Context.It has long been known that the high-speed stream (HSS) peak velocity at Earth directly depends on the area of the coronal hole (CH) on the Sun. Different degrees of association between the two parameters have been shown by many authors. In this study, we revisit this association in greater detail for a sample of 45 nonpolar CHs during the minimum phase of solar cycle 24. The aim is to understand how CHs of different properties influence the HSS peak speeds observed at Earth and draw from this to improve solar wind modeling.Aims.The CHs were extracted based on the Collection of Analysis Tools for Coronal Holes which employs an intensity threshold technique applied to extreme-ultraviolet filtergrams. We first examined all the correlations between the geometric characteristics of the CHs and the HSS peak speed at Earth for the entire sample. The CHs were then categorized in two different groups based on morphological criteria, such as the aspect ratio and the orientation angle. We also defined the geometric complexity of the CHs, a parameter which is often neglected when the formation of the fast solar wind at Earth is studied. The quantification of complexity was done in two ways. First, we considered the ratio of the maximum inscribed rectangle over the convex hull area of the CH. The maximum inscribed rectangle provides an estimate of the area from which the maximum speed of the stream originates. The convex hull area is an estimate of how irregular the CH boundary is. The second way of quantifying the CH complexity was carried out by calculating the CH’s fractal dimension which characterizes the raggedness of the CH boundary and internal structure.Methods.When treating the entire sample, the best correlations were achieved between the HSS peak speed observed in situ, and the CH longitudinal extent. When the data set was split into different subsets, based on the CH aspect ratio and orientation angle, the correlations between the HSS maximum velocity and the CH geometric characteristics significantly improved in comparison to the ones estimated for the whole sample. By further dividing CHs into subsets based on their fractal dimension, we found that the Pearson’s correlation coefficient in the HSS peak speed – CH area plot decreases when going from the least complex toward the most complex structures. Similar results were obtained when we considered categories of CHs based on the ratio of the maximum inscribed rectangle over the convex hull area of the CH. To verify the robustness of these results, we applied the bootstrapping technique. The method confirmed our findings for the entire CH sample. It also confirmed the improved correlations, compared to the ones found for the whole sample, between the HSS peak speed and the CH geometric characteristics when we divided the CHs into groups based on their aspect ratio and orientation angle. Bootstrapping results for the CH complexity categorizations are, nonetheless, more ambiguous.Results.Our results show that the morphological parameters of CHs such as the aspect ratio, orientation angle, and complexity play a major role in determining the HSS peak speed at 1 AU. Therefore, they need to be taken into consideration for empirical models that aim to forecast the fast solar wind at Earth based on the observed CH solar sources.