Search Results (26)

The Hengduan Mountains region (HMR) is vulnerable to flash flood disasters, which account for the largest proportion of flood-related fatalities in China. Flash flood regionalization, which divides a region into homogeneous subdivisions based on flash flood-inducing factors, provides insights for the spatial distribution patterns of flash flood risk, especially in ungauged areas. However, existing methods for flash flood regionalization have not fully reflected the spatial topology structure of the inputted geographical data. To address this issue, this study proposed a novel framework combining a state-of-the-art unsupervised Graph Neural Network (GNN) method, Dink-Net, and Shapley Additive exPlanations (SHAP) for flash flood regionalization in the HMR. A comprehensive dataset of flash flood inducing factors was first established, covering geomorphology, climate, meteorology, hydrology, and surface conditions. The performances of two classic machine learning methods (K-means and Self-organizing feature map) and three GNN methods (Deep Graph Infomax (DGI), Deep Modularity Networks (DMoN), and Dilation shrink Network (Dink-Net)) were compared for flash-flood regionalization, and the Dink-Net model outperformed the others. The SHAP model was then applied to quantify the impact of all the inducing factors on the regionalization results by Dink-Net. The newly developed framework captured the spatial interactions of the inducing factors and characterized the spatial distribution patterns of the factors. The unsupervised Dink-Net model allowed the framework to be independent from historical flash flood data, which would facilitate its application in ungauged mountainous areas. The impact analysis highlights the significant positive influence of extreme rainfall on flash floods across the entire HMR. The pronounced positive impact of soil moisture and saturated hydraulic conductivity in the areas with a concentration of historical flash flood events, together with the positive impact of topography (elevation) in the transition zone from the Qinghai–Tibet Plateau to the Sichuan Basin, have also been revealed. The results of this study provide technical support and a scientific basis for flood control and disaster reduction measures in mountain areas according to local inducing conditions. Full article

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. Full article

This paper describes the modelling of suspended sediment yield in a plains region in the European part of Russia (EPR) and its prediction for ungauged catchments. The studied plains area, excluding the Caucasus and Ural Mountains, covers 3.5 × 10⁶ km² of the total area of about 3.8 × 10⁶ km². Multiple regression methods, such as a generalized linear model (GLM) and a generalized additive model (GAM), are used to construct the models. The research methodology is based on a catchment approach. There are 49,516 river basins with an average area of about 75 km² in the plain regions. The suspended sediment yield geodatabase contains data from 385 gauging stations. The linear GLM model of suspended sediment yield explains about 50% and the GAM model about 65% of the data variability (R-squared adjusted). The models include mean slope steepness, percentage of arable land, runoff per unit area, catchment area, soil rank and catchment soil erodibility as significant predictors. They also include a zonal-sectoral gradient (the sum of active temperatures and the standard deviation of air temperature, or directly by geographic coordinates). A GAM model is trained to predict suspended sediment yields for unexplored areas of the area. The paper presents the results of extrapolating suspended sediment yield values to ungauged river basins in a plains region of the EPR. For the first time for such a large area, the models built and the use of the basin approach made it possible to predict runoff values for hydrologically unexplored river basins. Full article

Many applications of water resources planning and management depend on continuous streamflow predictions. A lack of data sources makes it difficult to predict stream flows in many world regions, including Saudi Arabia. Therefore, using simple, parsimonious models is more attractive in areas where data is scarce since they contain few parameters and require minimal input data. This study investigates the ability of simple, parsimonious water balance model models to simulate monthly time series of stream flows for poorly gauged catchments. The modified Schreiber’s empirical model and SIXPAR monthly water balance model were applied to simulate monthly streamflow in six mountainous watersheds located southwest of Saudi Arabia. The SIXPAR model was calibrated on one single gauged catchment where adequate hydrological data were available. The calibrated parameters were then transferred to the ungauged catchments based on transferring information using a physical similarity approach to regionalization. The results show that the simplified Schreiber’s model was found to consistently underestimates the monthly discharge, especially at low and moderate flow. The monthly water balance model SIXPAR based on the regionalization approach was found more capable of producing the monthly streamflow at the ungauged site under all flow conditions. This study’s finding agrees with other studies conducted in the same area using different modeling approaches. Full article

