Papers for

water resource managers

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

RiverScan improves runoff prediction using river network structure

RiverScan: Learning to Accumulate on River Networks

Abstract: Accurate runoff prediction across large river basins requires models to capture heterogeneous local hydrological responses and their accumulation downstream. River-network connectivity provides a natural structure for this task, but using it effectively requires learning how to combine contributions from subcatchments of different sizes. We frame this challenge as learning to accumulate and introduce RiverScan, a spatial operator based on area-weighted selective state-space recurrence. RiverScan recursively combines local temporal representations with accumulated upstream states in topological order. Contributing-area fractions account for subcatchment size, while learned retention controls how upstream information persists downstream. Experiments on 608 stations in the Danube basin show that RiverScan outperforms the evaluated temporal and graph baselines in overall predictive performance. Structural controls and ablations demonstrate that composing heterogeneous subcatchment representations in a manner aligned with hydrological accumulation substantially improves runoff prediction, with the benefits of correct topology growing along the river-network hierarchy. These findings establish physically structured accumulation as an effective inductive bias for hydrological modeling and motivate finer-grained, fully distributed AI-based hydrological models.

Fri 13 NovMachine Learning
The gist
Predicting river runoff accurately is tricky because different parts of a river basin respond differently to rainfall and water flow. The authors created RiverScan, a method that learns how to combine water flow information from smaller parts of the river into bigger ones, respecting how rivers join together. They tested RiverScan in the Danube river basin and found it predicts runoff better than other methods. This approach uses how rivers are connected as a helpful guide for making predictions.
Open → 2611.18005v1

Mass conserving deep learning models simulate water flow in snow watershed

Advancing Mass-Conserving Deep Learning for Distributed Hydrologic Modeling: A Case Study in a Snow-Dominated Watershed

Abstract: Recent progress in physics-aware deep learning (DL) has embedded mass conservation within Long Short-Term Memory (LSTM) architecture for hydrologic modeling, improving physical consistency and enabling diagnostics of internal states. However, existing mass-conserving neural networks are primarily developed in lumped formulations. Consequently, previous studies provide only fragmented and often indirect evidence of when and why imposed physical constraints may lead to unrealistic hydrologic behavior. To address this limitation, we propose Physics-Enhanced Conservation-Augmented Net (PECAN), a mass-conserving convolutional LSTM-based framework for distributed hydrologic modeling and diagnosing structural limitations of physics-aware DL models. PECAN is implemented in two variants: (1) PECAN-MC with a strict precipitation-based conservation constraint and (2) PECAN-MCR with a relaxed conservation constraint allowing additional water inputs beyond precipitation. Both variants are trained solely on gauge streamflow and evaluated in a snow-dominated watershed in Colorado. We benchmark PECAN against lumped DL baselines and the distributed process-based Variable Infiltration Capacity (VIC) model. Results show that PECAN achieves competitive outlet streamflow simulation accuracy while producing spatially explicit latent states for interpretation. Comparisons with reference soil moisture indicate that PECAN cell states resemble meaningful temporal soil moisture dynamics. However, analyses also reveal that PECAN misrepresents snow processes by prematurely removing winter snow water storage from the system rather than retaining it for spring release. PECAN demonstrates accurate streamflow simulation at gauges and shows diagnostic potential to expose structural limitations in hydrologic process representation, thereby informing future physics-aware neural network development for hydrology.

Mon 9 NovMachine Learning
The gist
Modeling how water moves through landscapes, especially in snowy areas, is challenging because the amount of water must be conserved correctly. The authors created a new deep learning tool called PECAN that uses rules about water conservation to simulate water flow across different parts of a watershed. Their model performs well at predicting streamflow and can show hidden water storage patterns, but it struggles with accurately representing snow storage and melt timing. This helps scientists see where the model's understanding is limited and how to improve future water models.
Open → 2611.13346v1

Hydrology model based on physical laws matches neural networks accuracy

From hydrological laws to catchment models: A process-based model matches deep-learning accuracy on three continents

