August 2021 Hurricane activity, from NASA Earth Observatory
This past year marked a major milestone in SkyTruth’s history. As we celebrated our 20th anniversary, we looked back to our humble beginnings while developing new technologies to tackle today’s most pressing environmental challenges. Equally exciting, we set the stage for tremendous growth in SkyTruth’s capacity — not just in our technical capabilities, but in our outreach to partners and others who can use SkyTruth tools to change the world.
We are poised for impact in our 21st year in a way that would have been hard for me to imagine in 2001 when I founded SkyTruth. Back then, leaving my steady job at a consulting firm took a leap of faith — that trading in my work analyzing satellite imagery for energy and mineral companies would pay off in other ways. Namely, in helping to protect people and the planet by leveling the playing field and providing small conservation groups the same tools large corporations use to exploit natural resources. In those early years I worked alone in the unfinished basement of my home outside Washington, D.C. and analyzed images on my desktop computer, one at a time.
Despite the limitations of technology in those early years, SkyTruth managed to help citizen’s groups in the Rocky Mountain West fight intensive oil and gas drilling overtaking public lands. In time, SkyTruth and our reach grew; we influenced research and policy related to hydraulic fracturing (fracking), mountaintop mining for coal, and the Deepwater Horizon oil spill, among other issues. We took another big step forward in 2015 by creating Global Fishing Watch (GFW) in partnership with Google and the international conservation group Oceana. GFW tracks fishing vessels and other activities in the world’s oceans to provide transparency and accountability for sound management and protection. We launched GFW as an independent nonprofit in 2017, and since then it has grown into a team of more than 80 experts around the globe.
Now, we are poised for even greater impact. As we continue to refine our machine learning capabilities, we can process thousands of satellite images a day from all over the world, allowing us to detect activities even in remote locations – from the world’s oceans to tropical rainforests. Our innovation and success haven’t gone unnoticed: we’ve attracted multi-year funding to grow our team, develop our groundbreaking tools further, and reach out to even more partners in government, academia, industry, and the nonprofit world.
With the addition of a Director of Impact and Strategic Partnerships, a Chief Technology Officer, and a Communications Manager in 2022, SkyTruth is well positioned to take the next step forward, amplify our reach, and realize our vision:
Everyone around the world enjoys a healthy environment and sustainable livelihood because transparency is the norm: polluters know they will be seen and caught, industries have cleaned up their practices, and governments are vigorously enforcing environmental protection.
Our vision is bold and SkyTruth is road-tested: we’ve taken big leaps forward with great success. I hope you’ll join us for our next exciting and important phase!
Satellite image in SkyTruth Alerts of the Telfer Mine in Western Australia, from July 2021 Planet Basemap.
Over the past two decades, SkyTruth has emerged as [one of] the most creative and effective organizations in the emerging use of satellite imagery in support of the environmental movement.
2021 Highlights
- Deployed our machine learning applications for partners and an engaged public to stop oil pollution at sea and illegal gold mining in tropical ecosystems.
- Automated a process for identifying harmful algal blooms.
- Published our analyses of landscape recovery and regeneration in areas of Appalachia impacted by mountaintop mining, and published our analysis of the impact of mining pollution on imperiled species.
- Mapped vegetation trends in the Colorado River basin to help watershed managers protect riparian habitats.
- Enabled citizen scientists and environmental advocates to monitor human impacts on the environment using an enhanced SkyTruth Alerts platform with new data layers, additional maps, and improved performance.
- Strengthened our internship program to continue advancing diversity, equity, and inclusion in the fields of conservation and technology.
Satellite image in SkyTruth Alerts of the Tesiin Gol (Tes River) in Mongolia, from August 2021 Planet Basemap.
SkyTruth is the world’s leading sleuth of satellite imagery and the leading interpreter of what it all means.
Program Updates
Securing Healthy Waters and Oceans
Stopping Oil Pollution at Sea with Cerulean
Oil slicks detected during Cerulean’s first year of operation.
In 2021 we added new deep learning techniques to the toolkit of Cerulean – our project to detect bilge dumping and stop oil pollution at sea. Bilge dumping occurs when cargo vessels and tankers illegally dump oily bilge water into the ocean. Bilge waste forms in most modern cargo and container vessels when fuel oil produces a thick sludge that drains into the bottom of the vessel and needs to be emptied regularly. International law specifies how bilge water should be treated to protect ocean ecosystems, but SkyTruth has shown that many ships bypass costly pollution prevention equipment by simply flushing the bilge waste directly into the sea. We began developing project Cerulean in 2020 to automate the detection of oily slicks likely caused by bilge dumping, and highlight another – often overlooked – impact of fossil fuel use.
