@article{BubeckBerghaeuserHudsonetal.2020, author = {Bubeck, Philip and Bergh{\"a}user, Lisa and Hudson, Paul and Thieken, Annegret}, title = {Using panel data to understand the dynamics of human behavior in response to flooding}, series = {Risk analysis : an international journal}, volume = {40}, journal = {Risk analysis : an international journal}, number = {11}, publisher = {Wiley}, address = {Hoboken}, issn = {0272-4332}, doi = {10.1111/risa.13548}, pages = {2340 -- 2359}, year = {2020}, abstract = {Insights into the dynamics of human behavior in response to flooding are urgently needed for the development of effective integrated flood risk management strategies, and for integrating human behavior in flood risk modeling. However, our understanding of the dynamics of risk perceptions, attitudes, individual recovery processes, as well as adaptive (i.e., risk reducing) intention and behavior are currently limited because of the predominant use of cross-sectional surveys in the flood risk domain. Here, we present the results from one of the first panel surveys in the flood risk domain covering a relatively long period of time (i.e., four years after a damaging event), three survey waves, and a wide range of topics relevant to the role of citizens in integrated flood risk management. The panel data, consisting of 227 individuals affected by the 2013 flood in Germany, were analyzed using repeated-measures ANOVA and latent class growth analysis (LCGA) to utilize the unique temporal dimension of the data set. Results show that attitudes, such as the respondents' perceived responsibility within flood risk management, remain fairly stable over time. Changes are observed partly for risk perceptions and mainly for individual recovery and intentions to undertake risk-reducing measures. LCGA reveal heterogeneous recovery and adaptation trajectories that need to be taken into account in policies supporting individual recovery and stimulating societal preparedness. More panel studies in the flood risk domain are needed to gain better insights into the dynamics of individual recovery, risk-reducing behavior, and associated risk and protective factors.}, language = {en} } @misc{SchmidtHesseAttingeretal.2020, author = {Schmidt, Lennart and Heße, Falk and Attinger, Sabine and Kumar, Rohini}, title = {Challenges in applying machine learning models for hydrological inference: a case study for flooding events across Germany}, series = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, journal = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, number = {5}, issn = {1866-8372}, doi = {10.25932/publishup-52384}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-523843}, pages = {12}, year = {2020}, abstract = {Machine learning (ML) algorithms are being increasingly used in Earth and Environmental modeling studies owing to the ever-increasing availability of diverse data sets and computational resources as well as advancement in ML algorithms. Despite advances in their predictive accuracy, the usefulness of ML algorithms for inference remains elusive. In this study, we employ two popular ML algorithms, artificial neural networks and random forest, to analyze a large data set of flood events across Germany with the goals to analyze their predictive accuracy and their usability to provide insights to hydrologic system functioning. The results of the ML algorithms are contrasted against a parametric approach based on multiple linear regression. For analysis, we employ a model-agnostic framework named Permuted Feature Importance to derive the influence of models' predictors. This allows us to compare the results of different algorithms for the first time in the context of hydrology. Our main findings are that (1) the ML models achieve higher prediction accuracy than linear regression, (2) the results reflect basic hydrological principles, but (3) further inference is hindered by the heterogeneity of results across algorithms. Thus, we conclude that the problem of equifinality as known from classical hydrological modeling also exists for ML and severely hampers its potential for inference. To account for the observed problems, we propose that when employing ML for inference, this should be made by using multiple algorithms and multiple methods, of which the latter should be embedded in a cross-validation routine.}, language = {en} } @article{SchmidtHesseAttingeretal.2020, author = {Schmidt, Lennart and Hesse, Falk and Attinger, Sabine and Kumar, Rohini}, title = {Challenges in applying machine learning models for hydrological inference}, series = {Water resources research}, volume = {56}, journal = {Water resources research}, number = {5}, publisher = {American Geophysical Union}, address = {Washington}, issn = {0043-1397}, doi = {10.1029/2019WR025924}, pages = {10}, year = {2020}, abstract = {Machine learning (ML) algorithms are being increasingly used in Earth and Environmental modeling studies owing to the ever-increasing availability of diverse data sets and computational resources as well as advancement in ML algorithms. Despite advances in their predictive accuracy, the usefulness of ML algorithms for inference remains elusive. In this study, we employ two popular ML algorithms, artificial neural networks and random forest, to analyze a large data set of flood events across Germany with the goals to analyze their predictive accuracy and their usability to provide insights to hydrologic system