What’s that 5km from the station “location”

[This article was first published on Steven Mosher's Blog, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

In our last installment we looked at stations which were pitch black. The case I examined, Middlesboro Kentucky illustrated 1. The station location data used by Hansen2010 has inaccuracies. 2. While the purported station location was pitch dark, nearby within a couple 1/100ths of a degree there were urban lights. What this example illustrated was that you have no assurance that a station which has Zero lights is in fact in a rural location. the principle reason? station mislocation. The first screen I looked at was a 3km screen. That is I looked for pitch black stations with urban lights within 3km. There are only a handful

In this post I push the boundary out and examine stations that fit two criteria:

1. Nasanightlights =0

2 DMSP nightlights > 35 ( urban class 1) within 5km.

What we are essentially looking for are larger errors in the station location data.  The proceedure is exactly the same. We screen for these stations. We plot the google earth map and the light contour on top of each other ( to show the algorithm works) and then we investigate the details of the station by using a google earth tour that is loaded with all the station of this class.

here are some samples








There are several things that become clear looking at these examples.

1. Airport locations have the wrong coordinates in the ROW.

2. US location data has better quality, in fact, using the 4 digit accuracy that is available from NCDC we can see that some sites, when properly placed are STILL pitch dark.

3. Coastal locations are particularly sensitive to this type of error.

Those observations may afford us some remedies. In the USA the remedies are fairly easy. Since the location data is good ( if you use 4 digit accuracies we can see that the probability that location errors result in miscoding rural as urban is small.

for the ROW we have to be concerned with Coastal and airport stations. I will quickly survey the coastal issue and the airport in  subsequent posts.

Oh, and ISA has been added to metadata. more on that later


To leave a comment for the author, please follow the link and comment on their blog: Steven Mosher's Blog.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)