOM) (Ruppecht et al., 1992). Aerosol and meteorological data from the CAMS-12 site were obtained for each dust event.
Transport pathways of dust associated with each of the 68 events were determined by creating back trajectories of wind arriving at UTEP: 13 of which were described in Rivas et al. (2018). Trajectories were calculated using the NOAA HYSPLIT model (Stein et al., 2015). Default parameters were used with a total run time of 24 h (or the duration of the dust event), number of trajectories set to 24, and a height of 500 m above ground level (representing dust in the boundary layer which therefore could be deposited) for each event. The ending time for each HYSPLIT run was the hour when PM 10 concentration first decreased below 100 μg/m 3 at the TCEQ CAMS-12 site. The receptor site, UTEP, was selected by using the latitude and longitude coordinates of the traps.
Approximately 0.3 g of sample per event was tested following general procedures for direct wet grain size analysis by laser diffraction (Sperazza et al., 2004), using a Malvern Mastersizer 2000 Laser Diffractometer (Malvern Instruments, Worcestershire, UK). To prepare the sample, the sediment was sieved to remove any grains >2 mm (larger than sand size), placed in a clean plastic bottle and 20 ml of sodium hexametaphosphate solution (50 g/L) was added. The mixture was then placed on a reciprocating shaker for at least 8 h. After shaking the sample was then placed into the HydroG circulating fluid vessel of the Mastersizer and a standard operating protocol was initiated, including sonication of the sample for 60 s at 80% of full sonication. Obscuration of the sample during laser diffraction analysis was kept between 10 and 20 % to optimize sample analysis. The system was flushed with ultrapure distilled water before each analysis, and a test analysis was conducted prior to analyzing the dust sample using ISO standard 12103-1, A4 Coarse Test Dust (Powder Technology Incorporated, Arden Hills, MN, USA). Particle size classification scheme adapted from the USDA (Soil Survey Division Staff, 1993) was used to identify equivalent textural classes (sand, silt, clay) of each sample of wind-fallen sediment. Grain sizes were classified as clay (<2 μm), silt (2-50 μm), and sand (>50 μm).
If enough sample remained after particle sizing, approximately 0.25g was used for mineralogical characterization with a Rigaku Miniflex II Benchtop X-Ray Diffraction (XRD) instrument. Samples were prepared by pulverizing the material in a synthetic ruby mortar and pestle with 100% acetone until it obtained a fine consistency and a paste-like substance was formed. Jade 9.0 (Jade Software Corporation), DIFFRAC PLUS Bruker AXS (1996-2007) (Bruker Instruments Inc.), and EVA 14.0 (Bruker Instruments Inc.) programs were used for pattern processing and identification of mineral phases. Clay mineral analyses were not performed.
If sufficient sample remained after XRD, elemental analysis for likely major and minor elements (Na, Mg, Al, Si, P, K, Ca, Ti, Mn, Fe, Sr, Zr, Ba) was performed using inductively-coupled plasma optical emission spectroscopy (ICP-OES, PerkinElmer 5300 DV I). The lithium metaborate fusion technique (Feldman, 1983;Ingasmells, 1970;Medlin et al., 1969;Suhr and Ingasmells, 1966) was used for sample preparation. Approximately 0.1 g of sample was added to 1.0 g of lithium metaborate and shaken for 2 min. Samples were placed into a crucible and placed into an oven at 900 • C for 15 min. Samples were then poured into Teflon beakers with 100 ml of 5% nitric acid and stirred for 15 min. Prior to analysis samples were diluted with 5% nitric acid at a ratio of 1:10. The mixture was then transferred to polyethylene bottles until analysis by ICP-OES. Following the protocols of Nyachoti et al. (2019), twenty rock standards from NIST and USGS were digested and analyzed with each batch of samples using the same procedure and used as calibration standards for major elements along with a procedure blank, and the USGS reference rock material W-2 was digested and analyzed with each batch of samples as a quality check. The differences between measured and certified values in all major element concentrations on W-2 were <10%. There was an apparent system error in elemental analyses of the 2014-2015 and 2015-2016 samples; data for these samples are not included.
Several statistical techniques were used to analyze some of the datasets. One-way analysis of variance (ANOVA) was used to test for significant differences among annual average PM 10 concentrations and texture of sediment deposited into the MDCOs. ANOVA tests were used to determine the differences among mean of PM 10 concentrations for all sampling hours, collection hours, and dusty hours. SAS Proc GLM (SAS 9.4, 2019) and Duncan Multiple Comparison methods were used for the ANOVA with significance level for all analyses set at 0.05. Additionally, PM 2.5/10 ratios between MDCO and CAMS data were analyzed using SAS software version 9.4 (SAS Institute, Inc., 2019) as Two-Sample Independent Sample T-test. The assumption of equality of variances was checked using Levene's test (Zar, 2009) and then equal variances T-test (pooled) or unequal variances T-test (Satterthwaite T-Test) were used for PM 10 and PM 2.5 yearly comparisons (Zar, 2009).
