Abstract Waterlogging is a major yet understudied constraint to cropland productivity. Here, we apply an agrohydrological modeling framework to assess the climatology of waterlogging (1981–2010), its yield impacts on major grain crops, and the potential for mitigation at a 500m x 500m grid resolution across Ethiopia’s rainfed agriculture (RFA) region. We estimate that areas within 39% of the grid cells in the RFA region are prone to waterlogging. Chickpea is the most sensitive crop, with average yield losses from 34% under low-severity to 97% under extreme-severity conditions, followed by wheat, barley, and faba bean (18–92% of yield loss). In contrast, maize and sorghum are generally tolerant except under extreme conditions. Our results also show considerable potential to reduce losses through soil drainage and use of tolerant crops. For instance, improved drainage could increase the current wheat yield on average by 57% under low waterlogging conditions. In flat areas where drainage is infeasible, productivity could be enhanced through using tolerant native crops and the expansion of rice production.

Rapid movements of sediment mass (e.g. shallow landslides, debris flows, rockfall) pose an imminent risk to settlements, infrastructure and human life in mountain regions. Forecasting such intermittent hazards on a large-scale is still challenging, yet essential to ensure effective risk management. Landslide early warning systems can benefit from the predictive power of dynamic hydroclimatic controls to better anticipate the initiation of these events.This study characterizes distinct hydroclimatic triggering conditions for rapid alpine mass movements, their exceptionality, and their predictability in time.We base our analysis on an inventory of ca. 1900 observations of shallow landslides, debris flows, and rockfalls in the Swiss Alpine Rhine basin (approx. 4300 km²) over the past 25 years. Utilizing hydrometeorological time series derived from gridded soil and climate products at a 1×1 km spatial and daily temporal resolution, we retrieve distinct families of predisposing and triggering conditions allowing us to objectively identify different process types. Our results show that a significant proportion of events are not exclusively rainfall-driven: approximately 20% of both shallow landslides and debris flows occurred under the influence of snow cover and snowmelt, suggesting that the hillslope response to precipitation and soil wetness varies seasonally. This underscores the necessity of a multivariate and sequential modeling approach.In a second step, we expand the methodology into a data-driven modelling framework by employing a recurrent neural network (long short-term memory LSTM). It simulates the probability of mass movements occurring over time by decoding the temporal dynamics of the catchment’s hydroclimatic conditions. We demonstrate the algorithm’s potential to internally reproduce hydrogeomorphic catchment states based solely on input time series of precipitation, temperature, and soil wetness. We report an area under the curve-receiver operating characteristic (AUC-ROC) metric of 0.94 (landslides) and 0.84 (debris flows) for testing.The findings of this study offer novel insights into hydroclimatic and hydrogeomorphic controls on the predisposing and triggering conditions of rapid alpine mass movements. Modern computational techniques allow to simulate seasonally varying contributions of multivariate hydrometeorological variables to the initiation of such events. This will enable predictions of changes in sediment mass movement distributions under a future climate and will offer an opportunity for plugging into early warning systems for landslides.

Predicting changes in urban pluvial flood hazards under climate warming is crucial for risk mitigation and disaster management. A key challenge in simulating future urban flood hazards is the scarcity of high-resolution rainfall projections, particularly at the sub-daily and kilometer scales required for hydrodynamic modeling. We present a cascading process-informed framework that requires minimal observed climatic data, enabling scenario analysis even in data-scarce cities. This framework consists of a distribution‐based spatial quantile mapping (DSQM) method to morph observed rainfall fields conditioned on temperature changes, a stochastic storm transposition (SST) method to account for the spatial variability of urban rainfall, and a rain‐on‐grid hydrodynamic model (AUTOSHED) for efficient simulation of urban pluvial floods at high spatio-temporal resolution. The framework allows the generation of stochastic rainfall fields under different rainfall return levels and regional warming levels. It supports the quantification of changes in future urban flood statistics with detailed hazard maps of inundation depth, duration, and flow velocity. We select the metropolitan area of Beijing (300 km2) as a case study area and utilize gridded hourly and 1 km rainfall data to simulate flood evolution at 5 min and 5 m resolution under regional warming levels of 1, 3, and 5 °C relative to the period 1998–2019. Our results show that with rising temperatures, regional storms tend to become more intense but smaller in spatial extent, which may in turn drive increased local flood depth, accelerated flow velocity, and deeper inundation, collectively elevating pluvial flood risk. Specifically, mean rainfall intensity increases by 6 %, 11 %, and 20 % (respectively with the warming levels), peak flood depth exhibits a nonlinear increase of 4 %, 7 %, and 8 %, due to the complex interactions of reduced storm area, increased storm intensities, and rainfall spatial variability. The proposed DSQM-SST-AUTOSHED framework offers a data-driven, physically grounded, and efficient approach to assess urban flood risk under regional warming. It only requires observed rainfall fields and temperature datasets, which are readily accessible from public sources, making the approach easily extendable to other cities.

