A New Objective Typhoon Location Algorithm Considering a Perturbation Factor Based on FY-4A Brightness Temperature Data

Tao Xie aSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing, China
bLaboratory for Regional Oceanography and Numerical Modeling, Qingdao National Laboratory for Marine Science and Technology, Qingdao, China

Search for other papers by Tao Xie in
Current site
Google Scholar
PubMed
Close
,
Jiajun Chen aSchool of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, Nanjing, China

Search for other papers by Jiajun Chen in
Current site
Google Scholar
PubMed
Close
, and
Junjie Yan cBeijing Huayun Shinetek Science and Technology Co., Ltd., Beijing, China

Search for other papers by Junjie Yan in
Current site
Google Scholar
PubMed
Close
Open access

Abstract

In this paper, a new objective typhoon positioning algorithm was proposed. The algorithm uses L1 12-channel far-infrared data of the FY-4A geostationary meteorological satellite for objective positioning, verified against best path data provided by the Tropical Cyclone Data Center of the China Meteorological Administration (CMA). By calculating the tangential and radial perturbation values of infrared brightness temperature images, the perturbation factor can be obtained. By adopting the position of the maximum perturbation factor as the center of a circle and considering a radius of no more than 20 km, the position of the minimum perturbation factor was determined; this value represents the central position of the typhoon. Tropical cyclones in 2019 and 2020 were selected for objective positioning, and the objective positioning results were verified against the corresponding time in the best path dataset. The results included centralized verification results for 29 typhoons and optimal path data in 2019. The maximum average error reached 54.67 km, with an annual average typhoon positioning error of 16.15 km. The centralization verification results for 23 typhoons and optimal path data in 2020 indicated a minimum average error of 12.71 km, a maximum average error of 18.56 km, and an annual average typhoon positioning error of 14.82 km. The positioning results for these two years suggest that the longitude obtained with the perturbation factor positioning method is satisfactory, exhibiting an error of less than 20 km.

Significance Statement

The purpose of this study is to help researchers make scientific discoveries and help the development of typhoon center location technology in the future. This is important because accurate positioning of typhoon center can provide effective help for typhoon path prediction and typhoon intensity determination.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Junjie Yan, yanjj@cma.gov.cn

Abstract

In this paper, a new objective typhoon positioning algorithm was proposed. The algorithm uses L1 12-channel far-infrared data of the FY-4A geostationary meteorological satellite for objective positioning, verified against best path data provided by the Tropical Cyclone Data Center of the China Meteorological Administration (CMA). By calculating the tangential and radial perturbation values of infrared brightness temperature images, the perturbation factor can be obtained. By adopting the position of the maximum perturbation factor as the center of a circle and considering a radius of no more than 20 km, the position of the minimum perturbation factor was determined; this value represents the central position of the typhoon. Tropical cyclones in 2019 and 2020 were selected for objective positioning, and the objective positioning results were verified against the corresponding time in the best path dataset. The results included centralized verification results for 29 typhoons and optimal path data in 2019. The maximum average error reached 54.67 km, with an annual average typhoon positioning error of 16.15 km. The centralization verification results for 23 typhoons and optimal path data in 2020 indicated a minimum average error of 12.71 km, a maximum average error of 18.56 km, and an annual average typhoon positioning error of 14.82 km. The positioning results for these two years suggest that the longitude obtained with the perturbation factor positioning method is satisfactory, exhibiting an error of less than 20 km.

Significance Statement

The purpose of this study is to help researchers make scientific discoveries and help the development of typhoon center location technology in the future. This is important because accurate positioning of typhoon center can provide effective help for typhoon path prediction and typhoon intensity determination.

© 2022 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Junjie Yan, yanjj@cma.gov.cn

1. Introduction

Typhoons are highly disastrous weather systems posing serious threats to the safety of human lives and property. It is important to track and locate typhoons quickly and accurately. In addition, during the operational forecasting of typhoons, accurate positioning of the wind center is the basis for typhoon intensity determination and forecasting (Wimmers and Velden 2016). Regarding typhoon center location algorithms, most of the positioning methods used entail manual or automatic location via meteorological satellites or radar systems. With the continuous development of objective typhoon positioning methods, manual typhoon positioning data should be gradually applied to verify automatic typhoon positioning results. Moreover, series of automatic typhoon central positioning systems can be used for automatic or semiautomatic typhoon monitoring and analysis (Johnson et al. 2018). At present, the Dvorak method remains the most widely used method for objective typhoon positioning and strength determination. In this method, cloud models are applied for typhoons in different development processes to classify cloud images according to the typhoon type. Each cloud-type model has a unique central positioning template and yields a specific intensity estimation value (Park et al. 2016). In addition, the image segmentation principle can be used to identify typhoon centers. Sun et al. (2015) used maximum interclass variance and threshold segmentation technology to realize the positioning and tracking of typhoon centers. However, the typhoon systems are variable and complex, making it difficult to locate the typhoon center with a high accuracy based only on morphological characteristics. Liu et al. (1997) used the principle of dynamic image analysis and a cloud-induced wind vector to locate the typhoon center. A tropical cyclone center exhibits the characteristics of a zero-rotation vector; hence, the center is difficult to accurately locate. Wei et al. (2011) proposed the spiral belt model to locate the typhoon center. The parameters of this model were optimized with the particle swarm optimization algorithm. Considering the influence of clouds and rain belts around the wind eye on the positioning process, the experimental results for infrared satellite cloud images indicated that the best average error of the model in the longitude and latitude directions reached 0.1223° (approximately 13.58 km). Zhao et al. (2019) used the centroid algorithm from two-dimensional Fourier filtering to detect potential cyclone vortices based on eliminating small-scale perturbations and finally calculated the vortex center. The advantage of this algorithm is that the parameters are automatically adjusted according to the size of the target vortex. The use of SAR images is also an effective method for locating the typhoon center based on the following two characteristics of a clear tropical cyclone (TC) eye in SAR images: 1) the eye color is dark, and 2) the edge wind speed is discontinuous (Xu et al. 2016). First, the normalized radar cross section (NRCS) is used to approximately determine the dark area of the typhoon eye (TCE0). The geometric center of all pixels in TCE0 can be described as the initial TC center position (CP0). In the polar coordinate system with CP0 as the center, the radial radiation gradient of each pixel in the SAR image can be calculated. In each radial direction, the region surrounded by the pixel with the highest gradient is the new region of the TC eye (TCE1). The geometric center of all pixels in TCE1 is the new center position. Step 3 is repeated until the convergence condition is satisfied (the distance between CPi and CPi−1 is less than the predefined threshold), and the final eye and center positions are determined (Xu et al. 2016). Although SAR images provide the advantage of a high spatial resolution for objective typhoon positioning, typhoons are associated with fast-changing conditions, and the temporal resolution of SAR sensors is not sufficient for locating typhoons in quasi–real time. However, with the development of space sensors, the notable advantages of infrared brightness temperature data in regard to their temporal and spatial resolutions have become apparent. Jaiswal and Kishtawal (2011) used infrared images of a stationary satellite to extract spiral features of the center of a tropical cyclone (TCS), automatically determined the TCS center, estimated the center of the TC by fitting the helixes at different positions, and matched the fitting coefficient with a spiral template value that was higher than the set value. Given a set threshold value, the center of the typhoon could be determined. Subsequently, they calculated the flux of the brightness temperature gradient vector and drew lines parallel to the gradient vector at each pixel in the analysis scene of the image. The points where the lines intersected formed a density matrix, and it was found that the maximum value in the density matrix could be determined as the typhoon center (Jaiswal and Kishtawal 2013). With high spatial and temporal resolutions, infrared brightness temperature cloud images could more quickly provide effective original images for objective typhoon positioning algorithms and enhance the timeliness of objective typhoon positioning (Zheng et al. 2019).

