New Global Light Pollution Atlas
Posted 19 September 2013 - 09:16 PM
Posted 20 September 2013 - 12:15 AM
Good to hear from you. The 2001 maps you generated made quite a splash.
Any chance you could put them in a Google-Maps format?
Posted 20 September 2013 - 09:25 AM
I'll try but the maps are kind of big. I think there may be more advanced ways of putting them in Google Maps that makes the image sizes more manageable. I'll have to look into this.
Posted 20 September 2013 - 09:37 AM
At the opposite extreme, check out the oil and gas fields in northern Russia (south and east of the Gulf of Ob), the Persian Gulf, the Niger Delta and the mouth of the Congo. 2006 is too early for the big boom in the Bakken formation (western North Dakota), but I think this is about as bright as northern Russia at this point (although obviously not as extensive).
Also, look at the border between India and Pakistan.
Posted 22 September 2013 - 09:03 AM
Good news, the atlas is now available from my site in Google Maps! I've just learned all about tiling in Google maps, so the whole world is at one link. This is probably the easiest way to see the data.
Also, unfortunately there was an error within 5 degrees latitude of the equator. Light pollution was being "reflected" across the equator. The problem is now FIXED. Sorry.
Posted 22 September 2013 - 12:30 PM
I'm suspicious about the amount of light pollution depicted around Grantsville, WV, though. That area is *very* dark at night with a small light dome around Grantsville itself, yet this map shows a stream of "dark green" extending all along SR-16 south for about 15 miles.
Was this data taken on a football night in the fall? The local high school does have home games at a site about 6 miles south of Grantsville.
Posted 22 September 2013 - 12:55 PM
The satellite data is a composite over all available cloud free, new moon nights. I actually asked the data providers if there was any additional information about when the observations were taken because I was concerned about snow and there really is no more available information (that doesn't involve reprocessing all the data all over again, I assume).
Do you have SQM measurements from the Grantsville area? The area outside Grantville is indeed very dark (dark blue) and it makes sense that it would be brighter along the highway to the south. The atlas could very well be off but the relative darkness does appear to make sense.
Posted 22 September 2013 - 01:42 PM
Looks great in Google maps.
One last request... A downloadable Google Earth overlay?
Posted 22 September 2013 - 07:08 PM
Posted 22 September 2013 - 07:42 PM
According to this I'm no longer in the white zone but in the red zone (within walking distance of white) which I find a little interesting considering between 2001-2006 there was a lot more development in this area. Also between 2006-now there has been still even more so it probably has switched back to white.
Posted 23 September 2013 - 12:05 PM
Posted 26 September 2013 - 04:29 PM
in another thread i was asking about this couple of months ago.
Posted 26 September 2013 - 08:29 PM
Posted 27 September 2013 - 12:29 AM
It does appear that fast-growing corridors in the southern states have deteriorated, especially the Piedmont regions of VA, NC and Georgia. Deerlick Astronomy Village looks like it's been downgraded into a "green" zone from a "blue zone".
Posted 27 September 2013 - 03:33 PM
For inspection within a given map, I found the 2001 gray scale maps most useful, due to the greater granularity.
Dave, thanks for taking the time once again to do this work.
Posted 27 September 2013 - 05:18 PM
It may not be very useful to attempt detailed comparisons between the 2001 and 2006 maps.
Yes, I'd be very wary of comparing maps made from different data. First of all, there's an inevitable problem of calibrating new satellite measurements against old. More subtly, there are bound to be both systematic and random variations due to the times when the satellite measurements were made.
The systematic variations come from things like snow -- it's known that the original 1997 maps made northern areas seem much worse than they really are due to snow cover.
The random variation could come from things like the time of night when the measurements were taken. Artificial skyglow invariably decreases -- often by a factor of two or more -- from early evening to predawn as more lights are turned off and fewer cars drive the roads. So an area that just happens to have been imaged at 4 a.m. will look darker than an identical area that was imaged at 10 p.m.
I would expect only really strong trends to be reliably recorded. For instance, the really substantial population growth in the South and Southwest -- or in resort towns like Jackson, WY.
I imagine that the Midwest rust belt, like the Northeast, has remained broadly the same over the last 15 years, because both area's populations have remained much the same. Except in a few rapidly growing outer suburbs.
Posted 29 September 2013 - 08:23 PM
Posted 29 September 2013 - 09:39 PM
Canton, OH, is in the bright red in 2006, but white in 2001. Columbus has a core of white, but mostly bright red, in 2006. In 2001, it was white all the way out to I-270. Toledo and Dayton had white zones in 2001 that are roughly equal in size the bright red zones in their downtown regions in the 2006 atlas.
Posted 30 September 2013 - 10:05 AM
I have just been comparing the 2006 data to some published CCD measurements over a wide range of sky conditions in Italy. The biggest errors between the light pollution atlas and the real data are in bright areas. It turns out that the bright areas are actually much brighter than suggested by the 2006 light pollution atlas, which suggests that maybe the 2001 data is closer to the truth in the bright cities. I'll try to post this soon, but first I want to check a few things.
Posted 05 April 2014 - 02:30 PM
Posted 05 April 2014 - 10:26 PM
South 7° 0'5.54"N
East 51° 1'32.63"W
West 180° 0'0.00"W
Eyeballed but seems to work ok..