Search Results
You are looking at 1 - 8 of 8 items for
- Author or Editor: Diana J. M. Greenslade x
- Refine by Access: All Content x
Abstract
The operational consensus forecast (OCF) scheme uses past performance to bias correct and combine numerical forecasts to produce an improved forecast at locations where recent observations are available. Here, OCF uses past observations and forecasts of significant wave height from five numerical wave models available in real time at the Australian Bureau of Meteorology. In addition to OCF, different adaptive weighting and forecast combination strategies are investigated. At deep-water sites (ocean depth > 25 m), all of the interpolated raw model forecasts outperformed 24-h persistence and, after bias correction, one model was clearly best. Significant improvements over raw model significant wave height forecasts were achieved by bias correction, linear-regression methods, and combination strategies. The best forecasts were obtained from a “composite of composites” in which models with highly correlated errors were combined before being included in the performance-weighted bias-corrected forecast. This technique slightly outperformed the linear-regression-corrected best model. At shallow-water sites (ocean depth < 25 m), all raw models perform poorly relative to the 24-h persistence. The composited, corrected forecasts significantly improved on raw model significant wave height forecasts but only slightly outperformed the 24-h persistence. The raw models generated unrealistically large biases that tended to be amplified with larger observed values of significant wave height.
Abstract
The operational consensus forecast (OCF) scheme uses past performance to bias correct and combine numerical forecasts to produce an improved forecast at locations where recent observations are available. Here, OCF uses past observations and forecasts of significant wave height from five numerical wave models available in real time at the Australian Bureau of Meteorology. In addition to OCF, different adaptive weighting and forecast combination strategies are investigated. At deep-water sites (ocean depth > 25 m), all of the interpolated raw model forecasts outperformed 24-h persistence and, after bias correction, one model was clearly best. Significant improvements over raw model significant wave height forecasts were achieved by bias correction, linear-regression methods, and combination strategies. The best forecasts were obtained from a “composite of composites” in which models with highly correlated errors were combined before being included in the performance-weighted bias-corrected forecast. This technique slightly outperformed the linear-regression-corrected best model. At shallow-water sites (ocean depth < 25 m), all raw models perform poorly relative to the 24-h persistence. The composited, corrected forecasts significantly improved on raw model significant wave height forecasts but only slightly outperformed the 24-h persistence. The raw models generated unrealistically large biases that tended to be amplified with larger observed values of significant wave height.
Abstract
One of the main limitations to current wave data assimilation systems is the lack of an accurate representation of the structure of the background errors. One method that may be used to determine background errors is the observational method of Hollingsworth and Lönnberg. The observational method considers correlations of the differences between observations and the background. For the case of significant wave height (SWH), potential observations come from satellite altimeters. In this work, the effect of the irregular sampling pattern of the satellite on estimates of background errors is examined. This is achieved by using anomalies from a 3-month mean as a proxy for model errors. A set of anomaly correlations is constructed from modeled wave fields. The isotropic length scales of the anomaly correlations are found to vary considerably over the globe. In addition, the anomaly correlations are found to be significantly anisotropic. The modeled wave fields are then sampled at simulated altimeter observation locations, and the anomaly correlations are recalculated from the simulated altimeter data. The results are compared to the original anomaly correlations. It is found that, in general, the simulated altimeter data can capture most of the geographic and seasonal variability in the isotropic anomaly correlation length scale. The best estimates of the isotropic length scales come from a method in which correlations are calculated between pairs of observations from prior and subsequent ground tracks, in addition to along-track pairs of observations. This method was found to underestimate the isotropic anomaly correlation length scale by approximately 10%. The simulated altimeter data were not so successful in producing realistic anisotropic correlation functions. This is because of the lack of information in the zonal direction in the simulated altimeter data. However, examination of correlations along ascending and descending ground tracks separately can provide some indication of the areas on the globe for which the anomaly correlations are more anisotropic than others.
Abstract
One of the main limitations to current wave data assimilation systems is the lack of an accurate representation of the structure of the background errors. One method that may be used to determine background errors is the observational method of Hollingsworth and Lönnberg. The observational method considers correlations of the differences between observations and the background. For the case of significant wave height (SWH), potential observations come from satellite altimeters. In this work, the effect of the irregular sampling pattern of the satellite on estimates of background errors is examined. This is achieved by using anomalies from a 3-month mean as a proxy for model errors. A set of anomaly correlations is constructed from modeled wave fields. The isotropic length scales of the anomaly correlations are found to vary considerably over the globe. In addition, the anomaly correlations are found to be significantly anisotropic. The modeled wave fields are then sampled at simulated altimeter observation locations, and the anomaly correlations are recalculated from the simulated altimeter data. The results are compared to the original anomaly correlations. It is found that, in general, the simulated altimeter data can capture most of the geographic and seasonal variability in the isotropic anomaly correlation length scale. The best estimates of the isotropic length scales come from a method in which correlations are calculated between pairs of observations from prior and subsequent ground tracks, in addition to along-track pairs of observations. This method was found to underestimate the isotropic anomaly correlation length scale by approximately 10%. The simulated altimeter data were not so successful in producing realistic anisotropic correlation functions. This is because of the lack of information in the zonal direction in the simulated altimeter data. However, examination of correlations along ascending and descending ground tracks separately can provide some indication of the areas on the globe for which the anomaly correlations are more anisotropic than others.
