In 2019, indigenous communities in the Peruvian Amazon started receiving alerts on their phones about illegal gold mining operations appearing in their territory. The notifications came from satellites orbiting 400 miles above Earth, processed through an open-access platform called Global Forest Watch, and delivered within days of the trees being cut. Government rangers arrived within 48 hours. This wasn't a special program for a well-funded reserve. It was free technology available to anyone with internet access.
From Space Agencies to Smartphones
For most of satellite monitoring's history, only governments and large institutions could afford to collect and analyze Earth observation data. NASA's Landsat 1, launched in 1972, began the systematic documentation of Earth's surface, but accessing and interpreting that information required specialized expertise and expensive software. The data existed, but it sat largely unused outside academic and government circles.
That bottleneck has broken open. The cost of manufacturing and launching satellites has dropped dramatically, flooding the market with imagery. More importantly, the infrastructure for making sense of all that data has shifted from exclusive research labs to platforms anyone can use. Global Forest Watch now offers 65 datasets for forest monitoring at no cost. The GLOBE Program, NASA's longest-running citizen science initiative, has trained hundreds of thousands of people to contribute observations that validate what satellites see from space.
The shift isn't just about access to pretty pictures from orbit. Satellite data can be broken into microtasks—small, discrete pieces of work that don't require years of training. Can you identify whether an image shows a cloud or clear sky? Distinguish between forest and cleared land? These simple classifications, multiplied across thousands of volunteers, create datasets that would take professional researchers decades to compile.
The Three-Legged Stool
Crowdsourcing serves three distinct functions in climate monitoring, each addressing a different limitation of satellite technology. The first is analysis—processing the sheer volume of imagery that satellites now produce. Platforms like Zooniverse host projects where volunteers classify images, tag features, and identify changes. Between July 2021 and December 2022, NASA's CLOUD GAZE project (the acronym stands for "Community science project Leveraging Online and User Data through GLOBE And Zooniverse Engagement") had citizens classify cloud photographs to improve how satellites measure atmospheric conditions.
The second function is monitoring for specific changes. Algorithms can detect broad patterns, but human observers catch anomalies and context that automated systems miss. The GLAD deforestation alert system sends weekly notifications about tree cover loss in tropical regions, but it takes people on the ground to distinguish between illegal logging, permitted forestry, and natural disturbances.
The third function has become the most valuable: collecting complementary ground data. Satellites can't see through clouds, measure water quality directly, or capture what happens under forest canopies. Drones equipped with multispectral sensors and thermal imaging fill some gaps, providing resolution impossible from orbit. But the most powerful complement is human observation—temperature readings, species counts, water samples, and local knowledge about what's normal and what's changed.
When Machines Need Human Teachers
The relationship between artificial intelligence and crowdsourced data creates a feedback loop that strengthens both. Machine learning algorithms need training data—lots of it—to learn what healthy vegetation looks like versus stressed crops, or how to distinguish between a natural wetland and a flooded field. Volunteers provide that training by labeling thousands of images, teaching algorithms to recognize patterns.
Once trained, AI handles the repetitive work, flagging potential issues for human review. Natural language processing scans social media posts and citizen reports for climate-related information, surfacing local observations that satellites might miss entirely. A farmer's tweet about unusual frost, a hiker's photo of a dried-up stream, a resident's complaint about air quality—these fragments of information, aggregated and analyzed, reveal patterns that complement the view from space.
This division of labor matters because neither humans nor machines work well alone at this scale. Algorithms process data faster but lack context. Humans provide context but can't review millions of images. Together, they create monitoring systems that are both comprehensive and nuanced.
The Copenhagen Question
Not all crowdsourcing deserves the label "democratization." The Copenhagen Framework on Citizen Data draws a line: for data to truly qualify as citizen-generated, people must be "sufficiently engaged in design and/or collection stages," not just used as cheap labor to click through images someone else decided were important.
This distinction reveals a tension in how crowdsourced satellite monitoring has evolved. Early models treated volunteers as distributed processors—human computers clicking through tasks designed entirely by scientists. Newer approaches involve communities in deciding what to monitor and why. When indigenous groups in the Amazon use GLAD alerts, they're not just processing data; they're determining which threats matter most to their land and how to respond.
The business ecosystem reflects this evolution. Initially, satellite data platforms led crowdsourcing efforts, recruiting volunteers to analyze their imagery. Now third-party organizations coordinate projects, often working directly with affected communities to identify monitoring priorities. The shift moves crowdsourcing closer to genuine citizen science, where participants help shape research questions and interpret findings, not just collect data points.
What Free Data Actually Costs
Making satellite data open and free sounds unambiguously positive, but it creates new dependencies. Communities that rely on Global Forest Watch or similar platforms have little control over whether those services continue, how algorithms change, or what gets monitored. When a platform updates its detection methods, years of baseline data might become incompatible with new measurements.
The technical barriers have lowered, but they haven't disappeared. Interpreting satellite imagery still requires understanding what sensors can and can't detect, how atmospheric conditions affect readings, and which changes matter. A community might receive alerts about forest loss without the context to know whether it's a crisis or normal seasonal variation. Training programs help, but they require ongoing investment that doesn't always materialize.
Still, the alternative—returning to a world where only well-funded institutions could monitor environmental change—serves no one. The Amazon Conservation Association's ability to detect illegal mining within days, using free tools and volunteer networks, has prevented destruction that would have taken months to discover under the old model. Imperfect democratization beats perfect exclusivity.
Monitoring From Below
The most profound change isn't technological—it's about who gets to ask questions. For decades, climate monitoring meant scientists choosing what to measure based on research priorities, then occasionally sharing findings with affected communities. Crowdsourced satellite data inverts that relationship, at least partially. Communities identify threats, use satellite tools to document them, and present evidence to authorities or the public.
This bottom-up approach makes science more responsive to local needs, but it also shifts responsibility. When citizens can monitor their own environment, governments and institutions face pressure to act on what that monitoring reveals. The data becomes evidence, sometimes uncomfortably so. Deforestation alerts don't just inform—they implicate, showing exactly when and where violations occur.
That accountability cuts both ways. Communities using these tools must grapple with what they find and decide how to respond. Satellite data might reveal that a neighbor is clearing protected land, that a local industry is polluting a watershed, or that climate impacts are accelerating faster than anyone acknowledged. The democratization of monitoring also democratizes the burden of knowing—and choosing whether to act on that knowledge.