functioning. The results of the ML algorithms are contrasted against a parametric approach based on multiple linear regression. For analysis, we employ a model-agnostic framework named Permuted Feature Importance to derive the influence of models' predictors. This allows us to compare the results of different algorithms for the first time in the context of hydrology. Our main findings are that (1) the ML models achieve higher prediction accuracy than linear regression, (2) the results reflect basic hydrological principles, but (3) further inference is hindered by the heterogeneity of results across algorithms. Thus, we conclude that the problem of equifinality as known from classical hydrological modeling also exists for ML and severely hampers its potential for inference. To account for the observed problems, we propose that when employing ML for inference, this should be made by using multiple algorithms and multiple methods, of which the latter should be embedded in a cross-validation routine.}, language = {en} } @article{SchmidtHesseAttingeretal.2020, author = {Schmidt, Lennart and Heße, Falk and Attinger, Sabine and Kumar, Rohini}, title = {Challenges in applying machine learning models for hydrological inference: a case study for flooding events across Germany}, series = {Water Resources Research}, volume = {56}, journal = {Water Resources Research}, number = {5}, publisher = {John Wiley \& Sons, Inc.}, address = {New Jersey}, pages = {10}, year = {2020}, abstract = {Machine learning (ML) algorithms are being increasingly used in Earth and Environmental modeling studies owing to the ever-increasing availability of diverse data sets and computational resources as well as advancement in ML algorithms. Despite advances in their predictive accuracy, the usefulness of ML algorithms for inference remains elusive. In this study, we employ two popular ML algorithms, artificial neural networks and random forest, to analyze a large data set of flood events across Germany with the goals to analyze their predictive accuracy and their usability to provide insights to hydrologic system functioning. The results of the ML algorithms are contrasted against a parametric approach based on multiple linear regression. For analysis, we employ a model-agnostic framework named Permuted Feature Importance to derive the influence of models' predictors. This allows us to compare the results of different algorithms for the first time in the context of hydrology. Our main findings are that (1) the ML models achieve higher prediction accuracy than linear regression, (2) the results reflect basic hydrological principles, but (3) further inference is hindered by the heterogeneity of results across algorithms. Thus, we conclude that the problem of equifinality as known from classical hydrological modeling also exists for ML and severely hampers its potential for inference. To account for the observed problems, we propose that when employing ML for inference, this should be made by using multiple algorithms and multiple methods, of which the latter should be embedded in a cross-validation routine.}, language = {en} } @article{SiegThieken2022, author = {Sieg, Tobias and Thieken, Annegret}, title = {Improving flood impact estimations}, series = {Environmental research letters}, volume = {17}, journal = {Environmental research letters}, number = {6}, publisher = {IOP Publ. Ltd.}, address = {Bristol}, issn = {1748-9326}, doi = {10.1088/1748-9326/ac6d6c}, pages = {16}, year = {2022}, abstract = {A reliable estimation of flood impacts enables meaningful flood risk management and rapid assessments of flood impacts shortly after a flood. The flood in 2021 in Central Europe and the analysis of its impacts revealed that these estimations are still inadequate. Therefore, we investigate the influence of different data sets and methods aiming to improve flood impact estimates. We estimated economic flood impacts to private households and companies for a flood event in 2013 in Germany using (a) two different flood maps, (b) two approaches to map exposed objects based on OpenStreetMap and the Basic European Asset Map, (c) two different approaches to estimate asset values, and (d) tree-based models and Stage-Damage-Functions to describe the vulnerability. At the macro scale, water masks lead to reasonable impact estimations. At the micro and meso-scale, the identification of affected objects by means of water masks is insufficient leading to unreliable estimations. The choice of exposure data sets is most influential on the estimations. We find that reliable impact estimations are feasible with reported numbers of flood-affected objects from the municipalities. We conclude that more effort should be put in the investigation of different exposure data sets and the estimation of asset values. Furthermore, we recommend the establishment of a reporting system in the municipalities for a fast identification of flood-affected objects shortly after an event.