Reproducing flash flood warnings with Machine Learning

Quantifying uncertainty associated with flash flood warning or forecast systems is required to enable informed decision making by those responsible for operation and management of natural hazard protection systems. The current system used by the U.S. National Weather Service (NWS) to issue flash-flood warnings and watches over the Unites States is a purely deterministic system. The authors propose a simple approach to augment the Flash Flood Guidance System (FFGS) with uncertainty propagation components. The authors briefly discuss the main components of the system, propose changes to improve it, and allow accounting for several sources of uncertainty. They illustrate their discussion with examples of uncertainty quantification procedures for several small basins of the Illinois River basin in Oklahoma. As the current FFGS is tightly coupled with two technologies, that is, threshold-runoff mapping and the Sacramento Soil Moisture Accounting Hydrologic Model, the authors discuss both as sources of uncertainty. To quantify and propagate those sources of uncertainty throughout the system, they develop a simple version of the Sacramento model and use Monte Carlo simulation to study several uncertainty scenarios. The results point out the significance of the stream characteristics such as top width and the hydraulic depth on the overall uncertainty of the Flash Flood Guidance System. They also show that the overall flash flood guidance uncertainty is higher under drier initial soil moisture conditions. The results presented herein, although limited, are a necessary first step toward the development of probabilistic operational flash flood guidance forecast-response systems.

National Oceanic and Atmospheric Administration's National Weather Service (NWS) flash flood warnings are issued by Weather Forecast Offices and are underpinned by information from the Flash Flood Guidance (FFG) system operated by the River Forecast Centers (RFCs). This study focuses on the quantitative evaluation and limitations of the FFG system using reported flash flood cases in 2010 and 2011. The flash flood reports were obtained from the NWS Storm Event database for the Arkansas‐Red Basin RFC (ABRFC). The current FFG system at the ABRFC provides gridded flash flood guidance (GFFG) system using the NWS Hydrology Laboratory‐Research Distributed Hydrologic Model to translate the upper zone soil moisture to estimates of Soil Conservation Service Curve Numbers. Comparisons of the GFFG and real‐time Multisensor Precipitation Estimator‐derived Quantitative Precipitation Estimate for the same duration and location were used to analyze the success of the system. Typically, the six‐hour duration was characterized by higher probability of detection values than the three‐hour duration, which highlights the difficulty of hydrologic process estimation for shorter time scales. The current system does not take into account physical characteristics such as land use, including irrigated agricultural farm and urban areas, hence, overly dry soil moisture estimates over these areas can lower the success rate of the GFFG product.

This study quantifies the skill of the National Weather Service’s (NWS) flash flood guidance (FFG) product. Generated by River Forecast Centers (RFCs) across the United States, local NWS Weather Forecast Offices compare estimated and forecast rainfall to FFG to monitor and assess flash flooding potential. A national flash flood observation database consisting of reports in the NWS publication Storm Data and U.S. Geological Survey (USGS) stream gauge measurements are used to determine the skill of FFG over a 4-yr period. FFG skill is calculated at several different precipitation-to-FFG ratios for both observation datasets. Although a ratio of 1.0 nominally indicates a potential flash flooding event, this study finds that FFG can be more skillful when ratios other than 1.0 are considered. When the entire continental United States is considered, the highest observed critical success index (CSI) with 1-h FFG is 0.20 for the USGS dataset, which should be considered a benchmark for future research that seeks to improve, modify, or replace the current FFG system. Regional benchmarks of FFG skill are also determined on an RFC-by-RFC basis. When evaluated against Storm Data reports, the regional skill of FFG ranges from 0.00 to 0.19. When evaluated against USGS stream gauge measurements, the regional skill of FFG ranges from 0.00 to 0.44.

