Search Results

You are looking at 31 - 40 of 68 items for

  • Author or Editor: Jing Zhang x
  • Refine by Access: All Content x
Clear All Modify Search
Yi Shang
,
Lijuan Miao
,
Yunpeng Shan
,
Kaushal Raj Gnyawali
,
Jing Zhang
, and
Giri Kattel

Abstract

Reliable ultra-short-term and short-term wind speed forecasting is pivotal for clean energy development and grid operation planning. During the wind forecasting process, decomposing the measured wind speed into data with different frequencies is a solution for overcoming the nonlinearity and the randomness of the natural wind. Existing forecasting methods, a hybrid method based on empirical mode decomposition and the back propagation neural network optimized by genetic algorithm (EMD-GA-BPNN), rely on partial decomposing the measured wind speed into data with different frequencies and subsequently achieving forecasting results from machine learning algorithms. However, such methods can roughly divide IMF signals in different frequency domains, but each frequency domain contains signals with multiple frequencies. The condition reflects that the method cannot fully distinguish wind speed into data with different frequencies and thus it compromises the forecasting accuracy. A complete decomposition of measured wind speed can reduce the complexity of machine learning algorithm, and has become a useful approach for precise simulations of wind speed. Here, we propose a novel hybrid method (CEEMDAN-GA-BPNN) based on the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) by completely decomposing the measured wind speed. The decomposition results are put into the back propagation neural network optimized by a genetic algorithm (GA-BPNN), and the final forecasting results are achieved by combining all the output values by GA-BPNN for each decomposition result from CEEMDAN. We benchmark the forecasting accuracy of the proposed hybrid method against EMD-GA-BPNN integrated by EMD and GA-BPNN. From a wind farm case in Yunnan Province, China, both for ultra-short-term forecasting (15 min) and short-term forecasting (1 h), the performance of the proposed method exceeds EMD-GA-BPNN in several criteria, including root-mean-square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and coefficient of determination (R 2). The forecasting accuracy in decomposed components of low frequencies outperform components of high and middle frequencies. Fine improvement of the error metric (in percentage) in ultra-short-term/short-term forecasting is found by the complete decomposition method CEEMDAN-GA-BPNN: RMSE (7.0% and 8.6%), MAE (7.41% and 7.9%), MAPE (11.0% and 8.7%), and R 2 (2.2% and 11.0%), compared with the incomplete decomposing method EMD-GA-BPNN. Our result suggests that CEEMDAN-GA-BPNN could be an accurate wind speed forecasting tool for wind farms development and intelligent grid operations.

Significance Statement

Nonlinearity and randomness of natural wind speed data are the limitations for short-term and ultra-short-term wind speed forecasting. By decreasing forecasting error in machine learning training process, data decomposition for the measured wind speed has become an effective method for overcoming this issue. Nonetheless, the normal incomplete decomposition method will compromise the extent of forecasting accuracy. We introduce a novel hybrid and complete decomposition method CEEMDAN-GA-BPNN (the complete decomposition method). Measured wind speed data from a wind farm in Yunnan Province, China, has been utilized. CEEMDAN-GA-BPNN outperforms EMD-GA-BPNN (the partial decomposition method) in forecasting accuracy both in the ultra-short-term and the short-term wind speed forecasting.

Full access
Qiong Wu
,
Hong-Qing Wang
,
Yin-Jing Lin
,
Yi-Zhou Zhuang
, and
Yan Zhang

Abstract

An optical flow algorithm based on polynomial expansion (OFAPE) was used to derive atmospheric motion vectors (AMVs) from geostationary satellite images. In OFAPE, there are two parameters that can affect the AMV results: the sizes of the expansion window and optimization window. They should be determined according to the temporal interval and spatial resolution of satellite images. A helpful experiment was conducted for selecting those sizes. The limitations of window sizes can cause loss of strong wind speed, and an image-pyramid scheme was used to overcome this problem. Determining the heights of AMVs for semitransparent cloud pixels (STCPs) is challenging work in AMV derivation. In this study, two-dimensional histograms (H2Ds) between infrared brightness temperatures (6.7- and 10.8-μm channels) formed from a long time series of cloud images were used to identify the STCPs and to estimate their actual temperatures/heights. The results obtained from H2Ds were contrasted with CloudSat radar reflectivity and CALIPSO cloud-feature mask data. Finally, in order to verify the algorithm adaptability, three-month AMVs (JJA 2013) were calculated and compared with the wind fields of ERA data and the NOAA/ESRL radiosonde observations in three aspects: speed, direction, and vector difference.

