September 24, 2016
by chris

More images for mapping

Just put up a few additions to the OSM images for mapping:

A number of aerial images from about a year ago from Operation Icebridge overflights of the Thule Airbase (where most of these flights are conducted from). In contrast to the previous images these are partly without snow.

There are more such images from late Autumn from different parts of Greenland but they only cover small strips that are fairly insignificant is area compared to the whole country. With still plenty of things to map from the lower resolution imagery there is little use in adding them.

Next is a fairly large image of the Northern and Polar Ural mountains in Russia. This is very badly covered in Bing and Mapbox image layers. This image from Sentinel-2 was taken in August.

This should be helpful in mapping for example lakes, rivers, glaciers, ridges and cliffs but note although this is a late summer view not every white patch in it is a glacier. Also visible are roads, settlements and mining areas.

And finally there is a small image of Ushakov Island – not really much to see there but it can be used to update the coastline. Also you can see one of the meltwater lakes on the ice has drained.


September 17, 2016
by chris

Sun glint

In connection with assembly of satellite image mosaics i mentioned the effect of sun glint in satellite images in a recent post and want to elaborate a bit on this here.

Sun glint is a common phenomenon in satellite images – it essentially refers to the specular reflection of the sun on water surfaces. It is more or less the same as the sun reflection you see on a curved glass surface.

From the perspective of an earth observation satellite in sun synchronous orbit with a morning view (both Landsat and Sentinel-2 fall under this) sun glint looks like this:

This image shows a single day’s imagery of the Terra-MODIS camera from May this year. You can see the different orbits of the satellite each crossing the equator at the same local time and the small gaps between. Since the satellite photographs a morning view the sun is in the east and therefore the sun reflection is slightly to the east from the middle of the image swath. Now higher resolution satellites like Landsat and Sentinel-2 photograph a much more narrow field of view so sun glint is normally visible as a bright overlay of the water areas increasing in intensity from west to east. Here an example from Sentinel-2:

And here one from Landsat 8:

Since Sentinel-2 has a larger field of view than Landsat and a slightly later local imaging time (10:30 equator crossing time compared to 10:11) sun glint effects as well as their variability across the images are significantly stronger than with Landsat on average. The elongated form of the sun glint in flight direction is caused by the way satellites record images by scanning lines at a right angle to the direction of flight. The angle between view direction and the earth surface varies along this line but is not tilted backwards or forwards in flight direction so the glint varies strongly at a right angle to the flight path due to the view direction rotation across the field of view and changes much slower in flight direction due to just the varying orientation of the curved earth surface towards the sun.

Because of the way sun glint is affected by the view geometry as described it is also strongly subject to striping in the images due to the image sensors being split into several modules in most recent earth observation satellites each of which looks either slightly forward or backward and so they are affected somewhat differently by sun glint. Here an example from Sentinel-2 showing how this typically looks.

In the first example from the Canary Islands by the way striping is not visible because this image is taken near the maximum of the sun glint and therefore the forward and backward looking sensor modules record more or less the same amount of glint.

Sun glint is generally strongest at the latitude where the sun is highest depending on the seasons. It falls off north and south of this. During mid summer the effects of sun glint can be clearly seen up to about 50 degrees latitude with Landsat, somewhat further with Sentinel-2. Correspondingly during Winter sun glint free images can be recorded at lower latitudes as well.

Since sun glint is caused by a pretty basic geometric constellation you could assume that you could compensate for it but practically this is hardly possible because of several things:

  • Its strength and characteristics are highly dependent on how smooth the water surface is, i.e. waves. On the bright side this has the nice side effect that you can observe the sea state quite well on images with stong sun glint – as demonstrated by the Landsat example above.
  • Water areas are generally quite dark when viewed from above especially deep and clear water so the sun glint outshines the real reflection signal and even if you could properly estimate and compensate for the amount of sun glint you would still have the actual signal buried in the noise from the specular reflection.

Generally sun glint is usually considered an undesirable effect although not really a flaw or quality deficit like clouds. Practically it is one of the main reasons why you rarely see satellite image products that include water coverage of larger water areas at lower latitudes since it is difficult to uniformly render water in areas where sun glint is present.

