مدل‎سازي مكاني ـ زماني سه‎بعدي پراكنش آلاينده‌ي اكسيدهاي ازت هوا ناشي از ترافيك در تقاطع خيابان ولي‌عصر ـ فاطمي شهر تهران

3D Spatio-Temporal Modeling of NOx Air Pollution of Vehicular Traffic in Vali-e-Asr and Fatemi Streets Intersection, Tehran City

يكي از مشكلات اصلي شهرها افزايش سريع آلودگي هواست كه ترافيك وسايل نقليه يكي از مهم ترين عوامل آن به شمار مي رود. مديريت هدفمند اين عامل آلوده كننده نيازمند اطلاعات صحيح و دقيق از نحوه ي انتشار آلاينده ها در شرايط گوناگون مكاني و زماني است. در اين باره، پژوهش حاضر نحوه ي انتشار سه بعدي آلودگي حاصل از اكسيدهاي نيتروژن (NOx) را در مقياس ميكرو بررسي و با استفاده از مدل GRAL محدوده ي تقاطع ولي عصر ـ فاطمي شهر تهران را در فصل زمستان مطالعه مي كند. با توجه به اين، خودروها به منزله ي مهم ترين عامل آلاينده به مدل معرفي شد و فرآيند مدل سازي در نُه ارتفاع متفاوت (از 1.7 تا 52.5 متري) انجام گرديد. براي بررسي ويژگي هاي فضايي و زماني داده هاي ميزان غلظت آلاينده ي NOx از روش هاي خودهم بستگي فضايي عمومي و محلي موران استفاده گرديد. ميزان شاخص موران معادل 0.7 تا 0.9 در حالت دوبعدي و معادل 0.22 در حالت سه بعدي در نتايج حاصل نشان دهنده ي وجود سطح بالايي از خود هم بستگي فضايي مثبت معنادار است كه گواه صحت عملكرد شبيه سازي صورت گرفته است. تحليل شاخص موران محلي/انسلين نشان دهنده ي غلبه ي نقاط بالا ـ بالا در ارتفاعات پايين تا متوسط و افزايش نقاط پايين ـ پايين در ارتفاع هاي بالاتر است. همچنين، وجود خوشه هاي آلودگي نسبتاً پايدارتر در ارتفاع هاي مختلف در تقاطع ها و ناپايداري وضعيت خوشه بندي آلودگي هوا در نزديكي ساختمان ها در نتايج حاصل مشهود است.

Air pollution has become one of the main problems of cities. Among the sources of air pollution, vehicular traffic plays an important role. Planning for efficient management and control of the air pollution caused by vehicular traffic requires accurate information on spatiotemporal dispersion of the pollutions. This research studies 3D spatiotemporal dispersion of NOx pollution caused by vehicular traffic at ValieasrFatemi intersection resides in Tehran, Iran. It is selected for being crowded and having the required meteorological and pollution data sensed by the Air Quality Control Corp. of Tehran Municipality. This study uses GRAL that is a local microscale air dispersion model defined based on EuleranLagrangian dispersion models. It investigates the level of spatiotemporal autocorrelation generated by GRAL simulations at both 2D and 3D modes and discusses how it adapts with the reality. Adopting the GRAL air pollution dispersion model, streets are defined as the linear source of pollution of NOx caused by vehicular traffic. The traffic rate is estimated based on street areas and directions, the designed average traffic velocity, traffic volume and car passage counting at the intersection. The 3D geometry of the buildings is also added to the model. All the required data that were available for winter of 2007 are gathered and introduced into the model. The model is executed at 9 heights vary from 1.7 m to 52.5 m. These heights are defined covering a range from an average human level height to average building height and above. These levels are considered both separately in 2D mode and integrated into a 3D mode. The formation of NOx clusters is investigated analyzing their autocorrelation using Moran Index at global and local scale. The calculated MoranI at global scale at each 9 levels of heights, varies from 0.7 to 0.9 that depicts the validity of the GRAL model adopted to simulate the expected autocorrelation of pollution density affected by spatial issues. The MoranI increases at higher levels as less air turbulence happens. However the result show that the turbulence increases temporarily at about 10m to 15m which are the average building heights. At local scale, the MoranI/Anselin shows that HH clusters dominate at lower levels, around streets central areas that are farther from the buildings, and around the intersections. At higher levels, esp. higher than buildings average height, the LL clusters dominate. However the HH clusters formed around intersections, while are shrank, are still visible at high levels. The turbulence caused by building fronts and their down wash effect is also shown in the result as no definite cluster is formed near the buildings front and back. The autocorrelation analysis is also carried for an integrated 3D model consists of all the 9 levels of heights. Considering the weight matrix for a 20m 2D neighborhood and 1m/s dispersion of the pollution vertically, the global calculated MoranI equals 0.229 which shows existence of a spatiotemporal autocorrelation of the results generated by GRAL. At local scale the results show that the HH clusters have higher temporal dispersion rate than LL clusters.