Showing papers in "Hydrology and Earth System Sciences in 2023"

Continuous streamflow prediction in ungauged basins: long short-term memory neural networks clearly outperform traditional hydrological models

Abstract: Abstract. This study investigates the ability of long short-term memory (LSTM) neural networks to perform streamflow prediction at ungauged basins. A set of state-of-the-art, hydrological model-dependent regionalization methods are applied to 148 catchments in northeast North America and compared to an LSTM model that uses the exact same available data as the hydrological models. While conceptual model-based methods attempt to derive parameterizations at ungauged sites from other similar or nearby catchments, the LSTM model uses all available data in the region to maximize the information content and increase its robustness. Furthermore, by design, the LSTM does not require explicit definition of hydrological processes and derives its own structure from the provided data. The LSTM networks were able to clearly outperform the hydrological models in a leave-one-out cross-validation regionalization setting on most catchments in the study area, with the LSTM model outperforming the hydrological models in 93 % to 97 % of catchments depending on the hydrological model. Furthermore, for up to 78 % of the catchments, the LSTM model was able to predict streamflow more accurately on pseudo-ungauged catchments than hydrological models calibrated on the target data, showing that the LSTM model's structure was better suited to convert the meteorological data and geophysical descriptors into streamflow than the hydrological models even calibrated to those sites in these cases. Furthermore, the LSTM model robustness was tested by varying its hyperparameters, and still outperformed hydrological models in regionalization in almost all cases. Overall, LSTM networks have the potential to change the regionalization research landscape by providing clear improvement pathways over traditional methods in the field of streamflow prediction in ungauged catchments.

The natural abundance of stable water isotopes method may overestimate deep-layer soil water use by trees

Abstract: Abstract. Stable water isotopes have been used extensively to study the water use strategy of plants in various ecosystems. In deep vadose zone (DVZ) regions, the rooting depth of trees can reach several meters to tens of meters. However, the existence of roots in deep soils does not necessarily mean the occurrence of root water uptake, which usually occurs at a particular time during the growing season. Therefore, quantifying the contribution of deep-layer soil water (DLSW) in DVZ regions using the natural abundance of stable water isotopes may not be accurate because this method assumes that trees always extract shallow- and deep-layer soil water. We propose a multi-step method for addressing this issue. First, isotopic labeling in deep layers identifies whether trees absorb DLSW and determines the soil layer depths from which trees derive their water source. Next, we calculate water sources based on the natural abundance of stable isotopes in the soil layer determined above to quantify the water use strategy of trees. We also compared the results with the natural abundance of stable water isotopes method. The 11- and 17-year-old apple trees were taken as examples for analyses on China's Loess Plateau. Isotopic labeling showed that the water uptake depth of 11-year-old apple trees reached 300 cm in the blossom and young fruit (BYF) stage and only 100 cm in the fruit swelling (FSW) stage, whereas 17-year-old trees always consumed water from the 0–320 cm soil layer. Overall, apple trees absorbed the most water from deep soils (>140 cm) during the BYF stage, and 17-year-old trees consumed more water in these layers than 11-year-old trees throughout the growing season. In addition, the natural abundance of stable water isotopes method overestimated the contribution of DLSW, especially in the 320–500 cm soil layer. Our findings highlight that determining the occurrence of root water uptake in deep soils helps to quantify the water use strategy of trees in DVZ regions.

Regional significance of historical trends and step changes in Australian streamflow

Abstract: Abstract. The Hydrologic Reference Stations is a network of 467 high-quality streamflow gauging stations across Australia that is developed and maintained by the Bureau of Meteorology as part of an ongoing responsibility under the Water Act 2007. The main objectives of the service are to observe and detect climate-driven changes in observed streamflow and to provide a quality-controlled dataset for research. We investigate trends and step changes in streamflow across Australia in data from all 467 streamflow gauging stations. Data from 30 to 69 years in duration ending in February 2019 were examined. We analysed data in terms of water-year totals and for the four seasons. The commencement of the water year varies across the country – mainly from February–March in the south to September–October in the north. We summarized our findings for each of the 12 drainage divisions defined by Australian Hydrological Geospatial Fabric (Geofabric) and for continental Australia as a whole. We used statistical tests to detect and analyse linear and step changes in seasonal and annual streamflow. Monotonic trends were detected using modified Mann–Kendall (MK) tests, including a variance correction approach (MK3), a block bootstrap approach (MK3bs) and a long-term persistence approach (MK4). A nonparametric Pettitt test was used for step-change detection and identification. The regional significance of these changes at the drainage division scale was analysed and synthesized using a Walker test. The Murray–Darling Basin, home to Australia's largest river system, showed statistically significant decreasing trends for the region with respect to the annual total and all four seasons. Drainage divisions in New South Wales, Victoria and Tasmania showed significant annual and seasonal decreasing trends. Similar results were found in south-western Western Australia, South Australia and north-eastern Queensland. There was no significant spatial pattern observed in central nor mid-west Western Australia, with one possible explanation for this being the sparse density of streamflow stations and/or the length of the datasets available. Only the Tanami–Timor Sea Coast drainage division in northern Australia showed increasing trends and step changes in annual and seasonal streamflow that were regionally significant. Most of the step changes occurred during 1970–1999. In the south-eastern part of Australia, the majority of the step changes occurred in the 1990s, before the onset of the “Millennium Drought”. Long-term monotonic trends in observed streamflow and its regional significance are consistent with observed changes in climate experienced across Australia. The findings of this study will assist water managers with long-term infrastructure planning and management of water resources under climate variability and change across Australia.

Signal contribution of distant areas to cosmic-ray neutron sensors – implications for footprint and sensitivity

Abstract: Abstract. This paper presents a new theoretical approach to estimate the contribution of distant areas to the measurement signal of cosmic-ray neutron detectors for snow and soil moisture monitoring. The algorithm is based on the local neutron production and the transport mechanism, given by the neutron–moisture relationship and the radial intensity function, respectively. The purely analytical approach has been validated with physics-based neutron transport simulations for heterogeneous soil moisture patterns, exemplary landscape features, and remote fields at a distance. We found that the method provides good approximations of simulated signal contributions in patchy soils with typical deviations of less than 1 %. Moreover, implications of this concept have been investigated for the neutron–moisture relationship, where the signal contribution of an area has the potential to explain deviating shapes of this curve that are often reported in the literature. Finally, the method has been used to develop a new practical footprint definition to express whether or not a distant area's soil moisture change is actually detectable in terms of measurement precision. The presented concepts answer long-lasting questions about the influence of distant landscape structures in the integral footprint of the sensor without the need for computationally expensive simulations. The new insights are highly relevant to support signal interpretation, data harmonization, and sensor calibration and will be particularly useful for sensors positioned in complex terrain or on agriculturally managed sites.

