Short-range Ensemble Prediction Research Articles

A Comparison of Ensemble Strategies for Flash Flood Forecasting: The 12 October 2007 Case Study in Valencia, Spain

Abstract On 12 October 2007, several flash floods affected the Valencia region, eastern Spain, with devastating impacts in terms of human, social, and economic losses. An enhanced modeling and forecasting of these extremes, which can provide a tangible basis for flood early warning procedures and mitigation measures over the Mediterranean, is one of the fundamental motivations of the international Hydrological Cycle in the Mediterranean Experiment (HyMeX) program. The predictability bounds set by multiple sources of hydrological and meteorological uncertainty require their explicit representation in hydrometeorological forecasting systems. By including local convective precipitation systems, short-range ensemble prediction systems (SREPSs) provide a state-of-the-art framework to generate quantitative discharge forecasts and to cope with different sources of external-scale (i.e., external to the hydrological system) uncertainties. The performance of three distinct hydrological ensemble prediction systems (HEPSs) for the small-sized Serpis River basin is examined as a support tool for early warning and mitigation strategies. To this end, the Flash-Flood Event–Based Spatially Distributed Rainfall–Runoff Transformation–Water Balance (FEST-WB) model is driven by ground stations to examine the hydrological response of this semiarid and karstic catchment to heavy rains. The use of a multisite and novel calibration approach for the FEST-WB parameters is necessary to cope with the high nonlinearities emerging from the rainfall–runoff transformation and heterogeneities in the basin response. After calibration, FEST-WB reproduces with remarkable accuracy the hydrological response to intense precipitation and, in particular, the 12 October 2007 flash flood. Next, the flood predictability challenge is focused on quantitative precipitation forecasts (QPFs). In this regard, three SREPS generation strategies using the WRF Model are analyzed. On the one side, two SREPSs accounting for 1) uncertainties in the initial conditions (ICs) and lateral boundary conditions (LBCs) and 2) physical parameterizations are evaluated. An ensemble Kalman filter (EnKF) is also designed to test the ability of ensemble data assimilation methods to represent key mesoscale uncertainties from both IC and subscale processes. Results indicate that accounting for diversity in the physical parameterization schemes provides the best probabilistic high-resolution QPFs for this particular flash flood event. For low to moderate precipitation rates, EnKF and pure multiple physics approaches render undistinguishable accuracy for the test situation at larger scales. However, only the multiple physics QPFs properly drive the HEPS to render the most accurate flood warning signals. That is, extreme precipitation values produced by these convective-scale precipitation systems anchored by complex orography are better forecast when accounting just for uncertainties in the physical parameterizations. These findings contribute to the identification of ensemble strategies better targeted to the most relevant sources of uncertainty before flash flood situations over small catchments.

A retrospective streamflow ensemble forecast for an extreme hydrologic event: a case study of Hurricane Irene and on the Hudson River basin

Abstract. This paper investigates the uncertainties in hourly streamflow ensemble forecasts for an extreme hydrological event using a hydrological model forced with short-range ensemble weather prediction models. A state-of-the art, automated, short-term hydrologic prediction framework was implemented using GIS and a regional scale hydrological model (HEC-HMS). The hydrologic framework was applied to the Hudson River basin ( ∼ 36 000 km2) in the United States using gridded precipitation data from the National Centers for Environmental Prediction (NCEP) North American Regional Reanalysis (NARR) and was validated against streamflow observations from the United States Geologic Survey (USGS). Finally, 21 precipitation ensemble members of the latest Global Ensemble Forecast System (GEFS/R) were forced into HEC-HMS to generate a retrospective streamflow ensemble forecast for an extreme hydrological event, Hurricane Irene. The work shows that ensemble stream discharge forecasts provide improved predictions and useful information about associated uncertainties, thus improving the assessment of risks when compared with deterministic forecasts. The uncertainties in weather inputs may result in false warnings and missed river flooding events, reducing the potential to effectively mitigate flood damage. The findings demonstrate how errors in the ensemble median streamflow forecast and time of peak, as well as the ensemble spread (uncertainty) are reduced 48 h pre-event by utilizing the ensemble framework. The methodology and implications of this work benefit efforts of short-term streamflow forecasts at regional scales, notably regarding the peak timing of an extreme hydrologic event when combined with a flood threshold exceedance diagram. Although the modeling framework was implemented on the Hudson River basin, it is flexible and applicable in other parts of the world where atmospheric reanalysis products and streamflow data are available.