}, language = {en} } @article{WutzlerHudsonThieken2022, author = {Wutzler, Bianca and Hudson, Paul and Thieken, Annegret}, title = {Adaptation strategies of flood-damaged businesses in Germany}, series = {Frontiers in water}, volume = {4}, journal = {Frontiers in water}, publisher = {Frontiers Media}, address = {Lausanne}, issn = {2624-9375}, doi = {10.3389/frwa.2022.932061}, pages = {13}, year = {2022}, abstract = {Flood risk management in Germany follows an integrative approach in which both private households and businesses can make an important contribution to reducing flood damage by implementing property-level adaptation measures. While the flood adaptation behavior of private households has already been widely researched, comparatively less attention has been paid to the adaptation strategies of businesses. However, their ability to cope with flood risk plays an important role in the social and economic development of a flood-prone region. Therefore, using quantitative survey data, this study aims to identify different strategies and adaptation drivers of 557 businesses damaged by a riverine flood in 2013 and 104 businesses damaged by pluvial or flash floods between 2014 and 2017. Our results indicate that a low perceived self-efficacy may be an important factor that can reduce the motivation of businesses to adapt to flood risk. Furthermore, property-owners tended to act more proactively than tenants. In addition, high experience with previous flood events and low perceived response costs could strengthen proactive adaptation behavior. These findings should be considered in business-tailored risk communication.}, language = {en} } @article{WutzlerHudsonThieken2022, author = {Wutzler, Bianca and Hudson, Paul and Thieken, Annegret}, title = {Adaptation strategies of flood-damaged businesses in Germany}, series = {Frontiers in Water}, journal = {Frontiers in Water}, publisher = {Frontiers Media SA}, address = {Lausanne, Schweiz}, issn = {2624-9375}, doi = {10.3389/frwa.2022.932061}, pages = {13}, year = {2022}, abstract = {Flood risk management in Germany follows an integrative approach in which both private households and businesses can make an important contribution to reducing flood damage by implementing property-level adaptation measures. While the flood adaptation behavior of private households has already been widely researched, comparatively less attention has been paid to the adaptation strategies of businesses. However, their ability to cope with flood risk plays an important role in the social and economic development of a flood-prone region. Therefore, using quantitative survey data, this study aims to identify different strategies and adaptation drivers of 557 businesses damaged by a riverine flood in 2013 and 104 businesses damaged by pluvial or flash floods between 2014 and 2017. Our results indicate that a low perceived self-efficacy may be an important factor that can reduce the motivation of businesses to adapt to flood risk. Furthermore, property-owners tended to act more proactively than tenants. In addition, high experience with previous flood events and low perceived response costs could strengthen proactive adaptation behavior. These findings should be considered in business-tailored risk communication.}, language = {en} } @misc{WutzlerHudsonThieken2022, author = {Wutzler, Bianca and Hudson, Paul and Thieken, Annegret}, title = {Adaptation strategies of flood-damaged businesses in Germany}, series = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, journal = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, number = {1304}, issn = {1866-8372}, doi = {10.25932/publishup-57735}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-577350}, pages = {13}, year = {2022}, abstract = {Flood risk management in Germany follows an integrative approach in which both private households and businesses can make an important contribution to reducing flood damage by implementing property-level adaptation measures. While the flood adaptation behavior of private households has already been widely researched, comparatively less attention has been paid to the adaptation strategies of businesses. However, their ability to cope with flood risk plays an important role in the social and economic development of a flood-prone region. Therefore, using quantitative survey data, this study aims to identify different strategies and adaptation drivers of 557 businesses damaged by a riverine flood in 2013 and 104 businesses damaged by pluvial or flash floods between 2014 and 2017. Our results indicate that a low perceived self-efficacy may be an important factor that can reduce the motivation of businesses to adapt to flood risk. Furthermore, property-owners tended to act more proactively than tenants. In addition, high experience with previous flood events and low perceived response costs could strengthen proactive adaptation behavior. These findings should be considered in business-tailored risk communication.}, language = {en} } @article{MacdonaldMerzGuseetal.2022, author = {Macdonald, Elena and Merz, Bruno and Guse, Bj{\"o}rn and Wietzke, Luzie and Ullrich, Sophie and Kemter, Matthias and Ahrens, Bodo and Vorogushyn, Sergiy}, title = {Event and catchment controls of heavy tail behavior of floods}, series = {Water resources research}, volume = {58}, journal = {Water resources research}, number = {6}, publisher = {American Geophysical Union}, address = {Washington}, issn = {0043-1397}, doi = {10.1029/2021WR031260}, pages = {25}, year = {2022}, abstract = {In some catchments, the distribution of annual maximum streamflow shows heavy tail behavior, meaning the occurrence probability of extreme events is higher than if the upper tail decayed exponentially. Neglecting heavy tail behavior can lead to an underestimation of the likelihood of extreme floods and the associated risk. Partly contradictory results regarding the controls of heavy tail behavior exist in the