Decline curve analysis: methods, type curves, and software

What decline curve analysis does

Decline curve analysis forecasts how a well's production rate will fall over time by fitting a curve to its production history and extending it into the future. It is the most widely used method for producing-well forecasts because it needs only rate and time data, which every operator has. The forecast then supports reserves estimates, budgets, acquisition valuations, and development decisions.

The method is empirical. It assumes that the factors that controlled past production will continue to control it, so it works best on wells with a stable operating history and becomes less reliable after changes such as workovers, new artificial lift, offset-well interference, or extended shut-ins. Reserves definitions and classification are governed by frameworks such as the SPE Petroleum Resources Management System, which sets out how production forecasts relate to reported reserves and resources.

Decline analysis sits alongside other forecasting methods rather than replacing them. Rate-transient analysis uses flowing pressure as well as rate to infer reservoir properties, numerical simulation models fluid flow explicitly, and material balance relates production to reservoir pressure. Decline curves are fast and transparent, which is why they remain the default for large producing portfolios, but significant decisions on a single asset often justify a cross-check with one of these methods.

The Arps family of curves

Most decline work still starts from the relationships published by J. J. Arps in 1945. Exponential decline assumes the rate falls by a constant percentage per period. Harmonic decline assumes the decline rate itself falls in proportion to the production rate. Hyperbolic decline sits between them and is controlled by an exponent usually called the b-factor: a b-factor of zero gives exponential behavior, and a value of one gives harmonic behavior.

Unconventional wells complicate the picture. Fitted b-factors for shale wells often exceed one in early life, which, if extended indefinitely, produces unrealistically large ultimate recoveries. Practitioners therefore commonly switch from a hyperbolic segment to an exponential tail at a chosen terminal decline rate. The choice of b-factor, the switch point, and the terminal rate can change a well's estimated ultimate recovery materially, so they should be documented and reviewed rather than left as software defaults.

Gas and oil wells, and different products from the same well, can decline differently, so forecasts are often built per stream and checked for consistent ratios over time. Abrupt changes in the gas-oil ratio or water cut are a signal that the curve may no longer describe the same reservoir behavior and should be re-examined rather than extended.

Type curves for undeveloped locations

A producing well has its own history; an undrilled location does not. Type curves fill that gap by combining the normalized production of comparable wells into a representative profile. Building one involves choosing analog wells by geology, completion design, lateral length, and vintage, normalizing for differences such as lateral length or proppant volume, aligning wells to a common start, and fitting a curve to the combined data. Percentile curves are often used to express uncertainty.

Type curves are where data quality and selection bias matter most. Wells with incomplete histories, inconsistent normalization, or analogs drawn from a different part of a basin can produce a curve that looks precise but is not representative. Operators such as Devon Energy, EOG Resources, and EQT develop large multi-well programs where type-curve assumptions feed directly into capital allocation, which is why the analog set and normalization rules should be preserved alongside the curve.

Software for forecasting and reserves

Specialist software automates curve fitting, type-curve construction, and the handoff to economics. ComboCurve's cloud platform covers forecasting, reserves, economics, and scheduling; an EQT proxy filing with the SEC describes EQT's license for ComboCurve decline-curve analysis, reserves work, and economic scenario modeling. PHDwin and Halliburton's ARIES are established reserves and economics systems, and Enverus PRISM connects asset screening, forecasting, and economics on shared data. Novi Labs offers machine-learning production forecasts built on well-level data, and a Journal of Petroleum Technology case study describes an operator using Novi-developed software for development scenarios.

When comparing tools, test them on the same wells. Ask how each handles segmented curves and terminal decline, outliers and shut-ins, batch forecasting across thousands of wells, type-curve normalization, version history, and the export of forecasts with their parameters into economics. A fast automatic fit is useful only if a reviewer can see and challenge the assumptions behind it.

From forecast to valuation

A production forecast becomes a value only after it is combined with prices, differentials, ownership, operating costs, capital, and taxes. Keep the forecast and the price assumptions separate and versioned, so that a change in either can be traced. The workflow pages linked from this guide describe that handoff and the price-deck governance that supports it.

The companies and systems on this page are listed alphabetically and are included because their profiles carry public evidence for this work. Inclusion is not an endorsement or a ranking. Confirm current functionality, licensing, and implementation effort with each provider before relying on it for reported reserves.