No one can put a clean percentage on how much of the Amazon remains unexplored, because “explored” means different things depending on who is asking. Satellites have imaged virtually every square kilometer of the basin, but ground-level scientific surveys of species, soils, and archaeological sites have covered a surprisingly small fraction. Researchers estimate that the catalog of flowering plants in the Brazilian Amazon alone is roughly 20% incomplete, with thousands of species awaiting formal description. And entire categories of life, from soil microbes to canopy insects, have barely been inventoried at all. The Amazon is simultaneously one of the most watched and least understood landscapes on Earth.
What “Unexplored” Means When Satellites Can See Everything
From orbit, the Amazon has been mapped in remarkable detail. Radar instruments can measure tree height, optical sensors track deforestation in near-real time, and lidar pulses can penetrate the canopy to reveal the ground surface beneath. But satellite coverage does not equal scientific knowledge of what lives there, what lies beneath the soil, or what civilizations once occupied the land. The distinction matters because it shapes public perception: people assume that because the forest appears on Google Earth, scientists must know what is in it. They largely do not.
Even satellite data has gaps. Wet-season cloud cover in the Amazon can reach 90 to 99%, and conventional satellite processing misses a large number of usable observations as a result. Advanced atmospheric correction methods can recover 20 to 80% more cloud-free pixels depending on the season, but the sheer density of cloud cover means that continuous optical monitoring remains a challenge.1Remote Sensing of Environment. Remote sensing of tropical ecosystems: Atmospheric correction and cloud masking matter The forest resists easy observation from above, and the story only gets harder once you zoom in to ground level.
Why Scientists Have Only Sampled a Sliver
The single biggest factor shaping what we know about Amazonian biodiversity is access. Field biologists need to physically reach a site, set traps, collect specimens, and transport them to a research institution. In practice, that means the vast majority of biological surveys in the Amazon have been conducted within easy reach of rivers, roads, and cities. A study of amphibian sampling across the basin found that rivers were the primary driver of where researchers collected specimens, while roads had a surprisingly limited influence, reflecting the region’s historical reliance on river transport.2Biotropica. Biases in Amphibian Sampling in the Amazon: Using Infrastructure and Accessibility Data to Identify Sampling Gaps The result is that huge swaths of interfluvial forest, the upland areas between major rivers, have received little or no ground-level biodiversity assessment.
The pattern is not limited to amphibians. A large-scale assessment of ant diversity across the Brazilian Amazon recorded over 1,000 nominal species but found that sampling locations were strongly biased by access routes, urban centers, research institutions, and major infrastructure projects.3Ecography. A large‐scale assessment of ant diversity across the Brazilian Amazon Basin: integrating geographic, ecological and morphological drivers of sampling bias Hydroelectric dam construction from the mid-1970s onward, and especially after a Brazilian government infrastructure acceleration plan starting in 2008, generated a burst of environmental impact assessments that skewed the sampling record further. Those studies produced a high volume of records near dam sites, overestimating the influence of rivers and hydroelectric plants while underestimating highways as a source of sampling bias.2Biotropica. Biases in Amphibian Sampling in the Amazon: Using Infrastructure and Accessibility Data to Identify Sampling Gaps
This means that our picture of Amazonian species richness is warped by where it was convenient to do fieldwork. The areas we know best are those closest to human infrastructure, which also tend to be the most disturbed. The areas we know least are the most remote and intact, precisely the places most likely to harbor species found nowhere else.
Thousands of Plant Species Still Awaiting Description
The Amazon contains the richest assemblage of plant species of any region on the planet, but pinning down the actual number has proven difficult. A taxonomically verified checklist of seed plants from lowland Amazon rain forests compiled over 14,000 confirmed species, including roughly 6,700 tree species.4PubMed Central. Amazon plant diversity revealed by a taxonomically verified species list Separately, researchers estimating tree diversity in the Brazilian Amazon alone arrived at a figure of about 11,200 tree species reaching meaningful trunk diameter.5PubMed Central. How many tree species are there in the Amazon and how many of them will go extinct? These numbers are close to what some ecological models predict, but statistical modeling of how many flowering plant species remain undescribed suggests the Amazon’s catalog is far from complete.
A recent analysis estimated that the Brazilian angiosperm catalog is at least 19 to 23% incomplete, with roughly 7,300 to 9,600 species still awaiting formal scientific description across all of Brazil’s ecological regions. The Amazon and the dry scrublands of the Caatinga showed the greatest potential for new species descriptions, with statistical models predicting roughly 3,000 to 3,200 species in the Amazon alone that have not yet been named.6PLOS ONE. Protecting hidden treasures: Indigenous lands safeguard 50% of areas with the highest potential for angiosperm discoveries in Brazil—patterns and conservation priorities Those are not obscure mosses tucked into inaccessible cliff faces. Many are likely trees, shrubs, and herbs that no botanist has yet collected from a remote patch of forest, pressed into a herbarium sheet, and formally described.
