Results/Conclusions Four major findings arose from our study: 1) Mean ANPP was only mildly different among stands dominated by evergreen, deciduous, and an N-fixing tree species; 2) conifer dominated plots had an estimated 3.12 Mg ha-1 y-1 ± 0.95 SE higher ANPP than hardwood dominated plots (P<0.04), while conifer and hardwood dominated plots were not significantly different in estimated net changes in total above-ground C stocks (P>0.05, average 4.72 ± 0.74 SE Mg ha-1 y-1), or NCER (P>0.05; Average 4.0 ± 0.42 SE mmol C m-2 s-1). 3) We found a positive linear relationship between ANPP and tree species richness, where overstory richness explained 46% of NPP (r2=0.46, P<0.05). We also find a positive relationship between average growing-season (April-October) NCER and productivity (r2=0.75, P<0.05), additionally plots with more tree species generally had higher NCER (r2=0.30, P<0.05); and 4) variation in overstory stand dominance was strongly predictive of understory plant community composition (MRPP P<0.05). We found relationships between understory plant diversity and NCER which were weak for species richness (r2=0.004, P<0.05) and stronger for Shannon’s (r2=0.22, P<0.05) and Simpson’s (r2=0.28, P<0.05) diversity indices. However, ANPP, net changes in C stocks, and understory plant communities were unrelated (P>0.05). These data highlight variability in second-growth forests that may have strong implications for net ecosystem productivity (NEP). We estimate NPP at 7.75 Mg C ha-1 yr-1, and suggest that this and other plot level measurements imply an upper bounds on NEP of ~3.7 Mg C ha-1 yr-1, and a lower bounds of ~2.7 Mg C ha-1 yr-1.