Currently, regardless of whether SAR or infrared brightness temperature data retrieved from geostationary satellites are used to objectively locate typhoons, suitable algorithms have been developed and achieved a high accuracy (Jin et al. 2019; Leonardo and Colle 2020). The longitude and latitude errors of the positioning center can be controlled within 0.5°. Considering the uncertain movement of typhoons, the core process of typhoon center positioning is accurate and fast. The longitude and latitude are important for the subsequent typhoon intensity determination and track prediction operations (Li et al. 2020; Reppucci et al. 2010). The perturbation factor algorithm described in this paper is a new method for locating the center of TCs using Fengyun-4A (FY-4A) stationary satellite infrared images. The gradient of infrared images in the longitude and latitude directions is calculated. Tangential and radial perturbation values can be obtained from surface temperature gradients, and the perturbation factor of the wind field above the sea surface can be calculated. The lowest perturbation factor corresponding to the typhoon center position and the lowest far-infrared brightness temperature perturbation are considered to obtain the typhoon center.

We used this technology to locate typhoons in the northwest Pacific in 2019 and compared them to the typhoons in the China Meteorological Administration (CMA) dataset. The temporal resolution of the obtained FY-4A satellite images was 15 min, and the objective location of a typhoon in a single image could be determined in less than 5 s, which ensured that we could accurately obtain the longitude and latitude of the typhoon center in near–real time.

The perturbation factor algorithm proposed in this paper is a new method for locating the center of a tropical cyclone using FY-4A geostationary satellite infrared imaging data. In this method, the brightness temperature (BT) gradient in the longitude and latitude directions is calculated, and tangential and radial perturbation values are obtained to determine the perturbation factor. In turn, the perturbation factor corresponding to the typhoon center position is the lowest, as is the corresponding BT perturbation. We used this algorithm to objectively locate tropical cyclones in the northwest Pacific in 2019 and 2020, as verified via a comparison to the CMA best path dataset. The algorithm requires less than 10 s to determine the objective location of a typhoon in a single image and can thus locate typhoons in near–real time.

The remainder of this paper is organized as follows. In sections 2 and 3, the study area and data used in this study, and objective perturbation factor–based typhoon eye location algorithm are introduced. In section 4, four typhoons in the northwest Pacific are used to assess the algorithm. In section 5, the location centers in 13 typhoon infrared images are compared to 308 location times corresponding to the optimal path dataset, and finally, section 6 summarizes our main findings.

2. Study area and data

a. Study area

The study area is the northwest Pacific, and the longitude and latitude ranges are 100°–150°E and 0°–45°N, respectively. More tropical cyclones occur in the northwest Pacific than in any other sea area globally, and approximately 80% of these tropical cyclones will develop into typhoons (Cheng et al. 2012). On average, annually, cyclones reaching at least the tropical storm intensity account for approximately 31% of the global total amount, more than twice that in any other region. The northwest Pacific is also an area where typhoons occur every month of the year. Moreover, because a large part of the northwest Pacific is adjacent to East Asian and Southeast Asian countries, typhoons landing in the northwest Pacific can cause damage and casualties in coastal cities in these countries. Figure 1 shows a map of the study area.

Fig. 1.
Fig. 1.

Regional scope in this study.

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

b. Geostationary satellite data

The FY-4A geostationary meteorological satellite started operation on 1 May 2018. It is a second-generation geostationary meteorological satellite that replaces the FY-2 satellite. Its continuous and stable operation could greatly improve the geostationary orbit meteorological detection level of satellites. Compared to the FY-2 satellite, which had 5 radiation imagining channels, the FY-4A satellite increased has 14, covering the visible light, shortwave infrared, mediumwave infrared, and longwave infrared bands. The specific parameters are presented in Table 1. The L1 full disk data obtained using the FY-4A ARGI imager exhibit a spatial resolution of 4 km and a temporal resolution of 15 min. The detection band of the multichannel radiation imager ranges from 0.45 to 13.8 μm, the radiation sensitivity of the active detection bands from 10.3 to 11.3 μm reaches 0.06 K, and the absolute calibration accuracy reaches 0.3 K (Wang and Xia 2018). In this paper, BT data from the channel-12 near- and far-infrared bands (wavelength: 10.3–11.3 μm) were selected as the original typhoon positioning data.

Table 1

Comparison of the FY-4A/AGRI band settings and main detection objects with those of FY-2.

Table 1

c. Typhoon short-term and impending forecast data from the China Meteorological Administration

Typhoon short-term and impending prediction data provided by the Typhoon and Marine Meteorological Prediction Center of the China Meteorological Administration were used to determine the initial prediction centers of the typhoons and obtain the approximate typhoon ranges. Typhoon forecast data (babj*.dat) are updated once every 1 or 3 h, where babj is the typhoon message originating from the Central Meteorological Observatory. When a typhoon is generated, babj*.dat (* indicates the typhoon number) forecast data are reported; when the typhoon process ends, so does the reporting of forecast data. In Table 2, 20 February 2019, is selected as an example to introduce the content and format of babj data storage, and babj typhoon forecast data observed at 0200, 0500, 0800, 1400, 1700, and 2000 UTC on each day are selected as starting values to forecast the longitude, latitude, air pressure, and wind speed of the typhoon center in a backward manner at each time point.

Table 2

Typhoon forecast data from the Center for Typhoon and Ocean Weather Forecasts of the CMA and babj_1902.dat data recorded on 20 Feb 2019.

Table 2

d. Best path dataset (CMABSTDATA)

The best path dataset was retrieved from the official website of the Tropical Cyclone Data Center of the China Meteorological Administration. The current version of the best tropical cyclone path dataset of the CMA provides the location and intensity of tropical cyclones in the northwest Pacific (including the South China Sea, north of the equator and west of 180° longitude) at 6-h intervals since 1949. The data are stored in separate text files by year and can be amended on a yearly basis (Ying et al. 2014). The best path dataset before 1972 was supplemented with reanalysis data after 1972, and after 1972, historical atlases, station observations and ship weather reports, automatic surface observations, weather maps, radiosonde data, and aircraft reconnaissance surveys were used for data integration. Later, satellite and coastal radar observation data were added (Lu et al. 2021). Table 3 lists the data types and storage formats for the CMA data recorded on 20 February 2019. The typhoon best path dataset contains the longitude and latitude of each typhoon center, maximum average wind speed, minimum air pressure at the center, and maximum wind speed at the center. This dataset was used to verify the objective typhoon positioning results.

