Probabilistic radar rainfall nowcasts using empirical and theoretical uncertainty models

Weather radar has a potential to provide accurate short‐term (0–3 h) forecasts of rainfall (i.e. radar nowcasts), which are of great importance in warnings and risk management for hydro‐meteorological events. However, radar nowcasts are affected by large uncertainties, which are not only linked to limitations in the forecast method but also because of errors in the radar rainfall measurement. The probabilistic quantitative precipitation nowcasting approach attempts to quantify these uncertainties by delivering the forecasts in a probabilistic form. This study implements two forms of probabilistic quantitative precipitation nowcasting for a hilly area in the south of Manchester, namely, the theoretically based scheme [ensemble rainfall forecasts (ERF)‐TN] and the empirically based scheme (ERF‐EM), and explores which one exhibits higher predictive skill. The ERF‐TN scheme generates ensemble forecasts of rainfall in which each ensemble member is determined by the stochastic realisation of a theoretical noise component. The so‐called ERF‐EM scheme proposed and applied for the first time in this study, aims to use an empirically based error model to measure and quantify the combined effect of all the error sources in the radar rainfall forecasts. The essence of the error model is formulated into an empirical relation between the radar rainfall forecasts and the corresponding ‘ground truth’ represented by the rainfall field from rain gauges measurements. The ensemble members generated by the two schemes have been compared with the rain gauge rainfall. The hit rate and the false alarm rate statistics have been computed and combined into relative operating characteristic curves. The comparison of the performance scores for the two schemes shows that the ERF‐EM achieves better performance than the ERF‐TN at 1‐h lead time. The predictive skills of both schemes are almost identical when the lead time increases to 2 h. In addition, the relation between uncertainty in the radar rainfall forecasts and lead time is also investigated by computing the dispersion of the generated ensemble members. Copyright © 2013 John Wiley & Sons, Ltd.     

Hydrological uncertainty processor based on a copula function

Quantifying the uncertainty in hydrological forecasting is valuable for water resources management and decision-making processes. The hydrological uncertainty processor (HUP) can quantify hydrological uncertainty and produce probabilistic forecasts under the hypothesis that there is no input uncertainty. This study proposes a HUP based on a copula function, in which the prior density and likelihood function are explicitly expressed, and the posterior density and distribution obtained using Monte Carlo sampling. The copula-based HUP was applied to the Three Gorges Reservoir, and compared with the meta-Gaussian HUP. The Nash-Sutcliffe efficiency and relative error were used as evaluation criteria for deterministic forecasts, while predictive QQ plot, reliability, resolution and continuous rank probability score (CRPS) were used for probabilistic forecasts. The results show that the proposed copula-based HUP is comparable to the meta-Gaussian HUP in terms of the posterior median forecasts, and that its probabilistic forecasts have slightly higher reliability and lower resolution compared to the meta-Gaussian HUP. Based on the CRPS, both HUPs were found superior to deterministic forecasts, highlighting the effectiveness of probabilistic forecasts, with the copula-based HUP marginally better than the meta-Gaussian HUP.     

A variational method for correcting non-systematic errors in numerical weather prediction

A variational method based on previous numerical forecasts is developed to estimate and correct non-systematic component of numerical weather forecast error. In the method, it is assumed that the error is linearly dependent on some combination of the forecast fields, and three types of forecast combination are applied to identifying the forecasting error: 1) the forecasts at the ending time, 2) the combination of initial fields and the forecasts at the ending time, and 3) the combination of the forecasts at...     

Evaluation of Short-Range Forecasts from a Mesoscale Model Over the Indian Region During Monsoon 2006