Abstract: Dooge asked whether hydrology has laws at the catchment scale. We ask whether a low-dimensional system built under physical laws can match the accuracy of neural networks that outperform process-based models on the same inputs. Three laws constrain a lumped catchment model. Conservation makes it a flux-connected storage network, dissipation fixes the sign of every flux and aggregation determines how the unresolved composition enters the fluxes, through effective laws over threshold distributions and a memory of the forcing. These laws admit four kinds of refinement. A parsimonious model under them, CaRT (11 parameters per catchment), is refined in a loop assisted by AI agents. Each refinement addresses an observed deficit through a hypothesis from the literature. The result, CaRT+, has fourteen states, 35 parameters per catchment and no daily neural network. Most refinements follow from aggregation over thresholds or from variables that carry no water but modulate fluxes. On 602 British and 593 US gauged catchments, single CaRT+ models reach median Nash-Sutcliffe efficiencies of 0.863 and 0.765 in held-out years. A stateful LSTM reaches 0.869 and 0.751 and an entity-aware LSTM 0.862 and 0.754, while three-seed LSTM ensembles remain ahead of the CaRT+ ensemble by 0.003-0.012 in median paired NSE. On 505 drier Australian gauged catchments unused in development, single CaRT+ models reach 0.778 against 0.762 and 0.753 for the two LSTMs. The results suggest that catchment-scale laws exist as laws of form, with the constitutive families shared and only their parameters learned.

Wed 4 NovMachine LearningArtificial Intelligence
The gist
Predicting how water flows in rivers and catchment areas is important but complicated. The authors created a simple model based on physical laws that predicts water flow almost as well as advanced neural networks, which are complex AI algorithms. They tested this model on many river basins across Britain, the US, and Australia. The results show it is possible to use known physical principles to build accurate and understandable models for hydrology, rather than relying solely on AI.
Open → 2611.06382v1

Basin2Vec improves streamflow prediction in ungauged basins

Basin2Vec: Learning Basin-Centric Embeddings for Scalable Hydrological Modeling

Abstract: Reliable streamflow prediction across time and in ungauged basins requires representations that capture hydrologically relevant basin characteristics. Existing methods rely on either fixed handcrafted attributes or general-purpose Earth observation embeddings not tailored to basin structure. We introduce \textbf{Basin2Vec}, a self-supervised framework that learns compact basin representations from hydroclimatic and physiographic data without observed streamflow. Basin2Vec organizes representation learning around basin scale and relationships across space and time. Across multiple forecasting architectures, it consistently outperforms predefined attributes and general-purpose geospatial embeddings in temporal generalization and zero-shot spatial transfer. Basin2Vec provides average relative gains of up to 8.3% temporally and achieves 3.5x the median predictive performance of the strongest baselines on unseen basins. These results demonstrate the effectiveness of basin-centric pretraining for transferable hydrologic prediction. Code is available at https://github.com/Muhammad-Talha-MT/basin2vec

Tue 3 NovMachine LearningArtificial Intelligence
The gist
Predicting how much water will flow in streams over time is hard, especially in places without measurement instruments. The authors created Basin2Vec, a new way to learn special summaries of river basins from climate and land data without needing water flow measurements. These summaries help computer models predict streamflow better over time and at locations they haven’t seen before. Basin2Vec does this by focusing on the unique features and scales of each basin in training its representation.
Open → 2611.05164v1

Evolving benchmark reveals where environmental models fail over time

EvoBench: An Evolving, Diagnostic Benchmark Framework for Environmental Modeling

Abstract: Machine learning models are now widely used to predict environmental variable dynamics in domains like hydrology, climate science, agriculture, and ecology. These models are usually compared on fixed benchmark datasets using a single average accuracy score. As models improve, their scores become very close when averaged over large regions and long time periods, so the benchmark can no longer tell them apart, and a single number hides \emph{where} and \emph{when} a model fails. Three failures are common across environmental domains: models get worse as the climate shifts (non-stationarity), models fail at unmonitored sites (spatial extrapolation), and models miss rare but important regimes such as floods and droughts. We present \textbf{EvoBench}, a benchmark framework that shows where a model fails, tracks failures over time, and evolves to focus on model capability gaps. We apply EvoBench to daily streamflow prediction as a case study (\textbf{CAMELS-Evolve}, about 9{,}000 U.S. basins, 1985--2024) and compare an operational physical model (NWM), a conceptual reference (HBV), and several data-driven models. No model wins on every testing scenarios (e.g., in terms of average accuracy, performance at low flows and unmonitored basins). Also we observe that all data-driven models under-predict floods. Additionally, we show that targeted fix based on the reported failures could mitigate weaknesses the benchmark identified in certain scenarios. EvoBench turns model evaluation from a saturated ranking into a moving map of model limitations.

