Source: Journal of Geophysical Research: Earth Surface
Sediment transport shapes the Earth surface in different ways, by forming desert dunes and by sculpting the topography of rivers, but the physics of sediment transport initiation is still incompletely understood. For decades, models have generally assumed two basic entrainment mechanisms: a grain resting on the sediment bed is either lifted directly by fluid forces, or it is emitted from the soil indirectly, as product of a granular splash caused by the heavy impact of another grain.
However, recent breakthroughs in grain-based simulations and high-speed visualization have been offering a much clearer look at the processes that trigger grain motion. Insights from these recent advances have revealed a rather broad spectrum of indirect particle-particle and particle-fluid interactions driving entrainment, including the rearrangement of surface grains after splash and changes in near‐bed flow structure due to moving grains. These interactions exert non-local influences on transport thresholds, giving rise to a dynamic process known as collective particle entrainment—a mechanism that remains poorly understood at a fundamental level.
In a new study, Chartrand [2026] shows that collective particle entrainment is size-dependent: large grains interact primarily with their peers, while smaller grains are mobilized by both large and similar-sized particles. This distinction leads to divergent transport signatures, with a new stochastic model predicting temporally correlated motion for small grains and uncorrelated, white-noise entrainment statistics for larger particles.
Although theoretical modeling will be required to shed further light on the physics of collective entrainment, the author’s study is a step toward a quantitative model of sediment transport from a probabilistic perspective. Looking ahead, Chartrand’s ideas could now be extended to other environments, potentially transforming our understanding of entrainment in other contexts such as wind-blown transport and extraterrestrial atmospheric processes.
Citation: Chartrand, S. M. (2026). Collective particle entrainment explored with experimental data and coupled transfer functions. Journal of Geophysical Research: Earth Surface, 131, e2025JF008657. https://doi.org/10.1029/2025JF008657
The evolution of rivers that split into multiple channels is a scientific challenge in terms of modeling and prediction. On the other hand, these rivers are widespread and play a key role for ecosystems’ life, groundwater recharge, and therefore, water security. They are also extremely sensitive to hydroclimatic changes, leading to shifts in precipitation, erosion and sediment transport.
Zhao et al. [2026] investigate the drivers of river evolution for 97 multithread river reaches worldwide, spanning diverse climates and morphologies. The study reveals the key role of intermittency for river evolution. In particular, higher flow intermittency could lead to more even flow partitioning among threads, therefore impacting hydrology and ecosystems. With flow variability increasing after climate change, rivers are likely to increase their thread count, thus impacting livelihoods and ecosystems.
Two example multithread reaches shown in Landsat images from (b) the Irtysh River (wandering) and (c) the Yukon River (braided). Credit: Zhao et al. [2026], Figure 1(b,c)
Citation: Zhao, F., Ganti, V., Chadwick, A., Greenberg, E., McLeod, J., Liu, Y., et al. (2026). Global hydroclimatic controls on multithread River dynamics. AGU Advances, 7, e2025AV002166. https://doi.org/10.1029/2025AV002166
NEW YORK, June 4 — Wildlife experts found eight new species of dragonfly, three unknown grasshoppers and some 60 new butterflies and moths in vivid hues during a trip to Angola's Lisima plateau in February, a conservation group said yesterday.
The Wilderness Project visited the waters that flow through the plateau and which feed four of Africa's major rivers: the Congo, Okavango, Zambezi and Cuanza.
New species included an armoured, predatory cricket, a previously undescribed species of copper caterpillar and its adult butterfly, and a crowned crab spider that fluoresces under ultraviolet light.
Experts also found a new blood orange-hued species of ladybird orb-web spider which mimics ladybirds in signaling to predators with a bright colour - normally a darker red - that it is too bitter or toxic.
"The armoured crickets are very cool ... very fierce-looking," expedition leader Rob Taylor told Reuters. "As a defense mechanism, they can actually squirt fluid onto whoever's trying to attack them." Scientists the world over are frantically trying to record species as they reckon with a global ecological crisis that has put a million plant and animal species on the brink of extinction. They estimate there are 8.7 million species in the world, of which science has identified only 1.5 million.
