Medium-sized Catchment Research Articles

Error correction method based on deep learning for improving the accuracy of conceptual rainfall-runoff model

Due to the complex runoff and concentration situation, flood forecasting for small and medium-sized catchments is very difficult. To improve the accuracy of flood forecasting, this study constructs a distributed model for flood forecasting based on the Xainanjiang (XAJ) model and the North China (NC) model respectively, and takes the deep learning model including LSTM and transformer to compare. Taking the Podi Basin and Shibazi Basin as study cases. LSTM, Transformer, and LSTPencoder models were used to correct the error of distributed models, and the differential evolution (DE) algorithm was used to optimize the model parameters. Taking the observed rainfall and distributed model results as input, the residuals of the simulation flow are fitted to improve the accuracy of the distributed model. The research results show that the NC model performs better than the XAJ, LSTM, and transformer models. Compared with the XAJ model, the average Nash-Sutcliffe coefficient of the NC model for the two catchments increased by 25.7 % and 5.87 % respectively. The performance of the NC+LSTPencoder model is better than other models. Compared with the NC model, the average Nash-Sutcliffe coefficient of the NC+LSTPencoder model for two catchments increased by 89.7 % and 1.12 % respectively, and the Root Mean Square Errors (RMSE) of the two catchments reduced by 79.1 % and 63.4 % respectively. The model proposed in this paper has strong correction ability, which has important significance for improving the accuracy of flood forecasts.

Hydrogeomorphological approach for flood analyses at high- detailed scale: Narrow rivers with broad complex alluvial plains

Fluvial flood risk management requires precise knowledge of flooding processes; thus, there is a need to carry out a detailed hydrological, hydraulic, and geomorphologic analysis to understand floods and the factors that influence them. Although the integration of geomorphological methods and hydrological-hydraulic modelling is known to be crucial for flood risk analysis, it remains a challenge to apply these approaches to the understanding of floods in small and medium-sized catchments, focusing on complex fluvial systems that have been highly modified. In this work, a hydrogeomorphological methodology was developed to analyse fluvial floods in narrow rivers with broad complex alluvial plains. The procedure comprises different phases that combine hydrologic-hydraulic modelling and a geomorphological break-down. Critically, this allows for a detailed analysis of existing relationships among erosional and depositional fluvial landforms and hydraulic parameters, comprising geometric sections, discharges, depths, flood extents and, therefore, flood hazards. The approach involves the analysis of specific flood events beforehand and the study of fluvial floods associated with different probabilities of occurrence afterwards. The information employed in the present study incorporates high- detailed inputs comprising hourly time series of hydrometeorological data, LiDAR data to extract high-precision terrain information, fluvial geomorphic maps and flood event extents based on Sentinel-2 imagery. We demonstrate this methodology in the Carrión River, located in the province of Palencia, Spain, whose characteristics include a wide complex alluvial plain, reaches highly modified by anthropic interventions and recurrent flood episodes. This hydrogeomorphological approach allowed the integration of geomorphological knowledge and hydrologic-hydraulic modelling in flood analysis. As a result, it led to a deeper comprehension of the processes of inundation, which would be challenging to thoroughly understand without incorporating a holistic method, an essential factor for defining effective flood management strategies.

Towards multi-model soil erosion modelling: An evaluation of the erosion potential method (EPM) for global soil erosion assessments

Soil erosion is expected to increase in the future due to climate change. Soil erosion models are useful tools that can be used by decision makers and other stakeholders to deal with soil erosion problems or the implementation of soil protection measures. Most of the modelling applications are using Universal Soil Loss Equation (USLE)-type models. In this study, we evaluate the applicability of the Erosion Potential Model (EPM) and its modified version (mEPM) for the estimation of the gross and net erosion rates at a global scale. The sensitivity analysis shows that the model results have the highest variability due to the soil protection (land cover) coefficient followed by the soil erodibility parameter. The models’ evaluations indicate that that the EPM cannot be applied to cold regions while the mEPM overcomes this issue. The erosion rates based on the EPM were 1.5–2.5 times larger than the ones obtained from the mEPM. Increasing the number of catchment properties as inputs to the model may help in improving the performance of the tested EPM and mEPM. Moreover, a comparison of net soil losses by mEPM with long-term suspended sediment yield data for 116 catchments located around the globe indicates a median bias of less than 10%, although the bias for around 1/3 of catchments was above 100%. Furthermore, a direct comparison with other soil erosion models such as USLE-type models is not possible since the EPM and mEPM do take into consideration other processes such as soil slumps and gully erosion and not just sheet and rill erosion. Therefore, as expected, the gross erosion rates by the EPM and mEPM are higher compared to the USLE-type models. Hence, the mEPM, despite its limitations, could be regarded as an interesting approach for the describing erosion processes around the globe and should be further tested using small- and medium-sized catchments from various climate zones.

