Assessing vegetation recovery following wrack disturbances in salt marshes using UAS data

Wrack can smother underlying vegetation in salt marshes, reducing plant productivity. Recovery time for vegetation damaged from wrack disturbances is often dictated by specific wrack or environmental characteristics. We used Unoccupied Aerial Systems (UAS) imagery collected at Dean Creek marsh on Sapelo Island, GA, USA to evaluate vegetation recovery length following wrack disturbances and relate it to wrack characteristics. We implemented a buffering and kriging technique to estimate the NDVI values of wrack-damaged areas and assess vegetation recovery. This method was tested on 24 control polygons and we found that using Normalized Difference Vegetation Index (NDVI) values in a 1-m buffer offset by 1 m for kriging provided consistent estimates of NDVI. We used 176 wrack patches identified in 2021, and applied this method to evaluate recovery lengths using a survivability model. We found that ∼75% of damaged vegetation recovered within 6 months and ∼ 90% within 1 year. We also used a generalized additive model (GAM) to identify how wrack characteristics might influence recovery length and found duration and arrival season were both significant. We found that vegetation damaged by wrack patches that arrived in the fall recovered faster than vegetation damaged by wrack from other seasons; patches with a 1-month duration recovered 2–3 months faster than patches with durations of ≥2 months, and there was little difference in recovery length based on patch area. This study introduces a new framework for evaluating vegetation recovery, showcases the importance of frequent UAS data collection for disturbance assessment, and helps to explain potential drivers of recovery following wrack disturbances.