You might wonder how the Green Marble renders the ocean without visible sun glint. This is because the wide field of view of the MODIS instrument contains sufficient data far enough from the directions of sun glint and this data is used to determine the color in such areas. But this is not without problems either – as you probably know from looking at water surfaces from the shore reflectivity increases when you look at the surface at a flatter angle – so you might get less sun glint this way but on the other hand get a larger amount of reflected skylight.

What could help dealing with sun glint in imagery is if satellites were able to record polarization information in the light received. Specular reflection is selective regarding polarization after all. But common earth observation satellites do not do this at the moment.

Some further info and literature on sun glint can be found here.


September 15, 2016
by chris

OpenStreetMap at its worst

To get this out upfront: this post is not about data quality – what i am going to talk about here is the mechanisms how OpenStreetMap functions and how these principles can break down.

In a nutshell – how OpenStreetMap works is that everyone can contribute to the database and thereby create something valuable others can use, in maps and other applications using OSM data. The real key however is that when you do so others contributing in the same area will then build upon your work, supplementing it with additional detail, updating information, correcting inaccuracies etc. This is what makes OSM attractive to contributors – you know your contribution matters, often even many years after you make it because you can be sure subsequent contributions will be supported by the basis you created. And it is attractive for data users because through this mechanism the result is usually significantly more valuable than the sum of all the individual contributions.

Those who know me might already imagine that when i talk about these principles breaking down i am talking about Canada, in particular about the Canadian North, the area you might best recognize from the distorted appearance in the Mercator projection:

The Canadian Arctic in OpenStreetMap

This region is one of the most sparsely populated areas of the northern hemisphere – Nunavut and the Northwest Territories have a combined population of less than 100000 the vast majority of which live south of the area discussed here which has likely less than 10000 inhabitants.

This makes it a fairly difficult area for mapping in OpenStreetMap. This is how this area looks like in terms of OSM node density – in a different, less distorting map projection:

I separated the data into three categories:

  • legacy imports of coastline and larger waterbodies, mostly PGS, made about 8 years ago and not touched since then are shown in blue – overall 1.3 million nodes.
  • unmodified Canvec data imports are shown in red – about 5.2 million nodes.
  • everything else, meaning hand mapping as well as any imported nodes that have been modified afterwards are shown in green – about 500 thousand nodes.

Now if you ignore the red you could get the impression this looks reasonably healthy considering the remoteness of the area. If you look at the age of the data:

you can see most manual mapping activity is fairly recent and limited to smaller areas. The Canvec import stuff is shown in gray since node age and data age are not the same for imported data of course – Canvec imports were made during the last five years mostly. Now you cannot compare this to a densely populated area in Europe of course, there is very little local mapping on the ground and nearly all of the data – both imported and manually mapped – is produced remotely. But lets compare it to Greenland – an area with quite comparable population density, accessibility and geography:

When magnified both maps have the same scale by the way. Compared to northern Canada Greenland has much earlier and more extensive manual mapping activities. There are likewise legacy imports of coastline data, in particular at the west coast. Overall the data volume is comparable if you disregard the Canvec imports, Greenland is about 2 million nodes in total, legacy imports and manual mapping together is about 1.8 million nodes in the shown part of Canada.

So what causes the difference? The fairly obvious explanation are the Canvec imports. Except for the legacy coastline imports from many years ago there have been no imports of data in Greenland. If you look at the maps above and the data you can see while the manual editing activity supplements and replaces the legacy imports quite freely there is hardly any interaction between manual mapping and Canvec imports. About 200k of the 500k manual nodes have been edited after initial creation (are version >=2), most of these are manually refined legacy import coastlines. Significantly less than one percent of the Canvec import nodes have been edited afterwards and most of the manual editing activity you can see in Canvec import areas is simple mechanical cleanup. If you have ever tried doing manual mapping in an area where Canvec data has been imported you know why this is rare – i did so once in the far north and this is not something you really want to do. Canvec imports are essentially a foreign body in OSM regarding normal editing activites which then try to operate around these.