State updating in a distributed hydrological model by ensemble Kalman filtering with error estimation

For flood simulation in small- and medium-sized catchments, discharge observations may be used to update model states of a distributed hydrological model to improve performance. The ensemble Kalman filter (EnKF) has been widely used for hydrological assimilation due to its relative simplicity and robustness. An advantage of the EnKF is that it is easy to include different sources of uncertainty, therefore the choice of error model is crucial for the application of the EnKF assimilation. This paper describes an EnKF assimilation scheme for estimating error models using the maximum a posteriori estimation method (MAP). We test this scheme in two small and medium-sized catchments in China with different characteristics, and in addition compared the performance differences under two kinds of rainfall forcing. We show that MAP is beneficial in specifying error models and providing reliable ensemble spread. The assimilation scheme can effectively ameliorate the degradation of distributed hydrological model performance due to uncalibrated model parameters and/or poor quality of input data.

Research on Parameter Regionalization of Distributed Hydrological Model Based on Machine Learning

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.

Identification of a Function to Fit the Flow Duration Curve and Parameterization of a Semi-Arid Region in North China

The discharge process has undergone major changes in many river basins throughout the world as a result of the simultaneous influences of global climate change and human activity. Flow duration curves (FDCs) are crucial indicators of river basins’ hydrological processes. However, it is challenging to compare FDCs in a quantitative way. This study will identify the best function with which to fit the flow duration curve in a semi-arid region of North China, so as to quantify the FDC, and parameterize the function of the FDC of the region in order to describe the FDCs of ungauged basins. In this work, six small- and medium-sized catchments in North China are selected as the study area, and three functions, i.e., log normal, generalized Pareto and H2018 functions, were chosen to fit the FDC at nineteen hydrological stations. The relationship between the parameters of the FDC and the basin characteristics, such as the climatic factors and geographical features, were analyzed. A regression formula of the parameters of the FDC function was established, and its spatial and temporal distributions were examined. Based on the evaluation of four indicators, the Nash–Sutcliffe efficiency, the root mean square relative error, the logarithmic Nash efficiency coefficient and the coefficient of determination, the results demonstrate that the H2018 function can match FDCs the best. Through the annual runoff, annual precipitation, precipitation in summer, potential evapotranspiration, catchment area, mean elevation, length of the main channel and maximum flow frequency, the parameters of a, b, and k in the H2018 function can be formulated. The regression formula constructed in this study can obtain a regional flow duration curve with satisfactory performance, which provides a reference for the validation of remote-sensing-based runoff data in ungauged regions.

Comparison of process-based and statistical approaches for simulation and projections of rainfed crop yields

Accurate and comprehensive modelling aimed at investigating the impact of climate change on rainfed crop yields is of great importance due to the interconnected issues of water scarcity and food security. Because the process-based and statistical approaches to simulating crop yields are different in nature, a comparison between them is needed. This study investigates the accuracy of crop yield simulations in the historical period as well as future projections using two modelling approaches: 1) a process-based approach employing the Soil and Water Assessment Tool+ (SWAT+) model, and 2) a statistical approach employing a data-driven model, Feed Forward Back Propagation Neural Network (FFBPNN) over a medium-sized catchment in north-western Poland. The application of two potential evapotranspiration methods (Penman-Monteith and Hargreaves) in SWAT+ permitted calibration (2004–2011) and validation (2012–2019) of runoff and yields of winter wheat and spring barley. Different combinations of climatic parameters with a drought index based on Joint Deficit Index were applied to simulate and project rainfed crop yields (winter wheat, barley, potato, rye, rapeseed, sugar beets, cereals, maize for grain, maize for green forage, pulses) with FFBPNN. The results reveal that adding the new drought index helped increase the FFBPNN performance. This approach showed that future yields of the studied crops would slightly increase under RCP8.5 by 2060. Winter wheat and spring barley projections from SWAT+ showed very small changes using both the Penman-Monteith and Hargreaves method. Policy-wise, the results should be of interest to climate change adaptation practitioners and food security experts. Future studies should aim at more thorough investigation of the role of the downscaling technique and extreme events, as well as the effect of elevated CO2 on future crop yields.

