Application of the gas tracer method for measuring oxygen transfer rates in subsurface flow constructed wetlands

The oxygen transfer rate (OTR) has a significant impact on the design, optimal operation and modelling of constructed wetlands treating wastewater. Oxygen consumption is very fast in wetlands and the OTR cannot be determined using an oxygen mass balance. This problem is circumvented in this study by applying the gas tracer method. Experiments were conducted in an unplanted gravel bed (dimensions L × W × d 125 × 50 × 35 cm filled with a 30-cm layer of 10-11-mm gravel) and a planted horizontal subsurface flow constructed wetland (HSSFCW) (L × W × d 110 × 70 × 38 cm filled with a 30-cm layer of 3.5-mm gravel with Phragmites australis). Tap water saturated with propane as gas tracer (pure or commercial cooking gas, depending on the test) was used. The mass transfer ratio between oxygen and commercial propane gas was quite constant and averaged R = 1.03, which is slightly lower than the value of R = 1.39 that is usually reported for pure propane. The OTR ranged from 0.31 to 5.04 g O₂ m⁻² d⁻¹ in the unplanted gravel bed and from 0.3 to 3.2 g O₂ m⁻² d⁻¹ in the HSSFCW, depending on the hydraulic retention time (HRT). The results of this study suggest that the OTR in HSSFCW is very low for the oxygen demand of standard wastewater and the OTR calculations based on mass balances and theoretical stoichiometric considerations overestimate OTR values by a factor that ranges from 10 to 100. The gas tracer method is a promising tool for determining OTR in constructed wetlands, with commercial gas proving to be a viable low-cost alternative for determining OTR.     

Evaluation of the AISI S213 seismic design procedures through testing of strap braced cold-formed steel walls

The North American standard for the design of lateral systems constructed of cold-formed steel (CFS) (AISI S213), which is published by the American Iron and Steel Institute (AISI), was updated in 2007. Included in this update were changes to the provisions related to strap braced walls. The 2007 version of AISI S213 includes requirements for the brace material and the use of capacity design principles; as well, the engineer is directed towards using welded connections to avoid possible net section fracture of the braces. Screw-connected braces may be used if shown to perform in a ductile fashion. The research described herein was undertaken to evaluate typical weld and screw-connected single-storey strap braced wall configurations with respect to their ability to resist lateral in-plane loads in the inelastic range of behaviour; that is, the results were used to verify the new strap braced wall seismic design provisions in AISI S213. A total of 44 tension-only X-braced walls ranging in size from 610×2440 mm² to 2440×2440 mm² (aspect ratios from 4:1 to 1:1), designed and detailed following a capacity-based approach, were tested under lateral loading using monotonic and reversed cyclic protocols. The strap braces were expected to undergo gross cross-section yielding along their length, while the other elements in the seismic force resisting system (SFRS) were selected to be able to carry the probable brace capacity. A summary of the test program is provided, including design approach, failure modes and ductility measures. The scope of the research also included the determination of “test-based” seismic force modification factors based on the measured ductility and overstrength of the test walls for comparison with the R-values recommended in AISI S213.     

Empirical evaluation of ductility factors for the special steel moment‐resisting frames in view of soil condition

It is well known that the response modification factor (R) takes into account the ductility, over‐strength, redundancy and damping of structural systems. The ductility factor has played an important role in seismic design, as it is a key component of R. In this study, the ductility factors (R_μ,MDOF) of special steel moment‐resisting frames are calculated by multiplying the ductility factor of single degree of freedom (SDOF) systems (R_μ,SDOF) with the multi‐degree of freedom (MDOF) modification factors (R_M). The ductility factors (R_μ,SDOF) of SDOF systems are computed from non‐linear dynamic analysis undergoing different levels of displacement ductility demands and periods when subjected to a large number of recorded earthquake ground motions. To compute the R_μ,SDOF, a group of 1,860 ground motions recorded from 47 earthquakes were considered. R_M factors are proposed to account for the MDOF systems, based on previous studies. A total of 108 prototype steel frames were designed to investigate the ductility factors, considering design parameters such as the number of stories (4, 8 and 16), framing systems (perimeter frames and distributed frames), failure mechanisms (strong column‐weak beam and weak column‐strong beam), soil profiles (S_A, S_C and S_E in Uniform Building Code 1997) and seismic zone factors (Z = 0·075, 0·2, and 0·4 in UBC 1997). The effects of these design parameters on the R_μ,MDOF of special steel‐moment‐resisting frames were investigated. Copyright © 2009 John Wiley & Sons, Ltd.     

