Explore the vast range of titles available.

Mountain flood in small watershed is widely distributed disaster, which have the characteristics of strong suddenness, great harm, and frequently. The traditional hydrodynamic and manual forecasting methods have high error rates for hourly forecasting. In order to improve the accuracy and real-time of water level forecasting in small watershed, we extract effective disaster-causing information, integrate multi-dimensional disaster-causing factors (such as hydrology, meteorology, geography, etc.), use a short-term prediction window and loop multi-step input method to improve the Machine Learning (ML) regression models’ accuracy, which can reduce the ML model’s process error. The non-ensemble and ensemble ML regression models is constructed for forecasting by loop multi-step, the non-ensemble models including Linear Regression (LR), Support Vector Machine Regression (SVMR) and k -Nearest Neighbors Regression ( k- NNR), and the ensemble ML models include Random Forest Regression (RFR) and Gradient Boosting Regression (GBR). The loop multi-step ensemble ML regression models have the characteristics of high accurate and low time consumption than the general ML regression models for mountain flood forecasting in small watershed.

Seasonal snow is an essential water resource in many mountain regions. However, the spatio-temporal variability in mountain snow depth or snow water equivalent (SWE) at regional to global scales is not well understood due to the lack of high-resolution satellite observations and robust retrieval algorithms. We investigate the ability of the Sentinel-1 mission to monitor snow depth at sub-kilometer (100 m, 500 m, and 1 km) resolutions over the European Alps for 2017–2019. The Sentinel-1 backscatter observations, especially in cross-polarization, show a high correlation with regional model simulations of snow depth over Austria and Switzerland. The observed changes in radar backscatter with the accumulation or ablation of snow are used in an empirical change detection algorithm to retrieve snow depth. The algorithm includes the detection of dry and wet snow conditions. Compared to in situ measurements at 743 sites in the European Alps, dry snow depth retrievals at 500 m and 1 km resolution have a spatio-temporal correlation of 0.89. The mean absolute error equals 20 %–30 % of the measured values for snow depths between 1.5 and 3 m. The performance slightly degrades for retrievals at the finer 100 m spatial resolution as well as for retrievals of shallower and deeper snow. The results demonstrate the ability of Sentinel-1 to provide snow estimates in mountainous regions where satellite-based estimates of snow mass are currently lacking. The retrievals can improve our knowledge of seasonal snow mass in areas with complex topography and benefit a number of applications, such as water resource management, flood forecasting, and numerical weather prediction. However, future research is recommended to further investigate the physical basis of the sensitivity of Sentinel-1 backscatter observations to snow accumulation.

Destructive and costly flooding can occur when warm storm systems deposit substantial rain on extensive snowcover1–6, as observed in February 2017 with the Oroville Dam crisis in California7. However, decision-makers lack guidance on how such rain-on-snow (ROS) flood risk may respond to climate change. Here, daily ROS events with flood-generating potential8 are simulated over western North America for a historical (2000–2013) and future (forced under Representative Concentration Pathway 8.59) period with the Weather Research and Forecasting model; 4 km resolution allows the basin-scale ROS flood risk to be assessed. In the warmer climate, we show that ROS becomes less frequent at lower elevations due to snowpack declines, particularly in warmer areas (for example, the Pacific maritime region). By contrast, at higher elevations where seasonal snowcover persists, ROS becomes more frequent due to a shift from snowfall to rain. Accordingly, the water available for runoff10 increases for 55% of western North American river basins, with corresponding increases in flood risk of 20–200%, the greatest changes of which are projected for the Sierra Nevada, the Colorado River headwaters and the Canadian Rocky Mountains. Thus, flood control and water resource planning must consider ROS to fully quantify changes in flood risk with anthropogenic warming.

Numerical weather prediction (NWP) models are fundamental to extend forecast lead times beyond the concentration time of a watershed. Particularly for flash flood forecasting in tropical mountainous watersheds, forecast precipitation is required to provide timely warnings. This paper aims to assess the potential of NWP for flood early warning purposes, and the possible improvement that bias correction can provide, in a tropical mountainous area. The paper focuses on the comparison of streamflows obtained from the post-processed precipitation forecasts, particularly the comparison of ensemble forecasts and their potential in providing skilful flood forecasts. The Weather Research and Forecasting (WRF) model is used to produce precipitation forecasts that are post-processed and used to drive a hydrologic model. Discharge forecasts obtained from the hydrological model are used to assess the skill of the WRF model. The results show that post-processed WRF precipitation adds value to the flood early warning system when compared to zero-precipitation forecasts, although the precipitation forecast used in this analysis showed little added value when compared to climatology. However, the reduction of biases obtained from the post-processed ensembles show the potential of this method and model to provide usable precipitation forecasts in tropical mountainous watersheds. The need for more detailed evaluation of the WRF model in the study area is highlighted, particularly the identification of the most suitable parameterisation, due to the inability of the model to adequately represent the convective precipitation found in the study area.

