Abstract Freshwater salinization affects the availability of water for human use and ecosystem needs worldwide. It has been estimated that total dissolved solids (TDS) in the Colorado River Basin cause $350 million/year in damages and substantial resources are devoted to reducing TDS loading to streams. This study describes the development and application of coupled watershed models that enable TDS source tracking through the subsurface and across the landscape at a seasonal timestep for 35 years in the Upper Colorado River Basin. Results indicate that, on average, 75% of TDS loading to streams originates as baseflow, and 50% of loading is lagged in delivery by longer than one season. Snowmelt was identified as a dominant process controlling the transport of lagged TDS to streams. This approach informs when and where TDS mitigation efforts may be effective in a watershed that serves as a critical water supply for the southwestern United States.

Founding populations of the first Americans likely occupied parts of Beringia during the Last Glacial Maximum (LGM). The timing, pathways, and modes of their southward transit remain unknown, but blockage of the interior route by North American ice sheets between ~26 and 14 cal kyr BP (ka) favors a coastal route during this period. Using models and paleoceanographic data from the North Pacific, we identify climatically favorable intervals when humans could have plausibly traversed the Cordilleran coastal corridor during the terminal Pleistocene. Model simulations suggest that northward coastal currents strengthened during the LGM and at times of enhanced freshwater input, making southward transit by boat more difficult. Repeated Cordilleran glacial-calving events would have further challenged coastal transit on land and at sea. Following these events, ice-free coastal areas opened and seasonal sea ice was present along the Alaskan margin until at least 15 ka. Given evidence for humans south of the ice sheets by 16 ka and possibly earlier, we posit that early people may have taken advantage of winter sea ice that connected islands and coastal refugia. Marine ice-edge habitats offer a rich food supply and traversing coastal sea ice could have mitigated the difficulty of traveling southward in watercraft or on land over glaciers. We identify 24.5 to 22 ka and 16.4 to 14.8 ka as environmentally favorable time periods for coastal migration, when climate conditions provided both winter sea ice and ice-free summer conditions that facilitated year-round marine resource diversity and multiple modes of mobility along the North Pacific coast.

Sinkholes in karst and pseudokarst regions threaten infrastructure, property, and lives. We mapped closed depressions in karst and pseudokarst regions of the conterminous United States (U.S.) from 10-m-resolution elevation data using high-performance computing, and then created a heuristic additive model of sinkhole susceptibility that also included nationally consistent data for factors related to geology, soils, precipitation extremes, and development. Maps identify potential sinkhole hotspots based on current conditions and projections for 50 years into the future (the years 2070–2079) based on climate change and urban development scenarios. Areas characterized as having either high or very high sinkhole susceptibility contain 94%–99% of known or probable sinkhole locations from three U.S. state databases. States and counties with the highest amounts and percentages of land in zones of highest sinkhole susceptibility are identified. Projected changes in extreme precipitation and development did not substantially change current hotspots of highest sinkhole susceptibility. Results provide a uniform index of sinkhole potential that can support national planning, instead of existing assessments produced through various methods within individual states or smaller areas.

