Swedish Meteorological and Hydrological Institute

The surface turbulent transport processes in the central Baltic Sea during summer are investigated using Lagrangian trajectory simulations forced by an operational ocean model and Finite Size Lyapunov Exponent (FSLE) and related diagnostics of horizontal stirring. The spatial distribution of Lagrangian Coherent Structures (LCS) and the local prevalence of attracting versus repelling LCS are presented. Coastal areas are characterised by strong stirring in contrast to the weakly stirred offshore regions. The along-trajectory absolute value of relative vorticity and total strain attest to the important role of eddies, fronts and jets in driving horizontal stirring in the ocean model. Strong asymmetry in LCS distribution between the eastern and western coast of Gotland is found and its consequences on the transport of algae blooms and pollutants are discussed.

A multi-year, high-resolution simulation of the central Baltic Sea with additional numerical tracers to track saltwater concentration and age was performed. A horizontal grid resolution of 250 m can be regarded as submesoscale permitting in the Baltic Sea. Comparing this with a coarse-resolution simulation indicates significant differences in the propagation statistics of the tracers due to differences in the physical processes that are resolved. Ventilation of the halocline and overturning circulation are more intense in the high-resolution simulation, but lateral tracer transport is greater in the coarse-resolution simulation. The long-term mean vertical velocities are stronger in the high-resolution simulation, which indicates the presence of submesoscale processes contributing to vertical exchange in the Baltic Sea. Results of this study highlight the importance of submesoscale processes also for other semi-enclosed regional seas.

Data assimilation (DA) is a cornerstone of scientific and engineering applications, combining model forecasts with sparse and noisy observations to estimate latent system states. Classical high-dimensional DA methods, such as the ensemble Kalman filter, rely on Gaussian approximations that are violated for complex dynamics or observation operators. To address this limitation, we introduce DAISI, a scalable filtering algorithm built on flow-based generative models that enables flexible probabilistic inference using data-driven priors. The core idea is to use a stationary, pre-trained generative prior that first incorporates forecast information through a novel , before assimilating observations via guidance-based conditional sampling. This allows us to leverage any forecasting model as part of the DA pipeline without having to retrain or fine-tune the generative prior at each assimilation step. Experiments on challenging nonlinear systems show that DAISI achieves accurate filtering results in regimes with sparse, noisy, and nonlinear observations where traditional methods struggle.

The monitoring and modeling of riverine floods have been covered extensively in the scientific literature with a substantial number of scientific contributions related to calibration/validation of hydraulic and hydrological models and assimilation of Earth Observation (EO) data into them. These models, when used for flood forecasting purposes, rely heavily on ground-based hydrological networks along with numerical weather models which, particularly in data-scarce regions, are often challenged by data sparsity. In these situations, EO data offer a viable solution to enhance the skill of these flood forecasting systems by providing global-scale observations of key hydrological variables such as precipitation, soil moisture, river discharge, water levels, and flood extent. This manuscript reviews and discusses the capability of these EO data in enhancing flood forecasting systems, by analyzing their accuracy, lead time, and reliability, while at the same time highlighting key challenges such as data latency, spatial–temporal resolution trade-offs, and model assimilation constraints. By leveraging recent advancements in remote sensing, data assimilation techniques, and artificial intelligence, EO-based flood forecasting has the potential to bridge existing observational gaps, particularly in vulnerable regions. The paper also outlines future research directions and technological developments needed to maximize the impact of satellite data in operational flood forecasting systems.

Understanding the impact of climatic drivers on future catchment runoff is crucial for assessing how climate change will influence hydrological regimes. This knowledge is essential for evaluating water resources at the local and regional scales. Here, we assessed the contributions of climatic drivers to runoff changes under future conditions across Europe by analyzing the sensitivity of runoff to changes in precipitation and potential evapotranspiration using the Budyko framework. This analysis was conducted for 35,408 basins and based on ensembles of bias-adjusted climate projections and hydrological data. We further explored the link between runoff changes and catchment hydrological similarities, investigating the impact of climatic drivers on runoff across different hydrological regimes in three future periods (early, mid and late century) and emission scenarios. Results indicate precipitation as the primary driver of runoff changes across Europe under low and medium emission scenarios. However, by the mid and late century in high emission scenarios, potential evapotranspiration’s contribution becomes comparable to precipitation, significantly influencing runoff variations in central and southern Europe. Finally, the sensitivity of the runoff changes to climatic drivers is linked to the regime of the river systems. Runoff changes in basins with slow response to precipitation tended to show a more balanced contribution from both precipitation and potential evapotranspiration under high emission scenarios, while basins that are very responsive to precipitation, changes in precipitation primarily led to runoff changes. This Budyko-based diagnosis provides new insights into attribution analysis by linking hydrological regimes to runoff changes under global warming, supporting more effective local adaptation strategies.