Many are fast disappearing because of human activity, with more than 800 animal species going extinct since around 1500.
Taylor said wildlife in the Lisima plateau was threatened by "tree-felling, deforestation and ... the artisanal diamond mining industry," as well as by slash-and-burn agriculture, which razes natural forests to use the soil for planting, only to see the nutrients wash away. — Reuters
“Future winters promise less snow, more rain. Nobody’s prepared.” So proclaims the title of a recent article in the Proceedings of the National Academy of Sciences of the United States of America that frames adaptation to snow loss as the “million-dollar question” facing the western United States.
As the largest sectoral consumer of fresh water globally, agriculture is particularly vulnerable to snow loss.
Declining snowfall—and snowmelt—affects ecosystems, urban and rural water supplies, hydropower, recreation, tourism, and agriculture. As the largest sectoral consumer of fresh water globally, agriculture is particularly vulnerable to snow loss.
In response, water managers have developed a range of approaches for adapting to snow loss: infrastructure-based approaches like managed aquifer recharge, nature-based solutions such as forest management and beaver dam analogues, demand-side approaches like multibenefit land repurposing, and polarizing supply-side approaches like reservoir expansion and cloud seeding (Figure 1).
Fig. 1. Potential approaches to reduce negative impacts to agriculture from snow loss include a variety of adaptive strategies that address either water supply or demand. Click image for larger version.
However, efforts to identify which of these strategies to implement for different drainage basins, or watersheds, using the variety of available approaches seem to fall into one of two traps: either searching for unrealistic one-size-fits-all panaceas [Ostrom, 2007] or treating every basin as unique, which is costly and inefficient.
The “trillion-dollar question” isn’t how to adapt but, rather, where existing strategies may make the most—and fastest—difference.
Importantly, continuing along this trajectory means that we’re on track to offset only about a third of global climate-induced crop yield losses by 2100. For the western United States, previous work has estimated cumulative economic losses from declining snowfall of hundreds of billions to trillions of dollars while noting that rational adaptation decisions are hampered by the lack of financial analyses of the importance of snow [Sturm et al., 2017].
We thus suggest that the “trillion-dollar question” isn’t how to adapt but, rather, where existing strategies may make the most—and fastest—difference to offset projected losses. Answering this question requires an approach that matches strategies to the contexts where they are more likely to succeed—one that treats basins as neither uniform nor unique.
A Mismatch in Research and Operational Scales
Physical scientists tend to look at snow loss as a basin-scale problem, in part because this view aligns with hydrologic boundaries. However, as our colleague, applied economist Joey Blumberg, explains, “county lines were not drawn to follow watersheds, and rivers do not conform to political borders, creating a patchwork of mismatched boundaries.”
Scientists have long emphasized that mitigating climate change requires us to “think globally, assess regionally, act locally.” And in 1992, the authors of the Dublin Principles reasoned that moving the needle on “wicked water problems” requires targeting decisionmaking at the “lowest appropriate level,” where stakeholders can collaborate most effectively.
Lake Tahoe, pictured here, contains 37 trillion gallons of water, roughly half of which is supplied by snowmelt in the Sierra Nevada Mountains. Credit: Beatrice L. Gordon
Working at this scale, we found one-size-fits-all strategies often don’t hold up, even within the same hydrologic basin [Gordon et al., 2024; Boisramé et al., 2026]. In the Upper Colorado River Basin, for example, expanding reservoir storage could buffer agriculture in northeastern Utah against declining snowpack, but the same strategy may fail miles away in southwestern Wyoming, where a thirstier atmosphere may make it harder to refill existing reservoirs.
However, collecting detailed local-scale information for just 13 of the roughly 2,600 operational contexts nationwide took almost 3 years of searching websites, reading working papers, and calling water managers.
Scaling this approach across the entire western United States is understandably overwhelming. We need a more systematic approach to help managers identify which strategies could work most effectively, and where.