Mon 26 OctArtificial IntelligenceMachine Learning
The gist
Environmental models predict things like streamflows and climate, but they often fail under changing conditions, new locations, or rare events like floods. The authors created EvoBench, a tool that shows exactly when and where different models struggle instead of just giving one average score. They tested this on daily streamflow predictions from thousands of U.S. basins and found no single model is best in all situations. EvoBench also helps identify and fix model weaknesses by tracking errors as they change through time.
Open → 2610.35464v1

Deep learning system produces fast global rain maps from satellites

Huayu-Global: Skillful real-time quasi-global precipitation retrieval

Abstract: Satellite precipitation products face a persistent trade-off: real-time estimates are timely but less accurate, whereas the most reliable products are often released months after observation. This delay keeps high-quality precipitation information outside the decision window for flood warning and disaster response. Here we report Huayu-Global, a cross-constellation deep-learning system that combines three geostationary sensor streams from five satellites through a common retrieval framework and generates coherent quasi-global precipitation fields every thirty minutes. Across more than a million independent station records from two periods outside the training years, Huayu-Global improves event detection and precipitation-amount estimation relative to the real-time products evaluated and reaches delayed-product-level skill in key metrics. Huayu-Global therefore turns operational geostationary observations into a timely, high-skill precipitation layer for flood forecasting, water management and disaster response. Its cross-constellation design also provides a scalable foundation for incorporating future geostationary missions into hydrological and weather-prediction systems.

Wed 21 OctArtificial Intelligence
The gist
Getting accurate rainfall data quickly is hard because the best measurements usually take a long time to become available. The authors created Huayu-Global, a deep-learning system that combines data from multiple weather satellites to make detailed rain maps every thirty minutes. This system is almost as accurate as the best delayed rainfall products but works much faster, helping with quicker flood warnings and water management. It can also include new satellites easily to improve weather forecasting in the future.
Open → 2610.29213v1

Ensemble spread underestimates streamflow risks during extreme floods

Ranked Right, Sized Wrong: Ensemble Spread Fails Silently at Streamflow Extremes

Abstract: Deep learning models, particularly Long Short-Term Memory networks, are increasingly used for operational streamflow forecasting, with ensemble spread providing a signal of predictive uncertainty. As climate change increases the frequency of hydroclimatic extremes, these models will increasingly operate under flow conditions that are rare in their historical records. We therefore evaluate whether ensemble spread remains informative under increasing flows, examining both its magnitude relative to error and its ability to rank errors. Prediction error is approximately $2\times$ the corrected spread at ordinary flows and up to $4\times$ at the highest flows, with the discrepancy increasing with flow (log-log slope $1.30$, 95% CI: $[1.20,1.41]$). Under-dispersion is strongest in observed extreme flows beyond the training record, reaching an $18\times$ discrepancy. This is pervasive and not explainable by conditional mean bias. Area Under the Sparsification Error, a standard distribution-free metric of uncertainty ranking, describing how well a model knows when it is uncertain, indicates near-oracle ranking for ensemble spread ($0.071$), yet ordering by predicted discharge alone nearly matches it. Informative uncertainty ranking can coexist with severe under-dispersion. In deployment, error ranking cannot be verified, yet the spread is available, and it is relied on as the uncertainty signal under increasingly extreme conditions, while systematically understating risk, most severely for the extreme events that matter most.