Effects of dynamic changes of desiccation cracks on preferential flow: experimental investigation and numerical modeling

Abstract: Abstract. Preferential flow induced by desiccation cracks (PF-DC) has been proven to be an important hydrological effect that could cause various geotechnical engineering and ecological environment problems. Investigation on the PF-DC remains a great challenge due to the soil shrinking–swelling behavior. This work presents an experimental and numerical study of the PF-DC considering the dynamic changes of desiccation cracks. A soil column test was conducted under wetting–drying cycles to investigate the dynamic changes of desiccation cracks and their hydrological response. The ratios between the crack area and soil matrix area (crack ratio), crack aperture and depth were measured. The soil water content, matrix suction and water drainage were monitored. A new dynamic dual-permeability preferential flow model (DPMDy) was developed, which includes physically consistent functions in describing the variation of both porosity and hydraulic conductivity in crack and matrix domains. Its performance was compared to the single-domain model (SDM) and rigid dual-permeability model (DPM) with fixed crack ratio and hydraulic conductivity. The experimental results showed that the maximum crack ratio and aperture decreased when the evaporation intensity was excessively raised. The self-closure phenomenon of cracks and increased surficial water content was observed during low-evaporation periods. The simulation results showed that the matrix evaporation modeled by the DPMDy is lower than that of the SDM and DPM, but its crack evaporation is the highest. Compared to the DPM, the DPMDy simulated a faster pressure head building-up process in the crack domain and higher water exchange rates from the crack to the matrix domain during rainfall. Using a fixed crack ratio in the DPM, whether it is the maximum or the average value from the experiment data, will overestimate the infiltration fluxes of PF-DC but underestimate its contribution to the matrix domain. In conclusion, the DPMDy better described the underlying physics involving crack evolution and hydrological response with respect to the SDM and DPM. Further improvement of the DPMDy should focus on the hysteresis effect of the soil water retention curve and soil deformation during wetting–drying cycles.

Machine-learning- and deep-learning-based streamflow prediction in a hilly catchment for future scenarios using CMIP6 GCM data

Abstract: Abstract. The alteration in river flow patterns, particularly those that originate in the Himalaya, has been caused by the increased temperature and rainfall variability brought on by climate change. Due to the impending intensification of extreme climate events, as predicted by the Intergovernmental Panel on Climate Change (IPCC) in its Sixth Assessment Report, it is more essential than ever to predict changes in streamflow for future periods. Despite the fact that some research has utilised machine-learning- and deep-learning-based models to predict streamflow patterns in response to climate change, very few studies have been undertaken for a mountainous catchment, with the number of studies for the western Himalaya being minimal. This study investigates the capability of five different machine learning (ML) models and one deep learning (DL) model, namely the Gaussian linear regression model (GLM), Gaussian generalised additive model (GAM), multivariate adaptive regression splines (MARSs), artificial neural network (ANN), random forest (RF), and 1D convolutional neural network (1D-CNN), in streamflow prediction over the Sutlej River basin in the western Himalaya during the periods 2041–2070 (2050s) and 2071–2100 (2080s). Bias-corrected data downscaled at a grid resolution of 0.25∘ × 0.25∘ from six general circulation models (GCMs) of the Coupled Model Intercomparison Project Phase 6 GCM framework under two greenhouse gas (GHG) trajectories (SSP245 and SSP585) were used for this purpose. Four different rainfall scenarios (R0, R1, R2, and R3) were applied to the models trained with daily data (1979–2009) at Kasol (the outlet of the basin) in order to better understand how catchment size and the geo-hydromorphological aspects of the basin affect runoff. The predictive power of each model was assessed using six statistical measures, i.e. the coefficient of determination (R2), the ratio of the root mean square error to the standard deviation of the measured data (RSR), the mean absolute error (MAE), the Kling–Gupta efficiency (KGE), the Nash–Sutcliffe efficiency (NSE), and the percent bias (PBIAS). The RF model with rainfall scenario R3, which outperformed other models during the training (R2 = 0.90; RSR = 0.32; KGE = 0.87; NSE = 0.87; PBIAS = 0.03) and testing (R2 = 0.78; RSR = 0.47; KGE = 0.82; NSE = 0.71; PBIAS = −0.31) period, therefore was chosen to simulate streamflow in the Sutlej River in the 2050s and 2080s under the SSP245 and SSP585 scenarios. Bias correction was further applied to the projected daily streamflow in order to generate a reliable times series of the discharge. The mean ensemble of the model results shows that the mean annual streamflow of the Sutlej River is expected to rise between 2050s and 2080s by 0.79 % to 1.43 % for SSP585 and by 0.87 % to 1.10 % for SSP245. In addition, streamflow will increase during the monsoon (9.70 % to 11.41 % and 11.64 % to 12.70 %) in the 2050s and 2080s under both emission scenarios, but it will decrease during the pre-monsoon (−10.36 % to −6.12 % and −10.0 % to −9.13 %), post-monsoon (−1.23 % to −0.22 % and −5.59 % to −2.83 %), and during the winter (−21.87 % to −21.52 % and −21.87 % to −21.11 %). This variability in streamflow is highly correlated with the pattern of precipitation and temperature predicted by CMIP6 GCMs for future emission scenarios and with physical processes operating within the catchment. Predicted declines in the Sutlej River streamflow over the pre-monsoon (April to June) and winter (December to March) seasons might have a significant impact on agriculture downstream of the river, which is already having problems due to water restrictions at this time of year. The present study will therefore assist in strategy planning to ensure the sustainable use of water resources downstream by acquiring knowledge of the nature and causes of unpredictable streamflow patterns.

