Search Results
You are looking at 1 - 9 of 9 items for :
- Author or Editor: Wei Yu x
- Journal of Applied Meteorology and Climatology x
- Refine by Access: All Content x
Abstract
The capability of the Canadian land surface external modeling system known as the Global Environmental Multiscale Surface (GEM-SURF) system with respect to surface wind predictions is evaluated. Based on the Interactions between Soil, Biosphere, and Atmosphere (ISBA) land surface scheme, and an exponential power law adjusted to the local stability conditions for the prediction of surface winds, the system allows decoupling of surface processes from those of the free atmosphere and enables high resolutions at the surface as dictated by the small-scale heterogeneities of the surface boundary. The simulations are driven by downscaled forecasts from the Regional Deterministic Prediction System, the 15-km Canadian regional operational modeling system. High-resolution, satellite-derived datasets of orography, vegetation, and soil cover are used to depict the surface boundary. The integration domains cover Canada’s eastern provinces at resolutions ranging from that of the driving model to resolutions similar to those of the geophysical datasets. The GEM-SURF predictions outperform those of the driving operational model. Reduction of the standard error and improvement of the model skill is seen as resolution increases, for all wind speeds. Further, the bias error is reduced in association with a rise in the corresponding value of the roughness length. For all examined resolutions GEM-SURF’s predictions are shown to be superior to those obtained through a simple statistical downscaling. In the prospect of the future development of a multicomponent system that provides wind forecasts at levels of wind energy generation, GEM-SURF’s potential for improved scores at the surface and its limited requirements in computer resources make it a suitable surface component of such a system.
Abstract
The capability of the Canadian land surface external modeling system known as the Global Environmental Multiscale Surface (GEM-SURF) system with respect to surface wind predictions is evaluated. Based on the Interactions between Soil, Biosphere, and Atmosphere (ISBA) land surface scheme, and an exponential power law adjusted to the local stability conditions for the prediction of surface winds, the system allows decoupling of surface processes from those of the free atmosphere and enables high resolutions at the surface as dictated by the small-scale heterogeneities of the surface boundary. The simulations are driven by downscaled forecasts from the Regional Deterministic Prediction System, the 15-km Canadian regional operational modeling system. High-resolution, satellite-derived datasets of orography, vegetation, and soil cover are used to depict the surface boundary. The integration domains cover Canada’s eastern provinces at resolutions ranging from that of the driving model to resolutions similar to those of the geophysical datasets. The GEM-SURF predictions outperform those of the driving operational model. Reduction of the standard error and improvement of the model skill is seen as resolution increases, for all wind speeds. Further, the bias error is reduced in association with a rise in the corresponding value of the roughness length. For all examined resolutions GEM-SURF’s predictions are shown to be superior to those obtained through a simple statistical downscaling. In the prospect of the future development of a multicomponent system that provides wind forecasts at levels of wind energy generation, GEM-SURF’s potential for improved scores at the surface and its limited requirements in computer resources make it a suitable surface component of such a system.
Abstract
A new statistical–dynamical downscaling procedure is developed and then applied to high-resolution (regional) time series generation and wind resource assessment. The statistical module of the new procedure uses empirical orthogonal function (EOF) analysis for the generation of large-scale atmospheric component patterns. The dominant atmospheric patterns (associated with the EOF modes explaining most of the statistical variance) are then dynamically downscaled or adjusted to high-resolution terrain and surface roughness by using the Global Environmental Multiscale–Limited Area Model (GEM-LAM). Regional time series are constructed using the model outputs. The new method is applied to the Gaspé region of Québec in Canada. The dataset used is the NCEP–NCAR reanalysis of wind, temperature, humidity, and geopotential height during the period 1958–2004. Regional time series of wind speed and temperature are constructed, and a numerical wind atlas of the Gaspé region is generated. The generated time series and the numerical wind atlas are compared with observations at different masts located in the Gaspé Peninsula and are also compared with a numerical wind atlas for the same region generated in Yu et al. The results suggest that the newly developed procedure can be useful to generate regional time series and reasonably accurate numerical wind atlases using large-scale data with much less computational effort than previous techniques.