A statistical approach for identifying factors governing streamflow recession behaviour

AbstractCatchment storage‐release relation has been widely studied using the parameters of streamflow recession analysis. The governing factors of the recession parameters are poorly understood, particularly in snow‐dominated regions. Here, we tailor a newly developed statistical approach, marginal contribution feature importance, to a hydrologic context, and couple it with random forests, in order to investigate the governing physical and climatic factors of catchment recession behaviour. The coupled approach can incorporate the interactions among catchment climatic and physical attributes. In a large sample hydrology study, we identify and compare the governing factors of recession parameters, in rainfall versus snowmelt‐dominated catchments, and in medium‐size versus large catchments, across more than 1000 catchments in United States and Canada. Results show that streamflow recession behaviour, particularly recession nonlinearity, strongly depend on belowground attributes and slope in rain‐dominated medium size catchments, and strongly depend on slope and annual maximum snow water equivalent in snow‐dominated medium size catchments. As catchment scale increases (>1000 km2), the attributes related to the magnitude and timing of input water (e.g., water surplus, aridity index, maximum snow water equivalent) dictates the streamflow recession behaviour and the importance of belowground attributes is dropped. Furthermore, recession nonlinearity generally increases with an increase in catchment size. The findings of this study help improve our understanding of the governing factors and the interpretation of the spatial variability of recession behaviours. Such understanding could inform the development of a generalizable process‐based framework for estimating the sensitivity of catchment storage‐release relation to climate change in different environmental settings.

Spatio-temporal differences dominate suspended sediment dynamics in medium-sized catchments in central Germany

Suspended sediment transport, dominating material export from lowland river systems, is highly variable in space and time. Since suspended sediment transport can have adverse effects on human infrastructure, targeted management is imperative. To achieve this, a thorough understanding of suspended sediment dynamics is needed, but requires long-term monitoring data. Despite studies investigating suspended sediment dynamics on decadal time scales, an assessment of the spatio-temporal variability across adjacent river catchments is still lacking.To fill this gap, we analyse suspended sediment transport at twelve stations along German upland rivers, monitored on a daily basis for 28 to 54 years. All monitoring stations belong to the Rhine (Lahn, Lippe, Main, and Neckar) and Weser river systems (Fulda, Leine, Werra, and Weser) and have contributing areas between 2500 and 22,000 km2. Although located in comparable topo-climatic settings, average sediment yield for the monitoring period varied more than fourfold. Covering the range of 5.9 to 28.7 t km−2 yr−1, we found the highest and lowest suspended sediment yields for the Neckar and Lippe catchments, respectively. Using two different streamflow separation approaches, we estimated the contribution of stormflow discharge to total sediment transport to fall between 73 and 85 % in the Neckar and Lahn rivers while remaining ~15 % lower for the remaining catchments. Investigating the relevance of extreme events, we found the share of total suspended sediment load transported during the ten largest events to vary between ~28 % for the Neckar and ~8 % for the Lippe and Werra catchments. These events were predominantly characterized by clockwise hysteresis loops that hint to in-channel reworking as a significant material source, with the exception of the Neckar stations where anti-clockwise hysteresis patterns prevailed. Here, the hysteresis analysis indicates sediment delivery from more distant sources in the headwater regions, likely related to elevated rainfall intensities and the susceptibility of the escarpment of the Swabian Alb towards landsliding. We conclude that the combined effect of varying rainfall amounts and intensities, the fraction of steeper slopes as well as lithological differences explain the across-catchment variability on multiple time scales that is further modulated by river management practices.

