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.

Despite the proliferation of computer-based research on hydrology and water resources, such research is typically poorly reproducible. Published studies have low reproducibility due to incomplete availability of data and computer code, and a lack of documentation of workflow processes. This leads to a lack of transparency and efficiency because existing code can neither be quality controlled nor reused. Given the commonalities between existing process-based hydrologic models in terms of their required input data and preprocessing steps, open sharing of code can lead to large efficiency gains for the modeling community. Here, we present a model configuration workflow that provides full reproducibility of the resulting model instantiations in a way that separates the model-agnostic preprocessing of specific data sets from the model-specific requirements that models impose on their input files. We use this workflow to create large-domain (global and continental) and local configurations of the Structure for Unifying Multiple Modeling Alternatives (SUMMA) hydrologic model connected to the mizuRoute routing model. These examples show how a relatively complex model setup over a large domain can be organized in a reproducible and structured way that has the potential to accelerate advances in hydrologic modeling for the community as a whole. We provide a tentative blueprint of how community modeling initiatives can be built on top of workflows such as this. We term our workflow the “Community Workflows to Advance Reproducibility in Hydrologic Modeling” (CWARHM; pronounced “swarm”).

Abstract. Floods cause extensive damage, especially if they affect large regions. Assessments of current, local, and regional flood hazards and their future changes often involve the use of hydrologic models. A reliable hydrologic model ideally reproduces both local flood characteristics and spatial aspects of flooding under current and future climate conditions. However, uncertainties in simulated floods can be considerable and yield unreliable hazard and climate change impact assessments. This study evaluates the extent to which models calibrated according to standard model calibration metrics such as the widely used Kling–Gupta efficiency are able to capture flood spatial coherence and triggering mechanisms. To highlight challenges related to flood simulations, we investigate how flood timing, magnitude, and spatial variability are represented by an ensemble of hydrological models when calibrated on streamflow using the Kling–Gupta efficiency metric, an increasingly common metric of hydrologic model performance also in flood-related studies. Specifically, we compare how four well-known models (the Sacramento Soil Moisture Accounting model, SAC; the Hydrologiska Byråns Vattenbalansavdelning model, HBV; the variable infiltration capacity model, VIC; and the mesoscale hydrologic model, mHM) represent (1) flood characteristics and their spatial patterns and (2) how they translate changes in meteorologic variables that trigger floods into changes in flood magnitudes. Our results show that both the modeling of local and spatial flood characteristics are challenging as models underestimate flood magnitude, and flood timing is not necessarily well captured. They further show that changes in precipitation and temperature are not always well translated to changes in flood flow, which makes local and regional flood hazard assessments even more difficult for future conditions. From a large sample of catchments and with multiple models, we conclude that calibration on the integrated Kling–Gupta metric alone is likely to yield models that have limited reliability in flood hazard assessments, undermining their utility for regional and future change assessments. We underscore that such assessments can be improved by developing flood-focused, multi-objective, and spatial calibration metrics, by improving flood generating process representation through model structure comparisons and by considering uncertainty in precipitation input.

Abstract. Probabilistic methods are useful to estimate the uncertainty in spatial meteorological fields (e.g., the uncertainty in spatial patterns of precipitation and temperature across large domains). In ensemble probabilistic methods, “equally plausible” ensemble members are used to approximate the probability distribution, hence the uncertainty, of a spatially distributed meteorological variable conditioned to the available information. The ensemble members can be used to evaluate the impact of uncertainties in spatial meteorological fields for a myriad of applications. This study develops the Ensemble Meteorological Dataset for North America (EMDNA). EMDNA has 100 ensemble members with daily precipitation amount, mean daily temperature, and daily temperature range at 0.1∘ spatial resolution (approx. 10 km grids) from 1979 to 2018, derived from a fusion of station observations and reanalysis model outputs. The station data used in EMDNA are from a serially complete dataset for North America (SCDNA) that fills gaps in precipitation and temperature measurements using multiple strategies. Outputs from three reanalysis products are regridded, corrected, and merged using Bayesian model averaging. Optimal interpolation (OI) is used to merge station- and reanalysis-based estimates. EMDNA estimates are generated using spatiotemporally correlated random fields to sample from the OI estimates. Evaluation results show that (1) the merged reanalysis estimates outperform raw reanalysis estimates, particularly in high latitudes and mountainous regions; (2) the OI estimates are more accurate than the reanalysis and station-based regression estimates, with the most notable improvements for precipitation evident in sparsely gauged regions; and (3) EMDNA estimates exhibit good performance according to the diagrams and metrics used for probabilistic evaluation. We discuss the limitations of the current framework and highlight that further research is needed to improve ensemble meteorological datasets. Overall, EMDNA is expected to be useful for hydrological and meteorological applications in North America. The entire dataset and a teaser dataset (a small subset of EMDNA for easy download and preview) are available at https://doi.org/10.20383/101.0275 (Tang et al., 2020a).

