Tracking annual changes in the distribution and composition of saltmarsh vegetation on the Jiangsu coast of China using Landsat time series-based phenological parameters

Saltmarshes provide among the most valuable ecosystem services but have undergone continuous modification under the combined pressures of climate change and anthropogenic disturbance. Dynamic monitoring of the distribution and composition of saltmarsh vegetation is an urgent need for understanding the balance of saltmarshes. Previous studies have examined the possibility of identifying the plant species of saltmarshes using Land Surface Phenology (LSP). However, balancing species identification with the timespan of the monitoring remains challenging, given that the available observations are limited by the frequent occurrence of cloudy weather and periodic tidal fluctuations in coastal areas. To overcome the difficulty, we first filtered out observations contaminated by cloud cover and tidal flooding, then proposed an intra-annual time-series model by incorporating inter-annual observations using the inter-calibrated Landsat observations, finally classified the plant species of saltmarshes using phenological parameters derived from the intra-annual time-series model. The annual changes of the three dominant plant species-Spartina alterniflora (S. alterniflora), Suaeda salsa (S. salsa), and Phragmites australis (P. australis) during 2000-2020 in Jiangsu Province, the most abundant salt marsh area of China, were tracked. The results demonstrate the following: (i) The intrinsic phenological differences among the plant species (similar to 32 days between S. alterniflora and P. australis) are greater than the latitudinal and temporal phenological variations (9-24 days). Large-scale and long-term classification mapping using LSP was thus feasible. (ii) The annual classification maps achieved a favorable performance with an average overall accuracy of 84.8% with minor fluctuations of 5.5%. Sufficient available observations for the spring and autumn seasons (R-2 = 0.78, p < 0.05) and for locations far from the sea (>240 m) were the reasons for this good performance. (iii) The saltmarsh vegetation was resilient to tidal flat reclamation on the Jiangsu coast, as indicated by a slight increase in the area from 300.30 to 336.38 km(2). However, this was concentrated within the Nature Reserve (38.3%), with the increasing dominance of two species: S. alterniflora (49.6%) and P. australis (42.7%). Special conservation measures are needed for S. salsa to prevent its continuing decrease and possible future disappearance.

Automated summary

This study used Land Surface Phenology (LSP) to dynamically monitor and classify saltmarsh vegetation in the Jiangsu Province of China. The results showed the feasibility of LSP for large-scale and long-term classification mapping, despite limitations due to weather and tidal fluctuations. The study also discovered the resilience of saltmarsh vegetation to tidal flat reclamation, but the need for conservation measures to prevent the decline of Suaeda salsa species.