Abstract
One of the main limitations to current wave data assimilation systems is the lack of an accurate representation of the structure of the background errors. One method that may be used to determine background errors is the “NMC method.” This method examines the forecast divergence component of the background error growth by considering differences between forecasts of different ranges valid at the same time. In this paper, the NMC method is applied to global forecasts of significant wave height (SWH) and surface wind speed (U10).
It is found that the isotropic correlation length scale of the SWH forecast divergence (L SWH) has considerable geographical variability, with the longest scales just to the south of the equator in the eastern Pacific Ocean, and the shortest scales at high latitudes. The isotropic correlation length scale of the U10 forecast divergence (L U10) has a similar distribution with a stronger latitudinal dependence. It is found that both L SWH and L U10 increase as the forecast period increases. The increase in L SWH is partly due to L U10 also increasing. Another explanation is that errors in the analysis or the short-range SWH forecast propagate forward in time and disperse and their scale becomes larger. It is shown that the forecast divergence component of the background error is strongly anisotropic with the longest scales perpendicular to the likely direction of propagation of swell. In addition, in regions where the swell propagation is seasonal, the forecast divergence component of the background error shows a similar strong seasonal signal. It is suggested that the results of this study provide a lower bound to the description of the total background error in global wave models.
Abstract
One of the main limitations to current wave data assimilation systems is the lack of an accurate representation of the structure of the background errors. One method that may be used to determine background errors is the “NMC method.” This method examines the forecast divergence component of the background error growth by considering differences between forecasts of different ranges valid at the same time. In this paper, the NMC method is applied to global forecasts of significant wave height (SWH) and surface wind speed (U10).
It is found that the isotropic correlation length scale of the SWH forecast divergence (L SWH) has considerable geographical variability, with the longest scales just to the south of the equator in the eastern Pacific Ocean, and the shortest scales at high latitudes. The isotropic correlation length scale of the U10 forecast divergence (L U10) has a similar distribution with a stronger latitudinal dependence. It is found that both L SWH and L U10 increase as the forecast period increases. The increase in L SWH is partly due to L U10 also increasing. Another explanation is that errors in the analysis or the short-range SWH forecast propagate forward in time and disperse and their scale becomes larger. It is shown that the forecast divergence component of the background error is strongly anisotropic with the longest scales perpendicular to the likely direction of propagation of swell. In addition, in regions where the swell propagation is seasonal, the forecast divergence component of the background error shows a similar strong seasonal signal. It is suggested that the results of this study provide a lower bound to the description of the total background error in global wave models.
Abstract
The use of numerical guidance has become integral to the process of modern weather forecasting. Using various techniques, postprocessing of numerical model output has been shown to mitigate some of the deficiencies of these models, producing more accurate forecasts. The operational consensus forecast scheme uses past performance to bias-correct and combine numerical forecasts to produce an improved forecast at locations where recent observations are available. This technique was applied to forecasts of significant wave height (Hs ), peak period (Tp ), and 10-m wind speed (U 10) from 10 numerical wave models, at 14 buoy sites located around North America. Results show the best forecast is achieved with a weighted average of bias-corrected components for both Hs and Tp , while a weighted average of linear-corrected components gives the best results for U 10. For 24-h forecasts, improvements of 36%, 47%, and 31%, in root-mean-square-error values over the mean raw model components are achieved, or 14%, 22%, and 18% over the best individual model. Similar gains in forecast skill are retained out to 5 days. By reducing the number of models used in the construction of consensus forecasts, it is found that little forecast skill is gained beyond five or six model components, with the independence of these components, as well as individual component’s quality, being important considerations. It is noted that for Hs it is possible to beat the best individual model with a composite forecast of the worst four.