The atmospheric–hydrological coupling systems are essential to flood forecasting because they allow for more improved and comprehensive prediction of flood events with an extended forecast lead time. Achieving this goal requires a reliable hydrological model system that enhances both rainfall predictions and hydrological forecasts. This study evaluates the potential of coupling the mesoscale numerical weather prediction model, i.e., the weather research and forecasting (WRF) model, with different hydrological modeling systems to improve the accuracy of flood simulation. The fully distributed WRF-Hydro modeling system and the semi-distributed Hydrological Engineering Center Hydrological Modeling System (HEC-HMS) were coupled with the WRF model, and the lumped HEC-HMS model was also adopted using the observed gauge precipitation as a benchmark to test the model uncertainty. Four distinct storm events from two mountainous catchments in northern China characterized by varying spatial and temporal rainfall patterns were selected as case studies. Comparative analyses of the simulated flooding processes were carried out to evaluate and compare the performance of the coupled systems with different complexities. The coupled WRF–HEC-HMS system performed better for long-duration storm events and obtained optimal performance for storm events uniformly distributed both temporally and spatially, as it adapted to more rapid recession processes of floods. However, the coupled WRF–HEC-HMS system did not adequately capture the magnitude of the storm events as it had a larger flow peak error. On the other hand, the fully distributed WRF–WRF-Hydro system performed better for shorter-duration floods with higher flow peaks as it can adapt to the simulation of flash floods. However, the performance of the system became poor as uniformity decreased. The performance of the lumped HEC-HMS indicates some source of uncertainty in the hydrological model when compared with the coupled WRF–HEC-HMS system, but a larger magnitude error was found in the WRF output rainfall. The results of this study can help establish an adaptive atmospheric–hydrologic coupling system to improve flood forecasting for different watersheds and climatic characteristics.

Flash flood disaster is a prominent issue threatening public safety and social development throughout the world, especially in mountainous regions. Rainfall threshold is a widely accepted alternative to hydrological forecasting for flash flood warning due to the short response time and limited observations of flash flood events. However, determination of rainfall threshold is still very complicated due to multiple impact factors, particular for antecedent soil moisture and rainfall patterns. In this study, hydrological simulation approach (i.e., China Flash Flood-Hydrological Modeling System: CNFF-HMS) was adopted to capture the flash flood processes. Multiple scenarios were further designed with consideration of antecedent soil moisture and rainfall temporal patterns to determine the possible assemble of rainfall thresholds by driving the CNFF-HMS. Moreover, their effects on rainfall thresholds were investigated. Three mountainous catchments (Zhong, Balisi and Yu villages) in southern China were selected for case study. Results showed that the model performance of CNFF-HMS was very satisfactory for flash flood simulations in all these catchments, especially for multimodal flood events. Specifically, the relative errors of runoff and peak flow were within ± 20%, the error of time to peak flow was within ± 2 h and the Nash–Sutcliffe efficiency was greater than 0.90 for over 90% of the flash flood events. The rainfall thresholds varied between 93 and 334 mm at Zhong village, between 77 and 246 mm at Balisi village and between 111 and 420 mm at Yu village. Both antecedent soil moistures and rainfall temporal pattern significantly affected the variations of rainfall threshold. Rainfall threshold decreased by 8–38 and 0–42% as soil saturation increased from 0.20 to 0.50 and from 0.20 to 0.80, respectively. The effect of rainfall threshold was the minimum for the decreasing hyetograph (advanced pattern) and the maximum for the increasing hyetograph (delayed pattern), while it was similar for the design hyetograph and triangular hyetograph (intermediate patterns). Moreover, rainfall thresholds with short time spans were more suitable for early flood warning, especially in small rural catchments with humid climatic characteristics. This study was expected to provide insights into flash flood disaster forecasting and early warning in mountainous regions, and scientific references for the implementation of flash flood disaster prevention in China.

Rainfall forecasting is a high-priority research problem due to the complex interplay of multiple factors. Despite extensive studies, a systematic quantitative review of recent developments in rainfall forecasting is lacking in the literature. This study conducted a systematic quantitative review of statistical, numerical weather prediction (NWP) and machine learning (ML) techniques for rainfall forecasting. The review adopted the preferred reporting items for systematic reviews and meta-analyses (PRISMA) technique for screening keywords and abstracts, leading to 110 qualified papers from multiple databases. The impact of rainfall threshold, meteorological parameters, topography, algorithm techniques, geographic location, the horizontal resolution of the model, and lead time on rainfall forecast was examined. The review shows the importance of precipitable water vapor (PWV) along with other meteorological parameters for accurate nowcasting in coastal and mountainous regions. An increase in rainfall forecast uncertainty with an increase in the lead time makes the NWP model less popular for the short-term forecast. The pre-processing techniques increased the accuracy of ML techniques by considering extreme values and detecting the irregularly distributed multi-scale features of rainfall in space and time. Future research can focus on hybrid models with improved accuracy for nowcasting. The output from the hybrid model serves as input for the decision support system required for urban flood risk management.

