Seasonal streamflow forecasts using mixture-kernel GPR and advanced methods of input variable selection

2018 ◽  
Vol 50 (1) ◽  
pp. 200-214 ◽  
Author(s):  
Shuang Zhu ◽  
Xiangang Luo ◽  
Zhanya Xu ◽  
Lei Ye
Keyword(s):  
Variable Selection ◽  
Gaussian Process Regression ◽  
Quadratic Term ◽  
Streamflow Prediction ◽  
Factors Analysis ◽  
Streamflow Forecasts ◽  
Random Components ◽  
Flexible Framework ◽  
High Flows ◽  
Seasonal Streamflow

Abstract Gaussian Process Regression (GPR) is a new machine-learning method based on Bayesian theory and statistical learning theory. It provides a flexible framework for probabilistic regression and uncertainty estimation. The main effort in GPR modelling is determining the structure of the kernel function. As streamflow is composed of trend, period and random components. In this study, we constructed a mixture-kernel composed of squared exponential kernel, periodic kernel and a rational quadratic term to reflect different properties of streamflow time series to make streamflow forecasts. A relevant feature-selection wrapper algorithm was used, with a top-down search for relevant features by Random Forest, to offer a systematic factors analysis that can potentially affect basin streamflow predictability. Streamflow prediction is evaluated by putting emphasis on the degree of coincidence, the deviation on low flows, high flows and the error level. The objective of this study is to construct a seasonal streamflow forecasts model using mixture-kernel GPR and the advanced input variable selection method. Results show that the mixture-kernel GPR has good forecasting quality, and top importance predictors are streamflow at 12, 6, 5, 1, 11, 7, 8, 4 months ahead, Nino 1 + 2 at 11, 5, 12, 10 months ahead.

scholarly journals Skilful seasonal forecasts of streamflow over Europe?

2017 ◽  
Author(s):  
Louise Arnal ◽  
Hannah L. Cloke ◽  
Elisabeth Stephens ◽  
Fredrik Wetterhall ◽  
Christel Prudhomme ◽  
...  
Keyword(s):  
Lead Time ◽  
Climate Model ◽  
Streamflow Forecasting ◽  
Streamflow Prediction ◽  
Seasonal Forecasts ◽  
Seasonal Climate ◽  
Climate Forecasts ◽  
Streamflow Forecasts ◽  
Meteorological Observations ◽  
Seasonal Streamflow

Abstract. This paper presents a Europe-wide analysis of the skill of the newly operational EFAS (European Flood Awareness System) seasonal streamflow forecasts, benchmarked against the Ensemble Streamflow Prediction (ESP) forecasting approach. The results suggest that, on average, the System 4 seasonal climate forecasts improve the streamflow predictability over historical meteorological observations for the first month of lead time only. However, the predictability varies in space and time and is greater in winter and autumn. Parts of Europe additionally exhibit a longer predictability, up to seven months of lead time, for certain months within a season. The results also highlight the potential usefulness of the EFAS seasonal streamflow forecasts for decision-making. Although the ESP is the most potentially useful forecasting approach in Europe, the EFAS seasonal streamflow forecasts appear more potentially useful than the ESP in some regions and for certain seasons, especially in winter for most of Europe. Patterns in the EFAS seasonal streamflow hindcasts skill are however not mirrored in the System 4 seasonal climate hindcasts, hinting the need for a better understanding of the link between hydrological and meteorological variables on seasonal timescales, with the aim to improve climate-model based seasonal streamflow forecasting.

scholarly journals The benefit of using an ensemble of seasonal streamflow forecasts in water allocation decisions

2020 ◽  
Vol 24 (7) ◽  
pp. 3851-3870
Author(s):  
Alexander Kaune ◽  
Faysal Chowdhury ◽  
Micha Werner ◽  
James Bennett
Keyword(s):  
Water Allocation ◽  
Irrigation Season ◽  
Streamflow Forecasting ◽  
Streamflow Prediction ◽  
Additional Time ◽  
Available Water ◽  
Streamflow Forecasts ◽  
Reservoir Inflow ◽  
Seasonal Streamflow ◽  
Allocation Decisions

