Accounting for spatial patterns in deriving sea-level rise thresholds for salt marsh stability: More than just total areas?

Ecological threshold is an important concept to indicate the boundary of alternate states of ecosystems driven by environmental conditions and to facilitate evaluation of ecosystem resilience. Sea-level rise (SLR) thresholds for the stability of salt marshes, if studied in two dimensions, are generally derived based on total areas without systematic accounting for spatial patterns related to edges, shapes, and contagions of patches. As these spatial patterns are potentially important for functions and ecosystem services of salt marshes and they are likely to be impacted by SLR in a different way from the total areas, it is necessary to study SLR thresholds based on these spatial patterns to obtain a more comprehensive understanding of salt marsh resilience to SLR. This research compares the SLR thresholds based on these spatial patterns of salt marshes to those based on total areas alone across different spatial resolutions. The spatial patterns of salt marshes were quantified by 26 commonly used landscape metrics, predicted from a mechanistic wetland change model. At spatial resolutions of 2-100 m, SLR thresholds were first derived using individual landscape metrics and then the first principal component that explained> 80% of total variance of these metrics showing threshold responses to SLR. In order to separate the effect of spatial configuration from composition, a neutral model which simulated the same amount of salt marsh change as the mechanistic model but at the random locations was applied. The SLR thresholds were derived based on the simulations from the neutral model and compared to those from the mechanistic model. The results show that total area-based SLR thresholds do not comprehensively represent salt marshes' resilience to SLR. Particularly, I find 1) the derived SLR thresholds vary from 7.29 to 11.12 mm/yr for 2100 based on landscape metrics used, 2) the SLR threshold based on the first principal components (7.99 mm/yr) is smaller than that based on the total area only (8.40 mm/yr), 3) the SLR thresholds are scale dependent, and 4) the spatial configuration' effect on SLR thresholds is smaller for smaller salt marsh areas compared to larger salt marsh areas. This study highlights the need to account for different spatial patterns of salt marshes and apply wetland maps with a spatial resolutions of 30 m or finer in deriving SLR thresholds, as using total areas alone or coarserresolution maps may provide a biased interpretation that salt marshes are more resilient to SLR than they actually are.