The Potential of Satellite Remote Sensing Time Series to Uncover Wetland Phenology under Unique Challenges of Tidal Setting

Abstract

While growth history of vegetation within upland systems is well studied, plant phenology within coastal tidal systems is less understood. Landscape-scale, satellite-derived indicators of plant greenness may not adequately represent seasonality of vegetation biomass and productivity within tidal wetlands due to limitations of cloud cover, satellite temporal frequency, and attenuation of plant signals by tidal flooding. However, understanding plant phenology is necessary to gain insight into aboveground biomass, photosynthetic activity, and carbon sequestration. In this study, we use a modeling approach to estimate plant greenness throughout a year in tidal wetlands located within the San Francisco Bay Area, USA. We used variables such as EVI history, temperature, and elevation to predict plant greenness on a 14-day timestep. We found this approach accurately estimated plant greenness, with larger error observed within more dynamic restored wetlands, particularly at early post-restoration stages. We also found modeled EVI can be used as an input variable into greenhouse gas models, allowing for an estimate of carbon sequestration and gross primary production. Our strategy can be further developed in future research by assessing restoration and management effects on wetland phenological dynamics and through incorporating the entire Sentinel-2 time series once it becomes available within Google Earth Engine.

Keywords

• tidal wetlands
• enhanced vegetation index
• phenology
• modeling
• Google Earth Engine
• Landsat