What Is Log-Normal Distribution?

Log-normal distribution is a continuous probability distribution of a positive-valued random variable whose natural logarithm follows a normal (Gaussian) distribution. If a random variable X is log-normally distributed, then Y = ln(X) has a normal distribution with mean mu and standard deviation sigma. The distribution is fully characterized by these two parameters, which govern its shape and scale on a logarithmic axis. It appears whenever a quantity arises from the multiplicative product of many independent positive factors, by analogy to the central limit theorem, which applies to sums of independent variables.

The distribution was studied in the nineteenth century in the context of particle size measurement and income distributions, but it found particularly wide application in engineering, telecommunications, and reliability analysis throughout the twentieth century. Because it is defined only for positive values and exhibits right-skewed behavior, it is well suited to quantities that cannot go negative and that span several orders of magnitude.

Mathematical Properties

The log-normal distribution has a probability density function that is zero for non-positive arguments and positively skewed for all parameter values. Its shape parameter sigma controls the degree of skew: small sigma values produce a distribution that resembles a normal distribution, while large sigma values produce a heavily skewed distribution with a long right tail. The median of X equals exp(mu), while the mean equals exp(mu + sigma^2/2), so the mean always exceeds the median. Variance, skewness, and kurtosis all increase monotonically with sigma. The NIST/SEMATECH e-Handbook of Statistical Methods provides closed-form expressions for these moments and serves as a standard reference for engineering applications of the distribution.

Because the logarithm of a log-normal variable is normally distributed, most analytical operations reduce to working with the underlying Gaussian. This property makes maximum-likelihood parameter estimation straightforward: compute the sample mean and variance of the log-transformed observations to obtain estimates of mu and sigma directly.

Wireless Channel Shadowing

In wireless communications, log-normal distribution is the standard model for large-scale shadowing, which describes variations in received signal strength caused by buildings, terrain, and vegetation blocking the propagation path. As a mobile receiver moves through an environment, the cumulative obstruction effect of many randomly placed obstacles produces a received power that is log-normally distributed around the local median predicted by a path-loss model. This is the basis of the log-distance path-loss model widely used in cellular network planning. Stanford course materials on wireless channel modeling detail how the shadowing standard deviation, typically between 4 and 12 dB depending on the environment, enters into link budget calculations and determines the required fade margin.

The MDPI journal article on composite shadowing distributions for advanced cellular systems examines how log-normal shadowing combines with fast-fading distributions such as Rayleigh or Nakagami-m to form composite channel models used in coverage prediction for 4G and 5G deployments.

Applications

Log-normal distribution has applications in a wide range of fields, including:

• Wireless network planning, where it models received signal variation due to shadowing in cellular and Wi-Fi systems
• Reliability engineering, where component lifetimes under fatigue or corrosion often follow log-normal behavior
• Biomedical data analysis, where physiological measurements such as blood pressure and enzyme concentration exhibit log-normal patterns
• Environmental science, where particle size distributions in aerosols and pollutant concentrations are modeled log-normally
• Financial modeling, where asset prices and returns are approximated as log-normal over short time horizons