论文部分内容阅读
[背景]热浪与健康预警系统的激活是基于对健康具有危害性的高温天气的预报。[目的]根据在密歇根州底特律市预测和观察到的气象数据,评估高温与死亡率之间的关联;并比较预报结果在预测热浪中的准确性。[方法]从天气观测和6个不同的预报结果中得出表观温度(AT)和热浪天数(热浪的定义为:日平均AT≥温暖季节平均温度的第95百分位、并持续≥2 d),并对这些数据进行比较。使用泊松回归分析,调整或不调整臭氧和/或PM1(0颗粒物的空气动力学直径≤10μm),来估计并比较每天全因死亡率与观测所得以及预测所得的AT和热浪天数之间的关联。[结果]1 d前的局部观测结果,即经修正的数字预测,与所有其他的预测相比较,误报的数量约占一半。平均而言,在控制了热浪因素后,与AT=8.5℃的日子相比,观察到AT=25.3℃的日子与3.5%较高的死亡率相关联(95%CI:-1.6%~8.8%)。与非热浪日相比,观察到的热浪日均与6.2%较高的死亡率相关联(95%CI:-0.4%~13.2%)。相对于根据观测指标所得的关联,预测的准确性存在变化,但死亡率与预测高温之间的关联通常倾向于高估高温的影响,而与预测热浪的关联则倾向于低估热浪的影响。[结论]上述研究结果表明,结合局部地区的信息可以提高热浪和健康预警系统预测的准确性。
[Background] The activation of heatwaves and health warning systems is based on the forecast of hot, weather-threatening health. [Objectives] To assess the association between hyperthermia and mortality based on meteorological data predicted and observed in Detroit, Michigan; and to compare the accuracy of forecast results in predicting heatwaves. [Method] The apparent temperature (AT) and days of heat waves (heat waves were defined as: the average AT> 95th percentile of the warm season average temperature and continued for ≥2 from the weather observations and 6 different forecast results d) and compare these data. Poisson regression analysis was used to estimate and compare the daily all-cause mortality with the observed and the predicted number of days of AT and heat waves, with or without adjustment for ozone and / or PM1 (aerodynamic diameter of 0 particles ≤ 10 μm) Associated. [Results] The local observations before 1 d, ie, the revised figures predicted that the number of false positives was about half compared with all other predictions. On average, a day with AT = 25.3 ° C was observed to correlate with a 3.5% higher mortality (95% CI: -1.6% -8.8%) after days of controlled heatwaves compared to AT = 8.5 ° C ). The observed average daily heat wave rates were associated with a higher 6.2% mortality (95% CI: -0.4% to 13.2%) compared to days on non-heat waves. The accuracy of the predictions varies relative to the associations based on the observed indicators, but the association between mortality and predicted highs usually tends to overestimate the effect of high temperatures, whereas the association with predicted heatsters tends to underestimate the effects of heatwaves. [Conclusion] The above results show that the combination of local information can improve the prediction accuracy of heatwave and health warning system.