Abstract
A new statistical–dynamical downscaling procedure is developed and then applied to high-resolution (regional) time series generation and wind resource assessment. The statistical module of the new procedure uses empirical orthogonal function (EOF) analysis for the generation of large-scale atmospheric component patterns. The dominant atmospheric patterns (associated with the EOF modes explaining most of the statistical variance) are then dynamically downscaled or adjusted to high-resolution terrain and surface roughness by using the Global Environmental Multiscale–Limited Area Model (GEM-LAM). Regional time series are constructed using the model outputs. The new method is applied to the Gaspé region of Québec in Canada. The dataset used is the NCEP–NCAR reanalysis of wind, temperature, humidity, and geopotential height during the period 1958–2004. Regional time series of wind speed and temperature are constructed, and a numerical wind atlas of the Gaspé region is generated. The generated time series and the numerical wind atlas are compared with observations at different masts located in the Gaspé Peninsula and are also compared with a numerical wind atlas for the same region generated in Yu et al. The results suggest that the newly developed procedure can be useful to generate regional time series and reasonably accurate numerical wind atlases using large-scale data with much less computational effort than previous techniques.
Abstract
The accuracy of rain-rate estimation using polarimetric radar measurements has been improved as a result of better characterization of radar measurement quality and rain microphysics. In the literature, a variety of power-law relations between polarimetric radar measurements and rain rate are described because of the dynamic or varying nature of rain microphysics. A variational technique that concurrently takes into account radar observational error and dynamically varying rain microphysics is proposed in this study. Rain-rate estimation using the variational algorithm that uses event-based observational error and background rain climatological values is evaluated using observing system simulation experiments (OSSE), and its performance is demonstrated in the case of an epic Colorado flood event. The rain event occurred between 11 and 12 September 2013. The results from OSSE show that the variational algorithm with event-based observational error consistently estimates more accurate rain rate than does the “R(Z HH, Z DR)” power-law algorithm. On the contrary, the usage of ad hoc or improper observational error degrades the performance of the variational method. Furthermore, the variational algorithm is less sensitive to the observational error of differential reflectivity Z DR than is the R(Z HH, Z DR) algorithm. The variational quantitative precipitation estimation (QPE) retrieved more accurate rainfall estimation than did the power-law dual-polarization QPE in this particular event, despite the fact that both algorithms used the same dual-polarization radar measurements from the Next Generation Weather Radar (NEXRAD).
Abstract
The accuracy of rain-rate estimation using polarimetric radar measurements has been improved as a result of better characterization of radar measurement quality and rain microphysics. In the literature, a variety of power-law relations between polarimetric radar measurements and rain rate are described because of the dynamic or varying nature of rain microphysics. A variational technique that concurrently takes into account radar observational error and dynamically varying rain microphysics is proposed in this study. Rain-rate estimation using the variational algorithm that uses event-based observational error and background rain climatological values is evaluated using observing system simulation experiments (OSSE), and its performance is demonstrated in the case of an epic Colorado flood event. The rain event occurred between 11 and 12 September 2013. The results from OSSE show that the variational algorithm with event-based observational error consistently estimates more accurate rain rate than does the “R(Z HH, Z DR)” power-law algorithm. On the contrary, the usage of ad hoc or improper observational error degrades the performance of the variational method. Furthermore, the variational algorithm is less sensitive to the observational error of differential reflectivity Z DR than is the R(Z HH, Z DR) algorithm. The variational quantitative precipitation estimation (QPE) retrieved more accurate rainfall estimation than did the power-law dual-polarization QPE in this particular event, despite the fact that both algorithms used the same dual-polarization radar measurements from the Next Generation Weather Radar (NEXRAD).