Full access
Shen-Ming Fu
,
Jing-Ping Zhang
,
Jian-Hua Sun
, and
Tian-Bao Zhao

ABSTRACT

A 14-yr climatology is presented of the mesoscale vortices generated in the vicinity of the Dabie Mountains [Dabie vortices (DBVs)] in the Yangtze River valley. Analyzing these vortices using the Climate Forecast System Reanalysis (CFSR), DBVs were found to be a frequent type of summer mesoscale weather system, with a mean monthly frequency of 12.2. DBVs were mainly located in the middle and lower troposphere, and ~92% of them triggered precipitation. Most DBVs were short lived, and only 19.5% persisted for more than 12 h. Latent heat release associated with precipitation is a dominant factor for the DBV’s three-dimensional geometry features, life span, and intensity.

The long-lived DBVs, all of which triggered torrential rainfall, were analyzed using a composite analysis under the normalized polar coordinates. Results indicate that these vortices generally moved eastward and northeastward, which corresponded to the vortices’ orientation, divergence, vorticity budget, and kinetic energy budget. The evolution of long-lived DBVs featured significant unevenness: those octants located at the front and on the right side of the vortices’ moving tracks were more favorable for their development and maintenance, while those octants located at the back and on the left side acted conversely. Convergence-related shrinking was the most favorable factor for the vortices’ development and persistence, while the tilting effect was a dominant factor accounting for their attenuation. Long-lived DBVs featured strong baroclinity, and the baroclinic energy conversion acted as the main energy source for the vortices’ evolution. In contrast, the barotropic energy conversion favored the vortices’ development and maintenance at first, and later triggered their dissipation.

Full access
Wenjun Zhang
,
Fei-Fei Jin
,
Jing-Xia Zhao
,
Li Qi
, and
Hong-Li Ren

Abstract

A severe drought struck southwest China during autumn 2009, which had a huge impact on productivity and the lives of the affected population. A nonconventional El Niño, the so-called warm pool (WP) El Niño, was supposed to be a principal factor of this strong autumn drought. In sharp contrast to a conventional El Niño, in the 2009 WP El Niño year the maximum sea surface temperature (SST) anomalies are confined to the central equatorial Pacific Ocean. Moreover, this WP El Niño was characterized by the relatively farther westward location and the strongest intensity among the WP El Niño events in the past 60 years. Observations and modeling studies both indicate that the rainfall deficits over southwest China are significantly influenced by the combined effects of the location and intensity of the WP El Niño. That is, the drought over southwest China tends to be more severe when the warming SST anomalies associated with the WP El Niño are located farther westward and are stronger. Therefore, the strong autumn drought over southwest China in 2009 can be largely attributed to the concurrent distinctive WP El Niño, which generates a strongly anomalous cyclone over the west North Pacific and leads to a serious reduction in rainfall over southwest China. The influence of the Indian Ocean warming on autumn rainfall over southwest China was also examined but seems to have little contribution to this drought.

