State-of-the-art flood hazard maps in coastal cities are often obtained from simulating coastal or pluvial events separately. This method does not account for the seasonality of flood drivers and their mutual dependence. In this article, we include the impact of these two factors in a computationally efficient probabilistic framework for flood risk calculation, using Ho Chi Minh City (HCMC) as a case study. HCMC can be flooded sub-annually by high tide, rainfall and storm surge events or a combination thereof during the monsoon or tropical cyclones. Using long gauge observations, we stochastically model 10,000 years of rainfall and sea level events based on their monthly distributions, dependence structure and co-occurrence rate. The impact from each stochastic event is then obtained from a damage function built from selected rainfall and sea level combinations, leading to an expected annual damage (EAD) of $1.02B (95^th annual damage percentile of $2.15B). We find no dependence for most months and large differences in expected damage across months ($36M-166M) driven by the seasonality of rainfall and sea levels. Excluding monthly variability leads to a serious underestimation of the EAD by 72% to 83%. This is because high-probability flood events, which can happen multiple times during the year and are properly captured by our framework, contribute the most to the EAD. This application illustrates the potential of our framework and advocates for the inclusion of flood drivers’ dynamics in coastal risk assessments.

Operation feature extraction of flood discharge structures under ambient excitation has attracted increasing attention in recent years. However, the vibration signal of flood discharge structures is a nonstationary random signal with low signal-to-noise ratio, which is mixed with lots of low-frequency water flow noise and high-frequency white noise. It is difficult to excavate the hidden vibration characteristic information accurately. To solve the problem, we propose a novel denoising method called improved variational mode decomposition. As an improved method of variational mode decomposition, improved variational mode decomposition can effectively determine the decomposition mode number of variational mode decomposition by using the mutual information method. Furthermore, improved variational mode decomposition is combined with a variance dedication rate to extract the overall operation characteristic information of the structure. In order to evaluate the applicability and effectiveness of the proposed improved variational mode decomposition–variance dedication rate method, we compare the denoising results of simulation signals produced by an improved variational mode decomposition–variance dedication rate with those produced by digital filter, wavelet threshold, empirical mode decomposition, empirical wavelet transform, complete ensemble empirical mode decomposition with adaptive noise, and improved variational mode decomposition methods and find a better performance of the improved variational mode decomposition–variance dedication rate method. In addition, the proposed method is applied to the Three Gorges Dam, and the results show that the improved variational mode decomposition–variance dedication rate method can effectively remove heavy background noises and extract the operation characteristic information of the flood discharge structure, which contributes to health monitoring and damage identification of the flood discharge structure.

Non-point source pollution from excessive use of fertilizers in agriculture is a major cause of the eutrophication problem in China. Understanding farmers’ decision-making concerning fertilization and identifying the influencing factors in this process are key to tackling overfertilization and related pollution issues. This paper reports a study on modelling decisions about fertilizer use based on data collected from 200 farmer households in the Three Gorges Reservoir area of China, using a well-fitted artificial neural network (ANN) with incorporated variance-based sensitivity analysis. The rate of fertilizer use estimated from the model is in good agreement with observed data. The model is further validated and tested by comparing the simulated and observed values. Results show that the model is able to identify the influencing factors and their interactions causing the variation in fertilizer use and to help pinpoint the underlying reasons. It is found that the farmers’ fertilization behavior is greatly affected by the area of cultivated land, followed by the interaction among farmers’ education level, annual income, and awareness of the importance of environmental protection. Future land consolidation is one of several ways to achieve more sustainable fertilization strategies.

There are many mountain torrent disasters caused by melting icebergs and snow in Xinjiang, which are very different from traditional mountain torrent disasters. Most of the areas affected by snowmelt are in areas without data, making it very difficult to predict and warn of disasters. Taking the Lianggoushan watershed at the southern foot of Boroconu Mountain as the research subject, the key factors were screened by Pearson correlation coefficient and the factor analysis method, and the data of rainfall, water level, temperature, air pressure, wind speed, and snow depth were used as inputs, respectively, with support vector regression (SVR), random forest (RF), k-nearest neighbor (KNN), artificial neural network (ANN), recurrent neural network (RNN), and long short-term memory neural network (LSTM) models used to simulate the daily average water level at the outlet of the watershed. The research results showed that the root mean square error (RMSE) values of SVR, RF, KNN, ANN, RNN, and LSTM in the training period were 0.033, 0.012, 0.016, 0.022, 0.011, and 0.010, respectively, and in the testing period they were 0.075, 0.072, 0.071, 0.075, 0.075, and 0.071, respectively. The performance of LSTM was better than that of other models, but it had more hyperparameters that needed to be optimized. The performance of RF was second only to LSTM; it had only one hyperparameter and was very easy to determine. The RF model showed that the simulation results mainly depended on the average wind speed and average sea level pressure data. The snowmelt model based on machine learning proposed in this study can be widely used in iceberg snowmelt warning and forecasting in ungauged areas, which is of great significance for the improvement of mountain flood prevention work in Xinjiang.

This paper presents the spatial variation of area‐orientated annual maximum daily rainfall (AMDR), represented by well‐fitted generalized extreme value (GEV) distributions, over the last century in Great Britain (GB) and Australia (AU) with respect to three spatial properties: geographic locations, sizes, and shapes of the region‐of‐interest (ROI). The results show that the spatial variation of GEV location‐scale parameters is dominated by geographic locations and area sizes. In GB, there is an eastward‐decreasing banded pattern compared with a concentrically increasing pattern from the middle to coasts in AU. The parameters tend to decrease with increased area sizes in both studied regions. Although the impact of the ROI shapes is insignificant, the round‐shaped regions usually have higher‐valued parameters than the elongated ones. These findings provide a new perspective to understand the heterogeneity of extreme rainfall distribution over space driven by the complex interactions between climate, geographical features, and the practical sampling approaches.