Potential of a probabilistic hydrometeorological forecasting approach for the 28 September 2012 extreme flash flood in Murcia, Spain

An improved understanding, modeling and forecasting of hydrometeorological extremes over the flood-prone Western Mediterranean region is one of the milestones of the international HyMeX program. A set of severe hydrometeorological episodes affected various basins across south and eastern Mediterranean Spain from 27 to 29 September 2012. Flooding was particularly catastrophic in Andalusia and Murcia, where 10 fatalities occurred and material losses were estimated at 120M€. The predictability bounds set by the type and scales of the processes involved in such high-impact episodes require the explicit representation of uncertainty in the hydrometeorological forecasting chain. A short-range ensemble prediction system (EPS) provides the optimal framework to generate risk-based forecasts supporting valuable early warning procedures and mitigation measures. We explore the potential of this probabilistic forecasting approach on the 28 September 2012 flash flood in the Guadalentín river basin, a medium-sized catchment located in Murcia, southeastern Spain. After a rigorous calibration with rain-gauge data, the hydrological response of the basin to this flooding is accurately simulated by the Hydrologic Engineering Center's Hydrological Modeling System runoff model. Then, we explore the uncertainty transference from a collection of mesoscale meteorological deterministic and probabilistic 48h predictions provided by the Weather Research and Forecasting (WRF) model. The meteorological simulations are nested within the global EPS of the European Centre for Medium-Range Weather Forecasts, therefore inheriting the spread of the global system and providing probabilistic high-resolution precipitation structures to the hydrological model. By assuming the calibrated model as a good representation of a perfect hydrological model for this event, it becomes an advanced and user-oriented verification tool for quantitative precipitation forecasts. Results highlight the benefits of accounting for uncertainties in the precipitation forecasts and the value of the proposed set-up for the short-range prediction of quantitative discharge forecasts. The warn-on-forecast approach is shown to be possible within a probabilistic hydrometeorological forecasting chain for basins as small and fast-responsive as the Guadalentín basin, proving to be suitable for civil protection warning procedures.

Latest developments around the ALADIN operational short-range ensemble prediction system in Hungary

Latest developments around the ALADIN operational short-range ensemble prediction system in Hungary

Cloud-Resolving Ensemble Simulations of Mediterranean Heavy Precipitating Events: Uncertainty on Initial Conditions and Lateral Boundary Conditions

Abstract This study assesses the impact of uncertainty on convective-scale initial conditions (ICs) and the uncertainty on lateral boundary conditions (LBCs) in cloud-resolving simulations with the Application of Research to Operations at Mesoscale (AROME) model. Special attention is paid to Mediterranean heavy precipitating events (HPEs). The goal is achieved by comparing high-resolution ensembles generated by different methods. First, an ensemble data assimilation technique has been used for assimilation of perturbed observations to generate different convective-scale ICs. Second, three ensembles used LBCs prescribed by the members of a global short-range ensemble prediction system (EPS). All ensembles obtained were then evaluated over 31- and/or 18-day periods, and on 2 specific case studies of HPEs. The ensembles are underdispersive, but both the probabilistic evaluation of their overall performance and the two case studies confirm that they can provide useful probabilistic information for the HPEs considered. The uncertainty on convective-scale ICs is shown to have an impact at short range (under 12 h), and it is strongly dependent on the synoptic-scale context. Specifically, given a marked circulation near the area of interest, the imposed LBCs rapidly overwhelm the initial differences, greatly reducing the spread of the ensemble. The uncertainty on LBCs shows an impact at longer range, as the spread in the coupling global ensemble increases, but it also depends on the synoptic-scale conditions and their predictability.