Plant water storage contributes to transpiration, but it is unclear how its relevance in supporting transpiration depends on the stringency of stomatal regulation. Here, we show the compounding effect of stomatal regulation and hydraulic capacitance on plant water use, by means of a soil-plant hydraulic model and measurements of leaf water potential, sap flow, stomatal conductance and capacitance in beech and spruce in the field. We found that large capacitance led to a large buffering effect on leaf water potential, explained by increasing amounts of transpiration sourced from internal plant water storage. However, the extent to which capacitance allows plants to sustain transpiration depends on the stringency of stomatal regulation. For stomata that limit leaf water potential at a fixed threshold (as observed in spruce), large capacitance increased transpiration throughout all soil water conditions. By contrast, for flexible stomatal regulation mechanisms optimizing transpiration over leaf water potential (as observed in beech), large capacitance caused stomata to close earlier in the day under wet soil conditions. Our findings suggest a trade-off between developing tissues that can store large water volumes and stomatal regulation mechanisms that allow leaf water potential to reach more negative values during periods of high transpiration demand.

Successful drought identification and characterization are essential for effective drought risk assessment and management, requiring advanced characterization methods and the careful selection of drought indices and aggregation timescales capable of representing diverse drought features. Despite the wide range of existing drought indices, their general applicability is often constrained by dominant local conditions (climate regime, hydrology, land surface characteristics, and data availability) and the necessity to choose a suitable aggregation timescale for operational applications. This study aims to identify suitable drought indices to effectively characterize and monitor drought in the Horn of Africa (HoA). A combined cluster-area- and shape-based filtering approach, followed by three-dimensional (2D space and 1D time) connectivity, was employed to capture drought dynamics simultaneously in space and time. A range of drought indices with varying levels of complexity was evaluated and compared, including indices derived from single variables such as precipitation or soil moisture, as well as more complex multivariate indices based on combinations of multiple variables, including precipitation, potential evapotranspiration, soil moisture, normalized difference vegetation index (NDVI), and surface temperature. The performance of these indices was assessed against historical drought records reported by governmental and non-governmental organizations. The findings demonstrate that multivariate indices generally outperform univariate ones, with indices incorporating potential evapotranspiration showing high performance; however, no single index consistently excelled across all evaluation criteria. Considering both computational complexity and effectiveness in identifying drought-affected areas and capturing temporal characteristics, the combined use of the standardized precipitation evapotranspiration index (SPEI)–based indices, SPEI6 and SPEI9, is recommended for drought monitoring, planning, and management in the HoA, a region dominated by arid and semi-arid climates and recurrent, spatially extensive drought events.

Understanding drought dynamics jointly in space and time is essential for effective drought risk management and forecasting systems. However, the characterization of simultaneous space-time evolution of droughts has not been explored extensively, and studies describing drought in three dimensions are limited. Most previous studies commonly filter out small drought clusters by a cluster-area threshold. However, thresholds are often arbitrarily selected, leading to inconsistent definitions of large-scale drought across studies. Moreover, these studies consider only the area of drought clusters, overlooking shape differences and treating compact clusters the same as scattered or excessively elongated ones. Such scattered and excessively elongated clusters, with areas slightly above the threshold, result in many minor drought events and tend to prolong drought duration through tenuous spatial connections. This study introduces a combined cluster filtering method that incorporates cluster-shape parameters alongside area thresholds, identifies drought events using 26-cell connectivity, and tracks the events spatiotemporally. The method is applied to Ethiopia after quantifying drought level in each grid cell using the Standardized Precipitation Evapotranspiration Index (SPEI). The approach is shown to effectively identify meaningful drought events by removing clusters that would both prolong drought duration and produce minor drought events, while retaining moderate, severe, and persistent events, thereby enhancing drought pathway tracking. Evidence from the Ethiopia case study shows that the identified severe events align with documented cases, capturing affected regions and periods. The method offers a valuable framework for studying drought dynamics jointly across space and time, providing key insights to improve drought preparedness and management.

