James Bay Lithium Claims: A Sentinel-2 Outcrop Survey

Most writing about James Bay lithium is about individual projects. This page is about the ground itself. We counted the claims in the Quebec government’s public titles file, dated the lithium showings in its geoscience database, and measured how much rock Sentinel-2 can see through the boreal forest, peat and lakes of the Eastmain-La Grande corridor. The audience is the geologist, land manager or investor who has to decide where the next dollar of grassroots work goes, and wants numbers rather than adjectives.

If you want the mineral system itself, our explainer on lithium deposits: pegmatite, brine and clay covers how LCT pegmatites form, and the lithium in Canada overview covers the national picture. For licences in other countries, the mining cadastre portal guide by country is the hub. Here we stay with James Bay lithium claims and the data anyone can download.

Canada’s two producing lithium mines turned out an estimated 5,983 tonnes of lithium in 2024, about 2.5% of world mine output, according to Natural Resources Canada.

Why we built a James Bay lithium dataset from open sources

Ask ten people how big the James Bay lithium rush is and you will hear ten different numbers, usually taken from press releases. Quebec already publishes the raw material to answer the question properly. The Ministère des Ressources naturelles et des Forêts (MRNF) posts its GESTIM digital data as Shapefile, MapInfo and GeoPackage downloads. We used the active-titles file dated 13 October 2026. Its user agreement allows derived works as long as the source is credited, and it warns that the file is raw data behind the official titles map, drawn at a nominal 1:50,000 scale, not the official map itself.

The second source is SIGÉOM, Quebec’s geoscience database. Its showings, deposits and mines layer is published on Données Québec under a CC BY 4.0 licence. Each record lists the main and secondary substances and a discovery year, which lets us count lithium showings by decade without naming a single property.

The third source, and the one that turns this into a James Bay lithium survey rather than a claims census, is the Copernicus Sentinel-2 archive, read through Google Earth Engine’s harmonized surface-reflectance collection. Sentinel-2 records 13 spectral bands at 10, 20 and 60 m, according to the European Space Agency’s mission description. We worked at 20 m, the resolution of the short-wave infrared bands that matter most for rock.

We ran everything with scripts that anyone with an Earth Engine account and the same files can rerun. The claim and showing counts cover the whole territory. The satellite part covers a named sub-area, because processing every pixel of a territory larger than many countries would have been a much bigger job than this question needs.

How many James Bay lithium claims are there?

Strictly, the GESTIM file does not say which claims are for lithium. A map-designated claim gives the holder an exclusive right to explore for mineral substances in general, so the honest count is of all claims in the territory. On 13 October 2026, the file held 252,838 active map-designated claims in Quebec. Of these, 124,695, or 49.3%, sit in the territory that GESTIM records as the Gouvernement régional d’Eeyou Istchee Baie-James. Together they cover 6,546,362 hectares, about 16.2 million acres. The median claim is 53.1 hectares, roughly 131 acres, because cells shrink toward the north on the provincial grid.

That is not all lithium ground. The southern part of the territory overlaps gold country in the Abitibi greenstone belt, and some of the densest claim blocks in our half-degree grid sit there. But the growth of the last five years tracks the James Bay lithium story closely, and the chart below shows when today’s active claims were first issued.

Chart

Active James Bay claims by year of issue

Claims still active on 13 Oct 2026, Eeyou Istchee Baie-James, by issue year (selected years; 2026 to 13 October)

Read this chart with one caveat. It shows claims that are still active, sorted by the year they were issued. A claim staked in 2019 and dropped in 2022 does not appear. So the bars understate the older years and say nothing about turnover. Even so, 2023 stands out: 14,741 claims issued that year are still held, more than any other year. 2025 and 2026 (to 13 October) each added more than 10,000 that are still active.

The renewal count tells a related story. A first term lasts three years, and renewals run in two-year terms, according to the MRNF page on how to obtain an exclusive exploration right. In the territory, 39,635 active claims (31.8%) have never been renewed, 31,803 have been renewed once, and 28,032 have been renewed five times or more. That is a young James Bay lithium land package sitting on top of an older one.

From claim to exclusive exploration right

Quebec now calls the claim an exclusive exploration right, or EER. The MRNF says the EER is “formerly known as a claim” and is the only exploration title available in the province. You obtain one by map designation of grid cells through GESTIM. Most people in the industry, and most searchers, still say “claim”, so we use both words here.