Precipitation prediction in mountainous regions is one of the most challenging topics in numerical weather prediction (NWP) models. This study aims to compare two types of NWP models: the General Circulation Model–Global Forecast System (GCM–GFS) and the mesoscale Weather Research and Forecasting (WRF). The comparison is based on various early lead-times (1, 3, and 5 days) in precipitation prediction and, consequently, flood forecasting in the northern parts of the Zagros Mountains, Iran. For this purpose, five observational flood events were selected in the region. To optimize the WRF model’s configuration, twelve setups were tested by combining microphysical and planetary boundary layer schemes. The Morrison and YSU schemes demonstrated superior performance in precipitation prediction. Comparative analysis of WRF and GFS model outputs revealed WRF’s better performance in point analysis using the nearest-neighbors method, while GFS exhibited greater reliability for mean areal precipitation. Subsequently, flood simulation was performed using the HEC-HMS model. Precipitation predicted by the GFS and WRF models was introduced to the HEC-HMS model in three early lead-times for flood forecast in all three domains of 3, 9, and 27 km. The results showed that in both the precipitation forecast and flood hydrographs produced by the HEC-HMS model, in most cases, the forecasting performance decreased with increasing early lead-time. Overall, based on the results of this study, the third domain of the mesoscale WRF model did not demonstrate significant added value over GFS outputs in most events. This underscores the necessity of focusing on reducing uncertainties and applying bias correction to the model outputs before their use in hydrological simulations, particularly in regions with complex topography.

Typhoon‐induced rainfall can trigger floods and landslides that pose significant hazards in Zhejiang. However, accurately forecasting its magnitude remains a significant challenge for global Numerical Weather Prediction (NWP) models, particularly in mountainous regions where complex orographic effects play a critical role. Using forecast data from three global NWP models (ECMWF, NCEP, and CMA‐GFS) and observed station data from 2021 to 2024, this study revealed that these models consistently underestimated heavy rainfall in high‐altitude areas, associated with their limited resolution in representing terrain‐induced amplification. To address this, the terrain correction method proposed by Xu et al. (2019) was applied to the models to improve the estimation of typhoon‐induced orographic rainfall in Zhejiang. Significant improvements in forecast skill were evidenced by increased Threat Scores (TS) and Probabilities of Detection (POD). In the ECMWF model, TS for rainstorms (≥ 50 mm·day−1) increased from 0.33 to 0.35, while POD rose from 0.54 to 0.68. Larger gains were observed for heavy downpours (≥ 250 mm·day−1), with TS rising from 0.02 to 0.08 and POD from near zero to 0.34. Similar improvements were found in the NCEP and CMA‐GFS models. This study also discussed the limitations of terrain correction and identified two scenarios in which it underperformed. One occurred when the original forecast overestimated rainfall due to excessive moisture flux convergence, sometimes further amplifying errors (e.g., the forecast initialized at 00:00 UTC 24 July 2021). The other involved spatial displacement of the predicted rainfall field due to typhoon track errors, resulting in poor alignment with observations (e.g., 00:00 UTC 12 September 2021). Despite these limitations, the terrain correction notably improved forecasting skills, as indicated by TS and POD metrics, thereby enhancing local preparedness against typhoon‐induced heavy rainfall and helping mitigate the risks of flooding and other related hazards in Zhejiang. A dynamic terrain correction method is applied to three global NWP models to improve typhoon rainfall predictions over Zhejiang's complex terrain. The correction significantly enhances forecast skill, particularly for heavy rainfall with daily magnitudes exceeding 100 and 250 mm, and its practical utility and limitations are also discussed.

Flash floods and extreme rains are destructive phenomena and difficult to forecast. In 2011, the mountainous region of Rio de Janeiro state suffered one of the largest natural hazards in Brazil, affecting more than 300,000 people, leaving more than 900 dead. This article simulates this natural hazard through Quantitative Precipitation Forecasting (QPF) and streamflow forecast ensemble, using 18 combinations of parameterizations between cumulus, microphysics, surface layer, planetary boundary layer, land surface and lateral contour conditions of the Weather Research and Forecasting (WRF) Model, coupling to the Soil Moisture Accounting Procedure (SMAP) hydrological model, seeking to find the best set of parametrizations for the forecasting of extreme events in the region. The results showed rainfall and streamflow forecast were underestimated by the models, reaching an error of 57.4% to QPF and 24.6% error to streamflow, and part of these errors are related to the lack of skill of the atmospheric model in predicting the intensity and the spatial-temporal distribution of rainfall. These results bring to light the limitations of numerical weather prediction, possibly due to the lack of initiatives involving the adaptation of empirical constants, intrinsic in the parametrization models, to the specific atmospheric conditions of each region of the country.