Table 3

Best path dataset recorded on 20 Feb 2019.

Table 3

3. Methods

The perturbation algorithm can be divided into three steps: 1) calculating the BT gradient in the longitude and latitude directions, 2) using the BT gradient to calculate tangential and radial perturbation values, and 3) obtaining the corresponding perturbation value according to the determined radial and tangential perturbation values. First, according to the typhoon center position based on the short-term and impending typhoon forecast data of the CMA, the polynomial interpolation method was used to predict the longitude and latitude of the initial typhoon center at the time point of the current satellite-acquired image. With this point as the center, a window was constructed to frame the entire typhoon area, and the BT gradient was calculated within this range. The BT gradient in the longitude direction was calculated as follows:
GBT=BTxx+BTyy,
where GBT is the brightness temperature gradient.
Then, tangential and radial perturbation values were obtained. The radial perturbation value represents the rotation degree of microelements near a given point due to the vector field. The radial perturbation value was determined as
Radial_GBT=GBTxcosθ+GBTysinθ,
where Radial_GBT is the radial perturbation value in N m−3 × 107. When the tangential perturbation value is negative, convergence occurs, which is conducive to the development and enhancement of convective weather systems such as cyclones (Minobe et al. 2008; Putrasahan et al. 2013).
The tangential perturbation value was calculated as
Tangential_GBT=GBTxsinθ+GBTycosθ,
where Tangential_GBT is the tangential perturbation value in N m−3 × 107. In the atmosphere, vorticity is the result of air rotations. In the Northern Hemisphere, the counterclockwise direction indicates positive vorticity, which is conducive to the development of weather systems such as anticyclones (Maloney and Chelton 2006).
Considering the joint action of the radial and tangential perturbations in the typhoon eye area, the calculated radial and tangential perturbations include positive and negative values. If only the mean value of these two quantities is considered, the offset could reach zero, resulting in discrimination errors. Therefore, we first determined the square of the tangential and radial perturbation values and finally calculated the BT perturbation parameter (Xie et al. 2010) as follows:
P=Radial_GBT2+Tangential_GBT2,
where P is the perturbation factor, which is calculated in the area after determining the approximate typhoon range. The maximum perturbation factor position is considered the center of a circular area, and the value of the perturbation factor is again calculated in this area within a radius of 20 km. The minimum perturbation factor position is determined as the center of the typhoon.

4. Results

In this paper, 29 typhoons in 2019 and 23 typhoons in 2020 in the northwest Pacific were objectively located according to the proposed perturbation factor algorithm.

First, with Typhoon Halong as an example, the BT perturbation value in the center area of the typhoon was calculated using BT data from the FY-4A 12 channel, and the spatial distribution of the perturbation values in the typhoon center area was analyzed. Typhoon Halong, the strongest typhoon in 2019, remained over the sea throughout its entire life cycle, and the effects of land and other factors on the typhoon could thus be excluded to best reflect the BT perturbation characteristics of the typhoon area and reveal the BT perturbation characteristics of typhoons with and without eye areas based on the entire life cycle of Typhoon Halong.

The typhoon center at the corresponding time in the best path dataset was superimposed onto the obtained distribution map of the perturbation values, and the spatial distribution of the perturbation values was analyzed in steps of 8 km to generate a spatial distribution box diagram of the perturbation values in the center area of Typhoon Halong at 1800 UTC 5 November 2019. The best typhoon center is shown in Fig. 2. The first minimum perturbation value in the typhoon central area was 8 km, while the second minimum perturbation value reached 64 km. We calculated the spatial distribution of the perturbation values in the central area of Typhoon Halong throughout the entire life cycle with respect to the typhoon center in the best path dataset and determined the locations of the first and second minimum perturbation values, as summarized in Table 4. The first minimum perturbation value ranged from 8 to 40 km, while the second minimum perturbation value ranged from 40 to 104 km. Therefore, the perturbation value was set within a range of 40 km to prevent the minimum perturbation value outside the wind eye area from impacting the positioning accuracy.

Fig. 2.
Fig. 2.

Distribution characteristics of Typhoon Halong perturbation values in the center area based on the best path dataset at 1800 UTC 5 Nov 2019.

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

Table 4

Pixel difference between the first minimum perturbation value, second minimum perturbation value, and center in the best path dataset throughout the life cycle of Typhoon Halong.

Table 4

The brightness temperature perturbation algorithm was applied in typhoon positioning to identify a typhoon area with slight temperature fluctuations at the minimum pixel position within an 8-km range; in each area, a minimum perturbation value occurred within an average typhoon eye radius of 20 km (Wang et al. 2014). Values with minimum perturbations beyond 40 km from the approximate typhoon center were ignored; therefore, determining the appropriate perturbation value range could greatly influence the positioning accuracy of forecasts.

The position of the minimum perturbation value with the largest range was assessed at 1800 UTC 5 November 2019. As shown in Fig. 3, the pixel position of the minimum perturbation value (yellow circle) occurred close to the best path typhoon center (red cross), while the center position of the first prediction occurred far from the best path typhoon center (red cross). Regarding typhoon waves in the summer within the life cycle of an event, minimum perturbation value pixels and the center distance were calculated from the best path of the typhoon center, as listed in Table 5. The minimum perturbation value pixel location along the best path of the typhoon center was generally close to the true value, with a minimum value of 5.27 km and an overall mean of 15.41 km. The location of the first prediction center occurred far from the best path typhoon center, with a minimum value of 6.65 km and an average value of 20.31 km. Thus, the use of the minimum perturbation value provides an accuracy of improvement of approximately 5 km over that when the first prediction center is used.

Fig. 3.
Fig. 3.

Best path data–based center, maximum perturbation value pixel position, first minimum perturbation value pixel position and second minimum perturbation value pixel position of Typhoon Halong at 1800 UTC 5 Nov 2019.

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

Table 5

Distance between the minimum perturbation value, first prediction center and BST center throughout the life cycle of Typhoon Halong.