The fees depend on which side of the 52nd parallel the ground sits. On the MRNF’s rates page, as of 15 July 2026, a right of more than 50 hectares north of the 52nd parallel costs C$244.40 to register when a holder takes 1 to 150 rights in a day, and C$1,222.00 each above 150. South of the parallel, a right of 25 to 100 hectares costs C$104.98 for the first 40 in a day. Renewal fees follow the same area bands. Rates are indexed every 1 January, and 2026 indexation was 2.05%. General information, not legal advice; confirm with the MRNF before you designate.

The GESTIM file also records how much work each claim needs before its next renewal. For active claims north of the 52nd parallel in the territory, the median requirement rises with each term: C$135 in the first term, C$450 in the second, C$900 in the third, C$1,350 in the fourth, C$1,800 in the fifth and sixth, and C$2,500 from the seventh term on. Summed across every active claim in the territory, the work required for the current term comes to about C$153.9 million.

At the defaults, 50 northern claims held for three terms, or seven years, need C$36,660 in fees and C$74,250 of required work: about C$110,900 in total, or roughly C$43 a hectare. Most of that is work, which is the point of the system: James Bay lithium claims are cheap to stake and costly to keep idle. The cheapest way to meet it is to spend it on the ground that most deserves it, and that is where the satellite part of this study comes in.

▶ Video (2 min 25 s): how satellite and soil data are being combined in Canadian critical-mineral exploration. · 2:25

James Bay lithium showings: a 2020s story

SIGÉOM records each showing with its substances and a discovery year. We searched every main and secondary substance field for lithium (Li) and kept records inside the bounding box of the territory’s active claims, 49.0 to 55.0° N and 79.5 to 70.0° W. That box is a little wider than the territory, so a few showings near its edges may sit just outside. We counted each showing once.

The box holds 153 lithium showings. One was recorded in the 1950s, eleven in the 1960s, two in the 1970s, none in the 1980s, two in the 1990s, four in the 2000s and eighteen in the 2010s. Then 115, three in every four, were recorded in the 2020s. The chart below starts at the 1960s.

Chart

Lithium showings by decade of discovery

SIGÉOM showings listing Li, James Bay territory bounding box, by recorded discovery year

Two cautions apply to these James Bay lithium showing counts. A discovery year in SIGÉOM is the year a showing was recorded, which follows reporting, not geology. And a showing is a documented occurrence of a substance, often a single outcrop or sample, not a resource. What the curve does show is how fast the documented James Bay lithium inventory has grown, and how recently most of it was written down. Much of the ground around those new showings was mapped at regional scale long before anyone looked for spodumene.

▶ Video (3 min 11 s): satellite and geochemical screening in another Canadian district, British Columbia, for comparison with James Bay. · 3:11

The test corridor: Eastmain to La Grande

For the satellite part we chose a single rectangle, 52.0 to 53.5° N and 78.0 to 75.5° W, between the Eastmain and La Grande rivers. We picked it from the data, not from any project: it contains three of the five half-degree cells with the most lithium showings in the territory. It holds about 2.39 million hectares of land (5.9 million acres) after lakes and rivers are removed, 18,847 active claims that touch it, and 52 of the lithium showings counted above. On our rasterised count, 33.2% of its land sits under an active claim. The whole box lies north of the 52nd parallel, so the northern fee bands apply.

This is Cree territory. The James Bay and Northern Quebec Agreement divides it into Category I, II and III lands, and the Cree Nation Government’s Cree Nation Mining Policy (June 2024) states that no mineral claims can be made on Category I land without prior approval by the affected Cree First Nation. The same policy says the territory holds over 300 traplines, each managed by a designated tallyman. We come back to what that means for a field program below.

How the satellite classes were made

We took every Sentinel-2 surface-reflectance scene over the corridor from July to September in 2019 to 2022 with less than 40% cloud, 580 scenes in all. We masked cloud, cloud shadow, cirrus and snow with the scene classification band, then took the median of each band at every 20 m pixel. Using summers before 2023 was deliberate. Quebec’s 2023 fire season burned 42,883 km², according to Environment and Climate Change Canada, and fresh burn scars look like bare ground to a vegetation index.

