Landsat Turns 45!

False colour image of Dallas, Texas. The first fully operational Landsat image taken on July 25, 1972, Image courtesy: NASA’s Earth Observatory

Landsat has celebrated forty-five years of Earth observation this week. The first Landsat mission was Earth Resources Technology Satellite 1 (ERTS-1), which was launched into a sun-synchronous near polar orbit on the 23 July 1972. It wasn’t renamed Landsat-1 until 1975. It had an anticipated life of 1 year and carried two instruments: the Multi Spectral Scanner (MSS) and the Return-Beam Vidicon (RBV).

The Landsat missions have data continuity at their heart, which has given a forty-five year archive of Earth observation imagery. However, as technological capabilities have developed the instruments on consecutive missions have improved. To demonstrate and celebrate this, NASA has produced a great video showing the changing coastal wetlands in Atchafalaya Bay, Louisiana, through the eyes of the different Landsat missions.

In total there have been eight further Landsat missions, but Landsat 6 failed to reach its designated orbit and never collected any data. The missions have been:

  • Landsat 1 launched on 23 July 1972.
  • Landsat 2 launched on 22 January 1975.
  • Landsat 3 was launched on 5 March 1978.
  • Landsat 4 launched on 16 July 1982.
  • Landsat 5 launched on 1 March 1984.
  • Landsat 7 launched on 15 April 1999, and is still active.
  • Landsat 8 launched on 11 February 2013, and is still active.

Landsat 9 is planned to be launched at the end 2020 and Landsat 10 is already being discussed.

Some of the key successes of the Landsat mission include:

  • Over 7 million scenes of the Earth’s surface.
  • Over 22 million scenes had been downloaded through the USGS-EROS website since 2008, when the data was made free-to-access, with the rate continuing to increase (Campbell 2015).
  • Economic value of just one year of Landsat data far exceeds the multi-year total cost of building, launching, and managing Landsat satellites and sensors.
  • Landsat 5 officially set a new Guinness World Records title for the ‘Longest-operating Earth observation satellite’ with its 28 years and 10 months of operation when it was decommissioned in December 2012.
  • ESA provides Landsat data downlinked via their own data receiving stations; the ESA dataset includes data collected over the open ocean, whereas USGS does not, and the data is processed using ESA’s own processor.

The journey hasn’t always been smooth. Although established by NASA, Landsat was transferred to the private sector under the management of NOAA in the early 1980’s, before returning to US Government control in 1992. There have also been technical issues, the failure of Landsat 6 described above; and Landsat 7 suffering a Scan Line Corrector failure on the 31st May 2003 which means that instead of mapping in straight lines, a zigzag ground track is followed. This causes parts of the edge of the image not to be mapped, giving a black stripe effect within these images; although the centre of the images is unaffected the data overall can still be used.

Landsat was certainly a game changer in the remote sensing and Earth observation industries, both in terms of the data continuity approach and the decision to make the data free to access. It has provided an unrivalled archive of the changing planet which has been invaluable to scientists, researchers, book-writers and businesses like Pixalytics.

We salute Landsat and wish it many more years!

If no-one is there when an iceberg is born, does anyone see it?

Larsen C ice Shelf including A68 iceberg. Image acquired by MODIS Aqua satellite on 12th July 2017. Image courtesy of NASA.

The titular paraphrasing of the famous falling tree in the forest riddle was well and truly answered this week, and shows just how far satellite remote sensing has come in recent years.

Last week sometime between Monday 10th July and Wednesday 12th July 2017, a huge iceberg was created by splitting off the Larsen C Ice Shelf in Antarctica. It is one of the biggest icebergs every recorded according to scientists from Project MIDAS, a UK-based Antarctic research project, who estimate its area of be 5,800 sq km and to have a weight of more a trillion tonnes. It has reduced the Larsen C ice Shelf by more than twelve percent.