This study examines the short-range forecast accuracy of the Pennsylvania State University-National Center for Atmospheric Research Mesoscale Model (MM5) as applied to the July 2006 episode of the Indian summer monsoon (ISM) and the model's sensitivity to the choice of different cumulus parameterization schemes (CPSs), namely Betts-Miller, Grell (GR) and Kain-Fritsch (KF). The results showed that MM5 day 1 (0–24 h prediction) and day 2 (24–48 h prediction) forecasts using all three CPSs overpredicted monsoon rainfall over the Indian landmass, with the larger overprediction seen in the day 2 forecasts. Among the CPSs, the rainfall distribution over the Indian landmass was better simulated in forecasts using the KF scheme. The KF scheme showed better skill in predicting the area of rainfall for most of the rainfall thresholds. The root mean square error (RMSE) in day 1 and day 2 rainfall forecasts using different CPSs showed that rainfall simulated using the KF scheme agreed better with the observed rainfall. As compared to other CPSs, simulation using the GR scheme showed larger RMSE in wind speed prediction at 850 and 200 hPa over the Indian landmass. MM5 24-h temperature forecasts at 850 hPa with all the CPSs showed a warm bias of the order of 1 K over the Indian landmass and the bias doubled in 48-h model forecasts. The mean error in temperature prediction at 850 hPa over the Indian region using the KF scheme was comparatively smaller for all the forecast intervals. The model with all the CPSs overpredicted humidity at 850 hPa. The improved prediction by MM5 with the KF scheme is well complemented by the smaller error shown by the KF scheme in vertical distribution of heat and mean moist static energy in the lower troposphere. In this study, the KF scheme which explicitly resolve the downdrafts in the cloud column tended to produce more realistic precipitation forecasts as compared to other schemes which did not explicitly incorporate downdraft effects. This is an important result especially given that the area covered by monsoon-precipitating systems is largely from stratiform-type clouds which are associated with strong downdrafts in the lower levels. This result is useful for improving the treatment of cumulus convection in numerical models over the ISM region.     

Performance of operational systems with respect to water level forecasts in the Gulf of Finland

This paper is devoted to the validation of water level forecasts in the Gulf of Finland. Daily forecasts produced by four setups of operational, three-dimensional Baltic Sea oceanographic models are analyzed using statistical means and are compared with water level observations at three Finnish stations located on the northern coast of the Gulf of Finland. The overall conclusion is that the operational systems were skillful in forecasting water level variations during the study period from November 1, 2003, to January 31, 2005. The factors causing differences between the water level forecasts of different models are discussed as well. An important task of operational sea level forecasting services is to provide accurate and early information about extreme water levels, both positive and negative surges. During the study period, two major winter storms occurred which caused coastal flooding in the region. According to our analysis, the operational models forecast the rise of water levels during these events rather successfully. Nowadays, operational forecasts can provide early warnings of extreme water levels at least 1 day in advance, which may be regarded as a minimum requirement for an operational forecasting system. The paper concludes that the models generally performed very well, with over 93% of the hourly water level forecasts found to be within the range of ±15 cm of the observed water levels, and with the timing of the water level peaks accurately predicted. Further discussion and studies dealing with the assessment of the skills of both operational meteorological and oceanographic forecasts, especially in connection with rare surge events, will be necessary. Skill assessment of operational oceanographic models would be relatively easy if acceptable error limits or a quality system was developed for the Baltic Sea operational models.     

Correction of inertial oscillations by assimilation of HF radar data in a model of the Ligurian Sea

This article aims at analyzing if high-frequency radar observations of surface currents allow to improve model forecasts in the Ligurian Sea, where inertial oscillations are a dominant feature. An ensemble of ROMS models covering the Ligurian Sea, and nested in the Mediterranean Forecasting System, is coupled with two WERA high-frequency radars. A sensitivity study allows to determine optimal parameters for the ensemble filter. By assimilating observations in a single point, the obtained correction shows that the forecast error covariance matrix represents the inertial oscillations, as well as large- and meso-scale processes. Furthermore, it is shown that the velocity observations can correct the phase and amplitude of the inertial oscillations. Observations are shown to have a strong effect during approximately half a day, which confirms the importance of using a high temporal observation frequency. In general, data assimilation of HF radar observations leads to a skill score of about 30% for the forecasts of surface velocity.     

Hydrodynamic-phytoplankton model for short-term forecasts of phytoplankton in Lake Taihu, China

Phytoplankton biomass is an important factor for short-term forecasts of algal blooms. Our new hydrodynamic-phytoplankton model is primarily intended for simulating the spatial and temporal distribution of phytoplankton in Lake Taihu within a time frame of 1-5 days. The model combines two modules: a simple phytoplankton kinetics module for growth and loss; and a mass-transport module, which defines phytoplankton transport horizontally with a two dimensional hydrodynamic model. To adapt field data for model input and calibration, we introduce two simplifications: (a) exclusion of some processes related to phytoplankton dynamics like nutrient dynamics, sediment resuspension, mineralization and nitrification, and (b) use of monthly measured data of the nutrient state. Chlorophyll-α concentration, representing phytoplankton biomass, is the only state variable in the model. A sensitivity analysis was carried out to identify the most sensitive parameter set in the phytoplankton kinetics module. The model was calibrated with field data collected in 2008 and validated with additional data obtained in 2009. A comparison of simulated and observed chlorophyll-α concentration for 33 grid cells achieved an accuracy of 78.7%. However, mean percent error and mean absolute percent error were 13.4% and 58.2%, respectively, which implies that further improvement is necessary, e.g. by reducing uncertainty of the model input and by an improved parameter calibration.     