Abstract. The area to be cropped in irrigation districts needs to be planned according to the allocated water, which in turn is a function of the available water resource. Initially conservative estimates of future (in)flows in rivers and reservoirs may lead to unnecessary reduction of the water allocated. Though water allocations may be revised as the season progresses, inconsistency in allocation is undesirable to farmers as they may then not be able to use that water, leading to an opportunity cost in agricultural production. We assess the benefit of using reservoir inflow estimates derived from seasonal forecast datasets to improve water allocation decisions. A decision model is developed to emulate the feedback loop between simulated reservoir storage and water allocations to irrigated crops and is evaluated using inflow forecasts generated with the Forecast Guided Stochastic Scenarios (FoGSS) model, a 12-month ensemble streamflow forecasting system. Two forcings are used to generate the forecasts: ensemble streamflow prediction – ESP (historical rainfall) – and POAMA (calibrated rainfall forecasts from the POAMA climate prediction system). We evaluate the approach in the Murrumbidgee basin in Australia, comparing water allocations obtained with an expected reservoir inflow from FoGSS against the allocations obtained with the currently used conservative estimate based on climatology as well as against allocations obtained using observed inflows (perfect information). The inconsistency in allocated water is evaluated by determining the total changes in allocated water made every 15 d from the initial allocation at the start of the water year to the end of the irrigation season, including both downward and upward revisions of allocations. Results show that the inconsistency due to upward revisions in allocated water is lower when using the forecast datasets (POAMA and ESP) compared to the conservative inflow estimates (reference), which is beneficial to the planning of cropping areas by farmers. Overconfidence can, however, lead to an increase in undesirable downward revisions. This is more evident for dry years than for wet years. Over the 28 years for which allocation decisions are evaluated, we find that the accuracy of the available water estimates using the forecast ensemble improves progressively during the water year, especially 1.5 months before the start of the cropping season in November. This is significant as it provides farmers with additional time to make key decisions on planting. Our results show that seasonal streamflow forecasts can provide benefit in informing water allocation policies, particularly by earlier establishing final water allocations to farmers in the irrigation season. This allows them to plan better and use water allocated more efficiently.

scholarly journals An Efficient Approach for Estimating Streamflow Forecast Skill Elasticity

2017 ◽  
Vol 18 (6) ◽  
pp. 1715-1729 ◽  
Author(s):  
Louise Arnal ◽  
Andrew W. Wood ◽  
Elisabeth Stephens ◽  
Hannah L. Cloke ◽  
Florian Pappenberger
Keyword(s):  
Sensitivity Analysis ◽  
Forecast Skill ◽  
Sensitivity Analyses ◽  
Streamflow Prediction ◽  
Streamflow Forecast ◽  
Analysis Technique ◽  
Climate Forecasts ◽  
Modeling And Forecasting ◽  
Streamflow Forecasts ◽  
Seasonal Streamflow

Abstract Seasonal streamflow prediction skill can derive from catchment initial hydrological conditions (IHCs) and from the future seasonal climate forecasts (SCFs) used to produce the hydrological forecasts. Although much effort has gone into producing state-of-the-art seasonal streamflow forecasts from improving IHCs and SCFs, these developments are expensive and time consuming and the forecasting skill is still limited in most parts of the world. Hence, sensitivity analyses are crucial to funnel the resources into useful modeling and forecasting developments. It is in this context that a sensitivity analysis technique, the variational ensemble streamflow prediction assessment (VESPA) approach, was recently introduced. VESPA can be used to quantify the expected improvements in seasonal streamflow forecast skill as a result of realistic improvements in its predictability sources (i.e., the IHCs and the SCFs)—termed “skill elasticity”—and to indicate where efforts should be targeted. The VESPA approach is, however, computationally expensive, relying on multiple hindcasts having varying levels of skill in IHCs and SCFs. This paper presents two approximations of the approach that are computationally inexpensive alternatives. These new methods were tested against the original VESPA results using 30 years of ensemble hindcasts for 18 catchments of the contiguous United States. The results suggest that one of the methods, end point blending, is an effective alternative for estimating the forecast skill elasticities yielded by the VESPA approach. The results also highlight the importance of the choice of verification score for a goal-oriented sensitivity analysis.

