Any reservoir on Earth
40,000+ dams pre-catalogued from GRanD, plus user-drawn polygons for unlisted lakes, wetlands, and seasonal water bodies. No "supported reservoirs" list.
A HelioLink Technologies product · v1.5.5
Watermark every reservoir on Earth.
Pick any dam, lake, or wetland. Arariel pulls every cloud-free satellite image of it since 2000, measures the waterline, and tells you the wettest and driest moments of each year — the watermarks. Then it forecasts the next one.
- 25 years of monthly waterline history
- 40,000+ reservoirs catalogued worldwide
- Zero insider access required
01
What Arariel does, in plain English
Reservoirs are the world's water bank accounts. They get deposits (snowmelt, rain, river inflow) and withdrawals (drinking water, irrigation, hydropower). If you don't know the balance, you can't plan.
Until now, knowing the balance meant either trusting whatever the dam operator publishes — which is often delayed, redacted, or just not public for foreign reservoirs — or hiring a remote-sensing analyst at $50,000+ to read the satellite imagery by hand.
Arariel reads the satellite imagery automatically and gives you a 25-year history in minutes. No insider access. No manual labor. No waiting.
Who needs this
- Water utilities
- Plan for shortages 6–18 months ahead
- Hydropower operators
- Forecast generation capacity by season
- Defense & intelligence
- Track strategic water infrastructure abroad
- Insurance & reinsurance
- Price drought risk on agricultural & municipal portfolios
- Researchers
- Defensible 25-year time series with documented methodology
- Journalists & advocacy
- Fact-check official reservoir reports independently
02
How it works — four steps
-
1
Pick
Click any reservoir on a world map. Arariel ships with the Global Reservoir and Dam Database (GRanD) — about 40,000 reservoirs worldwide — plus user-defined polygons for lakes and wetlands that aren't in GRanD.
-
2
Pull
The app fetches every cloud-free Landsat 5/7/8/9 and Sentinel-2 scene of the reservoir since 2000. That's typically 200–600 usable scenes per reservoir.
-
3
Measure
For each scene, Arariel computes MNDWI — the Modified Normalized Difference Water Index, a published water-detection method — and traces the waterline. Out comes a surface area in km² for every observation.
-
4
Watermark
The annual peak and trough are the HIGH and LOW watermarks. The forecast is an ordinary-least-squares regression on the trend, validated by leave-one-out cross-validation (LOOCV). Every number is reproducible from the source scenes.
03
Live demo — Hoover Dam & Lake Mead
Same methodology Arariel applied to the Indus Basin's Tarbela and Mangla reservoirs in our flagship engagement (study HYD-26-014) — re-run here on the most-watched reservoir in the United States, with 25 years of U.S. Bureau of Reclamation elevation history as ground truth.
25 years of waterline — measured monthly
End-of-month elevation, Lake Mead, 2000–2025. The gold dots mark each year's LOW watermark; the blue dots mark each year's HIGH. Hover for exact values.
Monthly elevation (ft)
HIGH watermark
LOW watermark
Tier-1 shortage threshold (1,075 ft)
Annual HIGH / LOW watermarks
The full dataset is 26 years × 2 watermarks = 52 rows. Showing the most recent 12 below. Download all 52 (CSV) · XLSX
| Year | HIGH date | HIGH elev (ft) | HIGH area (km²) | LOW date | LOW elev (ft) | LOW area (km²) | Drop (ft) |
|---|
Forecast — where will Lake Mead be in 2026?
OLS regression on 25 years of HIGH watermark data, with a 95% prediction interval derived from leave-one-out cross-validation (LOOCV). The same forecast you'd get from Arariel on any reservoir, computed live in your browser from the dataset above.
2026 HIGH forecast
— ft
± — ft (95% CI)
2027 HIGH forecast
— ft
± — ft (95% CI)
LOOCV skill
r² = —
RMSE = — ft
Methodology
- Extract the annual HIGH watermark (max elevation) for each year 2000–2025. These are the 26 training points.
-
Fit an ordinary-least-squares regression:
elev = β₀ + β₁ · year. - For LOOCV: re-fit the OLS 26 times, each time holding out one year, predict the held-out year, and accumulate the squared residuals. The square root of the mean is the LOOCV RMSE.
- The 95% prediction interval at any future year is ±1.96 × RMSE.
