Moment: Statistical Moments, Tracer-Test Analysis, and Reservoir Property Distributions
A moment, in the statistical sense used throughout reservoir engineering and petrophysics, is the mean or expected value of the product formed by multiplying together a set of one or more variates or variables, each raised to a specified power. In its most common single-variable form, the nth moment of a distribution is the expected value of the variable raised to the power n, and moments taken about the mean rather than about the origin are called central moments. This compact definition underpins a surprising amount of quantitative oilfield work, because moments are the numbers that summarise the shape of any probability distribution or any time-varying response. The first moment about the origin is simply the mean, the centre of mass of the distribution. The second central moment is the variance, which measures spread. The third central moment, normalised, gives skewness, a measure of asymmetry, and the fourth gives kurtosis, a measure of how heavy the tails are. Reservoir characterization lives and dies by distributions, so the moments of those distributions are everywhere: the mean and variance of a porosity population define the input to a geostatistical model, the skewness of a permeability histogram warns that a log-normal transform is needed before kriging, and the moments of a grain-size or pore-throat distribution characterise reservoir quality in WCSB sandstones like the Cardium or Viking. Moments also have a powerful application in dynamic data analysis. In tracer-test and pressure-transient interpretation, the method of moments treats a measured response curve, for example the concentration of an injected chemical tracer arriving at a producing well, as a distribution in time, and computes its moments to extract physical reservoir properties without fitting a full analytical model. The zeroth moment of a tracer breakthrough curve gives recovered mass, the first moment gives the mean residence time and therefore the swept pore volume between an injector and a producer, and the second moment gives dispersion, a measure of reservoir heterogeneity and channeling. In nuclear magnetic resonance logging, the moments of the T2 relaxation-time distribution summarise pore-size character and bound-fluid volume. Because a handful of moments can capture the essential behaviour of an entire curve or population, they let engineers reduce noisy, high-dimensional data to a few robust, physically meaningful numbers, which is why the concept recurs across geostatistics, well testing, enhanced-recovery surveillance, and formation evaluation.
Key Takeaways
- Expected value of a power: The nth moment is the expected value of a variable raised to the power n. The first moment about the origin is the mean; central moments are taken about the mean. This single definition generates the family of descriptors, mean, variance, skewness, and kurtosis, that summarise the location, spread, and shape of any distribution used in reservoir work.
- Mean, variance, skewness, kurtosis: The first moment is the mean (centre), the second central moment is the variance (spread), the normalised third is skewness (asymmetry), and the fourth is kurtosis (tail weight). Together these four moments characterise a WCSB porosity or permeability population well enough to drive a geostatistical reservoir model.
- Method of moments in tracer tests: Treating a tracer breakthrough curve as a time distribution, its zeroth moment gives recovered mass, its first moment gives mean residence time and swept pore volume between injector and producer, and its second moment gives dispersion. This model-free analysis quantifies interwell connectivity and heterogeneity in WCSB waterfloods.
- Heterogeneity and the log-normal warning: A strongly skewed permeability histogram, a high third moment, signals that the data are closer to log-normal than normal, so a log transform is required before kriging or simulation. Ignoring the moment structure of WCSB permeability data leads to biased upscaling and unreliable flow predictions.
- NMR and curve summarisation: Moments of the NMR T2 relaxation distribution summarise pore-size character and partition bound-fluid from free-fluid volume in WCSB tight reservoirs. More broadly, reducing a noisy response curve to a few moments yields robust, physically interpretable numbers that are far less sensitive to data scatter than full curve fitting.
The Method of Moments in WCSB Tracer Surveillance
In a Western Canadian Sedimentary Basin waterflood, operators inject a chemical or radioactive tracer at an injection well and monitor its arrival at surrounding producers to map fluid movement. Rather than fit a complex transport model, the analyst computes the moments of each producer's breakthrough curve. The first temporal moment, the mean arrival time, multiplied by injection rate, estimates the swept pore volume along that flow path, which directly indicates how much of the reservoir between the wells is being contacted. A short mean residence time and a small dispersion moment point to a high-permeability channel or fracture short-circuiting the flood, while a long, broad response indicates good areal sweep. These moment-derived diagnostics guide infill drilling and conformance decisions on mature Cardium and Viking floods.
Moments as Inputs to Geostatistical Models
Before any geostatistical realisation of a WCSB reservoir is built, the moments of the underlying property distributions must be honoured. The mean and variance of porosity, the first and second moments, set the global statistics that kriging or sequential simulation must reproduce. The skewness, the third moment, dictates whether a normal-score or log transform is applied so the simulated field matches reality at the extremes that control flow. A model that reproduces the mean but ignores the higher moments of permeability will systematically misestimate connectivity and recovery, so quality-control workflows check the moments of the simulated cells against the moments of the input well data.
Fast Facts
The statistical concept of the moment borrows its name directly from mechanics, where a moment is the product of a force and its distance from a pivot. Karl Pearson, who systematised much of modern statistics around 1900, popularised the term because the formula for the centre of mass of a distribution is mathematically identical to the formula for the first moment of a probability density. The same arithmetic that locates the balance point of a beam also locates the mean of a permeability histogram, which is why the borrowed name stuck across a century of reservoir science.
Related Terms
Moments quantify the shape of a Probability Distribution, the function whose mean, variance, and skewness they measure. They are the foundation of Geostatistics, because honouring the moments of well data is what makes a simulated reservoir model statistically faithful. The method of moments is applied in Tracer testing to extract swept volume and dispersion from breakthrough curves, and the second central moment, the Variance, is the most-used single moment in reservoir uncertainty analysis.
Real-World WCSB Scenario: Diagnosing a Channeling Waterflood
An operator running a mature Viking waterflood near Provost suspects that injected water is short-circuiting to one producer. A CAD 180,000 interwell tracer program injects a unique tracer per injector and samples produced water for several months. Computing the moments of each breakthrough curve, the reservoir engineer finds one producer with a mean residence time less than a third of its neighbours and a very low dispersion moment, the signature of a high-permeability thief zone carrying water directly from injector to producer with little sweep.
Acting on the moment analysis, the operator runs a polymer conformance treatment on the offending injector to plug the channel, and follow-up tracer moments show a longer residence time and broader response, confirming improved areal sweep and incremental oil recovery from the pattern.