Sentinel-2 high-resolution data for river discharge monitoring

River monitoring is an open issue due to many intrinsic problems of the ground monitoring network. Over the last few decades, the role of satellite sensors in river discharge estimation is significantly increased thanks to the strong growth in technologies and applications. Focusing on daily river discharge measurements, a non-linear regression model has been used to link the near-infrared (NIR) reflectance ratio between a dry and a wet pixel around the section of a river to the ground measurements of river discharge. The use of medium-resolution satellite data, such as those from MODIS sensors, enables to monitor high and low flows in medium-sized catchments (<100,000 km2), thanks to satellite frequent revisit time and wide spatial coverage. However, such sensors are not suitable to provide information for medium-narrow rivers (< 250 m wide), nor to study river features and patterns that are averaged within a single pixel. Here, we investigated the use of Sentinel-2 NIR reflectances to support the hypothesis that a higher spatial resolution, i.e. 10 m, is able to better identify the wet pixels, more related to the river dynamics, with obvious advantages for river discharge estimation compared to the medium resolution sensors (e.g., MODIS at 250 m). Moreover, it also allows both a finer distinction between vegetation, soil and water and the characterization of water turbidity in the river area. A new formulation enriched by the sediment component is proposed together with a first step toward an uncalibrated procedure to select the wet pixels. Google Earth Engine (GEE) platform has been employed for the data analysis, allowing to avoid the download of big amounts of data, fostering the reproducibility of the analysis in different locations. The accuracy of the river discharges derived from Sentinel-2 reflectances is evaluated against the in-situ observations from selected gauging stations along two Italian rivers, Po and Tiber. The results confirm the good performances obtained with high-resolution images over the Po River, with average Nash-Sutcliffe efficiency ranging between 0.39 and 0.56 for the different configurations adopted. Relatively worse results were obtained over the Tiber River where the Nash-Sutcliffe efficiency ranged between 0.2 and 0.61, due to an issue on the registration of Sentinel-2 images.

Evaluating performance dependency of a geomorphologic instantaneous unit hydrograph-based hydrological model on DEM resolution

The digital elevation model (DEM) is a type of model that has been widely used in terrain analysis and hydrological modeling. DEM resolution influences the hydrological and geomorphologic features of delineated catchments and consequently affects hydrological simulations. This study investigated the impacts of DEM resolution on the performance of the XAJ-GIUH hydrological model, a model coupling the widely used Xinanjiang (XAJ) hydrological model with the geomorphologic instantaneous unit hydrograph (GIUH), in flood simulations in small and medium-sized catchments. To test the model performance, the model parameters were calibrated at a fine DEM resolution (30 m) and then directly transferred to the simulation runs using coarser DEMs. Afterwards, model recalibration was conducted at coarser DEM resolutions. In the simulation runs with the model parameters calibrated at the 30-m resolution, the DEM resolution slightly affected the overall shape of the simulated flood hydrographs but presented a greater impact on the simulated peak discharges in the two study catchments. The XAJ-GIUH model consistently underestimated the peak discharges when the DEM resolution became coarser. The qualified ratio of peak simulations decreased by 35% when the DEM resolution changed from 30 m to 600 m. However, model recalibration produced comparable model performances when DEMs with different resolutions were used. This study showed that the impact of DEM resolution on model performance can be mitigated by model recalibration to some extent, if the DEM resolution is not too coarse.

A Simple Approach to Account for Stage–Discharge Uncertainty in Hydrological Modelling

The effect of stage–discharge (H-Q) data uncertainty on the predictions of a MIKE SHE-based distributed model was assessed by conditioning the analysis of model predictions at the outlet of a medium-size catchment and two internal gauging stations. The hydrological modelling was carried out through a combined deterministic–stochastic protocol based on Monte Carlo simulations. The approach considered to account for discharge uncertainty was statistically rather simple and based on (i) estimating the H-Q data uncertainty using prediction bands associated with rating curves; (ii) redefining the traditional concept of residuals to characterise model performance under H-Q data uncertainty conditions; and (iii) calculating a global model performance measure for all gauging stations in the framework of a multi-site (MS) test. The study revealed significant discharge data uncertainties on the order of 3 m3 s−1 for the outlet station and 1.1 m3 s−1 for the internal stations. In general, the consideration of the H-Q data uncertainty and the application of the MS-test resulted in remarkably better parameterisations of the model capable of simulating a particular peak event that otherwise was overestimated. The proposed model evaluation approach under discharge uncertainty is applicable to modelling conditions differing from the ones used in this study, as long as data uncertainty measures are available.