Floods are one of the most devastating natural hazards around the world, and flash floods are extremely dangerous to property and human lives. This study reviews the current flash flood forecasting in China and proposes a new methodology to develop a flash flood potential index (FFPI) based on GIS to further improve China's flash flood guidance system and national weather forecasting services. We collected datasets required to calculate the FFPI and tested a series of schemes to determine the method to calculate the FFPI over China. The result indicates that the calculated China's FFPI distribution can represent the risk of flash flood well, especially in Central and Southeast China. It is also implied that the southwest and western complex terrains are more susceptible to rain storms. Then we chose Yunnan Province as a test-bed and also yielded good results which can offer reference to the current China's flash flood forecasting system. Further research can be placed on incorporation of soil moisture and rainfall frequency components into the FFPI calculation method for dynamic warning and forecasting of flash floods.

Flash floods have become the nation's number one stormy weather killer in the 1970s. Deaths from flash floods are now approaching 200 each year compared to an average of less than 70 per year during the preceding 30-year period. Losses from flash floods are now nearly 10 times what they were in the 1940s. To combat these escalating losses the National Weather Service (NWS) is expanding its efforts to improve forecasting and warning of flash floods and to improve the public's response to flash flood threats. This paper summarizes the present and planned NWS operational flash flood warning and disaster preparedness programs.

Flash floods have the potential to damage infrastructure and buildings, and pose a considerable threat to human life. The short lag times associated with flash floods (typically a few hours) present a significant challenge to existing flood warning systems. These systems currently rely on rainfall and runoff measurements, as well as hydrological models. Consequently, they are of limited applicability in ungauged catchments. This is a critical issue given that climate change is intensifying extreme rainfall and thereby increasing the potential for flash flooding.This study investigates the detection of flash floods based solely on rainfall characteristics, thus eliminating dependency on runoff measurements and hydrological infrastructure. Using high-resolution radar rainfall data and 15-minute runoff observations, 1,330 extreme rainfall-runoff events are selected across 147 German catchments with an area of up to 100 km². These events are subsequently classified as either flash or non-flash floods using a rainfall-runoff-based classification scheme as a reference, with 103 of the selected events identified as flash floods. For each event, various space-time rainfall characteristics are quantified. A random forest model for flash flood detection is then trained using only the rainfall metrics and static catchment attributes. The main aim is to assess the potential for rainfall-driven flash flood detection without relying on runoff. In addition, the most relevant rainfall characteristics associated with flash floods are identified, thereby improving our understanding of the underlying drivers.In future work, the developed detection approach will be combined with real-time rainfall nowcasting to enable earlier prediction and warning of flash floods.

[Departement_IRSTEA]Eaux [TR1_IRSTEA]ARCEAU

Aim: The majority of machine learning (ML) models in healthcare are built on retrospective data, much of which is collected without explicit patient consent for use in artificial intelligence (AI) and ML applications. The primary aim of this study was to evaluate whether clinicians and scientific researchers themselves consent to provide their own data for the training of ML models. Materials and Methods: This survey was conducted through an anonymous online survey, utilizing platforms such as Telegram, LinkedIn, and Viber. The target audience comprised specific international groups, primarily Russian, German, and English-speaking, of clinicians and scientific researchers. These participants ranged in their levels of expertise and experience, from beginners to veterans. The survey centered on a singular, pivotal question: “Do You Consent to the Use of Your Biological and Private Data for Training Machine Learning and AI Models?” Respondents had the option to choose from three responses: “Yes” and “No”. Results: The survey was conducted in January 2024. A total of 119 unique and verified individuals participated in the survey. The results revealed that only 50% of respondents (63 persons) expressed consent to provide their own data for the training of ML and AI models. Conclusion: In the development of ML and AI models, particularly open-source ones, it is crucial to ascertain whether participants are willing to provide their private data. While ML algorithms can transform the nature of data, it is important to remember that the primary owner of this data is the individual. Our findings show that in 50% of the cases, even participants from scientific research and clinical backgrounds – individuals typically accountable for ensuring data quality in AI and ML model development – do not consent to the use of their data in AI and ML settings. This highlights the need for more stringent consent processes and ethical considerations in the utilization of personal data in AI and ML research.