The Great Dismal Swamp wetland, spanning >400 km2 along the Virginia and North Carolina border, was shaped by a complex combination of geomorphic, climatic, and anthropogenic forcings during the last 14,000 years. Pollen, macrofossils, charcoal, and physical properties from sediment cores at seven sites provide a detailed record of the spatial heterogeneity of the wetland and the roles played by natural hydrologic variability, wildfire, and human modification of drainage in shaping vegetation and habitats. Cold-temperate forests occupied regional uplands from at least 13.5-10.3 cal ka BP. Marshes dominated by grasses and other herbaceous taxa began developing along low-elevation streams as early as 10.3 cal ka BP, resulting in accumulation of organic silts. Long-hydroperiod, peat accumulating marshes, with abundant floating aquatic plants, developed as early as 9.6 cal ka BP, as rapid rates of sea-level rise elevated the water table and facilitated wetland development and peat accumulation along stream courses. By the mid-Holocene (c. 7-6.5 cal ka BP), when local sea-level rise began slowing and reached about 12-15 m below present, shorter hydroperiod, peat-accumulating marshes dominated the land-scape, with increased wildfire activity. Great Dismal Swamp vegetation shifted from marshes to peat -accumulating forested wetlands by c. 3.7 cal ka BP; these were dominated by varying combinations of Nyssa (tupelo), Taxodium (cypress), and Chamaecyparis thyoides (Atlantic white cedar). Wildfires were infrequent during this time, and the forested wetlands persisted, with minor compositional changes related to climate-driven fluctuations in stream flow, until colonial ditching and logging began in the swamp during the late 18th century. These activities decreased cypress and cedar populations, and, by the mid-20th century, expanded ditching resulted in even drier conditions and expansion of maple-gum (dominated by Acer and Liquidambar), and pine-pocosin (dominated by Pinus) forests. The distribution of these forests differs from that of the late Holocene and represents a fundamental shift in hydrology, peat structure, vegetation, and fire regime due to landscape alterations of the last few centuries. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

Predictive modeling of submerged archaeological sites requires accurate sea-level predictions in order to reconstruct coastal paleogeography and associated geographic features that may have influenced the locations of occupation sites such as rivers and embayments. Earlier reconstructions of the paleogeography of parts of the western U.S. coast used an assumption of eustatic sea level, but this neglects the large spatial variations in relative sea level (RSL) associated with glacial isostatic adjustment (GIA) and tectonics. Subsequent work using a one-dimensional (1-D) solid Earth model showed that reconstructions that accounted for GIA result in significant differences from those based on eustatic sea level. However, these analyses neglected the complex three-dimensional (3-D) solid Earth structure associated with the Cascadia subduction zone that has also strongly influenced RSL along the Oregon-Washington (OR-WA) coast, requiring that the paleogeographic reconstructions must also account for this effect. Here we use RSL predictions from a 3-D solid Earth model that have been validated by RSL data to update previous paleogeographic reconstructions of the OR-WA coast for the last 12 kyr based on a 1-D solid Earth model. The large differences in the spatial variations in RSL on the OR-WA continental shelves predicted by the 3-D model relative to eustatic and 1-D models demonstrate that accurate reconstructions of coastal paleogeography for predictive modeling of submerged archaeological sites need to account for 3-D viscoelastic Earth structure in areas of complex tectonics.

Soil carbon (C) in permafrost peatlands is vulnerable to decomposition with thaw under a warming climate. The amount and form of C loss likely depends on the site hydrology following permafrost thaw, but antecedent conditions during peat accumulation are also likely important. We test the role of differing hydrologic conditions on rates of peat accumulation, permafrost formation, and response to warming at an Arctic tundra fen using a process-based model of peatland dynamics in wet and dry landscape settings that persist from peat initiation in the mid-Holocene through future simulations to 2100 CE and 2300 CE. Climate conditions for both the wet and dry landscape settings are driven by the same downscaled TraCE-21ka transient paleoclimate simulations and CCSM4 RCP8.5 climate drivers. The landscape setting controlled the rates of peat accumulation, permafrost formation and the response to climatic warming and permafrost thaw. The dry landscape scenario had high rates of initial peat accumulation (11.7 ± 3.4 mm decade−1) and rapid permafrost aggradation but similar total C stocks as the wet landscape scenario. The wet landscape scenario was more resilient to 21st century warming temperatures than the dry landscape scenario and showed 60% smaller C losses and 70% more new net peat C additions by 2100 CE. Differences in the modeled responses indicate the largest effect is related to the landscape setting and basin hydrology due to permafrost controls on decomposition, suggesting an important sensitivity to changing runoff patterns. These subtle hydrological effects will be difficult to capture at circumpolar scales but are important for the carbon balance of permafrost peatlands under future climate warming.

