Scythian Landscapes - Report

For the final Scythian mound lab we were tasked with creating a final predictive result based on our previous work over the last two weeks.  We looked at spatial distribution of burial mounds in the nearby region of Tuekta, Russia. Three secondary coverages were created for statistical analysis showing elevation, slope, and aspect. Additionally, we combined a shapefile showing known burial mounds and random points then followed that with an OLS regression analysis.

The results for the analysis showed an R-squared value of 0.714636, which suggests that the three surface variables of the analysis account for roughly 71.4% of the sites in the predictive model. The coefficient for elevation was 0.633116, the coefficient for slope was 0.073404, and the coefficient for aspect was 0.076252.  Because the coefficients for each variable are positive and not near zero, indicates the variables are contributing to the model.   Spatial Autocorrelation results show a z-score of 14.811686 and a p-value of 0.00.  A high z-score indicates a normal distribution of data. The p-score indicates that there is less that 1% likelihood that the clustered patern could be the result of random chance.

Because the results show clusters with valleys, it may be beneficial to add a variable showing access to waterways/hydrography. Further analysis could be done to confirm that the model is producing tangible data by adding additional variables, ground-truthing, and perhaps additional regression models.

http://arcgis.com/explorer/?open=195a401048ad4f51a49970fc5e8938ba&extent=9544873.05706802,6549970.40297947,9624367.56646057,6592469.39069318