In 2006, a group of Bahraini activists used Google Earth to view palaces and land owned by the elite that was situated next to overcrowded cities. This exploration revealed vast stretches of unused open spaces along with luxury properties and golf courses all owned by the ruling minority. The group compared these big, empty spaces with the crowded, built up areas that the majority of people lived in.
In response the Bahraini government blocked Google Earth in order to prevent further dissemination. The group circumvented this by distributing emails containing an anonymous PDF of these satellite images which contributed to growing unrest about the lack of affordable housing.
Without this view from above many of these properties were hidden behind walls and 'no trespassing' signs. They were out of sight, subject to the phenomenon of not being able to see the forest through the trees, something which an aerial perspective overcomes, moving the viewer above the trees to make observations and comparisons not possible on the ground.
Or as Mahmood Yousif, who is known as the grandfather of the Bahraini blogging community stated to the Financial Times, “Some palaces take up more space than three or four villages nearby and block access to the sea for fishermen. People knew this already. But they never saw it. All they saw were the surrounding walls."
Google Earth images from the PDF report circulated by activists in 2006. To see more about this campaign and the other 45 PDFs that were created see here.
Through using this example of land abuse in Bahrain we will look at how to start thinking about working with satellite imagery.
Anyone who has undertaken an investigation using satellite imagery will tell you to pose the problem or question first that you are interested in investigating, and then to justify why satellite imagery is a useful tool to demonstrate these questions. Is there something about the perspective, spectral bands, or time change that satellite imagery specifically would help to show?
Locate the coordinates or address of the geographic area of interest and start working with Google Maps and Google Earth. It doesn't hurt to browse through Google Maps, even if that is not your last destination. It is fast, and will quickly give you an idea of the resolution that is available. The downside of using Google Maps for sourcing satellite imagery is that there is very little flexibility and it is difficult to verify because the images are stitched together, and don't list the detailed source metadata.
Below you will see the result of searching for the city Malkiyya through the browser using Google Maps:
Screenshot of Malkiyya, Bahrain, Google Maps
As you can see, it is quite similar to the image used for the campaign material. However if we want to do more with this image, such as, comparing it to the original and finding out the actual dates the images were taken, then we have to move to Google Earth.
Google Earth is a desktop application that works on Linux, Mac and Windows operating systems. It can be easily downloaded and locations can be found through entering them in through the search bar.
This becomes instantly useful as we now have more metadata to work with. Through this we know that the image was taken on 12 October 2015. To see how this scene has changed, and in this case how the land use has changed over time, we can use the 'Historical Imagery' option, shown in the image below.
Screenshot of Malkiyya, Bahrain, 12 October, 2015, Digital Globe, Google Earth
By scrolling along the timeline of imagery we can observe changes over time. Something to be aware of when using any satellite imagery viewing platform: within a given frame of view there can be different images stitched together that actually have been taken on different dates. Sometimes it appears seamless, so that in this case you can only be sure by scrolling over the frame with the cursor and watching for a date change, but the example frame below clearly shows that we are working with two different images that were taken at two entirely different times.
Screenshot of Malkiyya, Bahrain, Digital Globe, Google Earth
It is important to keep in mind when using Google Earth that firstly, it can be a prognosis tool to lead you to the next step in an investigation, or a place for corroboration. In short, satellite imagery is a tool to be used together with other datasets and observations on the ground. When Mishka Henner created Feedlots, he used data from State and National Feedyard directories and associations available online to guide his investigation. In this case, accessing the property records to point out areas of interest and places to observe over time helped to reveal injustices in land ownership in Bahrain.
Occasionally one does stumble upon surprises by using satellite data. For example, we zoomed out to see if it was possible to see changes on a macro scale to Bahrain. By scrolling from 2003 to 2014 there is stark changes in land mass to be observed:
Screenshot of Bahrain, 10 February 2004, Google Earth Screenshot of Bahrain, 10 December 2015, Google Earth
When we zoomed in to investigate further there was what appeared to be a bright red rectangle with writing on it:
Screenshot of Diyar Al Muharraq, 29 November 2015, Google Earth
Screenshot of Diyar Al Muharraq, 29 November 2015, Google Earth
This 'message,' visible from satellite, reads “Diyar Al Muharraq,” which after a quick internet search, turns out to be a new city in the process of being built. In their own words, “Diyar Al Muharraq embraces the water and land elements with a breath-taking combination of an exclusive and stylish 21st century lifestyle built around interconnecting waterways, public beaches, parks and promenades. It offers a promising investment opportunity in an emerging commercial hub for a range of industries such as retail malls, five-star hotels and logistics centres.”