Gravitational mass movements in alpine regions, such as landslides, debris flows and rockfall, are driven by complex physical processes. While the translation and runout of these events can be reasonably well modelled once they occur, the predisposing and triggering mechanisms leading to failure are very challenging to assess. This is particularly demanding for practitioners who need to take decisions on the ground to ensure the safety of the population. There is potential to improve the situation by using a variety of new space-time climate and land surface datasets to describe the hydrogeomorphic system state and relate it to possible failure by confronting it with past observed events. In this work we focus on the local susceptibility to the initiation of mass wasting events (shallow landslides, debris flows and rockfall) in low- and subalpine regions by exploring the predictive power of various hydro-meteorological drivers related to rainfall, snowmelt, high soil moisture, freezing, etc.To provide spatially and temporally consistent information, we model all hydro-meteorological drivers governing the hydrogeomorphic catchment state of the Alpine Rhine (GR, Switzerland) over the period 1998-2022 based on globally available soil information (SoilGrids) as well as national climate (Federal Office of Meteorology and Climatology MeteoSwiss), snow (WSL Institute for Snow and Avalanche Research SLF) and terrain data (Federal Office of Topography Swisstopo). The temporal and spatial resolution of the analysis is daily over a 1x1km grid. We determine the seasonally varying contribution of each driver to the triggering of each individual mass movement type utilizing the concept of receiver operating characteristics (ROC) and its area under the curve (AUC) as performance metrics. The underlying events recorded in the Swiss natural hazard database comprise 459 shallow landslides, 295 debris flows and 761 rockfalls (StorMe, Swiss Federal Office for the Environment FOEN) in the study period. The best-performing hydro-meteorological drivers then serve as input to predict the occurrence of mass wasting events with data driven models. We test both a traditional statistical approach and machine learning algorithms to compare their capability of modelling the susceptibility to alpine mass movements.Compared to a purely rainfall-based prediction of landslide or debris flow activity, which is commonly done in the literature, this approach benefits from the availability of further spatially distributed climate variables and terrain characteristics. Our findings contribute to a better understanding of the role of catchment state on predisposing and triggering conditions of alpine mass movements, and illustrate also the limits of predictability for such events due to the inherent randomness in the triggering processes.

The AlpRhineS2S project, a collaboration between ETH Zurich and the University of Bern, researches the interplay of geological, geomorphological and hydrological processes within the sedimentary system of the Alpine Rhine in the canton of Grisons in Switzerland. Mechanisms of sediment erosion, transport and deposition determine the pathways of sediment from sources to sinks in a river basin. Long-term basin-averaged denudation rates serve to characterize the geomorphic properties of a catchment and to derive a sediment budget (Garipova et al., 2024), while specific hotspots of erosion considerably contribute to the short-term sediment supply into the fluvial system. Accordingly, mass wasting events play a crucial role in an Alpine geomorphic context by intermittently providing considerable amounts of sediment for transport in the river network. A large part of this sediment is transported in suspension, producing a complex turbidity signal at the outlet (Agostini et al., 2024) that features distinct tracers of source material composition (Garipova et al., 2024). In this contribution, we investigate the effects of precipitation as a triggering factor for frequent mass wasting events in the Alpine Rhine catchment. We correlate records of shallow landslide, debris flow, and rockfall events from the Swiss natural hazard database (StorMe, Swiss Federal Office for the Environment FOEN) to the gridded daily precipitation product RhiresD (Swiss Federal Office of Meteorology and Climatology MeteoSwiss). We estimate rainfall thresholds for those events by classifying consecutive rainfall days as either triggering or non-triggering events and performing jackknife cross-validation to assess the temporal bias of the event data following Leonarduzzi et al. (2017; 2020). We characterize the regional and seasonal effect of heavy precipitation events on increased sediment supply available for transport in the fluvial system. Finally, we also identify individual erosion hotspots and their link to sediment connectivity and slope stability assessments. Analyzing external drivers, we hypothesize on the effect of changes in climatic forcing on erosion mechanisms over the past decades, particularly due to increasing temperatures and precipitation intensities. References: Agostini, L., Demmel, S., Garipova, S., Sinclair, S., Schlunegger, F., Molnar, P. (2024): Suspended sediment transport in a river network: testing signal propagation and modelling approaches. EGU 2024. Garipova, S., Mair, D., Demmel, S., Agostini, L., Akçar, N., Molnar, P., Schlunegger, F. (2024): Source-to-Sink Sediment Tracing in the Glogn River Catchment. EGU 2024. Leonarduzzi, E., Molnar, P., McArdell, B.W. (2017): Predictive performance of rainfall thresholds for shallow landslides in Switzerland from gridded daily data. Water Resources Research 53(8): 6612–6625. https://doi.org/10.1002/2017WR021044. Leonarduzzi, E. and Molnar, P. (2020): Deriving rainfall thresholds for landsliding at the regional scale: daily and hourly resolutions, normalisation, and antecedent rainfall. Nat. Hazards Earth Syst. Sci., 20, 2905–2919. https://doi.org/10.5194/nhess-20-2905-2020.

