We constructed mass balances of both calcium and phosphorus for two watersheds in Big Cypress National Preserve in southwest Florida (USA) to evaluate the time scales over which its striking landscape pattern developed. This low-relief carbonate landscape is dotted with evenly spaced, evenly sized, shallow surface depressions that annually fill with surface water and thus support wetland ecosystems (e.g. cypress domes) embedded in a pine-dominated upland matrix with exposed bedrock. Local and landscape scale feedbacks between hydrology, ecological dynamics and limestone dissolution are hypothesized to explain this karst dissolution patterning. This hypothesis requires the region to be wet enough to initiate surface water storage, which constrains landscape formation to interglacial periods. The time scale therefore would be relatively recent if creation of the observed pattern occurred in the current interglacial period (i.e. Holocene), and older time scales could reflect inherited patterns from previous inter-glacial periods, or from other processes of abiotic karstification. We determined phosphorus stocks across four landscape compartments and estimated the limestone void space (i.e., wetland depression volume) across the landscape to represent cumulative calcium export. We calculated fluxes in (e.g., atmospheric deposition) and out (i.e., solute export) of the landscape to determine landscape denudation rates through mass balance. Comparing stocks and annual fluxes yielded independent estimates of landscape age from the calcium and phosphorus budgets. Our mass balance results indicate that the landscape began to develop in the early-mid Holocene (12,000-5000 ybp). Radiocarbon dating estimates implied similar rates of dissolution (~1 m per 3000-3500 years), and were in agreement with Holocene origin. This supports the hypothesis that ecohydrologic feedbacks between hydrology and vegetation occurring during the present interglacial period are sufficient to shape this landscape into the patterns we see today, and more broadly suggests the potential importance of biota in the development of macro-scale karst features.

Sea‐level rise should cause salt‐water intrusion into coastal aquifers and limit fresh submarine groundwater discharge. Pargos Spring offshore of Puerto Morelos, Quintana Roo, Mexico, intermittently discharges brackish water and allows intrusion of lagoon water with seawater salinity to the aquifer. Lagoon water intrusion occurred when sea level was > 0.08 m above mean observed values during the study period. Salt water intrusion will be permanent within a few decades at the current eustatic sea‐level rise rate of ∼ 3 mm/yr. A mixing model demonstrates that oxygen dissolved in the lagoon water is reduced as it intrudes the spring. Dissolved oxygen (DO) reduction is greater at the spring vent than at sensors ∼ 10 m inside the conduit, reflecting rapid reaction kinetics. DO reduction results from organic carbon remineralization, which also releases N and P to the water. Increased frequency of intrusion events or continuous intrusion may alter microbially mediated biogeochemical reactions, thereby increasing aquifer vulnerability to sea‐level rise.

Riverine input of terrestrial dissolved organic matter (DOM) is an important component of the marine carbon cycle and drives net carbon dioxide production in coastal zones. DOM exports to the Arctic Ocean are likely to increase due to melting of permafrost and the Greenland Ice Sheet, but the quantity and quality of DOM exports from deglaciated watersheds in Greenland, as well as expected changes with future melting, are unknown. We compare DOM quantity and quality in Greenland over the melt seasons of 2017-2018 between two rivers directly draining the Greenland Ice Sheet (meltwater rivers) and four streams draining deglaciated catchments that are disconnected from the ice (nonglacial streams). We couple these data with discharge records to compare dissolved organic carbon (DOC) exports. DOM sources and quality differ significantly between watershed types: fluorescence characteristics and organic molar C:N ratios suggest that DOM from deglaciated watersheds is derived from terrestrial vegetation and soil organic matter, while that in glacial watersheds contains greater proportions of algal and/or freshly produced biomass and may be more reactive. DOC specific yield is similar for nonglacial streams (0.1-1.2 Mg/km 2 /year) compared to a glacial meltwater river (0.2-1.1 Mg/km 2 /year), despite orders of magnitude differences in instantaneous discharge. Upscaling based on land cover leads to an estimate of total DOC contributions from Greenland between 0.2 and 0.5 Tg/year, much of which is derived from deglaciated watersheds. These results suggest that future warming and ice retreat may increase DOC fluxes from Greenland with consequences for the Arctic carbon cycle.

No abstract