A5059047105- Hydrology and Watershed Management Studies
- Flood Risk Assessment and Management
- Hydrological Forecasting Using AI
- Hydrology and Drought Analysis
- Meteorological Phenomena and Simulations
- Plant Water Relations and Carbon Dynamics
- Climate variability and models
- Cryospheric studies and observations
- Precipitation Measurement and Analysis
- Neural Networks and Applications
- Water resources management and optimization
- Time Series Analysis and Forecasting
- Landslides and related hazards
- Complex Systems and Time Series Analysis
- Soil Geostatistics and Mapping
- Soil Moisture and Remote Sensing
- Water-Energy-Food Nexus Studies
- Statistical Mechanics and Entropy
- Geology and Paleoclimatology Research
- Groundwater flow and contamination studies
- Soil erosion and sediment transport
- Computational Physics and Python Applications
- Reservoir Engineering and Simulation Methods
- Geochemistry and Geologic Mapping
- Tropical and Extratropical Cyclones Research
Karlsruhe Institute of Technology
2015-2024
Institute of Hydrology of the Slovak Academy of Sciences
2013
Norsk Hydro (Germany)
2011
Technical University of Munich
2009-2010
The Prediction in Ungauged Basins (PUB) initiative of the International Association Hydrological Sciences (IAHS), launched 2003 and concluded by PUB Symposium 2012 held Delft (23–25 October 2012), set out to shift scientific culture hydrology towards improved understanding hydrological processes, as well associated uncertainties development models with increasing realism predictive power. This paper reviews work that has been done under six science themes Decade outlines challenges ahead for...
Abstract. Despite considerable progress in recent years, output of both global and regional circulation models is still afflicted with biases to a degree that precludes its direct use, especially climate change impact studies. This well known, overcome this problem, bias correction (BC; i.e. the model towards observations post-processing step) has now become standard procedure In paper we argue BC currently often used an invalid way: it added GCM/RCM chain without sufficient proof...
Abstract. The July 2021 flood in central Europe was one of the five costliest disasters last half century, with an estimated total damage EUR 32 billion. aim this study is to analyze and assess within interdisciplinary approach along its entire process chain: synoptic setting atmospheric pressure fields, processes causing high rainfall totals, extraordinary streamflows water levels affected catchments, hydro-morphological effects, impacts on infrastructure society. In addition, we address...
Abstract. Heavy precipitation over western Germany and neighboring countries in July 2021 led to widespread floods, with the Ahr Erft river catchments being particularly affected. Following event characterization process analysis Part 1, here we put historical context regarding discharge records terms of temporal transformation valley morphology. Furthermore, evaluated role ongoing future climate change on modification rainfall totals associated flood hazard, as well implications for...
Abstract. Throughout its historical development, hydrology as an earth science, but especially a problem-centred engineering discipline has largely relied (quite successfully) on the assumption of stationarity. This includes assuming time invariance boundary conditions such climate, system configurations land use, topography and morphology, dynamics flow regimes flood recurrence at different spatio-temporal aggregation scales. The justification for this was often that when compared with...
Abstract. The organization of drainage basins shows some reproducible phenomena, as exemplified by self-similar fractal river network structures and typical scaling laws, these have been related to energetic optimization principles, such minimization stream power, minimum energy expenditure or maximum "access". Here we describe the dynamics systems using thermodynamics, focusing on generation, dissipation transfer free associated with flow sediment transport. We argue that reflects...
Abstract. According to Dooge (1986) intermediate-scale catchments are systems of organized complexity, being too and yet small be characterized on a statistical/conceptual basis, but large heterogeneous in deterministic manner. A key requirement for building structurally adequate models precisely this intermediate scale is better understanding how different forms spatial organization affect storage release water energy. Here, we propose that combination the concept hydrological response...
This paper investigates the relationship between expert judgement and numerical criteria when evaluating hydrological model performance by comparing simulated observed hydrographs. Using a web-based survey, we collected visual evaluations of 150 experts on set high- low-flow We then compared these answers with results from 60 criteria. Agreement was found to be more frequent in absolute terms (when rating models) than relative models), better for high flows low flows. When judgements...
Abstract. Surface topography is an important source of information about the functioning and form a hydrological landscape. Because its key role in explaining processes structures, also because wide availability at good resolution digital elevation models (DEMs), it frequently used to inform analyses. Not surprisingly, several indices have been proposed for linking geomorphic properties landscape with functioning; widely example “height above nearest drainage” (HAND) index. From...
Abstract. The July 2021 flood in central Europe was one of the five costliest natural disasters last half century with estimated total damage EUR 32 billion. This study investigates complex interactions between meteorological, hydrological, and hydro-morphological processes mechanisms that led to exceptional flood. Furthermore, we present our estimates impacts terms inundation areas, traffic disruptions, economic losses. estimation areas as well derived assessments were carried out during or...
