Carina Project: Robust Mineral Resource Upgrade Supported BY Extensive Drilling
PA
CARINA PROJECT: ROBUST MINERAL RESOURCE UPGRADE
SUPPORTED BY EXTENSIVE DRILLING
TORONTO, ON, October 1, 2025 – Aclara Resources Inc. (“Aclara” or the “Company”) (TSX: ARA) is pleased to
announce an updated mineral resource estimate (“MRE”) for the Carina Project, the Company ’s flagship ion
adsorption clay project located in Goiás, Brazil (“Carina”, “Carina Project” or the “Project”) . The MRE represents
a key component of the Carina pre -feasibility study ( “PFS”) and reflects considerable geological and technical
advancements compared to the previously reported inferred mineral resource statement (the “2024 Resource
Statement”). The MRE has been prepared in accordance with the Canadian Institute of Mining, Metallurgy and
Petroleum ( “CIM”) Definition Standards (2019) and National Instrument 43 -101 - Standards of Disclosure for
Mineral Projects (“NI 43-101”).
Highlights
• Updated MRE from “ inferred” to “ indicated” mineral resource c ategory: By applying the principles of
Reasonable Prospects for Economic Extraction (RPEE) and constraining the estimate within an optimized pit
shell, the mineral resources have been refined and upgraded, resulting in 236 million tonnes (Mt) of indicated
mineral resources and 48 Mt of inferred mineral resources, compared with the 297 Mt of inferred mineral
resources as reported in the 2024 Resource Statement.
• Consistent grades of magnetic rare earths: Indicated mineral resources report stable and consistent magnetic
element grades, including dysprosium oxide ( Dy₂O₃) at 42.7 ppm, terbium oxide ( Tb₄O₇) at 6.8 ppm, and
NdPr oxide (Nd₂O₃ & Pr₆O₁₁) at 292.6 ppm, closely aligned with the 2024 Resource Statement (Dy ₂O₃: 42.1
ppm, Tb₄O₇: 6.9 ppm, NdPr: 296.5 ppm).
• Significantly improved geological confidence: The MRE is supported by 24,564 meters (m) of drilling across
1,682 drillholes, representing a pproximately a 500% increase in drilling compared to the 2024 Resource
Statement, which was based on 4,104 m of drilling across 363 drillholes.
• Strong geo-metallurgical database supports robust process: The MRE is supported by 14,001 samples
analysed for total rare earth oxides (TREO), desorbable rare earth oxides (DREO), and impurit ies, providing
a granular understanding of the Project’s geo-metallurgical behaviour compared to 2024 Resource Statement,
which was based on 3,789 samples.
• High conversion rate of mineral resources demonstrates consistent mineralization: Approximately 79% of
inferred mineral resources have been successfully upgraded to the indicated mineral resource category.
• Declaration of reserves expected with PFS: The indicated mineral resources serve as the foundation for the
PFS mine plan and the subsequent estimation of mineral reserves.
PA
Aclara COO, Hugh Broadhurst, commented:
“Following a major drilling campaign, we have successfully converted the majority of our inferred mineral resources
into the indicated mineral resource category. This is an important step as we prepare for their assessment as
mineral reserves upon completion of the PFS in the coming weeks.
The quality of the Carina deposit has been affirmed. Grades of both heavy and light rare earths remain strong,
tonnage is consistent, and results continue to demonstrate the stability and continuity of mineralization. At the
same time, our knowledge of Car ina has advanced considerably, with metallurg ical confidence strengthened by
a comprehensive new database of more than 10,000 data points. This strong foundation will be essential to
optimizing the mine plan, reducing costs, and ensuring efficient manageme nt of operations into the future.
We are now fully focused on delivering the PFS within 45 days of the date of this news release and the feasibility
study in the second quarter of next year —key milestones in unlocking the full economic potential of this unique
heavy rare earths resource.”
Mineral Resource Statement
Table 1 summarizes the MRE, including TREO, NdPr, Dy, and Tb contents by geological domain. Further details
of the MRE are provided in Tables 7 and 8.