Simulating Discharge in a Non-Dammed River of Southeastern South America Using SWAT Model

Within a single region, it is possible to identify opposite changes in flow production. This proved to be the case for several basins in southeastern South America. It remains challenging to the causes this behavior and whether changes in streamflow will continue at current levels or decline in the coming decades. In this study, we used the Soil Water Assessment Tool to simulate monthly river discharge in the Ivaí River Basin, an unregulated medium-sized catchment and tributary of the Upper Paraná River Basin. After calibration, the simulated flow regime for the five streamflow stations based on the Nash-Sutcliffe Efficiency index (NSE) rated four of the streamflow stations Very Good (NSE between 0.86 and 0.89) and only one in the Good index (0.70). The overall flow behavior was well represented, although an underestimation was identified in four monitoring stations. Through assessment of its functionality and limitations in terms of specific flow duration curves percentages, the calibrated model could provide (to managers) the reliability needed for a realistic intervention. The results of this study may assist managers and support public policies for the use of water resources at the Ivaí River basin.

The Application of a Meteo-hydrological Forecasting System with Rainfall Bias Correction in a Small and Medium-sized Catchment

The Application of a Meteo-hydrological Forecasting System with Rainfall Bias Correction in a Small and Medium-sized Catchment

Sources of fine-sediment reservoir deposits from contrasting lithological zones in a medium-sized catchment over the past 60 years

Reservoirs influence the processes associated with the downstream transportation and deposition of sediments in various sizes of catchments. Patterns of sediment yield from different lithological zones in a catchment can differ greatly. Assessing the historical sources of fine sediment deposited in a reservoir derived from contrasting lithological zones aids our understanding of the characteristic pattern of sediment yields from different lithological environments and improves our knowledge of sediment sources in reservoir systems. However, few previous studies have addressed this issue. In this study, geochemical fingerprinting was used to identify the sources of the fine sediments derived from contrasting lithological zones and deposited in a reservoir in a medium-sized catchment in a rocky mountainous region of northern China over the past 60 years, using the chronology determined by 137Cs and 210Pbex tracers. The influence of grain-size sorting on the source of the fine sediments deposited in the reservoir was examined. Soil samples were collected from three types of lithological zones within the catchment, namely, the granite (205 samples), limestone (120 samples), and shale (86 samples) zones, and the elemental composition and grain size of these samples was determined. Six sediment cores (Cores 1–6) were collected from the perennial backwater area of the reservoir to determine the elemental composition, grain size, chronologies, and the source of fine sediments derived from the three types of lithological zones. Two further cores (Cores 7 and 8) were collected from the reservoir’s upstream floodplain, and the grain size distributions and chronologies of these deposits were also determined. Overall, the sources of fine sediment identified in Cores 1–6 reflected the changing trends in sediment contribution from the three types of lithological zones over the past 60 years, which was linked to the differing sediment yields from the contrasting lithological zones. The granite zone occupies half of the catchment, has thicker soils, is relatively susceptible to soil erosion, and has been subjected to greater disturbance by human activity than the limestone and shale zones, but contributed only 25% of the fine sediment in Cores 1–6. In contrast, the limestone and shale zones occupy 23% and 15%, respectively, of the catchment but contributed 42% and 33% of the fine sediment in Cores 1–6. This result is related to grain-size sorting during downstream sediment transportation and deposition in the reservoir system, as demonstrated by a comparison of the grain size distributions of the soil samples from the three types of lithological zones, Cores 7 and 8, and Cores 1–6. These findings suggest the sediment source has been strongly influenced by grain-size sorting. This study provides new insights into the source of sediments deposited in a reservoir system from contrasting lithological zones at the scale of a medium-sized catchment over multi-decadal timescales.

