Konstantinos M. Andreadis

Given the insurmountable challenge of measuring all rivers in situ, global river models serve as the foundation of freshwater knowledge past, present, and future. We adopt an inductive empirical framework based on Surface Water and Ocean Topography (SWOT) satellite measurements to assess these models. After controlling for SWOT data quality, we examine 68,347 individual river reaches representing ∼38% of global discharge. We find river models currently struggle in areas of heavy economic development, multi‐channel rivers, arid areas, and many Arctic rivers. After controlling for these expected errors, we find better skill as rivers get wider and that parts of Siberia and China are particularly difficult to model. We also find large variability and spatial heterogeneity to model performance, resisting oversimplification. Our results suggest that leveraging SWOT observations within river models will improve them, but river models must adapt their structure to represent realistic hydraulics to do so.

Satellite remote sensing is widely used to monitor lake surface water temperature (LSWT) due to its global coverage and relatively long‐term record. A common practice in previous studies is to exclude observations during cloudy periods, as most satellite‐based LSWT products rely on optical sensors that cannot penetrate cloud cover. Using synthetic LSWT data sets, we demonstrate that cloud cover tends to coincide with specific thermal conditions, introducing geographically structured biases in annual mean LSWT estimates. We further conducted global‐scale synthetic numerical experiments to examine the impact of cloud cover propagates into model predictions when satellite‐derived LSWTs are used for model development. Two modeling approaches were evaluated: a physically‐based model (Air2water) and a data‐driven model based on a Long Short‐Term Memory (LSTM) neural network. While the LSTM model achieved an overall higher predictive accuracy, it exhibited significantly amplified biases when trained on cloud‐affected data sets, particularly in the warmest‐month temperature and ice duration. In contrast, Air2water showed relatively stable bias patterns, reflecting its resistance to data gaps due to its physical constraints. Our findings underscore the importance of accounting for cloud‐induced selection bias in optical satellite remote sensing and environmental modeling.

Accurate spatiotemporal monitoring of inland water bodies is crucial, since, apart from numerous ecosystem services they also provide valuable water resources. This is particularly true in water-limited Mediterranean regions where detailed characterization of lake water extent, level and storage could facilitate sustainable water resources management under climate extremes (e.g., droughts). Here, focusing on the Trichonida – Lysimachia lake complex in Western Greece, we synthesized remote sensing observations, in-situ measurements, and auxiliary environmental and geospatial datasets, in order to characterize the spatiotemporal dynamics in their water extent, level, and storage. The Trichonida – Lysimachia lake complex is a sensitive ecosystem, protected as part of the Natura 2000 network; lake Trichonida is the largest natural lake in Greece (surface area of ~93 km2 and maximum depth of ~52 m) and is connected through an open channel with the much smaller and shallower lake Lysimachia (surface area of ~10 km2 and maximum depth of ~8 m). The analysis of optical (Landsat 5, 7, 8, 9, Sentinel 2) and microwave (Sentinel 1) satellite imagery revealed that both lakes displayed significant changes in their areal extent at the seasonal and annual time scale, with these results being more pronounced for the lake with shallower bathymetry (i.e., Lysimachia). The surface area of lake Trichonida (Lysimachia) decreased significantly during the period 1985 – 2024 at a rate equal of 32.3 m2 yr-1 (36.4 m2 yr-1) with hotspots that displayed more than 100 m shift in the shoreline. In-situ water level measurements agreed well with estimates from satellite altimetry (ICESat, SWOT), and, when combined with the detailed bathymetries of the two lakes, detailed water level-area-volume curves were derived. Such curves, synthesize multivariate observations, in-situ measurements, and cross-disciplinary hydrogeodetic techniques and reveal lake-specific 3D patterns. The obtained results offer valuable insights not only towards the sustainable management of the two lakes but can also contribute to the refinement of regional- and global-scale initiatives on satellite-based 3D lake monitoring.

