Storm Detection Research Articles

An Objective Scoring Method for Evaluating the Comparative Performance of Automated Storm Identification and Tracking Algorithms

Abstract While storm identification and tracking algorithms are used both operationally and in research, there exists no single standard technique to objectively determine performance of such algorithms. Thus, a comparative skill score is developed herein that consists of four parameters, three of which constitute the quantification of storm attributes—size consistency, linearity of tracks, and mean track duration—and the fourth that correlates performance to an optimal postevent reanalysis. The skill score is a cumulative sum of each of the parameters normalized from zero to one among the compared algorithms, such that a maximum skill score of four can be obtained. The skill score is intended to favor algorithms that are efficient at severe storm detection, i.e., high-scoring algorithms should detect storms that have higher current or future severe threat and minimize detection of weaker, short-lived storms with low severe potential. The skill score is shown to be capable of successfully ranking a large number of algorithms, both between varying settings within the same base algorithm and between distinct base algorithms. Through a comparison with manually created user datasets, high-scoring algorithms are verified to match well with hand analyses, demonstrating appropriate calibration of skill score parameters. Significance Statement With the growing number of options for storm identification and tracking techniques, it is necessary to devise an objective approach to quantify performance of different techniques. This study introduces a comparative skill score that assesses size consistency, linearity of tracks, mean track duration, and correlation to an optimal postevent reanalysis to rank diverse algorithms. This paper will show the capability of the skill score at highlighting algorithms that are efficient at detecting storms with higher severe potential, as well as those that closely resemble human-perceived storms through a comparison with manually created user datasets. The novel methodology will be useful in improving systems that rely on such algorithms, for both operational and research purposes focusing on severe storm detection.

Application of Random Forest Algorithm on Tornado Detection

Tornadoes are highly destructive small-scale extreme weather processes in the troposphere. The weather radar is one of the most effective remote sensing devices for the monitoring and early warning of tornadoes. The existing tornado detection algorithms based on radar data are unsupervised and have strict multi-altitude constraints, such as the tornado detection algorithm based on tornado vortex signatures (TDA-TVS), which may lead to high false alarm rates, and the performance of the detection algorithm is greatly affected by the radar data quality control algorithm. A novel TDA-RF algorithm based on the random forest (RF) classification algorithm is proposed for real-time tornado identification of the S-band China new generation of Doppler weather radar (CINRAD-SA). The TDA-RF algorithm uses velocity features to identify tornadoes and adds features related to reflectivity and velocity spectrum width in radar level-II data. Historical CINRAD-SA tornado data from 2006–2015 are used to construct the tornado dataset and train the TDA-RF model. The performance of TDA-RF is evaluated using CINRAD-SA data from five tornadoes of 2016–2020 with enhanced Fujita(EF) scale ratings ranging from EF0 to EF4 and distances from 10 to 130 km to the radar. TDA-RF performs well overall with the probability of detection (POD), false alarm ratio (FAR), and critical success index (CSI) of 71%, 29%, and 55%, respectively. Moreover, the TDA-RF improves POD and CSI, and reduces FAR compared to the TDA-TVS. The maximum tornado early-warning time of TDA-RF is 17 min, and the average is 6 min; TDA-RF can provide classification probability according to the tornado generation and development process to facilitate tracking ability.

On the potential use of infrasound for tornado detection and monitoring

The notion that tornadoes emit an infrasonic signal that can be detected from great distances; and thus be used to aid in the detection, monitoring, and potentially tracking of active tornadoes; has been discussed for decades. To test this notion, we deployed a network of infrasound sensor arrays in northern Alabama for the 2017, 2018, and 2019 tornado seasons. Infrasound propagation is complicated by vertical temperature gradients and wind shear. Efficient propagation depends on ducts caused by positive temperature and wind speed gradients. Infrasound signal detection depends on a signal arriving at the detection array with amplitude larger than the turbulent pressure amplitudes (wind noise) in the neighborhood of the array. Focusing on the 1 Hz to 10 Hz frequency band, we show that from each tornado that passed through northern Alabama during that period for which propagation and wind nose analysis suggested that a signal should be detected, a signal was detected. To make a truly convincing argument that the observed signals are produced by active tornadoes two things are needed. One needs to demonstrate that storms from which no tornadoes were spawned do not produce such signals and one must show that there is a physical mechanism related directly to tornadoes through which the infrasound is emitted. We will report on progress on both these fronts.