A total of 68 events were sampled. Six samples from Spring 2011 are not included in collection year deposition rate data since they represented only part of a collection year.
Summary data including number of events sampled, number of days traps were deployed, and sediment deposition rate for each year are shown in Table 1.
Annual deposition rates of falling dust and sand trapped atop the UTEP Biology Building during synoptic dust events over collection seasons ranged from 85 g/m 2 /yr to 164 g/m 2 /yr for the five-year period (Fig. 2). Average annual deposition rate was 111 g/m 2 /yr. The highest deposition rate occurred in 2012-2013, the year with the greatest number of events, as well as the culmination of an extensive drought. The reduced rainfall over the Southwestern USA during prior seasons through 2011, intensifying into 2012, and ecosystem "memory" of soil moisture and drought resulted in increased dust emission in the Chihuahuan Desert (Eibedingil et al., 2024;Hoerling et al., 2014). Conversely, during the 2014-2015 collection season, less of the surrounding region was in drought (Eibedingil et al., 2024), and there were fewer wind events; the deposition rate was the lowest, suggesting that this season was somewhat transport capacity and/or sediment availability limited.
Deposition rates from passive aeolian sediment traps should be considered a general range best compared to other deposition measurements obtained with the same type of collector. With any kind of passive falling-sediment collector, deposition rates and characteristics such as particle size of material collected are dependent on factors including emission rate at the source (Shao, 2008), dust/sand concentration, wind velocity (Gong et al., 2024;Sow et al., 2006), orographic features of the environment (Lawrence and Neff, 2009), wind gustiness (Yan et al., 2017), land use/land cover changes (Belnap et al., 2009), and overall weather conditions throughout the source area (Reheis and Kihl, 1995). While MDCOs are widely used and their data are widely reported, the amount and characteristics of sediment they collect will depend on the specific design and dimensions of collector used (Gong et al., 2024;Goossens, 2007;Sow et al., 2006), orientation of the trap with regards to the wind direction (Sow et al., 2006), height above the land surface (Goossens and Rajot, 2008) and urban topography including location within the urban area and density and orientation of buildings (Erell and Tsoar, 1997;Putman et al., 2022). With these provisos in mind, Table 2 shows rates of aeolian sediment deposition obtained with marble dust
Selected aeolian sediment deposition rates (dust + sand) from passive aboveground dust traps with marble substrate, including standard MDCO trap (Goossens and Offer, 2000) and USGS marble dust trap (Reheis and Kihl, 1995).
Where four or more distinct sites were sampled, range and average are given if available.
Location Deposition rate (g m -2 yr -1 ) Source Kuwait City, Kuwait 340-940 Average: 590 Al-Awadhi and Al-Shuaibi (2013) Riyadh, Saudi Arabia 454 Modaihsh et al. (2017) Farmington Bay playa, Great Salt Lake, Utah 315-404 Putman et al. (2023), Blakowski et al. (2022, 2023) Southern Owens Lake playa, California ~290 Reheis (2003) Lake Qinghai, China 266 Wan et al. (2012) Jabal Haroun, Jordan ≥250 Lucke et al. (2019) El Paso-UTEP, El Paso County, Texas 111 This work Lubbock, Texas 103 Crabtree (2004) Big Spring, Texas 77 Crabtree (2004) Beer Sheva, Israel (urban sites) Long term: 49-95 Average: 74 single event: 1393 Erell and Tsoar (1997) White Sands Missile Range, New Mexico 69 Velarde (2011) Hueco Tanks State Park and Historic Site, El Paso County, Texas 56 Rivas (2019) Northern Utah-urban regions 54 Scholz et al. (2019), quoted in Brahney (2019) Southeast El Paso County, Texas 50 Ortiz and Jin (2021) Salt Flat Basin, Texas 50 Perez and Gill (2009) Northern Utah, 2022-excluding playa sites 16-74 Average: 35 Blakowski et al. (2023) Northern Utah-urban regions 25-35 Average: 31 Goodman et al. (2019) Northern Utah, 2018-2019, urban regions 8-82 Average: 24 Blakowski et al. (2022) Rio Grande Valley, southern New Mexico 10-60 Average: 23 Gile and Grossman (1979) Colorado Plateau, Utah 14-32 Average: 22 Belnap et al. (2009) Canyonlands region, Utah, 1999-2008 (average of multiple sites) 22 Reheis and Urban (2011) Southern Nevada, Southern California, 1984-1999 8-207 Average: 18 Reheis (2003) Southern Sierra Nevada, California 6-36 Average: 16
Aciego et al. (2017) Northern Sacramento Mountains, New Mexico 11 Rea et al. (2020) Mojave Desert and Southern Great Basin, 1999-2008 (average of multiple sites) 9 Reheis and Urban (2011) Perry Mesa, Yavapai County, northern Arizona 3-14 Nakase et al. (2014) Front Range, Colorado 5 Ley et al. (2004)
collectors worldwide, including this study.