Innovative self-powered submersible microbial electrolysis cell (SMEC) for biohydrogen production from anaerobic reactors

A self-powered submersible microbial electrolysis cell (SMEC), in which a specially designed anode chamber and external electricity supply were not needed, was developed for in situ biohydrogen production from anaerobic reactors. In batch experiments, the hydrogen production rate reached 17.8 mL/L/d at the initial acetate concentration of 410 mg/L (5 mM), while the cathodic hydrogen recovery (RH₂) and overall systemic coulombic efficiency (CE_os) were 93% and 28%, respectively, and the systemic hydrogen yield (YH₂) peaked at 1.27 mol-H₂/mol-acetate. The hydrogen production increased along with acetate and buffer concentration. The highest hydrogen production rate of 32.2 mL/L/d and YH₂ of 1.43 mol-H₂/mol-acetate were achieved at 1640 mg/L (20 mM) acetate and 100 mM phosphate buffer. Further evaluation of the reactor under single electricity-generating or hydrogen-producing mode indicated that further improvement of voltage output and reduction of electron losses were essential for efficient hydrogen generation. In addition, alternate exchanging the electricity-assisting and hydrogen-producing function between the two cell units of the SMEC was found to be an effective approach to inhibit methanogens. Furthermore, 16S rRNA genes analysis showed that this special operation strategy resulted same microbial community structures in the anodic biofilms of the two cell units. The simple, compact and in situ applicable SMEC offers new opportunities for reactor design for a microbial electricity-assisted biohydrogen production system.     

Development of an empirical model for determining G*/sin δ in crumb rubber modified binders

This paper describes the development of an empirical model which may be used for predicting the G*/sin δ for neat and crumb rubber modified (CRM) binders. The model was developed using 36 unique CRM binder combinations, crumb rubber concentrations were varied at 5% intervals between 5% and 20%. The effects of crumb rubber particle size on model accuracy were also studied, ultimately a model was produced with the capability of predicting G*/sin δ values over a range of temperatures and crumb rubber concentrations. By definition, the upper limit of the performance grade is dependent on the G*/sin δ value; therefore, the relationship was also considered in terms of high end failure temperature.The rubber coefficient for G*/sin δ (R_cg) was identified as an important parameter in the estimation of G*/sin δ in addition to the CRM. This term is a quantitative representation of the increase typically witnessed in G*/sin δ values with the addition of CRM. Ambient ground CRM exhibited higher R_cg values than cryogenically ground particles. Additionally, 95% confidence intervals were generated for the predictive model, thus providing a range of accuracy for the model. The resulting confidence intervals were approximately ±1300 Pa, these confidence intervals were seen to capture 92.6% of the 462 data points used. Findings from this research suggest that the differences between cryogenic and ambient CRM binder are accurately described using the R_cg, furthermore binder properties may be predicted using an empirical equation.     

