Influenced by human, land, and oceanic drivers, coastal wetlands exist within highly dynamic environments. Coastal wetlands store proportionally larger amount of carbon (C) compared to their terrestrial counterparts and are increasingly important to stakeholders. Recent advances have been made to simulate ecosystem-wide coastal wetland C and nutrient cycling using land surface models (LSMs), along with observational data. Most LSMs simulate different ecosystems based on large-scale vegetation types known as Plant Functional Types (PFTs). While recent modeling incorporates salinity and inundation responses and improves biomass allocation parameterization to more accurately represent coastal wetland PFTs, existing coastal wetland PFTs fail to capture the biodiversity of coastal wetland vegetation, such as mangroves and succulent marshes. Mangroves have uniquely adapted growth forms within highly dynamic environments, making them challenging to represent within models. Our study adapted a tropical broadleaf evergreen PFT into a mangrove PFT within the Energy Exascale Earth System Model (E3SM) Land Model (ELM). Specifically, we address the following research question: How can the integration of observational data enhance the representation of sub-tropical mangrove traits and environmental controls in ELM within the context of the Florida Coastal Everglades? We selected the Florida Coastal Everglades (FCE) Long-Term Ecological Research (LTER) site to test our newly defined mangrove PFT. Using data from the Shark River Slough (SRS-6) flux tower, we calibrate ELM parameters and validate model outputs. Our study highlights the value of representing coastal wetlands diversity and shows the utility of model-experimental integration of intensive field sampling in hard-to-access flooded coastal zones to enhance the computational accuracy of Earth system models.