Estimating propagation probability from meteorological to ecological droughts using a hybrid machine learning copula method

Abstract: Abstract. The impact of droughts on vegetation is essentially manifested as the transition of water shortage from the meteorological to ecological stages. Therefore, understanding the mechanism of drought propagation from meteorological to ecological drought is crucial for ecological conservation. This study proposes a method for calculating the probability of meteorological drought to trigger ecological drought at different magnitudes in northwestern China. In this approach, meteorological and ecological drought events during 1982–2020 are identified using the three-dimensional identification method; the propagated drought events are extracted according to a certain spatiotemporal overlap rule, and propagation probability is calculated by coupling the machine learning model and C-vine copula. The results indicate that (1) 46 drought events are successfully paired with 130 meteorological and 184 ecological drought events during 1982–2020, and ecological drought exhibits a longer duration but smaller affected area and severity than meteorological drought; (2) a quadratic discriminant analysis (QDA) classifier performs the best among the 11 commonly used machine learning models which are combined with four-dimensional C-vine copula to construct the drought propagation probability model; and (3) the hybrid method considers more drought characteristics and a more detailed propagation process which addresses the limited applicability of the traditional method to regions with large spatial extent.

Estimating leaf moisture content at global scale from passive microwave satellite observations of vegetation optical depth

Abstract: Abstract. The moisture content of vegetation canopies controls various ecosystem processes such as plant productivity, transpiration, mortality, and flammability. Leaf moisture content (here defined as the ratio of leaf water mass to leaf dry biomass, or live-fuel moisture content, LFMC) is a vegetation property that is frequently used to estimate flammability and the danger of fire occurrence and spread, and is widely measured at field sites around the globe. LFMC can be retrieved from satellite observations in the visible and infrared domain of the electromagnetic spectrum, which is however hampered by frequent cloud cover or low sun elevation angles. As an alternative, vegetation water content can be estimated from satellite observations in the microwave domain. For example, studies at local and regional scales have demonstrated the link between LFMC and vegetation optical depth (VOD) from passive microwave satellite observations. VOD describes the attenuation of microwaves in the vegetation layer. However, neither were the relations between VOD and LFMC investigated at large or global scales nor has VOD been used to estimate LFMC. Here we aim to estimate LFMC from VOD at large scales, i.e. at coarse spatial resolution, globally, and at daily time steps over past decadal timescales. Therefore, our objectives are: (1) to investigate the relation between VOD from different frequencies and LFMC derived from optical sensors and a global database of LFMC site measurements; (2) to test different model structures to estimate LFMC from VOD; and (3) to apply the best-performing model to estimate LFMC at global scales. Our results show that VOD is medium to highly correlated with LFMC in areas with medium to high coverage of short vegetation (grasslands, croplands, shrublands). Forested areas show on average weak correlations, but the variability in correlations is high. A logistic regression model that uses VOD and additionally leaf area index as predictor to account for canopy biomass reaches the highest performance in estimating LFMC. Applying this model to global VOD and LAI observations allows estimating LFMC globally over decadal time series at daily temporal sampling. The derived estimates of LFMC can be used to assess large-scale patterns and temporal changes in vegetation water status, drought conditions, and fire dynamics.

Spatial variability in Alpine reservoir regulation: deriving reservoir operations from streamflow using generalized additive models

Abstract: Abstract. Reservoir regulation affects various streamflow characteristics, from low to high flows, with important implications for downstream water users. However, information on past reservoir operations is rarely publicly available, and it is hardly known how reservoir operation signals, i.e. information on when water is stored in and released from reservoirs, vary over a certain region. Here, we propose a statistical model to reconstruct reservoir operation signals in catchments without information on reservoir operation. The model uses streamflow time series observed downstream of a reservoir that encompass a period before and a period after a known year of reservoir construction. In a first step, a generalized additive model (GAM) regresses the streamflow time series from the unregulated pre-reservoir period on four covariates including temperature, precipitation, day of the year, and glacier mass balance changes. In a second step, this GAM, which represents natural conditions, is applied to predict natural streamflow, i.e. streamflow that would be expected in the absence of the reservoir, for the regulated period. The difference between the observed regulated streamflow signal and the predicted natural baseline should correspond to the reservoir operation signal. We apply this approach to reconstruct the seasonality of reservoir regulation, i.e. information on when water is stored in and released from a reservoir, from a dataset of 74 catchments in the central Alps with a known reservoir construction date (i.e. date when the reservoir went into operation). We group these reconstructed regulation seasonalities using functional clustering to identify groups of catchments with similar reservoir operation strategies. We show how reservoir management varies by catchment elevation and that seasonal redistribution from summer to winter is strongest in high-elevation catchments. These elevational differences suggests a clear relationship between reservoir operation and climate and catchment characteristics, which has practical implications. First, these elevational differences in reservoir regulation can and should be considered in hydrological model calibration. Furthermore, the reconstructed reservoir operation signals can be used to study the joint impact of climate change and reservoir operation on different streamflow signatures, including extreme events.

Advancing measurements and representations of subsurface heterogeneity and dynamic processes: towards 4D hydrogeology

Abstract: Abstract. Essentially all hydrogeological processes are strongly influenced by the subsurface spatial heterogeneity and the temporal variation of environmental conditions, hydraulic properties, and solute concentrations. This spatial and temporal variability generally leads to effective behaviors and emerging phenomena that cannot be predicted from conventional approaches based on homogeneous assumptions and models. However, it is not always clear when, why, how, and at what scale the 4D (3D + time) nature of the subsurface needs to be considered in hydrogeological monitoring, modeling, and applications. In this paper, we discuss the interest and potential for the monitoring and characterization of spatial and temporal variability, including 4D imaging, in a series of hydrogeological processes: (1) groundwater fluxes, (2) solute transport and reaction, (3) vadose zone dynamics, and (4) surface–subsurface water interactions. We first identify the main challenges related to the coupling of spatial and temporal fluctuations for these processes. We then highlight recent innovations that have led to significant breakthroughs in high-resolution space–time imaging and modeling the characterization, monitoring, and modeling of these spatial and temporal fluctuations. We finally propose a classification of processes and applications at different scales according to their need and potential for high-resolution space–time imaging. We thus advocate a more systematic characterization of the dynamic and 3D nature of the subsurface for a series of critical processes and emerging applications. This calls for the validation of 4D imaging techniques at highly instrumented observatories and the harmonization of open databases to share hydrogeological data sets in their 4D components.

The origin of hydrological responses following earthquakes in a confined aquifer: insight from water level, flow rate, and temperature observations

Abstract: Abstract. Although many mechanisms of earthquake-induced hydrological response have been proposed in recent decades, the origins of these responses remain enigmatic, and a quantitative understanding of them is lacking. In this study, we quantitatively analyze the mechanism of coseismic response in water level and flow rate from an artesian well in southwestern China before and after multiple earthquakes and reveal the origin of the earthquake-induced hydrological response based on the monitoring data of water temperature. Water level and temperature always show coseismic step-like increases following earthquakes, which are independent of the earthquakes' epicentral distances and magnitudes. Tidal analysis finds changes in aquifer and aquitard permeability following these earthquakes, which corresponds to the post-seismic total discharge of 85–273 m3 in 20 d after earthquakes. Furthermore, we couple the flow rate and temperature data to model the mixing processes that occurred following each earthquake. The results indicate that coseismic temperature changes are the result of the mixing of different volumes of water from shallow and deep aquifers, with the mixing ratio varying according to each earthquake.

