Executive Summary
Habitat Evolution Mapping Project (HEMP) and the Baylands Habitat Map (BHM) are both effective mapping products that can be used for longitudinal habitat change analyses, but direct comparisons between the two require adjustments due to differences in methodology. HEMP is created using pixel-based supervised classification on spectral imagery, while BHM is created using object-based image classification of elevation, imagery, and auxiliary information. BHM uses elevation and tidal data to specify a region’s position within the intertidal range, while HEMP strictly maps the directly observable land cover. HEMP focuses on specific vegetation classes and uses those to define habitat types, while BHM defines habitat types directly. Defining vegetation species into higher-level habitat types like upland vegetation or high/low marsh can be nuanced, and HEMP often makes generalizations about these classes in order to aggregate the mapping into higher-level habitats.
BHM uses auxiliary data to define management status of certain marshes, while HEMP does not distinguish tidal connectivity. On occasion, BHM will identify certain regions as one managed class regardless of the habitat types present in the managed area. Because BHM maps tidal connectivity, it distinguishes between tidally connected marshes and managed marshes. HEMP’s methods do not differentiate between vegetation in a tidal marsh versus the same species of vegetation in a managed marsh, or in upland areas. HEMP results can be more spatially heterogeneous as they are mapped on a per-pixel basis (at 50 cm resolution), while BHM uses minimum mapping units (usually 5 square meters), which results in more spatially contiguous classified areas.
The greatest difference in results between HEMP and BHM are mostly located in managed or restored regions, which reflect methodological differences and limitations on data availability. While a direct comparison between HEMP and BHM can be messy, simplified metrics such as total tidal marsh extent and the unvegetated to vegetated ratio can allow for foundational assessments of the success of restoration projects. We found that HEMP and BHM generally align at this level of analysis and there are no unexplained systematic differences.