Full access
Jing-Wu Liu
,
Shang-Ping Xie
,
Joel R. Norris
, and
Su-Ping Zhang

Abstract

A sharp sea surface temperature front develops between the warm water of the Gulf Stream and cold continental shelf water in boreal winter. This front has a substantial impact on the marine boundary layer. The present study analyzes and synthesizes satellite observations and reanalysis data to examine how the sea surface temperature front influences the three-dimensional structure of low-level clouds. The Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) satellite captures a sharp low-level cloud transition across the Gulf Stream front, a structure frequently observed under the northerly condition. Low-level cloud top (<4 km) increases by about 500 m from the cold to the warm flank of the front. The sea surface temperature front induces a secondary low-level circulation through sea level pressure adjustment with ascending motion over the warm water and descending motion over cold water. The secondary circulation further contributes to the cross-frontal transition of low-level clouds. Composite analysis shows that surface meridional advection over the front plays an important role in the development of the marine atmospheric boundary layer and low-level clouds. Under cold northerly advection over the Gulf Stream front, strong near-surface instability leads to a well-mixed boundary layer over the Gulf Stream, causing southward deepening of low-level clouds across the sea surface temperature front. Moreover, the front affects the freezing level by transferring heat to the atmosphere and therefore influences the cross-frontal variation of the cloud phase.

Full access
Jing-Wu Liu
,
Shang-Ping Xie
,
Shuang Yang
, and
Su-Ping Zhang

Abstract

The East China Sea Kuroshio (ECSK) flows in the East Asian monsoon region where the background atmospheric circulation varies significantly with season. A sea surface temperature (SST) front associated with the ECSK becomes narrower and sharper from winter to spring. The present study investigates how low clouds respond to the ECSK front in different seasons by synthesizing spaceborne lidar and surface visual observations. The results reveal prominent cross-frontal transitions in low clouds, which exhibit distinct behavior between winter and spring. In winter, cloud responses are generally confined below 4 km by the strong background descending motion and feature a gradual cloud-top elevation from the cold to the warm flank of the front. The ice clouds on the cold flank of the ECSK front transform into liquid water clouds and rain on the warm flank. The springtime clouds, by contrast, are characterized by a sharp cross-frontal transition with deep clouds reaching up to 7 km over the ECSK. In both winter and spring, the low-cloud morphology exhibits a large transformation from the cold to the warm flank of the ECSK front, including increases in cloud-top height, a decline in smoothness of cloud top, and the transition from stratiform to convective clouds. All this along with the atmospheric soundings indicates that the decoupling of the marine atmospheric boundary layer (MABL) is more prevalent on the warm flank of the front. Thus, long-term observations reveal prominent cross-frontal low-cloud transitions in morphology associated with MABL decoupling that resemble a large-scale cloud-regime transition over the eastern subtropical Pacific.

Full access
Jing Li
,
Barbara E. Carlson
,
William B. Rossow
,
Andrew A. Lacis
, and
Yuanchong Zhang

Abstract

Because of the importance of clouds in modulating Earth’s energy budget, it is critical to understand their variability in space and time for climate and modeling studies. This study examines the consistency of the spatiotemporal variability of cloud amount (CA) and cloud-top pressure (CTP) represented by five 7-yr satellite datasets from the Global Energy and Water Cycle Experiment (GEWEX) cloud assessment project, and total cloud fraction observation from the Extended Edited Cloud Reports Archive (EECRA). Two spectral analysis techniques, namely combined maximum covariance analysis (CMCA) and combined principal component analysis (CPCA), are used to extract the dominant modes of variability from the combined datasets, and the resulting spatial patterns are compared in parallel. The results indicate that the datasets achieve overall excellent agreement on both seasonal and interannual scales of variability, with the correlations between the spatial patterns mostly above 0.6 and often above 0.8. For seasonal variability, the largest differences are found in the Northern Hemisphere high latitudes and near the South African coast for CA and in the Sahel region for CTP, where some differences in the phase and strength of the seasonal cycle are found. On interannual scales, global cloud variability is mostly associated with major climate modes, including El Niño–Southern Oscillation (ENSO), the Pacific decadal oscillation (PDO), and the Indian Ocean dipole mode (IODM), and the datasets also agree reasonably well. The good agreement across the datasets supports the conclusion that they are describing cloud variations with these climate modes.