Remember above i wrote than OpenStreetMap works by contributions supporting and forming a basis for further subsequent contributions. Canvec data imports do not work this way, especially not in the high Arctic. In addition they neither work the other way round, i.e. by integrating and making use of previous manual contributions. If at all such imports bury previously mapped stuff under tons of data of questionable quality. And the outlook of this happening is not exactly an incentive for mappers to contribute, especially not if they can also do so a few hundred kilometers further east where no such problems exist.

Now i wrote initially this is not about data quality but still i want to deal with one of the key arguments of Canvec import proponents: that the Canvec data is of good quality and much better than what can in most cases be manually mapped from available imagery sources. This is wrong. Canvec data in this area is in most parts somewhat more detailed than available image sources but it is worse in about every other aspect:

  • it is less up to date which is of particular importance in the Arctic due to glacier retreat and climate change. Most of the source data Canvec is based on in this area is at least ten years old, significant parts are much older (like 1980s).
  • it is often factually incorrect, partly due to incorrect original mapping, partly due to incorrect attribute conversion.

Everyone who does not believe that i would highly recommend to look at the recent images from the OSM images for mapping in the area and compare them to the Canvec data.

Due to these problems the imported data does not even give valuable hints to mappers unfamiliar with the area how to map things – on the contrary it suggests incorrect tagging in many cases.

Another argument frequently brought up is that having additional data in the OSM database is an advantage on its own. In reality this hardly is the case – if data users find the Canvec data useful it is generally much easier to take it directly from the source where it is available in uniform quality and with all original attributes for the whole country. And if you consume data on this scale the slight possible advantage of having it in the OSM format you are used to already is usually not significant.

Long story short – the only way the Canadian OSM community could in the long term make sure the Canadian North is a valuable part of the OSM database and an area where it feels to be rewarding to contribute for mappers would be to put a stop to Canvec imports in the area and make an effort in removing the previously imported data. Otherwise the Canadian Arctic will likely continue to fall back behind the rest of the world in terms of community building as well as data usefulness – not despite the imports but because of the imports.

Some probably read into this i am generally against data imports in OSM but i am not – the key question for such however has to be if they support further mapping in OSM in the area of the import or not and in this case the answer is quite certainly no.

Now one thing i asked myself in the matter is if this is actually a deeper cultural difference between Europe and America, the old world and the new world. Being from Europe i am probably not unbiased on this – despite extensive experience in mapping in the Arctic. It is possible that what i wrote about mapper motivation and incentives applies to the typical European mapper but not the North American one. Since much of the manual mapping in Northern Canada is done by people from abroad even what can be observed from a neutral standpoint could be distorted in that direction. OpenStreetMap is built upon the principle of primacy of the local mappers – they decide on their own how things are mapped in their area and if data is imported. But is someone sitting in Toronto, Montreal or Vancouver really more of a local mapper on Devon Island or Ellesmere Island than someone from Britain, Germany or Russia?


August 23, 2016
by chris

Some new satellite image mosaics

Recently i completed a number of new cloud free satellite image mosaics i want to introduce here.

The first one is an image of Cuba based on Landsat imagery. Seen from above Cuba is one of the most photogenic parts of the earth with very diverse landscapes and colorful coastal reefs and wetlands.

Then i have two smaller images of the Canary islands and Madeira.

The Canary islands mosaic is made from Sentinel-2 imagery while the Madeira image is a mixture of Sentinel-2 and Landsat 8. These are still somewhat experimental – you might notice the overall color balance being somewhat different from other mosaics. There are two special problems here – one is these islands feature extremely steep relief so to avoid long shadows preference was given to summer images with a high sun position while normally i target the maximum plant growth. This brings up a second issue: sun glint – which affects appearance of water surfaces. I will write about this in more detail in a separate post.

Finally there is also a new Sentinel-2 based mosaic of the Falkland Islands.

Click on the full image previews above to visit the mosaics’ pages on where you can find more information, previews and can inquire for using them.