Fate of aniline and sulfanilic acid in UASB bioreactors under denitrifying conditions

Two upflow anaerobic sludge blanket (UASB) reactors were operated to investigate the fate of aromatic amines under denitrifying conditions. The feed consisted of synthetic wastewater containing aniline and/or sulfanilic acid and a mixture of volatile fatty acids (VFA) as the primary electron donors. Reactor 1 (R1) contained a stoichiometric concentration of nitrate and Reactor 2 (R2) a stoichiometric nitrate and nitrite mixture as terminal electron acceptors. The R1 results demonstrated that aniline could be degraded under denitrifying conditions while sulfanilic acid remains. The presence of nitrite in the influent of R2, caused a chemical reaction that led to immediate disappearance of both aromatic amines and the formation of an intense yellow coloured solution. HPLC analysis of the influent solution, revealed the emergence of three product peaks: the major one at retention time (R_t) 14.3 min and two minor at R_t 17.2 and 21.5 min. In the effluent, the intensity of the peaks at R_t 14.3 and 17.2 min was very low and of that at R_t 21.5 min increased (∼3-fold). Based on the mass spectrometry analysis, we propose the structures of some possible products, mainly azo compounds. Denitrification activity tests suggest that biomass needed to adapt to the new coloured compounds, but after a 3 days lag phase, activity is recovered and the final (N₂ + N₂O) is even higher than that of the control.     

Experimental study of high strength concrete-filled circular tubular columns under eccentric loading

The paper describes 37 tests conducted on slender circular tubular columns filled with normal and high strength concrete subjected to eccentric axial load. The test parameters were the nominal strength of concrete (30, 70 and 90 MPa), the diameter to thickness ratio D/t, the eccentricity ratio e/D and the column slenderness (L/D). The experimental ultimate load of each test was compared with the design loads from Eurocode 4, which limits the strength of concrete up to 50 MPa. The aim of the paper is to establish the advisability of the use of high strength concretes as opposed to that of normal strength concretes by comparing three performance indices: concrete contribution ratio, strength index and ductility index. The results show for the limited cases analyzed that the use of high strength concrete for slender composite columns is interesting since this achieves ductile behavior despite the increase in load-carrying capacity is not greatly enhanced.     

New approach to evaluate the response modification factors for steel moment resisting frames

The design force levels currently specified by most seismic codes are calculated by dividing the base shear for elastic response by the response modification factor (R). This is based on the fact that the structures possess significant reserve strength, redundancy, damping and capacity to dissipate energy. This paper proposed the evaluation methodology and procedure of the response modification factors for steel moment resisting frames. The response modification factors are evaluated by multiplying ductility factor (R _μ) for SDOF systems, MDOF modification factor (R _M ) and strength factor (R _S ) together. The proposed rules were applied to existing steel moment resisting frames. The nonlinear static pushover analysis was performed to estimate the ductility (R _μ), MDOF modification (R _M ) and strength factors (R _S ). The results showed that the response modification factors (R) have different values with various design parameters such as design base shear coefficient (V/W), failure mechanism, framing system and number of stories.     

Efficiency of phenol biodegradation by planktonic Pseudomonas pseudoalcaligenes (a constructed wetland isolate) vs. root and gravel biofilm

In the last two decades, constructed wetland systems gained increasing interest in wastewater treatment and as such have been intensively studied around the world. While most of the studies showed excellent removal of various pollutants, the exact contribution, in kinetic terms, of its particular components (such as: root, gravel and water) combined with bacteria is almost nonexistent.In the present study, a phenol degrader bacterium identified as Pseudomonas pseudoalcaligenes was isolated from a constructed wetland, and used in an experimental set-up containing: plants and gravel. Phenol removal rate by planktonic and biofilm bacteria (on sterile Zea mays roots and gravel surfaces) was studied. Specific phenol removal rates revealed significant advantage of planktonic cells (1.04 × 10⁻⁹ mg phenol/CFU/h) compared to root and gravel biofilms: 4.59 × 10⁻¹¹-2.04 × 10⁻¹⁰ and 8.04 × 10⁻¹¹-4.39 × 10⁻¹⁰ (mg phenol/CFU/h), respectively.In batch cultures, phenol biodegradation kinetic parameters were determined by biomass growth rates and phenol removal as a function of time. Based on Haldane equation, kinetic constants such as μ_max = 1.15/h, K_s = 35.4 mg/L and K_i = 198.6 mg/L fit well phenol removal by P. pseudoalcaligenes.Although P. pseudoalcaligenes planktonic cells showed the highest phenol removal rate, in constructed wetland systems and especially in those with sub-surface flow, it is expected that surface associated microorganisms (biofilms) will provide a much higher contribution in phenol and other organics removal, due to greater bacterial biomass.Factors affecting the performance of planktonic vs. biofilm bacteria in sub-surface flow constructed wetlands are further discussed.     