In this study we propose a new methodology for pluvial flood risk estimation, combining stochastic rainfall modelling, climate projection based adaptations of the rainfall frequency-intensity relations and DEM data sets, along with hydrodynamic modelling. New global precipitation datasets, such as CMORPH, GSMaP or MERRA2 offer an affordable and accessible solution for water resource and water-hazard risk management in data-scarce regions and enable comprehensive global comparative studies. However, these datasets, often derived from satellite observations and coarse-scale climate modelling, consistently underestimate short-duration, high-intensity rainfall events, particularly those lasting one hour or less, that belong to the tails of the distributions (i.e., return levels higher than 30-year). This underestimation goes beyond spatial scale considerations, commonly addressed by areal reduction factors. Consequently, utilizing these global datasets for pluvial flood risk analysis results in conservative flood risk estimates. The availability of global terrain models and mapped man-made structures like buildings, channels, and roads enables the generation of wide-coverage digital surface models. These can be used for flood inundation modelling in combination with corrected extremes of the global precipitation data sets, allowing near-global rough flood risk estimates. In this study, we introduce a methodology for estimating pluvial flood risk using openly available global datasets. To achieve this, we derive hourly-scale Intensity-Duration-Frequency (IDF) curves suitable for pluvial flood inundation modeling in ungauged areas using global precipitation datasets. The first step uses high temporal resolution satellite remote sensing rainfall data (GSMaP) to train a stochastic rainfall generator model - the point process Bartlet-Lewis model. Subsequently, the weather generator is used to disaggregate daily global precipitation data (GPCC) through stochastic ensemble simulation. The resulting disaggregated ensemble data is then utilized to generate more accurate IDF curves including uncertainty, forming the basis for pluvial flooding risk assessments. Our approach integrates the openly available FabDEM terrain model with OpenStreetMap to generate digital surface models for flood risk modeling analysis. Discrepancies in flood inundation risk estimates in urban environments, attributable to underestimated rainfall intensity, are demonstrated using CADDIES, a 2-dimensional hydrodynamic model. The workflow allows the IDF curves for the current climate to be adapted based on climate model projections of temperatures using the Clausius–Clapeyron relation, and to study their impact on future flood risk. A comparative risk analysis is presented for several tropical coastal cities, including future pluvial risk projections. All analytical steps adhere to FAIR principles, utilizing publicly available datasets. The proposed workflow provides globally applicable first order estimates of pluvial flood risk, especially in data-poor areas, with better quality than existing global IDF studies or IDF curves derived directly from global precipitation datasets.