Although the MDCO dust and sand deposition rate in urban El Paso is lower than those of most sites in the "global dust belt" (region defined by Prospero et al., 2002), it is higher than rates reported elsewhere in North America, except for at the downwind edges of dust-emitting playas including Owens Lake (Reheis, 2003), which has been one of the continent's most intense dust sources (Cahill et al., 1996;Reheis, 2003), and the Farmington Bay playa of Great Salt Lake, an intensifying dust source (Blakowski et al., 2022(Blakowski et al., , 2023;;Putman et al., 2023). Bulk deposition rates are relatively higher in the eastern Chihuahuan Desert sites (El Paso, White Sands Missile Range, Hueco Tanks State Park and Historic Site, Southeast El Paso County, Salt Flat Basin) extending into the Southern High Plains (Lubbock, Big Spring) than reported for the rest of North America (Table 2). This region also broadly matches a springtime (synoptic dust season) zone of enhanced coarse aerosol mass (CM = PM 10 -PM 2.5 ) consistent with synoptic-scale airflow patterns (Hand et al., 2019). Deposition rates in the urbanized Wasatch Front of Utah are slightly lower than in the Chihuahuan Desert but slightly higher than those reported for the interior mountains and Colorado Plateau. The Mojave Desert and southern Great Basin appear to have lower deposition rates of falling dust and sand than other USA drylands. High-elevation sites (northern Sacramento Mountains, Front Range) have low vertical flux of aeolian sediment, much as they have reduced CM due to increased land cover, distance from dust sources, less bare soil, and/or elevation-limited transport (Hand et al., 2019). No MDCO measurements appear to have been reported from some other dusty regions, such as the Sahara and Sahel of Africa, the Aral Sea basin, or the Sonoran Desert lowlands and Columbia Plateau of the United States, so data from those areas are not included in Table 2.
Table 1 and Fig. 1 in Lawrence and Neff (2009) show a comprehensive list of annual deposition rates from 52 sites worldwide ranging from 0.05 g/m 2 /yr at Penny Ice Cap, Canada (representing minimum global background transport) to 450 g/m 2 /yr (local transport) in China. Compared to Lawrence and Neff (2009)'s findings, El Paso-UTEP average annual deposition would place above the 80th percentile of the reported areas. Fig. 3 shows El Paso-UTEP data in the context of their results. By Lawrence and Neff (2009)'s classifications, El Paso-UTEP dust would be considered "local," very near to active source areas, having travelled ~10 km or less from the source. While most of the dust sources identified by remote sensing as impacting El Paso (Baddock et al., 2011(Baddock et al., , 2021;;Rivera Rivera et al., 2010) would be considered "regional" in nature (~100 km upwind) and expected to be associated with lower deposition fluxes, the deposition rates measured at UTEP were approximately twice those measured in rural portions of El Paso County (Ortiz and Jin, 2021;Rivas, 2019) (also shown on Fig. 3), suggesting urban enhancement of deposition rates. There are a wide variety of fugitive dust sources throughout the Paso del Norte metropolitan area (Garcıa et al., 2003), and the MDCO deployment site is < 1 km downwind of Ciudad Juarez, Mexico, a city with a high percentage of roads that are unpaved and highly fugitive-dust-emissive (Kavouras et al., 2016). The high rate of dust and sand deposition at the UTEP-El Paso site is consistent with increased dust deposition rates in urban areas, due to local anthropogenic dust sources (Sorooshian et al., 2011), as well as deceleration of particle-laden winds as they enter a roughness-enhanced urban area which causes dust deposition (Putman et al., 2022;Zhang et al., 2001).
On the other hand, the annual sediment deposition rates calculated for UTEP-El Paso must be underestimates, perhaps large ones. Due to the original purpose of the sampling (the collection of dry microinvertebrate propagules; Rivas et al., 2018Rivas et al., , 2019)), sediment traps were not deployed for synoptic events when precipitation was forecasted, nor for any convective dust events which are almost always accompanied by thunderstorms. Thus, multiple events were not sampled each year. In addition, a certain amount of background deposition takes place over the majority of the year outside of high-wind storm events.