From this observation, it appears that not much has changed with regard to inequality and housing distribution since the original campaign in 2006. Indeed, in our research done in 2013 we already found that, “Despite the wide circulation and exposure of the visual evidence, there was no direct change in the distribution of land in Bahrain, or in administrative transparency. This illustrates one of the traps that campaigners and activists often fall into. In campaigning terms, we tend to look for a direct 'cause and effect' reaction, and we ask when the impact of such an intervention may have been. But we may be missing an important point: some interventions are essential catalysts, but when an issue needs a long-term approach to change, with incremental steps along the way, our campaigns may be best understood and appreciated for their ability to get the fire started or for the role they play in affecting direction."
Thus, be specific about the intent. This might be a stopping point and provide enough evidence to have impact or raise awareness – or it may just be the tip of the iceberg and you'll want to dig deeper.
Google Earth can be a great exploratory tool for the onset of an investigation, especially with features such as 'Historical Imagery.' There are however some shortcomings. There are less ways to explore the image further, or see more metadata, as the only export file format is jpg. In traditional satellite imagery files (GeoTIFF, for example) there are different light bands to observe different elements other that RGB True-colour.For example, near infrared (Band 5 in Landsat imagery) is often used in observing changes in vegetation or forestry. For an example of the increased options of other satellite imagery platforms, see the images below from the Landsat download tool
Downloading options for spectral bands from the Libra tool for downloading Landsat imagery
This table below is a useful starting point of the different satellite imagery sources, what type of images they are most suited for, differences in resolution and time spans and also, importantly, the cost.
|Type of imagery||Provider||Resolution (m)||Temporal coverage (frequency)||Temporal span||Cost||Mostly suitable for|
|MODIS||NASA||500||1-2 days||1999- present||Free||Vegetation monitoring, atmospheric conditions|
|Landsat 1-8||NASA||30 (varies)||16 days||1972- present||Free||Urban extant, land use change, vegetation monitoring, atmospheric conditions|
|ASTER (infrared bands)||NASA, METI, ERSDAC||
|16 days||1999- present||Free in the US||Urban extant, land use change, vegetation monitoring, atmospheric conditions|
|CBERS 1-2||China, Brazil||20||26 days||1999- present||Free||Urban extant, land use change, vegetation monitoring, atmospheric conditions|
|ICESAT/GLAS (LIDAR; elevation data only)||NSDIC||60||91 days||2003-2009||Free||Ice, cloud, and land elevation|
|Skybox||Skybox Imaging, Google||0.9||N/A||2014- present||Case-by-case||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters, houses, roads, automobiles, large groups of people|
|Planet Labs||Planet Labs||4||N/A||2014- present||Case-by-case||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters, houses, roads|
|GeoEye-1||DigitalGlobe||1.65||8.3 days||2008- present||Varies; Around $25 per km2; Grants available||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters houses, roads, automobiles, large groups of people|
|IKONOS||DigitalGlobe||3.2||3 days||1999- present||Varies; Around $25 per km2; Grants available||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters houses|
|Worldview 1- 3||DigitalGlobe||1.24||1.7 days||2007- present||Varies; Around $25 per km2; Grants available||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters houses, roads, automobiles, large groups of people|
|Quickbird||DigitalGlobe||2.62||2.5-5.6 days||2001- present||Varies; Around $25 per km2; Grants available||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters houses, roads, automobiles, large groups of people|
|RapidEye||Blackbridge||6.5||5.5 days||2009- present||Area dependent||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters, houses|
|SPOT 1-6||Airbus||1.5 (varies)||daily||1986- present||Varies; Around $5,000 per scene (60km x 60km)||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters, houses, roads, automobiles, large groups of people|
|Pleiades 1A- 1B||Airbus||2||26 days (taskable to daily)||2011- present||Varies; Around $20 per km2||High-detail agricultural monitoring, urban-area land use, refugee movements, commodity storage, natural disasters, houses, roads, automobiles, large groups of people|
Copyright 2016 Innovations for Poverty Action. Micro-satellite Data: Measuring Impact from Space is made available under a Creative Commons Attribution-NonCommercialNoDerivatives 4.0 International License.
It is also worth mentioning that there is the option of NGOs and researchers collaborating directly with satellite imagery providers. This publication on using satellite imagery from Innovations for Poverty Action provides a great overview of this.
Lisa Gutermuth, is a Programme Coordinator at Tactical Technology Collective. Her research focuses on the intersection of satellite imagery and agriculture.