A robust gap-filling approach for European Space Agency Climate Change Initiative (ESA CCI) soil moisture integrating satellite observations, model-driven knowledge, and spatiotemporal machine learning

Abstract: Abstract. Spatiotemporally continuous soil moisture (SM) data are increasingly in demand for ecological and hydrological research. Satellite remote sensing has potential for mapping SM, but the continuity of satellite-derived SM is hampered by data gaps resulting from inadequate satellite coverage, snow cover, frozen soil, radio-frequency interference, and so on. Therefore, we propose a new gap-filling approach to reconstruct daily SM time series using the European Space Agency Climate Change Initiative (ESA CCI). The developed approach integrates satellite observations, model-driven knowledge, and a machine learning algorithm that leverages both spatial and temporal domains. Taking SM in China as an example, the reconstructed SM showed high accuracy when validated against multiple sets of in situ measurements, with a root mean square error (RMSE) and a mean absolute error (MAE) of 0.09–0.14 and 0.07–0.13 cm3 cm−3, respectively. Further evaluation with a 10-fold cross-validation revealed median values of the coefficient of determination (R2), RMSE, and MAE of 0.56, 0.025, and 0.019 cm3 cm−3, respectively. The reconstructive performance was noticeably reduced both when excluding one explanatory variable and keeping the other variables unchanged and when removing the spatiotemporal domain strategy or the residual calibration procedure. In comparison with gap-filled SM data based on a satellite-derived diurnal temperature range (DTR), the gap-filled SM data from bias-corrected model-derived DTRs exhibited relatively lower accuracy but higher spatial coverage. Application of our gap-filling approach to long-term SM datasets (2005–2015) produced a promising result (R2=0.72). A more accurate trend was achieved relative to that of the original CCI SM when assessed with in situ measurements (i.e., 0.49 versus 0.28, respectively, in terms of R2). Our findings indicate the feasibility of integrating satellite observations, model-driven knowledge, and spatiotemporal machine learning to fill gaps in short- and long-term SM time series, thereby providing a potential avenue for applications to similar studies.

Projected changes in droughts and extreme droughts in Great Britain strongly influenced by the choice of drought index

Abstract: Abstract. Droughts cause enormous ecological, economical and societal damage, and they are already undergoing changes due to anthropogenic climate change. The issue of defining and quantifying droughts has long been a substantial source of uncertainty in understanding observed and projected trends. Atmosphere-based drought indicators, such as the Standardised Precipitation Index (SPI) and the Standardised Precipitation Evapotranspiration Index (SPEI), are often used to quantify drought characteristics and their changes, sometimes as the sole metric representing drought. This study presents a detailed systematic analysis of SPI- and SPEI-based drought projections and their differences for Great Britain (GB), derived from the most recent set of regional climate projections for the United Kingdom (UK). We show that the choice of drought indicator has a decisive influence on the resulting projected changes in drought frequency, extent, duration and seasonality using scenarios that are 2 and 4 ∘C above pre-industrial levels. The projected increases in drought frequency and extent are far greater based on the SPEI than based on the SPI. Importantly, compared with droughts of all intensities, isolated extreme droughts are projected to increase far more with respect to frequency and extent and are also expected to show more pronounced changes in the distribution of their event durations. Further, projected intensification of the seasonal cycle is reflected in an increasing occurrence of years with (extremely) dry summers combined with wetter-than-average winters. Increasing summer droughts also form the main contribution to increases in annual droughts, especially using the SPEI. These results show that the choice of atmospheric drought index strongly influences the drought characteristics inferred from climate change projections, with a comparable impact to the uncertainty from the climate model parameters or the warming level; therefore, potential users of these indices should carefully consider the importance of potential evapotranspiration in their intended context. The stark differences between SPI- and SPEI-based projections highlight the need to better understand the interplay between increasing atmospheric evaporative demand, moisture availability and drought impacts under a changing climate. The region-dependent projected changes in drought characteristics by two warming levels have important implications for adaptation efforts in GB, and they further stress the need for rapid mitigation.

Past and future climate change effects on the thermal regime and oxygen solubility of four peri-alpine lakes

Abstract: Abstract. Long-term effects of climate change on lakes globally will include a substantial modification in the thermal regime and the oxygen solubility of lakes, resulting in the alteration of ecosystem processes, habitats, and concentrations of critical substances. Recent efforts have led to the development of long-term model projections of climate change effects on lake thermal regimes and oxygen solubility. However, such projections are hardly ever confronted with observations extending over multiple decades. Furthermore, global-scale forcing parameters in lake models present several limitations, such as the need of significant downscaling. In this study, the effects of climate change on thermal regime and oxygen solubility were analyzed in the four largest French peri-alpine lakes over 1850–2100. We tested several one-dimensional (1D) lake models' robustness for long-term variations based on up to 63 years of limnological data collected by the French Observatory of LAkes (OLA). Here, we evaluate the possibility of forcing mechanistic models by following the long-term evolution of shortwave radiation and air temperature while providing realistic seasonal trends for the other variables for which local-scale downscaling often lacks accuracy. Based on this approach, MyLake, forced by air temperatures and shortwave radiations, predicted accurately the variations in the lake thermal regime over the last 4 to 6 decades, with RMSE < 1.95 ∘C. Over the previous 3 decades, water temperatures have increased by 0.46 ∘C per decade (±0.02 ∘C) in the epilimnion and 0.33 ∘C per decade (±0.06 ∘C) in the hypolimnion. Concomitantly and due to thermal change, O2 solubility has decreased by −0.104 mg L−1 per decade (±0.005 mg L−1) and −0.096 mg L−1 per decade (±0.011 mg L−1) in the epilimnion and hypolimnion, respectively. Based on the shared socio-economic pathway SSP370 of the Intergovernmental Panel on Climate Change (IPCC), peri-alpine lakes could face an increase of 3.80 ∘C (±0.20 ∘C) in the next 70 years, accompanied by a decline of 1.0 mg L−1 (±0.1 mg L−1) of O2 solubility. Together, these results highlight a critical alteration in lake thermal and oxygen conditions in the coming decades, and a need for a better integration of long-term lake observatories data and lake models to anticipate climate effects on lake thermal regimes and habitats.