An iterative runoff propagation approach to identify priority locations for land cover change minimizing downstream river flood hazard

Local infiltration and surface runoff generation depends on local land use, soil type, antecedent soil water content and slope, while runoff accumulation downstream is furthermore determined by re-infiltration along the flow paths. Hence, land use changes can both mitigate and exacerbate runoff accumulation and flood hazard, providing an opportunity to identify optimal locations for land use changes. An optimization method is presented, encompassing the iterative application of a spatially explicit rainfall-runoff model. This method ranks eligible locations (pixels) according to their modelled contribution to accumulated runoff downstream for a given land use change and given rainfall events, thereby guaranteeing maximal or minimal impact. This method was tested for two medium-sized catchments, located in Flanders, Belgium. Three land use changes were considered: afforestation, sealing and practicing winter cover crops. Results show the considerable impact of these land use changes and their locations on runoff accumulation at the outlet: afforestation of all eligible pixels reduces runoff volumes with 67% to 84 %, cover crops reduce runoff volumes in winter with 42% to 37%, while sealing triples runoff. The priority pixels have a larger impact on downstream runoff volume: afforesting or sealing the 20% highest or lowest ranked pixels leads to a reduction of 71% to 54%, resp. an increase of 102% to 115%. These priority pixels are characterized by high flow accumulation, highlighting the importance of enhancing the infiltration capacity in river valleys. The presented procedure allows spatial planners to consider the impact of local land use interventions to flood resistance downstream.

Near-Realtime Quantitative Precipitation Estimation and Prediction (RealPEP)

Flash floods in small- to medium-sized catchments and intense precipitation over cities caused by severe local storms pose increasing threats to our society. For the timely prediction of such events, the value of high-resolution and high-quality QPE and corresponding forecasts cannot be overrated. Seamless predictions harmonizing nowcasting and numerical weather prediction (NWP) across forecast lead times from minutes to days would greatly help to improve the value and efficiency of warnings. Organized by the Research Unit on Near-Realtime Precipitation Estimation and Prediction (RealPEP, www2.meteo.uni-bonn.de/realpep) and supported by the Project on Seamless Integrated Forecasting System (SINFONY, www.dwd .de/DE/forschung/forschungsprogramme/sinfony_iafe/sinfony_node.html) of the German Meteorological Service (DWD), an international 3-day online conference was held from 5 to 7 October 2020, dedicated to Precipitation and Flash-Flood Predictions from Minutes to Days (https://indico .scc.kit.edu/event/883/). Most speakers agreed to have their presentations recorded, which we uploaded to YouTube for further distribution (see, e.g., on the conference homepage, https:// indico.scc.kit.edu/event/883/page/588-recorded-talks). The speakers were both invited experts in the respective research fields and researchers from the RealPEP and SINFONY projects. Talks and discussions could be followed on video stream. Interaction between the about 250 participants was enabled by entering written questions and comments via a dedicated tool, which allowed for voting and thus also ranking questions. Registered participants could enter chat rooms from where they could be moved to the speaker room for posing the questions directly to the speakers and the auditorium. On the last day of the conference podium discussions with selected speakers summarized talks and discussions and elaborated on overarching problems, ideas, and developments in the fields of quantitative precipitation estimation (QPE), quantitative precipitation nowcasting (QPN), quantitative precipitation forecasting (QPF), flash-flood prediction (FFP), and their organization into seamless prediction systems, which also constituted the topics of the five sessions during the conference. We report here in particular on the outcomes of the panel discussions.

How reliable are the satellite-based precipitation estimations in guiding hydrological modelling in South China?