The iceberg has been named A68, which is a pretty boring name for such a huge iceberg. However, icebergs are named by the US National Ice Centre and the letter comes from where the iceberg was originally sited – in this case the A represents area zero degrees to ninety degrees west covering the Bellingshausen and Weddell Seas. The number is simply the order that they are discovered, which I assume means there have been 67 previous icebergs!

After satisfying my curiosity on the iceberg names, the other element that caught our interest was the host of Earth observation satellites that captured images of either the creation, or the newly birthed, iceberg. The ones we’ve spotted so far, although there may be others, are:

  • ESA’s Sentinel-1 has been monitoring the area for the last year as an iceberg splitting from Larsen C was expected. Sentinel-1’s SAR imagery has been crucial to this monitoring as the winter clouds and polar darkness would have made optical imagery difficult to regularly collect.
  • Whilst Sentinel-1 was monitoring the area, it was actually NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) instrument onboard the Aqua satellite which confirmed the ‘birth’ on the 12th July with a false colour image at 1 km spatial resolution using band 31 which measures infrared signals. This image is at the top of the blog and the dark blue shows where the surface is warmest and lighter blue indicates a cooler surface. The new iceberg can be seen in the centre of the image.
  • Longwave infrared imagery was also captured by the NOAA/NASA Visible Infrared Imaging Radiometer Suite (VIIRS) on the Suomi NPP satellite on July 13th.
  • Similarly, NASA also reported that Landsat 8 captured a false-colour image from its Thermal Infrared Sensor on the 12th July showing the relative warmth or coolness of the Larsen C ice shelf – with the area around the new iceberg being the warmest giving an indication of the energy involved in its creation.
  • Finally, Sentinel-3A has also got in on the thermal infrared measurement using the bands of its Sea and Land Surface Temperature Radiometer (SLSTR).
  • ESA’s Cryosat has been used to calculate the size of iceberg by using its Synthetic Aperture Interferometric Radar Altimeter (SIRAL) which measured height of the iceberg out of the water. Using this data, it has been estimated that the iceberg contains around 1.155 cubic km of ice.
  • The only optical imagery we’ve seen so far is from the DEMIOS1 satellite which is owned by Deimos Imaging, an UrtheCast company. This is from the 14th July and revealed that the giant iceberg was already breaking up into smaller pieces.

It’s clear this is a huge iceberg, so huge in fact that most news agencies don’t think that readers can comprehend its vastness, and to help they give a comparison. Some of the ones I came across to explain its vastness were:

  • Size of the US State of Delaware
  • Twice the size of Luxembourg
  • Four times the size of greater London
  • Quarter of the size of Wales – UK people will know that Wales is almost an unofficial unit of size measurement in this country!
  • Has the volume of Lake Michigan
  • Has the twice the volume of Lake Erie
  • Has the volume of the 463 million Olympic-sized swimming pools; and
  • My favourite compares its size to the A68 road in the UK, which runs from Darlington to Edinburgh.

This event shows how satellites are monitoring the planet, and the different ways we can see the world changing.

Two Fantastic Remote Sensing Innovations

Aberdeenshire (Scotland) January 2016 flooding captured by Sentinel-1; Data courtesy of Copernicus/ESA

Aberdeenshire (Scotland) January 2016 flooding captured by Sentinel-1

Two academic remote sensing research announcements caught our eye this week. To be fair most remote sensing announcements catch our eye, but these two were intriguing as they are repurposing remote sensing techniques.

Remote Sensing the Human Body
Researchers at Kyoto University Centre of Innovation have developed a system based on spread-spectrum radar technology to remotely sense signals from the human body. They have focussed on heartbeats, although they acknowledge that other elements such as breathing and movement are also measured by the system. It uses a unique signal analysis algorithm to extract the beats of the heart from the radar signals, and then calculates the intervals to give the heartbeat.

Anyone who has ever needed to wear a Holter monitor for twenty-four or forty-eight hours will appreciate the advantage of having measurements taken remotely, in real time. In addition, under controlled conditions, the system has worked with a similar accuracy to an electrocardiographs (ECG). This will be music to the ears of regular ECG takers who know how much removing those sticky electrode pads can hurt!