Accurate precipitation data are essential for understanding hydrological processes in high mountainous regions with limited observations and highly variable precipitation events. While frequency-corrected precipitation data are expected to aid in understanding hydrological processes, its performance in ungauged high mountain hydrological simulation remains unclear. To clarify this issue, we conducted a numerical experiment that used reanalysis precipitation, frequency-corrected precipitation, and gridded precipitation to drive a distributed cold region hydrological model. We selected an ungauged basin in high mountain Asia (Manas River Basin in China) as the study area and employed a statistical parameter optimization method to avoid subjectivity in parameter selection. Our findings indicate that the frequency information from the few existing stations can aid in correcting the reanalysis precipitation data. The frequency correction approach can reduce the total volume of errors in the reanalysis precipitation data, especially when severe biases occur. Our findings show that frequency-corrected precipitation performs better in modeling discharge, runoff depth, and evaporation. Furthermore, the improvement in precipitation using frequency correction bears clear altitude differences, which implies that having more stations at different altitudes is necessary to measure precipitation accurately in similar areas. Our study provides a feasible flow for future precipitation preparation for similar ungauged high mountain areas. Frequency correction, instead of direct interpolation, may be a viable option for precipitation preparation. Our work has reference implications for future hydrological simulations in similar ungauged high mountains. Full article

Obtaining accurate precipitation data in mountainous regions is important but challenging. In ungauged areas, remotely sensed precipitation products are useful supplements and alternatives to measured precipitation products. However, their ability to detect solid precipitation and light precipitation in mountain areas is still unclear. The primary objective of this study is to evaluate two satellite precipitation products, Global Precipitation Measurement (GPM) and Tropical Precipitation Measuring Mission (TRMM), in the headwaters of an inland river on the northeastern Tibetan Plateau (the Heihe river basin), with a specific focus on their performance regarding light precipitation and solid precipitation. The achieved results reveal that both GPM and TRMM perform poorly over the Heihe river basin, with low Correlation Coefficient value and Critical Success Index value, particularly in winter. Based on the coupled Time-Variant Gain Model-Degree Day Factor Model (TVGM-DDF) initiated in this paper, the GPM is more applicable in terms of running hydrological models. With the aim of detecting solid precipitation, the GPM is more capable of detecting solid precipitation but still unsatisfactory at two stations. In the case of light precipitation, both products underestimate light precipitation. In general, the performance of the two products in the Heihe river basin is not satisfactory and should be enhanced in upcoming explorations. This study provides a strong foundation for choosing alternate precipitation data for related research in the mountain basin. Full article

Lake Ulansuhai, one of the main water sources for semi-arid areas of China, has a local deposit caused by soil erosion during past decades. However, a lack of monitor stations prevents better estimation of soil erosion levels. Therefore, we try to estimate soil erosion in the Huangtuyaozi (HTYZ) watershed, an ungauged small watershed of the lake’s eastern watershed, by using the revised universal soil loss equation (RUSLE) model and multi-source remote sensing data, and analyze its key drivers and effect on the lake siltation. The result showed that the soil erosion rate in the HTYZ watershed ranged from 0 to 129.893 t ha⁻¹ yr⁻¹ with an average of 6.45 t ha⁻¹ yr⁻¹ during 1986–2015. In particular, 80.06% of the area was less than 10 t ha⁻¹ yr⁻¹, and just 0.06% was over 50 t ha⁻¹ yr⁻¹, mainly in the mountain area, the southern part of the HTYZ watershed. Moreover, rainfall erosivity factor is the key factor, and rainfall during flood season plays a key role in soil erosion. Due to the soil erosion of HTYZ, siltation in Lake Ulansuhai reached 223.83 ha, with the annual siltation area increasing at a rate of 7.46 ha/yr. The results could provide a reference for estimating soil erosion of ungauged small watershed in semi-arid areas. Full article