August 8, 2016
by chris

Snow in different places

I have two satellite images to show here featuring unusual amounts of snow for this time of year. The first in from already a few weeks back in July showing fresh summer snow in the southwestern Alps:

Depending on the weather snow can fall in the Alps at higher altitude all around the year but it does not last long usually once the weather changes. Here fresh snow reaches down to about 2000m which is already pretty far for this time of year.

The second image is from Lesotho in southern Africa which received large amounts of winter snow this year.

Winter snow is common in Lesotho as well as in South African mountains further west but most of it usually does not last very long. Much of the snow that originally fell has already vanished in this image. See also here for more detailed explanations.

First image is created from Landsat 8 data, second from Copernicus Sentinel 2 data.


August 5, 2016
by chris

ALOS AW3D30 relief data review

When NASA after the SRTM (shuttle radar topography mission) in 2000 released the first SRTM data (the latest version of which i reviewed previously) this was for large parts of the wold the first detailed relief information available at all and for even larger parts the first relief data that was accessible at reasonable costs and with a uniform quality standard. There probably is no other single data set that was of such fundamental influence for today’s geodata landscape as SRTM.

Today, 15 years later, SRTM is still in some part of the world the best freely available relief data source but there is an increasing number of other open elevation data products as well. Most of them locally produced data covering only smaller areas, there are still very few free data sets with a more or less global scope. The newest one is the ALOS AW3D30 Global Digital Surface Model from the Japanese space agency JAXA. I now wrote a detailed review of this data set.


July 19, 2016
by chris

Arctic off-Nadir Landsat 2016

As mentioned previously when writing about images of the Antarctic the USGS announced some time ago they’d be starting to regularly record off-nadir images with Landsat 8 of the polar areas beyond the regular Landsat recording limit. As likewise mentioned these areas are also not covered by Sentinel-2.

A few days ago the first set of such images since the early ones from May 2013 i also wrote about previously was recorded in the Arctic in northern Greenland.

The path chosen is slightly different from the one from May 2013 which is shown here for comparison.

Note both images are in polar stereographic projection so they are not north oriented. Both cover all land areas that are not in the regular Landsat coverage.

The new images are not completely free of clouds but fairly clear overall. Mid July is not yet the snow and ice minimum, especially ice of lakes and at the coast is going to melt further in the coming weeks. But the time chosen is a good compromise between snow and ice on the one hand and sun position on the other. And it is possible that further images are going to be recorded in the coming weeks although likely not as many as in the Antarctic half a year ago since this much more interferes with regular image recording on the northern hemisphere because you cannot instantly switch between normal and off-nadir image recording.

Here a few crops from this new imagery showing the northmost land on earth.

You can also find these on the OSM images for mapping – together with a few ASTER images of northern Ellesmere Island likewise from the past few weeks. Overall with regards to mapping in OSM these areas, which a few years ago were hardly mappable at all, have now a significantly better and more accessible image coverage than many other parts of the planet.

July 16, 2016
by chris

What satellites see

In the last post here i showed an illustration of the spectral sensitivities of current earth observation satellites. I like to comment this a bit more in detail separately here.

The list of satellites is not meant to be complete, in particular not listed are military or otherwise secret systems, older geostationary weather satellites and also various systems with unclear operational status. Except for Landsat 5 (no more) and GCOM-C (not yet) i only included operational systems – Sentinel-3 however is not yet producing data for the public.

Satellites are grouped into three groups:

  • At the bottom the commercial and limited access systems
  • The center group with brighter background are the higher resolution open data systems
  • On top the lower resolution continuously recording systems

I also listed the revisit and global coverage intervals. Revisit is somewhat ill defined for satellites with pointing capability but is generally considered to be 1-6 days for a single satellite with small field of view – depends somewhat on the orbit of course.

Global coverage is only specified for those satellites which regularly record a global or near global coverage. Several commercial operators claim this but there is no reliable data to verify those claims. Landsat 5 and 7 have both in the past been used to produce global image collections over the course of several years. Landsat 8 now usually records all land surfaces within its coverage area in the 16 day revisit cycle with only occasional gaps. The lower resolution polar orbiting systems (MODIS/VIIRS/Sentinel-3) all have a well defined global coverage interval since they are capable of continuously recording. AVHRR is meant – using several satellites – to achieve a 6 hour revisit frequency and of course geostationary systems can record their area of view in much higher frequency. DSCOVR EPIC finally – by not being a real satellite – has its own special recording scenario.