Abstract
The use of numerical guidance has become integral to the process of modern weather forecasting. Using various techniques, postprocessing of numerical model output has been shown to mitigate some of the deficiencies of these models, producing more accurate forecasts. The operational consensus forecast scheme uses past performance to bias-correct and combine numerical forecasts to produce an improved forecast at locations where recent observations are available. This technique was applied to forecasts of significant wave height (Hs ), peak period (Tp ), and 10-m wind speed (U 10) from 10 numerical wave models, at 14 buoy sites located around North America. Results show the best forecast is achieved with a weighted average of bias-corrected components for both Hs and Tp , while a weighted average of linear-corrected components gives the best results for U 10. For 24-h forecasts, improvements of 36%, 47%, and 31%, in root-mean-square-error values over the mean raw model components are achieved, or 14%, 22%, and 18% over the best individual model. Similar gains in forecast skill are retained out to 5 days. By reducing the number of models used in the construction of consensus forecasts, it is found that little forecast skill is gained beyond five or six model components, with the independence of these components, as well as individual component’s quality, being important considerations. It is noted that for Hs it is possible to beat the best individual model with a composite forecast of the worst four.
Abstract
A formalism recently developed for determining the effects of sampling errors on objectively smoothed fields constructed from an irregularly sampled dataset is applied to investigate the relative merits of single and multiple satellite altimeter missions. For small smoothing parameters, the expected squared error of smoothed fields of sea surface height (SSH) varies geographically at any particular time and temporally at any particular location. The philosophy proposed here for determining the resolution capability of SSH fields constructed from altimeter data is to identify smoothing parameters that are sufficiently large to satisfy two criteria: 1) the expected squared errors of the estimates of smoothed SSH over the space–time estimation grid must be either spatially and temporally homogeneous to within some a priori specified degree of tolerance or smaller than some a priori specified threshold, and 2) the space–time estimation grid on which smoothed SSH estimates are constructed must satisfy the Nyquist criteria for the wavenumbers and frequencies included in the smoothed fields.
The method is illustrated here by adopting a specified tolerance of 10% variability and a nominal expected squared error threshold of 1 cm2 to determine the resolution capabilities of SSH fields constructed from 10 single and multiple combinations of altimeter measurements by TOPEX/Poseidon, the ERS Earth Resource Satellites, and Geosat. Because of the lack of coordination of the orbit configurations of these satellites (different repeat periods and different orbit inclinations), the mapping resolution capabilities of the combined datasets are not significantly better than those of fields constructed from TOPEX/Poseidon data alone. The benefits of coordinated multiple missions are demonstrated by consideration of several multiple combinations of 10-, 17-, and 35-day orbit configurations.
Abstract
A formalism recently developed for determining the effects of sampling errors on objectively smoothed fields constructed from an irregularly sampled dataset is applied to investigate the relative merits of single and multiple satellite altimeter missions. For small smoothing parameters, the expected squared error of smoothed fields of sea surface height (SSH) varies geographically at any particular time and temporally at any particular location. The philosophy proposed here for determining the resolution capability of SSH fields constructed from altimeter data is to identify smoothing parameters that are sufficiently large to satisfy two criteria: 1) the expected squared errors of the estimates of smoothed SSH over the space–time estimation grid must be either spatially and temporally homogeneous to within some a priori specified degree of tolerance or smaller than some a priori specified threshold, and 2) the space–time estimation grid on which smoothed SSH estimates are constructed must satisfy the Nyquist criteria for the wavenumbers and frequencies included in the smoothed fields.
The method is illustrated here by adopting a specified tolerance of 10% variability and a nominal expected squared error threshold of 1 cm2 to determine the resolution capabilities of SSH fields constructed from 10 single and multiple combinations of altimeter measurements by TOPEX/Poseidon, the ERS Earth Resource Satellites, and Geosat. Because of the lack of coordination of the orbit configurations of these satellites (different repeat periods and different orbit inclinations), the mapping resolution capabilities of the combined datasets are not significantly better than those of fields constructed from TOPEX/Poseidon data alone. The benefits of coordinated multiple missions are demonstrated by consideration of several multiple combinations of 10-, 17-, and 35-day orbit configurations.
Abstract
Satellite altimetry provides an immensely valuable source of operational significant wave height (Hs ) data. Currently, altimeters on board Jason-1 and Envisat provide global Hs observations, available within 3–5 h of real time. In this work, Hs data from these altimeters are validated against in situ buoy data from the National Data Buoy Center (NDBC) and Marine Environmental Data Service (MEDS) buoy networks. Data cover a period of three years for Envisat and more than four years for Jason-1.
Collocation criteria of 50 km and 30 min yield 3452 and 2157 collocations for Jason-1 and Envisat, respectively. Jason-1 is found to be in no need of correction, performing well throughout the range of wave heights, although it is notably noisier than Envisat. An overall RMS difference between Jason-1 and buoy data of 0.227 m is found. Envisat has a tendency to overestimate low Hs and underestimate high Hs . A linear correction reduces the RMS difference by 7%, from 0.219 to 0.203 m.
