tag:blogger.com,1999:blog-5337555368793819627.post5196252147129512654..comments2023-06-07T09:04:36.390-04:00Comments on More Grumbine Science: What cooling trend?Robert Grumbinehttp://www.blogger.com/profile/10783453972811796911noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-5337555368793819627.post-77317301500118790202010-01-25T17:16:31.428-05:002010-01-25T17:16:31.428-05:00Welcome Mike. Never too late to add a good though...Welcome Mike. Never too late to add a good thought. Mining is far from my experience, so I appreciate your example. Yet another field that has arrived at a similar principle.Robert Grumbinehttps://www.blogger.com/profile/10783453972811796911noreply@blogger.comtag:blogger.com,1999:blog-5337555368793819627.post-63070711076244397682010-01-23T19:25:48.447-05:002010-01-23T19:25:48.447-05:00This comment is months late, but I have only disco...This comment is months late, but I have only discovered your excellent blog today.<br /><br />As a geologist with some background in the geochemistry of mineral deposits, I would note that the concept of a proper scale for statisically valid sampling should be familiar to anyone with experience in the mining industry. Sampling too small a volume or rock will give widely varying results that are dominated by "nuggets". Sampling too large a volume will dilute the ore, but is nonetheless an essential calculation to determine the economics of a potential mine. I therefore find it very odd that a certain high-profile AGW skeptic with an extensive background in the mining industry continues to use inappropriate sampling sizes when it comes to his "audits" of temperature records!Dr Mike at Otagohttps://www.blogger.com/profile/17002470523618539327noreply@blogger.comtag:blogger.com,1999:blog-5337555368793819627.post-1314918749351638132009-08-01T16:08:01.041-04:002009-08-01T16:08:01.041-04:00Very nice illustration of the problem with short t...Very nice illustration of the problem with short term "trends".<br /><br />Horatio shows this concept with an "animation" of sorts in <a rel="nofollow">See-saw science</a>Horatio Algeranonhttps://www.blogger.com/profile/12988805467080448954noreply@blogger.comtag:blogger.com,1999:blog-5337555368793819627.post-87468897471595873352009-07-21T09:08:35.174-04:002009-07-21T09:08:35.174-04:00Thanks Gordon. I actually started with fluid dyna...Thanks Gordon. I actually started with fluid dynamics as the guide -- the continuum approximation volume. Exactly the same idea as your Representative Elemental Volume, for exactly the same reasons. That's one of the things I like about these fundamental principles -- you can apply them in many areas. The fluid dynamics-guided post is http://moregrumbinescience.blogspot.com/2008/08/what-is-climate-2.html<br /><br />I also took a look from a couple calculus ideas you'll recognize, and arrived at the same conclusions as to time scales:<br />http://moregrumbinescience.blogspot.com/2008/12/how-to-decide-climate-trends.html<br /><br />Something I haven't yet done is look at the long period end. Same as you don't want your REV to be too large (encompassing different watersheds, for instance), we need the climate averaging period to be short enough not to be averaging different climate periods. (Ok, by 'look at' I only mean that I haven't written it up for the blog. I've definitely looked at the issue.)Robert Grumbinehttps://www.blogger.com/profile/10783453972811796911noreply@blogger.comtag:blogger.com,1999:blog-5337555368793819627.post-60686837672536407722009-07-20T21:18:54.971-04:002009-07-20T21:18:54.971-04:00Excellent post. Perhaps the best discussion of th...Excellent post. Perhaps the best discussion of this topic, at the layman's level, that I've yet seen. Very clear and simple as to why you need a certain number of data points to arrive at a meaningful representation. <br /><br />It reminds me of the Representative Elemental Volume used by hydrogeologists in groundwater flow (fluid flow through porous media). The REV is the appropriate sample size required to accurately represent the media as homogenous for a particular parameter and problem (in groundwater flow, typically hydraulic conductivity or porosity). For a small sample size (or area or volume), you may have "wild fluctuations" in the parameter of interest (e.g., in groundwater flow, small sample volume may be influence by one or more large pores, voids, or aggregate, causing porosity to be unusually great or little in relation to the volume selected). As the sample size passes the REV, however, variation in the parameter of interest converges to a stable value (until, in hydrogeologic problems, the value may diverge or trend to a new value - indicating a sample size that is now so large as to include, and evidence, true spatial variability).<br /><br />http://unrulednotebook.wordpress.com/2008/08/19/porous-medium-homogeneity-and-representative-elemental-volume/<br /><br />I applied this concept to some geotechnical sampling and analyses I completed for my master's degree, and the engineering professors rather ate that up, as it were. It seems to me to be a similar concept here in climatology - how many temporal data points do you need to have a "representative" timeframe, and as you let that time extend further, until you are incorporating historical trends, I would assume you'd get steady changes to different values. But for the purposes of understanding climate, in this sense, your Representative Elemental Volume for global temperature trend, as you just demonstrated, would be on the order of 15 to 30 years.<br /><br />I would also expect that if someone were to pick a random monthly value from the historical record, say May 1995, and repeat your procedure, the result would look similar - lots of fluctuation on the scale of months to a few years, very likely both positive and negative, then a gradual settling into a stable value...Gordon Parishnoreply@blogger.com