Why do our rainfall–runoff models keep underestimating the peak flows?

Abstract: Abstract. In this paper, the question of how the interpolation of precipitation in space by using various spatial gauge densities affects the rainfall–runoff model discharge if all other input variables are kept constant is investigated. The main focus was on the peak flows. This was done by using a physically based model as the reference with a reconstructed spatially variable precipitation model and a conceptual model calibrated to match the reference model's output as closely as possible. Both models were run with distributed and lumped inputs. Results showed that all considered interpolation methods resulted in the underestimation of the total precipitation volume and that the underestimation was directly proportional to the precipitation amount. More importantly, the underestimation of peaks was very severe for low observation densities and disappeared only for very high-density precipitation observation networks. This result was confirmed by using observed precipitation with different observation densities. Model runoffs showed worse performance for their highest discharges. Using lumped inputs for the models showed deteriorating performance for peak flows as well, even when using simulated precipitation.

Hyper-resolution PCR-GLOBWB: opportunities and challenges from refining model spatial resolution to 1 km over the European continent

Abstract: Abstract. The quest for hydrological hyper-resolution modelling has been on-going for more than a decade. While global hydrological models (GHMs) have seen a reduction in grid size, they have thus far never been consistently applied at a hyper-resolution (<=1 km) at the large scale. Here, we present the first application of the GHM PCR-GLOBWB at 1 km over Europe. We thoroughly evaluated simulated discharge, evaporation, soil moisture, and terrestrial water storage anomalies against long-term observations and subsequently compared results with the established 10 and 50 km resolutions of PCR-GLOBWB. Subsequently, we could assess the added value of this first hyper-resolution version of PCR-GLOBWB and assess the scale dependencies of model and forcing resolution. Eventually, these insights can help us in understanding the current challenges and opportunities from hyper-resolution models and in formulating the model and data requirements for future improvements. We found that, for most variables, epistemic uncertainty is still large, and issues with scale commensurability exist with respect to the long-term yet coarse observations used. Merely for simulated discharge, we can confidently state that model output at hyper-resolution improves over coarser resolutions due to better representation of the river network at 1 km. However, currently available observations are not yet widely available at hyper-resolution or lack a sufficiently long time series, which makes it difficult to assess the performance of the model for other variables at hyper-resolution. Here, additional model validation efforts are needed. On the model side, hyper-resolution applications require careful revisiting of model parameterization and possibly the implementation of more physical processes to be able to resemble the dynamics and spatial heterogeneity at 1 km. With this first application of PCR-GLOBWB at 1 km, we contribute to meeting the grand challenge of hyper-resolution modelling. Even though the model was only assessed at the continental scale, valuable insights could be gained which have global validity. As such, it should be seen as a modest milestone on a longer journey towards locally relevant model output. This, however, requires a community effort from domain experts, model developers, research software engineers, and data providers.

Contribution of rock glacier discharge to late summer and fall streamflow in the Uinta Mountains, Utah, USA

Abstract: Abstract. Water draining from rock glaciers in the Uinta Mountains of Utah (USA) was analyzed and compared with samples of groundwater and water from the primary stream in a representative 5000 ha drainage. Rock glacier water resembles snowmelt in the early summer but evolves to higher values of d-excess and greatly elevated Ca and Mg content as the melt season progresses. This pattern is consistent with models describing a transition from snowmelt to melting of seasonal ice to melting of perennial ice in the rock glacier interior in late summer and fall. Water derived from this internal ice appears to have been the source of ∼25 % of the streamflow in this study area during September of 2021. This result emphasizes the significant role that rock glaciers can play in the hydrology of high-elevation watersheds, particularly in summers following a winter with below-average snowpack.

Comparison of artificial neural networks and reservoir models for simulating karst spring discharge on five test sites in the Alpine and Mediterranean regions

Abstract: Abstract. Hydrological models are widely used to characterize, understand and manage hydrosystems. Lumped parameter models are of particular interest in karst environments given the complexity and heterogeneity of these systems. There is a multitude of lumped parameter modelling approaches, which can make it difficult for a manager or researcher to choose. We therefore conducted a comparison of two lumped parameter modelling approaches: artificial neural networks (ANNs) and reservoir models. We investigate five karst systems in the Mediterranean and Alpine regions with different characteristics in terms of climatic conditions, hydrogeological properties and data availability. We compare the results of ANN and reservoir modelling approaches using several performance criteria over different hydrological periods. The results show that both ANNs and reservoir models can accurately simulate karst spring discharge but also that they have different advantages and drawbacks: (i) ANN models are very flexible regarding the format and amount of input data, (ii) reservoir models can provide good results even with a few years of relevant discharge in the calibration period and (iii) ANN models seem robust for reproducing high-flow conditions, while reservoir models are superior in reproducing low-flow conditions. However, both modelling approaches struggle to reproduce extreme events (droughts, floods), which is a known problem in hydrological modelling. For research purposes, ANN models have been shown to be useful for identifying recharge areas and delineating catchments, based on insights into the input data. Reservoir models are adapted to understand the hydrological functioning of a system by studying model structure and parameters.

A comprehensive open-source course for teaching applied hydrological modelling in Central Asia

Abstract: Abstract. Numerical modelling is often used for climate impact studies in water resources management. It is, however, not yet highly accessible to many students of hydrology in Central Asia. One big hurdle for new learners is the scripting requirement for the preparation of relevant data prior to the actual modelling. We present a robust, open-source workflow and comprehensive teaching material that aim at bridging the gap between theoretical hydrological modelling know-how and applied modelling for climate impact studies in Central Asia. The teaching material has been refined over 2 consecutive years and is being taken up by several professors teaching hydrological modelling in Central Asia.