A Diagnostic for Agriculture and Snow Loss
Ostrom [2007] argued that complex systems, such as Western agriculture, “are partially decomposable in their structure.” This insight is woven into archetype analysis, an approach for identifying recurrent patterns across otherwise diverse systems.
Like workplace assessments—which are genuinely useful, albeit imperfect, tools for understanding successful management styles—archetypes draw on qualitative, quantitative, or hybrid approaches to group diverse operational contexts on the basis of shared characteristics [Sietz et al., 2019]. These groupings enable systematic knowledge transfer about, for example, how management strategies that work in one context can also guide adaptation elsewhere.
Three main characteristics interact to define operational contexts in snow-dependent agriculture in the western United States: physical constraints, governance systems, and human behavior.
“Researchers can empirically derive building blocks or components that comprise archetypes to represent key features of a system,” explains Elizabeth Koebele, who studies urban water sustainability [Garcia et al., 2019] and has begun applying archetypes in that context. However, she notes, these building blocks “vary based on the system context, available data, and study goal.”
We propose three main characteristics that interact to define operational contexts in snow-dependent agriculture in the western United States: physical constraints, governance systems, and human behavior. Physical constraints, including biophysical setting, infrastructure, and climate, determine available water supplies. Governance capacity relative to governance complexity shapes how those supplies are allocated across competing uses. Human behaviors influence both water demand and how users respond to supply conditions and governance rules.
Using these characteristics to establish archetypes of water management contexts could define a path forward for operationalizing an approach to accelerate successful adaptations to declining snowpacks in the West.
Constraining How Snowmelt Becomes Water Supply
Physical constraints stem from biophysical processes that influence how, when, and how much snow becomes streamflow; infrastructure that stores and conveys water; and hydrologic and climatic uncertainties about future supplies. These constraints can vary substantially from basin to basin.
Consider the Walker River Basin and California’s San Joaquin Valley, both of which rely on Sierra Nevada snowpack but have different biophysical settings. In some parts of the central Sierra, forest management can reduce wildfire risk and increase streamflow by up to 14% during low-snow years. Elsewhere, however, water made available by forest management may be consumed by remaining vegetation, limiting downstream gains.
These biophysical differences interact with uses of built infrastructure, including irrigation systems, reservoir outlets, and canals, to determine how and when water is stored and released. As temperatures warm and snowmelt declines, officials in both the Walker River and San Joaquin Valley basins must increasingly manage for a wider range of extremes, including “cold-water droughts.” However, the infrastructure to manage these trade-offs through reservoir storage and operations that balance agricultural deliveries with aquatic habitat needs is more developed in the highly managed San Joaquin than in the Walker.
Thankfully, measuring physical constraints on snowmelt at basin scales is becoming more feasible today with newly developed tools.
Layered on top of biophysical and infrastructural constraints are climatic and hydrological uncertainties, such as whether snow loss will lead to more evapotranspiration and less streamflow. These uncertainties complicate management decisions based on cost-benefit modeling of individual strategies: Should districts expand reservoir storage if precipitation is predicted to increase or decrease depending on the model? Frameworks like Decision Making Under Deep Uncertainty emphasize the need to select strategies that are robust across many possible futures.
Thankfully, measuring physical constraints on snowmelt at basin scales—a means, along with improved modeling, to reduce hydroclimatic uncertainties—is becoming more feasible today with newly developed tools. Water managers can turn, for example, to databases like the U.S. Geological Survey’s ResOpsUS [Steyaert et al., 2022], which catalogs historical reservoir operations across the contiguous United States, and to publicly available hydrologic projections such as those from Oak Ridge National Laboratory’s Coupled Models Intercomparison Project phase 6 (CMIP6) ensemble.