Table 1. Carina Project MRE (as of July 29, 2025) compared to the 2024 Resource Statement (as of May 3, 2024)
Updated MRE (PFS 2025)
*Other: Stacked material (AT)
2024 Resource Statement (PEA 2024)
Notes:
1. TREO means total rare earth oxides (La2O3, CeO2, Pr6O11, Nd2O3, Sm2O3, Eu2O3, Gd2O3, Tb4O7, Dy2O3, Ho2O3, Er2O3, Tm2O3,
Yb2O3, Lu2O3, and Y2O3).
2. NdPr means neodymium and praseodymium (Nd2O3 and Pr6O11).
3. Dy means dysprosium (Dy2O3) and Tb means terbium (Tb4O7).
4. Updated mineral resources were estimated above an NSR cut-off of 10.0 US$/t, using average long term metal prices and
metallurgical recoveries outlined below under “Mineral Resource Determination and Selection”.
5. The 2024 Resource Statement was estimated above an NSR cut-off of 7.4 US$/t, using average long term metal prices and
metallurgical recoveries reported on the 2024 Resource Statement on August 9, 2024.
6. PEA means Preliminary Economic Assessment.
7. PFS means Pre-Feasibility Study.
8. Totals may not be balanced due to rounding of figures.
Mass
Mt Total REO NdPr Dy Tb Total REO NdPr Dy Tb
0.3 1,392 243 36 5.8 473 83 12 2
26.8 994 138 20 3.0 26,660 3,695 532 80
63.6 1,576 292 36 5.8 100,176 18,555 2,292 370
60.6 1,847 369 52 8.5 111,858 22,338 3,147 513
54.3 1,595 299 50 7.9 86,581 16,247 2,711 429
30.7 1,488 268 46 7.1 45,744 8,232 1,404 220
236.3 1,572 293 43 6.8 371,492 69,150 10,099 1,614
48 1,288 236 41 6.4 61,675 11,316 1,949 307
Oxide Content (Tonnes)
Other
Upper Pedolith
Lower Pedolith
Upper Saprolite
Saprock
Lower Saprolite
Geological Domain Oxide Total Grade (ppm)
TOTAL INDICATED
TOTAL INFERRED
Mass
Mt Total REO NdPr Dy Tb Total REO NdPr Dy Tb
298 1,452 284 39 6.4 432,003 84,565 11,573 1,897
Category
Oxide Total Grade (ppm) Oxide Content (Tonnes)
TOTAL INFERRED
*
PA
Geological Model and Estimation Domains
Aclara developed the geological model for the Carina Project based on 1,682 drillholes totalling 24,564 m of auger,
reverse circulation (RC), and sonic drilling, using Leapfrog Geo (v2024.1.3). The model integrates lithological
alteration and regolith logging, ensuring consistency with the block model resolution and supporting robust mineral
resource estimation.
The vertical weathering profile, interpreted from geological logging and supported by geochemical transitions,
comprises seven distinct horizons: Stacked Material (AT), Upper Pedolit h (UP), Lower Pedolit h (LP), Upper
Saprolite (US), Lower Saprolite (LS), Saprock (SR), and BR. The regolith model (Figure 1) was constructed as a
sequential stack of erosive surfaces, honouring the natural weathering progression. The LS–SR interface was key
to defining the base of REE-enriched horizons.
Figure 1. Three-dimensional regolith model; UP: Upper Pedolit; LP: Lower Pedolith; US: Upper Saprolite; LS:
Lower Saprolite (Source: Aclara, 2025).
PA
Three alteration domains (Figure 2) were modeled using visual logging and elemental proxies (Na ₂O, Fe ₂O₃,
CaO):
• High Hydrothermal Alteration (HH), typically associated with the highest REE enrichment;
• Albitization (ALB), characterized by elevated Na₂O and partial preservation of feldspar textures; and
• No Hydrothermal Alteration (NH), representing unaltered granite. The NH domain, together with BR, was
excluded from the resource estimation process.
Figure 2. Plan view of interpreted alteration model (Source: Aclara, 2025).