Table 5

Next, four strong typhoons and super typhoons, namely, Wutip, Lekima, Hagibis, and Halong, were selected as cases for further analysis.

a. Typhoon Wutip

First, Typhoon Wutip was studied in February 2019. One specific time was selected as an example to employ the location algorithm at each stage, namely, the initial formation period, strong period, and attenuation period. Typhoon Wutip entered the typhoon stage at 1400 UTC 21 February 2019. Previously, the system occurred at the tropical storm level and then gradually weakened. Finally, the central meteorological station (CMS) stopped tracking this system at 1700 UTC 28 February, after a period of 7 days. The initial stage of typhoon formation lasted 21–23 days. On 21 February 2019, the system was upgraded to a typhoon by the CMS. The maximum wind speed of 58 m s−1 was reached at 2000 UTC 23 February. The strong period ranged from 24 to 25 days, and the system was upgraded to a super typhoon by the CMS at 0800 UTC 25 February. Then, the system was upgraded to level 5 by the Joint Typhoon Warning Center (JTWC) on 25 February. The attenuation period ranged from 26 to 27 days, and the maximum wind speed gradually decreased. Figure 4 shows FY-4A 12-channel infrared typhoon images from the early and later stages of the event, as determined by the typhoon center (red dot) based on the perturbation factor and the typhoon center (blue dot) at the corresponding time in the best path dataset at 1200 UTC 24 February, 1200 UTC 25 February, and 1200 UTC 26 February. Figure 5a shows the path of the Typhoon Wutip center (blue square) recorded in the best path dataset at 0000, 0600, 1200, and 1800 UTC from 21 to 28 February 2019. The figure also shows the objective positioning center (red dot) obtained with the perturbation factor algorithm at the corresponding time based on CMA data. The background typhoon time was 1200 UTC 25 February. Figure 5b shows an enlarged comparison of local paths at 0000, 0600, 1200, and 1800 UTC from 23 to 25 February.

Fig. 4.
Fig. 4.

Objective typhoon positioning result at 1200 UTC (a) 24 Feb (b) 25 Feb, and (c) 26 Feb 2019 (red cross) compared to the CMA track dataset (blue × symbol).

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

Fig. 5.
Fig. 5.

(a) Track of Typhoon Wutip based on the objective positioning results from 21 to 28 Feb 2019 (red dashed line) and the CMA positioning path (blue solid line). The background image time is at 1200 UTC 25 Feb 2019. (b) Comparison of the objective positioning path (red dashed line) and CMA positioning path (blue solid line) at 0000, 0600, 1200, and 1800 from 23 to 25 Feb 2019.

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

b. Typhoon Lekima

Typhoon Lekima No. 9 in 2019 was named by the Japan Meteorological Agency (JMA) on 4 August 2019. On 7 August, the system was upgraded to a typhoon by the CMS. On 7 August, the system was further upgraded to a super typhoon by the CMS. This typhoon continued to move northwest and close to the coastal area of Zhejiang Province. On 10 August, the system made landfall in the coastal area of Zhejiang Province. The typhoon then crossed Zhejiang and Jiangsu Provinces, moved into the Yellow Sea, and again made landfall along the coast of Shandong Province on 11 August. The maximum wind force near the center reached level 9 (23 m s−1) before the system moved into the Bohai Sea and weakened. Finally, tracking was stopped by the CMS on 13 August.

Figure 6a shows the track (red dot) and CMA positioning path (blue square) of Typhoon Lekima at 0000, 0300, 0600, 0900, 1200, 1500, 1800, and 2100 UTC from 7 to 10 August 2019; the background typhoon time was 1200 UTC 9 August. Figure 6b shows an enlarged comparison between the objective and CMA positioning paths of Typhoon Lekima at 0000, 0300, 0600, 0900, 1200, 1500, 1800, and 2100 UTC from 7 to 9 August.

Fig. 6.
Fig. 6.

(a) Track of Typhoon Lekima based on the objective positioning results from 7 to 10 Aug 2019 (red dashed line) and CMA positioning path (blue solid line). The background image time is at 1200 UTC 8 Aug 2019; and (b) comparison of the objective positioning path (red dashed line) and CMA positioning path (blue solid line) at 0000, 0300, 0600, 0900, 1200, 1500, 1800, and 2100 UTC from 9 to 10 Aug 2019.

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

c. Typhoon Hagibis

Typhoon Hagibis was named by the JMA on 6 October 2019, and the system experienced a process of rapid strengthening. On 7 October, the system was upgraded to a typhoon by the CMS and then upgraded to a super typhoon by the CMS on 7 October. The typhoon continued to strengthen, becoming the strongest typhoon since 2019, as recognized by the CMS. Subsequently, the system passed through the Northern Mariana Islands and turned northward on 12 October, making landfall on the coast of Japan. The maximum wind near the center reached level 14 (42 m s−1), and the CMS finally stopped tracking this typhoon on 13 October.

Figure 7a shows the objective positioning results (red dot) and CMA positioning path (blue square) of Typhoon Hagibis at 0000, 0600, 1200, and 1800 UTC from 7 to 12 October 2019. The background typhoon time was 1200 UTC 9 October. Figure 7b shows an enlarged comparison of the objective and CMA positioning paths of Typhoon Hagibis at 0000, 0600, 1200, and 1800 UTC from 9 to 10 October.

Fig. 7.
Fig. 7.

(a) Track of Typhoon Hagibis based on the objective positioning results at 0000, 0600, 1200, and 1800 UTC from 7 to 12 Oct 2019 (red dashed line) and the CMA positioning path (blue solid line). The background image time is 1200 UTC 7 Oct 2019; and (b) comparison of the objective positioning path (red dashed line) and CMA positioning path (blue solid line) at 0000, 0600, 1200, and 1800 UTC from 9 to 10 Oct 2019.

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

d. Typhoon Halong

Typhoon Halong was named by the JMA on 3 November 2019. On 4 November, the Central Meteorological Observatory (CMO) upgraded this system to a typhoon. On 5 November, the CMS upgraded it to a super typhoon. On 7 November, the CMO then downgraded this system to a strong typhoon. On 8 November, the CMO confirmed that it had become an extratropical cyclone in the southeast ocean of Japan and stopped tracking this system. Figure 8a shows the objective positioning results (red dot) and CMA positioning path (blue square) of Typhoon Halong at 0000, 0600, 1200, and 1800 UTC from 3 to 8 November 2019. The background typhoon time was 1200 UTC 6 November. Figure 8b shows an enlarged comparison between the objective and CMA positioning paths of Typhoon Halong at 0000, 0600, 1200, and 1800 from 6 to 7 November.

Fig. 8.
Fig. 8.

(a) Track of Typhoon Halong based on the objective positioning results at 0000, 0600, 1200, and 1800 UTC from 3 to 8 Nov 2019 (red dashed line) and CMA positioning path (blue solid line); the background image time is at 1200 UTC 6 Nov 2019; and (b) comparison of the objective positioning path (red dashed line) and CMA positioning path (blue solid line) at 0000, 0600, 1200, and 1800 UTC from 6 to 7 Nov 2019.

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

The positioning results revealed that the perturbation factor could be used to objectively locate the typhoon, and the positioning center coincided with the CMA typhoon center, indicating that the typhoon could be objectively and accurately located.