- Skill is reported as r² (coefficient of determination on the LOOCV predictions).
This is the same forecast the Arariel desktop app generates when you click "Forecast" on any reservoir. The math is transparent on purpose — you can re-derive it from the CSV.
04
Get the Hoover Dam deliverables
Every Arariel engagement produces this same artifact set. The Hoover Dam bundle below is free — same format, same depth as our paid client deliveries.
Full PDF study
Methodology, findings, forecast, references —
same template as the paid client studies.
Study (DOCX)
The same study as an editable Microsoft Word file.
Drop into your own templates.
Watermark dataset (XLSX)
26 years × HIGH/LOW = 52 rows with scene IDs,
sensors, cloud cover, and MNDWI thresholds.
Watermark dataset (CSV)
Same data, plain text. Drop into pandas, R,
Excel, or your tool of choice.
Annotated satellite imagery
1024×1024 PNG with crosshair, coordinates,
and Esri attribution baked in.
References
Every source cited in the demo study —
USBR, Pekel, Udall & Overpeck, Williams.
05
References — Lake Mead demo
- U.S. Bureau of Reclamation, 2024. Boulder Canyon Operations Office — Lake Mead Elevation Records. Lower Colorado River Operations. usbr.gov/lc/region/g4000/hourly/mead-elv.html
- Bureau of Reclamation, 2023. Near-term Colorado River Operations: Draft Supplemental Environmental Impact Statement. Modifications to the December 2007 Record of Decision.
- Castle, S. L., Thomas, B. F., Reager, J. T., Rodell, M., Swenson, S. C., Famiglietti, J. S., 2014. "Groundwater depletion during drought threatens future water security of the Colorado River Basin." Geophysical Research Letters 41 (16), 5904–5911.
- Pekel, J.-F., Cottam, A., Gorelick, N., Belward, A. S., 2016. "High-resolution mapping of global surface water and its long-term changes." Nature 540, 418–422.
- Udall, B., Overpeck, J., 2017. "The twenty-first century Colorado River hot drought and implications for the future." Water Resources Research 53 (3), 2404–2418.
- Williams, A. P., Cook, B. I., Smerdon, J. E., 2022. "Rapid intensification of the emerging southwestern North American megadrought in 2020–2021." Nature Climate Change 12, 232–234.
- Xu, H., 2006. "Modification of normalised difference water index (NDWI) to enhance open water features in remotely sensed imagery." International Journal of Remote Sensing 27 (14), 3025–3033.
- Lehner, B., Liermann, C. R., Revenga, C., et al., 2011. "Global Reservoir and Dam Database (GRanD), version 1.3." Frontiers in Ecology and the Environment 9 (9), 494–502.
- USGS, 2024. Landsat Collection 2 Level-2 Science Products. Earth Resources Observation and Science (EROS) Center. usgs.gov/landsat-missions
- Colorado River Basin Drought Contingency Plans, 2019. Public Law 116-14. Upper and Lower Basin DCPs.
06
What Arariel can do for you
Any timeframe
Landsat record goes back to 1972. Modern Collection-2 calibration covers 2000–present at 30 m resolution. Sentinel-2 fills in at 10 m from 2015 onward.
Any deliverable
CSV dataset, branded PDF study, editable DOCX, annotated satellite imagery, time-series charts, forecast with uncertainty bands. Pick what you need on the SOW.
Reproducible
Every result traces back to specific Landsat scene IDs and an explicit MNDWI threshold. Methodology is published, not proprietary. Defensible in court, in peer review, in IG audits.
Open-source data
NASA / USGS / NOAA / ESA / ECMWF / OpenStreetMap. No subscription fees, no licensing gotchas, no foreign-data-access restrictions.
SDVOSB
HelioLink Technologies is a Service-Disabled Veteran-Owned Small Business, eligible for set-aside and sole-source contracting under FAR 19.14.
07
Run this on your reservoir
Single-reservoir study
from $18,500
- 25-year HIGH / LOW watermark dataset (CSV + XLSX)
- Annotated satellite imagery (PNG)
- 10-page PDF study with forecast & references
- Editable DOCX for your own templates
- Perpetual internal-use license
- 50% deposit, 50% on delivery, NET 0 on balance
Volume discount on multi-reservoir engagements · quarterly refresh subscriptions available · GovCloud hosting on request