Each pixel then went into one of four classes. Water: the scene classification marked it as water, or the modified water index was above zero. Bare: the vegetation index (NDVI) was below 0.25. Bright low-iron outcrop: bare, with mean visible reflectance above 0.12, a red-to-blue ratio below 1.6 and short-wave infrared (band 11) reflectance above 0.20. Everything else counted as vegetated land. The thresholds were set by hand, not trained on known pegmatites. The idea follows published work: Cardoso-Fernandes, Teodoro and Lima used Landsat, Sentinel-2 and ASTER to map lithium-bearing pegmatites in scenes with low vegetation cover. White, iron-poor felsic rock is what a spodumene pegmatite often looks like at surface. So does a quartz vein, a granite dome, a sand esker or a gravel pit.

Infographic

From open data to a James Bay field program

1. 1ScreenGESTIM map, SIGÉOM showings and satellite outcrop classes.
2. 2DesignateMap designation in GESTIM; three-year first term.
3. 3Plan and informAnnual work plan to the municipality and Cree community 30 days ahead.
4. 4Work and reportOutcrop mapping and sampling; report the work to the MRNF.
5. 5Renew or dropRenew in two-year terms with fees and required work.

The flow above is how this kind of screen fits a real James Bay lithium program in Quebec. The steps after the screen come from the MRNF’s own pages, and the order matters: the screen tells you where to designate and where to walk, while the rules decide when you may walk there.

How much rock can Sentinel-2 see on James Bay lithium claims?

Very little. Across the corridor, 0.78% of land pixels were bare: about 18,600 hectares out of 2.39 million. Claimed land was barely different from unclaimed land, at 0.81% against 0.76%. Across 328 grid cells of 0.1 degree, the rank correlation between how much of a cell is claimed and how much bare rock it shows was 0.08, close to none. When we split the cells into quarters by bare-rock share, the share under claim stayed between 31.9% and 34.2% in every quarter.

Chart

How much bare rock a 0.1-degree cell shows

Grid cells in the Eastmain-La Grande corridor by bare-rock share of land (328 cells)

Most cells show almost nothing. 187 of the 328 cells, 57%, have less than 0.5% bare land; only 21 reach 2% or more. That is the reality of exploring a lowland of black spruce, lichen, peat and lakes. It also explains why James Bay lithium claims cannot be valued from imagery alone. Where the rock is hidden, the satellite has nothing to measure, and the decision has to rest on geology, geophysics, till and lake-sediment geochemistry and boots on the ground.

The bright-outcrop proxy near known lithium showings

The second number is more encouraging. Our bright, low-iron proxy covered 0.048% of corridor land. On unclaimed land it was 0.041%; on claimed land it was 0.062%, half again as much. Within 500 m of the 52 known lithium showings, a combined 3,811 hectares, it rose to 0.126%: 2.6 times the corridor average and 3.1 times the unclaimed rate.

Chart

Bright low-iron outcrop pixels, by zone

Share of land pixels flagged as the felsic-outcrop proxy (%), Eastmain-La Grande corridor

For James Bay lithium explorers, note what did not change. Bare ground of any kind was no more common near the showings (0.65%) than elsewhere. What changed was the type of bare surface: brighter, less iron-stained, more reflective in the short-wave infrared. That is consistent with white pegmatite outcrops, or with the stripped and trenched outcrop that follows a discovery. Our test cannot separate the two, and with 52 showings it is a small sample. Treat it as a reason to look, not as a result.

The limits: forest, peat, fire, snow and pixel size

Every James Bay lithium number on this page depends on what the ground surface allows. In plain words, here are the limits we know about.

• Vegetation: a thin cover of lichen or moss hides rock completely from an optical sensor. The NDVI threshold sends those pixels to the vegetated class.
• Glacial cover: much of the lowland is blanketed by till, clay and peat left by the last ice sheet. Rock under a metre of till is invisible to Sentinel-2 at any resolution.
• Fire: burned stands look bare for years. We used pre-2023 summers to avoid the 2023 scars, but older burns can still add false bare ground.
• Snow and cloud: only July to September scenes were used, with clouds and snow masked. Thin haze that the mask misses can still brighten pixels.
• Pixel size: 20 m is coarse next to a dyke a few metres across. Narrow pegmatites will be missed or diluted.
• Grid geometry: claims were rasterised at 20 m, and the GESTIM file is drawn at a nominal 1:50,000, so edge pixels are approximate.

None of this makes satellite screening useless on James Bay lithium ground. It makes it a ranking tool for the small fraction of land that is exposed, and a planning tool for access, camps and traverses on the rest. On the exposed fraction, higher-resolution and hyperspectral imagery can say far more than our four simple classes.

