Page 31 - GIS for Science, Volume 3 Preview
P. 31
The intended practical value of the CSU data is that any 1 km segment on the global coastline can be queried in this manner, supporting the understanding of place-based coastal zone dynamics, and facilitating the comparison of stretches of coastline at local scales. Coastal classifications are used in many ways, including coastline vulnerability studies;6,23,56 conservation planning;4 ecosystem services assessments;5 understanding of coastal ecosystem structure, function, and distributions;24,27,28 design of marine protected area (MPA) networks;15 and other applications. Our team hopes that the practical value of the CSU classification will be realized by its adoption and use in these and other applications.
Ecological coastal units (ECUs)
The global assessment of optimal cluster numbers identified 16 clusters, shown in the world map of ECUs. Fewer than 16 clusters appeared largely responsive to latitudinal gradients, whereas more than 16 clusters did not identify new regional groupings, instead demonstrating successive partitioning of the parent clusters. The 16 clusters represent the identification of a preliminary set of standardized and replicable ECUs with potential utility for understanding global scale differences in coastal ecological settings and for global scale conservation assessments and priority setting. The world map below depicts the ECUs at a glance with maximum color separation. An easier interpretation of the individual ECU distributions is presented in the set of 16 maps on the next pages showing the global distributions of each ECU.
Several ECU distributions appear to have strong climate region associations, mainly driven by latitudinal temperature gradients. ECUs 1 and 4 are clearly tropical systems. ECU 2 has a more subtropical distribution with extensions into the warm temperate climate regions. ECU 6 is largely distributed in warm temperate regions. ECUs 7, 9, 15, and 16 are largely boreal. ECUs 8, 10, 11, 12, and 14 are largely distributed in polar regions, and of these, two (ECUs 8 and 12) are only distributed in the Northern Hemisphere, while ECUs 10, 11, and 14 are found in both hemispheres.
While none of the ECUs show a strong pole-to-pole distribution, ECUs 3 and 5 span multiple latitudes, indicating that the climate region influence in these clusters is low. ECU 5 exhibits a strong pattern of distribution along major estuaries/river mouths. ECU 16 shows an interesting distribution in coastal areas characterized by numerous coastal islands (Chile, Norway) and island chains.
The 10-dimensional nature of the dataset complicates visual identification of additional patterns in ECU distributions. The exact nature of each ECU, including the relative influences of the input variables driving their existence and distributions, can be quantitatively described, and the data are available. We summarize the main properties and descriptive statistics of the ECUs in the appendix. The team did not present a rigorous assessment and comparison of the ECU compositional properties herein, as our purpose is rather to describe the development of the ECUs
(continued on page 24)
The global distributions of the 16 ECUs, with maximum color separation for easier visual differentiation.
Earth’s Coastlines 19