The AlpRhineS2S project, a collaboration between ETH Zürich and the University of Bern, researches the interplay of geological, geomorphological and hydrological processes within the sedimentary system of the Alpine Rhine in the canton of Grisons, Switzerland. Distributed river network hydrology-sediment models are being used in Alpine basins for the prediction of source activation and transport rates, for both fine and course sediment. Fine sediment input in such models may be generated by hillslope mass movements in hotspots of erosion (Demmel et al., 2024) which can be tracked, facilitating the development of sediment budgets (Garipova et al., 2024). However, despite the utility of hydrology-sediment models, the propagation of the sediment signal along channels is rarely tested against exact solutions and observations.In this contribution, we investigate the propagation of suspended sediment signals along channels and compare modelling simplifications with observations and theory. Averaged over long timescales, suspended sediment load represents the erosion rates of the catchment. At shorter timescales, from seasonal to hourly, sediment fluxes can describe the spatial distribution and activation of sediment sources and sinks across the basin. Active sediment sources and sinks constitute points of discontinuity in the basin, which create turbidity signals along the river network. Here we ask the questions: To what extent can channel flood wave propagation describe the sediment dynamics? Do current modelling approximations capture the richness of turbidity signals carried across the river network?The observation data used here are retrieved from flushing events and environmental flow releases across selected Alpine rivers. The turbidity signal properties of the different events are compared in non-dimensional terms, and synthetic common properties across the samples are determined. Modelling is compared through a successive approximation approach starting with a 1D solution for unsteady flow with the model 1D BASEMENT (Vanzo et al., 2021) for a range of channel geometries and slopes. Then the sediment propagation is analysed with the steady flow assumption of the parabolic and kinematic flood wave, in analytical form and in the TOPKAPI-ETH model, which we plan to use in the AlpRhineS2S Project for sediment fluxes and sediment source identification (Battista et al., 2020).Results highlight the extent to which numerical models can represent the channel sediment dynamics and what is consecutively missing from the introduced approximations. Findings also show that the suspended sediment propagation, even during controlled release events, cannot be described as a boundary condition problem: the interplay of deposition and resuspension along with local morphology and vegetation also play a fundamental role in the signal description. ReferencesBattista, G., Schlunegger, F., Burlando, P., Molnar, P. (2020): Modelling localized sources of sediment in mountain catchments for provenance studies, https://doi.org/10.1002/esp.4979.Demmel, S., Agostini, L., Garipova, S., Leonarduzzi, E., Schlunegger, F., Molnar, P. (2024): Climatic triggering of landslide sediment supply in the Alpine Rhine, EGU24.Garipova, S., Mair, D., Demmel, S., Agostini, L., Akçar, N., Molnar, P., Schlunegger, F. (2024): Source-to-Sink Sediment Tracing in the Glogn River Catchment, EGU24.Vanzo, Davide, et al. "BASEMENT v3: A modular freeware for river process modelling over multiple computational backends." (2021)

We aim at exploring the sedimentary source-to-sink pathways in the Alpine Rhine, Switzerland through integrating hydrological modeling, connectivity mapping, and field observations. We hypothesize that either rainfall-driven overland flow erosion or landsliding (Battista et al., 2020) are the main mechanisms contributing to the generation of sediment and controlling the source-to-sink transport of sediment in the basin. We test this hypothesis through mapping such sediment sources in the field and on lidar DEMs, and we conduct conceptual models to characterize the sensitivity of these sources to temporally and spatially varying rainfall rates (Demmel et al, 2024). We complement this analysis with a coupled hydrology-erosion model, through which we predict how the water and suspended sediment waves propagate downstream from the source through the channel network (Agostini et al, 2024). We then test these model-based predictions on rainfall-dependent source-to-sink sedimentary pathways with field data. We start with the 370 km²-large Glogn river catchment, which is a tributary of the Alpine Rhine. In the headwater reaches, the Glogn catchment is made up of a dense network of channels that are perched on the hillslopes, whereas farther downstream, the basin hosts several deep-seated landslides that potentially supply large volume of sediment to the channel network (Cruz Nuñes et al, 2015). We proceed upon collecting data about the size of clasts and their petrographic composition to characterize the source signal for the bedload of the Glogn River, and we trace these signals from upstream to downstream. We complement this dataset with a petrographic characterization of the suspension load including the bulk geochemical and mineralogical composition of sand and the measurements of concentrations of cosmogenic 10Be and 26Al in riverine quartz minerals. We then apply a principal component analysis to this dataset to identify the material signals of the different sediment sources, and we estimate the relative contribution of material from tributary basins through mixing modelling. We postulate that in the upstream, less dissected part of the basin, overland flow erosion constitutes the major mechanism of the sediment production, whereas in the downstream area where the Glogn has deeply dissected into the substratum, mass failure processes such as landsliding is the most important mechanism contributing to the production of sediment. References: Agostini, L., Demmel, S., Garipova, S., Sinclair, S., Schlunegger, F., Molnar, P. (2024) Suspended sediment transport in river network models: testing signal propagation and modelling approaches. EGU24. Battista, G., Schlunegger, F., Burlando, P., Molnar, P. (2020) Modelling localized sources of sediment in mountain catchments for provenance studies. Earth Surf. Process. Landforms, 45, 3475– 3487. Cruz Nuñes, F., Delunel, R., Schlunegger, F., Akçar, N., Kubik, P.W. (2015) Bedrock bedding, landsliding and erosional budgets in the Central European Alps. Terra Nova, 1-10. Demmel, S., Agostini, L., Garipova, S., Leonarduzzi, E., Schlunegger, F., Molnar, P. (2024) Climatic triggering of landslide sediment supply. EGU24.