Abstract. Floods cause large damages, especially if they affect large regions. Assessments of current, local and regional flood hazards and their future changes often involve the use of hydrologic models. However, uncertainties in simulated floods can be considerable and yield unreliable hazard and climate change impact assessments. A reliable hydrologic model ideally reproduces both local flood characteristics and spatial aspects of flooding, which is, however, not guaranteed especially when using standard model calibration metrics. In this paper we investigate how flood timing, magnitude and spatial variability are represented by an ensemble of hydrological models when calibrated on streamflow using the Kling–Gupta efficiency metric, an increasingly common metric of hydrologic model performance. We compare how four well-known models (SAC, HBV, VIC, and mHM) represent (1) flood characteristics and their spatial patterns; and (2) how they translate changes in meteorologic variables that trigger floods into changes in flood magnitudes. Our results show that both the modeling of local and spatial flood characteristics is challenging. They further show that changes in precipitation and temperature are not necessarily well translated to changes in flood flow, which makes local and regional flood hazard assessments even more difficult for future conditions. We conclude that models calibrated on integrated metrics such as the Kling–Gupta efficiency alone have limited reliability in flood hazard assessments, in particular in regional and future assessments, and suggest the development of alternative process-based and spatial evaluation metrics.

Abstract. Probabilistic methods are very useful to estimate the spatial variability in meteorological conditions (e.g., spatial patterns of precipitation and temperature across large domains). In ensemble probabilistic methods, equally plausible ensemble members are used to approximate the probability distribution, hence uncertainty, of a spatially distributed meteorological variable conditioned on the available information. The ensemble can be used to evaluate the impact of the uncertainties in a myriad of applications. This study develops the Ensemble Meteorological Dataset for North America (EMDNA). EMDNA has 100 members with daily precipitation amount, mean daily temperature, and daily temperature range at 0.1° spatial resolution from 1979 to 2018, derived from a fusion of station observations and reanalysis model outputs. The station data used in EMDNA are from a serially complete dataset for North America (SCDNA) that fills gaps in precipitation and temperature measurements using multiple strategies. Outputs from three reanalysis products are regridded, corrected, and merged using the Bayesian Model Averaging. Optimal Interpolation (OI) is used to merge station- and reanalysis-based estimates. EMDNA estimates are generated based on OI estimates and spatiotemporally correlated random fields. Evaluation results show that (1) the merged reanalysis estimates outperform raw reanalysis estimates, particularly in high latitudes and mountainous regions; (2) the OI estimates are more accurate than the reanalysis and station-based regression estimates, with the most notable improvement for precipitation occurring in sparsely gauged regions; and (3) EMDNA estimates exhibit good performance according to the diagrams and metrics used for probabilistic evaluation. We also discuss the limitations of the current framework and highlight that persistent efforts are needed to further develop probabilistic methods and ensemble datasets. Overall, EMDNA is expected to be useful for hydrological and meteorological applications in North America. The whole dataset and a teaser dataset (a small subset of EMDNA for easy download and preview) are available at https://doi.org/10.20383/101.0275 (Tang et al., 2020a).

Abstract. Station-based serially complete datasets (SCDs) of precipitation and temperature observations are important for hydrometeorological studies. Motivated by the lack of serially complete station observations for North America, this study seeks to develop an SCD from 1979 to 2018 from station data. The new SCD for North America (SCDNA) includes daily precipitation, minimum temperature (Tmin⁡), and maximum temperature (Tmax⁡) data for 27 276 stations. Raw meteorological station data were obtained from the Global Historical Climate Network Daily (GHCN-D), the Global Surface Summary of the Day (GSOD), Environment and Climate Change Canada (ECCC), and a compiled station database in Mexico. Stations with at least 8-year-long records were selected, which underwent location correction and were subjected to strict quality control. Outputs from three reanalysis products (ERA5, JRA-55, and MERRA-2) provided auxiliary information to estimate station records. Infilling during the observation period and reconstruction beyond the observation period were accomplished by combining estimates from 16 strategies (variants of quantile mapping, spatial interpolation, and machine learning). A sensitivity experiment was conducted by assuming that 30 % of observations from stations were missing – this enabled independent validation and provided a reference for reconstruction. Quantile mapping and mean value corrections were applied to the final estimates. The median Kling–Gupta efficiency (KGE′) values of the final SCDNA for all stations are 0.90, 0.98, and 0.99 for precipitation, Tmin⁡, and Tmax⁡, respectively. The SCDNA is closer to station observations than the four benchmark gridded products and can be used in applications that require either quality-controlled meteorological station observations or reconstructed long-term estimates for analysis and modeling. The dataset is available at https://doi.org/10.5281/zenodo.3735533 (Tang et al., 2020).