Hybrid forecasting: blending climate predictions with AI models

Abstract: Abstract. Hybrid hydroclimatic forecasting systems employ data-driven (statistical or machine learning) methods to harness and integrate a broad variety of predictions from dynamical, physics-based models – such as numerical weather prediction, climate, land, hydrology, and Earth system models – into a final prediction product. They are recognized as a promising way of enhancing the prediction skill of meteorological and hydroclimatic variables and events, including rainfall, temperature, streamflow, floods, droughts, tropical cyclones, or atmospheric rivers. Hybrid forecasting methods are now receiving growing attention due to advances in weather and climate prediction systems at subseasonal to decadal scales, a better appreciation of the strengths of AI, and expanding access to computational resources and methods. Such systems are attractive because they may avoid the need to run a computationally expensive offline land model, can minimize the effect of biases that exist within dynamical outputs, benefit from the strengths of machine learning, and can learn from large datasets, while combining different sources of predictability with varying time horizons. Here we review recent developments in hybrid hydroclimatic forecasting and outline key challenges and opportunities for further research. These include obtaining physically explainable results, assimilating human influences from novel data sources, integrating new ensemble techniques to improve predictive skill, creating seamless prediction schemes that merge short to long lead times, incorporating initial land surface and ocean/ice conditions, acknowledging spatial variability in landscape and atmospheric forcing, and increasing the operational uptake of hybrid prediction schemes.

Bayesian calibration of a flood simulator using binary flood extent observations

Abstract: Abstract. Computational simulators of complex physical processes, such as inundations, require a robust characterization of the uncertainties involved to be useful for flood hazard and risk analysis. While flood extent data, as obtained from synthetic aperture radar (SAR) imagery, have become widely available, no methodologies have been implemented that can consistently assimilate this information source into fully probabilistic estimations of the model parameters, model structural deficiencies, and model predictions. This paper proposes a fully Bayesian framework to calibrate a 2D physics-based inundation model using a single observation of flood extent, explicitly including uncertainty in the floodplain and channel roughness parameters, simulator structural deficiencies, and observation errors. The proposed approach is compared to the current state-of-practice generalized likelihood uncertainty estimation (GLUE) framework for calibration and with a simpler Bayesian model. We found that discrepancies between the computational simulator output and the flood extent observation are spatially correlated, and calibration models that do not account for this, such as GLUE, may consistently mispredict flooding over large regions. The added structural deficiency term succeeds in capturing and correcting for this spatial behavior, improving the rate of correctly predicted pixels. We also found that binary data do not have information on the magnitude of the observed process (e.g., flood depths), raising issues in the identifiability of the roughness parameters, and the additive terms of structural deficiency and observation errors. The proposed methodology, while computationally challenging, is proven to perform better than existing techniques. It also has the potential to consistently combine observed flood extent data with other data such as sensor information and crowdsourced data, something which is not currently possible using GLUE calibration framework.

Improved soil evaporation remote sensing retrieval algorithms and associated uncertainty analysis on the Tibetan Plateau

Abstract: Abstract. Actual evapotranspiration (ET) is the key link between water and energy cycles. However, accurate evaporation estimation in alpine barren areas remains understudied. In this study, we aimed to improve the satellite-driven Process-based Land Surface ET/Heat fluxes algorithm (P-LSH) for better satellite retrieval of ET on the Tibetan Plateau by introducing two effective soil moisture constraint schemes in which normalized surface soil moisture and the ratio of cumulative antecedent precipitation to cumulative antecedent equilibrium evaporation are used to represent soil water stress, respectively, based on the intercomparison and knowledge-learning of the existing schemes. We first conducted intercomparison of six existing soil evaporation algorithms and sorted out the two most effective soil moisture constraint schemes. We then introduced the modified versions of the two constraint schemes into the P-LSH algorithm and further optimized the parameters using the differential evolution method. As a result, it formed two improved P-LSH algorithms. We systematically assessed the performances of the two improved P-LSH algorithms and six existing remote sensing ET retrieval algorithms on two barren-dominated basins of the Tibetan Plateau using reconstructed ET estimates derived from the terrestrial water balance method as a benchmark. The two moisture constraint schemes largely improved the performance of the P-LSH algorithm and showed better performance in both basins (root mean square error (RMSE) = 7.36 and 7.76 mm per month; R2=0.86 and 0.87), resulting in a higher simulation accuracy than all six existing algorithms. We used five soil moisture datasets and five precipitation datasets to further investigate the impact of moisture constraint uncertainty on the improved P-LSH algorithm. The ET estimates of the improved P-LSH algorithm, driven by the GLDAS_Noah soil moisture, performed best compared with those driven by other soil moisture and precipitation datasets, while ET estimates driven by various precipitation datasets generally showed a high and stable accuracy. These results suggest that high-quality soil moisture can optimally express moisture supply to ET, and that more accessible precipitation data can serve as a substitute for soil moisture as an indicator of moisture status for its robust performance in barren evaporation.

Reconstructing five decades of sediment export from two glacierized high-alpine catchments in Tyrol, Austria, using nonparametric regression

Abstract: Abstract. Knowledge on the response of sediment export to recent climate change in glacierized areas in the European Alps is limited, primarily because long-term records of suspended sediment concentrations (SSCs) are scarce. Here we tested the estimation of sediment export of the past five decades using quantile regression forest (QRF), a nonparametric, multivariate regression based on random forest. The regression builds on short-term records of SSCs and long records of the most important hydroclimatic drivers (discharge, precipitation and air temperature – QPT). We trained independent models for two nested and partially glacier-covered catchments, Vent (98 km2) and Vernagt (11.4 km2), in the upper Ötztal in Tyrol, Austria (1891 to 3772 m a.s.l.), where available QPT records start in 1967 and 1975. To assess temporal extrapolation ability, we used two 2-year SSC datasets at gauge Vernagt, which are almost 20 years apart, for a validation. For Vent, we performed a five-fold cross-validation on the 15 years of SSC measurements. Further, we quantified the number of days where predictors exceeded the range represented in the training dataset, as the inability to extrapolate beyond this range is a known limitation of QRF. Finally, we compared QRF performance to sediment rating curves (SRCs). We analyzed the modeled sediment export time series, the predictors and glacier mass balance data for trends (Mann–Kendall test and Sen's slope estimator) and step-like changes (using the widely applied Pettitt test and a complementary Bayesian approach). Our validation at gauge Vernagt demonstrated that QRF performs well in estimating past daily sediment export (Nash–Sutcliffe efficiency (NSE) of 0.73) and satisfactorily for SSCs (NSE of 0.51), despite the small training dataset. The temporal extrapolation ability of QRF was superior to SRCs, especially in periods with high-SSC events, which demonstrated the ability of QRF to model threshold effects. Days with high SSCs tended to be underestimated, but the effect on annual yields was small. Days with predictor exceedances were rare, indicating a good representativity of the training dataset. Finally, the QRF reconstruction models outperformed SRCs by about 20 percent points of the explained variance. Significant positive trends in the reconstructed annual suspended sediment yields were found at both gauges, with distinct step-like increases around 1981. This was linked to increased glacier melt, which became apparent through step-like increases in discharge at both gauges as well as change points in mass balances of the two largest glaciers in the Vent catchment. We identified exceptionally high July temperatures in 1982 and 1983 as a likely cause. In contrast, we did not find coinciding change points in precipitation. Opposing trends at the two gauges after 1981 suggest different timings of “peak sediment”. We conclude that, given large-enough training datasets, the presented QRF approach is a promising tool with the ability to deepen our understanding of the response of high-alpine areas to decadal climate change.