The predictability of meteo-oceanographic events

We have explored the predictability of storms in a small enclosed basin with a complicated surrounding orography. We have considered two exceptional storms in the far past and three mild events happened in recent years. A posteriori forecasts have been done up to 6 days before the events. The results have been compared versus measured data and the related analysis. Good predictability (10–15% error in surface wind speed and wave height) have been found up to day 4, mildly larger (<30%) up to day 6 before the event. In no case was a storm missed. This suggests that the effective predictability in more open basins may extend to even larger ranges.     

Probabilistic forecasting of drought: a hidden Markov model aggregated with the RCP 8.5 precipitation projection

The creeping characteristics of drought make it possible to mitigate drought’s effects with accurate forecasting models. Drought forecasts are inevitably plagued by uncertainties, making it necessary to derive forecasts in a probabilistic framework. In this study, we proposed a new probabilistic scheme to forecast droughts that used a discrete-time finite state-space hidden Markov model (HMM) aggregated with the Representative Concentration Pathway 8.5 (RCP) precipitation projection (HMM-RCP). The standardized precipitation index (SPI) with a 3-month time scale was employed to represent the drought status over the selected stations in South Korea. The new scheme used a reversible jump Markov chain Monte Carlo algorithm for inference on the model parameters and performed an RCP precipitation projection transformed SPI (RCP-SPI) weight-corrected post-processing for the HMM-based drought forecasting to perform a probabilistic forecast of SPI at the 3-month time scale that considered uncertainties. The point forecasts which were derived as the HMM-RCP forecast mean values, as measured by forecasting skill scores, were much more accurate than those from conventional models and a climatology reference model at various lead times. We also used probabilistic forecast verification and found that the HMM-RCP provided a probabilistic forecast with satisfactory evaluation for different drought categories, even at long lead times. In a drought event analysis, the HMM-RCP accurately predicted about 71.19 % of drought events during the validation period and forecasted the mean duration with an error of less than 1.8 months and a mean severity error of <0.57. The results showed that the HMM-RCP had good potential in probabilistic drought forecasting.     

Improving the forecasts of extreme streamflow by support vector regression with the data extracted by self‐organizing map

During typhoons or storms, accurate forecasts of hourly streamflow are necessary for flood warning and mitigation. However, hourly streamflow is difficult to forecast because of the complex physical process and the high variability in time. Furthermore, under the global warming scenario, events with extreme streamflow may occur that leads to more difficulties in forecasting streamflows. Hence, to obtain more accurate hourly streamflow forecasts, an improved streamflow forecasting model is proposed in this paper. The computational kernel of the proposed model is developed on the basis of support vector machine (SVM). Additionally, self‐organizing map (SOM) is used to analyse observed data to extract data with specific properties, which are capable of providing valuable information for streamflow forecasting. After reprocessing, these extracted data and the observed data are used to construct the SVM‐based model. An application is conducted to clearly demonstrate the advantage of the proposed model. The comparison between the proposed model and the conventional SVM model, which is constructed without SOM, is performed. The results indicate that the proposed model is better performed than the conventional SVM model. Moreover, as regards the extreme events, the result shows that the proposed model reduces the forecasting error, especially the error of peak streamflow. It is confirmed that because of the use of data extracted by SOM, the improved forecasting performance is obtained. The proposed model, which can produce accurate forecasts, is expected to be useful to support flood warning systems. Copyright © 2012 John Wiley & Sons, Ltd.     

Incorporating Natural Changes in Global Climate in Very-Long-Range Forecasting of River Runoff

A method is proposed for incorporating possible global changes in the state of the atmosphere, basing on K. Hasselmann’s theory of stochastic climate models, for assessing the significance of forecasts of variations of annual river runoff depth in the XXI century. The data used includes the results of river runoff simulation at warming, obtained using 21 IPCC climate models along with six IPCC scenarios of greenhouse gas emission, and MEI scenario. The significance index of forecasted runoff variations, i.e., the values of runoff depth increments divided by the standard error of forecasts was mapped. To demonstrate the role of the maps of significance index, which have been constructed taking into account forecast uncertainty because of the natural changes in global climate, those maps were compared with the maps of significance index calculated basing on other sources of errors. At large time scales, the uncertainty of runoff forecasts owing to natural changes in global climate plays the main role in assessing the reliability of forecasts in areas where greenhouse effect is strongest. Estimates of the significance index show that statistically significant changes in the annual runoff depth in the extreme northeast of Eurasia can be expected to occur not earlier than the late XXI century. In other RF regions, as well as in the majority of world areas, the forecasted changes in the annual runoff depth are comparable with the standard errors of the respective estimates or are less than they are.     