Climate change is expected to influence future agricultural water availability, posing particular challenges in rainfed agricultural systems. This study aims to analyze the climatology of green water availability and water-limited attainable yield (AY) – the maximum crop yield achieved with available green water under optimal soil nutrient and crop management, considering four major cereal crops (teff, maize, sorghum, and wheat) produced in Ethiopia. An agrohydrological modeling framework was developed to simulate climatic–hydrological–crop interactions. The model was applied to a reference period (1981–2010) and a future period (2020–2099) under scenarios of low, intermediate, and high greenhouse gas emissions with the following aims: (i) evaluate the current green water availability and AY potential, (ii) assess their climate-driven changes, and (iii) analyze the sensitivity of changes in AY to changes in rainfall and atmospheric evaporative demand. With regional variations based on climatic regimes, the main growing season (Meher, May to September) has an average AY of 79 % of a fully irrigated potential yield, with an average soil moisture deficit of 29 % of moisture content at full water-holding capacity. AY of the short growing season (Belg, February to May) is, on average, 37 % of the potential yield, with a soil moisture deficit of 56 %. Under the future climate, Meher is expected to experience small changes in AY the range of ±5 %, with dominantly positive trends in the 2030s and decreases in the 2060s and 2080s, mainly driven by changes in the atmospheric evaporative demand due to rising temperatures. The Belg regions are expected to experience increased AY that is dominantly controlled by increases in rainfall. On the other hand, a substantial yield gap is identified between actual and water-limited yields. This points to the need for combining green water management practices with nutrient and tillage management, plant protection, and cultivar improvement to close the yield gaps and to build up the climate resilience of farmers.

Fine sediment transported in suspension is an important part of the total sediment yield in most rivers with erodible upland sediment sources. Fine sediment has positive effects on the stabilization of riverbanks, the accretion of floodplains, nutrient transport, and carbon sequestration. However, when fine sediment load is excessive, it can also clog the streambed, reduce invertebrate and fish habitat, prevent river-aquifer exchange and hyporheic flows, and damage hydropower infrastructure. To effectively design sediment management policies in rivers, it is fundamental to understand the fine sediment dynamics at the catchment scale. This study focuses on the washload, the fine sediment fraction that, once entrained, remains in suspension until its deposition.Washload dynamics are typically quantified by concurrently measuring stage or discharge (Q) and turbidity, from which suspended sediment concentration (SSC) is derived. Q-SSC pairs often create a hysteretic relationship, allowing us to infer the distance of fine sediment sources upstream of the station.This contribution adopts a reach-scale perspective on Q-SSC analysis, moving beyond single-station hysteresis loops and leveraging Q-SSC data from two stations, one upstream and one downstream. The core idea is that we can study washload as a passive tracer to gain further information about the hydraulic variables of roughness and water velocity, for each event separately. We can then integrate this information to further describe the fine sediment sources dynamics and the washload regime of the studied reach. Combining the subsequent reach-scale information we can completely reconstruct the washload production timing and yields across the whole catchment.For this purpose, we developed new tools which ought to become the new standard for Q-SSC analysis. First, we identify in the discharge timeseries the flood and sediment pulse events through a new algorithm based on Empirical Mode Decomposition. Second, we study the virtual velocity of the flood and sediment signal by a new definition of cross correlation, analysing the Hilber transforms of the signal. Third, we study the presence and the nature (e.g. intensity and seasonality) of suspended sediment sources and tributaries through a time dependant boundary condition analytical solution of the advection-diffusion equation, both for discharge and washload concentration. The development of these three new tools and their application to the Arc-Isere (France) with six stations and four reaches, allowed us to identify the fine sediment sources and sinks in the river network. We also gained insights into seasonal fine sediment yields, the deposition and re-suspension dynamics of riverbed sediment stocks, and their progressive depletion during the spring-summer season. These methods are generalizable and applicable wherever discharge (Q) or stage and SSC data are available at two or more locations.