This system is still at an early developmental stage and further testing and validation is necessary, but it offers a potential new use of remote sensing technology.

Remote Sensing & Social Media
Researchers from Pennsylvania State University have led a project developing an innovative way of combining social media and remote sensing. The research was undertaken on a flood in Boulder, Colorado in September 2013 with a particular focus on urban locations.

The team identified over 150,000 flood related tweets and used a cloud-based geo-social networking application called CarbonScanner, from The Carbon Project, to cluster the pictures from Twitter and Flickr to identify flooding hotspots. These were then used to obtain optical data, in this case from the high resolution commercial satellite Worldview 2 and the lower resolution, but freely available, Landsat 8.

A machine learning algorithm was developed to perform a semi-automated classification to identify individual pixels that contained water. As the data was optical it used the near infrared band as, due to its strong absorption, water is easily distinguishable from soil and vegetation. The researchers believe that this methodology has the potential to give emergency teams near real-time data, which could make live-saving differences to their work.

This is a particularly interesting development for us, given our current work on flood-mapping using synthetic aperture radar (SAR) data as part of the Space for Smarter Government Programme.

These two current examples show that remote sensing is an exciting, innovative and developing field, and one that is not solely related to Earth observation.

Evolution of Coastal Zones

Lost Lake Area of Louisiana, USA. Landsat 5 image from 1985 on left, Landsat 8 from 2015 on right. Data courtesy of NASA/USGS.

Lost Lake Area of Louisiana, USA. Landsat 5 image from 1985 on left, Landsat 8 from 2015 on right. Data courtesy of NASA/USGS.

Coastal zones are the place where the sea and the land meet, and they’ve played a massive role in the life of Pixalytics. From a personal standpoint we’re based, and live, in Plymouth on the south-west coast and anyone who saw the Dawlish railway tracks swinging in midair eighteen months ago will know how these areas can affect our transport links. In addition, Sam’s PhD was focussed on the ‘Remote Sensing of Suspend Sediment in the Humber Estuary’, and so Pixalytics has effectively been grown from a coastal zone!

Last week the BBC carried a report highlighting the erosion of the Louisiana coastal wetlands; in particular, it noted that more than an area the size of a football pitch was disappearing every hour. This statistic caught our attention, and our next steps were obvious! We downloaded two images of the Lafourche Bayou in Louisiana; the first was a Landsat 5 image acquired on the 31st August 1985, and the second was a Landsat 8 image acquired twenty years later on the 02nd August 2015.

Mouth of Atchafalya River, Louisiana, USA. Landsat 5 image on left from 1985, Landsat 8 image from 2015 on right. Data courtesy of NASA/USGS.

Mouth of Atchafalya River, Louisiana, USA. Landsat 5 image on left from 1985, Landsat 8 image from 2015 on right. Data courtesy of NASA/USGS.

The image at the top of the blog shows the area around the Lost Lake, in the bottom left hand corner, just off the coast of Louisiana; with the 1985 image on the left, and the 2015 image on the right. The loss of land, described in the BBC report, can be seen in the northern portion of the image with a lot more water visible. However, the image on the right shows the mouth of the Atchafalya River in Louisiana; again, the 1985 image is on the left. Coastal evolution is again clearly visible, but this time there are islands that have risen from the water.

Swamplands, like in Louisiana, aren’t the only coastal zones changing. In 2011, the United Nations Environmental Programme estimated that over the last 40 years Jamaica’s Negril beaches have experienced average beach erosion of between 0.5 m and 1 m per year. Another coastal zone in decline are mangroves and wetland forests; a 2007 report noted that the areal extent of mangrove forests had declined by between 35 % and 86 % over the last quarter half century (Duke et al. 2007).

Coastal zones have social, economic and environmental importance as they attract both human settlements and economic activity; however, they are also particularly susceptible to the impacts of climate change and their evolution will have impacts on the human, flora and fauna populations of those areas. So when you’re next at the coast have a good look around; the view in front of you may never be seen again!