Development,implementation, and skill assessment of the NOAA/NOS Great Lakes Operational Forecast System

The NOAA Great Lakes Operational Forecast System (GLOFS) uses near-real-time atmospheric observations and numerical weather prediction forecast guidance to produce three-dimensional forecasts of water temperature and currents, and two-dimensional forecasts of water levels of the Great Lakes. This system, originally called the Great Lakes forecasting system (GLFS), was developed at The Ohio State University and NOAA’s Great Lakes Environmental Research Laboratory (GLERL) in 1989. In 1996, a workstation version of the GLFS was ported to GLERL to generate semi-operational nowcasts and forecasts daily. In 2004, GLFS went through rigorous skill assessment and was transitioned to the National Ocean Service (NOS) Center for Operational Oceanographic Products and Services (CO-OPS) in Silver Spring, MD. GLOFS has been making operational nowcasts and forecasts at CO-OPS since September 30, 2005. Hindcast, nowcast, and forecast evaluations using the NOS-developed skill assessment software tool indicated both surface water levels and temperature predictions passed the NOS specified criteria at a majority of the validation locations with relatively low root mean square error (4–8 cm for water levels and 0.5 to 1°C for surface water temperatures). The difficulty of accurately simulating seiches generated by storms (in particular in shallow lakes like Lake Erie) remains a major source of error in water level prediction and should be addressed in future improvements of the forecast system.     

Updating Real-Time Flood Forecasting Using a Fuzzy Rule-Based Model/Mise à Jour de Prévision de Crue en Temps Réel Grâce à un Modèle à Base de Règles Floues

Abstract

An updating technique is a tool to update the forecasts of mathematical flood forecasting model based on data observed in real time, and is an important element in a flood forecasting model. An error prediction model based on a fuzzy rule-based method was proposed as the updating technique in this work to improve one- to four-hour-ahead flood forecasts by a model that is composed of the grey rainfall model, the grey rainfall—runoff model and the modified Muskingum flow routing model. The coefficient of efficiency with respect to a benchmark is applied to test the applicability of the proposed fuzzy rule-based method. The analysis reveals that the fuzzy rule-based method can improve flood forecasts one to four hours ahead. The proposed updating technique can mitigate the problem of the phase lag in forecast hydrographs, and especially in forecast hydrographs with longer lead times.     

ARIMA forecasting of ambient air pollutants (O3, NO, NO2 and CO)

In the present study, a stationary stochastic ARMA/ARIMA [Autoregressive Moving (Integrated) Average] modelling approach has been adapted to forecast daily mean ambient air pollutants (O₃, CO, NO and NO₂) concentration at an urban traffic site (ITO) of Delhi, India. Suitable variance stabilizing transformation has been applied to each time series in order to make them covariance stationary in a consistent way. A combination of different information-criterions, namely, AIC (Akaike Information Criterion), HIC (Hannon–Quinn Information Criterion), BIC (Bayesian Information criterion), and FPE (Final Prediction Error) in addition to ACF (autocorrelation function) and PACF (partial autocorrelation function) inspection, has been tried out to obtain suitable orders of autoregressive (p) and moving average (q) parameters for the ARMA(p,q)/ARIMA(p,d,q) models. Forecasting performance of the selected ARMA(p,q)/ARIMA(p,d,q) models has been evaluated on the basis of MAPE (mean absolute percentage error), MAE (mean absolute error) and RMSE (root mean square error) indicators. For 20 out of sample forecasts, one step (i.e., one day) ahead MAPE for CO, NO₂, NO and O₃, have been found to be 13.6, 12.1, 21.8 and 24.1%, respectively. Given the stochastic nature of air pollutants data and in the light of earlier reported studies regarding air pollutants forecasts, the forecasting performance of the present approach is satisfactory and the suggested forecasting procedure can be effectively utilized for short term air quality forewarning purposes.     

