大约一年前,我终于实现了一个机制,每隔五分钟,从SMHI获取天气数据(瑞典官方组织,记录温度和自1700年代中期)的GPS坐标很靠近我的房子。隐私。(小于1公里)以来我真的爱它下雨的时候,然后提取所有的预测下雨,把它们放在我的日历,我经常看到我的屏幕上,这样我就能一眼就看到雨什么时候预计当前天,13天到未来。(很明显,每一次,它首先删除旧数据。)但这就是变得奇怪。现在会说明天要下雨了,但昨天,或者前几小时,5天到下次下雨。同一天,它可能说要下雨了10分钟,然后它增加到20分钟,然后一个小时离开,然后3个小时离开,然后3天没有下雨。然后为7天没有雨。然后明天下雨。似乎变化很大,完全不可靠的和无用的。 It feels like I might as well just remove this mechanism as I could just as well just make my own random guesses and probably be as accurate. But why is this? I thought that weather predictions were extremely reliable at this point, with global, advanced models processing enormous amounts of information with powerful computers around the clock and whatnot. Yet they seem to have no clue? Is predicting rain especially difficult, perhaps? Are just SMHI specifically very incompetent?