Landsat Showing What The Eye Can’t See

Landsat 8 True colour composite of Paris from 11/11/14. Courtesy NASA/USGS.

Landsat 8 True colour composite of Paris from 11 November 2014. Courtesy NASA/USGS.

In our recent blog we described the five simple steps to select, download and view LandsatLook Natural Colour Images. However, did you know that the Natural Colour Image isn’t actually a single image? Instead, it’s a combination of three separate images!

This is because remote sensing works by exploiting the fact that the Earth’s surfaces, and the substances on it, reflect electromagnetic energy in different ways. Using different parts of the electromagnetic spectrum makes it possible to see details, features and information that aren’t obvious to the naked eye. Some remote sensing satellites carry instruments that can measure more than part of the electromagnetic spectrum, with each different measurement known as a spectral band.

Landsat 8 currently has two instruments, measuring eleven different spectral bands:

  • Three visible light bands that approximate red, green and blue
  • One near infrared band
  • Two shortwave infrared bands
  • Two thermal bands used for sensing temperature
  • Panchromatic band with a higher spatial resolution
  • The two final bands focus on coastal aerosols and cirrus clouds.

Combing the red, green and blue bands produces a single image that is very similar to what your eye would see; and this composite is the Natural Colour Image product that Landsat offers. However, you can also create your own colour composites using Image Processing Software, as Landsat offers the possibility of downloading an image for each of the individual spectral bands, known as the Level 1 GeoTIFF files.

Once imported into an image processing package, it’s straightforward to create different composites by combining different variations of the spectral bands. For example, combing the red, green and blue bands creates an image like the one at top of the blog showing the eastern edge of Paris, with the Bois de Vincennes, the largest public park in Paris, on the left hand side.

This image has colours your eyes expect to see, for example, trees are green, water is blue, etc, known as a true colour or RGB composite. Combining other spectral bands produces images where the colours are different to what you would expect, these are known as false colour composites. As they use different parts of the electromagnetic spectrum, the surface of the earth reacts differently to the light and allows features hidden when showing true colour to become far more prominent.

Landsat 8 False colour composite of Paris from 11 November 2014. Courtesy NASA/USGS.

Landsat 8 False colour composite of Paris from 11 November 2014. Courtesy NASA/USGS.

An example of a false composite can be seen on the right, it uses the near infrared, red and green bands. Like in the RGB image, the park is easily distinguishable from the surrounding Paris; but in the false colour image, the park’s water features of the Lac Daumesnil and the Lac des Minimes have become visible as black swirls.

Landsat 8 False colour composite of Paris from 11 November 2014. Courtesy NASA/USGS.

Landsat 8 False colour composite of Paris from 11 November 2014. Courtesy NASA/USGS.

A second example of a false colour composite is shown on the right, which this time combines the near infrared, shortwave infrared 2 and the coastal aerosol band. In this case, the vegetation of Paris appears orange and jumps out of the image when compared to urbanisation shown in blue.

Using different combinations of spectral bands is just one remote sensing technique to create valuable information and knowledge from an image. However, every satellite measures different spectral bands and you need to be aware of what you are looking at. For example, we’ve described Landsat 8 in this blog, previous Landsat missions have measured similar, but slightly different spectral bands; full details of all Landsat missions and their spectral bands can be found here.

Using the individual spectral bands, rather than relying on the set Landsat products, means you may gain new insights into the area you are looking at and you can great some fantastic images. You can literally make things appear before your eyes!

Is space a good investment?

Space is an expensive, and uncertain, environment to work in, and decisions to invest in space technology and missions are frequently questioned in the current global economic climate. Headline figures of tens of millions, or billions, do little to counter the accusations that there are more appropriate things to be investing in. Is the cost of investing in space worthwhile?