An evaluation and improvement of tropical cyclone prediction in the western North Pacific basin from global ensemble forecasts

Forecasts of tropical cyclones(TCs) of the western North Pacific basin during the period of July to August 2018,especially of Rumbia(2018), Ampil(2018) and Jongdari(2018) that made landfall over Shanghai, have opposed great challenges for numerical models and forecasters. The predictive skill of these TCs are analyzed based on ensemble forecasts of ECMWF and NCEP. Results of the overall performance show that ensemble forecasts of ECMWF generally have higher predictive skill of track and intensity forecasts than those of NCEP. Specifically, ensemble forecasts of ECMWF have higher predictive skill of intensity forecasts for Rumbia(2018) and Ampil(2018) than those of NCEP, and both have low predictive skill of intensity forecasts for Jongdari(2018) at peak intensity. To improve the predictive skill of ensemble forecasts for TCs, a method that estimates adaptive weights for members of an ensemble forecast is proposed. The adaptive weights are estimated based on the fit of ensemble priors and posteriors to observations. The performances of ensemble forecasts of ECMWF and NCEP using the adaptive weights are generally improved for track and intensity forecasts. The advantages of the adaptive weights are more prominent for ensemble forecasts of ECMWF than for those of NCEP.     

Artificial neural networks as routine for error correction with an application in Singapore regional model

This research presents an error correction scheme based on artificial neural networks, and demonstrates its application on water level forecast for the Singapore water. The error correction scheme combines the numerical model outputs with the in situ measurements on a two-step basis: (1) predicting the model errors at the measurement stations and (2) distributing the predicted errors to the nonmeasurement stations. Artificial neural networks are used in both error prediction and error distribution as the mapping function approximators. The efficiency of this scheme is tested on six water level stations in the Singapore regional model domain with four prediction horizons. The results show that this error correction scheme produces high-precision forecasts, and improves the forecast accuracy at both measurement and nonmeasurement stations.     

Nowcasting with INCA During SNOW-V10

A Central-European nowcasting system which has been developed for use in mountainous terrain is tested in the Whistler/Vancouver area as part of the SNOW-V10 experiment. The integrated nowcasting through comprehensive analysis system provides hourly updated gridded forecasts of temperature, humidity, and wind, as well as precipitation forecasts which are updated every 15 min. It is based on numerical weather prediction (NWP) output and real-time surface weather station and radar data. Verification of temperature, relative humidity, and wind against surface stations shows that forecast errors are significantly reduced in the nowcasting range compared to those of the driving NWP model. The main contribution to the improvement comes from the implicit bias correction due to use of the latest observations. Relative humidity shows the longest lasting effect, with >50 % reduction of mean absolute error up to +4 h. For temperature and wind speed this percentage is reached after +2 and +3 h, respectively. Two cases of precipitation nowcasting are discussed and verified qualitatively.     

Clustering and forecasting of dissolved oxygen concentration on a river basin

The aim of this contribution is to combine statistical methodologies to geographically classify homogeneous groups of water quality monitoring sites based on similarities in the temporal dynamics of the dissolved oxygen (DO) concentration, in order to obtain accurate forecasts of this quality variable. Our methodology intends to classify the water quality monitoring sites into spatial homogeneous groups, based on the DO concentration, which has been selected and considered relevant to characterize the water quality. We apply clustering techniques based on Kullback Information, measures that are obtained in the state space modelling process. For each homogeneous group of water quality monitoring sites we model the DO concentration using linear and state space models, which incorporate tendency and seasonality components in different ways. Both approaches are compared by the mean squared error (MSE) of forecasts.     

Forecasting river flow using nonlinear dynamics

A Nearest Neighbor Method (NNM) is used to forecast daily river flows that were measured at a single location over a time period spanning about seventy years. A parsimonious three parameter NNM is developed in the context of Nonlinear Dynamics and the dependence between forecast error and length of history used to construct forecasts is investigated. Comparison is made to Auto-Regressive Integrated Moving Average (ARIMA) models. The NNM is found to provide improved forecasts.     

Space–time forecasting using soft geostatistics: a case study in forecasting municipal water demand for Phoenix, Arizona

Managing environmental and social systems in the face of uncertainty requires the best possible forecasts of future conditions. We use space–time variability in historical data and projections of future population density to improve forecasting of residential water demand in the City of Phoenix, Arizona. Our future water estimates are derived using the first and second order statistical moments between a dependent variable, water use, and an independent variable, population density. The independent variable is projected at future points, and remains uncertain. We use adjusted statistical moments that cover projection errors in the independent variable, and propose a methodology to generate information-rich future estimates. These updated estimates are processed in Bayesian Maximum Entropy (BME), which produces maps of estimated water use to the year 2030. Integrating the uncertain estimates into the space–time forecasting process improves forecasting accuracy up to 43.9% over other space–time mapping methods that do not assimilate the uncertain estimates. Further validation studies reveal that BME is more accurate than co-kriging that integrates the error-free independent variable, but shows similar accuracy to kriging with measurement error that processes the uncertain estimates. Our proposed forecasting method benefits from the uncertain estimates of the future, provides up-to-date forecasts of water use, and can be adapted to other socio-economic and environmental applications.