We have trained Cerulean’s machine learning model to sift through hundreds of thousands of satellite images per year and identify oil slicks from vessels. By using Automatic Identification System (AIS) broadcasts, we’re often able to identify vessels potentially responsible for these slicks. In 2021 we began conversations with several government, nonprofit, and investigative journalism groups – including the U.S. National Oceanic and Atmospheric Administration and Greenpeace – hoping to use our dataset to expose and combat vessel pollution at sea. We also worked with commercial imagery providers to gain access to more satellite radar imagery and cover larger swaths of open ocean.
According to our estimates, vessels might be dumping more than 1.3 million barrels of engine oil and sludge per year; the equivalent of the entire Exxon Valdez spill every three months. Our machine learning model identified over 100 vessel slicks each month, giving us a better feel for the true scale and distribution of the problem. In addition to geographic insights, like the hotspot analysis shown above, Cerulean also positively identified 153 of the responsible vessels and found that 56% of polluters were tankers and 30% were cargo vessels. The majority of the polluting vessels were flagged in Russia followed closely by Panama.
Detecting Harmful Algal Blooms
The Muskingum Watershed Conservancy District in eastern Ohio analyzed with the KSU spectral decomposition method. Red = high levels of harmful algae.
Over the past three years SkyTruth has collaborated with Professor Joseph Ortiz’s lab at Kent State University (KSU) to implement a novel water quality analysis that identifies harmful algal blooms (HABs) using satellite imagery in Google Earth Engine. Harmful algal blooms affect every coastal and Great Lakes state in the U.S. and are a problem worldwide. Usually algae are harmless. But under the right conditions, they can grow out of control, and some release toxins that kill wildlife and make people sick. Different types of algae, suspended sediment, and other aquatic material can make it difficult to pull the signal (such as cyanobacteria, the algae often responsible for algal blooms) out of the noise (all of the other stuff) to detect when blooms might be harmful. The KSU method makes it possible to separate these different components – a process called spectral decomposition.
In 2020 we implemented this process in Google Earth Engine, but could only analyze a single image at a time. And so, in 2021, we focused on adapting the existing analysis to work on multiple images. With multiple images, we could illuminate conditions over time — such as an entire season — rather than simply a snapshot in time. This seasonal information provides water resource managers with additional information about the composition of algae in water bodies to guide their decisionmaking.
We completed this initial phase of work in the early summer, and then started the process of automating the entire workflow to analyze individual images using the combined power of Earth Engine and Google Cloud Platform (GCP). This allowed us to automate the entire process, eliminating intermediary steps that required human analysis. We also completed work for the Florida Department of Environmental Protection that used the KSU spectral decomposition method, Earth Engine, and GCP. Although that work focused on mapping the benthic characteristics of Biscayne Bay in southern Florida, we ended up making several improvements to the spectral decomposition workflow in Earth Engine. We rewrote our code to process several images at a time, make it easier to review summary plots and graphs, and to average intermediate outputs to create a final product. Ultimately, our work in Biscayne Bay paved the way for us to automate the analysis of multiple images without errors, and we are now able to run images through KSU’s algorithm without manual input.
Our work with KSU could easily be scaled up and applied globally and seasonally to reveal the presence of different types of algae, inform water managers, and control pollution in other imperiled water bodies.
Restoring Clean Air and a Stable Climate
Measuring Recovery from Coal Mining
SkyTruth remains committed to updating and maintaining our inventory of surface coal mining in Appalachia to document many of the harmful impacts of this dangerous fossil fuel. In 2021, we published our latest annual update to our mining extent dataset. This work showed the continuation of a trend we’ve seen since 2008: although coal production is in decline, the area of land disturbed to extract coal is increasing. But our work in recent years goes beyond mapping coal mining’s footprint.
In collaboration with our partners at Appalachian Voices and Defenders of Wildlife, we began examining the long-term ecological impacts of mines on local ecosystems. In May, working with our partners at Appalachian Voices, we released the Central Appalachian Mine Reforestation Assessment report and data, the first comprehensive assessment of mine land recovery across the region. The dataset and associated webmap allow citizens and regulators to examine the ecological conditions of a site, and compare them to a healthy forest. This work creates transparency in the process of bond release — in which mining companies are freed from further reclamation requirements — and landscape recovery.
Throughout the year, we worked closely with researchers at Appalachian Voices, the University of Nevada, and Colorado State University to expand our mine land recovery analysis. Currently, we’re preparing a peer-reviewed publication that offers a nuanced understanding of how these areas recover and the factors that contribute to recovery.