In addition to wave height–dependent biases in the altimeter Hs estimate, a wave state–dependent bias is also identified, with steep (smooth) waves producing a negative (positive) bias relative to buoys.
A systematic difference in the Hs being reported by MEDS and NDBC buoy networks is also noted. Using the altimeter data as a common reference, it is estimated that MEDS buoys are underestimating Hs relative to NDBC buoys by about 10%.
Abstract
Satellite altimetry provides an immensely valuable source of operational significant wave height (Hs ) data. Currently, altimeters on board Jason-1 and Envisat provide global Hs observations, available within 3–5 h of real time. In this work, Hs data from these altimeters are validated against in situ buoy data from the National Data Buoy Center (NDBC) and Marine Environmental Data Service (MEDS) buoy networks. Data cover a period of three years for Envisat and more than four years for Jason-1.
Collocation criteria of 50 km and 30 min yield 3452 and 2157 collocations for Jason-1 and Envisat, respectively. Jason-1 is found to be in no need of correction, performing well throughout the range of wave heights, although it is notably noisier than Envisat. An overall RMS difference between Jason-1 and buoy data of 0.227 m is found. Envisat has a tendency to overestimate low Hs and underestimate high Hs . A linear correction reduces the RMS difference by 7%, from 0.219 to 0.203 m.
In addition to wave height–dependent biases in the altimeter Hs estimate, a wave state–dependent bias is also identified, with steep (smooth) waves producing a negative (positive) bias relative to buoys.
A systematic difference in the Hs being reported by MEDS and NDBC buoy networks is also noted. Using the altimeter data as a common reference, it is estimated that MEDS buoys are underestimating Hs relative to NDBC buoys by about 10%.
Abstract
A method for routinely verifying numerical weather prediction surface marine winds with satellite scatterometer winds is introduced. The marine surface winds from the Australian Bureau of Meteorology’s operational global and regional numerical weather prediction systems are evaluated. The model marine surface layer is described. Marine surface winds from the global and limited-area models are compared with observations, both in situ (anemometer) and remote (scatterometer). A 2-yr verification shows that wind speeds from the regional model are typically underestimated by approximately 5%, with a greater bias in the meridional direction than the zonal direction. The global model also underestimates the surface winds by around 5%–10%. A case study of a significant marine storm shows that where larger errors occur, they are due to an underestimation of the storm intensity, rather than to biases in the boundary layer parameterizations.
Abstract
A method for routinely verifying numerical weather prediction surface marine winds with satellite scatterometer winds is introduced. The marine surface winds from the Australian Bureau of Meteorology’s operational global and regional numerical weather prediction systems are evaluated. The model marine surface layer is described. Marine surface winds from the global and limited-area models are compared with observations, both in situ (anemometer) and remote (scatterometer). A 2-yr verification shows that wind speeds from the regional model are typically underestimated by approximately 5%, with a greater bias in the meridional direction than the zonal direction. The global model also underestimates the surface winds by around 5%–10%. A case study of a significant marine storm shows that where larger errors occur, they are due to an underestimation of the storm intensity, rather than to biases in the boundary layer parameterizations.
Abstract
This study examines the application of three different variations of linear-regression corrections to the surface marine winds from the Australian Bureau of Meteorology’s recently implemented operational atmospheric model. A simple correction over the entire domain is found to inadequately account for geographical variation in the wind bias. This is addressed by considering corrections that vary in space. Further, these spatially varying corrections are extended to vary in time. In an operational environment, the error characteristics of the wind forcing can be expected to change over time with the evolution of the atmospheric model. This in turn requires any applied correction to be monitored and maintained. Motivated by a desire to avoid this manual maintenance, a self-learning correction method is proposed whereby spatially and temporally varying corrections are calculated in real time from a moving window of historical comparisons between observations and preceding forecasts. This technique is shown to effectively remove both global and regionally varying wind speed biases.
Abstract
This study examines the application of three different variations of linear-regression corrections to the surface marine winds from the Australian Bureau of Meteorology’s recently implemented operational atmospheric model. A simple correction over the entire domain is found to inadequately account for geographical variation in the wind bias. This is addressed by considering corrections that vary in space. Further, these spatially varying corrections are extended to vary in time. In an operational environment, the error characteristics of the wind forcing can be expected to change over time with the evolution of the atmospheric model. This in turn requires any applied correction to be monitored and maintained. Motivated by a desire to avoid this manual maintenance, a self-learning correction method is proposed whereby spatially and temporally varying corrections are calculated in real time from a moving window of historical comparisons between observations and preceding forecasts. This technique is shown to effectively remove both global and regionally varying wind speed biases.