Performance of multi-model AEMET-SREPS precipitation probabilistic forecasts over Mediterranean area

Abstract. Spanish Meteorological Agency (AEMET) runs a daily experimental multi-model Short-Range Ensemble Prediction System (AEMET-SREPS). The role of the system horizontal resolution (0.25 degrees) on the performance of 24-h precipitation probabilistic forecasts, and its relation with mesoscale events, are assessed comparing the performance over the Mediterranean area and over an European Atlantic area. Gridded high resolution rain observations and standard verification measures have been used at different precipitation thresholds, while studying the dependency on seasons for a one year period (May 2007 to June 2008). As a general result, performance over the Mediterranean area is higher than over the Atlantic one, albeit some relative loss of skill is found in autumn, when mesoscale convective organization is assumed to play a more important role. So it is suggested that AEMET-SREPS system precipitation predictability over the Mediterranean in autumn could be expected to improve if the horizontal and vertical resolution is increased in order to take into account the effect of meso-beta scale, especially important for convective organization.

Linear and non-linear response to parameter variations in a mesoscale model

Parameter uncertainty in atmospheric model forcing and closure schemes has motivated both parameter estimation with data assimilation and use of pre-specified distributions to simulate model uncertainty in short-range ensemble prediction. This work assesses the potential for parameter estimation and ensemble prediction by analysing 2 months of mesoscale ensemble predictions in which each member uses distinct, and fixed, settings for four model parameters. A space-filling parameter selection design leads to a unique parameter set for each ensemble member. An experiment to test linear scaling between parameter distribution width and ensemble spread shows the lack of a general linear response to parameters. Individual member near-surface spatial means, spatial variances and skill show that perturbed models are typically indistinguishable. Parameter—state rank correlation fields are not statistically significant, although the presence of other sources of noise may mask true correlations. Results suggest that ensemble prediction using perturbed parameters may be a simple complement to more complex model-error simulation methods, but that parameter estimation may prove difficult or costly for real mesoscale numerical weather prediction applications.

Predictability of short-range forecasting: a multimodel approach

Numerical weather prediction (NWP) models (including mesoscale) have limitations when it comes to dealing with severe weather events because extreme weather is highly unpredictable, even in the short range. A probabilistic forecast based on an ensemble of slightly differentmodel runs may help to address this issue. Among other ensemble techniques, Multimodel ensemble prediction systems (EPSs) are proving to be useful for adding probabilistic value to mesoscale deterministic models. A Multimodel Short Range Ensemble Prediction System (SREPS) focused on forecasting the weather up to 72 h has been developed at the SpanishMeteorological Service (AEMET). The system uses five different limited area models (LAMs), namely HIRLAM (HIRLAM Consortium), HRM (DWD), the UM (UKMO), MM5 (PSU/NCAR) and COSMO (COSMO Consortium). These models run with initial and boundary conditions provided by five different global deterministic models, namely IFS (ECMWF), UM (UKMO), GME (DWD), GFS (NCEP) and CMC (MSC). AEMET-SREPS (AE) validation on the large-scale flow, using ECMWF analysis, shows a consistent and slightly underdispersive system. For surface parameters, the system shows high skill forecasting binary events. 24-h precipitation probabilistic forecasts are verified using an up-scaling grid of observations from European high-resolution precipitation networks, and compared with ECMWF-EPS (EC).

The U.S. Air ForceWeather Agency’s mesoscale ensemble: scientific description and performance results

This work evaluates several techniques to account for mesoscale initial-condition (IC) and model uncertainty in a short-range ensemble prediction system based on the Weather Research and Forecast (WRF) model. A scientific description and verification of several candidate methods for implementation in the U.S. Air Force Weather Agency mesoscale ensemble is presented. Model perturbation methods tested include multiple parametrization suites, landsurface property perturbations, perturbations to parameters within physics schemes and stochastic ‘backscatter’ streamfunction perturbations. IC perturbations considered include perturbed observations in 10 independent WRF-3DVar cycles and the ensemble-transform Kalman filter (ETKF). A hybrid of ETKF (for IC perturbations) and WRF-3DVar (to update the ensemble mean) is also tested. Results show that all of the model and IC perturbation methods examined are more skilful than direct dynamical downscaling of the global ensemble. IC perturbations are most helpful during the first 12 h of the forecasts. Physical parametrization diversity appears critical for boundary-layer forecasts. In an effort to reduce system complexity by reducing the number of suites of physical parametrizations, a smaller set of parametrization suites was combined with perturbed parameters and stochastic backscatter, resulting in the most skilful and statistically consistent ensemble predictions.