Abstract
A variational algorithm for estimating measurement error covariance and the attenuation of X-band polarimetric radar measurements is described. It concurrently uses both the differential reflectivity Z DR and propagation phase ΦDP. The majority of the current attenuation estimation techniques use only ΦDP. A few of the ΦDP-based methods use Z DR as a constraint for verifying estimated attenuation. In this paper, a detailed observing system simulation experiment was used for evaluating the performance of the variational algorithm. The results were compared with a single-coefficient ΦDP-based method. Retrieved attenuation from the variational method is more accurate than the results from a single coefficient ΦDP-based method. Moreover, the variational method is less sensitive to measurement noise in radar observations. The variational method requires an accurate description of error covariance matrices. Relative weights between measurements and background values (i.e., mean value based on long-term DSD measurements in the variational method) are determined by their respective error covariances. Instead of using ad hoc values, error covariance matrices of background and radar measurement are statistically estimated and their spatial characteristics are studied. The estimated error covariance shows higher values in convective regions than in stratiform regions, as expected. The practical utility of the variational attenuation correction method is demonstrated using radar field measurements from the Taiwan Experimental Atmospheric Mobile-Radar (TEAM-R) during 2008’s Southwest Monsoon Experiment/Terrain-Influenced Monsoon Rainfall Experiment (SoWMEX/TiMREX). The accuracy of attenuation-corrected X-band radar measurements is evaluated by comparing them with collocated S-band radar measurements.
Abstract
A variational algorithm for estimating measurement error covariance and the attenuation of X-band polarimetric radar measurements is described. It concurrently uses both the differential reflectivity Z DR and propagation phase ΦDP. The majority of the current attenuation estimation techniques use only ΦDP. A few of the ΦDP-based methods use Z DR as a constraint for verifying estimated attenuation. In this paper, a detailed observing system simulation experiment was used for evaluating the performance of the variational algorithm. The results were compared with a single-coefficient ΦDP-based method. Retrieved attenuation from the variational method is more accurate than the results from a single coefficient ΦDP-based method. Moreover, the variational method is less sensitive to measurement noise in radar observations. The variational method requires an accurate description of error covariance matrices. Relative weights between measurements and background values (i.e., mean value based on long-term DSD measurements in the variational method) are determined by their respective error covariances. Instead of using ad hoc values, error covariance matrices of background and radar measurement are statistically estimated and their spatial characteristics are studied. The estimated error covariance shows higher values in convective regions than in stratiform regions, as expected. The practical utility of the variational attenuation correction method is demonstrated using radar field measurements from the Taiwan Experimental Atmospheric Mobile-Radar (TEAM-R) during 2008’s Southwest Monsoon Experiment/Terrain-Influenced Monsoon Rainfall Experiment (SoWMEX/TiMREX). The accuracy of attenuation-corrected X-band radar measurements is evaluated by comparing them with collocated S-band radar measurements.
Abstract
Icing is a weather phenomenon that is typical of cold climates. It impacts human activities through ice accretion on tower structures, transmission lines, and the blades of wind turbines. Icing on turbine blades, in particular, results in wind turbine performance degradation and/or safety shutdowns. The objective of this study is to explore the feasibility of using a coupled atmospheric and ice load model to simulate icing start-up, duration, and amount while also quantitatively evaluating power loss in wind plants related to icing events and mechanisms. Eight of 27 icing episodes identified for a wind plant in the Gaspé region of Québec (Canada) during the period 2008–10 were simulated using a mesoscale model (the Global Environmental Multiscale Limited-Area Model, or GEM-LAM). The simulations were verified using near-surface temperature, relative humidity, and wind speed, all of which compared well to in situ observations. Simulated wind speed, precipitation, cloud liquid water content, and median volume diameter of the droplets were used to drive ice load models to simulate the total ice load on a cylindrical structure. The three ice load models accounted for freezing rain, wet snow, and in-cloud icing, respectively, and in all three cases a sink term was added to account for melting due to radiation. The start-up and duration of ice were well captured by the coupled model, and a positive correlation was found between icing episodes and wind power reduction. This study demonstrates the improvements of the icing forecasts by using three ice load models, and provides a framework for both qualitative and quantitative evaluation of icing impact on wind turbine operations.