5. Verification and discussion

In this paper, the CMA best path dataset provided by the Tropical Cyclone Data Center of the CMA was compared to the latitude and longitude of the typhoon center obtained via the objective perturbation factor positioning algorithm. Table 6 provides the longitude and latitude of the objective positioning center obtained with the perturbation factor algorithm, the best path center throughout the life cycle of Typhoon Wutip and the error between these paths; the calculation error is based on the spherical distance between paths. As indicated in Table 6, the positioning longitude and latitude errors remained within 0.2°. The objective typhoon positioning center suitably matched that in the best path dataset of the CMA, with an average positioning error of less than three pixels. Thus, typhoon positioning was accurately achieved. The root-mean-square error (RMSE) was 6.16 km, with an average value of 4.36 km.

Table 6

Comparison of the longitude and latitude of Typhoon Wutip in 1902 and the values at the corresponding time in the CMA best path dataset.

Table 6

Four cases were selected to illustrate the relationship between the p value and the typhoon location center. For Typhoon Wutip at 0600 UTC 25 February as an example, we observed from the p value and the original satellite BT image that the maximum p value was mainly concentrated in the wind eye area and at the edge of the cloud area. In addition, the minimum p value was observed in the wind eye area and in the cloud area outside the wind eye region. In theory, the minimum p value obtained should be the minimum p value in the eye area. Figure 9a shows a 21 × 21 pixel window of the BT in the eye region. In this study, the average p value of the 21 × 21 window, as shown in Fig. 9b, was used to screen the entire window, and pixels with values larger than the average value for the whole window were selected, as shown in Fig. 9c. The minimum area box depicted in Fig. 9c, as highlighted in Fig. 9d, still contained the minimum p value outside the eye area, leading to a large positioning error. Therefore, we averaged the values contained in the window that were screened in Fig. 9c and eliminated pixels with values exceeding the average value, as shown in Fig. 9e. Consequently, the minimum area selected by the box contained the most suitable eye area. As shown in Fig. 9f, the pixel with the minimum p value in the eye area was the best eye center. Similarly, for Typhoon Lekima, Fig. 10a shows a BT map of the wind eye region within a 21 × 21 window, and the average p value of the 21 × 21 window in Fig. 10b was used to screen the entire window. In the figure, pixels with values larger than the average value of the entire window were selected, as shown in Fig. 10c, and the minimum area box contained in Fig. 10c was used, as shown in Fig. 10d. Figure 10d shows the minimum p value outside the eye area, which could still result in a large positioning error. Therefore, we averaged the values contained in the window that were screened in Fig. 10c and eliminated pixels with values greater than the average value, as shown in Fig. 10e. Consequently, the minimum area selected in the box contained the most suitable wind eye area, as shown in Fig. 10f. This figure shows that the typhoon center with the minimum p value was the best typhoon center for objective positioning. The results for Typhoons Hagibis and Halong are shown in Figs. 11 and 12, respectively. First, the average value was used to screen the p value within the 21 × 21 range, and the average value was then used to screen the values close to the typhoon eye area. Based on two screening frames, the pixel with the smallest p value in the eye area was determined as the best objective positioning center.

Fig. 9.
Fig. 9.

(a) Brightness temperature in the eye area of Typhoon Wutip at 0600 UTC 25 Feb, selected with a 21 × 21 window; (b) calculated p value in the typhoon eye area selected with a 21 × 21 window; (c) distribution of pixels with values larger than the overall average value of the window in (b); (d) minimum pixel selection range for the pixels in (c); the distribution of pixels with values larger than the average value of the window in (c) represents the pixel distribution of the overall average value of the window; and (f) minimum pixel frame selection range for the pixels in (e).

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

Fig. 10.
Fig. 10.

(a) Brightness temperature in the eye area of Typhoon Lekima at 2000 UTC 8 Aug, selected with a 21 × 21 window; (b) calculated p value in the typhoon eye area selected with a 21 × 21 window; (c) distribution of pixels with values larger than the overall average value of the window in (b); (d) minimum pixel selection range of the pixels in (c); (e) the distribution of pixels with values larger than the average value of the window in (c) represents the pixel distribution of the overall average value of the window; and (f) minimum pixel frame selection range of the pixels in (e).

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

Fig. 11.
Fig. 11.

(a) Brightness temperature in the eye area of Typhoon Hagibis at 2000 UTC 7 Oct, selected with a 21 × 21 window; (b) calculated p value in the typhoon eye area selected with a 21 × 21 window; (c) distribution of pixels with values larger than the overall average value of the window in (b); (d) minimum pixel selection range of the pixels in (c); (e) the distribution of pixels with values larger than the average value of the window in (c) represents the pixel distribution of the overall average value of the window; and (f) minimum pixel frame selection range of the pixels in (e).

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

Fig. 12.
Fig. 12.

(a) Brightness temperature in the eye area of Typhoon Halong at 0200 UTC 6 Nov, selected with a 21 × 21 window; (b) calculated p value in the typhoon eye area selected with a 21 × 21 window; (c) distribution of pixels with values larger than the overall average value of the window in (b); (d) minimum pixel selection range in (c); (e) the distribution of pixels with values larger than the average value of the window in of the pixels (c) represents the pixel distribution of the overall average value of the window; and (f) minimum pixel frame selection range of the pixels in (e).

Citation: Journal of Atmospheric and Oceanic Technology 39, 12; 10.1175/JTECH-D-22-0016.1

Finally, the perturbation factor algorithm was used to locate 29 typhoons in 2019 and 23 in 2020 in the northwest Pacific Ocean; the results are summarized in Tables 7 and 8, respectively. Table 7 provides the centralized verification results for 29 typhoons and optimal path data for 2019. The maximum average error was 54.67 km, and the annual average typhoon positioning error was 16.15 km. Table 8 provides the centralization verification results for 23 typhoons and optimal path data in 2020. The minimum average error was 12.71 km, the maximum average error was 18.56 km, and the annual average typhoon positioning error was 14.82 km. The positioning results for these two years suggest that the longitude determined with the perturbation factor positioning method is satisfactory, at less than 20 km.

Table 7

Error between the typhoon positioning results and best path dataset in 2019.

Table 7
Table 8

Error between the typhoon positioning results and best path dataset in 2020.

Table 8

6. Conclusions

In this paper, a new objective typhoon positioning algorithm was proposed. With the use of the perturbation factor positioning algorithm and BT data from the FY-4A satellite, the tangential and radial perturbation values of the BT were calculated, and the perturbation factor was obtained. The position of the minimum perturbation factor was determined within a radius of no more than 20 km with the maximum perturbation factor as the center of the circle. The following conclusions were obtained in this study.