Properties of singular vectors using convective available potential energy as final time norm

We study the feasibility of using the singular vector technique to create initial condition perturbations for shortrange ensemble prediction systems (SREPS) focussing on predictability of severe local storms and in particular deep convection. For this a newfinal time semi-norm based on the convective available potential energy (CAPE) is introduced. We compare singular vectors using the CAPE-norm with SVs using the more common total energy (TE) norm for a 2-week summer period in 2007, which includes a case of mesoscale extreme rainfall in the south west of Finland. The CAPE singular vectors perturb the CAPE field by increasing the specific humidity and temperature of the parcel and increase the lapse rate above the parcel in the lower troposphere consistent with physical considerations. The CAPE-SVs are situated in the lower troposphere. This in contrast to TE-SVs with short optimization times which predominantly remain in the high troposphere. By examining the time evolution of the CAPE singular values we observe that the convective event in the south west of Finland is clearly associated with high CAPE singular values.

Latest developments around the ALADIN operational short-range ensemble prediction system in Hungary

Ensemble prediction systems (EPSs) are an essential part of numerical weather prediction for the provision of probabilistic forecast guidance. The Hungarian Meteorological Service has implemented a limited area EPS (called ALADIN HUNEPS) based on the ALADIN mesoscale limited area model coupled to the French global ARPEGE EPS (PEARP). The dynamical downscaling method is assessed in terms of ensemble verification scores taking also into account the recent upgrade of the PEARP global system. The verification results point towards some weaknesses of the ALADIN HUNEPS, mainly with respect to the near-surface parameters. Therefore, some improvements are needed so as to provide better and more reliable ensemble predictions for the Carpathian Basin. The application of near-surface perturbations into the surface data assimilation scheme is implemented and tested. The first results show that the surface perturbation method slightly improves the ALADIN HUNEPS, however, further experiments should be made to find out the optimal settings of the method for definite robust improvements of ALADIN HUNEPS

Probabilistic prediction of raw and BMA calibrated AEMET-SREPS: the 24 of January 2009 extreme wind event in Catalunya

Abstract. At 00:00 UTC of 24 January 2009 (24Jan09) an explosive cyclogenesis placed at the Gulf of Vizcaya, reached its maximum intensity with observed surface pressures below 970 hPa on its center. During its path through the south of France there were strong westerly and north-westerly winds over Iberian Peninsula (above 150 km/h). These extreme winds leaved 8 casualties in Catalunya, the north-east region of Spain. The aim of this work is validating the skill of the Spanish Meteorological Agency (AEMET) Short Range Ensemble Prediction System (SREPS) in forecasting this event. Two probabilistic forecasts of wind are compared, a non-calibrated (or raw) and a calibrated one using the Bayesian Model Averaging (BMA). AEMET runs a daily experimental SREPS twice a day (00:00 and 12:00 UTC). This system consists on 25 members that are constructed by integrating five different Limited Area Models (LAMs) at 0.25 degrees of horizontal resolution. Each model uses five different initial and boundary conditions from five Global Models (GMs). Thus it is obtained a probabilistic forecast that takes into account initial, contour and model uncertainties. BMA is a statistical tool for combining predictive Probability Distribution Functions (PDFs) from different sources. BMA predictive PDF is a weighted average of PDFs centered on the individual bias-corrected forecasts. Each weight is a measure of the corresponding forecast skill. Here BMA is applied to calibrate probabilistic forecasts of wind speed. In this work two time forecast ranges (H+60 and H+36) of 10-m wind speed over Catalonia are verified subjectively at 12:00 UTC of 24Jan09 valid time. We focus on the location and intensity of 10-m wind speed maximum values. Observations at 29 automatic ground stations of AEMET are used for the verification. On one hand results indicate that raw AEMET-SREPS is able to forecast 60 h ahead mean winds higher than 36 and 54 km/h and that it correctly locates them in three different areas. On the other hand, predicted probability loses its skill after BMA calibration of the ensemble. This is due to the fact that BMA bias correction underestimates the intensity of wind.