HYSPLIT back trajectories associated with collected samples are shown in Fig. 4, corresponding to synoptic-scale dry deposition events only and excluding the long term or "background" sample. A total of 67 trajectories were run, 13 of which were described in Rivas et al. (2018); trajectory results indicate airflow primarily from the westerly to southwesterly direction with several events originating from the northerly or easterly directions. The dominant transport direction during synoptic dust events was from the southwest (57% of the trajectories), in accordance with the long-term record of prevailing dust-raising winds impacting El Paso (Novlan et al., 2007) and consistent with the direction of trajectories associated with anomalously high PM 2.5 concentrations in El Paso (Dayalu et al., 2024). An additional 22% of the trajectories came from the west, 7% from the south, 6% from the northwest, 3% from the southeast, 3% from the north and 2% from the northeast. The few trajectories from the north and northeast suggested dust advecting to El Paso from the White Sands of New Mexico (White et al., 2015) and the Great Plains (Stout, 2001).
The surface horizons of soils underlying the trajectories west of El Paso on the U.S. side of the international border are generally mapped as loamy fine sands including poorly-developed soils developed from windblown sands (Soil Survey Staff, 2024) with ecological site descriptions of "sandy" and "deep sand" (Bestelmeyer et al., 2016). Trajectories arriving from Mexico cross soils mapped (using the International Union of Soil Sciences (2022) system) as calcic and eutric regosols in desert plains, and solonchak/solonetz in paleolake basins and alluvial lowlands (Instituto Nacional de Estadística, Geografía e Informática, n.d.). To summarize, trajectories arriving from the west and southwest over the Chihuahuan Desert (approx. 80% of trajectories during synoptic dust events) traverse erodible sandy and lacustrine soils that would readily entrain materials with compositions consistent with the subsequent mineral and elemental analysis. Remote sensing studies have identified multiple hotspots of dust emission within the southwestern and western transport pathways into El Paso, including alluvial lowlands of the Casas Grandes River (Rivera Rivera et al., 2010) and Mimbres River (Kandakji et al., 2020), playas such as the Paleolake Palomas Basin (Baddock et al., 2011(Baddock et al., , 2016(Baddock et al., , 2021)), and sand sheets and dunes surrounding El Paso (Langford, 2000). Fig. 4. HYSPLIT 24-h back trajectories for each of 67 dust events (the long-term sampled background event of 12/13/14 to 2/15/15 not included) over the five-year collection period, ending at the collection site in El Paso (yellow star). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
Grain size (mean ± SD) of sediments trapped in the MDCOs for each collection year and entire collection period. Proportion of particles in the MDCOs in PM 10 and PM 2.5 size ranges and simultaneous PM 2.5/10 ratios for MDCO sediments and measured by CAMS-12 TEOM for collection hours also shown.
Collection year Mean MDCO grain size (μm) % PM 10 MDCO (Mean ± SD) % PM 2.5 MDCO (Mean ± SD) PM 2.5/10 Ratio MDCO (Mean ± SD) PM 2.5/10 Ratio TEOM (Mean ± SD) 2011-2012 106 ± 6.15 10.2 ± 5.47 4.12 ± 2.50 0.403 ± 0.05 0.13 ± 0.05 2012-2013 124 ± 5.26 9.50 ± 4.78 3.76 ± 1.66 0.395 ± 0.04 0.19 ± 0.08 2013-2014 99 ± 5.47 11.30 ± 5.76 4.12 ± 1.59 0.365 ± 0.05 0.13 ± 0.03 2014-2015 156 ± 4.68 6.20 ± 3.07 2.00 ± 0.67 0.323 ± 0.06 0.22 ± 0.09 2015-2016 122 ± 4.52 8.20 ± 4.80 3.02 ± 1.58 0.366 ± 0.05 0.17 ± 0.03 Entire period 108 ± 5.28 9.30 ± 4.15 3.50 ± 1.85 0.376 ± 0.05 0.14 ± 0.06
A total of 54 events yielded sufficient sample to perform grain size analysis, 10 of which were reported in Rivas et al. (2018). Table 3 shows mean grain size of wind-carried sediments deposited in the MDCOs, PM 2.5 /PM 10 ratios for those sediments, and corresponding PM 2.5 /PM 10 ratios from the CAMS-12 TEOM. Table 4 shows the proportion of clay, silt, and sand sized grains in the MDCO-deposited sediment by collection year. Fig. 5 shows grain size curves for MDCO data for each collection year.
Tables 5A and 5B show mean and peak hourly PM 10 and PM 2.5 concentrations at the CAMS-12 site for each collection year and the fiveyear collection period for all hours; collection hours (all hours in which MDCOs were deployed); and dust hours (all hours in which dust was reported by weather observers at El Paso International Airport). Table 6 shows mean peak hourly PM 10 and PM 2.5 concentrations per event at CAMS-12 for each collection year and the five-year sampling period.