scholarly journals Skill of a global forecasting system in seasonal ensemble streamflow prediction

2016 ◽  
Author(s):  
Naze Candogan Yossef ◽  
Rens van Beek ◽  
Albrecht Weerts ◽  
Hessel Winsemius ◽  
Marc F. P. Bierkens
Keyword(s):  
Information Service ◽  
Brier Score ◽  
Streamflow Prediction ◽  
Seasonal Forecasts ◽  
Ensemble Forecasts ◽  
Streamflow Forecasts ◽  
Forecasting System ◽  
Set Up ◽  
Framework Programme Project ◽  
Seasonal Streamflow

Abstract. In this study we assess the skill of seasonal streamflow forecasts with the global hydrological forecasting system FEWS-World which has been set up within the European Commission 7th Framework Programme project Global Water Scarcity Information Service (GLOWASIS). FEWS-World incorporates the global hydrological model PCR-GLOBWB. We produce ensemble forecasts of monthly discharges for 20 large rivers of the world, with lead times of up to 6 months, forcing the system with bias-corrected seasonal meteorological forecast ensembles from the ECMWF and with probabilistic meteorological ensembles obtained following the ESP procedure. Here, the skill from the ESP ensembles, which contain no actual information on weather, serves as a benchmark to assess the additional skill that may be obtained using ECMWF seasonal forecasts. We use the Brier Score to quantify the skill of the system in forecasting high and low flows, defined as discharges higher than the 75th and lower than the 25th percentiles for a given month respectively. We determine the theoretical skill by comparing the results against model simulations and the actual skill in comparison to discharge observations. We calculate the ratios of actual to theoretical skill in order to quantify the percentage of the theoretical skill that is achieved. The results suggest that the skill of ECMWF S3 forecasts is close to that of the ESP forecasts. While better meteorological forecasts could potentially lead to an improvement in hydrological forecasts, this cannot be achieved yet using the ECMWF S3 dataset.

The role of parameter estimation strategies on ensemble streamflow prediction results across extratropical Andean catchments

2020 ◽  
Author(s):  
Eduardo Muñoz-Castro ◽  
Pablo A. Mendoza ◽  
Ximena Vargas
Keyword(s):  
Hydrologic Model ◽  
Snow Accumulation ◽  
Weighting Coefficient ◽  
Parameter Estimates ◽  
Model Parameters ◽  
High Mountain ◽  
Streamflow Prediction ◽  
Streamflow Forecasts ◽  
Ensemble Streamflow Prediction ◽  
Seasonal Streamflow

<p>In catchments with a highly variable flow regime, an accurate and reliable hydrological forecasting framework is critical to support water resources management. However, due to model structural deficiencies and changing climatic conditions, the parameter estimates during the calibration period are expected to vary with hydrological conditions. This work aims to test the added value of incorporating potential non-stationarities in hydrologic model parameters on seasonal streamflow forecasts in high-mountain environments, using the ensemble streamflow prediction (ESP) methodology. To this end, we apply the GR4J rainfall-runoff model coupled with the snow accumulation and ablation CemaNeige module in six basins located in Central Chile (30-36° S). We explore the effects of four parameter selection strategies on the quality of seasonal streamflow forecasts produced with the ESP method: (i) a single set of parameters for the entire hindcast period (our benchmark), (ii) using parameters calibrated with a ‘leave-one-year-out’ approach, (iii) using parameter sets based on expected hydroclimatic conditions, and (iv) dual data assimilation to improve the initial condition and parameters before the forecast initialization. Results show that parameters related to production store capacity in GR4J model, and degree-day melt coefficient and weighting coefficient for snow pack thermal state in the CemaNeige module have a high inter-annual variability, with variations of 50% with respect to the benchmark scenario.</p>

scholarly journals Seasonal streamflow forecasting by conditioning climatology with precipitation indices