Governance Controls Supply Allocations
We frame governance around capacity and complexity. Capacity in this context is the ability of stakeholders “to mobilize resources in order to make equitable and fair decisions around shared challenges,” according to governance scholar Gina Gilson. Complexity refers to the number and intricacy of jurisdictions, authorities, regulations, and stakeholders involved. As governance complexity increases, the effectiveness of adaptation strategies becomes more sensitive to capacity constraints, particularly regarding timescales and funding.
For example, infrastructure in the Walker is controlled locally by a single water district, and jurisdictional coordination involves two states and the Walker River Paiute Tribe. Coordination on water management is never simple, but fewer jurisdictions generally means faster decisionmaking and clearer authority, allowing the single water district to implement strategies like multibenefit land repurposing more readily. Such implementations, in turn, enable reduced agricultural water use, directly supporting restoration of Walker Lake and recovery of endangered species.
The San Joaquin Valley is vastly different in scale and complexity, covering eight California counties, one of which alone has 22 water districts and seven cities. Following the passage of the state’s Sustainable Groundwater Management Act, water users in the basin formed more than 120 groundwater sustainability agencies. Agricultural water management thus involves overlapping federal and state systems that operate under different rules, contracts, and regulatory requirements. Whileland repurposing programs can be implemented, more substantial capacity, time, and resources are typically needed to do so.
Emerging efforts like the Western States Water Data Access and Analysis Tool (WestDAAT) and the Harmonized Database of Western U.S. Water Rights make it easier to assess governance in a basin by standardizing data about rules, regulations, and water rights across states. Combined with mapping of irrigation service areas and water transfers [Siddik et al., 2023], these resources help stakeholders identify the jurisdictions involved, how authority is distributed, and what coordination mechanisms exist for agricultural water management.
Human Behavior Shapes Demand Responses
Once snowmelt reaches water users, behavioral dynamics—how people respond to crises, policies, and changing conditions—determine how effectively management strategies achieve desired results.
Water demand is influenced by consumption choices and by economic, political, and cultural factors.
Water demand is influenced by consumption choices and by economic, political, and cultural factors. It is also influenced by factors that typical hydrologic models rarely account for, including social structure, social memory, and affluence. More affluent users are less likely to modify their behavior to reduce water use under conditions of scarcity.
The dynamics of water demand in the South Platte River Basin, for example, are especially complex, as they are balanced across cities, agriculture, and ecosystems across parts of Colorado, Nebraska, and Wyoming. Water prices in the basin’s Big Thompson project, a federal water diversion system in northern Colorado, jumped from $1,500 per acre-foot in 1990 to more than $30,000 in 2018, driven by economic factors that resulted in cities owning 70% of water originally intended for agriculture.
Even with reliable projections of future climate and water supply, carefully planned strategies can be overwhelmed by economic and behavioral factors, resulting in transfers and reallocations of water. What’s more, behavioral responses to adaptation strategies can paradoxically increase demand when users perceive that scarcity problems are solved.
The “reservoir effect” occurs when water security perceptions encourage expansion of water-intensive activities [Di Baldassarre et al., 2018]. Similarly, the irrigation efficiency paradox shows how efficiency gains can lead to expanded production and reduced return flows (how much irrigation water returns to streams and aquifers) downstream [Grafton et al., 2018].
Conceptual frameworks, models, and global case studies have all been used as approaches to study the effects of human behavior on hydrology. With sufficient training data, we believe tools like machine learning could be used to further explore how behaviors influence adaptation and to anticipate shifts as snow loss continues.
Archetypes in Practice
By evaluating how physical factors, governance systems, and human behavior shape outcomes across places like the Walker, South Platte, and San Joaquin basins, researchers and practitioners can establish archetypes to help identify patterns in what strategies are most effective in different places and assess how to transfer lessons from one setting to another (Figure 2).
Fig. 2. An archetype-based diagnostic grounded in evaluating the physical constraints, governance, and human behavioral dynamics affecting hydrologic basins could facilitate more rapid transfer of learning about successful adaptation approaches across snowmelt-dependent agriculture in the western United States.
