Longleaf pine (Pinus palustris Mill.) forests in the southeastern United States are considered endangered ecosystems, because of their dramatic decrease in area since European colonization and poor rates of recovery related to episodic natural regeneration. Sporadic seed production constrains restoration efforts and complicates sustainable management of this species. Previous studies of other tree species found invariant scaling properties in seed output. Here, using long-term monitoring data for cone production at seven sites across the native range of longleaf pine, we tested the possible presence of two types of power laws. Findings indicate that (i) the frequency distribution of cone production at seven sites, from 1958 to 2014, follows power law relationships with high level of significance; (ii) although there is no general trend in the dynamics of scaling exponents among all sites, there are dynamics of scaling exponents at each site, with sudden changes in scaling exponents generally corresponding to the years of higher or lower cone production; and (iii) Taylor‟s power laws explain cone production at different locations, but the scaling exponents vary among these. Results from this computational approach provide new insight into the irregular cone production of longleaf pine at spatial and temporal scales. Integrated ecosystem monitoring will be necessary to more fully understand future changes in cone production.