One of the major challenges posed by climate change in agriculture is the long-term alteration in cropland suitability. This alteration has serious consequences for food security and economic stability at global, regional, and local scales. especially in smallholder and rainfed agricultural systems like in Ethiopia. A comprehensive understanding of the current state of croplands and future changes under the warming temperature and increasing rainfall uncertainty is critical for national climate adaptation planning. Here, we evaluated cropland suitability (CLS) for four major cereal crops (teff, maize, sorghum, and wheat), under both current and future climates across the rainfed agriculture (RFA) landscapes of Ethiopia. We utilized a novel suitability modeling approach that establishes mathematical relationships between crop yield, and climatic factors (rainfall and temperature) and soil factors (texture, pH, and organic carbon). Furthermore, we analyzed the relative influences of the growing season rainfall and temperature on the changes in CLS. The results show that ~50% of the RFA area has a suitability index of 0.6 or higher (moderately to highly suitable) for teff and that 64%, 68%, and 46% of the grid cells are suitable for maize, sorghum, and wheat crops, respectively. The suitable agroecologies of the four crops will likely undergo altitudinal shifts and areal reduction, with magnitudes of the changes depending on the emission scenarios. Under the SSP2-4.5, the suitable areas are projected to decrease by 23% for teff, 13% for maize, 14% for sorghum, and 16% for wheat in the 2080s. In semi-arid and hyper-humid climates, CLS is sensitive to changes in the growing season rainfall, whereas in the most lowland and highland regions, it is temperature-sensitive. In light of our results, we argue that adaptation actions that are tailored to climatic conditions and topographic locations are vitally necessary to offset climate change's long-term impacts on Ethiopia's rainfed agriculture.

<p>Forests modulate precipitation and evapotranspiration fluxes. One important – yet often overlooked - component in the forest water cycle is the forest-floor litter layer. Leaves and deadwood retain significant amounts of annual precipitation and enhance subcanopy humidity. At the “Waldlabor Zurich” ecohydrology field site we conducted numerous experiments to quantify the water fluxes from and to the forest-floor litter layer. We estimated the total retention capacities of needle, broadleaf and deadwood litter, assessed the litter water content before and after precipitation events, and measured soil moisture in litter-covered and litter-free plots. We used micro lysimeters to estimate evaporation from the litter layer and measured subcanopy humidity and temperature at different heights above the forest floor to assess the effect of evaporation on subcanopy microclimate.</p> <p>Storage capacities of needle litter and broadleaf litter averaged 3.1 and 1.9 mm, respectively, with evaporation timescales exceeding 2 days, whereas deadwood stored ~0.7 mm of precipitation, and retained water for >7 days. Deadwood water retention increased with more advanced decomposition. Together the forest floor litter layer reduced soil water recharge, reduced soil evaporation rates, and insulated against ground heat fluxes thus impacting snowmelt patterns. Timeseries of deadwood water content revealed a diel cycle of stored water, water content increased during nighttime due to condensation of dew and fog and decreased during the day when vapor pressure deficit and evaporation were high. The water evaporating from the forest‐floor litter layer increased humidity, decreased temperature, and reduced vapor pressure deficit in the subcanopy atmosphere. Although, the absolute amounts of water storage in the forest-floor litter layer are relatively small, these storages were frequently filled and emptied with every precipitation event, thus effecting the overall soil water recharge. Overall, 18% of annual precipitation, or 1/3 of annual evapotranspiration, were retained in the forest-floor litter layer suggesting that overlooking litter interception may lead to substantial overestimates of recharge and transpiration in many forest ecosystems.</p>