Wed 14 OctMachine Learning
The gist
Streamflow forecasts use deep learning models to predict river flows and how uncertain those predictions are. The authors show that during very high and extreme flows, these models underestimate how uncertain they are, making their predictions look more confident than they should be. Although the models can still rank which predictions are more or less uncertain, the actual amount of uncertainty is much larger than the model says, especially for rare, extreme events. This means that relying on the model’s uncertainty estimates could lead to underpreparedness for serious floods.
Open → 2610.20947v1

Hydrology model simulates rainfall runoff with interpretable water flow states

A Gated Catchment Hydrology Model (GCHM) for Physically-Interpretable Rainfall-Evapotranspiration-Runoff Simulation Across Diverse Hydroclimatic Regimes

Abstract: Recent advances in machine learning have improved rainfall-runoff prediction, yet most data-driven models remain difficult to interpret and do not explicitly enforce physical conservation laws. We present the Gated Catchment Hydrology Model (GCHM), a mass-conserving architecture based on the Mass-Conserving Perceptron (MCP). GCHM uses five storage states representing canopy interception, snow, soil moisture, groundwater, and streamflow generation. Constrained gates govern precipitation partitioning, evapotranspiration, infiltration, seepage, runoff, and baseflow while learning context-dependent process relationships. We evaluate GCHM across 50 CAMELS catchments spanning snowmelt-dominated, recent-rainfall, historical-rainfall, and arid regimes, using NOAA AORC meteorological forcing and remotely sensed vegetation and land-surface data. Performance is assessed using the Jawad-Kling-Gupta Efficiency (JKGE) metric, which accounts for stationary and non-stationary hydrologic behavior. We compare the full architecture with ablated variants and recurrent neural network benchmarks. GCHM reproduces streamflow dynamics across diverse hydroclimatic regimes, including arid catchments, while maintaining interpretable states and fluxes. Ablation experiments show that simplified architectures may perform adequately under wet conditions but struggle to represent flow persistence and recession during dry periods, when groundwater storage sustains low flows. Learned gates recover hydrologically plausible relationships governing snowmelt, runoff, seepage, and evapotranspiration. GCHM combines competitive predictive performance with mass conservation and interpretability, providing a foundation for hydrologic modeling and counterfactual analysis.

Sat 10 OctMachine Learning
The gist
Predicting how rainfall turns into river flow is important but usually hard to explain and make sure water is conserved. The authors built a model called GCHM that tracks water through different parts of a landscape like snow and soil, making sure water amounts add up correctly. It works well across many different climates and pours, and the model helps people understand how rainfall, evaporation, and runoff interact. This makes it easier to trust and use in studying water systems.
Open → 2610.14493v1

Physics guides stable long-term forecasting on open hydrologic graphs

Physics-Refined Spatiotemporal Forecasting on Open-Boundary Hydrologic Graphs

Abstract: Spatiotemporal forecasting on hydrologic graphs is especially prone to instability in open-boundary systems, where the forecast domain exchanges fluxes with an unobserved exterior. In such systems, boundary nodes receive external forcing, e.g., upstream inflows in rivers or tidal signals in coastal regions, that is typically unavailable at prediction time. The absence of this information can compound errors as forecasts unfold in an autoregressive fashion, leading to inferior long-horizon performance. This paper dissects this instability issue by exploring two questions. 1) What boundary forcing enters the forecast domain when information beyond the boundary is missing? 2) How should this forcing propagate through the domain without incurring error amplification under autoregressive rollout? To address both, we propose a new computing framework comprising two key components. First, to compensate for the boundary forcing, our framework learns ghost node proxies from the boundary and interior nodes, striving to approximate unobserved external inputs. Second, to control error accumulation from these learned proxies, we leverage two physics refiners. In particular, one refiner enforces local consistency by aligning ghost proxies with their two-hop neighbors (i.e., boundary nodes and their immediate interiors). The other refiner enhances global stability by correcting the model forecasts through a physics-guided graph neural operator, reducing long-horizon numerical drift. Two real-world hydrologic graphs are employed for empirical evaluation. Comparative results show that our proposal enjoys higher prediction accuracy and long-horizon stability over both learning-based and physics-informed model competitors.