Determination of sorption of seventy-five pharmaceuticals in sewage sludge

Sorption of 75 active pharmaceutical ingredients (APIs) to three different types of sludge (primary sludge, secondary sludge with short and long sludge age respectively) were investigated. To obtain the sorption isotherms batch studies with the APIs mixture were performed in four nominal concentrations to water containing 1 g of sludge. The range of APIs concentrations was between ng L⁻¹ to μg L⁻¹ which are found in the wastewater effluents. Isotherms were obtained for approximately 45 of the APIs, providing distribution coefficients for linear (K_d), Freundlich (K_f) and Langmuir (K_L) isotherms. K_d, K_f and K_L ranging between 7.1 × 10⁴ and 3.8 × 10⁷, 1.1 × 10⁻² and 6.1 × 10⁴ and 9.2 × 10⁻³ and 1.1 L kg⁻¹, respectively. The obtained coefficients were applied to estimate the fraction of APIs in the water phase (see Abstract Graphic). For 37 of the 75 APIs, the predicted presence in the liquid phase was estimated to >80%. 24 APIs were estimated to be present in the liquid phase between 20 and 80%, and 14 APIs were found to have <20% presence in the liquid phase, i.e. high affinity towards sludge. Furthermore, the effect of pH at values 6, 7 and 8 was evaluated using one way ANOVA-test. A significant difference in K_ds due to pH changes were found for 6 of the APIs (variation 10-20%).     

Tests on high-strength rectangular concrete-filled steel hollow section stub columns

An experimental programme on high-strength rectangular concrete-filled steel hollow section columns is reported. A total of 22 specimens were tested under concentric loading. The test variables include the material strengths (f_y=495 MPa; and 89 MPa), cross-sectional aspect ratio (1.0-2.0) and volumetric steel-to-concrete ratio (0.13-0.17). Favourable ductility performance was observed for the specimens in the tests. The experimental results are employed to calibrate the specifications in the codes (EC4, ACI and AISC) for the design of high-strength composite columns. Comparison indicates that the specifications in ACI and AISC conservatively estimate the failure loads of the specimens by 9 and 11%, respectively. The EC4 method gives a close and conservative estimate of the ultimate capacities with a difference of 1%.     

Modeling approaches and performance for estimating personal exposure to household air pollution: A case study in Kenya

This study assessed the performance of modeling approaches to estimate personal exposure in Kenyan homes where cooking fuel combustion contributes substantially to household air pollution (HAP). We measured emissions (PM_2.5, black carbon, CO); household air pollution (PM_2.5, CO); personal exposure (PM_2.5, CO); stove use; and behavioral, socioeconomic, and household environmental characteristics (eg, ventilation and kitchen volume). We then applied various modeling approaches: a single-zone model; indirect exposure models, which combine person-location and area-level measurements; and predictive statistical models, including standard linear regression and ensemble machine learning approaches based on a set of predictors such as fuel type, room volume, and others. The single-zone model was reasonably well-correlated with measured kitchen concentrations of PM_2.5 (R² = 0.45) and CO (R² = 0.45), but lacked precision. The best performing regression model used a combination of survey-based data and physical measurements (R² = 0.76) and a root mean-squared error of 85 µg/m³, and the survey-only-based regression model was able to predict PM2.5 exposures with an R² of 0.51. Of the machine learning algorithms evaluated, extreme gradient boosting performed best, with an R² of 0.57 and RMSE of 98 µg/m³.     

