Mangrove expansion exceeds loss across South Florida despite major storm damage in the last 38 years

Mangrove forests are disturbance adapted ecosystems that provide high levels of ecosystem services. Mangrove survival depends on either increasing soil elevation to survive in place or migrating to higher ground in the wetland interior. Due to their coastal location, mangrove ecosystems bear the brunt of sea level rise, enhanced storms and drought driven by climate change. However, with adequate accommodation space mangroves are also expanding their range inland. We compiled Landsat imagery from 1983 to 2021 to identify changes in normalized difference vegetation index (NDVI) in mangrove ecosystems in South Florida, USA, (centered on the Everglades, but including the Florida Keys, Biscayne National Park, Ten Thousand Islands National Wildlife Refuge and surrounding areas). We also applied a random forest algorithm to identify changes in mangrove extent over time to identify mangrove encroachment and conversion to alternative ecosystem types across the landscape. By identifying changes in NDVI and vegetation classes over 38 years, we quantify four categories of change: 1) mangrove loss to other ecosystem types, 2) mangrove degradation, 3) mangrove gain through encroachment into non-mangrove ecosystems and 4) increased NDVI within mangrove ecosystems. Our results show widespread damage and loss in mangrove areas punctuated by major storm impacts, particularly along the coast and in low-elevation basins. However, the overall area of mangrove gain exceeded mangrove area loss by two-fold from 1983 to 2021, largely through encroachment into interior wetlands. We illustrate the role of storms on mangrove loss, degradation and expansion, with storms creating immediate mangrove loss and degradation, but also driving a delayed expansion of mangroves toward wetland interiors. Our findings suggest that mangroves are increasingly at risk due to climate change and storms in their current locations, but are currently expanding their range inland. Finally, we model net carbon dynamics across the region and discuss the uncertainty associated with scaling disturbance-driven carbon change.