Vegetation and Habitat Classification of Created and Natural Brackish Marshes via Unoccupied Aerial Systems (UAS): A Case Study of the Lake Hermitage Marsh Creation Project Estuaries and Coasts

Much of Louisiana’s coastal wetlands have been lost over the last century, leading to federal and state agencies allocating billions of dollars towards coastal restoration, flood protection, and marsh creation projects. Traditional post-construction monitoring of marshes involves in situ vegetation sampling and aerial imagery from fixed-wing occupied aircraft, but these methods can be logistically intensive and limited in spatial and temporal resolution. To address these limitations, we evaluated the use of Unoccupied Aerial Systems (UAS) in post-construction monitoring of the Lake Hermitage Marsh Creation Project in Plaquemine Parish, Louisiana, as a case study. Specifically, we used UAS-derived habitat classification maps to compare vegetation cover between created and reference (i.e., natural) marsh sites and conducted a power analysis to quantify the number of in situ plots needed to reliably characterize site-wide vegetation cover. Habitat classification accuracies of UAS-derived maps ranged from 77.9 to 84.5% with slightly lower accuracies at created relative to reference marsh sites due to their more heterogenous vegetation cover. UAS-derived maps discriminated between created and reference marsh sites based on vegetation community similarity, while in situ vegetation monitoring plots did not. Furthermore, our case study illustrates the ability of UAS-based habitat and vegetation classifications to complement, inform, and optimize plot-based, in situ vegetation sampling in future post-construction marsh monitoring plans.