Anthropogenic climate change during the 21st century presents a significant challenge to the protection of cultural resources (CRs) on federal lands that encompass ∼ 28% of the U.S. In particular, CRs on this land base may be adversely affected by a wide range of climate-change hazards, including damage by sea-level rise, enhanced deterioration by increasing temperature and precipitation, and destruction by more-frequent and severe wildfire. Most current measures to manage the impacts of hazards on CRs use vulnerability assessments, but because these require that all CRs be treated as having an equal chance of being affected by climate-change hazards (i.e., equal exposure) across large landscapes, the cost and resources required for such analyses are overwhelming to land management agencies. Projections of changes in many hazards, however, show that the probability of hazard occurrence will be unevenly distributed on the landscape. Incorporating this information into a risk assessment thus allows CR managers to prioritize their efforts on assessing impacts to CRs in those areas where the probability of the hazard is greatest, thus increasing efficiency. We provide several heuristic examples of implementing the first part of a CR risk assessment by using 21st-century projections of several hazards most likely to adversely affect CRs on nine National Forests (NFs) managed by the U.S. Forest Service in northern Idaho and Montana. Overlaying the projected distribution of hazards on these NFs with the distribution of CRs identifies CR exposure that, with information on their vulnerability, is required to determine risk. Additional policy and field studies will be needed to determine how to prioritize those CRs that are most at risk according to their significance as well as identify how impacts can be reduced and managed through adaptation planning and implementation. Adaptation will follow the iterative risk management process particularly by improving projection resolution. Finer scale, process-based modeling informed by the highest priority CRs would also provide a means to assess various adaptation options that might change the estimated risk and increase the odds of CRs being as little affected as possible.

Climate warming in high‐latitude regions is thawing carbon‐rich permafrost soils, which can release carbon to the atmosphere and enhance climate warming. Using a coupled model of long‐term peatland dynamics (Holocene Peat Model, HPM‐Arctic), we quantify the potential loss of carbon with future climate warming for six sites with differing climates and permafrost histories in Northwestern Canada. We compared the net carbon balance at 2100 CE resulting from new productivity and the decomposition of active layer and newly thawed permafrost peats under RCP8.5 as a high‐end constraint. Modeled net carbon losses ranged from −3.0 kg C m −2 (net loss) to +0.1 kg C m −2 (net gain) between 2015 and 2100. Losses of newly thawed permafrost peat comprised 0.2%–25% (median: 1.6%) of “old” C loss, which were related to the residence time of peat in the active layer before being incorporated into the permafrost, peat temperature, and presence of permafrost. The largest C loss was from the permafrost‐free site, not from permafrost sites. C losses were greatest from depths of 0.2–1.0 m. New C added to the profile through net primary productivity between 2015 and 2100 offset ∼40% to >100% of old C losses across the sites. Differences between modeled active layer deepening and flooding following permafrost thaw resulted in very small differences in net C loss by 2100, illustrating the important role of present‐day conditions and permafrost aggradation history in controlling net C loss.

Baseflow is critical to sustaining streamflow in the Upper Colorado River Basin. Therefore, effective water resources management requires estimates of baseflow response to climatic changes. This study provides the first estimates of projected baseflow changes from historical (1984–2012) to thirty‐year periods centered around 2030, 2050, and 2080 under warm/wet, median, and hot/dry climatic conditions using a hybrid statistical‐deterministic baseflow model. Total baseflow supplied to the Lower Colorado River Basin may decline by up to 33%, although this value may increase in the near future by 6% under warm/wet conditions. The percentage of baseflow lost during in‐stream transport is projected to increase by 1%–5% relative to historical conditions. Results highlight that climate‐driven changes in high‐elevation hydrology have impacts on basinwide water availability. Study results have implications for human and ecological water availability in one of the most heavily managed watersheds in the world.