Mean grain size of MDCO-deposited sediment in all events was greater than 50 μm, the boundary between silt (dust) and sand, even though collectors were deployed 21 m above ground level. It is commonly assumed that suspension of sand-size particles is unusual and that most sand transport occurs as saltation (summarized in Zhang et al., 2021), though the height of the collectors was roughly an order of magnitude higher than the maximum height that sand grains generally move in saltation (Fryberger et al., 1979). Analysis of mean grain size of deposited sediment by collection year indicated that 2011-2012 (historic drought year) and 2014-2015 (low wind and dust) collection periods were significantly different from the rest of the years. Annual average mean grain sizes for all years (Table 3) would be classified as fine or very fine sand, and based on clay, silt, and sand percentages (Table 4), the material deposited in the MDCOs for every collection year as well as the overall period would have a loamy sand texture (Soil Survey Division Staff, 1993). Mean grain size by event ranged from 56 μm (3/6-3/8/12) to 371 μm (1/23-1/26/16). These results are consistent with grain sizes of sand sheet and nabkha dune sediments surrounding El Paso (Langford, 2000), and the surface textures of wind-erodible soils in the Chihuahuan Desert lowlands to the west (Soil Survey Staff, 2024). Therefore, at least as collected at UTEP, since most of the grains in wind events are sand falling from suspension, these events in El Paso could be characterized as blowing sand or sandstorms, rather than blowing dust or dust storms. A sample of falling aeolian sediment in the Texas Panhandle from the 1930s "Dust Bowl" was also comprised predominantly of sand-sized particles (Gill et al., 2000). Deposition over five to six years in the Southern High Plains dust source region of Texas was evaluated by Crabtree (2004): MDCO samples collected in Big Spring were predominantly sand, while samples from Lubbock were predominantly silt. El Paso has some of the highest average PM 10 concentrations in the USA. In a study of average annual PM 2.5 and PM 10 concentrations across the United States (Li et al., 2013), two sites in El Paso were the only locations with PM 10 exceeding 40 μg/m 3 . Values for CAMS-12 for our five-year collection period overall were not so high (28 μg/m 3 ), but during hours when dust was reported at El Paso International Airport, PM 10 averaged 200 μg/m 3 . PM 10 concentrations during all hours, dust hours, and peak concentration per event were highest in the historic drought period of 2011-2013, showing its influence on local aerosol concentrations. For all years except 2014-2015 (a period of low wind and dust), the standard deviation of PM 10 concentrations exceeded the mean value, illustrating the extreme variability of particulate matter concentrations in El Paso as driven by dust events. These relationships were not so clear-cut for PM 2.5 , and ANOVA analyses indicated significant differences between mean PM 10 concentrations and mean PM 2.5 concentrations per collection year, suggesting windblown dust's primary effect in El Paso is on PM coarse (particles between 2.5 μm and 10 μm in mean aerodynamic diameter) concentrations. The mean peak hourly PM 10 concentration over more than five years of synoptic-scale dust events in El Paso exceeded 550 μg/m 3 and the mean peak hourly PM 2.5 concentration during such events exceeded 65 μg/m 3 , with large standard deviations. All numeric PM values should be considered general upper bounds due to TEOM PM concentrations being higher than those obtained with other samplers during high winds (Sharratt and Pi, 2018) and heavy dust conditions (Ono et al., 2000).
At the scale of an individual windstorm, the dust concentration (represented by peak hourly PM 10 ) will be related to sediment availability (the amount of material available to be eroded from the land surface, highly responsive to moisture levels and land cover) and transport capacity (the velocity of the wind) (Mockford et al., 2018). The greatest peak PM 10 concentrations for individual events (Tables 5A and 6) occurred in 2011-2012, the height of one of the most extreme drought periods of record in the region (Hoerling et al., 2014), which reduced regional land cover, temporarily increasing sediment availability and dust emission (Eibedingil et al., 2024;Munson et al., 2011;White et al., 2015). As drought conditions diminished in the Chihuahuan Desert and land cover recovered over the following years, dust emissions decreased (Eibedingil et al., 2024) and so did peak event PM 10 concentrations at the El Paso receptor site (Tables 5A and 6). Peak event PM 10 concentrations dropped to approximately an order of magnitude lower in 2014-2015, the year of fewest events and ~40% lower total deposition, than during 2011-2013. Note, however, that total deposition was highest in 2012-2013, when there was a greater total number of synoptic wind events (see section 3.1). Total deposition of falling aeolian sediment does not seem to be related to short-term PM 10 concentration.
