AquaScan Pro — Transparency Charter
AquaScan Pro is a new class of tool, so we do not ask for blind trust — we publish exactly how every report is built, what it can and cannot claim, and how its predictions are being validated against real drilled boreholes. This page is the standing reference for customers, hydrogeologists, drillers and financiers who want to check our work.
What AquaScan Pro is — and is not
It IS
- A desktop pre-feasibility screening — the evidence-gathering stage a competent hydrogeologist performs before any site visit, automated and priced honestly.
- A fusion of real registries, satellite physics and published regional drilling statistics into one audited report.
- A filter that catches siting-stage errors (wrong geology, no offset evidence, dry-well clusters) before money is spent on fieldwork or drilling.
It is NOT
- NOT a statutory hydrogeological survey, and never sold as one — Kenya law requires a licensed survey and WRA authorisation before drilling.
- NOT authority to mobilise a drilling rig. No AquaScan score, however high, releases the mobilisation gate without field ERT, a survey-grade peg, professional sign-off and WRA approval.
- NOT a measurement of your site — depths, yields and water quality are calibrated estimates until field validation, and estimated figures are labelled as such (for example “modelled”, “regional est.” or “not measured”).
The three-score architecture
Most siting reports blur “how good is this target” with “are you allowed to drill”. AquaScan separates them so each question gets an honest answer:
DDTR — Desktop Pre-Feasibility Readiness (0–95). How complete and convergent the desktop evidence stack is. Hard-capped at 95: desktop evidence alone can never reach 100 — the final points belong to fieldwork.
FRR — Field & Regulatory Readiness (0–100). Ten gates only a site visit can close: reconnaissance, survey-grade peg, field ERT/VES, hydrogeologist sign-off, WRA authorisation, pump test, laboratory analysis.
MAG — Mobilisation Authorisation Gate. BLOCKED → CONDITIONALLY RELEASED (field confirmation only) → RELEASED FOR DRILLING. A high desktop score never auto-releases this gate.
Every evidence family, and where it comes from
| Evidence family | Sources | Role in the report |
|---|
| Water-point registries | WPDx (Water Point Data Exchange), OpenStreetMap, UNESCO IHP-WINS and KenyaRapid+ field records, owner-supplied WRA/county records | Real named water points (springs, wells, boreholes) within 50 km — the strongest desktop evidence in weathered basement terrain. No water point is ever invented. |
| Regional drilling statistics | Compiled county drilling records (all 47 Kenya counties) + national statistics (RWSN, UNICEF WASH, BGS) | Drilled-depth bands, tested-yield ranges and success rates that anchor and bound every depth/yield estimate. |
| Climate & recharge | NASA POWER, ERA5-Land, GLDAS, multi-year precipitation series | Water-balance (banded Budyko) recharge model with quickflow separation — sustainable abstraction is capped by recharge. |
| Terrain & structure | SRTM 30 m DEM (slope, aspect, TWI, drainage density), DEM lineament analysis | Topographic wetness and structural screening. Desktop lineaments are weighted LOW by design — published inter-analyst reproducibility is poor. |
| Satellite remote sensing | MODIS NDVI (MOD13A2), Landsat multispectral, Sentinel-1 (proxy), GRACE-FO, SMAP/MERRA-2, thermal IR | 10-method fusion (Saaty AHP weights). Proxy-grade methods are LABELLED proxy and barred from making site-specific claims. |
| Soil & geology | ISRIC SoilGrids v2.0, rock classification, Kenya hydrogeological province priors (5 provinces, literature-banded) | Infiltration, aquifer setting and the province prior that keeps every estimate inside published literature ranges. |
| Water-quality screening | Regional hydrogeochemistry + WHO 2011 guideline comparison | MODELLED screening only — every report states that an ISO 17025 laboratory analysis is the only proof. |
Every report carries a full data-provenance annex naming the source of each governing value. Where a source did not answer, the report says so. A few inputs (rainfall, elevation, satellite storage and vegetation) then fall back to a regional or latitude-based estimate; each one is listed in the report under “data sources that used regional estimates”, and none of them is presented as a measurement.