In November, working with our partners at Defenders of Wildlife, we published new research that links mountaintop removal mining to water quality degradation in Appalachian rivers and streams that are home to endangered species. This research revealed that thresholds for toxic pollutants were exceeded thousands of times since 1985 in streams across the region. This result represents the next step in linking the practice of surface mining to its true impact on surrounding ecosystems, and hopefully will lead to stricter permitting for new or expanded mines to help protect native species and ecosystems in one of the most biodiverse regions on Earth.
Ending Natural Gas Venting and Flaring
In 2021, the Colorado School of Mines’ Earth Observation Group (EOG), which provides the satellite data that powers SkyTruth’s flaring maps, limited access to their data so that it was only available to partners they have approved. SkyTruth applied and was granted permission to continue using EOG’s data. In addition, for the first time SkyTruth can provide users with EOG flaring data in a summary form. These changes allow users to zoom in on their areas of concern more easily and analyze month-to-month changes directly.
Also in 2021, SkyTruth rewrote the flaring backend – where data is downloaded from EOG and processed for the maps – and moved it off of two Rackspace servers onto Google Cloud Platform. In doing so, we were able to greatly reduce the amount of code and resources required. This also allows staff to more easily keep the server-side software updated and to implement new security features recently put into place by the EOG.
Safeguarding lands and communities
Mapping Vegetation Trends and Recovery in the Colorado River Basin
In 2021 our work in the Colorado River basin focused on connecting vegetation trends over the last 35 years to watershed management and restoration practices in a region experiencing a prolonged megadrought – caused in part by climate change.
In June of 2021, we released our 35-year time series dataset of vegetation trends for the Colorado River basin using Landsat satellite imagery. And in November, we released our Verde River map — a demonstration of what’s possible on a larger scale – that combines our 35-year trend dataset with high-resolution satellite imagery basemaps from Planet Labs. This map allows users to investigate “hot” and “cool” spots of vegetation loss and recovery in detail, enabling them to easily monitor even remote areas.
The dataset includes seven different spectral indices for the entire Colorado River drainage which provide crucial information about soil and vegetation conditions on the ground: the Normalized Difference Vegetation Index, Enhanced Vegetation Index, Soil Adjusted Vegetation Index, Modified Soil Adjusted Vegetation Index, Normalized Difference Moisture Index, Normalized Burn Ratio, and Normalized Burn Ratio 2. More details about the different indices can be found here, and the data can be downloaded as a series of CSV files here.
The CSV file contains 7,558 rows (including every subwatershed in the Colorado River drainage) and 471 columns – all 35 years of all spectral indices plus some basic identifying information for each subwatershed. We’ve also included a shapefile containing geometries for all of the subwatersheds in the dataset. We hope that watershed managers can use this new data to identify which particular restoration techniques generate the landscape responses that they want, so they can make better choices in the future about targeting their efforts and resources.
Expanding and Improving SkyTruth Alerts
Many skytruthers are familiar with our SkyTruth Alerts platform. It provides an easy way for anyone to subscribe to an Area of Interest (AOI), and then receive alerts whenever certain environmental incidents, such as a new drilling permit, a hazardous chemical spill, or a change in land cover, occurs in those areas.
Throughout 2021, we made numerous upgrades to make the platform faster and more user-friendly. First, we’ve used a new Javascript API called WebGL to speed up the way that Google Maps renders datasets, and we’ve added support for vector tiles to make rendering large datasets much quicker and easier. Secondly, we’ve replaced Sentinel Hub’s map tiles with map tiles pulled directly from Google Earth Engine. These technical changes speed-up the response time in Alerts so users get their results quicker.
We’ve also added several new features to the Alerts platform. Most importantly, we’ve added many new layers, including layers that identify Exclusive Economic Zones in the world’s oceans, protected areas around the world, and 12 layers related to energy production from the National Energy Technology Lab, including gas pipelines and wells, storage sites, ports, and more. We also produced several new public maps that highlight a variety of features of interest, including large oil spills, oil pipeline incidents in Louisiana, oil spills from the Taylor Energy site, vessel strandings, and more — maps that are available to any Alerts user. We’ve also made it possible for users to create their own custom, user-generated Alerts (and maps) and publish them to other users of the Alerts system. And we’ve made it easier for users to organize, collaborate, and save their maps. Users can group themselves into organizations, and those organizations can create their own features such as their own Alert types. We’ve also deployed a Spanish-language version of the Alerts platform to better support our users in Central and South America.
Map of the October oil pipeline spill off the coast of Huntington Beach, California created in SkyTruth Alerts. This map appeared in at least 18 media stories, including the CBS Evening News.