Flash floods occur frequently and distribute widely in mountainous areas because of complex geographic and geomorphic conditions and various climate types. Effective flash flood forecasting with useful lead times remains a challenge due to its high burstiness and short response time. Recently, machine learning has led to substantial changes across many areas of study. In hydrology, the advent of novel machine learning methods has started to encourage novel applications or substantially improve old ones. This study aims to establish a discharge forecasting model based on Long Short-Term Memory (LSTM) networks for flash flood forecasting in mountainous catchments. The proposed LSTM flood forecasting (LSTM-FF) model is composed of T multivariate single-step LSTM networks and takes spatial and temporal dynamics information of observed and forecast rainfall and early discharge as inputs. The case study in Anhe revealed that the proposed models can effectively predict flash floods, especially the qualified rates (the ratio of the number of qualified events to the total number of flood events) of large flood events are above 94.7% at 1–5 h lead time and range from 84.2% to 89.5% at 6–10 h lead-time. For the large flood simulation, the small flood events can help the LSTM-FF model to explore a better rainfall-runoff relationship. The impact analysis of weights in the LSTM network structures shows that the discharge input plays a more obvious role in the 1-h LSTM network and the effect decreases with the lead-time. Meanwhile, in the adjacent lead-time, the LSTM networks explored a similar relationship between input and output. The study provides a new approach for flash flood forecasting and the highly accurate forecast contributes to prepare for and mitigate disasters.

Machine learning-based water quality prediction using octennial in-situ Daphnia magna biological early warning system data

Machine Learning Can Accurately Predict Overnight Stay, Readmission, and 30-Day Complications Following Anterior Cruciate Ligament Reconstruction

Flash flood mitigation: recommendations for research and applications

Study question Can a machine learning (ML) model, developed using modern neural network architecture produce comparable annotation data; utilisable for algorithmic outcome prediction, to manual time-lapse annotations? Summary answer The model automatically annotated unseen embryos with comparable results to manual methods, generating morphokinetic data to enable comparably predictive outputs from an embryo selection algorithm. What is known already The application of artificial intelligence across healthcare industries, including fertility, is increasing. Several ML models are available that seek to generate or analyse embryo images and morphokinetic data, and to determine embryo viability potential. Along with photographic images, the use of time-lapse in IVF laboratories has amassed numeric data, resulting predominantly from annotated manual assessment of images over time. Embryo annotation practice is variable in quality, can be subjective and is time-consuming; commonly taking several minutes per embryo. The development of rapid, accurate automatic annotation would represent a significant time-saving as well as an increase in reproducibility and accuracy. Study design, size, duration Multicentre quality assured annotation data from 63,383 time-lapse monitored embryos (EmbryoScope®), comprising over 400 million individual images, were used to train a ML model to automatically generate morphokinetic annotations. Data was derived from 8 UK clinics within a cohesive group between 2012-2021. Accuracy was assessed using 900 unseen embryos (with live birth outcome) by comparing the output of an established in-house, prospectively validated embryo selection model when the input was either ML-automated, or manual annotations. Participants/materials, setting, methods Multi-focal plane images were processed on the Azure cloud (Microsoft) and resampled to 300x300 pixels. A Laplacian-based focal stacking algorithm merged frames into a single image. The model consisted of an EfficientNetB4 Convolutional Neural Network classifier to extract features and classify the stage of embryo images. A Temporal Convolutional Network interpreted a time-series of image features; producing annotations from pronuclear fading through to blastocyst. Soft localisation loss function used QA data to integrate annotation subjectivities. Main results and the role of chance The ML model rapidly and automatically generated annotations. Efficacy and comparability of the ML model to automate reliable, utilisable annotations was demonstrated by comparison with manual annotation data and the ML model’s ability to auto-generate annotations which could be used to predict live birth by providing annotation data to an established, validated in house embryo selection model. Live birth-predictive capability was measured, and benchmarked against manual annotation, using the area under the receiver operating characteristic curve (AUC). When tested on time-lapse images, collected from pronuclear fading to full blastulation, representing 900 previously unseen, transferred blastocysts where live birth outcomes were blinded, the in-house developed auto-annotation ML model resulted in an AUC of 0.686 compared with 0.661 for manual annotations, for live birth prediction. Auto annotation using the developed model took only milliseconds to complete per embryo. The developed auto-annotation model, built and tested on large data, is considered suitable for productionisation with the aim of being validated and integrated into an application to support IVF laboratory practice. Limitations, reasons for caution Whilst this model was trained to recognise key morphokinetic events, there are other morphokinetic variables that may be useful in the prediction of live birth and further improve embryo selection, or deselection, ability. Akin to manual interpretation, some embryos may fail to be annotated or need second opinion. Wider implications of the findings There is increasing evidence supporting the application of ML to utilise big data from time-lapse imaging and fertility care generally. Whilst promising benefits to IVF clinics and patients, responsible use of data is required alongside large high-quality datasets, and rigorous validation, to ensure safe and robust applications. Trial registration number N/A

