Kenneth R. Mylne

Decision‐making from probability forecasts based on forecast value

A method of estimating the economic value of weather forecasts for decision-making is described. This method has recently been used for user-oriented verification of probability forecasts, but is here applied to aid forecast users in optimising their decision-making from probability forecasts. Value may be calculated in the same way for either probability forecasts or deterministic forecasts, and thus provides the user with a direct comparison of the value of each in terms of money saved, which is more relevant than most standard verification scores. The method is illustrated using site-specific probability forecasts generated from the ECMWF ensemble prediction system and deterministic forecasts from the ECMWF high-resolution global model. It is found that for most forecast events and for most users the probability forecasts have greater value than the deterministic forecasts from a higher resolution model.

Joint Medium-Range Ensembles from The Met. Office and ECMWF Systems

Abstract One possible method of incorporating model sensitivities into ensemble forecasting systems is to combine ensembles run from two or more models. Furthermore, the use of more than one analysis, to which perturbations are added, may provide further unstable directions for error growth not present with a single analysis. Results are presented from recent investigations into the potential benefit of combining ensembles from the systems of the European Centre for Medium-Range Weather Forecasts and The Met. Office of the United Kingdom. The multimodel and multianalysis ensemble significantly outperforms either individual system in many performance aspects, including deterministic and probabilistic forecast skill, spread–skill correlations, and breadth of synoptic information. It is demonstrated that these improvements are achieved through the combination of independent, useful information contained in the individual systems, and not through simple cancellation of biases that could occur when ensembles f...

Concentration fluctuation measurements in a plume dispersing in a stable surface layer

A set of tracer experiments studying concentration fluctuations in a pollutant plume dispersing near the surface in a stably stratified nocturnal boundary layer is described, and the results are compared with those obtained in near-neutral stability conditions by Mylne and Mason (1991). The results highlight the importance of slow meandering of the plume which is characteristic of stable conditions. This meandering makes it impossible to conduct experiments under near-stationary conditions, resulting in considerable statistical variability in the results, but is important in reducing time-averaged concentrations. Spectral characteristics of the plume and general fluctuation statistics are qualitatively similar to those in near-neutral stability, but there are significant quantitative differences. Fluctuation time scales are shown to be substantially longer under stable conditions. This difference cannot be fully explained by the reduced windspeed alone, indicating that the length scale of plume elements is also longer. Some of the differences observed in stable conditions, particularly the longer time scales, are shown to substantially increase the potential hazard due to fluctuations in practical applications. A conceptual model of plume dispersion is described, which explains the observed plume structure under different conditions by relating it to the turbulent velocity spectra.

Multi-model multi-analysis ensembles in quasi-operational medium-range forecasting

Ensemble prediction systems (EPS) for medium-range forecasting attempt to account for uncertainty in numerical weather prediction (NWP) by sampling the distribution function of future atmospheric states. Forecast uncertainty derives from uncertainty in both the analysed initial conditions (analysis errors) and in the forecast evolution (model errors). Current operational systems are primarily based on sampling the analysis errors through initial-condition perturbations with, at best, only limited sampling of model errors. One approach to sampling model errors and also to widening the sampling of analysis errors, is to include more than one NWP model, and more than one operational analysis to which perturbations are added, in the ensemble system. Previous work has demonstrated from a small number of case-studies that this multi-model multi-analysis ensemble (MMAE) approach can perform significantly better than a single-model system such as the Ensemble Prediction System (EPS) run by the ECMWF (European Centre for Medium-Range Weather Forecasts). In this study a MMAE was created by combining the ECMWF ensemble with an ensemble using the Met Office model and analysis, and was run daily for a year to assess the benefits over a larger, quasi-operational sample of forecasts. The results are compared with the operational ECMWF EPS which includes the latest upgrades, including stochastic physics which makes some allowance for uncertainty due to model errors. Results show that both for probabilistic forecasts (assessed by Brier skill scores and relative operating characteristics) and for deterministic forecasts based on the ensemble mean (assessed by root-mean square errors) the MMAE has increased forecast skill relative to the EPS. These improvements are obtained with no overall increase in ensemble size. Ensemble spread is also greater in the MMAE, and the increased skill is believed to be due to the additional model producing solutions which are synoptically more different than those produced by a single model ensemble. Benefits of the MMAE vary both in time and with geographical region, depending on which individual ensemble system performs better in particular synoptic situations. It is found that the MMAE almost always performs as well as the best individual ensemble, and on occasions better than either of them.

Test of a Poor Man’s Ensemble Prediction System for Short-Range Probability Forecasting

Abstract Current operational ensemble prediction systems (EPSs) are designed specifically for medium-range forecasting, but there is also considerable interest in predictability in the short range, particularly for potential severe-weather developments. A possible option is to use a poor man’s ensemble prediction system (PEPS) comprising output from different numerical weather prediction (NWP) centers. By making use of a range of different models and independent analyses, a PEPS provides essentially a random sampling of both the initial condition and model evolution errors. In this paper the authors investigate the ability of a PEPS using up to 14 models from nine operational NWP centers. The ensemble forecasts are verified for a 101-day period and five variables: mean sea level pressure, 500-hPa geopotential height, temperature at 850 hPa, 2-m temperature, and 10-m wind speed. Results are compared with the operational ECMWF EPS, using the ECMWF analysis as the verifying “truth.” It is shown that, despite...