Mapping of background air pollution at a fine spatial scale across the European Union

Background

There is a need to understand much more about the geographic variation of air pollutants. This requires the ability to extrapolate from monitoring stations to unsampled locations. The aim was to assess methods to develop accurate and high resolution maps of background air pollution across the EU.

Methods

We compared the validity of ordinary kriging, universal kriging and regression mapping in developing EU-wide maps of air pollution on a 1 × 1 km resolution. Predictions were made for the year 2001 for nitrogen dioxide (NO₂), fine particles < 10 µm (PM₁₀), ozone (O₃), sulphur dioxide (SO₂) and carbon monoxide (CO) using routine monitoring data in Airbase. Predictor variables from EU-wide databases were land use, road traffic, population density, meteorology, altitude, topography and distance to sea. Models were developed for the global, rural and urban scale separately. The best method to model concentrations was selected on the basis of predefined performance measures (R², Root Mean Square Error (RMSE)).

Results

For NO₂, PM₁₀ and O₃ universal kriging performed better than regression mapping and ordinary kriging. Validation of the final universal kriging estimates with results from all validation sites gave R²-values and RMSE-values of 0.61 and 6.73 µg/m³ for NO₂; 0.45 and 5.19 µg/m³ for PM₁₀; and 0.70 and 7.69 µg/m³ for O₃. For SO₂ and CO none of the three methods was able to provide a satisfactory prediction.

Conclusion

Reasonable prediction models were developed for NO₂, PM₁₀ and O₃ on an EU-wide scale. Our study illustrates that it is possible to develop detailed maps of background air pollution using EU-wide databases.     

Response modification factor for steel moment-resisting frames by different pushover analysis methods

The earthquake loads imposed to the structures are generally much more than what they are designed for. This reduction of design loads by seismic codes is through the application of response modification factor (R-factor). During moderate to severe earthquakes, structures usually behave inelastically, and therefore inelastic analysis is required for design. Inelastic dynamic analysis is time consuming and interpretation of its results demands high level of expertise. Pushover analysis, recently commonly used, is however, a simple way of estimating inelastic response of structures. Despite its capabilities, conventional pushover analysis (CPA) does not account for higher mode effects and member stiffness changes. Adaptive pushover analysis (APA) method however, overcomes these drawbacks. This research deals with derivation and comparison of some seismic demand parameters such as ductility based reduction factor, R_μ, overstrength factor, Ω, and in particular, response modification factor, R, from capacity curves obtained from different methods of APA and CPA. Three steel moment-resisting frames of 3, 9 and 20 stories adopted from SAC steel project are analyzed. In pushover analyses for each frame, eight different constant as well as adaptive lateral load patterns are used. Among the main conclusions drawn is that the maximum relative difference for response modification factors was about 16% obtained by the methods of conventional and adaptive pushover analyses.     

Synergistic effect of nickel ions on the coupled dechlorination of trichloroethylene and 2,4-dichlorophenol by Fe/TiO₂ nanocomposites in the presence of UV light under anoxic conditions

The coupled removal of priority pollutants by nanocomposite materials has recently been receiving much attention. In this study, trichloroethylene (TCE) and 2,4-dichlorophenol (DCP) in aqueous solutions were simultaneously removed by Fe/TiO₂ nanocomposites under anoxic conditions in the presence of nickel ions and UV light at 365 nm. Both TCE and DCP were effectively dechlorinated by Fe/TiO₂ nanocomposites, and the pseudo-first-order rate constants (k_obs) for TCE and DCP dechlorination were (1.39 ± 0.05)×10⁻² and (1.08 ± 0.05)×10⁻² h⁻¹, respectively, which were higher than that by nanoscale zerovalent iron alone. In addition, the k_obs for DCP dechlorination was enhanced by a factor of 77 when Fe/TiO₂ was illuminated with UV light for 2 h. Hydrodechlorination was found to be the major reaction pathway for TCE dechlorination, while DCP could undergo reductive dechlorination or react with hydroxyl radicals to produce 1,4-benzoquinone and phenol. TCE was a stronger electron acceptor than DCP, which could inhibit the dechlorination efficiency and rate of DCP during simultaneous removal processes. The addition of nickel ions significantly enhanced the simultaneous photodechlorination efficiency of TCE and DCP under the illumination of UV light. The k_obs values for DCP and TCE photodechlorination by Fe/TiO₂ in the presence of 20-100 μM Ni(II) were 30.4-136 and 13.2-192 times greater, respectively, when compared with those in the dark. Electron spin resonance analysis showed that the photo-generated electron-hole pairs could be effectively separated through Ni ions cycling, leading to the improvement of electron transfer efficiency of TCE and DCP by Fe/TiO₂.     