The MRE is constrained to domains that combine favourable regolith horizons with either HH or ALB alteration.
The intersection of both models resulted in 12 estimation units : HH_AT, HH_UP, HH_LP, HH_US, HH_LS,
HH_SR, ALB_AT, ALB_UP, ALB_LP, ALB_US, ALB_LS, and ALB_SR. NH and BR zones were excluded due to
lack of enrichment or insufficient geostatistical support.
For statistical analysis and reporting purposes, REEs were grouped into light REEs (LREEs), defined as the sum
of La, Pr, Nd, Sm, Eu, and Ce; and heavy REEs (HREEs), defined as the sum of Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu,
and Y. The total REE (TREE) content is the sum of LREEs and HREEs. All chemical assay data were originally
provided in elemental form (pure metal content, in ppm), not as oxides. This distinction is important, as oxide
grades are typically higher due to the added molecular weight of the oxygen component.
PA
Assay data were composited to 2 m intervals, with approximately 75% of samples meeting target length. Shorter
intervals, mostly near unit boundaries, did not materially affect the overall estimate. Each estimation unit was
assigned a numerical code for consistent tracking in statistical analysis and final reporting.
Estimation Methodology
Mineral resource estimation was conducted using ordinary kriging (OK) to interpolate 18 variables (15 individual
REEs, Al2O3, Th, and U) across the 12 estimation units . Each unit was estimated independently using strict
geological boundaries and a 2 × 2 × 2 node block model. A two-pass estimation approach was applied consistently,
with a short-range search in the first pass and an extended search in the second, limiting the number of composites
per drillhole to ensure data quality.
Different variogram models were applied depending on the alteration type of each unit, and a high -grade spatial
constraint was implemented. Units with lower drill density had adjusted composite selection criteria to maintain
accuracy.
A detailed kriging neighbourhood analysis demonstrated that using between 5 to 16 composites in the first pass
and 4 to 20 in the second pass balances local variability preservation and model robustness effectively. Using
fewer than 12 neighbours increased estimation bias and variability, while more than 20 neighbours offered minimal
improvements.
Statistical validation metrics confirmed high kriging efficiency (~0.90) and low bias, supporting the chosen
estimation parameters as optimal for accurate, reliable resource modelling with manageable computational
demands.
Mineral Resource Classification
Mineral resources were classified following international reporting standards, including NI 43 -101 and CIM
definitions, ensuring transparency, materiality and competence. The classification is based on geological
continuity, grade consistency and data qual ity, using a geometric approach known as the “three drillhole rule,”
which evaluates the average distance from each block to its three nearest drillholes.
The following classification criteria were applied:
• Inferred Mineral Resources: Blocks supported by drill spacing up to 250 m, indicating reasonable but
lower confidence geological continuity.
• Indicated Mineral Resources: Blocks within a tighter 125 m × 125 m drill spacing, providing sufficient
confidence for geological and grade assumptions; many blocks were upgraded from Inferred to Indicated.
• Measured Mineral Resources: Although the nominal 60 m grid meets geometric requirements, current
drilling only defines small, isolated sectors in the southwestern portion of the deposit, lacking sufficient
lateral continuity. In line with CIM guidelines, no measured mineral resources are reported in this update.
Blocks with theoretically adequate spacing were conservatively classified as indicated mineral resources.
Updated drilling and improved geological understanding enabled the reclassification of 236 Mt as indicated, while
48 Mt remain inferred. Drill spacing thresholds align with variogram analyses showing spatial continuity of 200 –
500 m depending on alteration zones, supporting the adopted classification standards.
Areas lacking sufficient drill support remain unclassified, with detailed spatial distribution reflecting drill density
and confidence levels.
PA
In Figure 3, the estimated block model grades for the DY variable are displayed along section NS = 8495000,
coloured by grade ranges. The Z-axis has been vertically exaggerated by a factor of five to enhance visualization
of the stratigraphy and grade distribution. Composites are shown as points with matching colour ranges, allowing
a direct visual comparison. This section illustrates the spatial relationship between them. A strong visual
correlation is observed between composite values and their corresponding block estimates.