Thu 1 OctMachine Learning
The gist
Forecasting how water behaves in rivers or coastal areas is tricky when parts of the system outside the observed area affect it, but their data is missing. The authors study how missing this outside information causes errors to build up over time in predictions. They propose a method that creates 'ghost' nodes to estimate unknown outside inputs and uses physics-based rules to keep those estimates consistent and stable. This approach improves the accuracy and stability of long-term water flow forecasts in real-world river networks.
Open → 2610.01765v1

Rainatlas dataset improves local forecast accuracy across continents

RainAtlas: A Multi-Continental Dataset for Precipitation Downscaling

Abstract: Extreme rainfall events are increasing in intensity and frequency as climate change accelerates. While kilometer-scale precipitation forecasts are critical for supporting local decision-making, the limited availability of high-resolution precipitation observations hinders their accuracy, especially in under-resourced regions. Machine learning models are widely used to downscale precipitation data to km-scale, but their application to unseen geographies presents challenges. First, processing raw high-resolution precipitation datasets across regions requires significant engineering and domain expertise. Second, generalization across regions remains difficult. To help overcome these barriers, we release RainAtlas, a large-scale, ML-ready and multi-continental dataset for precipitation downscaling. Covering three continents, RainAtlas harmonizes heterogeneous hourly km-scale observations to a common 2-km grid. Each regional partition contains around 210,000 aligned low- and high-resolution precipitation pairs, respectively from ERA5 reanalysis and direct observations. We benchmark state-of-the-art ML-based downscaling models across RainAtlas using a wide range of metrics. Our evaluation reveals substantial variance in out-of-domain generalization depending on the training regions. This underscores the need for cross-regional, multi-source km-scale evaluation, establishing RainAtlas as a well-positioned benchmark for precipitation downscaling research.

Wed 30 SeptMachine Learning
The gist
More intense and frequent heavy rainfall is happening because of climate change, making it important to predict rain at a fine scale for local areas. The paper discusses RainAtlas, a large dataset that combines detailed hourly rain data from three different continents on a common grid. This makes it easier to train and test machine learning models for improving rain forecasts at about 2 kilometers resolution, even in places without many observations. The authors found that models trained in one region don’t always work well in others, highlighting the importance of testing across multiple regions. RainAtlas provides a standardized way to study and improve rain downscaling methods worldwide.
Open → 2609.39833v1

Neural methods guide when satellite water use estimates should change

When Should a Satellite Estimate Be Changed? Stress-Testing Neural Corrections for Evapotranspiration

Abstract: Neural residuals can improve satellite evapotranspiration (ET) estimates, but selectors must predict when a correction helps and reject unsupported inputs. We evaluate ten-member models on 16,366 flux-tower observations from 151 stations paired with OpenET, across nine rolling years and five spatial folds. At one held-out station, Gain accepted corrections on all 32 physically invalid records: it predicted a mean benefit of 0.83 mm/day, but the corrections increased mean absolute error by 21.6 mm/day versus OpenET. On spatially held-out unit errors, SupportGain reduced station-macro MAE versus Gain by 0.148 mm/day under wind x3.6 (simultaneous 95% interval, 0.070 to 0.226), with 9.3% acceptance versus Gain's 51.8%; on clean inputs, its 0.006 mm/day advantage had an interval that includes zero. These fault analyses are exploratory; none of 40 preplanned temporal comparisons passed Holm correction, while a separate predeclared cropland contrast found 0.041 mm/day lower station-macro MAE with crop-only training (95% interval, 0.009 to 0.079).

Mon 28 SeptMachine Learning
The gist
Satellites estimate how much water plants release into the air, but these estimates can sometimes be wrong. The authors studied ways to use neural networks to fix these errors, but only if the fix actually helps. They found that some methods wrongly accepted bad corrections, while others were better at rejecting harmful fixes. Their tests show cautious use of neural corrections can improve accuracy, especially for crops.
Open → 2609.35314v1

SegFormer model estimates snow water and height from satellite data

Learning Regional Snow Water Equivalent and Snow Height Variations from Sentinel-1 InSAR Acquisitions

Abstract: Managing water resources in mountainous regions depends heavily on reliable Snow Water Equivalent (SWE) and Snow Height (HS) data, yet these variables remain difficult to track at scale. This study evaluates three machine learning architectures (XGBoost, U-Net and SegFormer) for the joint estimation of SWE and HS variations from Sentinel-1 InSAR data over the Italian Alps, using the IT-SNOW reanalysis as reference. SegFormer achieves the best results on both targets, with an MAE of 10.391 cm for HS and 27.113 mm w.e. for SWE and the lowest variability across initializations. A feature sensitivity analysis shows that including all available features does not guarantee the lowest error, with model- and task-specific sensitivities. Spatial metrics (R2, Pearson's r) separate the three architectures far more clearly than mean error (MAE, RMSE) does, and decomposing the error per window attributes most of it to a systematic offset in the estimated mean variation rather than to the spatial pattern.