The Walker River Basin exemplifies an archetype common to agriculturally dominated headwaters in the western United States with low governance complexity (few jurisdictions), adequate capacity (resources), low behavioral complexity (more predictable and unified user groups), and substantial physical constraints (significant future snow loss and limited infrastructure for water storage and supplementation).
With this profile, the Walker is an ideal testing ground for evaluating how effectively different strategies offset changes in snowmelt. Does cloud seeding increase snowpack? Could beaver dam analogues—a nature-based solution reminiscent of Idaho Fish and Game’s mid-20th century effort to parachute beavers into the wilderness—meaningfully increase water retention? Could multibenefit land repurposing buffer people and ecosystems against supply volatility while restoring ecosystem functionality?
The value of organizing operational contexts by archetypes is that each context need not be treated as unique.
The value of organizing operational contexts by archetypes is that each context need not be treated as unique. Lessons learned from the Walker could be systematically transferred to other areas with similar characteristics and could be incrementally tested in others.
The South Platte has physical constraints similar to Walker’s but features greater governance complexity because of multiple interstate compacts, as well as greater behavioral complexity. Modeling analyses indicate that demand-side strategies could adapt to more volatile water supply in the South Platte [Gharib et al., 2023]. But implementing them requires balancing perspectives from both agricultural and urban water users—a behavioral dynamic absent in Walker.
Crop switching to cultivate higher-value crops on less acreage could reduce water use. However, options for what crops can be grown where are constrained by factors like elevation and climate. Even where feasible, new crops would require investments in education, new infrastructure, risk management, and agronomic knowledge.
Through iterative expansion and testing, broad archetypes like “high behavioral complexity” could be specified to reflect dynamics like rural-urban competition or concerns around buy-and-dry economics. Archetypes may also point to contexts where governance complexity signals that decisionmaking is occurring above the lowest appropriate level.
The San Joaquin, with its extremely complex governance involving numerous local, state, and federal agencies managing surface and groundwater, is one potential example. Recognizing this pattern can help identify where substantial resources and long timelines may be required to implement programs (e.g., LandFlex) requiring legislative authorization, multiagency coordination, and stakeholder engagement. It may also signal the need to identify smaller operational contexts within larger settings so implementations proceed more rapidly.
Operationalizing Archetypes from Diagnosis to Action
Developing a systematic approach to match adaptation strategies with areas where they are most likely to succeed in operation is only a first step. Applying diagnostics without mechanisms to implement new strategies is often insufficient to drive timely action.
An instructive precedent of success in water quality management comes from the 1970s. By then, pollution controls on factories had improved compared with the early 20th century, yet water quality in surface waters across the country still declined because of pollution in agricultural runoff. The breakthrough came with the EPA’s total maximum daily load (TMDL) program, which created a structured process that set measurable goals for reducing pollution and assigned responsibility for meeting those goals to the sources of the pollution, allowing for local control over adaptation.
Archetypes could play a similar role in facilitating beneficial snow-loss adaptations, and a structure like the TMDL program could start by assessing supply-demand risks across operational areas, setting performance targets such as reservoir reliability and shortage frequency, and then using the diagnostic to identify which strategies fit each archetype. Results and lessons could be shared region-wide, while implementation would remain locally driven.
This suggestion is, emphatically, not a prescription for specific policy mechanisms. But it serves as a reminder that—just as few of us engage with workplace assessments or change behavior on the basis of their results without organizational support—archetypes will need to be paired with implementation structures to translate diagnosis into action.
Beyond Silver Bullets
There is no single answer to our trillion-dollar question, but one path forward for sustaining complex Western ecosystems lies in developing archetypes of different types of basins.
Nearly 20 years ago, Ostrom [2007] warned against seeking panaceas for complex environmental problems. There is no silver bullet for snow loss or single answer to our trillion-dollar question, but one path forward for sustaining complex Western ecosystems lies in developing archetypes of different types of basins.
A small irrigation district, for example, wouldn’t need to independently test every strategy in Figure 1 or develop complex decision support tools when a similar archetype already evaluated which strategies work under comparable governance, behavioral, and physical conditions.