Animals in the Blind Spots
New vertebrate species continue to turn up in the Amazon with striking regularity. A fish-collecting expedition in the lower Vaupés River basin in Colombia, focusing specifically on rapids and fast-water habitats, registered 95 species across 30 families and identified the first records of four fish species in Colombia along with six species thought to be entirely new to science.7PubMed Central. Fishes (Actinopterygii) of the rapids and associated environments in the lower Vaupés River Basin: an undiscovered Colombian Amazon diversity That a single targeted expedition to one river system could yield half a dozen potentially undescribed fish gives a sense of how much remains to be found.
Invertebrates tell an even more dramatic story. A study of insect diversity in an Amazonian forest measured species richness at different heights, from the ground to the upper canopy, and found that about 62% of the fly species collected were never sampled at ground level.8PubMed Central. Vertical stratification of insect abundance and species richness in an Amazonian tropical forest In other words, standard ground-based insect traps miss the majority of species in a single group of common insects. Since most biodiversity surveys do not include canopy sampling, the implication is that a huge fraction of Amazonian insect life has never been formally documented.
The Microbial Frontier
If plants and animals are undercounted, the situation for microorganisms is orders of magnitude worse. Amazonian soils harbor some of the most diverse microbial communities on Earth, but we know remarkably little about what is in them. A study of soil metagenomes from forest and pasture sites in the eastern Brazilian Amazon recovered dozens of assembled microbial genomes spanning two archaeal and eleven bacterial phyla, and taxonomic classification suggested that most of them represented potential novel microbial taxa.9PubMed Central. Genome-resolved metagenomics reveals novel archaeal and bacterial genomes from Amazonian forest and pasture soils These are not subtle genetic variants of known organisms. They are lineages different enough from anything previously characterized that they may belong to entirely new groups.
The biotechnological implications are just as striking. Researchers who sequenced soil bacteria from an Amazon conservation unit and mined their genomes for biosynthetic gene clusters, the stretches of DNA that encode the production of biologically active chemicals, found that about 63% of those clusters were related to unknown metabolites. The strains also contained pathways for the production of antibiotic and antitumor agents.10PubMed Central. A metabologenomics approach reveals the unexplored biosynthetic potential of bacteria isolated from an Amazon Conservation Unit The Amazon’s soil contains a chemical library that we have barely begun to catalog, let alone understand or use.
Ancient Cities Hidden by the Canopy
Perhaps the most surprising dimension of the Amazon’s unexplored character has nothing to do with biology. It is archaeological. For centuries, the prevailing view was that the Amazon’s poor soils could never have supported large, complex societies. That assumption has been systematically demolished over the past two decades. Lidar surveys in the Bolivian Amazon revealed a previously undescribed form of tropical low-density urbanism belonging to the Casarabe culture, dating from roughly 500 to 1400 CE. Two large sites covering 147 and 315 hectares contained stepped platforms, conical pyramids up to 22 meters tall, concentric defensive earthworks, raised causeways stretching several kilometers, and massive canal and reservoir systems.11Nature. Lidar reveals pre-Hispanic low-density urbanism in the Bolivian Amazon
In the upper Amazon of Ecuador, fieldwork combined with lidar analysis uncovered an anthropized landscape with clusters of monumental platforms, plazas, streets, agricultural drainage systems, and wide straight roads running over great distances, dating back roughly 2,000 years.12PubMed. Two thousand years of garden urbanism in the Upper Amazon Along the southern rim of the Amazon basin, researchers predicted that earthwork sites in an area comprising just 7% of the Amazon could have sustained a population in the hundreds of thousands, which alone discredits early estimates that the entire basin held only 1.5 to 2 million people before European contact.13Nature Communications. Pre-Columbian earth-builders settled along the entire southern rim of the Amazon
The soils themselves carry evidence of large-scale human management. Amazonian dark earths, patches of unusually fertile, carbon-rich soil scattered across the basin, have been shown to result from intentional practices by ancient Indigenous populations who enriched the soil to improve crop productivity.14PubMed Central. Intentional creation of carbon-rich dark earth soils in the Amazon These dark earths form a widespread pattern, found as ring-shaped mounds surrounding flat terraces across large areas of prehistoric settlements in multiple regions.15Journal of Archaeological Science. Dark earths and the human built landscape in Amazonia: a widespread pattern of anthrosol formation The archaeology of the Amazon is still in its early stages. Most of the basin has never been surveyed with lidar, and given the density of discoveries wherever researchers do look, the implication is that many more sites remain hidden beneath the forest canopy.
People Living Beyond Contact
The Amazon is also home to the world’s largest concentration of Indigenous peoples who have chosen to remain isolated from outside society. These communities live in some of the most remote parts of the basin, and their very existence highlights how inaccessible certain areas remain. Researchers using remote sensing documented 28 isolated villages within four Brazilian states, estimated to host over 1,700 individuals, and found that these villages consistently occur at mid-elevations, near the tops of watersheds, and far from existing roads and trails.16PLOS ONE. Geographic Distribution of Isolated Indigenous Societies in Amazonia and the Efficacy of Indigenous Territories
Some of these communities appear to be growing. The world’s largest known isolated tribe, identified by its watershed name Rio Humaitá, has expanded into at least four separate longhouse and garden clusters in recent decades, spreading in three directions and undergoing a visible fissioning event around 2016.17Scientific Reports. Remote sensing evidence for population growth of isolated indigenous societies in Amazonia Machine learning models trained on satellite data have shown a strong ability to distinguish between contacted and uncontacted villages based on features like smaller cleared areas, greater distance from artificial lights and populated places, higher elevation, and lower regional population density.18PeerJ Computer Science. Machine learning with remote sensing data to locate uncontacted indigenous villages in Amazonia These models suggest that additional uncontacted groups likely exist in areas not yet surveyed.