Figure 3: Estimated block model grade distribution for dysprosium along section NS = 8495000; Z -axis has been
vertically exaggerated by a factor of five to enhance visualization of the stratigraphy and grade distribution.
(Source: ABelco, 2025)
Figure 4: Plan view of resource classification map; yellow: indicated mineral resource blocks; red: inferred
mineral resource blocks; light brown: blocks that remain unclassified due to insufficient drill support; black dots:
drillhole composites used in the estimation. (Source: Aclara, 2025)
PA
Mineral Resource Determination & Selection
The MRE for the Carina Project has been prepared in accordance with the CRIRSCO International Reporting
Template, applying the principle of Reasonable Prospects for Economic Extraction (RPEE). The reported
resources are constrained within a Lerchs–Grossmann optimized pit shell ensuring that only mineralized material
with a credible potential for economic recovery is classified as mineral resources.
Grades were initially reported in elemental form (ppm of La, Ce, Nd, etc.) but were converted to their equivalent
oxides (REO) to align with industry standards and market compar isons. This was achieved by applying
stoichiometric oxide factors for each element, a standard practice supported by published technical literature (see
Table 4). The conversion was applied directly to the block model, ensuring consistency in grade reporting.
Table 4. Conversion factors from elemental rare earth elements to rare earth oxides.
Element Conversion Factor Rare Earth Oxide
La 1.1728 La2O3
Pr 1.2082 Pr6O11
Nd 1.1664 Nd2O3
Sm 1.1596 Sm2O3
Eu 1.1579 Eu2O3
Ce 1.2284 CeO2
Gd 1.1526 Gd2O3
Tb 1.1762 Tb4O7
Dy 1.1477 Dy2O3
Ho 1.1455 Ho2O3
Er 1.1435 Er2O3
Tm 1.1421 Tm2O3
Yb 1.1387 Yb2O3
Lu 1.1371 Lu2O3
Y 1.2699 Y2O3
Pit Optimization
Conducted with Geovia Whittle ™, the optimization used cost assumptions and metallurgical recoveries
representative of the deposit.
Economic inputs were validated by ABelco Consulting SpA and shown in Table 5.
PA
Table 5. Key economic inputs to the Lerchs-Grossmann pit optimization model.
Parameter Value
Mining cost US$3.1/t mined
Processing & G&A US$10.0/t processed
Separation cost US$15.0/kg concentrate
Selling cost US$0.03/kg concentrate
Royalty 2%
Plant efficiency 93%
Pit slope angle 25°
Recoveries were calculated as the median of the ratio desorbable to total grade for each REE within each regolith
unit and shown in Table 6 . Metallurgical recovery results were obtained from analyses performed by AGS
Laboratory in La Serena, Chile, and SGS Geosol Laboratory in Vespasiano Minas Gerais, Brazil from a total of
14,001 drilling samples.
Table 6. Metallurgical recoveries per regolith.
Regolith Unit Recovery (%)
Upper Pedolith 28.2
Lower Pedolith 31.4
Upper Saprolite 28.8
Lower Saprolite 23
Saprock 17.2
Only blocks with NSR ≥ US$10.0/t and within the optimized pit are reported. Average recoveries per domain are
considered appropriate for PFS–level evaluation.
The selling prices applied in this MRE are based on long -term assumptions, ensuring that all resources with
potential future economic viability are captured. For pit optimization, price estimates (reported in US$/kg on an
oxide basis) were applied only to the magnetic rare earth elements —dysprosium (Dy), terbium (Tb), and
neodymium-praseodymium (NdPr)—as follows:
• Dy = 1,302
• Tb = 3,826
• NdPr = 151
All other REEs were assigned a zero value. For the eventual conversion of Indicated mineral resources into
mineral reserves, the Company intends to apply more conservative price assumptions.
Importantly, since the 2024 Resource Statement in which ~15% of the mineral resources were located outside of
Aclara’s mineral concessions, the current update confirms that 100% of the reported mineral resources are now
within titled properties fully owned by the Company, eliminating this previous constraint and notably reducing risks
associated with the Project.