2017 ◽  
Vol 21 (3) ◽  
pp. 1573-1591 ◽  
Author(s):  
Louise Crochemore ◽  
Maria-Helena Ramos ◽  
Florian Pappenberger ◽  
Charles Perrin
Keyword(s):  
Long Range ◽  
Precipitation Forecast ◽  
Low Flow ◽  
Drought Risk ◽  
Streamflow Forecasting ◽  
Streamflow Forecasts ◽  
Precipitation Indices ◽  
Seasonal Streamflow ◽  
Prediction Approach ◽  
The Impact

Abstract. Many fields, such as drought-risk assessment or reservoir management, can benefit from long-range streamflow forecasts. Climatology has long been used in long-range streamflow forecasting. Conditioning methods have been proposed to select or weight relevant historical time series from climatology. They are often based on general circulation model (GCM) outputs that are specific to the forecast date due to the initialisation of GCMs on current conditions. This study investigates the impact of conditioning methods on the performance of seasonal streamflow forecasts. Four conditioning statistics based on seasonal forecasts of cumulative precipitation and the standardised precipitation index were used to select relevant traces within historical streamflows and precipitation respectively. This resulted in eight conditioned streamflow forecast scenarios. These scenarios were compared to the climatology of historical streamflows, the ensemble streamflow prediction approach and the streamflow forecasts obtained from ECMWF System 4 precipitation forecasts. The impact of conditioning was assessed in terms of forecast sharpness (spread), reliability, overall performance and low-flow event detection. Results showed that conditioning past observations on seasonal precipitation indices generally improves forecast sharpness, but may reduce reliability, with respect to climatology. Conversely, conditioned ensembles were more reliable but less sharp than streamflow forecasts derived from System 4 precipitation. Forecast attributes from conditioned and unconditioned ensembles are illustrated for a case of drought-risk forecasting: the 2003 drought in France. In the case of low-flow forecasting, conditioning results in ensembles that can better assess weekly deficit volumes and durations over a wider range of lead times.

scholarly journals Seasonal streamflow forecasts in the Ahlergaarde catchment, Denmark: effect of preprocessing and postprocessing on skill and statistical consistency

2017 ◽  
Author(s):  
Diana Lucatero ◽  
Henrik Madsen ◽  
Jens C. Refsgaard ◽  
Jacob Kidmose ◽  
Karsten H. Jensen
Keyword(s):  
Hydrological Model ◽  
Predictive Distribution ◽  
Low Flow ◽  
Model Errors ◽  
Streamflow Prediction ◽  
Initial State ◽  
Lower Accuracy ◽  
Spatially Distributed ◽  
Streamflow Forecasts ◽  
Statistical Consistency

Abstract. In the present study we analyze the effect of bias adjustments in both meteorological and streamflow forecasts on skill and reliability of monthly average streamflow and low flow forecasts. Both raw and pre-processed meteorological seasonal forecast from the European Center for Medium-Range Weather Forecasts (ECMWF) are used as inputs to a spatially distributed, coupled surface – subsurface hydrological model based on the MIKE SHE code in order to generate streamflow predictions up to seven months in advance. In addition to this, we postprocess streamflow predictions using an empirical quantile mapping that adjusts the predictive distribution in order to match the observed one. Bias, skill and statistical consistency are the qualities evaluated throughout the forecast generating strategies and we analyze where the different strategies fall short to improve them. ECMWF System 4-based streamflow forecasts tend to show a lower accuracy level than those generated with an ensemble of historical observations, a method commonly known as Ensemble Streamflow Prediction (ESP). This is particularly true at longer lead times, for the dry season and for streamflow stations that exhibit low hydrological model errors. Biases in the mean are better removed by postprocessing that in turn is reflected in the higher level of statistical consistency. However, in general, the reduction of these biases is not enough to ensure a higher level of accuracy than the ESP forecasts. This is true for both monthly mean and minimum yearly streamflow forecasts. We highlight the importance of including a better estimation of the initial state of the catchment, which will increase the capability of the system to forecast streamflow at longer leads.

scholarly journals Bias correcting precipitation forecasts to improve the skill of seasonal streamflow forecasts