Full access
Suhua Liu
,
Hongbo Su
,
Jing Tian
,
Renhua Zhang
,
Weizhen Wang
, and
Yueru Wu

Abstract

Surface air temperature is a basic meteorological variable to monitor the environment and assess climate change. Four remote sensing methods—the temperature–vegetation index (TVX), the univariate linear regression method, the multivariate linear regression method, and the advection-energy balance for surface air temperature (ADEBAT)—have been developed to acquire surface air temperature on a regional scale. To evaluate their utilities, they were applied to estimate the surface air temperature in northwestern China and were compared with each other through regressive analyses, t tests, estimation errors, and analyses on estimations of different underlying surfaces. Results can be summarized into three aspects: 1) The regressive analyses and t tests indicate that the multivariate linear regression method and the ADEBAT provide better accuracy than the other two methods. 2) Frequency histograms on estimation errors show that the multivariate linear regression method produces the minimum error range, and the univariate linear regression method produces the maximum error range. Errors of the multivariate linear regression method exhibit a nearly normal distribution and that of the ADEBAT exhibit a bimodal distribution, whereas the other two methods display negative skewness distributions. 3) Estimates on different underlying surfaces show that the TVX and the univariate linear regression method are significantly limited in regions with sparse vegetation cover. The multivariate linear regression method has estimation errors within 1°C and without high levels of errors, and the ADEBAT also produces high estimation errors on bare ground.

Full access
Yu Wang
,
Hong-Qing Wang
,
Lei Han
,
Yin-Jing Lin
, and
Yan Zhang

Abstract

This study was designed to provide basic information for the improvement of storm nowcasting. According to the mean direction deviation of storm movement, storms were classified into three types: 1) steady storms (S storms, extrapolated efficiently), 2) unsteady storms (U storms, extrapolated poorly), and 3) transitional storms (T storms). The U storms do not fit the linear extrapolation processes because of their unsteady movements. A 6-yr warm-season radar observation dataset was used to highlight and analyze the differences between U storms and S storms. The analysis included geometric features, dynamic factors, and environmental parameters. The results showed that storms with the following characteristics changed movement direction most easily in the Beijing–Tianjin region: 1) smaller storm area, 2) lower thickness (echo-top height minus base height), 3) lower movement speed, 4) weaker updrafts and the maximum value located in the mid- and upper troposphere, 5) storm-relative vertical wind profiles dominated by directional shear instead of speed shear, 6) lower relative humidity in the mid- and upper troposphere, and 7) higher surface evaporation and ground roughness.

Full access
Jingzhuo Wang
,
Jing Chen
,
Jun Du
,
Yutao Zhang
,
Yu Xia
, and
Guo Deng

Abstract

This study demonstrates how model bias can adversely affect the quality assessment of an ensemble prediction system (EPS) by verification metrics. A regional EPS [Global and Regional Assimilation and Prediction Enhanced System-Regional Ensemble Prediction System (GRAPES-REPS)] was verified over a period of one month over China. Three variables (500-hPa and 2-m temperatures, and 250-hPa wind) are selected to represent “strong” and “weak” bias situations. Ensemble spread and probabilistic forecasts are compared before and after a bias correction. The results show that the conclusions drawn from ensemble verification about the EPS are dramatically different with or without model bias. This is true for both ensemble spread and probabilistic forecasts. The GRAPES-REPS is severely underdispersive before the bias correction but becomes calibrated afterward, although the improvement in the spread’s spatial structure is much less; the spread–skill relation is also improved. The probabilities become much sharper and almost perfectly reliable after the bias is removed. Therefore, it is necessary to remove forecast biases before an EPS can be accurately evaluated since an EPS deals only with random error but not systematic error. Only when an EPS has no or little forecast bias, can ensemble verification metrics reliably reveal the true quality of an EPS without removing forecast bias first. An implication is that EPS developers should not be expected to introduce methods to dramatically increase ensemble spread (either by perturbation method or statistical calibration) to achieve reliability. Instead, the preferred solution is to reduce model bias through prediction system developments and to focus on the quality of spread (not the quantity of spread). Forecast products should also be produced from the debiased but not the raw ensemble.

Open access