A discrete step model of low-altitude atmospheric turbulence is described. It has been derived from a commonly used continuous Gaussian process model by considering transitions occurring over discrete intervals. The consequent elimination of fine structure enables additional statistical properties to be calculated. A primary objective is to provide additional data for the assessment of Gaussian process models of low-altitude atmospheric turbulence, which are used for evaluating aircraft take-orf and landing performance. A comparison is made with some measured fluctuations in a strong wind.