Knowledge-informed deep learning for hydrological model calibration: an application to Coal Creek Watershed in Colorado

Abstract: Abstract. Deep learning (DL)-assisted inverse mapping has shown promise in hydrological model calibration by directly estimating parameters from observations. However, the increasing computational demand for running the state-of-the-art hydrological model limits sufficient ensemble runs for its calibration. In this work, we present a novel knowledge-informed deep learning method that can efficiently conduct the calibration using a few hundred realizations. The method involves two steps. First, we determine decisive model parameters from a complete parameter set based on the mutual information (MI) between model responses and each parameter computed by a limited number of realizations (∼50). Second, we perform more ensemble runs (e.g., several hundred) to generate the training sets for the inverse mapping, which selects informative model responses for estimating each parameter using MI-based parameter sensitivity. We applied this new DL-based method to calibrate a process-based integrated hydrological model, the Advanced Terrestrial Simulator (ATS), at Coal Creek Watershed, CO. The calibration is performed against observed stream discharge (Q) and remotely sensed evapotranspiration (ET) from the water year 2017 to 2019. Preliminary MI analysis on 50 realizations resulted in a down-selection of 7 out of 14 ATS model parameters. Then, we performed a complete MI analysis on 396 realizations and constructed the inverse mapping from informative responses to each of the selected parameters using a deep neural network. Compared with calibration using observations covering all time steps, the new inverse mapping improves parameter estimations, thus enhancing the performance of ATS forward model runs. The Nash–Sutcliffe efficiency (NSE) of streamflow predictions increases from 0.53 to 0.8 when calibrating against Q alone. Using ET observations, on the other hand, does not show much improvement on the performance of ATS modeling mainly due to both the uncertainty of the remotely sensed product and the insufficient coverage of the model ET ensemble in capturing the observation. By using observed Q only, we further performed a multiyear analysis and show that Q is best simulated (NSE > 0.8) by including in the calibration the dry-year flow dynamics that show more sensitivity to subsurface characteristics than the other wet years. Moreover, when continuing the forward runs till the end of 2021, the calibrated models show similar simulation performances during this evaluation period as the calibration period, demonstrating the ability of the estimated parameters in capturing climate sensitivity. Our success highlights the importance of leveraging data-driven knowledge in DL-assisted hydrological model calibration.

Local moisture recycling across the globe

Abstract: Abstract. Changes in evaporation over land affect terrestrial precipitation via atmospheric moisture recycling and, consequently, freshwater availability. Although global moisture recycling at regional and continental scales is relatively well understood, the patterns of local moisture recycling and the main variables that impact it remain unknown. We calculate the local moisture recycling ratio (LMR) as the fraction of evaporated moisture that precipitates within a distance of 0.5∘ (typically 50 km) of its source, identify variables that correlate with it over land globally, and study its model dependency. We derive the seasonal and annual LMR using a 10-year climatology (2008–2017) of monthly averaged atmospheric moisture connections at a scale of 0.5∘ obtained from a Lagrangian atmospheric moisture tracking model. We find that, annually, an average of 1.7 % (SD of 1.1 %) of evaporated moisture returns as precipitation locally, although with large temporal and spatial variability, and the LMR peaks in summer and over wet and mountainous regions. Our results show that wetness, orography, latitude, convective available potential energy, wind speed, and total cloud cover correlate clearly with the LMR, indicating that wet regions with little wind and strong ascending air are particularly favourable for a high LMR. Finally, we find that spatial patterns of local recycling are consistent between different models, yet the magnitude of recycling varies. Our results can be used to study the impacts of evaporation changes on local precipitation, with implications for, for example, regreening and water management.

Water level variation at a beaver pond significantly impacts net CO₂ uptake of a continental bog

Abstract: Abstract. The carbon (C) dynamics of northern peatlands are sensitive to hydrological changes owing to ecohydrological feedbacks. We quantified and evaluated the impact of water level variations in a beaver pond (BP) on the CO2 flux dynamics of an adjacent, raised Sphagnum–shrub-dominated bog in southern Canada. We applied the CoupModel to the Mer Bleue bog, where the hydrological, energy and CO2 fluxes have been measured continuously for over 20 years. The lateral flow of water from the bog to the BP was estimated by the hydraulic gradient between the peatland and the BP's water level and the vertical profile of peat hydraulic conductivity. The model outputs were compared with the measured hydrological components, CO2 flux and energy flux data (1998–2019). CoupModel was able to reproduce the measured data well. The simulation shows that variation in the BP water level (naturally occurring or due to management) influenced the bog net ecosystem exchange (NEE) of CO2. Over 1998–2004, the BP water level was 0.75 to 1.0 m lower than during 2017–2019. Simulated net CO2 uptake was 55 gCm-2yr-1 lower during 1998–2004 compared to 2017–2019 when there was no BP disturbance, which was similar to the differences in measured NEE between those periods. Peatland annual NEE was well correlated with water table depth (WTD) within the bog, and NEE also shows a linear relation with the water level at the BP, with a slope of −120 gCO2-Cm-2yr-1m-1. The current modelling predicts that the bog may switch from CO2 sink to source when the BP water levels drop lower than ∼ 1.7 m below the peat surface at the eddy covariance (EC) tower, located on the bog surface 250 m from the BP. This study highlights the importance of natural and human disturbances to adjacent water bodies in regulating the net CO2 uptake function of northern peatlands.

