应用多元线性逐步回归分析方法探讨气象因素与气传致敏花粉飘散的关系(英文)

Relationship between dispersal of airborne pollen and meteorological factors: Analysis by using the multivariate linear stepwise regression equation

背景: 气传致敏花粉是导致季节性变态反应疾病的最重要因素, 花粉的产生及飘散与气候因素密切相关。目的: 观察南昌市城区大气气传花粉与气象七要素( 气压、气温、湿度、降水、风速、蒸发量、日照) 的相关性, 为提高该地区过敏性疾病的防治水平提供理论依据。设计: 观察性实验。单位: 江西医学院变态反应疾病研究室。材料: 采用 Durham 重量法进行为期一年的大气气传花粉调查, 收集南昌市城区气传致敏花粉、气象资料。方法: 选择南昌市中心的江西医学院南院实验大楼四层楼顶为曝片点,采用 Durham 重量法于 2000- 04- 01/2001- 03- 31 进行为期 1 年的大气气传花粉调查。收集气象资料, 每日气象七要素的详细资料由江西省气象局提供。应用 SAS 6.12 进行多元线性逐步回归分析大气气传花粉与气象因素的相关性。主要观察指标: 南昌市城区气传致敏花粉的种类、数量、季节性分布以及和气象七要素的相关性。结果: 调查发现南昌市长年有花粉飘散, 花粉种类达 47 个种属。柏科花粉主要在 3 月飘散; 松科花粉主要在 3 和 4 月份飘散; 杉木属花粉主要在 3 月飘散; 藜苋科花粉主要在 8～11 月飘散; 豚草属花粉主要在 8～11月, 蒿属花粉主要在 9～11 月飘散; 葎草属以 9 月飘散为主; 禾本科在 12个月中都有出现, 但以 6～10 月飘散为主。在大气气传花粉飘散与气象七要素相关性分析中, 应用几种主要花粉飘散稳定期的花粉计数与气象七要素作多元线性逐步回归分析。松科花粉飘散与气压、气温、风速、日照呈正相关, 与蒸发量、降水呈负相关; 柏科花粉飘散与气压、气温、日照呈正相关, 与湿度、降水、蒸发量、风速呈负相关; 豚草属花粉飘散与气压、气温、风速呈正相关, 与湿度、降水、蒸发量、日照呈负相关; 蒿属花粉飘散与气压、气温、风速呈正相关, 与湿度、降水、蒸发量、日照呈负相关; 禾本科花粉飘散与气压、气温呈正相关, 与湿度、降水、蒸发量呈负相关。但以主要花粉及花粉总量的全年计数进行分析和以主要花粉及花粉总量的半年计数量进行分析时纳入回归方程的因素要少得多。结论: 花粉飘散与气象七要素具有相关性; 在分析气象因素与空中飘散花粉粒数的相关性时, 最好选择单种花粉粒, 并选择其飘散较为稳定的时段进行。

BACKGROUND ： Airborne pollen is the most important factor to induce the seasonal allergic diseases. The production and dispersal of pollen are closely correlated with the meteorological factors. OBJECTIVE： To observe the correlation between the seven meteorological factors （air pressure, air temperature, humidity, precipitation, wind speed, evaporation capacity and sunshine） and the airborne pollen in Nanchang so as to provide theoretic evidences for preventing and treating anaphylactic disease in that region. DESIGN ： Observational experiment. SETTING： Department of Allergic Reaction, Jiangxi Medical College MATERIALS： An investigation on airborne pollen was performed for a year by Durham gravity method. Data of airborne pollen and the seven meteorological factors in urban area of Nanchang city were collected. METHODS： The top of the 4^th experimental building in the southern yard of Jiangxi Medical College located in the center of Nanchang city was taken as exposed point. The pollen was surveyed by Durham gravity method from April 1^st, 2000 to March 31^st 2001. Meteorological data of the seven factors were collected every day and supplied by the Weather Bureau of Jiangxi Province. The correlation between meteorological factors and airborne pollen was analyzed by the multivariate linear stepwise regression analysis with SAS 6.12. MAIN OUTCOME MEASURES ： The type, amount, seasonal distribution of airborne pollen and the correlation between the seven meteorological factors and the airborne pollen in urban area of Nanchang city. RESULTS ： There were airborne pollens in Nanchang city all the year round, which included about 47 types. The main pollen season of Cupressaceae was in March, Pinaceae in March and April, and Cunninghamia R.Br in March, Chenopodium-Amaranthus and Ambrosia from August to November, Artemisia from September to November and Humulus in September. Gramineae had been scattering all seasons but mostly from June to October. In the correlation analysis between dispersal of airborne pollen and seven meteorological factors, multivariate linear stepwise regression analysis was conducted in some major pollen counting in steady dispersal period and seven meteorological factors. The positive correlation was found in Pinaceae with air pressure, air temperature, wind speed and sunlight, and the negative correlation was found with precipitation and evaporation capacity. The positive correlation was found in Cupressaceae with air pressure, air temperature and sunlight, and negative correlation was found with humidity, precipitation, wind speed and evaporation capacity. The positive correlation was found in Ambrosia with air pressure, air temperature and wind speed, while negative correlation was found with humidity, precipitation, evaporation capacity and sunshine. The positive correlation was found in Artemisia with air pressure, air temperature and wind speed, while negative correlation was found with humidity, precipitation, evaporation capacity and sunshine. The positive correlation was found in Gramineae with air pressure and air temperature, while negative correlation was found with humidity, precipitation and evaporation capacity. However, if using count of total pollen in a year or half a year to analyze the relationship, the correlated factors were much fewer than that mentioned above. CONCLUSION ： The dispersal of airborne pollen is related with the seven meteorological factors. It is better to use the grain amount of single pollen in its main pollen season to analyze the relationship with meteorological factors.