Harvesting wild seaweeds has a long history and is still relevant today, even though aquaculture now supplies >96% of global seaweed production. Current wild harvests mostly target canopy-forming kelp, rockweed and red macroalgae that provide important ecosystem roles, including primary production, carbon storage, nutrient cycling, habitat provision, biodiversity and fisheries support. Harvest methods range from selective hand-cutting to bottom trawling. Resulting ecosystem impacts depend on extraction method and scale, ranging from changes in primary production to habitat disruption, fragmentation, food-web alterations and bycatch of non-target species. Current management often aims for sustainable harvesting in a single-species context, although some agencies acknowledge the wider ecosystem structure, functions and services seaweeds provide. We outline potential ecosystem-based management approaches that would help sustain productive and diverse seaweed-based ecosystems. These include maintaining high canopy biomass, recovery potential, habitat structure and connectivity, limiting bycatch and discards, while incorporating seasonal closures and harvest-exclusion zones into spatial management plans. Other sustainability considerations concern monitoring, enforcement and certification standards, a shift to aquaculture, and addressing cumulative human impacts, invasive species and climate change. Our review provides a concise overview on how to define and operationalize ecosystem-based management of seaweed harvesting that can inform ongoing management and conservation efforts.

<p>Glacial erosion has often been parameterized as proportional to glacier sliding velocity, while the role played by local geology, hydrology and climate remain largely unquantified. As a result, our understanding of the links between global climate, tectonics and glacial erosion is limited. To address this shortcoming, we present a comprehensive synthesis of previously published Quaternary glacial erosion rates from six different measurement techniques integrated over 10<sup>-2</sup> to 10<sup>6</sup> years: (i) instrumental measurements beneath active glaciers, (ii) sediment fluxes derived from meltwater streams or (iii) ice-marginal deposits, (iv) terrestrial cosmogenic nuclide dating (TCN), (vi) luminescence thermochronometry, and (v) relief generation of chronologically constrained surfaces. Our synthesis includes 1065 empirical data points and 465 erosion rates from ice sheets, ice caps, and topographically confined glaciers that range over six orders of magnitude, between 10<sup>-4</sup> and 100 mm yr<sup>-1</sup>. Using a filtered dataset of contemporary erosion rates, we apply machine learning tools, using available environmental, glaciological, and geological datasets to assess the dominant controls on subgroups of nominal data categories.  On a global scale, while glacial sliding velocity is an important control, we also discover equally strong or stronger correlations between other glaciological, environmental, and lithological parameters and glacial erosion rate, some of which have not been previously documented. </p><p> </p>