literature and the knowledge is still very dispersed and limited. To better understand the drivers, we analyze the upper tail behavior and its controls for 480 catchments in Germany and Austria over a period of more than 50 years. The catchments span from quickly reacting mountain catchments to large lowland catchments, allowing for general conclusions. We compile a wide range of event and catchment characteristics and investigate their association with an indicator of the tail heaviness of flood distributions, namely the shape parameter of the GEV distribution. Following univariate analyses of these characteristics, along with an evaluation of different aggregations of event characteristics, multiple linear regression models, as well as random forests, are constructed. A novel slope indicator, which represents the relation between the return period of flood peaks and event characteristics, captures the controls of heavy tails best. Variables describing the catchment response are found to dominate the heavy tail behavior, followed by event precipitation, flood seasonality, and catchment size. The pre-event moisture state in a catchment has no relevant impact on the tail heaviness even though it does influence flood magnitudes.}, language = {en} } @phdthesis{Banerjee2022, author = {Banerjee, Abhirup}, title = {Characterizing the spatio-temporal patterns of extreme events}, doi = {10.25932/publishup-55983}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-559839}, school = {Universit{\"a}t Potsdam}, pages = {xiv, 91}, year = {2022}, abstract = {Over the past decades, there has been a growing interest in 'extreme events' owing to the increasing threats that climate-related extremes such as floods, heatwaves, droughts, etc., pose to society. While extreme events have diverse definitions across various disciplines, ranging from earth science to neuroscience, they are characterized mainly as dynamic occurrences within a limited time frame that impedes the normal functioning of a system. Although extreme events are rare in occurrence, it has been found in various hydro-meteorological and physiological time series (e.g., river flows, temperatures, heartbeat intervals) that they may exhibit recurrent behavior, i.e., do not end the lifetime of the system. The aim of this thesis to develop some sophisticated methods to study various properties of extreme events. One of the main challenges in analyzing such extreme event-like time series is that they have large temporal gaps due to the paucity of the number of observations of extreme events. As a result, existing time series analysis tools are usually not helpful to decode the underlying information. I use the edit distance (ED) method to analyze extreme event-like time series in their unaltered form. ED is a specific distance metric, mainly designed to measure the similarity/dissimilarity between point process-like data. I combine ED with recurrence plot techniques to identify the recurrence property of flood events in the Mississippi River in the United States. I also use recurrence quantification analysis to show the deterministic properties and serial dependency in flood events. After that, I use this non-linear similarity measure (ED) to compute the pairwise dependency in extreme precipitation event series. I incorporate the similarity measure within the framework of complex network theory to study the collective behavior of climate extremes. Under this architecture, the nodes are defined by the spatial grid points of the given spatio-temporal climate dataset. Each node is associated with a time series corresponding to the temporal evolution of the climate observation at that grid point. Finally, the network links are functions of the pairwise statistical interdependence between the nodes. Various network measures, such as degree, betweenness centrality, clustering coefficient, etc., can be used to quantify the network's topology. We apply the methodology mentioned above to study the spatio-temporal coherence pattern of extreme rainfall events in the United States and the Ganga River basin, which reveals its relation to various climate processes and the orography of the region. The identification of precursors associated with the occurrence of extreme events in the near future is extremely important to prepare the masses for an upcoming disaster and mitigate the potential risks associated with such events. Under this motivation, I propose an in-data prediction recipe for predicting the data structures that typically occur prior to extreme events using the Echo state network, a type of Recurrent Neural Network which is a part of the reservoir computing framework. However, unlike previous works that identify precursory structures in the same variable in which extreme events are manifested (active variable), I try to predict these structures by using data from another dynamic variable (passive variable) which does not show large excursions from the nominal condition but carries imprints of these extreme events. Furthermore, my results demonstrate that the quality of prediction depends on the magnitude of events, i.e., the higher the magnitude of the extreme, the better is its predictability skill. I show quantitatively that this is because the input signals collectively form a more coherent pattern for an extreme event of higher magnitude, which enhances the efficiency of the machine to predict the forthcoming extreme events.}, language = {en} }