  1. 1) The advantage of the perturbation factor positioning technique is that it could objectively locate a typhoon in satellite remote sensing images at any time. While obtaining typhoon prediction data, the algorithm searches available satellite remote sensing data to achieve quasi-real-time positioning, which can provide a suitable reference for typhoon early warning systems. The perturbation factor–based objective typhoon positioning algorithm exhibited a short operation time, a single positioning time less than 10 s, and a high positioning accuracy. Moreover, the overall error was controlled within two to three pixels, providing accurate information for typhoon monitoring and prediction.
  2. 2) When the structure was unclear at the initial stage of typhoon formation and during the attenuation period, on the premise of accurately obtaining the typhoon range, the perturbation algorithm could still effectively locate the typhoon center, which resolved the inaccurate positioning issue due to an unclear eye structure at the initial and final stages of the typhoon.
  3. 3) The disadvantage was that when determining the first prediction center of a typhoon, the fitting function used in this paper could only roughly determine the typhoon center, but typhoon movement is affected by complex factors such as the sea surface temperature and internal motion of the cyclone. The location of the fitted typhoon center often differed from that of the actual typhoon center. If the objective location of the typhoon was determined based on an inaccurate initial prediction center, the resultant error would be large. Therefore, identifying the initial prediction center is very important for the perturbation factor algorithm.
  4. 4) FY-4A infrared BT data may be missing, resulting in the inability to obtain typhoon centers at a certain time. For subsequent algorithm improvement, multisource satellite data fusion must be performed. When FY-4A data are missing, other satellite data, such as FY-4B/FY-2, could be used as a supplement during positioning.

Acknowledgments.

This work was partially supported by the National Key R&D Program of China (2021YFC2803302), the National Natural Science Foundation of China Project (42176180), the Natural Science Foundation of Jiangsu Province (Grant YJGL-YF-2020-16), the Natural Science Foundation of Jiangsu Province (Grant JSZRHYKJ202114), and the Jiangsu Provincial Innovation Foundation for Postgraduates (KYCX21_0976).

Data availability statement.

FY-4A geostationary meteorological satellite data are publicly available in the Fengyun service network (http://satellite.nsmc.org.cn/portalsite/default.aspx#). The best path dataset can be retrieved from the official website of the Tropical Cyclone Data Center of the China Meteorological Administration and is publicly available in the service network (http://g.hyyb.org/systems/TY/info/tcdataCMA/wxfxzl_zlhq.html).

REFERENCES

  • Cheng, Y.-H., S. Huang, A. K. Liu, C. Ho, and N. Kuo, 2012: Observation of typhoon eyes on the sea surface using multi-sensors. J. Remote Sen. Environ., 123, 435441, https://doi.org/10.1016/j.rse.2012.04.009.

    • Search Google Scholar
    • Export Citation
  • Jaiswal, N., and C. M. Kishtawal, 2011: Automatic determination of center of tropical cyclone in satellite-generated IR images. IEEE Geosci. Remote Sens. Lett., 8, 460463, https://doi.org/10.1109/LGRS.2010.2085418.

    • Search Google Scholar
    • Export Citation
  • Jaiswal, N., and C. M. Kishtawal, 2013: Objective detection of center of tropical cyclone in remotely sensed infrared images. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 6, 10311035, https://doi.org/10.1109/JSTARS.2012.2215016.

    • Search Google Scholar
    • Export Citation
  • Jin, X. L., X. Yang, J. A. Zhang, and D. Shen, 2019: Identification of tropical cyclone centers in SAR imagery based on template matching and particle swarm optimization algorithms. IEEE Trans. Geosci. Remote Sens., 57, 598608, https://doi.org/10.1109/TGRS.2018.2863259.

    • Search Google Scholar
    • Export Citation
  • Johnson, B., S. Thomas, and J. S. Rani, 2018: A novel framework for objective detection and tracking of TC center from noisy satellite imagery. Adv. Space Res., 62, 4454, https://doi.org/10.1016/j.asr.2018.04.017.

    • Search Google Scholar
    • Export Citation
  • Leonardo, N. M., and B. A. Colle, 2020: An investigation of large along-track errors in extratropical transitioning North Atlantic tropical cyclones in the ECMWF ensemble. Mon. Wea. Rev., 148, 457476, https://doi.org/10.1175/MWR-D-19-0044.1.

    • Search Google Scholar
    • Export Citation
  • Li, Z. C., L. Zhang, and Q. F. Qian, 2020: The development and consideration of typhoon forecast operation of Central Meteorological Center (in Chinese). J. Trans. Atmos. Sci., 43, 1019.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., K. Lin, and A. Guo, 1997: Morphological feature extraction of satellite cloud images. J. Comp. Res. Dev., 34, 689693.

  • Lu, X., H. Yu, M. Ying, B. K. Zhao, S. Zhang, L. M. Lin, L. N. Bai, and R. J. Wan, 2021: Western North Pacific tropical cyclone database created by the China Meteorological Administration. Adv. Atmos. Sci., 38, 690699, https://doi.org/10.1007/s00376-020-0211-7.

    • Search Google Scholar
    • Export Citation
  • Maloney, E. D., and D. B. Chelton, 2006: An assessment of the sea surface temperature influence on surface in numerical weather prediction and climate models. J. Climate, 19, 27432762, https://doi.org/10.1175/JCLI3728.1.

    • Search Google Scholar
    • Export Citation
  • Minobe, S., A. Kuwano-Yoshida, N. Komori, S. Xie, and R. Small, 2008: Influence of the Gulf Stream on the troposphere. J. Nature, 452, 206209, https://doi.org/10.1038/nature06690.

    • Search Google Scholar
    • Export Citation
  • Park, M., M. Kim, M. Lee, J. Im, and S. Park, 2016: Detection of tropical cyclone genesis via quantitative satellite ocean surface wind pattern and intensity analyses using decision trees. J. Remote Sens. Environ., 183, 205214, https://doi.org/10.1016/j.rse.2016.06.006.

    • Search Google Scholar
    • Export Citation
  • Putrasahan, D. A., A. J. Miller, and H. Seo, 2013: Isolating mesoscale coupled ocean–atmosphere interactions in the Kuroshio Extension region. Dyn. Atmos. Oceans, 63, 6078, https://doi.org/10.1016/j.dynatmoce.2013.04.001.

    • Search Google Scholar
    • Export Citation
  • Reppucci, A., S. Lehner, J. Schulz-Stellenfleth, and S. Brusch, 2010: Tropical cyclone intensity estimated from wide-swath SAR images. J. IEEE Trans. Geosci. Remote Sens., 48, 16391649, https://doi.org/10.1109/TGRS.2009.2037143.

    • Search Google Scholar
    • Export Citation
  • Sun, Q., C. Wang, and Y. Wang, 2015: Research on typhoon identification and location method. 32nd Annual Meeting of China Meteorological Society, Tianjin, China, Chinese Meteorological Society, 128–133.

  • Wang, G., and S. Xia, 2018: Application of FY-4 radiation imager and its data in satellite meteorology. J. Nature, 40, 111.

  • Wang, Y., Z. Ding, and X. Li, 2014: Mechanism and evolution analysis of the structure and track of the Tropical Cyclone Morakot before and after landing on Taiwan Island (in Chinese). J. Nat. Disasters, 23, 4757, https://doi.org/10.13577/j.jnd.2014.0606.