The integrity architecture — why you can trust the numbers
- One governing value. Yield, depth and probability are reconciled once — through aquifer physics bounded by the region's published tested-yield band — and every page of the report is stamped from that single result. Sub-model outputs appear only in a labelled diagnostic annex.
- A consistency validator that can block its own product. Before export, every report passes a set of automated audits (range containment, cross-engine reconciliation, water-point arithmetic, grade gating). If a check fails, the report refuses to export — customers have seen this happen, because it is real.
- No invented data — enforced in code. No synthetic wells, no fabricated satellite scenes, no “ground is stable” claims from proxy data, no drilled-outcome credit for estimated records. Estimated values are labelled as estimates, and anything not measured says so.
- Automated verification tests run against the physics and scoring engines on every change — Theis drawdown limits (Driscoll ⅔ rule), Budyko recharge banding, consensus-tracks-governing, county-record lookup, score discrimination (a data-poor site scores <40; no participation trophies).
- An integrity fingerprint on every PDF. Each report prints a fingerprint computed from its governing values and generation time. If a figure in a circulated copy has been altered, the fingerprint no longer matches — ask us to verify any AquaScan report you are shown.
Validation ledger — our track record, published honestly
AquaScan Pro is a new tool. Rather than pretend otherwise, we publish the state of its validation and update this ledger as drilled outcomes come in:
Measured benchmark (northern Kenya): against real drilled boreholes in validated counties, the modelled depth fell within the drilled band ~40% of the time — and the reports say exactly that, including that yield could not be reliably predicted without a pump test. This benchmark is stated on the cover of every report.
Prediction-vs-outcome register: collecting. Every borehole drilled off an AquaScan report — with the owner's consent — will have its predicted vs actual depth, yield and outcome published here, successes and failures alike. No cherry-picking: the register is append-only.
Founding pilot programme: the first drilling clients who share their WRA completion records (drill log, pump test, lab certificate) receive priority engineering support from EmersonEIMS — and calibrate the tool for everyone who comes after them.
Join the pilot → Known limitations — stated plainly
- No satellite can see individual fractures. In basement terrain the difference between a disappointing and an excellent borehole is often a single fracture zone that only field ERT/VES resolves — this is physics, not a software limit.
- Registry water points prove groundwater occurrence, but most registries do not publish true drilled depths. Depth/yield calibration is registry- and literature-based until WRA completion records are ingested.
- Modelled water quality is a screening, not a test. Only an ISO 17025 laboratory analysis proves potability.
- A strong desktop score is not a wet borehole. A 64% success probability means roughly one site in three still disappoints — the report states its uncertainty ranges for exactly this reason.
Scientific references
- MacDonald A.M., Bonsor H.C., Dochartaigh B.É.Ó., Taylor R.G. (2012). Quantitative maps of groundwater resources in Africa. Environmental Research Letters 7(2).
- MacDonald A.M., Davies J., Calow R.C., Chilton P.J. (2005). Developing Groundwater: A Guide for Rural Water Supply. ITDG Publishing.
- Driscoll F.G. (1986). Groundwater and Wells, 2nd ed. Johnson Filtration Systems. (Usable drawdown ≤ ⅔ of saturated thickness in unconfined aquifers.)
- Theis C.V. (1935). The relation between the lowering of the piezometric surface and the rate and duration of discharge of a well using ground-water storage. Trans. AGU 16.
- Cooper H.H., Jacob C.E. (1946). A generalized graphical method for evaluating formation constants. Trans. AGU 27.
- Budyko M.I. (1974). Climate and Life. Academic Press. (Water-balance recharge banding, Fu-parameterised.)
- WHO (2011, 4th ed. + addenda). Guidelines for Drinking-water Quality.
- Saaty T.L. (1980). The Analytic Hierarchy Process. (Multi-sensor weighted-overlay fusion weights.)
- RWSN / UNICEF (2016+). Guidance on professional borehole siting and supervision.
- Kenya Water Act 2016 and the Water Resources Regulations in force — statutory hydrogeological survey and WRA authorisation requirements (confirm current provisions with WRA).
Check our work
Run a site, read the audit trail, challenge a number. If you are a licensed hydrogeologist, we invite you to review this methodology — credited reviewers are acknowledged here.
EmersonEIMS · AquaScan Pro Transparency Charter · This page is updated whenever the methodology changes.