In the past decade, more than 300 people have died per year on average due to mountain torrents in China. Mountain torrents mostly occur in ungauged small and medium-sized catchments, so it is difficult to maintain high accuracy of flood prediction. In order to solve the problem of the low accuracy of flood simulation in the ungauged areas, this paper studies the influence of different methods on the parameter regionalization of distributed hydrological model parameters in hilly areas of Hunan Province. According to the terrain, landform, soil and land use characteristics of each catchment, we use Shortest Distance, Attribute Similarity, Support Vector Regression, Generative Adversarial Networks, Classification and Regression Tree and Random Forest methods to create parameter regionalization schemes. In total, 426 floods of 25 catchments are selected to calibrate the model parameters, and 136 floods of 8 catchments are used for verification. The results showed that the average values of the Nash–Sutcliffe coefficients of each scheme were 0.58, 0.64, 0.60, 0.66, 0.61 and 0.68, and the worst values were 0.27, 0.31, 0.25, 0.43, 0.35 and 0.59. The random forest model is the most stable solution and significantly outperforms other methods. Using the random forest model to regionalize parameters can improve the accuracy of flood simulation in ungauged areas, which is of great significance for flash flood forecasting and early warning. Full article

Flow duration curves (FDCs) that represent streamflow regime function through an empirical relationship between the FDC parameters and basin descriptors are widely adopted for hydrologic applications. However, the applications of this method are highly dependent on the availability of observation data. Hence, it is still of great significance to explore the process controls of underpinning regional patterns on streamflow regimes. In this study, we developed a new regionalization method of FDCs to solve the problem of runoff prediction for ungauged mountainous basins. Five empirical equations (power, exponential, logarithmic, quadratic, and cubic) were used to fit the observed FDCs in the 64 mountainous basins in eastern China, and the power model outperforms other models. Stepwise regression was used to explore the differentiated control of 23 basin descriptors on the 13 percentile flows of FDCs, and seven descriptors remained as independent variables for further developing the regional FDCs. Application results with different combinations of these selected descriptors showed that five indices, i.e., average annual rainfall (P), average elevation (H), average gradient (β), average topographic index (TI), and maximum 7d of annual rainfall (Max7d), were the main control factors of FDCs in these areas. Through the regional method, we found that 95.31% of all the basins have NSE values greater than 0.60 and $\epsilon$ (namely the relative mean square error) values less than 20%. In conclusion, our study can guide runoff predictions to help manage booming demands for water resources and hydropower developments in mountainous areas. Full article

