Meana Petro
SR-11IntermediateWave 2AI

Forecasting with Uncertainty

In one sentence

Produces P10, P50 and P90 production forecasts whose spread is calculated, not chosen to look reasonable.

The problem

Field forecasts are usually a single deterministic line with a percentage band added by convention. When actuals fall outside it, nobody can say whether the model was wrong or the outcome was simply within the real uncertainty.

How it works

Probabilistic forecasting combines decline-curve, material-balance and simulation-based methods, weighting each by its demonstrated skill on that field’s own history in a rolling backtest. The backtest is published with the forecast so the reader sees the method’s track record.

Data required

  • Production history by well and by field with downtime separated
  • Well events: workovers, shut-ins, artificial-lift changes
  • Reservoir pressure history
  • Planned future activity and facility constraints

What you get

  • P10/P50/P90 forecasts at well, reservoir and field level
  • Rolling backtest showing historical forecast accuracy
  • Variance attribution when actuals deviate from plan

What it cannot do

Uncertainty bands are conditional on the future activity assumption. A forecast is not a prediction of what a company will do, and most large historical forecast misses in this industry were activity misses, not reservoir misses. We report the two sources separately so the miss can be attributed honestly.

Every capability page states a real limitation. If a vendor cannot tell you where their product stops, you are not being told the truth about where it starts.

Related capabilities

  • PR-01Foundational

    Production Forecasting

    Forecasts what each well will produce next month, and tells you which wells it is not confident about.

    Production & Artificial LiftAI
  • SR-14Foundational

    Automated Reserves Estimation & DCA

    Fits decline curves across every well in the field on a consistent basis and keeps the audit trail that a reserves report needs.

    Subsurface & Reservoir IntelligenceDomainAI
  • SR-01Frontier

    Smart Proxy Reservoir Simulation

    A learned stand-in for the full reservoir simulator that answers in seconds instead of hours, so an engineer can test fifty scenarios instead of three.

    Subsurface & Reservoir IntelligenceAIDomain
  • SR-02Frontier

    Automated History Matching

    Tunes the reservoir model until it reproduces what the field actually produced, automatically and with the uncertainty kept rather than discarded.

    Subsurface & Reservoir IntelligenceAIDomain
  • SR-12Intermediate

    Automated Well Test Interpretation

    Reads pressure build-up data, proposes the reservoir model behind it, and flags the tests that are not interpretable.

    Subsurface & Reservoir IntelligenceAIDomain