Mon 28 SeptMachine Learning
The gist
Measuring how much water is stored in mountain snow is hard but important for managing water resources. This study checked three computer programs to figure out snow measurements from satellite images taken over the Alps. The SegFormer program did the best job estimating both snow depth and water content. The researchers also found that including every possible input didn’t always improve the results, and that most errors were due to consistent differences in average estimates rather than errors in detailed patterns.
Open → 2609.34614v1

Hapi improves continental flood forecasts with fast multivariable AI model

Hapi: A Multivariable Land-Surface Transformer for Medium-Range Hydrological Forecasting at Continental Scale

Abstract: Accurate flood forecasts several days in advance are essential for flood control, water-resource management, and emergency response. Producing them at high resolution over a continental domain calls for local hydrological detail together with spatial context extending from river basins to synoptic weather systems. We developed Hapi, a U-Net Swin Transformer that uses fine three-dimensional patches and hierarchical shifted-window attention to forecast discharge, surface runoff, snow water equivalent, and soil wetness across the contiguous United States. The model produces 24--72-hour forecasts at $0.05^{\circ}$ resolution, with learned Laplacian task weights adjusting each variable's contribution to training. On 2024 test data using reconstructed weather and land-surface inputs from ERA5-Land, Hapi outperformed an operational physics-based model and a state-of-the-art AI model in flood detection. Independent validation against 3,881 U.S. Geological Survey gauges and a Hurricane Helene case study supported its advantage over the physics-based model in reproducing daily discharge. Controlled experiments showed that learned task weighting strengthens rare-flood detection, which is particularly sensitive to changes in precipitation inputs. Hapi produced a four-variable, 72-hour forecast across the contiguous United States with an average inference time of 0.11 seconds on a single A100 GPU.

Sat 19 SeptArtificial Intelligence
The gist
Flood forecasting helps people prepare for heavy rains and floods days before they happen. The authors created Hapi, an AI model that quickly predicts water flow, runoff, snow, and soil wetness over the United States with fine details and wide context. It works faster and more accurately than current physics models and other AI models, especially detecting rare floods. The model runs quickly on a powerful GPU and could support better flood control and emergency responses.
Open → 2609.22702v1

Physical knowledge improves historic data forecasting of groundwater levels

Physical knowledge on historical data matters more than enforcing physical constraints on the forecast

Abstract: Time series forecasting has seen signicant advancements with the emergence of new deep learning models. However, forecasting time series in applications involving physical processes remains a major challenge. Despite the apparition of Physics Informed Neural Networks (PINN), recent models do not estimate unobservable intermediate physical variables, which are important for domain experts to understand the target behavior. To this end, we propose a Physics Informed Recurrent Neural Network (PIRNN) which predicts, along the target, unobservable variables on both historic data and forecast target. This approach enhances the model robustness and results interpretation using domain knowledge. Our method is easily adaptable to any physical model using several equations, each having its own set of unobservable variables, to describe it-self. As a case study, we incorporate physical equations used for groundwater levels predictions by the physical model called Gardenia. This model uses transfers equations between reservoirs, optimized with data assimilation, to simulate the evolution of groundwater levels. Evaluation includes several well known neural network models and the Gardenia model compared on twelve real world datasets. In addition, we study the impact of each component through an ablation study. Our model outperforms other models on ve out of the twelve datasets and our ablation study underlines the importance of having a physical background in our time series forecasting task. Finally, the coherence of the physical variables predicted by our neural network is assessed by a domain expert.