▶ Video (2 min 38 s): soil-based methods for covered Canadian Shield ground, where outcrop is scarce, as in most of James Bay. · 2:38

What the data means for a James Bay lithium field program

Put the James Bay lithium numbers to work. Think of a junior holding 50 claims, about 2,600 hectares, in the corridor. On our averages it might show about 20 hectares of bare ground and about one hectare of bright, low-iron outcrop, around 30 pixels at 20 m. Those few hectares are where a prospecting crew should land first. The rest of the block needs other tools before anyone spends drill money on it.

Here is an order of work that follows from the numbers and the rules:

1. Desk screen in winter. Pull GESTIM to see open and held ground, SIGÉOM for showings and old reports, and a summer satellite composite for outcrop and access. Rank the exposed patches.
2. Designate with intent. If you are adding ground, map designation through GESTIM gives a three-year first term. North of 52°, fees jump above 150 rights in one day, so plan the batch.
3. Plan and inform. The MRNF’s annual work planning rule requires the plan to reach the local municipality and the Indigenous community concerned at least 30 days before any work, airborne surveys included, with an information session on request.
4. Walk the bright patches first. July to September is the clear-sky window our composite used. Map, sample and photograph the flagged outcrops before you move to covered ground.
5. Report and renew. Outcrop work goes into the MRNF’s simplified exploration work report; reports over C$5,000 must be signed by the qualified professional who supervised the work.

Where Farmonaut fits, and where it does not

We are a satellite data company. Our satellite-based mineral detection work analyses multispectral and hyperspectral imagery for alteration, lithology and structure, and lithium is on our list of detectable materials. A client sends an area as coordinates, KML/KMZ or a polygon, names the target mineral, and gets ranked high-potential zones with GIS files, usually in 5 to 20 business days depending on size and complexity. We have scanned more than 100,000 hectares across more than 20 minerals and 25 countries. Remote screening of this kind can cut early-exploration timelines from months to days and reduce early-stage cost by up to 80 to 85%.

What we do not do matters as much on James Bay lithium ground. Our results are exploration targets that sit ahead of any resource estimate. They are not NI 43-101 disclosure, and a qualified person still has to sample, drill and report; our NI 43-101 explainer covers that line. We are not a licensing agent or a land broker, and a land screen is not an environmental or social clearance. On covered ground, where this study shows most of the territory sits, a satellite report helps you decide where to send the geophysics and the crew, not whether a deposit exists. For the geophysical side, see geophysical survey methods for mining, and for public geology data across Canada, our guide to geological survey agencies.

There are over 300 traplines covering the territory, and the harvesting activities on each are managed by a designated responsible tallyman.” (Cree Nation Mining Policy, 2024)

Method notes and how to reproduce this James Bay lithium survey

All inputs, scripts and result tables are kept with this article’s working files. The steps are short enough to rerun on a newer GESTIM file.

• Claims: active titles shapefile from the GESTIM download directory, file of 13 October 2026. Kept map-designated claims (title code CDC) whose municipality field is Eeyou Istchee Baie-James. Holder fields were deleted on load. Claims were counted by issue year, renewal count and half-degree cell, and rasterised at 20 m for the corridor.
• Showings: SIGÉOM metallic substances table from Données Québec, downloaded 15 October 2026. Kept records listing Li in any substance field, inside 49.0 to 55.0° N and 79.5 to 70.0° W, one record per showing number.
• Imagery: Sentinel-2 L2A harmonized, July to September 2019 to 2022, scene cloud below 40%, scene-classification masks for cloud, shadow, cirrus and snow, per-band median, 20 m in UTM zone 18N, 120 tiles.
• Statistics: class shares of land pixels for the whole corridor, claimed and unclaimed land, and 500 m buffers around showings; and per 0.1° cell, sampled every 25th pixel, with cells under 200 sampled land pixels dropped.

Because the GESTIM file updates often, James Bay lithium claim counts will drift between versions. Anyone rerunning the scripts on a newer file should expect different totals and the same broad picture: a territory with half the province’s claims, a young land package, and very little rock at surface.

Frequently asked questions about James Bay lithium

General information, not legal, tax or investment advice. Claim rules and fees change; confirm with the MRNF and, for work in Eeyou Istchee, the Cree Nation Government. Contains modified Copernicus Sentinel data. GESTIM data used under the MRNF digital data user agreement; it is not the official titles map. Farmonaut results are exploration targets ahead of any resource estimate.