Critically, these archetypes can be developed and refined by managers and scientists to capture more nuanced realities. Physically constrained systems, for example, could include areas facing high future uncertainty or limited reservoir flexibility. Governance and behavioral dimensions could likewise evolve to represent contexts where subsidies lead to incoherent incentives or where cultural norms link water use to local identities and traditions.
Like workplace assessments, the goal isn’t to diminish unique personalities but to work with them more strategically. Archetypes can show where we don’t need to reinvent the wheel to adapt and where the wheel might need to be tweaked. By leveraging collective knowledge and learning across regions facing similar challenges, rather than crafting new solutions basin by basin, we can reduce the time and resources needed to implement equitable and sustainable adaptation solutions.
Acknowledgments
This work is supported by the National Science Foundation (NSF) under grants 1828902 and OIA-2148788. Where We Live is funded by a grant from NSF’s Established Program to Stimulate Competitive Research (EPSCoR) RII Track-2 program and features partnerships across the University of Idaho (award 2316126); the University of Nevada, Reno (award 2316127); and the University of South Carolina (award 2316128). Work was also supported by internal funds from the Division of Hydrologic Resources at the Desert Research Institute.
References
Boisramé, G. F., et al. (2026), Think globally, model locally: Complex responses of agricultural water supplies to different climate projections, J. Am. Water Resour. Assoc., 62(3), e70117, https://doi.org/10.1111/1752-1688.70117.
Garcia, M., et al. (2019), Towards urban water sustainability: Analyzing management transitions in Miami, Las Vegas, and Los Angeles, Global Environ. Change, 58, 101967, https://doi.org/10.1016/j.gloenvcha.2019.101967.
Gharib, A. A., et al. (2023), Assessment of vulnerability to water shortage in semi-arid river basins: The value of demand reduction and storage capacity, Sci. Total Environ., 871, 161964, https://doi.org/10.1016/j.scitotenv.2023.161964.
Gordon, B. L., et al. (2024), The essential role of local context in shaping risk and risk reduction strategies for snowmelt‐dependent irrigated agriculture, Earth’s Future, 12(6), e2024EF004577, https://doi.org/10.1029/2024EF004577.
Ostrom, E. (2007), A diagnostic approach for going beyond panaceas, Proc. Natl. Acad. Sci. U. S. A., 104(39), 15,181–15,187, https://doi.org/10.1073/pnas.0702288104.
Sietz, D., et al. (2019), Archetype analysis in sustainability research: Methodological portfolio and analytical frontiers, Ecol. Soc., 24(3), 34, www.jstor.org/stable/26796999.
Steyaert, J. C., et al. (2022), ResOpsUS, a dataset of historical reservoir operations in the contiguous United States, Sci. Data, 9, 34, https://doi.org/10.1038/s41597-022-01134-7.
Sturm, M., et al. (2017), Water and life from snow: A trillion dollar science question, Water Resour. Res., 53(5), 3,534–3,544, https://doi.org/10.1002/2017WR020840.
Author Information
Beatrice L. Gordon (beatrice.gordon@dri.edu), Gabrielle F. S. Boisrame, Christine M. Albano, and Rosemary W. H. Carroll, Desert Research Institute, Reno, Nev.; and Adrian A. Harpold, University of Nevada, Reno
Citation: Gordon, B. L., G. F. S. Boisrame, C. M. Albano, R. W. H. Carroll, and A. A. Harpold (2026), Archetypes could accelerate agricultural adaptation to less snowpack, Eos, 107, https://doi.org/10.1029/2026EO260184. Published on 9 June 2026.
This article does not represent the opinion of AGU, Eos, or any of its affiliates. It is solely the opinion of the author(s).
SINGAPORE: Singapore is prepared to pay more to attract new local bus captains, but the harder task is convincing them to stay.
From next year, new Singaporean and permanent resident bus captains will receive a S$450 monthly starting salary increase, along with a higher sign-on bonus of S$2,000. The changes could lift first-year earnings by around S$600 a month, pushing average monthly pay beyond S$4,000 when overtime, allowances and bonuses are included.
