Forest spatial patterns influence many ecological processes in dry conifer forests. Thus, understanding and replicating spatial patterns is critically important in order to make these forests sustainable and more resilient to fire and other disturbances. The labor and time required to stem-map trees and the large plot size (> 0.5 ha) needed to study tree spatial patterns have limited our examination of how these patterns change as a function of site conditions and tree densities. We stem-mapped all trees>40 cm DBH within two large relict (minimally logged) pure ponderosa pine study sites on experimental forests at Long Valley (73 ha) on sedimentary soils and Fort Valley (32 ha) on basalt soils in northern Arizona, USA. We also simulated 1,000 4-ha plots from models of each study site incorporating field data parameters. Using cluster analysis and field data, we found that an inter-tree distance (ITD) of 9-11m best separated single trees and groups within our study sites. Using a fixed 10-m ITD, the more productive Long Valley (LV) site had 62 trees ha-1 and groups of up to 113 trees, compared to the Fort Valley (FV) site, which averaged 41 trees ha-1 and had 22 trees in the largest group. However, the sites differed only slightly in terms of single trees ha-1 (LV 7.3; FV 5.6) and group of tree ha-1 (LV 7.2; FV 8.1). Simulation results indicated that when tree densities are equal, the spatial patterns were very similar between the two sites, suggesting that tree spatial pattern variability is a function of tree densities and only indirectly related to site productivity. As the number of trees increased, the additional trees integrated into existing groups rather than creating new groups. In addition to tree spatial patterns, we quantified gaps (defined as>30m wide stem-to-stem) and openings (defined as ≥30m wide stem-to-stem) within the two study sites. Although both sites were dominated by small openings most of the open area was found within a few large openings. Our large plots allowed us to incorporate variability and capture a larger range of tree and openings spatial patterns than have been captured in previous studies to provide insights on spatial heterogeneity that can inform management of this important forest type in North America.