The mean PM 2.5 /PM 10 ratio from CAMS-12 TEOM measurements during all collection periods was 0.14 (annual range 0.13-0.22) and during synoptic dust event hours was 0.13 (annual range 0.10-0.17, individual event range 0.06-0.36). This very low PM 2.5 /PM 10 ratio in El Paso is consistent with high amounts of PM coarse being dispersed by windblown dust and fugitive dust. It is broadly equivalent to the PM 2.5 / PM 10 ratio during synoptic dust events at CAMS-12 for 2001-2010 (0.14) (Rivera Rivera, 2012), and consistent with ratios of 0.12 (measured with TEOM) during a windblown dust event in urban Spokane, Washington (Haller et al., 1999), 0.15 during a synoptic dust event in Lubbock, Texas (Ardon-Dryer et al., 2022), and ratios of ~0.07 to ~0.2 during high-dust periods in Phoenix, Arizona (Huang et al., 2015). PM 2.5 /PM 10 ratio collection year averages for MCDO samples ranged from 0.32 to 0.40, with ratios for individual events ranging from 0.25 to 0.46. MDCO PM 2.5 /PM 10 ratios were significantly larger than TEOM PM 2.5 /PM 10 ratios for all collection years and the entire collection period (two-sample independent sample T-test, pooled method for all periods except for 2012-2013 (Satterthwaite method), with P < 0.05). MDCO PM ratios are much larger due to the sample preparation which breaks down compound grains (aggregates) into their component particles in a detergent solution, as well as the different sampling methodologies of the two devices. TEOMs tend to overestimate PM 10 concentrations during conditions of high wind and/or dust (Sharratt and Pi, 2018), while MDCOs have many issues including decrease of efficiency as wind speed and particle size increase (Sow et al., 2006).
Mean percent clay, silt, and sand of sediment trapped in the MDCOs for each collection year and for the entire collection period.
Collection Year % clay % silt % sand 2011-2012 4.0 ± 2.3 21.0 ± 8.1 75.0 ± 10.0 2012-2013 3.2 ± 1.7 19.2 ± 6.1 77.6 ± 7.6 2013-2014 3.7 ± 1.4 22.7 ± 4.3 73.6 ± 5.1 2014-2015 1.7 ± 0.5 20.8 ± 7.5 77.5 ± 7.9 2015-2016 2.7 ± 1.5 19.2 ± 8.6 78.1 ± 9.8 Entire period (Mean ± SD) 3.17 ± 1.75 20.02 ± 7.08 76.79 ± 8.46
Measuring the mineralogy and chemistry of aeolian dust and sand deposition not only helps understand its role in nutrient delivery to ecosystems, but also helps in understanding the geochemical flux from aeolian transport (Lawrence and Neff, 2009). XRD analysis results for MDCO-collected sediment from 46 events (Fig. 6) indicated that quartz was the dominant mineral, present in all samples analyzed. In addition to quartz, measurable amounts of anorthite (37% of samples), albite (22% of samples), calcite (22% of samples), and microcline (17% of samples) were identified from samples. Gypsum, orthoclase, and muscovite were present in only a few samples.
XRD results indicate aeolian deposition at UTEP is comprised of minerals found in wind-erodible soils of the Chihuahuan Desert (Peinado et al., 2014), as well as northern Chihuahuan Desert soils in general (Gile et al., 1981) and dune sands immediately to the SW (upwind) of El Paso (Langford, 2000); quartz, feldspars, and calcite, with occasional presence of gypsum and micas. Clay minerals were not analyzed for in this study, though the MDCO samples likely contain clays. Aeolian sediment composition in El Paso is consistent with the mineralogy of locally-derived dusts worldwide (Lawrence and Neff, 2009). Quartz is the dominant mineral and silicon the dominant element
Mean PM 10 concentrations from TEOM data at CAMS-12 site by collection year for all hours, collection (MDCOs deployed) hours, and dust hours (as per weather observations at El Paso International Airport). Concentrations are given in μg/ m 3 , mean ± SD.
Collection year PM 10, All hours PM 10, Collection hours PM 10, Dust hours Peak hourly PM 10 2011-2012 31.0 ± 76.1 a 67.2 ± 212.2 ab 278.9 ± 528.1 a 4739 2012-2013 31.5 ± 59.4 a 66.6 ± 143.3 a 242.9 ± 281.2 a 1955 2013-2014 27.3 ± 47.7 b 68.7 ± 47.7 a 194.7 ± 225.0 a 1467 2014-2015 22.2 ± 20.7 c 25.2 ± 20.5 c 51.9 ± 47.5 b 281 2015-2016 28.7 ± 34.7 b 52.0 ± 107.8 b 116.9 ± 107.8 b 608 Entire collection period 27.9 ± 53.9 c 56.0 ± 140.1 200.2 ± 315.9 4739
Letters in each column indicate significant differences (ANOVA, Tukey range test, P < 0.05).