Thu 17 SeptArtificial Intelligence
The gist
Forecasting things that depend on physical processes, like groundwater levels, is hard because some important parts can't be measured directly. The authors created a new method that predicts these hidden physical variables alongside the usual forecast, using knowledge about how the physical process works. This makes their predictions more reliable and easier to understand. They tested their approach on real groundwater data and found it often works better than other models, and experts confirmed the predictions make physical sense.
Open → 2609.19871v1

Model and dataset improve water level estimates in sparse river networks

A Dataset and Model for Imputing Water Surface Elevation on a Large and Extremely Sparse Spatiotemporal Graph

Abstract: Continuous monitoring of water surface elevation across river networks is critical for flood forecasting, water resource management, and understanding the global water cycle. Yet, the scarcity of in situ gauges across much of the globe constrains the development of reliable modeling frameworks. Satellite altimetry has the potential to alleviate this problem but its use is currently hindered by sparse temporal coverage. To this end, we introduce AmazonSWE, a dataset for training and evaluating large-scale spatiotemporal graph imputation methods that integrates processed satellite altimetry measurements from a range of sources, including the recent wide-swath SWOT sensor. The dataset covers over 19K river sections and 10 years (2016-2026) in the Amazon river basin, with in situ gauges held out for evaluation. Besides contributing a novel real-world use case with the potential for societal impact, AmazonSWE introduces significant technical challenges: with fewer than 1% of sections observed per day, the dataset is far sparser than existing imputation benchmarks, and its directed acyclic river topology is both structurally different from and larger than graphs in existing datasets. We show that prior spatiotemporal graph imputation methods are not adapted to this topology, scale and sparsity, and propose a simple bidirectional selective state space model that outperforms them by sampling connected subgraphs and flattening space and time into a single token sequence with topology-aware positional encodings. Compared to the state-of-the-art published method for SWOT-based WSE densification, which integrates statistics with physical modeling, our model reduces RMSE against in situ gauges by 18-39%, while producing predictions for every river section rather than only those with sufficient nearby satellite coverage.

Thu 10 SeptMachine Learning
The gist
Measuring water levels in rivers is important for predicting floods and managing water but is hard because there are very few sensors on many rivers. The authors created a large dataset of satellite measurements over many rivers in the Amazon basin to help estimate water levels where data is missing. They found existing methods did not work well for this sparse and complex data, so they made a new model that uses river network structure and time together to make better guesses. Their model predicts water levels more accurately than previous approaches, especially in places without nearby satellite data.
Open → 2609.11580v1

Neptune emulates ocean conditions for up to 60 days globally

Neptune: An AI model for Global Ocean Subseasonal Prediction

Abstract: Subseasonal-to-seasonal (S2S) forecasting is societally critical, supporting decision-making in sectors ranging from water and agricultural management to disaster risk reduction, energy planning, and insurance. Achieving reliable predictions at these timescales requires representing the ocean and its dynamics, but traditional physics-based Ocean General Circulation Models (OGCMs), are computationally expensive and difficult to develop and improve because of the code complexity. In this work, we propose Neptune, an end-to-end data-driven framework for global ocean and sea-ice components emulation tailored for S2S timescales, up to 60 days. Neptune combines Convolutional Neural Networks (CNNs) and Spherical Fourier Neural Operators (SFNOs) to effectively capture local features and global cross-scale interactions, thereby obtaining a coherent representation of the ocean state. Forced by prescribed daily atmospheric fields, Neptune emulates ocean state variables, from temperature and salinity, to zonal and meridional currents, from sea surface height to sea ice thickness and concentration, with daily outputs at the ocean surface and through the water column. Specifically, we propose two variants of Neptune, Neptune-1 and Neptune-025, capable of emulating the ocean state at 1° and 0.25° resolution, respectively. Evaluated against a suite of metrics, including statistics (RMSE, CRPS and ACC), physical coherency (Ocean Heat Content, Eddy Kinetic Energy and Ice Brier Score) and climate indices (ENSO and Z20 metric, IOD), Neptune successfully reproduces the spatio-temporal evolution of the oceanic fields up to 60 days, and is stable over long timescales. Neptune provides compelling evidence that end-to-end data-driven ocean emulators can become a powerful component of next-generation S2S forecasting systems, emulating ocean state at high spatio-temporal resolution.

Tue 8 SeptArtificial Intelligence
The gist
Predicting the ocean's state for weeks ahead is important for managing resources and disasters. Traditional models are slow and complex. The authors developed Neptune, an AI system that uses neural networks to quickly estimate ocean temperatures, currents, and ice conditions daily for up to 60 days worldwide. Neptune matches key ocean patterns well and can run stably over long periods. This shows AI can help improve subseasonal ocean forecasts.
Open → 2609.08606v1