Accurate knowledge of the rain amount is a crucial driver in several hydrometeorological applications. This is especially true in complex orography territories, which are typically impervious, thus, leaving most mountain areas ungauged. Due to their spatial and temporal coverage, weather radars can potentially overcome such an issue. However, weather radar, if not accurately processed, can suffer from several limitations (e.g., beam blocking, altitude of the observation, path attenuation, and indirectness of the measurement) that can hamper the reliability of the rain estimates performed. In this study, a comparison between rain gauge and weather radar retrievals is performed in the target area of the Abruzzo region in Italy, which is characterized by a heterogeneous orography ranging from the seaside to Apennine ridge. Consequently, the Abruzzo region has an inhomogeneous distribution of the rain gauges, with station density decreasing with the altitude reaching approximately 1500 m a.s.l. Notwithstanding, pluviometric inflow spatial distribution shows a subregional dependency as a function of four climatic and altimetric factors: coastal, hilly, mountain, and inner plain areas (i.e., Marsica). Such areas are used in this analysis to characterize the radar retrieval vs. rain gauge amounts in each of those zones. Compared to previous studies on the topic, the analysis presented the importance of an accurate selection of the climatic and altimetric subregional areas where the radar vs. rain gauge comparison is undertaken. This aspect is not only of great importance to correct biases in radar retrieval in a more selective way, but it also paves the way for more accurate hydrometeorological applications (e.g., hydrological model initialization and quantification of aquifer recharge), which, in general, require the accurate knowledge of rain amounts upstream of a basin. To fill the gap caused by the uneven rain gauge distribution, ordinary Kriging (OK) was applied on a regional scale to obtain 2D maps of rainfall data, which were cumulated on a monthly and yearly basis. Weather radar data from the Italian mosaic were also considered, in terms of rain rate retrievals and cumulations performed on the same time frame used for rain gauges. The period considered for the analysis was two continuous years: 2017 and 2018. The output of the elaborations included raster maps for both radar and interpolated rain gauges, where each pixel contained a rainfall quantity. Although the results showed a general underestimation of the weather radar data, especially in mountain and Marsica areas, they were within the 95% confidence interval of the OK estimation. Our analysis highlighted that the average bias between radar and rain gauges, in terms of precipitation amounts, was a function of altitude and was almost constant in each of the selected areas. This achievement suggests that after a proper selection of homogeneous target areas, radar retrieval can be corrected using the denser network of rain gauges typically distributed at lower altitudes, and such correction can be extended at higher altitudes without loss of generality. Full article

Calibrating spatially distributed hydrological models in ungauged mountain basins is complicated due to the paucity of information and the uncertainty in representing the physical characteristics of a drainage area. In this study, an innovative method is proposed that incorporates the Budyko framework and water balance equation derived water yield (WYLD) in the calibration of the Soil and Water Assessment Tool (SWAT) with a monthly temporal resolution. The impact of vegetation dynamics (i.e., vegetation coverage) on Budyko curve shape parameter ω was considered to improve the Budyko calibration. The proposed approach is applied to the upstream Lancang-Mekong River (UL-MR), which is an ungauged mountain basin and among the world’s most important transboundary rivers. We compared the differences in SWAT model results between the different calibration approaches using percent bias (PBIAS), coefficient of determination (R²), and Nash–Sutcliffe efficiency (NSE) coefficient. The results demonstrated that the Budyko calibration approach exhibited a significant improvement against an unfitted priori parameter run (the non-calibration case) though it did not perform as good as fitting of the calibration by the observed streamflow. The NSE value increased by 44.59% (from 0.46 to 0.83), the R² value increased by 2.30% (from 0.87 to 0.89) and the PBIAS value decreased by 55.67% (from 39.7 to 17.6) during the validation period at the drainage outlet (Changdu) station. The outcomes of the analysis confirm the potential of the proposed Budyko calibration approach for runoff predictions in ungauged mountain basins. Full article

A large number of 2D models were originally developed as 1D models for the calculation of water levels along the main course of a river. Due to their development as 2D distributed models, the majority have added precipitation as a source term. The models can now be used as quasi-2D hydrodynamic rainfall–runoff models (‘HDRRM’). Within the direct rainfall method (‘DRM’), there is an approach, referred to as ‘rain-on-grid’, in which input precipitation is applied to the entire catchment area. The study contains a systematic analysis of the model behavior of HEC-RAS (‘Hydrologic Engineering Center—River Analysis System’) with a special focus on spatial resolution. The rain-on-grid approach is applied in a small, ungauged, low-mountain-range study area (Messbach catchment, 2.13 km²) in Central Germany. Suitable model settings and recommendations on model discretization and parametrization are derived therefrom. The sensitivity analysis focuses on the influence of the mesh resolution’s interaction with the spatial resolution of the underlying terrain model (‘subgrid’). Furthermore, the sensitivity of the parameters interplaying with spatial resolution, like the height of the laminar depth, surface roughness, model specific filter-settings and the precipitation input-data temporal distribution, is analyzed. The results are evaluated against a high-resolution benchmark run, and further criteria, such as 1. Nash–Sutcliffe efficiency, 2. water-surface elevation, 3. flooded area, 4. volume deficit, 5. volume balance and 6. computational time. The investigation showed that, based on the chosen criteria for this size and type of catchment, a mesh resolution between 3 m to 5 m, in combination with a DEM resolution from 0.25 m to 1 m, are recommendable. Furthermore, we show considerable scale effects on flooded areas for coarser meshing, due to low water levels in relation to topographic height. Full article