Suspended sediment concentration, flux, and river discharge are essential indicators of river ecosystem health and reflect watershed‐scale processes. Monitoring these variables is labor‐intensive, leading to sparse and geographically biased observations and the development of models to fill in the observational gaps. These models generally use either climatological data or satellite images to estimate one of these variables. In this work, we present a novel deep learning model that can leverage multiple data sources with different temporal characteristics to produce continuous daily estimates of suspended sediment concentration (SSC), suspended sediment flux (SSF), and discharge. The model first encodes daily hydrological data from the ERA5‐Land reanalysis using a Long Short‐Term Memory network and water color data from Landsat satellites using a Multi‐Layer Perceptron network, then merge these encoded data sources using a cross‐attention decoder. We train and test the model on a large data set of in situ observations from 630 river sites over 43 years in the contiguous United States, covering a wide range of watersheds and conditions. We produce SSC, SSF, and discharge predictions with respective relative errors of 49%, 57%, and 44%, and relative bias of −2.5%, 2.6%, and 3.7%. We use our model to create a data set of continuous daily SSC, SSF, and discharge for all large rivers in the contiguous United States. This new model architecture provides a valuable tool for monitoring river systems, addressing limitations of single‐source models and offering a framework applicable to other Earth systems monitoring problems where integrating diverse data streams may be useful.

The water in Earth's rivers propagates as waves through space and time across hydrographic networks. A detailed understanding of river dynamics globally is essential for achieving accurate knowledge of surface water storage and fluxes to support water resources management and water‐related disaster forecasting and mitigation. Global in situ information on river flows are crucial to support such an investigation but remain difficult to obtain at adequate spatiotemporal scales, if they even exist. Many expectations are placed on remote sensing techniques as key contributors. Despite a rapid expansion of satellite capabilities, however, it remains unclear what temporal revisit, spatial coverage, footprint size, spatial resolution, observation accuracy, latency time, and variables of interest from satellites are best suited to capture the space‐time propagation of water in rivers. Additionally, the ability of numerical models to compensate for data sparsity through model‐data fusion remains elusive. We review recent efforts to identify the type of remote sensing observations that could enhance understanding and representation of river dynamics. Key priorities include: (a) resolving narrow water bodies (finer than 50–100 m), (b) further analysis of signal accuracy versus hydrologic variability and relevant technologies (optical/SAR imagery, altimetry, microwave radiometry), (c) achieving 1–3 days observation intervals, (d) leveraging data assimilation and multi‐satellite approaches using existing constellations, and (e) new variable measurement for accurate water flux and discharge estimates. We recommend a hydrology‐focused, multi‐mission observing system comprising: (a) a cutting‐edge single or dual‐satellite mission for advanced surface water measurements, and (b) a constellation of cost‐effective satellites targeting dynamic processes.

Suspended sediment concentration, flux, and river discharge are essential indicators of river ecosystem health and reflect watershed-scale processes. Monitoring these variables is labor-intensive, leading to sparse and geographically biased observations and the development of models to fill in the observational gaps. These models generally use either climatological data or satellite images to estimate one of these variables. In this work, we present a novel deep learning model that can leverage multiple data sources with different temporal characteristics to produce continuous daily estimates of suspended sediment concentration (SSC), suspended sediment flux (SSF), and discharge. The model first encodes daily hydrological data from the ERA5-Land reanalysis using a Long Short-Term Memory network and water color data from Landsat satellites using a Multi-Layer Perceptron network, then merge these encoded data sources using a cross-attention decoder. We train and test the model on a large dataset of in-situ observations from 630 river sites over 43 years in the contiguous United States, covering a wide range of watersheds and conditions. We produce SSC, SSF, and discharge predictions with respective relative errors of 54\%, 73\%, and 28\%, and relative bias of -15\%, -19\%, and -3\%. We use our model to create a dataset of continuous daily SSC, SSF, and discharge for all large rivers in the contiguous United States. This new model architecture provides a valuable tool for monitoring river systems, addressing limitations of single-source models and offering a framework applicable to other Earth systems monitoring problems where integrating diverse data streams may be useful.