A snow and glacier hydrological model for large catchments – case study for the Naryn River, central Asia

Abstract: Abstract. In this paper we implement a degree day snowmelt and glacier melt model in the Dynamic fluxEs and ConnectIvity for Predictions of HydRology (DECIPHeR) model. The purpose is to develop a hydrological model that can be applied to large glaciated and snow-fed catchments yet is computationally efficient enough to include model uncertainty in streamflow predictions. The model is evaluated by simulating monthly discharge at six gauging stations in the Naryn River catchment (57 833 km2) in central Asia over the period 1951 to a variable end date between 1980 and 1995 depending on the availability of discharge observations. The spatial distribution of simulated snow cover is validated against MODIS weekly snow extent for the years 2001–2007. Discharge is calibrated by selecting parameter sets using Latin hypercube sampling and assessing the model performance using six evaluation metrics. The model shows good performance in simulating monthly discharge for the calibration period (NSE is 0.74

Assessing specific differential phase (KDP)-based quantitative precipitation estimation for the record- breaking rainfall over Zhengzhou city on 20 July 2021

Abstract: Abstract. Although radar-based quantitative precipitation estimation (QPE) has been widely investigated from various perspectives, very few studies have been devoted to extreme-rainfall QPE. In this study, the performance of specific differential phase (KDP)-based QPE during the record-breaking Zhengzhou rainfall event that occurred on 20 July 2021 is assessed. Firstly, the OTT Parsivel disdrometer (OTT) observations are used as input for T-matrix simulation, and different assumptions are made to construct R(KDP) estimators. KDP estimates from three algorithms are then compared in order to obtain the best KDP estimates, and gauge observations are used to evaluate the R(KDP) estimates. Our results generally agree with previous known-truth tests and provide more practical insights from the perspective of QPE applications. For rainfall rates below 100 mm h−1, the R(KDP) agrees rather well with the gauge observations, and the selection of the KDP estimation method or controlling factor has a minimal impact on the QPE performance provided that the controlling factor used is not too extreme. For higher rain rates, a significant underestimation is found for the R(KDP), and a smaller window length results in a higher KDP and, thus, less underestimation of rain rates. We show that the QPE based on the “best KDP estimate” cannot reproduce the gauge measurement of 201.9 mm h−1 with commonly used assumptions for R(KDP), and the potential factors responsible for this result are discussed. We further show that the gauge with the 201.9 mm h−1 report was in the vicinity of local rainfall hot spots during the 16:00–17:00 LST period, while the 3 h rainfall accumulation center was located southwest of Zhengzhou city.

SMPD: a soil moisture-based precipitation downscaling method for high-resolution daily satellite precipitation estimation

Abstract: Abstract. As a key component in the water and energy cycle, estimates of precipitation with high resolution and accuracy is of great significance for hydrological, meteorological, and ecological studies. However, current satellite-based precipitation products have a coarse spatial resolution (from 10 to 50 km) not meeting the needs of several applications (e.g., flash floods and landslides). The implementation of spatial downscaling methods can be a suitable approach to overcome this shortcoming. In this study, we developed a soil moisture-based precipitation downscaling (SMPD) method for spatially downscaling the integrated multisatellite retrievals for global precipitation measurement (IMERG) V06B daily precipitation product over a complex topographic and climatic area in southwestern Europe (Iberian Peninsula) in the period 2016–2018. By exploiting the soil-water balance equation, high-resolution surface soil moisture (SSM) and normalized difference vegetation index (NDVI) products were used as auxiliary variables. The spatial resolution of the IMERG daily precipitation product was downscaled from 10 to 1 km. An evaluation using 1027 rain gauge stations highlighted the good performance of the downscaled 1 km IMERG product compared to the original 10 km product, with a correlation coefficient of 0.61, root mean square error (RMSE) of 4.83 mm and a relative bias of 5 %. Meanwhile, the 1 km downscaled results can also capture the typical temporal and spatial variation behaviors of precipitation in the study area during dry and wet seasons. Overall, the SMPD method greatly improves the spatial details of the original 10 km IMERG product also with a slight enhancement of accuracy. It shows good potential to be applied for the development of high-quality and high-resolution precipitation products in any region of interest.

Seasonal forecasting of snow resources at Alpine sites

Abstract: Abstract. Climate warming in mountain regions is resulting in glacier shrinking, seasonal snow cover reduction, and changes in the amount and seasonality of meltwater runoff, with consequences on water availability. Droughts are expected to become more severe in the future with economical and environmental losses both locally and downstream. Effective adaptation strategies involve multiple timescales, and seasonal forecasts can help in the optimization of the available snow and water resources with a lead time of several months. We developed a prototype to generate seasonal forecasts of snow depth and snow water equivalent with a starting date of 1 November and a lead time of 7 months, so up to 31 May of the following year. The prototype has been co-designed with end users in the field of water management, hydropower production and mountain ski tourism, meeting their needs in terms of indicators, time resolution of the forecasts and visualization of the forecast outputs. In this paper we present the modelling chain, based on the seasonal forecasts of the ECMWF and Météo-France seasonal prediction systems, made available through the Copernicus Climate Change Service (C3S) Climate Data Store. Seasonal forecasts of precipitation, near-surface air temperature, radiative fluxes, wind and relative humidity are bias-corrected and downscaled to three sites in the Western Italian Alps and finally used as input for the physically based multi-layer snow model SNOWPACK. Precipitation is bias-corrected with a quantile mapping method using ERA5 reanalysis as a reference and then downscaled with the RainFARM stochastic procedure in order to allow an estimate of uncertainties due to the downscaling method. The impacts of precipitation bias adjustment and downscaling on the forecast skill have been investigated. The skill of the prototype in predicting the deviation of monthly snow depth with respect to the normal conditions from November to May in each season of the hindcast period 1995–2015 is demonstrated using both deterministic and probabilistic metrics. Forecast skills are determined with respect to a simple forecasting method based on the climatology, and station measurements are used as reference data. The prototype shows good skills at predicting the tercile category, i.e. snow depth below and above normal, in the winter (lead times: 2–3–4 months) and spring (lead times: 5–6–7 months) ahead: snow depth is predicted with higher accuracy (Brier skill score) and higher discrimination (area under the relative operating characteristics (ROC) curve skill score) with respect to a simple forecasting method based on the climatology. Ensemble mean monthly snow depth forecasts are significantly correlated with observations not only at short lead times of 1 and 2 months (November and December) but also at lead times of 5 and 6 months (March and April) when employing the ECMWFS5 forcing. Moreover the prototype shows skill at predicting extremely dry seasons, i.e. seasons with snow depth below the 10th percentile, while skills at predicting snow depth above the 90th percentile are model-, station- and score-dependent. The bias correction of precipitation forecasts is essential in the case of large biases in the global seasonal forecast system (MFS6) to reconstruct a realistic snow depth climatology; however, no remarkable differences are found among the skill scores when the precipitation input is bias-corrected, downscaled, or bias-corrected and downscaled, compared to the case in which raw data are employed, suggesting that skill scores are weakly sensitive to the treatment of the precipitation input.