<p>Simulation of the catchment rainfall-runoff transformation with physically based watershed models is a traditional way to predict streamflow and other hydrological variables at catchment scales. However, the calibration of such models requires large data inputs and computational power and contains many parameters which are often impossible to constrain or validate. An alternative approach is to use data-driven machine learning for streamflow prediction.</p> <p>In the past few years, LSTM (long short-term memory) models and its variants have been explored in rainfall-runoff modelling. Typical applications use daily climate variables as inputs and model the rainfall-runoff transformation processes with different timescales of memory. This is especially useful as delays in runoff production by snow accumulation and melt, soil water storage, evapotranspiration, etc., can be included. In contrast to feed-forward ANNs (artificial neural networks), LSTMs are capable of maintaining the sequential temporal order of inputs, and compared to RNNs (recurrent neural networks), of learning the long-term dependencies. [1]</p> <p>However, current work on LSTMs mostly focuses on the USA, the UK and Brazil, where CAMELS datasets are available [1, 2, 3]. Catchments at higher altitudes with snow-driven dynamics and sometimes glaciers are present in small number in these datasets (if at all). Systematic applications of LSTMs for streamflow prediction in climates where a significant part of the catchments are snow and ice dominated are missing. In this work, an FS-LSTM (fast slow-LSTM) previously applied in Brazil is adapted for Swiss catchments to fill this gap [3]. The FS-LSTM explored builds on the work of Hoedt et al. (2021) that imposed mass constraints on an LSTM, called MC-LSTM [4]. FS-LSTM adds a fast and slow part for streamflow, containing rainfall and soil moisture respectively. We will discuss benchmark results against an existing semi-distributed conceptual model widely used in Switzerland for streamflow simulation [5].</p> <p> </p> <p>References:</p> <p>[1]: Kratzert et al., Rainfall-runoff modelling using Long Short-Term Memory (LSTM) networks, 2018.</p> <p>[2]: Lees et al., Hydrological concept formation inside long short-term memory (LSTM) networks, 2022.</p> <p>[3]: Quinones et al., Fast-Slow Streamflow Model Using Mass-Conserving LSTM, 2021.</p> <p>[4]: Hoedt et al., MC-LSTM: Mass-Conserving LSTM, 2021.</p> <p>[5]: Viviroli et al., An introduction to the hydrological modelling system PREVAH and its pre- and post-processing-tools, 2009.</p>

<p>Successful application of hydrological models requires data to assess the validity, as well as the inherent uncertainty, of the outputs, most importantly streamflow. In parts of Sub-Saharan Africa (SSA), such data is often lacking. Therefore, it is frequently challenging to find the necessary resources for setting up a robust hydrological model and hydrological monitoring platforms. In data-scarce regions within SSA, ground data required to model and make water resources decisions are not always available and therefore, some form of alternative data sources and simplified modelling approaches are required. In recent years, satellite and climate reanalysis data have been intensely explored for watershed modelling in poorly gauged regions with variables such as precipitation, evapotranspiration, soil moisture, runoff etc. Very good potential is provided by the ERA5-Land dataset which is considered one of the best freely available global products for hydrology given its 0.1° x 0.1° spatial resolution and an hourly to monthly temporal resolution spanning from 1950 till present. Here, ERA-5 Land input on a monthly resolution was assessed in the Berg River Basin, South Africa using the Modified PITMAN model. Total precipitation, runoff, and potential evapotranspiration for each of the basin’s 12 quaternary catchments were retrieved using the Google Earth Engine platform for a study period of 40 years (1981-2021). A validation period of 20 years (1985-2005) was used corresponding to the freely available streamflow data. The assimilation of ERA5-Land precipitation data showed satisfactory results across the basin with the best results in the upstream catchment (G10A) with a 0.634 coefficient of efficiency and 0.404 KGE during the initial run. However, runoff for the downstream catchments (G10K) gave positive biases in high-flow months. This paper gives a detailed analysis of the performance of remotely sensed datasets (ERA5-Land) on catchments with varying climatic, land use and cover, water use, and geomorphological characteristics, therefore, offering a valuable reference for its applications in understanding hydrological processes in different river basins across SSA.</p>