    • Search Google Scholar
    • Export Citation
  • Wei, K., Z. Jing, Y. Li, and S. Liu, 2011: Spiral band model for locating tropical cyclone centers. J. Pattern Recognit. Lett., 32, 761770, https://doi.org/10.1016/j.patrec.2010.12.011.

    • Search Google Scholar
    • Export Citation
  • Wimmers, A. J., and C. S. Velden, 2016: Advancements in objective multisatellite tropical cyclone center fixing. J. Appl. Meteor. Climatol., 55, 197212, https://doi.org/10.1175/JAMC-D-15-0098.1.

    • Search Google Scholar
    • Export Citation
  • Xie, T., W. Perrie, and W. Chen, 2010: Gulf Stream thermal fronts detected by synthetic aperture radar. Geophys. Res. Lett., 37, L06601, https://doi.org/10.1029/2009GL041972.

    • Search Google Scholar
    • Export Citation
  • Xu, Q., G. Zhang, X. Li, and Y. Cheng, 2016: An automatic method for tropical cyclone center determination from SAR. 2016 IEEE Int. Geoscience and Remote Sensing Symp., Beijing, China, IEEE, 2250–2252, https://doi.org/10.1109/igarss.2016.7729581.

  • Ying, M., W. Zhang, H. Yu, X. Lu, J. Feng, Y. Fan, Y. Zhu, and D. Chen, 2014: An overview of the China Meteorological Administration tropical cyclone database. J. Atmos. Oceanic Technol., 31, 287301, https://doi.org/10.1175/JTECH-D-12-00119.1.

    • Search Google Scholar
    • Export Citation
  • Zhao, C., J. Song, H. Leng, and J. Zhao, 2019: Objective center-finding algorithm for tropical cyclones in numerical models. Atmosphere, 10, 376, https://doi.org/10.3390/atmos10070376.

    • Search Google Scholar
    • Export Citation
  • Zheng, G., J. Liu, J. Yang, and X. Li, 2019: Automatically locate tropical cyclone centers using top cloud motion data derived from geostationary satellite images. IEEE Trans. Geosci. Remote Sens., 57, 10 17510 190, https://doi.org/10.1109/TGRS.2019.2931795.

    • Search Google Scholar
    • Export Citation
Save
  • Cheng, Y.-H., S. Huang, A. K. Liu, C. Ho, and N. Kuo, 2012: Observation of typhoon eyes on the sea surface using multi-sensors. J. Remote Sen. Environ., 123, 435441, https://doi.org/10.1016/j.rse.2012.04.009.

    • Search Google Scholar
    • Export Citation
  • Jaiswal, N., and C. M. Kishtawal, 2011: Automatic determination of center of tropical cyclone in satellite-generated IR images. IEEE Geosci. Remote Sens. Lett., 8, 460463, https://doi.org/10.1109/LGRS.2010.2085418.

    • Search Google Scholar
    • Export Citation
  • Jaiswal, N., and C. M. Kishtawal, 2013: Objective detection of center of tropical cyclone in remotely sensed infrared images. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 6, 10311035, https://doi.org/10.1109/JSTARS.2012.2215016.

    • Search Google Scholar
    • Export Citation
  • Jin, X. L., X. Yang, J. A. Zhang, and D. Shen, 2019: Identification of tropical cyclone centers in SAR imagery based on template matching and particle swarm optimization algorithms. IEEE Trans. Geosci. Remote Sens., 57, 598608, https://doi.org/10.1109/TGRS.2018.2863259.

    • Search Google Scholar
    • Export Citation
  • Johnson, B., S. Thomas, and J. S. Rani, 2018: A novel framework for objective detection and tracking of TC center from noisy satellite imagery. Adv. Space Res., 62, 4454, https://doi.org/10.1016/j.asr.2018.04.017.

    • Search Google Scholar
    • Export Citation
  • Leonardo, N. M., and B. A. Colle, 2020: An investigation of large along-track errors in extratropical transitioning North Atlantic tropical cyclones in the ECMWF ensemble. Mon. Wea. Rev., 148, 457476, https://doi.org/10.1175/MWR-D-19-0044.1.

    • Search Google Scholar
    • Export Citation
  • Li, Z. C., L. Zhang, and Q. F. Qian, 2020: The development and consideration of typhoon forecast operation of Central Meteorological Center (in Chinese). J. Trans. Atmos. Sci., 43, 1019.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., K. Lin, and A. Guo, 1997: Morphological feature extraction of satellite cloud images. J. Comp. Res. Dev., 34, 689693.

  • Lu, X., H. Yu, M. Ying, B. K. Zhao, S. Zhang, L. M. Lin, L. N. Bai, and R. J. Wan, 2021: Western North Pacific tropical cyclone database created by the China Meteorological Administration. Adv. Atmos. Sci., 38, 690699, https://doi.org/10.1007/s00376-020-0211-7.

    • Search Google Scholar
    • Export Citation
  • Maloney, E. D., and D. B. Chelton, 2006: An assessment of the sea surface temperature influence on surface in numerical weather prediction and climate models. J. Climate, 19, 27432762, https://doi.org/10.1175/JCLI3728.1.

    • Search Google Scholar
    • Export Citation
  • Minobe, S., A. Kuwano-Yoshida, N. Komori, S. Xie, and R. Small, 2008: Influence of the Gulf Stream on the troposphere. J. Nature, 452, 206209, https://doi.org/10.1038/nature06690.

    • Search Google Scholar
    • Export Citation
  • Park, M., M. Kim, M. Lee, J. Im, and S. Park, 2016: Detection of tropical cyclone genesis via quantitative satellite ocean surface wind pattern and intensity analyses using decision trees. J. Remote Sens. Environ., 183, 205214, https://doi.org/10.1016/j.rse.2016.06.006.

    • Search Google Scholar
    • Export Citation
  • Putrasahan, D. A., A. J. Miller, and H. Seo, 2013: Isolating mesoscale coupled ocean–atmosphere interactions in the Kuroshio Extension region. Dyn. Atmos. Oceans, 63, 6078, https://doi.org/10.1016/j.dynatmoce.2013.04.001.

    • Search Google Scholar
    • Export Citation
  • Reppucci, A., S. Lehner, J. Schulz-Stellenfleth, and S. Brusch, 2010: Tropical cyclone intensity estimated from wide-swath SAR images. J. IEEE Trans. Geosci. Remote Sens., 48, 16391649, https://doi.org/10.1109/TGRS.2009.2037143.

    • Search Google Scholar
    • Export Citation
  • Sun, Q., C. Wang, and Y. Wang, 2015: Research on typhoon identification and location method. 32nd Annual Meeting of China Meteorological Society, Tianjin, China, Chinese Meteorological Society, 128–133.

  • Wang, G., and S. Xia, 2018: Application of FY-4 radiation imager and its data in satellite meteorology. J. Nature, 40, 111.

  • Wang, Y., Z. Ding, and X. Li, 2014: Mechanism and evolution analysis of the structure and track of the Tropical Cyclone Morakot before and after landing on Taiwan Island (in Chinese). J. Nat. Disasters, 23, 4757, https://doi.org/10.13577/j.jnd.2014.0606.

