Determining Changes in Land Use Structure in Malopolska Using Statistical Methods

For this paper we studied changes in areas of four land use variables were in the Malopolska Voivodeship of Poland. The following annual land use time series were used: surface area of agricultural lands, surface area of forest lands and wooded and bushy areas, surface area of built-up and urbanized lands, and surface area of wasteland. The autocorrelation structure and the temporal changes were studied using the autocorrelation function (ACF) and the Maim-Kendall test for monotonic trend, respectively. The parameters of the deterministic component of the time series were identified using the method of least squares. The Akaike information criterion was used in the final choice of this component. The stochastic component was studied using statistical tests. Coefficient of determination and mean absolute relative error were used as diagnostic measures of the models. Results show a significant dependence of the current land use values on the past values from the previous two years. A negative trend was observed in the series of agricultural land surface area and a positive trend in the other three series. Temporal evolution of the series includes both quadratic trend and Gaussian white noise. In conclusion, temporal monotonic changes of the four land use variables were demonstrated in Malopolska. Some options for land use decision-making are clarified.