Study on the relationship between airborne pollen concentration and vegetation spatial structure in Beijing in spring based on CART decision tree model

  Objective  The research on the relationship between airborne pollen concentration and vegetation spatial structure, aiming to summarize the vegetation spatial structure under different pollen concentration sections, would be applied to the design and transformation of plant community. The conclusion can provide data support for the design, construction and transformation of urban green space in Beijing and has practical guiding significance for the construction of urban green space environment.

  Method  The research applied the Durham pollen collector to monitor the airborne pollen concentration of 92 sampling points in Beijing from April 3, 2021 to April 9, 2021. The vegetation spaces with different pollen concentrations were classified by CART, which obtained the vegetation spatial structure model maps corresponding to high, medium and low pollen concentration sections.

  Result  (1) Coverage of arbors was the main control index to distinguish the pollen concentration of different squares. And the relationship between coverage of arbors with pollen concentration was not a single linearity but a certain degree of volatility. When the coverage of arbors was between 0.698 and 0.777, the pollen concentration was the lowest, and the predicted value was 74 grain/(10³ mm²). When the coverage of arbors was greater than 0.777, the pollen concentration was the largest and the predicted value was 285 grain/(10³ mm²). When the coverage of arbors was between 0.600 and 0.698, the predicted value of pollen concentration was 207 grain/(10³ mm²). When the coverage of arbors was less than 0.600, the predicted value of pollen concentration was 133 grain/(10³ mm²). (2) In addition to the coverage of arbors, the control indicators for pollen concentrations of different vegetation space included the average height of shrubs, average height of arbors, average height of ground cover plant, average height under branches of arbors, structure types of plant communities, percentage of arbors and three-dimensional green quantity of ground cover plant, etc. (3) The vegetation spatial structure with low pollen concentration was mainly with high trees in the upper layer and low shrubs in the middle and lower layer. The vegetation type was mainly large trees with high branch points. There were two main types of the vegetation spatial structure with high pollen concentration. One was that when the coverage of upper trees was lower, the vegetation in the middle and lower layers was dense, the other was that the coverage of upper trees was very high, and the vegetation community was very dense. The vegetation types were mainly evergreen trees, deciduous trees with dense canopy, and shrubs with rich branches.

  Conclusion  The scientific basis for the planting and optimization of airborne pollen allergenic plants was provided. For new urban green space, the allergenic plants should be used as much as possible in communities with similar structures to vegetation with high pollen concentrations. For established communities containing airborne pollen allergenic plants, the original low pollen concentration vegetation space could be transformed into the high pollen concentration vegetation space by increasing plants. The pollen would be retained and settled in the vegetation space as far as possible, so the amount of pollen dispersed into the city would be reduced and the health threat of airborne pollen allergenic plants to allergic people would be alleviated.