With increasing spatiotemporal resolution and accuracy, the satellite-based precipitation estimations have made their exploitation in hydrological applications possible, especially in poorly gauged regions. However, their hydrological utilities still need to be evaluated in different catchments to provide useful knowledge to the end-users. This study investigates the performance of latest version (V06) of Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (IMERG) products and their capacity to guide hydrological modelling in South China. Both near-real-time (i.e., “Early” run, IMERG-E) and post-real-time (i.e., “Final” run, IMERG-F) IMERG products are evaluated, and the post-real-time TRMM Multi-satellite Precipitation Analysis (TMPA 3B42) product is employed for intercomparison. The hydrological utilities of abovementioned products are further explored over six different medium-sized catchments in South China based on the grid-based Xinanjiang model. The results indicate that the IMERG-F outperforms the IMERG-E in South China, and the quality of TMPA 3B42 falls between them in terms of KGE' and its components. For the hydrological utility, using gauge-based model parameters, the IMERG-E forced simulations show poor performance in most of the selected catchments, whereas much better performance is observed for the IMERG-F forced simulations. Implementation of input-specific calibration significantly improves the simulation performance of three SPEs, and such improvement is much more pronounced for the IMERG-E. The TMPA 3B42 forced simulations generally illustrate comparable reliability to these of IMERG-F but with a slightly degraded performance. Overall, the IMERG-E shows high values in facilitating hydrological modelling over most ungauged catchments, and the IMERG-F shows much better hydrological performance. However, the best performing IMERG-F forced simulations are still inferior to these driven by the benchmark precipitation, and it’s thus still hard to replace gauge observations completely with the SPEs. This study provides a helpful knowledge about hydrological performance of satellite-based precipitation estimations, which potentially promote their widespread applications in hydrological modelling.

Can we estimate flood frequency with point-process spatial-temporal rainfall models?

Stochastic rainfall models are commonly used in practice for long-term flood risk management. One of the most widely used model types is based on point processes. Despite the widespread use of such models, whether their known simplifications in describing the space–time structure of rainfall will affect the accuracy of flood estimation has not been quantified. In this study, we quantify the biases introduced by the rainfall model limitations to flood estimates in two medium-sized river catchments (717 km2 and 844 km2) in the South East of the UK. To achieve this, we used nine years of hourly radar rainfall data, a dense network of hourly rain gauges, a spatial–temporal rainfall stochastic model based on point processes, and a fully distributed hydrological model. We modelled the corresponding catchment water dynamics using observed and simulated hourly rainfall and then assessed whether the errors introduced by the stochastic model will propagate in the river flow dynamics. Our results show that the stochastic rainfall model properly captures the point-scale rainfall statistics, including point extremes and the cross-site spatial correlations. However, the model results in a bias on extremes of areal statistics, including an overestimation of the areal reduction factor, extreme areal mean precipitation, and the areal fraction of rain (wet area ratio). Using this as input for continuous hydrological simulations, we find that the flow duration curves are well preserved, particularly in the high flow seasons (relative bias is less than 7%). The model also reproduces well the flood frequency curves at a daily scale with an averaged relative bias of 0.36–16.9% at 10-year return levels, confirming its ability to infer the long-term flood risk for medium-sized catchments. However, the summer-season hourly peak discharge is highly overestimated with a relative bias of over 163.5% at the same return level. The overestimation in summer hourly peak discharge is explained by the dominating convective systems in summer and the misrepresented spatial structure in the model.

The impact of recent gully filling practices on wheat yield at the Campiña landscape in Southern Spain

The significance of gully filling practices on agricultural production has been only rarely evaluated in literature. In this study, a medium-sized catchment was selected, for being representative of the conventional farming practices on annual crops in the Campiña landscape in Southern Spain. During two wheat campaigns, an analysis of the spatial and temporal evolution of satellite-based vegetation indexes (NDVI, EVI and VH/VV radar polarization ratio) was conducted before and after gully filling practices associated to a rehabilitation project. The development of a linear model of topographic factors along with the calibration of a yield equation using a field survey allowed to determine the expected site-based productivities in the study area. The maximum EVI proved to be the best proxy for wheat yield in a location and was controlled mainly by rainfall and irradiation conditions. A significant yield reduction was found in the vicinity of those gullies affected by soil scraping within the rehabilitation project. Indirect yield losses in the productive land doubled the direct cost of unproductive areas caused by gullying and, when added up to the latter, between 15–25% of the expected yield was found to be lost. Although longer and broader studies are needed to confirm our results, this case study shows that current agricultural practices are not environmentally unsustainable but also economically detrimental. Moreover, it highlights that the impact assessment of gully erosion cannot only be circumscribed to the eroded area but the degradation of gully-adjacent locations must be also considered.