Image of East Devon, UK taken by Landsat 8 on 4th November 2013.  The River Exe flows from top to bottom and the River Teign from left to right. Plumes of suspended sediment are clearly visible following periods of heavy rainfall in late October and early November 2013.  Image courtesy of the U.S. Geological Survey

Image of East Devon, UK taken by Landsat 8 on 4th November 2013.
The River Exe flows from top to bottom and the River Teign from left to right. Plumes of suspended sediment are clearly visible following periods of heavy rainfall in late October and early November 2013.
Image courtesy of the U.S. Geological Survey

Last week the Landsat Advisory Group, a sub-committee of the US Government’s National Geospatial Advisory Committee, issued a report looking at the economic value of Landsat data to America. As Landsat data is freely available, quantifying the value of that data isn’t easy; and the Group approached it by considering the cost of providing alternative solutions for Landsat data.

They considered sixteen applications, linked to US Government departments, which use Landsat data. These ranged from flood mitigation, shoreline mapping and coastal change; through forestry management, waterfowl habitats and vineyard management; to mapping, wildfire assessment and global security support. The report estimated that these sixteen streams alone produced savings of between $350 million and $436 million to the US economy. The report concluded that the economic value of just one year of Landsat data far exceeds the multi-year total cost of building, launching, and managing Landsat satellites and sensors.

This conclusion was interesting given reports in 2014 that Landsat 8 cost around $850m to build and launch, a figure which will increase to almost $1 billion with running costs; and that NASA were estimating that Landsat 9 would cost in excess of the $650m budget they had been given. These figures are significantly in excess of the quantified figures in the Advisory Group report; however work undertaken by US Geological Survey in 2013 identified the economic benefit of Landsat data for the year 2011 is estimated to be $1.70 billion for US users, and $400 million for international users.

The discrepancy between the two figures is because the Advisory Group did not include private sector savings; nor the fact that Landsat data is also collected, and disseminated, by the European Space Agency; nor did it include unquantified societal benefits or contribution to scientific research. For example, it highlighted that humanitarian groups use Landsat imagery to monitor human rights violations at low cost and without risking staff entering dangerous, and often inaccessible, world regions.

Last week also demonstrated the uncertain side of space, with the discovery of the Beagle-2 spacecraft on the surface of Mars. The UK led probe mission was assumed to have crash landed on Christmas Day 2003, however recent images indicate it landed successfully but its solar panels did not unfurl successfully. The Beagle 2 discovery has obvious echoes with the recent shady site of the Philea comet landing, and demonstrates that space exploration is a risky business. Given the Beagle 2 mission cost £50 million and the Philea mission was estimated to cost around region of €1.4 billion, is the cost of investing in space worthwhile?

Consider satellite television, laptops, smoke detectors, tele-medicine, 3D graphics and satellite navigation – all of these developments came through the space industry, and so now think about the jobs and economic activity generated by these sectors. Working in space is expensive and challenging, but it’s precisely because of this that the space industry is innovative and experimental. The space sector works at the technological cutting edge, investment in space missions benefits and enhances our life on earth. So if anyone ever asks whether space is a good investment, tell them about the financial benefits of Landsat, the development of laptops, the number of lives saved by smoke detectors or the humanitarian support provided to Amnesty International.

Remote Sensing Big Data: Possibilities and dangers

Remote sensing is an industry riding the crest of the big data wave. It offers great opportunities to those that can harness the power, but it’s also fraught with dangers. Big data is a blanket term used to describe datasets that are large and complex, due to the quantity of data, the speed at which new data becomes available or the variety of data. Remote sensing ticks all three of these boxes!

Sentinel-1 Netherlands

Sentinel-1 image of the coast of the Netherlands; courtesy of ESA

When I first started working with remote sensing, I approached the IT department to ask for 100 megabytes of disk space for my undergraduate project and was told nobody ever needs that much storage! Currently, the amount of Earth observation data available to the community is growing exponentially. To give you some examples, the recently launched Copernicus Sentinel 1-A satellite collects around 1.7 terabytes of data daily, the number of daily images collected by Landsat 8 has been increased by 18% this month and DigitalGlobe estimates it captures two petabytes of data each year. This quantity of data gives two key challenges; firstly, where to store it? Secondly, how do you know what data is valuable to enhance your decision-making?