Bio-reaction of nitrobenzene with Microcystis aeruginosa: characteristics, kinetics and application

The bio-reaction of nitrobenzene (NB) with Microcystis aeruginosa was investigated at different initial algal densities and NB concentrations by performing static experiments. The results showed that the elimination of NB was enhanced by the bio-reaction, and the reaction rate varied as a function of the reaction time. Moreover, the reaction rate was significantly affected by the algal density and NB concentration. A kinetic analysis showed that the elimination of NB in a solution without algae appeared to be pseudo-first-order with respect to the NB concentration, whereas a first-order model was too oversimplified to describe the elimination of NB in a solution with algae. Assuming that different algal cells have the same effect on the bio-reaction under the same conditions, the bio-reaction can be described as dC_NB = −k₀C_A^mA_NBⁿdt (where k₀ is the reaction rate constant, C_A is the algae density and C_NB is the concentration of NB). When the growth of algae was not considered, the value of k₀, m and n were 8.170 × 10⁻⁴, 0.5887 and 1.692, respectively. Alternatively, when algae were in the exponential growth phase, the value of k₀, m and n were 1.6871 × 10⁻⁵, 0.7248 and 2.5407, respectively, according to a nonlinear fitting analysis. The kinetic model was also used to elucidate the effect of nutritional limitation on the bio-reaction.     

A photoelectrocatalytic process that disinfects water contaminated with Mycobacterium kansasii and Mycobacterium avium

Nontuberculous mycobacteria are resistant to conventional water treatment; indeed, they have been recovered from a wide variety of environmental sources. Here, we applied the photoelectrocatalytic technique using a Ti/TiO₂–Ag photoanode to inactivate mycobacteria. For a mycobacteria population of 5 × 10⁸ CFU mL⁻¹, we achieved 99.9 and 99.8% inactivation of Mycobacterium kansasii and Mycobacterium avium with rate constant of 6.2 × 10⁻³ and 4.2 × 10⁻³ min⁻¹, respectively, after 240 min. We compared the proposed method with the photolytic and photocatalytic methods. Using a mycobacteria population of 7.5 × 10⁴ CFU mL⁻¹, the proposed Ti/TiO₂–Ag photoanode elicited total mycobacteria inactivation within 3 min of treatment; the presence of Ag nanoparticles in the electrode provided 1.5 larger degradation rate constant as compared with the Ti/TiO₂ anode (1.75 × 10⁻² for M. kansassi and 1.98 × 10⁻² for M. avium). We monitored the degradation of the metabolites released during cellular lysis by TOC removal, sugar release, chromatography, and mass spectrometry measurements; photoelectrocatalysis and Ti/TiO₂–Ag photoanodes furnished the best results.     