This study aims to build and test the adaptability and reliability of the Soil and Water Assessment Tool hydrological model in a small mountain forested watershed. This ungauged watershed covers 184 km² and supplies 90% of blue water for the Brașov metropolitan area, the second largest metropolitan area of Romania. After building a custom database at the forest management compartment level, the SWAT model was run. Further, using the SWAT-CUP software under the SUFI2 algorithm, we identified the most sensitive parameters required in the calibration and validation stage. Moreover, the sensitivity analysis revealed that the surface runoff is mainly influenced by soil, groundwater and vegetation condition parameters. The calibration was carried out for 2001–2010, while the 1996–1999 period was used for model validation. Both procedures have indicated satisfactory performance and a lower uncertainty of model results in replicating river discharge compared with observed discharge. This research demonstrates that the SWAT model can be applied in small ungauged watersheds after an appropriate parameterisation of its databases. Furthermore, this tool is appropriate to support decision-makers in conceiving sustainable watershed management. It also guides prioritising the most suitable measures to increase the river basin resilience and ensure the water demand under climate change. Full article

Soil erosion from arable lands removes the top fertile soil layer (comprised of humus/organic matter) and therefore requires fertilizer application which affects the overall sustainability. Hence, determination of soil erosion from arable lands is crucial to planning conservation measures. A modeling approach is a suitable alternative to estimate soil loss in ungauged catchments. Soil erosion primarily depends on soil texture, structure, infiltration, topography, land uses, and other erosive forces like water and wind. By analyzing these parameters, coupled with geospatial tools, models can estimate storm wise and annual average soil losses. In this study, a hilly watershed called Nongpoh was considered with the objective of prioritizing critical erosion hazard areas within the micro-catchment based on average annual soil loss and land use and land cover and making appropriate management plans for the prioritized areas. Two soil erosion models namely Revised Universal Soil Loss Equation (RUSLE) and Modified Morgan–Morgan–Finney (MMF) models were used to estimate soil loss with the input parameters extracted from satellite information and automatic weather stations. The RUSLE and MMF models showed similar results in estimating soil loss, except the MMF model estimated 7.74% less soil loss than the RUSLE model from the watershed. The results also indicated that the study area is under severe erosion class, whereas agricultural land, open forest area, and scrubland were prioritized most erosion prone areas within the watershed. Based on prioritization, best management plans were developed at catchment scale for reducing soil loss. These findings and the methodology employed can be widely used in mountainous to hilly watersheds around the world for identifying best management practices (BMP). Full article

The use of the Mike11 one-dimensional (1D) hydraulic model, together with official hydrology, represents a standard approach of the National Water Management Authority (NWMA) in Poland for flood mapping procedures. A different approach, based on the hydrological Event-Based Approach for Small and Ungauged Basins (EBA4SUB) model and the Flood-2 Dimensional (FLO-2D) hydraulic model has here been investigated as an alternative procedure. For the analysis, two mountainous rivers in Poland were selected: Kamienica Nawojowska is characterized by a narrow valley, while Skawinka has a broad valley. It was found that the flood zones can enormously differ locally, with larger zones generated by the Mike11/NWMA model in some cases and by the EBA4SUB/FLO-2D model in other situations. The benefits of using the two-dimensional (2D) model are consistent in areas without drainage and where the connection to the main channel is insufficient. The use of 1D modeling is preferred for the possibility of mapping the entire river network in a short computational time. Full article