Abstract. Hydrological hybrid models have been proposed as an option to combine the enhanced performance of deep learning methods with interpretability process-based models. Among various available, dynamic parameterization conceptual using long short-term memory (LSTM) networks has shown high potential. We explored this method further evaluate specifically if flexibility given by overwrites physical part. conducted our study a subset CAMELS-GB dataset. First, we show that model can reach...
Long Short-Term Memory (LSTM) networks have demonstrated state-of-the-art performance for rainfall-runoff hydrological modeling. However, most studies focus on daily-scale predictions, limiting the benefits of sub-daily (e.g. hourly) predictions in applications like flood forecasting. Moreover, training an LSTM exclusively data is computationally expensive, and may lead to model-learning difficulties due extended sequence lengths. In this study, we introduce a new architecture,...
Abstract. Applying metrics to quantify the similarity or dissimilarity of hydrographs is a central task in hydrological modelling, used both model calibration and evaluation simulations forecasts. Motivated by shortcomings standard objective such as Root Mean Square Error (RMSE) Absolute Peak Time (MAPTE) advantages visual inspection powerful tool for simultaneous, case-specific multi-criteria (yet subjective) evaluation, we propose new metric termed Series Distance, which close accordance...
Abstract. This study investigates whether a thermodynamically optimal hillslope structure can, if existent, serve as first guess for uncalibrated predictions of rainfall–runoff. To this end we propose thermodynamic framework to link rainfall–runoff processes and dynamics potential energy, kinetic energy capillary binding in catchments hillslopes. The starting point is that hydraulic equilibrium soil corresponds local (LTE), characterized by maximum entropy/minimum free water. Deviations from...
Abstract. The 2 June 2008 flood-producing storm on the Starzel river basin in South-West Germany is examined as a prototype for organized convective systems that dominate upper tail of precipitation frequency distribution and are likely responsible flash flood peaks Central Europe. availability high-resolution rainfall estimates from radar observations rain gauge network, together with indirect peak discharge detailed post-event survey, provided opportunity to study detail...
Abstract. The increasing diversity and resolution of spatially distributed data on terrestrial systems greatly enhance the potential hydrological modeling. Optimal parsimonious use these sources requires, however, that we better understand (a) which system characteristics exert primary controls dynamics (b) to what level detail do those need be represented in a model. In this study develop test an approach explore questions draws upon information theoretic thermodynamic reasoning, using...
Quantitative precipitation estimation and forecasting (QPE QPF) are among the most challenging tasks in atmospheric sciences. In this work, QPE based on numerical modelling data assimilation is investigated. Key components Weather Research Forecasting (WRF) model combination with its 3D variational scheme, applied convection-permitting scale sophisticated physics over central Europe. The system operated a 1-hour rapid update cycle processes large set of situ observations, from French radar...
The potential role of rural land use in mitigating flood risk and protecting water supplies continues to be great interest regulators planners. ability hydrologists quantify the impact change on cycle is however limited we are not able provide consistently reliable evidence support planning policy decisions. This shortcoming stems mainly from lack data, but also modelling methods tools. Numerous research projects over last few years have been attempting address underlying challenges. paper...
Abstract. In this study we propose and demonstrate a data-driven approach in an “information-theoretic” framework to quantitatively estimate precipitation. context, predictive relations are expressed by empirical discrete probability distributions directly derived from data instead of fitting applying deterministic functions, as is standard operational practice. Applying probabilistic relation has the benefit providing joint statements about rain rate related estimation uncertainty. The...
Using information-theoretic quantities in practical applications with continuous data is often hindered by the fact that probability density functions need to be estimated higher dimensions, which can become unreliable or even computationally unfeasible. To make these useful more accessible, alternative approaches such as binned frequencies using histograms and k-nearest neighbors (k-NN) have been proposed. However, a systematic comparison of applicability methods has lacking. We wish fill...
Abstract. Data-driven techniques have shown the potential to outperform process-based models in rainfall–runoff simulation. Recently, hybrid models, which combine data-driven methods with approaches, been proposed leverage strengths of both methodologies, aiming enhance simulation accuracy while maintaining a certain interpretability. Expanding set test cases evaluate under different conditions, we their generalization capabilities for extreme hydrological events, comparing performance...
Deep learning methods in hydrology have traditionally focused on deterministic models, limiting their ability to quantify prediction uncertainty. Recent advances generative modeling opened new possibilities for probabilistic modelling various applied fields, including hydrological forecasting (Jahangir & Quilty, 2024). These models learn represent underlying probability distributions using neural networks, enabling uncertainty quantification through sampling a very flexible...
Meteorological observations (e.g. from weather radar) and the output of meteorological models reanalyses or forecasts) are often stored used in form time series 2-d spatial gridded fields. With increasing temporal resolution these products, with transition providing single deterministic fields to ensembles, their size has dramatically increased, which makes use, transfer archiving a challenge. Efficient compression such - lossy lossless is required solve this problem.The goal work was...