A global biophysical typology of mangroves version 3

This dataset in an updated version of:

Worthington, T. A. et al. A global biophysical typology of mangroves and its relevance for ecosystem structure and deforestation. Sci. Rep. 10, 14652 (2020)

which delineates the world’s mangroves into geomorphic units based on their biophysical setting. Each unit consists of one or more patches of mangrove, grouped based on their proximity to macroscale coastal features, with these features determining their geomorphic class – deltaic, estuarine, lagoonal, and open coast.

With the development of an updated mangrove extent timeseries (Global Mangrove Watch (GMW) v3.12), we updated the mangrove biophysical typology¹ to match this new extent. We firstly created an overlay between GMW v3.12 (all years combined to produce a 24-year composite extent) and the mangrove typology (v2.2) and identified those patches that were present in both. These patches were assigned to the same geomorphic class and individual geomorphic unit as the mangrove typology and provided the basis for the updated version. The patches present in the typology v2.2 but not in GMW v3.12 were deleted as they were no longer being mapped as mangrove in the GMW dataset.

The patches now being mapped as mangrove in GMW v3.12 but had not been identified as such in the previous extent used to create the typology were then assigned to a geomorphic type and individual geomorphic unit using an iterative approach. Firstly, we identified patches that intersected with a single geomorphic unit and merged those patches to that unit, creating an enlarged unit extent. We repeated this procedure with the enlarged units, again enlarging their extent with patches only intersecting a single unit.

We then used a series of distance buffers to identify unassigned patches that were within a certain distance of a single unit. After each step patches that were within the buffer distance of a single geomorphic unit were merged with that unit, and then the buffer was recalculated. The buffer distances were 1000m, 1000m, 1000m, 500m, 250m and 100m. Following the buffer, for the remaining unassigned patches we split them into those whose centroid was ≤ 10,000m from a geomorphic unit and those whose centroid was >10,000m for a geomorphic unit. As some of the patches were close (≤ 10,000m) from multiple geomorphic units, they were manually assessed and their assignment was corrected where necessary.

The remaining patches (>10,000m from a geomorphic unit) were then visually assessed and can be split into three groups, 1) those part of large existing geomorphic units (only deltas, estuaries and lagoons) that were merged with that unit, 2) patches near deltas, estuaries and lagoons not mapped in the original GMW dataset, and 3) areas of open coast. The patches near deltas, estuaries and lagoons not mapped in the original GMW dataset resulted in the creation of 81 new geomorphic units. The open coast patches were aggregated into 268 clusters using a distance of 10,000m. Thirty-eight of the clusters were within 10,000m of an original open coast geomorphic unit and were merged with that unit. The remaining 230 were designated as new geomorphic units.

We then undertook a process to merge open coast geomorphic units, by finding those small (<1km²) open coast geomorphic units that were within 10,000m of a larger one. Repeating the procedure to merge small open coast geomorphic units that were within 10,000m of another small open coast geomorphic unit. The final step was to do a visual assessment of all the units to remove errors. This was based around merging neighbouring geomorphic units of the same class if they represent the same system (e.g., one contiguous estuary or lagoon unit), assessed using high resolution imagery and the fluvial boundaries of the Hydrosheds basins². Manually editing errors at unit boundaries where patches of one unit were surrounded by another unit. Splitting open coast units that overlapped another class of geomorphic unit e.g., an open coast unit with an estuary in the middle of it. Merging open coast units into large extents, particular those of the same section or aspect of the coast, using a distance of 10,000m as an approximate guide and trying not to create extents >100km².

A final publication version of the GMW dataset³ (v3.14) was released https://zenodo.org/record/6894273, which differed slightly from v3.12. Firstly, a number of small areas of mangrove were removed at the edges of polygons, these were also removed from the typology. Secondly, additional areas of mangrove were mapped in the Persian Gulf, these new areas were merged with existing geomorphic units. These steps resulted in a final dataset ‘Mangrove Typology v3’ consisting of 3983 geomorphic units.

1.         Worthington, T. A. et al. A global biophysical typology of mangroves and its relevance for ecosystem structure and deforestation. Sci. Rep. 10, 14652 (2020).

2.         Lehner, B. & Grill, G. Global river hydrography and network routing: Baseline data and new approaches to study the world’s large river systems. Hydrol. Process. 27, 2171–2186 (2013).

3.         Bunting, P. et al. Global mangrove extent change 1996-2020: Global Mangrove Watch version 3.0. Remote Sens. 14, 3657 (2022).