Mean PM 2.5 concentrations from TEOM data at CAMS-12 site by collection year for all hours, collection (MDCOs deployed) hours, and dust hours (as per weather observations at El Paso International Airport). Concentrations are given in μg/ m 3 , mean ± SD.
Collection year PM 2.5, All hours PM 2.5, Collection hours PM 2.5, Dust hours Peak hourly PM 2.5 2011-2012 7.2 ± 7.9 c 9.0 ± 15.2 b 27.4 ± 47.2 a 414.2 2012-2013 8.0 ± 9.8 c 12.6 ± 21.6 b 37.2 ± 39.5 a 288.3 2013-2014 6.9 ± 7.5 c 9.0 ± 11.7 b 29.3 ± 31.5 a 195.2 2014-2015 6.1 ± 4.7 b 5.5 ± 5.0 c 9.0 ± 5.5 a 37.1 2015-2016 7.8 ± 7.1 c 8.7 ± 8.0 b 19.7 ± 15.5 a 97.1 Entire collection period 7.2 ± 7.6 c 7.8 ± 11.3 b 26.3 ± 33.3 a 414.2
Letters in each column indicate significant differences (ANOVA, Tukey range test, P < 0.05).
Peak hourly PM 10 and PM 2.5 per event from TEOM data at CAMS-12 site averaged over all events for each collection year and entire collection period. Concentrations are given in μg/m 3 , mean ± SD.
Collection Year Average Peak Hourly PM 10 Per Event Average Peak Hourly PM 2.5 Per Event 2011-2012 1204.5 ± 1506 93.5 ± 134.2 2012-2013 723.2 ± 544.5 95.0 ± 75.2 2013-2014 523.8 ± 418.5 62.1 ± 59.0 2014-2015 156.8 ± 62.6 28.7 ± 16.0 2015-2016 287.0 ± 133.2 39.7 ± 22.5 Entire collection period 556.8 ± 707.1 66.3 ± 72 in dusts, sands and soils from arid regions globally (Lawrence and Neff, 2009;Rashki et al., 2013), and these data show El Paso is no exception. Calcite likely occurs as fragments of pedogenic carbonate (caliche) which are present in local surface sands (Instituto Nacional de Estadística, Geografía e Informática, n.d.; Langford, 2000). Dust deposition is the major source of calcium for the formation of pedogenic carbonate in local desert soils (Capo and Chadwick, 1999;Ortiz and Jin, 2021), thus some of this calcium carbonate formed from dust is apparently cycled back into the atmosphere in local windstorms.
Elemental concentrations from ICP-OES analyses of 33 samples, presented as oxide equivalents, are shown in Table 7. Data show large amounts of Si (63%-98% oxide equivalent, mean 76.3%, standard deviation 10.2%) reflecting the dominance of quartz and silicate minerals throughout all samples collected at UTEP. More than 10% Al as Al 2 O 3 oxide equivalent, likely from feldspars and clay minerals, was detected along with 5% CaO likely from calcite, clay minerals, and feldspars. Minor amounts of Fe 2 O 3 , K 2 O, Na 2 O, MgO, TiO 2 , P 2 O 5 and MnO, were also detected, along with trace amounts (<1000 ppm) of Sr, Ba, and Zr.
ICP-OES analyses indicated elements associated with rock-forming minerals predominated in fallen dust and sand (e.g., Si, Al, Ca, Fe, Na, and K), consistent with the XRD results. Compared to averages for global aeolian deposition from Lawrence and Neff (2009), El Paso's samples are higher in Si and lower in Al, Fe, Ti, and Mn. Concentrations of Si in the MDCO samples are consistent with Si content of regional nebkha and sheet sands (Dominguez Acosta et al., 2010). Concentrations of Na, Mg, Al, K, Ca, Ti and Fe are slightly enriched compared to the composition of dune sands within the predominant wind transport pathway from the southwest into El Paso (Dominguez Acosta et al., 2010), while concentrations of Na, Mg, Al, Ti, Mn and Fe are slightly depleted compared to playa "dust hotspot" sediments along the same dominant transport path (Dominguez Acosta and Gill, 2007). The overall elemental composition of bulk aeolian sediments falling at UTEP suggests a dominant contribution from local sands, representing "quartz dilution," the outsized impact of larger particles on bulk sediment chemistry (Putman et al., 2022), with a subsidiary contribution from regional playa and alluvial dust sources. Chihuahuan Desert dust tends to have a whitish to grayish appearance (Fig. 7), lighter in color than dust from other North American dryland regions, likely due to its lower Fe and Mn content.