Paleoceanographic reconstructions from the North Pacific reveal a dynamic environment during the last deglacial transition, including rapid changes in sea surface temperature (SST) and sea-ice extent, pulses of iceberg and meltwater discharge from the Cordilleran Ice Sheet, and ocean hypoxic events. Understanding how these abrupt changes impacted marine ecosystems and shaped human migrations provides insight into past histories and possible future scenarios arising from climate-ecosystem interactions. Here we synthesize new and previously published paleoceanographic reconstructions of SST, sea-ice, paleo-salinity, ice-rafted debris (IRD), and marine hypoxia, together with model simulations of changes in ocean current strength from the last glacial maximum (LGM) to early Holocene, to assess the most viable time periods for a coastal migration from Beringia into North America. Our synthesis indicates that the period between 19-17 ka was the coldest of the last 20 ka, with a 1-2°C cooling relative to LGM conditions, extensive seasonal sea-ice, episodic reductions in sea surface salinities, and peak deposition of IRD during the ‘Siku 1’ glacial surge event, indicating cold and ice-choked conditions that would have made a coastal passage highly challenging. Model simulations indicate that the velocity of the Alaska Coastal Current (ACC) would have been strengthened during glacial conditions (relative to modern), and in response to freshwater input, making a southward transit against this cyclonic current particularly difficult for seafarers during the LGM and early deglacial period. Starting around 16.5-16.0 ka, SST began to increase in the Northeast Pacific, sea-ice retracted, and deposition of IRD began to wane, reflecting the most likely first window for human migration into North America along a coastal route. Abrupt ocean warming occurred at the onset of the Bølling-Allerød (14.7-12.9 ka), accompanied by a transition to severe hypoxia at intermediate depths, reflecting an expansion of the oxygen minimum zone (OMZ). It remains unknown whether the expanded OMZ may have negatively impacted upper ocean ecosystems, such as the kelp forest ecosystems that have been hypothesized to have supported early coastal seafarers, possibly limiting coastal habitation during this time.

Society and the environment in the arid southwestern United States depend on reliable water availability, yet current water use outpaces supply. Water demand is projected to grow in the future and climate change is expected to reduce supply. To adapt, water managers need robust estimates of future regional water supply to support management decisions. To address this need, we estimate future streamflow in seven water resource regions in the southwestern U.S. using a new SPAtially Referenced Regressions On Watershed attributes (SPARROW) streamflow model. We present streamflow projections corresponding to input data from seven climate models and two greenhouse gas Representative Concentration Pathways (RCP4.5 and 8.5) for three, thirty-year intervals centered on the 2030s, 2050s, and 2080s, and for a historical thirty year interval centered on the 1990s. Across water resource regions, about half of the RCP4.5 models (51%) and two thirds of the RCP8.5 models (67%) indicate decreases in streamflow in the 2080s relative to the historical period. Models project maximum decreases in streamflow of 36–80% in all water resource regions for all periods and RCPs relative to historical streamflow, and maximum streamflow decreases of up to 20–45% in the 2080s at sites along the Colorado River used for measuring compliance with interstate and international water agreements. Headwaters are projected to experience the greatest declines, with substantial downstream implications. Among these estimates, the streamflows from models forced with RCP8.5 tend to be lower than those forced with RCP4.5. Not all climate models, times, and RCPs project widespread streamflow declines. The most ubiquitous streamflow increases are projected to occur in the 2030s under RCP4.5. Later time periods and enhanced greenhouse gas forcings indicate smaller regions of streamflow increase and lower accumulated streamflows, suggesting that limiting or reducing greenhouse gas concentrations could support future water availability. Although some possible streamflow increases are promising, the modest and spatially limited increases in streamflow projected for later time periods are still unlikely to be sufficient to meet the projected water demand. These results inform the likelihood of future water agreement compliance, and support developing strategies to balance water supply and demand.

New radiocarbon and sedimentological results from the Gulf of Alaska document recurrent millennial-scale episodes of reorganized Pacific Ocean ventilation synchronous with rapid Cordilleran Ice Sheet discharge, indicating close coupling of ice-ocean dynamics spanning the past 42,000 years. Ventilation of the intermediate-depth North Pacific tracks strength of the Asian monsoon, supporting a role for moisture and heat transport from low latitudes in North Pacific paleoclimate. Changes in carbon-14 age of intermediate waters are in phase with peaks in Cordilleran ice-rafted debris delivery, and both consistently precede ice discharge events from the Laurentide Ice Sheet, known as Heinrich events. This timing precludes an Atlantic trigger for Cordilleran Ice Sheet retreat and instead implicates the Pacific as an early part of a cascade of dynamic climate events with global impact.