Abstract
Icing is a weather phenomenon that is typical of cold climates. It impacts human activities through ice accretion on tower structures, transmission lines, and the blades of wind turbines. Icing on turbine blades, in particular, results in wind turbine performance degradation and/or safety shutdowns. The objective of this study is to explore the feasibility of using a coupled atmospheric and ice load model to simulate icing start-up, duration, and amount while also quantitatively evaluating power loss in wind plants related to icing events and mechanisms. Eight of 27 icing episodes identified for a wind plant in the Gaspé region of Québec (Canada) during the period 2008–10 were simulated using a mesoscale model (the Global Environmental Multiscale Limited-Area Model, or GEM-LAM). The simulations were verified using near-surface temperature, relative humidity, and wind speed, all of which compared well to in situ observations. Simulated wind speed, precipitation, cloud liquid water content, and median volume diameter of the droplets were used to drive ice load models to simulate the total ice load on a cylindrical structure. The three ice load models accounted for freezing rain, wet snow, and in-cloud icing, respectively, and in all three cases a sink term was added to account for melting due to radiation. The start-up and duration of ice were well captured by the coupled model, and a positive correlation was found between icing episodes and wind power reduction. This study demonstrates the improvements of the icing forecasts by using three ice load models, and provides a framework for both qualitative and quantitative evaluation of icing impact on wind turbine operations.
Abstract
As the new-generation geostationary satellite Himawari-8 provides a greater frequency and more observation channels than its predecessor, the Multifunctional Transport Satellite series (e.g., MTSAT-2), an opportunity arises to generate atmospheric motion vectors (AMVs) with an increased accuracy and extensive distribution over eastern Asia. In this work AMVs were derived from consecutive images of an infrared-window channel (IR1) of the Himawari-8 satellite using particle image velocimetry (PIV) based on the theory of cross-correlation schemes. A multipass scheme and an adaptive interrogation scheme were also employed to increase spatial resolution and accuracy. For height assignment, an infrared-window method was applied for opaque cloud, while an H2O-intercept method was employed for semitransparent cloud. Validation was conducted by comparing the PIV-derived AMVs with wind fields obtained from NWP analysis, radiosonde observations, and the operational system from the Meteorological Satellite Center (MSC) of the Japan Meteorological Agency (JMA) or JMA/MSC. The comparison of wind velocity maps with the NWP data shows that the PIV-derived AMVs are capable of quantitatively depicting full-field wind field maps and strong jets in atmospheric circulation. Through comparisons with radiosonde observations, the root-mean-square error and wind speed bias (4.29 and −1.05 m s−1) of the PIV-derived AMVs are comparable to, although slightly greater than, that of the NWP data (3.88 and −0.26 m s−1). Based on comparison between the PIV-derived AMVs and wind fields obtained from the JMA/MSC operational system, the PIV-derived AMVs are again comparable, producing a slightly lower error but a larger wind speed bias (−1.05 vs 0.20 m s−1). This also implies that a better height assignment algorithm is necessary.
Abstract
As the new-generation geostationary satellite Himawari-8 provides a greater frequency and more observation channels than its predecessor, the Multifunctional Transport Satellite series (e.g., MTSAT-2), an opportunity arises to generate atmospheric motion vectors (AMVs) with an increased accuracy and extensive distribution over eastern Asia. In this work AMVs were derived from consecutive images of an infrared-window channel (IR1) of the Himawari-8 satellite using particle image velocimetry (PIV) based on the theory of cross-correlation schemes. A multipass scheme and an adaptive interrogation scheme were also employed to increase spatial resolution and accuracy. For height assignment, an infrared-window method was applied for opaque cloud, while an H2O-intercept method was employed for semitransparent cloud. Validation was conducted by comparing the PIV-derived AMVs with wind fields obtained from NWP analysis, radiosonde observations, and the operational system from the Meteorological Satellite Center (MSC) of the Japan Meteorological Agency (JMA) or JMA/MSC. The comparison of wind velocity maps with the NWP data shows that the PIV-derived AMVs are capable of quantitatively depicting full-field wind field maps and strong jets in atmospheric circulation. Through comparisons with radiosonde observations, the root-mean-square error and wind speed bias (4.29 and −1.05 m s−1) of the PIV-derived AMVs are comparable to, although slightly greater than, that of the NWP data (3.88 and −0.26 m s−1). Based on comparison between the PIV-derived AMVs and wind fields obtained from the JMA/MSC operational system, the PIV-derived AMVs are again comparable, producing a slightly lower error but a larger wind speed bias (−1.05 vs 0.20 m s−1). This also implies that a better height assignment algorithm is necessary.