2016 ◽  
Author(s):  
Louise Crochemore ◽  
M.-H. Ramos ◽  
Florian Pappenberger
Keyword(s):  
Bias Correction ◽  
Linear Scaling ◽  
Drought Risk ◽  
Seasonal Precipitation ◽  
Streamflow Forecasting ◽  
Seasonal Forecasts ◽  
Ensemble Forecasts ◽  
Streamflow Forecasts ◽  
Distribution Mapping ◽  
Seasonal Streamflow

Abstract. Meteorological centres make sustained efforts to provide seasonal forecasts that are increasingly skilful, which has the potential to benefit streamflow forecasting. Seasonal streamflow forecasts can help to take anticipatory measures for a range of applications, such as water supply or hydropower reservoir operation and drought risk management. This study assesses the skill of seasonal precipitation and streamflow forecasts in France to provide insights into the way bias correcting precipitation forecasts can improve the skill of streamflow forecasts at extended lead times. We apply eight variants of bias correction approaches to the precipitation forecasts prior to generating the streamflow forecasts. The approaches are based on the linear scaling and the distribution mapping methods. A daily hydrological model is applied at the catchment scale to transform precipitation into streamflow. We then evaluate the skill of raw (without bias correction) and bias corrected precipitation and streamflow ensemble forecasts in sixteen catchments in France. The skill of the ensemble forecasts is assessed in reliability, sharpness, accuracy, and overall performance. A reference prediction system, based on historical observed precipitation and catchment initial conditions at the time of forecast (i.e., ESP method), is used as benchmark in the computation of the skill. The results show that, in most catchments, raw seasonal precipitation and streamflow forecasts are often more skilful than the conventional ESP method in terms of sharpness. However, they are not significantly better in terms of reliability. Forecast skill is generally improved when applying bias correction. Two bias correction methods show the best performance for the studied catchments, each method being more successful in improving specific attributes of the forecasts: the simple linear scaling of monthly values contribute mainly to increasing forecast sharpness and accuracy, while the empirical distribution mapping of daily values is successful in improving forecast reliability.

Why is large sample hydrology important in hydrological forecasting? 

2021 ◽  
Author(s):  
Ilias Pechlivanidis ◽  
Louise Crochemore ◽  
Marc Girons Lopez
Keyword(s):  
River Flow ◽  
Forecast Skill ◽  
Machine Learning Techniques ◽  
Precipitation Forecast ◽  
River Systems ◽  
Local Conditions ◽  
Large Sample ◽  
Streamflow Forecasts ◽  
Seasonal Streamflow

<p>The scientific community has made significant progress towards improving the skill of hydrological forecasts; however, most investigations have normally been conducted at single or in a limited number of catchments. Such an approach is indeed valuable for detailed process investigation and therefore to understand the local conditions that affect forecast skill, but it is limited when it comes to scaling up the underlying hydrometeorological hypotheses. To advance knowledge on the drivers that control the quality and skill of hydrological forecasts, much can be gained by comparative analyses and from the availability of statistically significant samples. Large-scale modelling (at national, continental or global scales) can complement the in-depth knowledge from single catchment modelling by encompassing many river systems that represent a breadth of physiographic and climatic conditions. In addition to large sample sizes which cover a gradient in terms of climatology, scale and hydrological regime, the use of machine learning techniques can contribute to the identification of emerging spatiotemporal patterns leading to forecast skill attribution to different regional physiographic characteristics.</p><p>Here, we draw on two seasonal hydrological forecast skill investigations that were conducted at the national and continental scales, providing results for more than 36,000 basins in Sweden and Europe. Due to the large generated samples, we are capable of demonstrating that the quality of seasonal streamflow forecasts can be clustered and regionalized, based on a priori knowledge of the local hydroclimatic conditions. We show that the quality of seasonal streamflow forecasts is linked to physiographic and hydroclimatic descriptors, and that the relative importance of these descriptors varies with initialization month and lead time. In our samples, hydrological similarity, temperature, precipitation, evaporative index, and precipitation forecast biases are strongly linked to the quality of streamflow forecasts. This way, while seasonal river flow can generally be well predicted in river systems with slow hydrological responses, predictability tends to be poor in cold and semiarid climates in which river systems respond immediately to precipitation signals.</p>

Sign in / Sign up

Export Citation Format

Share Document