The vertical profile of concentration fluctuations in near-surface plumes

Field experiments on concentration fluctuations have frequently measured horizontal cross-sections of fluctuation statistics through plumes at fixed heights “near the surface”, but have not considered the effect of height above the ground in any detail. A set of tracer experiments designed to measure vertical profiles of concentration fluctuations in plumes, with a range of source heights, is described, and profiles of statistics are presented. Considerable variation of the statistics with both source and detector height is found. Near the surface, fluctuation intensity is a minimum and the time and length scales of the fluctuations are greatly increased. Profiles are consistent with the idea that concentration fluctuations near the surface are like those higher up at a greater distance from the source. Lowering the source height reduces the fluctuation intensity at all heights, and also alters the form of the concentration PDF. Results may be explained by the reduced length scale of sheargenerated turbulence near the surface causing enhanced small-scale mixing, which rapidly smooths out much of the fine structure with the plume.

Concentration fluctuation measurements in tracer plumes using high and low frequency response detectors

A set of tracer experiments designed to compare two concentration fluctuation detectors and measure fluctuation statistics at high frequencies is described. A detector which has been used in several previous fluctuation experiments (the TIP photoionisation detector manufactured by Photovac of Canada) is compared with another with a much higher frequency response (the flame ionisation detector — FID — made by Cambustion of the UK). Good agreement is found and results show that the signal optimization system used in previous work with the TIP provides an accurate enhancement of the instrument output, thus improving confidence in the results of previous papers. They also confirm that the TIP detector is able to resolve most of the concentration variance in most situations of interest, but not at very short range. Measurements of the high frequency end of the fluctuation spectrum using the FID show inertial-convective subrange behaviour at frequencies not resolved by the TIP, supporting earlier work. Fluctuation spectra measured very close to the source are also shown to have a characteristic +2/3 power law behaviour (when nSc (n) is plotted against n) at lower frequencies, in agreement with theoretical predictions.

The Benefits of Multianalysis and Poor Man's Ensembles

Abstract A new approach to probabilistic forecasting is proposed, based on the generation of an ensemble of equally likely analyses of the current state of the atmosphere. The rationale behind this approach is to mimic a poor man’s ensemble, which combines the deterministic forecasts from national meteorological services around the world. The multianalysis ensemble aims to generate a series of forecasts that are both as skillful as each other and the control forecast. This produces an ensemble mean forecast that is superior not only to the ensemble members, but to the control forecast in the short range even for slowly varying parameters, such as 500-hPa height. This is something that it is not possible with traditional ensemble methods, which perturb a central analysis. The results herein show that the multianalysis ensemble is more skillful than the Met Office’s high-resolution forecast by 4.5% over the first 3 days (on average as measured for RMSE). Similar results are found for different verification ...

Use of medium-range ensembles at the Met Office I: PREVIN –a system for the production of probabilistic forecast information from the ECMWF EPS

This is the first of a pair of papers covering the production and use of probability forecasts at ranges of 3 to 10 days at the Met Office using the ECMWF Ensemble Prediction System (EPS). The use of ensembles is intended to provide a set of forecasts which cover the range of possible uncertainty, recognising that it is impossible to obtain a single deterministic forecast which is always correct. We present a brief review of ensemble forecasting techniques and their use for the generation of probability forecasts. A wide range of probability forecasting products and tools are now available to forecasters at the Met Office, generated from the EPS. These will be described, and their use and interpretation discussed, both for site-specific forecast data and for fields of data covering a wider area. An important part of any forecasting system is verification: this is covered in some detail using several different methods, for the site-specific forecasts of surface weather parameters. Comparisons are made between the probabilistic forecasts and equivalent deterministic forecasts generated from the high-resolution ECMWF model, and it is evident that the former are more skilful according to most assessment methods for forecasts more than three days ahead. The companion paper shows how forecasters use the vast amounts of information available in forecast production (Young and Carroll 2002). Copyright

Experimental Measurements of Concentration Fluctuations

With few exceptions, air pollution models are designed to predict dosage or ensemble mean concentrations averaged over time scales of tens of minutes, hours, or more. This is adequate for the study of long-range transport problems and also for short-range dispersion of, for example, radioactive contaminants, for which the time-averaged dosage is the important factor in hazard assessment. The models are also widely used to predict mean concentrations in toxic accidents, for which their time scales are less satisfactory. The toxicity of many gases does not vary linearly with concentration C and exposure time t. For example, the toxicity of Chlorine (Cl 2) varies approximately as C 2.75 t (see Griffiths and Megson (1984)). In this case the use of a time averaged dosage could lead to a dangerous underestimation of the hazard over short and medium range. Time averaged concentrations are also unsuitable for the assessment of the inflammability or odours of a gas plume, for which time scales of a few seconds are applicable. While modellers recognise that fluctuations of concentration occur on short time scales, they have not attempted to include them, partly due to a lack of good validation data. Recently technology has advanced to the stage at which tracer experiments may be conducted using continuous chemical analysers capable of measuring concentration time series with a frequency response of up to about 10Hz. Jones (1983) achieves a resolution better than 100Hz using ionized air as a tracer, but is limited to short range experiments and has to account for ionic repulsion in analysis of results. The response required in practise varies according to the application, but 10Hz is adequate for most. purposes. In particular, it is fast relative to the time scale of human breathing.