It’s assumed the storage issue has been resolved by cloud computing, but there is a cost for getting the data to, and from, the cloud. An interesting recent study by the University of British Columbia discovered that over 80% of scientific data is lost within 20 years, mostly due to obsolete storage devices and email addresses. I have first-hand experiences of this. My PhD data was stored on hundreds of floppy disks and when I came to use them recently most didn’t work; fortunately I have a zip drive backup – although I still need to work out how to read Quattro Pro spreadsheets! I also have several Sun workstations with associated data on tapes which will only read from the machines they were written on; so how much of this data is accessible is debatable.

How often do you think about your old and archived data? Take a moment to consider how, and where, your critical data is stored. Is all of your data available and accessible? When was the last time the back-up procedures for your scientific or business data were tested? Does your IT department know which email addresses are critical for the receipt of satellite data?

The second challenge is knowing what data to use, particularly for people new to remote sensing. There is free data, paid for data, various satellites, various data types, various formats and the list can go on. The remote sensing community needs to help by providing more bridges between the data and the user community. The datasets available can offer huge benefits for business and science, but if people have to spend hours hunting round and trying to find the right image for them, they won’t stay users for long.

You can hire remote sensing companies, like us, who can offer impartial advice to help you select the right information. Pixalytics is striving to find more ways to make data more available, more accessible and more understandable. Remote sensing data belongs to everyone, and we need to support users to get it.

The Question of Dredging

Dredging has been a hugely contentious issue in the UK ever since the St Jude’s storm hit the country on the 28 October 2013; this marked the beginning of a relentless winter weather pattern of heavy rainfall and high winds. This severe weather coupled with coastal surges breaching flood defences led to large parts of the UK to be under water – a situation that still exists for significant parts of Somerset. Satellite data was used to map the flooded areas as part of the flood response by UK government agencies; more details can be found in our post Is the Southern UK Flooding a Disaster?

As the flooding occurred local communities bemoaned the lack of river dredging in recent times, and they felt this was a significant contributing factor for the rising water levels. In Prime Minister’s Questions on the 29th January this year, David Cameron announced that once flood waters in Somerset had drained away, rivers in the county will be dredged. Dredging itself also creates problems, releasing suspended sediment into the river water and secondly the need to get rid of the dredged material.

Image of East Devon, UK taken by Landsat 8 on 4th November 2013.  The River Exe flows from top to bottom and the River Teign from left to right. Plumes of suspended sediment are clearly visible following periods of heavy rainfall in late October and early November 2013.  Image courtesy of the U.S. Geological Survey

Image of East Devon, UK taken by Landsat 8 on 4th November 2013.

The River Exe flows from top to bottom and the River Teign from left to right. Plumes of suspended sediment are clearly visible following periods of heavy rainfall in late October and early November 2013.

Image courtesy of the U.S. Geological Survey

In addition to flood mapping, satellite data can also be used to map and monitor sediment transported. The Landsat 8 image on the right shows the plumes of sediment visible around the east Devon coastline just one week after the St. Jude’s storm. Since 1972 the Landsat mission has continuously monitoring the Earth’s surface; and makes this information freely accessible for use across a range of sectors.

This week it was announced that dredging of the River Tamar, on the border between Devon and Cornwall, will continue for the next two years in order to keep the channels clear for access to Devonport Dockyard. The silt from this process will be deposited in Whitsand Bay, Cornwall, despite the area being designated as a Marine Conservation Zone.

Dredging is a tool coming back into the UK flood defence armoury; the benefits, and potential harm, will be monitored closely in the coming months and years.

Next week’s we’ll be looking at the accelerated coastal erosion from the winter storms.

Blog produced by Bryony Hanlon, work placement student with Pixalytics, and Andy Lavender.