    • Search Google Scholar
    • Export Citation
  • Wei, K., Z. Jing, Y. Li, and S. Liu, 2011: Spiral band model for locating tropical cyclone centers. J. Pattern Recognit. Lett., 32, 761770, https://doi.org/10.1016/j.patrec.2010.12.011.

    • Search Google Scholar
    • Export Citation
  • Wimmers, A. J., and C. S. Velden, 2016: Advancements in objective multisatellite tropical cyclone center fixing. J. Appl. Meteor. Climatol., 55, 197212, https://doi.org/10.1175/JAMC-D-15-0098.1.

    • Search Google Scholar
    • Export Citation
  • Xie, T., W. Perrie, and W. Chen, 2010: Gulf Stream thermal fronts detected by synthetic aperture radar. Geophys. Res. Lett., 37, L06601, https://doi.org/10.1029/2009GL041972.

    • Search Google Scholar
    • Export Citation
  • Xu, Q., G. Zhang, X. Li, and Y. Cheng, 2016: An automatic method for tropical cyclone center determination from SAR. 2016 IEEE Int. Geoscience and Remote Sensing Symp., Beijing, China, IEEE, 2250–2252, https://doi.org/10.1109/igarss.2016.7729581.

  • Ying, M., W. Zhang, H. Yu, X. Lu, J. Feng, Y. Fan, Y. Zhu, and D. Chen, 2014: An overview of the China Meteorological Administration tropical cyclone database. J. Atmos. Oceanic Technol., 31, 287301, https://doi.org/10.1175/JTECH-D-12-00119.1.

    • Search Google Scholar
    • Export Citation
  • Zhao, C., J. Song, H. Leng, and J. Zhao, 2019: Objective center-finding algorithm for tropical cyclones in numerical models. Atmosphere, 10, 376, https://doi.org/10.3390/atmos10070376.

    • Search Google Scholar
    • Export Citation
  • Zheng, G., J. Liu, J. Yang, and X. Li, 2019: Automatically locate tropical cyclone centers using top cloud motion data derived from geostationary satellite images. IEEE Trans. Geosci. Remote Sens., 57, 10 17510 190, https://doi.org/10.1109/TGRS.2019.2931795.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Regional scope in this study.

  • Fig. 2.

    Distribution characteristics of Typhoon Halong perturbation values in the center area based on the best path dataset at 1800 UTC 5 Nov 2019.

  • Fig. 3.

    Best path data–based center, maximum perturbation value pixel position, first minimum perturbation value pixel position and second minimum perturbation value pixel position of Typhoon Halong at 1800 UTC 5 Nov 2019.

  • Fig. 4.

    Objective typhoon positioning result at 1200 UTC (a) 24 Feb (b) 25 Feb, and (c) 26 Feb 2019 (red cross) compared to the CMA track dataset (blue × symbol).

  • Fig. 5.

    (a) Track of Typhoon Wutip based on the objective positioning results from 21 to 28 Feb 2019 (red dashed line) and the CMA positioning path (blue solid line). The background image time is at 1200 UTC 25 Feb 2019. (b) Comparison of the objective positioning path (red dashed line) and CMA positioning path (blue solid line) at 0000, 0600, 1200, and 1800 from 23 to 25 Feb 2019.

  • Fig. 6.

    (a) Track of Typhoon Lekima based on the objective positioning results from 7 to 10 Aug 2019 (red dashed line) and CMA positioning path (blue solid line). The background image time is at 1200 UTC 8 Aug 2019; and (b) comparison of the objective positioning path (red dashed line) and CMA positioning path (blue solid line) at 0000, 0300, 0600, 0900, 1200, 1500, 1800, and 2100 UTC from 9 to 10 Aug 2019.

  • Fig. 7.

    (a) Track of Typhoon Hagibis based on the objective positioning results at 0000, 0600, 1200, and 1800 UTC from 7 to 12 Oct 2019 (red dashed line) and the CMA positioning path (blue solid line). The background image time is 1200 UTC 7 Oct 2019; and (b) comparison of the objective positioning path (red dashed line) and CMA positioning path (blue solid line) at 0000, 0600, 1200, and 1800 UTC from 9 to 10 Oct 2019.

  • Fig. 8.

    (a) Track of Typhoon Halong based on the objective positioning results at 0000, 0600, 1200, and 1800 UTC from 3 to 8 Nov 2019 (red dashed line) and CMA positioning path (blue solid line); the background image time is at 1200 UTC 6 Nov 2019; and (b) comparison of the objective positioning path (red dashed line) and CMA positioning path (blue solid line) at 0000, 0600, 1200, and 1800 UTC from 6 to 7 Nov 2019.

  • Fig. 9.

    (a) Brightness temperature in the eye area of Typhoon Wutip at 0600 UTC 25 Feb, selected with a 21 × 21 window; (b) calculated p value in the typhoon eye area selected with a 21 × 21 window; (c) distribution of pixels with values larger than the overall average value of the window in (b); (d) minimum pixel selection range for the pixels in (c); the distribution of pixels with values larger than the average value of the window in (c) represents the pixel distribution of the overall average value of the window; and (f) minimum pixel frame selection range for the pixels in (e).

  • Fig. 10.

    (a) Brightness temperature in the eye area of Typhoon Lekima at 2000 UTC 8 Aug, selected with a 21 × 21 window; (b) calculated p value in the typhoon eye area selected with a 21 × 21 window; (c) distribution of pixels with values larger than the overall average value of the window in (b); (d) minimum pixel selection range of the pixels in (c); (e) the distribution of pixels with values larger than the average value of the window in (c) represents the pixel distribution of the overall average value of the window; and (f) minimum pixel frame selection range of the pixels in (e).

  • Fig. 11.

    (a) Brightness temperature in the eye area of Typhoon Hagibis at 2000 UTC 7 Oct, selected with a 21 × 21 window; (b) calculated p value in the typhoon eye area selected with a 21 × 21 window; (c) distribution of pixels with values larger than the overall average value of the window in (b); (d) minimum pixel selection range of the pixels in (c); (e) the distribution of pixels with values larger than the average value of the window in (c) represents the pixel distribution of the overall average value of the window; and (f) minimum pixel frame selection range of the pixels in (e).

  • Fig. 12.

    (a) Brightness temperature in the eye area of Typhoon Halong at 0200 UTC 6 Nov, selected with a 21 × 21 window; (b) calculated p value in the typhoon eye area selected with a 21 × 21 window; (c) distribution of pixels with values larger than the overall average value of the window in (b); (d) minimum pixel selection range in (c); (e) the distribution of pixels with values larger than the average value of the window in of the pixels (c) represents the pixel distribution of the overall average value of the window; and (f) minimum pixel frame selection range of the pixels in (e).

All Time Past Year Past 30 Days
Abstract Views 158 0 0
Full Text Views 655 347 14
PDF Downloads 500 265 29