Verification of a short-range ensemble precipitation prediction system over Iberia

Abstract. The purpose of this paper is the verification of a short-range ensemble prediction system (SREPS) built with five different model physical process parameterization schemes and two different initial conditions from global models, allowing to construct several versions of the non-hydrostatic mesoscale MM5 model for a 1-month period of October 2006. From the SREPS, flow-dependent probabilistic forecasts are provided by means of predictive probability distributions over the Iberian Peninsula down to 10-km grid spacing. In order to carry out the verification, 25 km grid of observational precipitation records over Spain from the Spanish Climatic Network has been used to evaluate the ensemble accuracy together with the mean model performance and forecast variability by means of comparisons between such records and the ensemble forecasts. This verification has been carried out upscaling the 10 km probabilistic forecast to the observational data grid. Temporal evolution of precipitation forecasts for spatial averaged ensemble members and the ensemble mean is shown, illustrating the consistency of the SREPS. Such evolutions summarize the SREPS information, showing each of the members as well as the ensemble mean evolutions. The Talagrand diagram derived from the SREPS results shows underdispersion which indicates some bias behaviour. The Relative Operating Characteristic (ROC) curve shows a very outstanding area, indicating potential usefulness of the forecasting system. The forecast probability and the mean observed frequency present good agreement with the SREPS results close to the no-skill line. Because the probability has a good reliability and a positive contribution to the brier skill score, a positive value of this skill is obtained. Moreover, the probabilistic meteogram of the spatial daily mean precipitation values shows the range of forecast values, providing discrete probability information in different quantile intervals. The epsgram shows different daily distributions, indicating the predictability of each day.

Impact of including moisture perturbations on short‐range ensemble forecasts

We are developing a Short‐Range Ensemble Prediction (SREP) system based on the Eta Model for use over South America. The Eta Model SREP system uses the CPTEC global model Ensemble Prediction System (EPS) forecasts as initial and lateral boundary conditions. The objectives of this work are to verify the impacts of including moisture perturbations in the global EPS on the SREP and to evaluate the forecast quality from the resulting SREP. We compare the SREP constructed with and without moisture perturbations. We chose four cases of South Atlantic Convergence Zone events that produced heavy rainfall for the tests and evaluation. The Eta Model was set with a horizontal resolution of 10 km and integrated for 6 days. The mean errors of the forecasts based on the two perturbation methodologies are similar, which indicates that including moisture did not increase the forecast error. Precipitation forecasts showed major improvement when moisture perturbation was included. The root mean square error (RMSE) of the SREP ensemble mean forecast from both initial condition perturbations is smaller than the RMSE of the control run. The constructed SREP system exhibits forecast RMSE growth rate larger than the ensemble forecast spread, on the other hand, this difference is reduced compared to the driver global model ensemble forecast system.

Early Warnings of Severe Weather from Ensemble Forecast Information

A system has been developed to give probabilistic warnings of severe-weather events for the United Kingdom (UK) on a regional and national basis, based on forecast output from the European Centre for Medium-Range Weather Forecasts (ECMWF) Ensemble Prediction System (EPS). The First-Guess Early Warnings (FGEW) project aims to give guidance to operational forecasters, to help them give earlier warning of severe weather in support of the UK National Severe Weather Warning Service (NSWWS). Calibration was applied to the EPS model output to optimize the probabilistic early warnings over an initial training period of one winter season, and the resulting warnings were then verified over a 16-month period spanning two winter seasons. The skill of warnings from several versions of FGEW is assessed using a range of probabilistic skill scores, and is also compared with that of warnings issued by forecasters. Results show that the system is capable of providing useful warnings 3-4 days ahead with some probabilistic skill. Most of the skill is attributable to warnings issued at low probabilities, but when higher probabilities do occur, this provides a valuable signal that has been used by forecasters on a number of occasions to issue warnings earlier than was done previously. Maximum skill of the FGEW warnings is found at a lead time of 4 days, with virtually no skill at shorter lead times of 1 or 2 days. This behavior is found to also occur in equivalent deterministic forecasts and so is not attributable to the ensemble perturbation strategy. Nevertheless it is suggested that while the EPS perturbations work well for the medium range, alternative perturbation strategies may be required for successful short-range ensemble prediction.