The objective of this document is to identify the most accurate machine learning (ML) models for predicting the value of a real estate property, based on a systematic literature review (SLR). It was conducted on research published between 2022 and 2023 that analyzed the accuracy of ML models in real estate valuation. Information was extracted regarding the ML models used, the databases employed, and the models highlighted for their accuracy. A variety of ML models used in real estate valuation were identified, including Random Forest (RF), XGBoost, Gradient Boosting Machine (GBM), Linear Regression (LR), and Lasso Regression. RF and LR models stood out as the most accurate in the analyzed research, showing that the accuracy of ML models varies depending on the database, property characteristics, and valuation context. It was concluded that ML models like RF and LR are promising tools for improving the accuracy of real estate valuation. Additionally, the choice of the best ML model depends on factors such as the database, property characteristics, and specific valuation objectives. Further research is needed on the application of ML in real estate valuation, considering aspects such as model interpretability and the real estate market context.

Metabolic dysfunction-associated steatotic liver disease (MASLD) can progress to metabolic dysfunction-associated steatohepatitis (MASH) and liver fibrosis, contributing to a heavier global health burden. Non-invasive diagnostic tools developed using machine learning (ML) and deep learning (DL), two representative artificial intelligence algorithms, are increasingly being explored for MASH and its related fibrosis assessment. This study aimed to compare the diagnostic performance of different ML and DL models and identify the top-performing models for diagnosing MASH and associated liver fibrosis. A systematic review and meta-analysis were conducted across PubMed, Web of Science, Embase and Cochrane Library from inception to May 18, 2025. Pooled area under the receiver operator characteristic curve (AUROC) values with 95 % confidence interval (CI) were calculated. Accuracy, specificity, sensitivity, positive predictive values, and negative predictive values were also recorded. Of 4,314 studies initially identified, 106 met the inclusion criteria, with 35 studies (ML: n=28; DL: n=7) providing data for analysis. Logistic Regression and Neural Network are the most commonly algorithms applied in ML and DL, respectively. The pooled AUROCs for diagnosing MASH were 0.833 (95 %CI: 0.806-0.860) for ML models and 0.841 (95 %CI: 0.782-0.900) for DL models. Light Gradient Boosting Machine (LightGBM) and ResNet50 were the best-performing models for diagnosing MASH within ML and DL algorithms, respectively, achieving corresponding AUROCs of 0.920 (95 %CI: 0.916-0.924) and 0.960 (95 %CI: 0.951-0.969). For fibrosis diagnosis, ML models had a pooled AUROC of 0.826 (95 %CI: 0.792-0.860), with Categorical Boosting (CatBoost) achieving the highest AUROC of 0.960 (95 %CI: 0.950-0.970). DL models yielded the pooled AUROC of 0.875 (95 %CI: 0.816-0.934) for fibrosis diagnosis. Both ML and DL models demonstrated strong diagnostic performance for MASH and liver fibrosis, with DL achieving marginally higher AUROCs. AI-driven approaches show promise in MASLD management.