The move comes as the public bus sector grapples with a shrinking share of local drivers. The proportion of Singaporean and permanent resident bus captains fell from 54 per cent in 2021 to 41 per cent in 2025, according to Channel NewsAsia (CNA )’s June 5 report. For many existing drivers, however, salary has never been the only issue.
The job starts at 3 am, long before sunrise
Several bus captains said that while better pay would attract newcomers, the realities of the job catch people off guard.
Bus drivers may begin work as early as 3 am to prepare for the first buses leaving depots before dawn. Working hours can be irregular, meal times unpredictable and shifts physically draining.
One common complaint is the split-shift arrangement. Drivers may work the morning rush, take an unpaid break lasting several hours, then return for the evening peak period.
Former public bus captain Muhammad Naz Farihin said these long breaks can make an entire day feel like it’s consumed by work, even though part of it is unpaid downtime. Many newcomer drivers also said they leave within their first year after discovering the demands involved.
Bus drivers don’t just drive a bus
The public usually sees bus captains as people who move passengers from one stop to another. Drivers say the role involves far more.
Besides operating large vehicles safely, bus captains handle customer service issues, manage emergencies and keep services running on tight schedules. Some are trained in cardiopulmonary resuscitation (CPR) and may be among the first to respond during medical incidents.
One bus captain said he hopes the higher salaries will help raise public appreciation of the profession. He argued that the job requires a set of skills that many commuters may not fully notice.
Bus operators are promising changes in bus driving schedules
Associate Professor Walter Theseira from the Singapore University of Social Sciences said bus driving is a specialised role that demands discipline and reliability. Unlike gig work, drivers cannot simply decide not to show up. A missing bus captain can disrupt an entire service.
He also said salaries needed to be competitive enough to attract people who have other job options, including mid-career workers who may already earn higher wages elsewhere.
Bus operators are also promising changes beyond pay. Measures under consideration include reducing split shifts, shortening continuous driving periods and offering better career progression opportunities.
Higher salaries may bring more people through the door. Retaining skilled drivers will likely depend on operators’ ability to make the work more sustainable over the long term.
Buses don’t run on engines alone. They run on people willing to show up before sunrise, navigate traffic safely and carry thousands of commuters to their destinations every day.
From rivers and oxbow lakes to crop-field patchworks and mineral sediments, Landsat has seen it all. A program of NASA and USGS, the satellite initiative has documented the Earth’s surface since 1972, making it the longest continuous record of our planet’s ever-evolving landscapes. And to mark Earth Day 2026, the organizations launched a playful way to interact with some of their findings collected over the past five-and-a-half decades—a name generator.
Using the tool is simple: type in your name, or any word, and Landsat returns it in the form of vertical snapshots of a wide range of terrain. Just like we see with composites of Mars, for instance, scientists have digitally enhanced some images to highlight specific features. Those used for “Your Name in Landsat” sport a wide array of hues, textures, and patterns that glimpse the diversity of our planet’s surface.
“Colossal”
Landsat is an incredible resource that features time-lapses of changing land use over several decades. Even this playful name generator allows you to hover over individual images and learn the exact locations—down to the coordinates—and all of the program’s data is publicly accessible. For example, the “C” in “Colossal” above is a vertical view of a cloud-speckled Deception Island in Antarctica, and the “A” is the uniquely shaped Lake Mjøsa in Norway.
You might also enjoy Overview, a book that chronicles how the landscape has changed over time. Learn more about Landsat from NASA. (via PetaPixel)
Atmospheric rivers act like “rivers in the sky,” shuttling intense bands of warm, heavy moisture from lower to higher latitudes. When an atmospheric river encounters cold air or mountainous terrain, the moisture it carries condenses and falls as heavy rain or snow. In Antarctica, the arrival of an atmospheric river can help build surface ice mass. Much of Antarctica is very dry; an atmospheric river can bring the moisture needed to potentially offset some ice loss.