The existence of these communities underscores a key point about Amazonian exploration. The forest is not empty of people just because it is empty of roads. Indigenous communities, both contacted and uncontacted, possess deep knowledge of areas that outside scientists have never visited. What counts as “unexplored” depends entirely on whose perspective you take.
New Tools That Reach Where People Cannot
Several technologies are beginning to fill in the knowledge gaps without requiring researchers to physically visit every stream and hillside. Environmental DNA, or eDNA, is one of the most promising. By collecting water samples and sequencing the genetic fragments shed by organisms into the environment, researchers can detect species that would be difficult or impossible to find with traditional methods. An eDNA study of vertebrate diversity in the Colombian Amazon and Orinoco basins identified 709 distinct taxonomic units across all sampling locations and found new records at every site, with the highest number of previously undetected fish species appearing in rivers that had already been surveyed with conventional nets and traps.19Frontiers in Ecology and Evolution. eDNA metabarcoding: an effective tool for vertebrate diversity studies in the Colombian Amazon and Orinoco basins Parallel work in small streams and ponds near Iquitos, Peru, has confirmed that eDNA metabarcoding can rapidly inventory local fish communities in habitats too small or remote for conventional sampling efforts.20PubMed Central. Assessing fish diversity in small streams and ponds of the Peruvian Amazon using environmental DNA metabarcoding
Airborne lidar, the same technology that revealed hidden urban sites, is equally transformative for ecology. It can map forest structure, identify individual large trees, and detect landforms invisible to optical sensors. Genome sequencing of soil and water samples, automated acoustic monitoring of animal soundscapes, and satellite-based machine learning models are all expanding the reach of Amazonian science far beyond what a field team with boots and nets can cover in a season. The bottleneck is no longer purely technological. It is funding, logistics, and the sheer scale of the basin.
Roads and the Shrinking Window
The tragedy of unexplored Amazonia is that some of it may be destroyed before it is ever studied. Road construction is one of the most powerful predictors of deforestation. In Peru’s Amazon region, 83% of deforestation between 1999 and 2005 occurred within 20 kilometers of a road, and nearly a quarter of that loss was concentrated near a single stretch of the Interoceanic Highway connecting southwestern Amazonia to Atlantic and Pacific ports.21Cambridge University Press. The threat of road expansion in the Peruvian Amazon Roads bring settlers, cattle ranchers, and illegal mining operations into previously inaccessible areas, and the ecological damage radiates outward from the pavement.
The areas most at risk from road expansion overlap heavily with the areas least surveyed by scientists, creating a perverse dynamic. Indigenous territories provide a partial buffer: one analysis found that existing Indigenous territories encompassed all of the identified isolated villages and half of the areas with the highest predicted probability of isolated village occurrence.16PLOS ONE. Geographic Distribution of Isolated Indigenous Societies in Amazonia and the Efficacy of Indigenous Territories The same study estimating undescribed plant species found that Indigenous lands safeguard roughly half of the areas with the highest potential for new angiosperm discoveries in Brazil.6PLOS ONE. Protecting hidden treasures: Indigenous lands safeguard 50% of areas with the highest potential for angiosperm discoveries in Brazil—patterns and conservation priorities Protected territories, in other words, are not just shielding known biodiversity. They are protecting species, soils, archaeological sites, and microbial lineages that science has not yet had the chance to document.
Quartzite Caves and Other Physical Frontiers
Beyond the forest floor, canopy, and rivers, the Amazon basin contains geological features that remain largely uninvestigated. Southern Venezuela and parts of the Guiana Shield host quartzite and quartz sandstone outcrops with cave systems that develop through extremely slow silica dissolution, potentially over timescales orders of magnitude longer than limestone cave formation.22ScienceDirect (Academic Press). Chapter 102 – Quartzite and quartz sandstone caves of South America These caves can display complex branching and network patterns similar to those found in carbonate systems, and many are tucked into remote tabletop mountains that are difficult to reach by foot or helicopter. The fauna living in and around these formations is essentially unknown. Isolated mountaintop ecosystems in the Amazon, sometimes called “sky islands,” have their own endemic species, but most have never been biologically surveyed.
The deep channels of Amazon tributaries present another frontier. River bottoms dozens of meters deep are home to specialized fish species adapted to strong currents and low light, and collecting in those habitats requires equipment and techniques that most survey teams do not carry. Every time researchers invest in sampling an unusual habitat, whether rapids, caves, deep channels, or upper canopy, they find species not recorded anywhere else. The pattern is consistent enough to be a rule: wherever scientists look for the first time in the Amazon, they find something new.