The relative iron content is a key characteristic of dust from different source regions, due to the ecological, radiative, and human health impacts of iron in atmospheric particles (Al-Abadleh et al., 2023). Dust in the Southern High Plains tends to be reddish (Park et al., 2007) due to its high Fe content, where Fe/Si ratios in total suspended particulate samples in Lubbock approach 0.2 (Gill et al., 2009). Elsewhere, MDCO deposition in northern Arizona averaged 2.9% elemental Fe (Nakase et al., 2014), slightly depleted from global averages, while iron content in windblown sediment fallout on a roof in the Phoenix, Arizona metropolitan area averaged 3.3% (Péwé et al., 1981) and dust trap samples in southern Nevada and southern California averaged approximately 3.5% Fe, very close to the global mean (Reheis et al., 2009). Silt-dominated aeolian deposition into the snowpack of the San Juan Mountains, Colorado, likely originating from the Colorado Plateau, averaged 2.4% Fe (Reynolds et al., 2020), and silt-dominated dust from the Great Basin deposited in snow of the Wasatch Range, Utah, during the April 14-15 dust storm in northern Utah averaged 2.3% Fe (Nicoll et al., 2020), lower than global averages. In comparison, iron content in aeolian deposition measured in El Paso averages lower than that from the aforementioned North American sites.
More than five years of data from passive aeolian deposition traps revealed the characteristics of falling dust and sand during synopticscale wind events in urban El Paso, Texas. Deposition rates were higher than reported for similar collectors in almost all other locations in North America, including surrounding rural sites in the Chihuahuan Desert, but lower than values reported for sites in the Global Dust Belt. Aeolian sediment deposition rates and PM 10 concentrations at an adjacent air monitoring site were highest during a strong drought period, further suggesting the land-atmosphere feedback between drought and dust in the Chihuahuan Desert. The reported rates are underestimates of actual annual deposition, since samples were not obtained during convective dust events (haboobs) or most background (non-dust event) conditions.
El Paso, Texas appears to be one of the dustiest and sandiest cities in North America. The high deposition rates in El Paso may be due to (1) its location near and downwind of major Chihuahuan Desert dust sources, as indicated by back trajectories and remote sensing; and (2) enhancement of dustfall in the urban area due to local sources of fugitive dust and urban topography which enhances deposition. Windstorms advect into El Paso primarily from the southwest and west, crossing erodible sandy soils and numerous dust emission hotspots in the northern Chihuahuan Desert across the international border. Analyses show that the aeolian sediment falling on El Paso, Texas is (1) primarily sand, not dust, such that events could be characterized as sandstorms, not dust storms, even though the collectors in this study were deployed 21 m above ground level, well above the height of saltation; and is (2) similar in mineralogy (dominated by quartz, feldspars, calcite), and (3) broadly similar in major and minor element concentrations to other global dustproducing regions, though depleted in iron.
It has been established that dust events in El Paso are associated with increased hospitalizations for many conditions (Herrera-Molina et al., 2021), especially respiratory diseases, with increased risk to children (Grineski et al., 2011). El Paso residents should be particularly aware of limiting their exposure, especially children's exposure, to the high particulate matter concentrations prevailing in the city during dusty and sandy conditions.
This study has provided important initial data on bulk aeolian deposition in El Paso, Texas, but much more research is needed on this topic. Windblown dust and sand deposition and its characteristics should be measured in other dust-prone regions and cities of North America, such as in the Columbia Plateau and central Great Plains, as well as emerging dust hotspots such as the Salton Sea Basin, to further contextualize deposition rates and physical and chemical properties at the continental scale. Analyses of aeolian deposition samples in El Paso have already shown that windstorms are a method of dispersal of rotifers, nematodes, tardigrades, crustaceans (Rivas et al., 2019), algae and fungi (Rivas et al., 2018); additional analyses could examine which other biota are dispersed with windblown sand and dust in El Paso and other localities. The presence of microplastics in dry deposition (e.g., Wright et al., 2020) is a rising topic of concern: samples in El Paso and other cities should be carefully analyzed for microplastic distribution. Full annual deposition data are needed, including wet deposition and dry deposition during convective storms and "background" (non-windy) periods. More detailed data on the elemental (trace elements, rare earth elements) and chemical (quantitative XRD, clay mineralogy, organic components, etc.) composition of bulk windblown sediment deposition are also needed. Monitoring fallout of dust and sand at different heights above the ground and different locations both within and outside the urban area in dust-prone dryland cities such as El Paso could clarify the amounts and properties of deposited dust and sand emanating from natural desert sources vs. those contributed by fugitive emissions within urban areas. Additional sampling within the metropolitan area would also improve our understanding of aeolian deposition flux relationships to land use and topography, as well as potential differential impacts on different communities and neighborhoods. Finally, correlating event-by-event deposition and its physical and chemical characteristics to trajectories and weather variables (wind speed, wind gust, moisture) would improve our understanding of meteorological impacts on dust and sand deposition.