Apart from DSCOVR EPIC and the geostationary satellites nearly all systems shown are in a sun synchronous orbit, they record earth at approximately the same time of the day locally. The only exception are the Planetscope satellites in ISS orbit.


July 15, 2016
by chris

On false colors

On various occasions i have written about the problem of accurate colors in satellite images before. However what often makes satellite imagery such a seemingly complicated and difficult matter are usually not these subtle details of color fidelity and perception. What makes images frequently difficult to understand is that what is widely advertised as satellite imagery is often not images or photos is a strict sense but a broad range of visualizations and artwork (if you want to call it art) based on remote sensing data. Such pictures are generally indiscriminately presented as satellite images, often without explanation and no established boundary exists where the domain of photos ends and abstract visualizations begin.

Compare that with what you see in a newspaper for example – you have photos there, you have occasionally drawings of real world situations like from court room procedures where no cameras are allowed. And you of course have things like diagrams, caricatures and other graphics that are not meant to directly depict a real world view. Having these clearly distinguished types of pictures is in many ways central to the concept and self image of journalism.

Ironically science marketing and science journalism are one of the fields where this distinction tends to get blurred most, especially when it comes to satellite data.

It is difficult to make reliable estimates but i would say at least half of what you usually run across in newspapers, television, on websites etc. as satellite images are not photos in the sense of an attempt to create a reproduction of a certain imaging measurement with even the most rudimentary effort of faithfulness.

About half of these non-photographic satellite images are completely abstract depictions, usually some kind of 2d data set based on remotely sensed data visualized in some map coordinate system that allows you to recognize geographic features. This includes for example shaded depictions of relief data (yes, i have seen those being called ‘satellite image’) and radar data visualizations (which are usually based on run time/phase information and not a 2d imaging process).

The other half is what you usually call false color images and i want to discuss this topic in more detail here.

False color

False color images are a concept that predates digital imagery. It has its origin in classic chemical photography, specifically color infrared films that are sensitive in various parts of the visible and infrared spectrum but develop into visual range colors. This concept was then later reused in digital images.

To show what false color images mean first for comparison a Landsat image in normal visual colors:

This shows parts of the western Caucasus mountains with a large variety of surface features in the image. You can see good contrast between some surface types, like rock/snow and vegetation and even forests and grassland/agricultural areas can be well differentiated. On the other hand contrast between vegetation and water is rather small.

The visual range sensor channels of the satellite are here directly mapped to the red, green and blue channels of the image file and ultimately the display device. No matter how well the sensor sensitivities match the definition of the display color space and the sensitivities of the human eye – this is generally called true color or visual color. This does not necessarily mean the image shows realistic colors of course.

In addition to the blue, green and red color channels the satellites also record a number of other channels across the electromagnetic spectrum, most of them in the infrared (that is longer wavelengths than the visible light):

Just like with color infrared film you can map these to visual range colors to visualize these measurements. A traditional mapping is

NIR → red
red → green
green → blue

which suited early earth observation satellites which often had a green, red and near infrared channel and which is also close to the characteristics of color infrared film. Water surfaces are in much better contrast to vegetation here:

The important thing to realize is that since such a mapping has absolutely no physical basis it is ultimately completely arbitrary – you can map your data to any combination of visual range colors, as i for example showed in context of the ASTER instrument it is sometimes preferred to have the vegetation in green and not red so the same data can also be mapped like this:

And you do not even need to have a one-to-one mapping of the spectral channels to the red, green and blue color components – you can use an arbitrary linear or even non-linear mapping. For example here is a mapping that retains brightness from the visual range but defines the color tone using infrared data:

For false color images to be of any practical use and not just abstract art there need to be conventions on their definition and you need to learn to read them. This is a huge problem because the definitions are often quite hairy and the need to learn how to read these images is widely ignored and underestimated. Showing a false color image is a bit like printing an article in Chinese in a German newspaper – the main difference is that with the text the reader obviously realizes it is not understandable while with the false color image you can easily get a false impression of understanding. But in fact it is even worse – even with a lot of experience and training at using a certain type of false color image you always have a high risk of erroneously believing to correctly interpret certain features while this can easily be an optical illusion or a cognitive fallacy. Since training and experience with false color satellite image is always based exclusively on looking at satellite images and lacks the real life context of normal visual color perception it can never reach the level of reliability and robustness of reading true color images.