Quantifying lake ice loss is crucial for understanding the impact of climate change on lake ecosystems. In this study, we trained a deep learning model (Long‐Short Term Memory with Landsat observations, 1984–2012) to simulate Northern Hemisphere lake ice changes at a fine spatial scale from 1980 to 2022. The model achieved good performance overall during the test period (2013–2022), and the derived ice‐on and ice‐off matched well with two independent ice phenology data sets. Results reveal a 76.8% increase in intermittently ice‐covered lakes from the 1980s to the 2010s, alongside a 10.7‐day shorter ice duration and a 3.9 percentage‐points reduction in annual mean ice cover fractions. The model can track daily partial ice cover changes, providing a novel contribution to understanding shifts in lake ice cover with climate change. These findings can provide valuable insights for future limnology studies, such as improving estimates of greenhouse gas emissions from lakes.

The Surface Water and Ocean Topography (SWOT) satellite has the potential to transform global hydrologic science by offering simultaneous and synoptic estimates of river discharge and other hydraulic variables. Discharge is estimated from SWOT observations of water surface elevation, width, and slope. A first assessment using just the highest quality SWOT measurements, over the first 15 months (March 2023–July 2024) of the mission evaluated at 65 gauged reaches shows results consistent with pre‐launch expectations. SWOT estimates track discharge dynamics without relying on any gauge information: median correlation is 0.73, with a correlation interquartile range of 0.51–0.89. SWOT estimates capture discharge magnitude correctly in some cases but are biased (median bias is 50%) in others. There are already a total of 11,274 ungauged global locations with highest quality SWOT measurements where SWOT discharge is expected to accurately track discharge variations: this value will increase as SWOT data record length grows, algorithms are refined and SWOT measurements are reprocessed. This first look indicates that SWOT discharge is performing as expected for SWOT data that achieve performance requirements, providing observed information on discharge variations in ungauged basins globally.

Land-use intensification and climate change are increasing pressure on water availability and use around the world. It is becoming urgent to understand hydrological cycles to manage water availability for natural and human systems. Forests cover 31% of global land area and are crucial for storing and releasing precipitation, however, it is difficult to quantify forest hydrology processes and apply the learnings from one watershed to another. In New Zealand, the 5-year Forest Flows MBIE Endeavour Research Programme (https://www.scionresearch.com/science/sustainable-forest-and-land-management/Forest-flows-research-programme) investigated these challenges with the novel integration of various terrestrial and remote sensing data in Pinus radiata (D. Don) plantation forests. At total of 1,717 terrestrial sensors were deployed above and below ground in wireless IoT sensor networks across five watersheds with a range of climatic and physiographic regions. The Kafka Big Data Pipeline streamed, cleaned and stored the 360,000 observations collected every 24-hours. The fusion of temporally rich terrestrial data and spatially rich remote sensing data provided new insights into the mechanistic drivers of forest hydrological processes at the point (tree), watershed, forest scales. Forest Flows used both traditional and machine learning methodologies, as well as process-based modelling, to quantify tree water use, watershed water storage and release. This presentation will introduce a novel deep learning (DL) framework applied to Big Data in environmental science, with a particular focus on the DL-based Neural Ordinary Differential Equations (NODE) Hydrological Framework. This innovative approach enabled high-precision super-resolution predictions of forest soil moisture derived from NASA's Soil Moisture Active Passive (SMAP) Mission, downscaling from a 9 km to a 1 km spatial resolution. Additionally, the framework provided reliable predictions for regions lacking direct observational products. We will demonstrate how this DL methodology can be leveraged to predict evapotranspiration, as well as surface and subsurface water fluxes, at fine spatial and temporal resolutions within forest ecosystems. The potential applications of this approach extend beyond forest environments, offering insights for other complex environmental Big Data challenges.