Abstract
Area-averaged estimates of Cn 2 from high-resolution numerical weather prediction (NWP) model output are produced from local estimates of the spatial structure functions of refractive index with corrections for the inherent smoothing and filtering effects of the underlying NWP model. The key assumptions are the existence of a universal statistical description of small-scale turbulence and a locally universal spatial filter for the NWP model variables. Under these assumptions, spatial structure functions of the NWP model variables can be related to the structure functions of the atmospheric variables and extended to the smaller underresolved scales. The shape of the universal spatial filter is determined by comparisons of model structure functions with the climatological spatial structure function determined from an archive of aircraft data collected in the upper troposphere and lower stratosphere. This method of computing Cn 2 has an important advantage over more traditional methods that are based on vertical differences because the structure function–based estimates avoid reference to the turbulence outer length scale. To evaluate the technique, NWP model–derived structure-function estimates of Cn 2 are compared with nighttime profiles of Cn 2 derived from temperature structure-function sensors attached to a rawinsonde (thermosonde) near Holloman Air Force Base in the United States.
Abstract
Area-averaged estimates of Cn 2 from high-resolution numerical weather prediction (NWP) model output are produced from local estimates of the spatial structure functions of refractive index with corrections for the inherent smoothing and filtering effects of the underlying NWP model. The key assumptions are the existence of a universal statistical description of small-scale turbulence and a locally universal spatial filter for the NWP model variables. Under these assumptions, spatial structure functions of the NWP model variables can be related to the structure functions of the atmospheric variables and extended to the smaller underresolved scales. The shape of the universal spatial filter is determined by comparisons of model structure functions with the climatological spatial structure function determined from an archive of aircraft data collected in the upper troposphere and lower stratosphere. This method of computing Cn 2 has an important advantage over more traditional methods that are based on vertical differences because the structure function–based estimates avoid reference to the turbulence outer length scale. To evaluate the technique, NWP model–derived structure-function estimates of Cn 2 are compared with nighttime profiles of Cn 2 derived from temperature structure-function sensors attached to a rawinsonde (thermosonde) near Holloman Air Force Base in the United States.
Abstract
The southwest vortex (SWV) is a critical weather system in China, but our knowledge of this system remains incomplete. Here, we investigate the cloud properties in the SWV. First, we search for the SWVs with time steps and center locations that are consistent between the SWV yearbook and ERA-Interim reanalysis data. Second, we supplement these SWVs’ life spans and movement paths. Third, we relocate the Fengyun (FY) satellite FY-4A cloud retrievals in the 10° × 10° region centered on each SWV and analyze the cloud occurrence frequency (COF), cloud-top height (CTH), and cloud optical thickness (COT). A distribution mode of cloud types is summarized from the COFs, with water clouds, supercooled clouds, mixed clouds, ice clouds, cirrus clouds, and overlap clouds occurring sequentially from west to east. The CTH probability density (PD) distribution features a significant north–south difference. In addition, the COT PD distributions exhibit a common trend: with increasing COT, the PD increases rapidly and then slowly before peaking, whereupon the PD decreases abruptly. From spring to summer, the region with the highest convective COF shifts from the northeast to the northwest, and an east–west gradient of the convective COF appears in autumn and winter. Furthermore, we investigate the cloud properties during SWV-related heavy rainfall. Heavy rain occurs mainly in the west of the SWV, and convective clouds are mainly in the northwest, partly in the southwest and near the SWV center. The average CTH in heavy rainfall is generally higher than 6 km, and the average COT is greater than 20.
Significance Statement
The southwest vortex (SWV) is an important weather system in China. However, we do not yet comprehensively know this weather system. The cloud properties can indicate the structures of weather systems and are key parameters in numerical weather prediction (NWP) models. Thus, investigating cloud properties is necessary and meaningful to understand the SWV and accurately predict SWV-related precipitation in NWP models. In this paper, a typical distribution mode of six cloud types in the SWV is summarized from the cloud occurrence frequency, and the distribution features of convective clouds, cloud-top height, and cloud optical thickness in the SWV are analyzed. Furthermore, the cloud properties in SWV-related heavy rain are also studied.