Floods often affect large regions and cause adverse societal impacts. Regional flood hazard and risk assessments therefore require a realistic representation of spatial flood dependencies to avoid the overestimation or underestimation of risk. However, it is not yet well understood how spatial flood dependence, that is, the degree of co-occurrence of floods at different locations, varies in space and time and which processes influence the strength of this dependence. We identify regions in the United States with seasonally similar flood behavior and analyze processes governing spatial dependence. We find that spatial flood dependence varies regionally and seasonally and is generally strongest in winter and spring and weakest in summer and fall. Moreover, we find that land-surface processes are crucial in shaping the spatiotemporal characteristics of flood events. We conclude that the regional and seasonal variations in spatial flood dependencies must be considered when conducting current and future flood risk assessments.

It is challenging to develop observationally based spatial estimates of meteorology in Alaska and the Yukon. Complex topography, frozen precipitation undercatch, and extremely sparse in situ observations all limit our capability to produce accurate spatial estimates of meteorological conditions. In this Arctic environment, it is necessary to develop probabilistic estimates of precipitation and temperature that explicitly incorporate spatiotemporally varying uncertainty and bias corrections. In this paper we exploit the recently developed ensemble Climatologically Aided Interpolation (eCAI) system to produce daily historical estimates of precipitation and temperature across Alaska and the Yukon Territory at a 2 km grid spacing for the time period 1980–2013. We extend the previous eCAI method to address precipitation gauge undercatch and wetting loss, which is of high importance for this high-latitude region where much of the precipitation falls as snow. Leave-one-out cross-validation shows our ensemble has little bias in daily precipitation and mean temperature at the station locations, with an overestimate in the daily standard deviation of precipitation. The ensemble is statistically reliable compared to climatology and can discriminate precipitation events across different precipitation thresholds. Long-term mean loss adjusted precipitation is up to 36% greater than the unadjusted estimate in windy areas that receive a large fraction of frozen precipitation, primarily due to wind induced undercatch. Comparing the ensemble mean climatology of precipitation and temperature to PRISM and Daymet v3 shows large interproduct differences, particularly in precipitation across the complex terrain of southeast and northern Alaska.

Abstract. Calibration is an essential step for improving the accuracy of simulations generated using hydrologic models. A key modeling decision is selecting the performance metric to be optimized. It has been common to use squared error performance metrics, or normalized variants such as Nash–Sutcliffe efficiency (NSE), based on the idea that their squared-error nature will emphasize the estimates of high flows. However, we conclude that NSE-based model calibrations actually result in poor reproduction of high-flow events, such as the annual peak flows that are used for flood frequency estimation. Using three different types of performance metrics, we calibrate two hydrological models at a daily step, the Variable Infiltration Capacity (VIC) model and the mesoscale Hydrologic Model (mHM), and evaluate their ability to simulate high-flow events for 492 basins throughout the contiguous United States. The metrics investigated are (1) NSE, (2) Kling–Gupta efficiency (KGE) and its variants, and (3) annual peak flow bias (APFB), where the latter is an application-specific metric that focuses on annual peak flows. As expected, the APFB metric produces the best annual peak flow estimates; however, performance on other high-flow-related metrics is poor. In contrast, the use of NSE results in annual peak flow estimates that are more than 20 % worse, primarily due to the tendency of NSE to underestimate observed flow variability. On the other hand, the use of KGE results in annual peak flow estimates that are better than from NSE, owing to improved flow time series metrics (mean and variance), with only a slight degradation in performance with respect to other related metrics, particularly when a non-standard weighting of the components of KGE is used. Stochastically generated ensemble simulations based on model residuals show the ability to improve the high-flow metrics, regardless of the deterministic performances. However, we emphasize that improving the fidelity of streamflow dynamics from deterministically calibrated models is still important, as it may improve high-flow metrics (for the right reasons). Overall, this work highlights the need for a deeper understanding of performance metric behavior and design in relation to the desired goals of model calibration.