tag:blogger.com,1999:blog-7729093380675162051.post2532586844363504056..comments2024-03-16T02:27:38.423+11:00Comments on moyhu: New local station trends - comments.Nick Stokeshttp://www.blogger.com/profile/06377413236983002873noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-7729093380675162051.post-30303096251325024312017-05-21T01:00:41.308+10:002017-05-21T01:00:41.308+10:00Not so much of a problem today but more significan...Not so much of a problem today but more significant in the past when coverage was poor. But I guess with less stations, locating discontinuities becomes more difficult. But with all of the independent sources and measurements available today, getting really good homogenization methods would at least improve error estimates of the older data. Correlation between SAT, MSU/AMSU, Radiosondes, Ships, Buoys, AVHHR. Correleation by distance, region, biome, etc. Lots of different ways to take everything that we do know, and constrain what we don't know.ccehttps://www.blogger.com/profile/03646816472336349526noreply@blogger.comtag:blogger.com,1999:blog-7729093380675162051.post-81555457904388955902017-05-20T11:29:18.930+10:002017-05-20T11:29:18.930+10:00CCE,
I wouldn't tackle homogenization myself. ...CCE,<br />I wouldn't tackle homogenization myself. I use unadjusted results basically to show it makes little difference in aggregate, but I do think it is better to homogenize. I think these plots do a good job of showing up inhomogeneities, and the extent to which homogenization fixes them (less than total). But I'd be wary of overdoing it. The basic need is to counter bias in the global (and regional) average rather than get perfect smoothness.Nick Stokeshttps://www.blogger.com/profile/06377413236983002873noreply@blogger.comtag:blogger.com,1999:blog-7729093380675162051.post-3429951998034067022017-05-20T09:14:42.816+10:002017-05-20T09:14:42.816+10:00I believe you said in the past that you preferred ...I believe you said in the past that you preferred not to do any homogenization. Have you reconsidered that? Seems like there is still room for improvement, at least over NOAA's methods.ccenoreply@blogger.com