So with these immense disadvantages why do people none the less use false color images? This is because

  • there is a lot of information available in the infrared data that is not present in the visual bands,
  • using the full range of a color display and our visual perception can be helpful when visualizing complex multidimensional data and
  • the human mind can be trained to recognize similarities and patterns in color images, even if they are based on artificial colors.

Here is – for the area previously shown – the today most common false color band combination for multispectral satellite images. This is based on the mapping

SWIR → red
NIR → green
red → blue

This band combination is so popular because it manages to compress a real lot of information into a single color image. Since it does not use the blue and green visual range bands there is fairly little atmospheric influence, you can see quite well even through veiling clouds. For basic orientation the following basic properties:

  • Vegetation is generally green – different tones of green primarily indicate different vegetation density and different states in the growth cycle, different types of vegetation are often not so easy to distinguish.
  • Water is generally dark blue, dry snow and ice is bright blue to cyan, wet snow is darker.
  • Bare ground appears in various tones of brown, red and gray. Differences between different geological settings are often better differentiated than in the visible range, for example in the area shown here the limestone mountains south of the main ridge appear significantly brighter than the crystalline/metamorphic rocks further north. You need to be careful however since soil humidity also influences this quite a lot.
  • Clouds are usually fairly colorless bright gray and white although they can sometimes also be somewhat bluish or reddish.

Although this type of band combination is routinely used with data from many different satellites the results vary since where exactly the NIR and SWIR bands are located varies, much more than it does for the visual range bands which are obviously often tied to the human perception. Here is a diagram of current earth observation satellites, an extension of what i showed for the visual range previously.

There are a huge number of other false color band mappings – many of them variants of the SWIR-NIR-visual mapping above with either the other SWIR band or another visual range band substituted – these all look fairly similar at the first glance but vary in the details. There are also a few completely different mappings you occasionally see, for example the pure infrared mapping SWIR2-SWIR1-NIR and a blue-SWIR1-SWIR2 combination:


All of this is only with the most common spectral bands – ultimately the possibilities are endless and it is easy to turn this into a kind of secret code with images only a few initiated experts can read. But the use of such endeavors is highly questionable – as i hope i made clear reliably interpreting even the most common false color image types is full of difficulties. If you think certain things can be reliably determined from the appearance in a false color depiction it is often better to formulate this conclusion in a quantitative way based on the data values. This also has the advantage you can take into account more than the three variables you can independently represent in a color image.


To sum things up the following recommendations

  • Do not use false color satellite images in publications or marketing.
  • If you use false color satellite images publicly include also a true color image for reference and include a clear warning that colors are artificial.
  • Consider using single band images or single value processings like band ratios or difference indices, possibly with a well designed color gradient (no rainbow colors!) instead of a false color image.
  • When interpreting false color satellite images be wary of the possibility of misconceptions and verify your interpretation is consistent with the true color appearance if possible.
  • Be aware that there are serious differences between different satellites with equivalent false color band combinations.

July 7, 2016
by chris

Beyond Mercator

FOSSGIS 2016 in Salzburg is over – many interesting talks you can all watch on the FOSSGIS Youtube channel. You can now also find the slides for my talk on Frab.

Many people showed interest in the Peirce quincuncial projection for which i showed a demo at the end – you can find the proj.4 implementation on github and the demo map on – or included directly below, pressing shift while panning gives you automated north orientation.

I will try to write up a bit more on this in the future – there are a lot of interesting aspects about this projection, both from a technical viewpoint as well as from a user perspective.