The goal of the U.S. Climate Resilience Toolkit’s (CRT) Climate Explorer (CE) is to provide information at appropriate spatial and temporal scales to help practitioners gain insights into the risks posed by climate change. Ultimately, these insights can lead to groups of local stakeholders taking action to build their resilience to a changing climate. Using CE, decision-makers can visualize decade-by-decade changes in climate conditions in their county and the magnitude of changes projected for the end of this century under two plausible emissions pathways. They can also check how projected changes relate to user-defined thresholds that represent points at which valued assets may become stressed, damaged, or destroyed. By providing easy access to authoritative information in an elegant interface, the Climate Explorer can help communities recognize—and prepare to avoid or respond to—emerging climate hazards. Another important step in the evolution of CE builds on the purposeful alignment of the CRT with the U.S. Global Change Research Program’s (USGCRP) National Climate Assessment (NCA). By closely linking these two authoritative resources, we envision that users can easily transition from static maps and graphs within NCA reports to dynamic, interactive versions of the same data within CE and other resources within the CRT, which they can explore at higher spatial scales or customize for their own purposes. The provision of consistent climate data and information—a result of collaboration among USGCRP’s federal agencies—will assist decision-making by other governmental entities, nongovernmental organizations, businesses, and individuals.

We assess monthly temperature and precipitation data derived by six statistically downscaled data sets for 14 general circulation models (GCMs) from the Climate Model Intercomparison Program Phase 5. We use a simple monthly water balance model to quantify and decompose uncertainties associated with the GCMs and statistical techniques in projections for four snow‐dominated regions in the western United States. The end‐of‐century projections from GCMs exhibit substantial variation in change over the regions (temperature range of 2.8–8.0 °C and precipitation range of −22–31%). The six downscaled data sets exhibit disparate high‐resolution representations of the magnitude and spatial patterns of future temperature (up to 2.2 °C) and precipitation (up to 30%) for a common GCM. Two data sets derived by the same downscaling method (Multivariate Adaptive Constructed Analogs) produce median losses of snow water equivalent over the Upper Colorado of 51% and 81%. The principal causes of the differences among the downscaled projections are related to the gridded observations used to bias correct the historical GCM output. Specifically, (1) whether a fixed atmospheric lapse rate (−6.5 °C/km) or a spatially and temporally varying lapse rate is used to extrapolate lower elevation observations to high‐elevations and (2) whether high‐elevation station data (e.g., SNOTEL) are included in the observations. The GCM projections are the largest source of uncertainty in the monthly water balance model simulations; however, the differences among seasonal projections produced by downscaled data sets in some regions highlight the need for careful evaluation of the statistically downscaled data in climate impact studies.

We apply the Community Ice Sheet Model (CISM2) to determine the extent to which the Last Glacial Maximum (LGM) temperature and precipitation climatologies from the Paleoclimate Modelling Intercomparison Project 3 (PMIP3) simulations support the large North American ice sheets that were prescribed as a boundary condition. We force CISM2 with eight PMIP3 general circulation models (GCMs), and an additional model, GENMOM. Seven GCMs simulate LGM climatologies that support positive surface mass balances of the Laurentide and Cordilleran ice sheets (LIS, CIS) consistent with where ice was prescribed in the GCMs. Two GCMs simulate July temperatures that are too warm to support the ice sheets. Four of the nine CISM2 simulations support ice sheets in Beringia, in absence of prescribed ice in the driving GCMs and in disagreement with geologic evidence that indicates the area remained ice-free during the LGM. We test the sensitivity of our results to a range of snow and ice positive degree-day factors, daily, monthly, and climatological temperature and precipitation inputs, and we evaluate the role of albedo and snow in the simulations. Areas with perennial snow in the GCM simulations correspond well to the presence of ice in the CISM2 simulation. GCMs with unrealistically low surface albedos over the LIS yield simulations that fail to simulate realistic ice sheets.