Measurements and Prediction of Particulate Number Concentrations and their Chemical Composition over Yanbu Industrial City, Saudi Arabia

  • Jaafar Al-Mahmodi

Student thesis: Phd


Many recent studies have highlighted the substantial health-related impacts of particle number (PMno) rather than particle mass. The aim of this study is to determine the correlation of trace gases with PMno, to identify the chemical composition of particle different sizes and to predict the NOx and PMno over Yanbu Industrial City (YIC). Trace gases (NOx, SO2, H2S, O3, CO), PMno with diameter (7nm-10micro metre), traffic and meteorological parameters were measured at three sampling sites in YIC. The maximum PMno (333,971 cm-3) at downwind site#1 was about 2.5 times higher than that (123,842 cm-3) at upwind site#2 and about 1.2 times higher than that (263,572 cm-3) at downwind site#3. The average PMno distribution at downwind sites consisted of one distinguishable mode (nucleation mode0.7) are generally stronger than with NO2 at sites#1 and 2, whereas for site #3 the correlation between PMno with NO2/NOx are better than with NO. PMno has generally either weak or poor correlation with SO2 and CO, respectively. Particle samples of different sizes (7nm-10micro metre) were chemically analysed using an ion chromatograph (IC) for inorganic ions and inductively coupled plasma mass spectrometry (ICP-MS) for trace metals at site#3. The ionic analysis revealed that sulfate and ammonium was mainly present in particle of size < 0.38micro metre while nitrate and chloride was mainly present in particles of size > 0.38micro metre. Non-sea salt sulfate was dominant in all particle sizes compared to the marine sulfate which is minor. The total sulfate and nitrate contributed 50.3% and 24.4% of the total ionic mass respectively followed by chloride (13.3%) and ammonium (10.6%). The trace-metals analysis results indicated that Na represented more than 94% of the total mass and the contributions of the remaining metals (Al, Sr, Zn, V, Cr, Fe, etc) were about 6%. A further part of this study consisted of the coupling of the WRF/CALMET system with the CALPUFF model, which was applied to predict NOx and PMno concentrations. The WRF model was employed to generate the meteorological input data for CALMET. WRF predictions were evaluated with surface data and upper air profiles using RASS/SODAR and radiosondes. WRF tends to underestimate the surface temperature on average with biases of up to -3.4oC and also underestimates temperature profiles with average biases ranging between -2.7 and -5.2oC when compared to the RASS profiler, but with a lower bias (< -2.4oC) when compared to radiosonde profiles. The mean wind speed bias for the majority of the cases was close to the benchmark of ±0.5m/s, but the mean wind direction bias for half of the cases exceeded the benchmark of 10o. It was concluded that WRF predictions can be used for air dispersion modeling to produce reasonable outputs. NOx predictions by CALPUFF showed that the contribution of the traffic to the highest concentrations during the nighttime was up to 80%, but after sunrise the contribution from industries became higher (up to 70%). The highest predicted NOx concentration (~313μg/m3) was much lower than the national ambient standard (660μg/m3) and the community area is affected much by industries during mid-morning hours when the wind shifting from land breeze to sea breeze. The fractional bias (FB) ranged between -0.1 and 1.06 indicating that the model tends to under-predict the NOx observations. PMno predictions of two sizes (7-40nm and 7nm-10micro metre) were derived based on the NOx predictions. All FB values were ranged between -0.1 and 0.5. It was concluded that PMno predictions were generally better than those of the NOx due mainly to adding the background term (intercept) for the PMno predictions.
Date of Award31 Dec 2011
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorThomas Choularton (Supervisor)


  • Particle Number, Particle Mass, NOx, Road Traffic, Industries, Chemical Composition
  • Weather prediction, WRF, CALPUFF, NOx Prediction

Cite this