July 1, 2016
by chris

Talk Announcement

On Monday the the FOSSGIS conference starts in Salzburg – after a more informal OSM-Sunday. I will give the opening talk there (in German) – with the title Jenseits von Mercator (Beyond Mercator) where i will be talking about the meanwhile omnipresent and seemingly unavoidable mercator map projection. I am going to discuss its virtues and disadvantages and show some alternatives. Should be interesting for anyone involved in use and production of maps somehow.

Those who would like to come by spontaneously can do so – OSM and Open Source community members can visit for free. Those who can’t be there is person will probably be able to watch it live or as recording on the internet – on

Salzburg by the way currently looks like this:


June 28, 2016
by chris

The new Google Landsat mosaic

It has been some time since i wrote the last review of a satellite image mosaic product. In fact the last one i did was for the last Google image mosaic based on Landsat 7 data three years ago. Now they have rolled out a new version of their base mosaic i am going to review here.

But first a general look at the three years in between. Why did i not do any reviews since then? Apparently because there was not much to review. Of course there were a few cases where map services put in some local image updates and improvements in their image layers. For example Bing recently filled some of the most obvious gaps in image coverage. But technologically these three years have been pretty uninteresting, nothing really revolutionary was released in the field of global mosaics, at least not publicly – so nothing to review really.

In terms of data – we remember Landsat 8 began operations a bit more than three years ago so during all this time a veritable heap of free high quality data has been piling up. And Google apparently thought it was time to put this to use in their base image product.

Has something changed?

If you look at the new image compared to the old at a coarse scale (new left):


you can hardly see a difference. Overall apparently not much has changed in their processing system. This especially means they apparently use the same color homogenization and color transfer techniques they have used before which leads to a globally fairly consistent appearance (for which – in light of the many other mosaics with colors grossly off – i commended them in my last review). But this also means locally color nuances get lost and the fairly good color reproduction of Landsat 8 is – i hate to say it – mostly wasted.

Comparison new-old from northern Ellesmere Island

Comparison new-old from Siberia

But before we have a closer look locally some basic facts overall. The image mosaic, just like the old one, is nearly global – as far as the Landsat coverage goes but excluding the Antarctic. In areas where they previously already used other images at the low zooms they continue to do so (Svalbard, South Georgia and a few other places). For the Antarctic they continue using LIMA, in the north beyond the Landsat coverage they now have a fairly poor low resolution image with the source not identifiable.

At the Landsat limit in northern Greenland

My own Greenland Mosaic for comparison

As you can also see in the sample above they still use the same inaccurate coastline mask cutting off their image data in the middle.

What is improved?

Well – since their main source of data now is Landsat 8 data from the last three years one thing that improved for sure is the imagery is more up-to-date. Previously with the Landsat 7 data basis the average image was likely from around 2007 and now it is likely from 2014. Note however that a three years data basis is not much data to deal with (which is why we have not seen any other global Landsat 8 mosaic so far). It seems Google did also continue to use Landsat 7 data to some extent – this is generally well visible in form of the characteristic striping which will not occur with Landsat 8 images.

Landsat 7 striping in the new Google mosaic

Their clouds handling seems to have in parts somewhat improved – possibly because they used the Landsat 8 data in addition to the images used before there

while in other areas it is significantly worse

While the areas shown previously are simply very hard concerning clouds there are also cases where they would have been completely avoidable:

Probably partly due to the relatively small overall image volume areas with strong seasonality like high latitudes and high mountain areas are often quite bad as well.

Google in the Alps

My Alps mosaic for comparison


Overall i am not really sure why Google produced this update. There was no real pressure from the competition making their imagery appear outdated – after all most map services still use the much older Landsat 7 images from before the SLC failure (1999-2003) as base imagery. I suppose they mostly see this as a demonstration of their satellite image processing platform which they also market independently.

In terms of technological development i don’t see much here though. If you know the difference between Landsat 7 and 8 regarding the source data quality, especially noise levels and dynamic range – the new mosaic seems rather disappointing. So the main thing this demonstrates is that their processing framework is sufficiently robust to produce a new mosaic from a changed data basis with comparable results.

And of course they show that they can still with their pinkie do what others have not been able to do for many years – namely produce a globally consistent and visually at least at a coarse look cloud free satellite mosaic in this resolution range.