Urban expansion is accelerating in the tropics, and may promote the spread of introduced plant species into urban-proximate forests. For example, soil disturbance can deplete the naturally high soil nitrogen pools in wet tropical soils, favoring introduced species with nitrogen-fixing capabilities. Also, forest fragmentation and canopy disturbance are likely to favor high-light species over shade-adapted rainforest species. We measured understory woody diversity, the abundance of introduced species, and soil nitrogen and carbon in urban, suburban, and rural secondary forests in Puerto Rico with canopies dominated by (1) native species, (2) introduced Fabaceae (potential nitrogen-fixers), and (3) introduced non-Fabaceae species. We hypothesized that forest stands with introduced Fabaceae in the canopy have higher soil nitrogen levels than stands with other introduced canopy species, and that this higher nitrogen is linked to increased native woody species diversity in the understory. We also predicted that more open canopies and smaller fragment sizes would be positively related with introduced species in the understory, and negatively related with total understory diversity. We found that stands with introduced Fabaceae in the canopy had significantly higher soil nitrogen levels than stands with other non-nitrogen fixing introduced species, and understory woody diversity in Fabaceae stands approached similar diversity levels as stands with native-dominated canopies. As predicted, aboveground stand structure and fragment size were also significantly associated with understory woody diversity across stands. These results suggest that introduced nitrogen-fixing trees may improve recruitment of native woody species in degraded tropical sites where native soil nitrogen is naturally high, particularly as Fabaceae stands mature and canopies close.

Urban areas are expanding rapidly in tropical regions, with potential to alter ecosystem dynamics. In particular, exotic grasses and atmospheric nitrogen (N) deposition simultaneously affect tropical urbanized landscapes, with unknown effects on properties like soil carbon (C) storage. We hypothesized that (H1) soil nitrate (NO 3 À ) is elevated nearer to the urban core, reflecting N deposition gradients. (H2) Exotic grasslands have elevated soil NO 3 À and decreased soil C relative to secondary forests, with higher N promoting decomposer activity. (H3) Exotic grasslands have greater seasonality in soil NO 3 À vs. secondary forests, due to higher sensitivity of grassland soil moisture to rainfall. We predicted that NO 3 À would be positively related to dissolved organic C (DOC) production via changes in decomposer activity. We measured six paired grassland/secondary forest sites along a tropical urban-to-rural gradient during the three dominant seasons (hurricane, dry, and early wet). We found that (1) soil NO 3 À was generally elevated nearer to the urban core, with particularly clear spatial trends for grasslands. (2) Exotic grasslands had lower soil C than secondary forests, which was related to elevated decomposer enzyme activities and soil respiration. Unexpectedly, soil NO 3 À was negatively related to enzyme activities, and was lower in grasslands than forests. (3) Grasslands had greater soil NO 3 À seasonality vs. forests, but this was not strongly linked to shifts in soil moisture or DOC.Our results suggest that exotic grasses in tropical regions are likely to drastically reduce soil C storage, but that N deposition may have an opposite effect via suppression of enzyme activities. However, soil NO 3 À accumulation here was higher in urban forests than grasslands, potentially related to of aboveground N interception. Net urban effects on C storage across tropical landscapes will likely vary depending on the mosaic of grass cover, rates of N deposition, and responses by local decomposer communities.