Did you ever wonder how we draw our station icons () on Google Maps? We simply convert our Earth-centered Earth-fixed cartesian station coordinates (X-Y-Z) to geodetic coordinates (latitude-longitude-height) using the WGS84 ellipsoid implemented in Google Maps (and using as many decimals as possible!) Then we use the Google Maps API to render the icon at the desired location.

But when we zoom in at the maximum level in Google Maps using satellite imagery we observe that most of the times the icon does not coincide with the actual position of the station antenna on the image. The image below shows the “gien” station antenna on the roof of INRiM, the Italian metrological institute in Turin. The station icon is around 5 meters away from the actual antenna!

How is this possible? Didn’t we say in a previous post that our coordinates have an accuracy of a couple of centimeters? The answer is that satellite images on Google Maps are in general georeferenced with an accuracy of just a few meters, which is of course enough for most users. For example if you use a normal handheld GPS receiver you get a positioning error of several meters, which is usually good enough to navigate through streets and roads without the need of a perfectly georeferenced map.

As from today we are moving to a new core station data management approach. We realize that in general for the user it is more useful to have recent data from many stations and with as little latency as possible. Unfortunately keeping old data from many stations on the magicGNSS server takes a lot of disk space (typically each station takes around 80 Mb of uncompressed RINEX data per day).

Until now we kept on the server core station data starting from July 1st 2008, to current time. As from now we will keep a moving window of one-month duration, i.e., the last 30 days of data will be always available on the server, and older data will be automatically deleted.