Abstract
The southwest vortex (SWV) is a critical weather system in China, but our knowledge of this system remains incomplete. Here, we investigate the cloud properties in the SWV. First, we search for the SWVs with time steps and center locations that are consistent between the SWV yearbook and ERA-Interim reanalysis data. Second, we supplement these SWVs’ life spans and movement paths. Third, we relocate the Fengyun (FY) satellite FY-4A cloud retrievals in the 10° × 10° region centered on each SWV and analyze the cloud occurrence frequency (COF), cloud-top height (CTH), and cloud optical thickness (COT). A distribution mode of cloud types is summarized from the COFs, with water clouds, supercooled clouds, mixed clouds, ice clouds, cirrus clouds, and overlap clouds occurring sequentially from west to east. The CTH probability density (PD) distribution features a significant north–south difference. In addition, the COT PD distributions exhibit a common trend: with increasing COT, the PD increases rapidly and then slowly before peaking, whereupon the PD decreases abruptly. From spring to summer, the region with the highest convective COF shifts from the northeast to the northwest, and an east–west gradient of the convective COF appears in autumn and winter. Furthermore, we investigate the cloud properties during SWV-related heavy rainfall. Heavy rain occurs mainly in the west of the SWV, and convective clouds are mainly in the northwest, partly in the southwest and near the SWV center. The average CTH in heavy rainfall is generally higher than 6 km, and the average COT is greater than 20.
Significance Statement
The southwest vortex (SWV) is an important weather system in China. However, we do not yet comprehensively know this weather system. The cloud properties can indicate the structures of weather systems and are key parameters in numerical weather prediction (NWP) models. Thus, investigating cloud properties is necessary and meaningful to understand the SWV and accurately predict SWV-related precipitation in NWP models. In this paper, a typical distribution mode of six cloud types in the SWV is summarized from the cloud occurrence frequency, and the distribution features of convective clouds, cloud-top height, and cloud optical thickness in the SWV are analyzed. Furthermore, the cloud properties in SWV-related heavy rain are also studied.
Abstract
Based on daily meteorological observation data in South China (SC) from 1967 to 2018, the spatiotemporal characteristics of the precipitation in SC over the past 52 years were studied. Only 8% of the stations showed a significant increase in annual rainfall, and there was no significant negative trend at any weather stations at a confidence level of 90%. Monthly rainfall showed the most significant decreasing and increasing trends in April and November, respectively. During the entire flooding season from April to September, the monthly rainfall at the weather stations in the coastal areas showed almost no significant change. The annual rainfall gradually decreased toward the inland area with the central and coastal areas of Guangdong Province as the high-value rainfall center. By using the empirical orthogonal function decomposition method, it was found that the two main monthly rainfall modes had strong annual signals. The first modal spatial distribution was basically consistent with the average annual rainfall distribution. Based on the environmental background analysis, it was found that during the flooding season the main water vapor to SC was transported by the East Asian summer monsoon and the Indian summer monsoon. In late autumn and winter, the prevailing wind from northeastern China could not bring much water vapor to SC and led to little precipitation in these two seasons. The spatial distribution of precipitation in SC during summer was more consistent with the moisture flux divergence distribution of the bottom layer from 925 to 1000 hPa rather than that of the layer from 700 to 1000 hPa.
Abstract
Based on daily meteorological observation data in South China (SC) from 1967 to 2018, the spatiotemporal characteristics of the precipitation in SC over the past 52 years were studied. Only 8% of the stations showed a significant increase in annual rainfall, and there was no significant negative trend at any weather stations at a confidence level of 90%. Monthly rainfall showed the most significant decreasing and increasing trends in April and November, respectively. During the entire flooding season from April to September, the monthly rainfall at the weather stations in the coastal areas showed almost no significant change. The annual rainfall gradually decreased toward the inland area with the central and coastal areas of Guangdong Province as the high-value rainfall center. By using the empirical orthogonal function decomposition method, it was found that the two main monthly rainfall modes had strong annual signals. The first modal spatial distribution was basically consistent with the average annual rainfall distribution. Based on the environmental background analysis, it was found that during the flooding season the main water vapor to SC was transported by the East Asian summer monsoon and the Indian summer monsoon. In late autumn and winter, the prevailing wind from northeastern China could not bring much water vapor to SC and led to little precipitation in these two seasons. The spatial distribution of precipitation in SC during summer was more consistent with the moisture flux divergence distribution of the bottom layer from 925 to 1000 hPa rather than that of the layer from 700 to 1000 hPa.