Applicability of the Small Diameter Sampler for Niigata Sand Deposits

Applicability of tube sampling for Niigata sand deposits is discussed through bender element and cyclic triaxial tests for samples obtained from two-chambered hydraulic piston samplers (Shogaki, 1997) with inner diameters of 48 mm and 50 mm, a one-chambered 70 mm diameter sampler, a 125-mm rotary triple-tube sampler and the frozen (FS) sampling method (Yoshimi et al., 1989). The relationship between the relative density (D_r) and normalized SPT N-value (N₁) obtained from small diameter samplers with inner diameters of 45 mm and 50 mm samplers was close to that of the FS and the N₁ coefficient was greater than those of the 70-mm and other tube samplers. The stress ratio at 20 cycles (R_L20) and the initial modulus of rigidity (G_CTX) of samples obtained from the 45-mm and 50-mm samplers were greater than those of the 70-mm, 125-mm rotary triple-tube and other tube samplers. However, the R_L20 values obtained from the 45-mm and 50-mm samplers were smaller than those of the FS sampler in the area of N₁>24. The G_BE and G_CTX values obtained from the 45-mm and 50-mm samplers were close to those of the FS sampling. Therefore, the 45-mm and 50-mm samplers can take equally high quality samples for Niigata sand deposits.     

Quantification of seismic performance factors for self‐centering controlled rocking special concentrically braced frame

Modern self‐centering controlled rocking special concentrically braced frame (SC‐CR SCBF) is capable of reducing structural damage compared with conventional buildings following an earthquake. This investigation quantifies three seismic performance factors, including over‐strength factor (Ω₀), period‐based ductility (μ_T) and response modification coefficient (R), for low‐ and mid‐rise SC‐CR SCBFs. Nonlinear static analysis is conducted to derive Ω₀ and μ_T factors for 12 SC‐CR archetypes. Validity of trial R coefficient is also evaluated using a collapse‐based assessment procedure by comparing adjusted collapse margin ratios with the established acceptance criteria. Results indicate that the Ω₀ and μ_T factors are in the range of 1.39 to 2.29 and 12.25 to 29.0, respectively, and R of 8 is proposed for design of SC‐CR archetypes. A reliability study is also performed to examine the effects of modeling and ground motion parameters on the safety margin of designed SC‐CR archetypes with the proposed R value. Results indicate that the design of mid‐rise space archetypes in high‐seismicity regions with the R coefficient of 8 is more reliable than that of the low‐rise perimeter ones in low‐seismicity regions. Copyright © 2016 John Wiley & Sons, Ltd.     

Improving spatial concentration estimates for nitrogen oxides using a hybrid meteorological dispersion/land use regression model in Los Angeles, CA and Seattle, WA

Predictions from a simple line source dispersion model, Caline3, were included as a covariate in a land use regression (LUR) model for NO_X/NO₂ in Los Angeles, CA and Seattle, WA. The Caline3 model prediction assumed a unit emission factor for all roadway segments (1.0 g/vehicle-mile). The NO_X and/or NO₂ measurements for LA and Seattle were obtained from a comprehensive measurement campaign that is part of the Multi-Ethnic Study of Atherosclerosis Air Pollution Study (MESA Air). The measurement campaigns in both cities were approximately 2 weeks in duration employing approximately 145 measurement sites in Greater LA and 26 sites in Seattle. The best “standard” LUR model (obtained without the inclusion of the Caline3 predictions) in LA had R² values of 0.53 for NO_X and 0.74 for NO₂. The leave-one-out cross-validated R² values for NO_X and NO₂ were 0.45 and 0.71, respectively. The equivalent “standard” NO₂ model for Seattle had an R² of 0.72 and a leave-one-out cross-validated R² of 0.63. When the Caline3 variable was included in the LA hybrid model, the R² values were 0.71 and 0.79 for NO_X and NO₂, respectively. The corresponding cross-validated R² values were 0.66 and 0.77, for NOX and NO2, respectively. In Seattle, the inclusion of the Caline3 variable resulted in a NO₂ model with an R² of 0.81 and a corresponding cross-validated R² of 0.67. In LA, hybrid model performance was not affected by excluding roadways with annual average daily traffic volumes (AADT) < 100,000. When the Caline3 predictions for heavy-duty trucks and lighter-duty vehicles were modelled as separate terms, the estimated fleet average NO_X emission factors were 8.9 (SE = 0.7) and 0.16 (SE = 0.12) grams NO_X/vehicle mile for heavy-duty and lighter-duty vehicles, respectively. These values are consistent with fleet average emission factors computed for LA with EMFAC 2007.