We incorporated more satellite imagery into SkyTruth Alerts this year, including high-resolution, monthly basemaps for the world’s tropics from Planet Labs that allow users to visually monitor land use changes in areas of interest month-to-month. We’ve also added support for creating spectral indices like the Normalized Difference Vegetation Index using Sentinel-2 and Landsat imagery to help detect land use changes. And we’ve included elevation data and greatly expanded map comparison options so that any layer or satellite image can now be compared to any other, allowing users to select and directly compare images from different dates. These changes allow users to identify places where the landscape has changed due to land clearing, logging, drilling, mining or other activities. We also made it easier for users to add notes, drawings, and other types of annotation to the images, and create and share their own custom maps based on their observations.
SkyTruth President John Amos used many of the new layers in Alerts to analyze the possible cause of the Huntington Beach oil spill in October 2021, including data on offshore oil pipelines and water depth. At least 18 media outlets used his annotated map, including the CBS Evening News.
Protecting a Legacy of Biodiversity
Detecting Illegal Mining in the Amazon
SkyTruth’s project Inambari was chosen as one of the winners of Conservation X Labs’ 2020 Artisanal Mining Grand Challenge. Small-scale mining for gold in the Amazon basin has become a dangerous threat to Indigenous people and biodiversity in recent decades. Miners scour waterways and separate tiny flecks of gold from sediment with toxic mercury, devastating broad swaths of river and forest habitat. The mercury poisons fish, wildlife, and drinking water for people downstream. Because it occurs in remote locations and is unregulated or illegal, enforcement agencies and conservation groups have little information about where new activity is occurring. Project Inambari, uses a combination of radar and multispectral satellite imagery to identify and track unregulated and illegal mining deep in the Amazon rainforest.
In 2021, we received additional support to conduct field testing in the Amazon with our colleagues at the Peruvian parks department, the Servicio Nacional de Áreas Naturales Protegidas por el Estado (SERNANP). Our work has focused on the area around Puerto Maldonado in the Peruvian department of Madre de Dios, a remote area near the Bolivian and Brazilian borders which is plagued by illegal mining. We made two trips to visit remote parts of the Amarakaeri Communal Reserve park affected by mining to collect data using drones, to ground truth and supplement our satellite data.
2019 — 2021 mining activity detected near Reserva comunal Amarakaeri in Peru. From the Artisanal Mining Challenge, displayed as a layer in SkyTruth Alerts.
Our trip to the Reserve revealed the scope of mining in the region and the challenge faced by park staff: informal mining occurs right at the park boundaries and park staff must mediate between local communities and miners operating in the buffer area around the park.
We also made two mapping trips in the Tambopata Reserve and confirmed illegal mining activity in the buffer area surrounding Tambopata. We conducted drone flights from the reserve’s reforestation post near La Pampa and will use imagery from these flights to evaluate the use of satellite imagery for monitoring the recovery of the reforested mining areas, as well as investigating several new incursions of suspected illegal activity.
We held a workshop in Puerto Maldonado with participation by personnel from SERNANP and other groups interested in monitoring mining in the region, including the Executor of the Administrative Contract of the Amarakaeri Communal Reserve, Conservación Amazónica, and the Center for Amazonian Scientific Innovation. At the workshop, we shared how these groups can use SkyTruth Alerts to help them monitor mining activity, and walked them step-by-step through the process of registering on the SkyTruth Alerts platform and doing some basic analysis. We also reviewed other types of deforestation information currently available, as well as various sources of satellite imagery. Participants were interested to see the capabilities of the platform, in particular how they could use it to determine when a deforestation alert was due to mining or other activities of interest.
Work on the Alerts platform has progressed significantly with the addition of new satellite image-based alerts (such as RADD and GLAD-S2) as well as high-repeat, high-resolution satellite imagery from commercial provider Planet Labs. We’ve also continued the development of our own mining-specific detection model that combines satellite radar imagery with multispectral imagery to generate monthly maps showing the location of mining regardless of cloud cover. We’re making the data publicly available to activists, government agencies, and others through our Alerts platform, and working to expand our analysis across the entire Amazon.
Training the Next Generation of Conservation Technology Specialists
Over the course of 2021, nine SkyTruth interns from around the world contributed to important ongoing SkyTruth projects, as well as to projects they completed independently. From developing a user-friendly dashboard for SkyTruth’s Colorado River basin work, to assessing potential worst case scenarios of mine tailings dam failures in Arizona, to automated road detection using machine learning, SkyTruth interns are a vital part of our day-to-day operations. Our interns included students at universities such as Harvard, the University of California at Berkeley, Georgia Tech and elsewhere. Many of our interns go on to work for some of the most impactful forces in the environmental movement, including Esri, Rocky Mountain Institute, Global Fishing Watch, and Encore Renewable Energy. In light of this, we view SkyTruth’s Internship program as a stepping stone towards long-term success in the environmental movement.
Geospatial Analyst Brendan Jarrell also manages the internship program.