Antarctica’s varied topography and dry conditions have made detecting atmospheric rivers over the continent challenging. Previous efforts to do so have suggested that atmospheric rivers contribute up to 30% of Antarctica’s total annual precipitation, but these methods may not be capturing the full picture of atmospheric river activity.
Takahashi et al. developed a new 3D atmospheric river detection algorithm to better capture how atmospheric rivers affect Antarctica’s complex terrain. Previous methods have mostly been 2D, meaning they do not accurately account for the vertical variations within an atmospheric river.
To evaluate the algorithm, the researchers applied it to two datasets: (1) daily snowfall totals measured during the 44th Japanese Antarctic Research Expedition (JARE44) at Dome Fuji from February 2003 to January 2004 and (2) the ERA5 (European Centre for Medium-Range Weather Forecasts atmospheric reanalysis) dataset of daily weather patterns and conditions in Antarctica from 1979 to 2023.
The results of the study’s new algorithm showed 16 significant snowfall events during the JARE44 expedition, all of which were not detected by the older 2D method. The new 3D method identified 17 days of atmospheric river activity, which corresponded with 10 heavy snowfall events and accounted for approximately 40% of the total precipitation. Between 1979 and 2023, atmospheric rivers occurred about 10% of the time yet contributed 30%–60% of total precipitation in the Antarctic interior.
The 3D method in the new study suggests that atmospheric river events contribute a greater proportion of total snowfall than previously thought—between 30% and 90%, depending on the Antarctic region. The researchers also suggest that long-term changes in Antarctic snowfall are closely linked with the changes in atmospheric river activity. This connection is especially apparent in East Antarctica, where the link between snowfall increases and atmospheric rivers had not yet been clearly identified in previous studies. (Geophysical Research Letters, https://doi.org/10.1029/2025GL120986, 2026)
Real-time hydrologic forecasting predicts river level and flooding inundation by combining continuously updated rainfall measurements, river gauge readings, and weather forecasts. Most of these flood forecasting systems depend on human interpretation and adjustments, or a “forecasters-in-the-loop” approach, which pairs computer models with a human expert on flood dynamics and local conditions. In contrast, in a “forecasters-over-the-loop” system, humans supervise automated forecasts and intervene only if necessary.
Recently, artificial intelligence (AI) and machine learning (ML) have become more integrated into flood prediction, and many of these systems are faster at processing large datasets and learning complex patterns from historical records than traditional models alone. But these new technologies also come with limitations—AI and ML require extensive data and may struggle to capture extreme, rare events.
Even though ML and AI are often touted as the future of flood forecasting, most studies have tested this technology against models that provide historical simulations, not the real-time operational systems that would be used during a flood. These simplified models may lack local details or are tested at daily rather than hourly resolution. Their effectiveness may be overestimated.
Tran et al. produced the first study comparing the performance of ML models to an actual flood forecasting system used at the California Nevada River Forecast Center (CNRFC) that uses professional forecasters and traditional hydrologic models. The study suggests that a forecasters-in-the-loop approach outperforms the ML models in several key ways, including streamflow predictions and flood event detection, because forecasters can recognize model errors and account for poor input data—actions models cannot take on their own.
The researchers used data gathered from CNRFC river stage forecasts across 50 California and Nevada locations between 2012 and 2022 and river condition lead times from 1 to 96 hours. Compared to the ML models, the Community Hydrologic Prediction System used at CNRFC generally performed better at predicting stream flow and flood peaks, especially with longer lead times. Though the ML models could perform better at very short lead times, their accuracy declined quickly. Though automated forecasting options may seem promising, they are not yet a suitable replacement for human expertise when it comes to protecting lives and livelihoods from damaging floods, the researchers say. (Geophysical Research Letters,https://doi.org/10.1029/2025GL118317, 2026)
Citation: Owen. R. (2026), Keeping humans in the loop improves flood forecasting, Eos, 107, https://doi.org/10.1029/2026EO260161. Published on 19 May 2026.
Login
