Research on Nanjing’s PM2.5 Pattern
Research on Nanjing’s PM2.5 Pattern
1.Background Information**
In recent years, China’s economy has experienced a huge boost. Along with improved living standards and increasing income, air pollution emerges as a major environmental problem. PM2.5, a minuscule and invisible pollutant, poses direct threat to human health. Therefore, this study chooses a specific target, Nanjing, as a typical example of eastern Chinese cities with prosperous economy to evaluate the extent of PM2.5 pollution in mainland China.
2. Data Source and Standard
In this research, data about PM2.5 value are extracted from measurements of China Meteorological Administration (CMA). This research uses PM2.5 measurements every day in 2018 from nine different monitoring stations located in Nanjing urban area. Using official standard of PM2.5 measurement (excellent:0-50, good:50-100, lightly polluted:100-150, moderately polluted:150-200, heavily polluted:200-300, dangerously polluted:>300), the average measurement of every month is calculated in Excel for every monitoring station.
3. PM2.5 Distribution Pattern
a) Monthly Pattern
After that, in software ArcGis, utilizing common kringing interpolation scheme, the research produces 12 pieces of picture that represent the overall PM2.5 distribution in Nanjing for every month in 2018. For March and September, the kringing interpolation scheme did not generate an effective picture because the overall pollution level is relatively low that the nine monitoring stations generated quite similar values. The following result shows the other 10 months’ value, which refers to Figure 3.1
b) Yearly Pattern
Then, by dividing the whole picture into numerous units of about 150000, from the calculated values of 10 months, the research derives a data table that contains both the location information (longitude, latitude) and the interpolation value of every point on the graph.
To further interpret the pollution level and range, the next step is to accomplish clustering analysis. This operation is done in R language. The method applied, in this process, is called Self-organizing Map (SOM) or Self-organizing Feature Map (SOFM), which is supported by R-package Kohonen. The basic principle involved, for this method, is to match every data point with the weighted vector that randomly generated. In addition, every modification in weights of vector, in order to fit closely with the general pattern of data points, is called training. Finally, according to which weighted vector they assemble, data points were divided into 16 groups. The following pictures are the training process (Figure 3.2) and counts plot (Figure 3.3). The graph for training process shows how the weighted vectors change their weight to match data points perfectly; the graph for counts plot, on the other hand, shows which weighted vectors are perfect representations of the data points after long process of training and value modification.
What’s next, in this research, is to derive the general pollution pattern of Nanjing for the whole year of 2018 (Figure 3.4). To achieve the final result, the prerequisite is to generate a graph with 16 groups of data points clearly on it. After matching with the location information of every point, the clustered data could be exported into ArcGIS for producing such graph. The part that shows scarlet is the area most seriously polluted as redness is proportional with the pollution.
4**. Research Findings**
From the analysis above, there are several research findings. From monthly record, we can clearly see that the overall pollution pattern seems to shrink in June and July, especially in July. This should be probably attributed to the climate type known as subtropical monsoon climate because this specific climate pattern would bring about a precipitation phenomenon known as “rainy season”. As precipitation level increases sharply, the corresponding rainfall would immerse most of PM2.5, which will finally deposit onto the ground. However, for other months, the overall pollution level seems to be relatively high. In winter months, such as December, January, and February, since monsoons from Siberia provide some conditions for spreading PM2.5, the general pollution area is not that broad; besides these months, the PM2.5 pollution took up most of the areas in Nanjing. Then, from the pollution graph for a whole year of 2018, something interesting is revealed. The part that showed most redness is exactly the location of downtown area in Nanjing. Therefore, it is easy to illustrate that skyscrapers, high population density, and heavy traffic would all contribute their power in the pollution level of PM2.5.
5. Suggestions and Further Steps
Considering all the factors that give a rise to the pollution level of PM2.5, the first and foremost step is to detect the level of pollution precisely. To achieve this goal, it is necessary to locate all the monitoring stations in the right place, which means some stations currently out of the serious pollution region should be removed, while it is important to build more stations in areas with relatively higher concentration of PM2.5. What’s more, to ensure the spread of PM2.5 pollutants, some high buildings should be removed. If they can not be removed, the distances between two of them must reach a standard to preserve a space for sufficient exchange of air flow, thereby takes away pollutants like PM2.5. Finally, asking people to stop using personal cars is an approach to alleviate any kind of artificial interference on the level of PM2.5. Besides the natural level of PM2.5, the reason why cities often owns a higher amount of pollutant lies on the additional sources of PM2.5 in cities, which refers to automobiles. This study also needs further work. On the one hand, since the research only investigates Nanjing, a wider places involved in the experiment should be considered to reach a more accurate, more general conclusion. On the other hand, if I can reach data includes every hours’ measurements, the spreading process of pollutant PM2.5 will be clearly shown. In this scenario, some detailed precautions and solutions of PM2.5 pollution can be derived based on the features of the spreading process.