North–South IMF Disturbance Detection via an Adaptive Filter Approach

Geomagnetic storm-detection algorithms are important for space-weather-warning services to provide reliable warnings (e.g., ionospheric disturbances). For that reason, a new approach using an adaptive filter (least mean squares algorithm) for the detection of geomagnetic storms based on the volatility of the north–south interplanetary magnetic field Bz is presented. The adaptive filter is not dependent on solar wind plasma measurements, which are more frequently affected by data gaps than Bz, and is less dependent on the magnitude of Bz disturbances compared with other detection algorithms (e.g., static thresholds). The configuration of the filter is discussed in detail with three geomagnetic storm events, and required optimization as well as possible extensions are discussed. However, the proposed configuration performs satisfactorily without further improvements, and good correlations are observed with geomagnetic indices. Long-term changes are also reflected by the filter (solar cycles 23 and 24), and thus the performance is not affected by different solar wind conditions during the solar minimum and maximum. Conclusively, the proposed filter provides a good solution when more complex approaches (e.g., solar-wind–magnetosphere coupling functions) that rely on solar wind plasma measurements are not available.

Earlier onset of North Atlantic hurricane season with warming oceans

Numerous Atlantic basin tropical cyclones have recently developed prior to the official start of hurricane season, including several pre-season landfalls in the continental United States. Pre-season and early-season tropical cyclones disproportionately affect populated landmasses, often producing outsized precipitation impacts. Here we show a significant trend towards earlier onset of tropical cyclone activity in the North Atlantic basin, with threshold dates of the first three percentiles of accumulated cyclone energy shifting earlier at a rate exceeding five days decade−1 since 1979, even correcting for biases in climatology due to increased detection of short-lived storms. Initial threshold dates of continental United States named storm landfalls have trended earlier by two days decade−1 since 1900. The trend towards additional pre-season and early-season activity is linked to spring thermodynamic conditions becoming more conducive for tropical cyclone formation. Genesis potential index value increases in the western Atlantic basin are primarily driven by warming ocean temperatures.

GNSS Storm Nowcasting Demonstrator for Bulgaria

Global Navigation Satellite System (GNSS) is an established atmospheric monitoring technique delivering water vapour data in near-real time with a latency of 90 min for operational Numerical Weather Prediction in Europe within the GNSS water vapour service (E-GVAP). The advancement of GNSS processing made the quality of real-time GNSS tropospheric products comparable to near-real-time solutions. In addition, they can be provided with a temporal resolution of 5 min and latency of 10 min, suitable for severe weather nowcasting. This paper exploits the added value of sub-hourly real-time GNSS tropospheric products for the nowcasting of convective storms in Bulgaria. A convective Storm Demonstrator (Storm Demo) is build using real-time GNSS tropospheric products and Instability Indices to derive site-specific threshold values in support of public weather and hail suppression services. The Storm Demo targets the development of service featuring GNSS products for two regions with hail suppression operations in Bulgaria, where thunderstorms and hail events occur between May and September, with a peak in July. The Storm Demo real-time Precise Point Positioning processing is conducted with the G-Nut software with a temporal resolution of 15 min for 12 ground-based GNSS stations in Bulgaria. Real-time data evaluation is done using reprocessed products and the achieved precision is below 9 mm, which is within the nowcasting requirements of the World Meteorologic Organisation. For the period May–September 2021, the seasonal classification function for thunderstorm nowcasting is computed and evaluated. The probability of thunderstorm detection is 83%, with a false alarm ration of 38%. The added value of the high temporal resolution of the GNSS tropospheric gradients is investigated for a storm case on 24–30 August 2021. Real-time tropospheric products and classification functions are integrated and updated in real-time on a publicly accessible geoportal.

Detecting dust loads in the atmosphere over Thar desert by using MODIS and INSAT-3D data

• Dust detection algorithm based on MODIS and INSAT-3D were developed for Thar desert. • Brightness temperature difference and dust indices were used to detect dust load. • Integrated dust detection approach using MODIS and INSAT-3D was complementary. • Dust load detection over Indian Thar desert was successfully validated and tested. Suspended dust particles in atmosphere have adverse impacts on environment, ecosystem as well as on human health. To avoid negative impacts of dust storm events, early warning system to predict it well in advance may be a suitable option. However, for this purpose, assessment on magnitude of dust load and its dynamics in atmosphere is a primary requirement. The present study aims to develop remote sensing based assessment of dust load in atmosphere specifically over the Indian Thar Desert region. The severe dust storm event occurred on 5 th June 2017 over the Indian Thar Desert has been used in this study to develop integrated dust detection algorithm using split window technique, mid-infrared technique and different dust indices derived from MODIS and INSAT-3D data. Evaluation of the developed algorithm revealed that the area classified under dust load depends on threshold value of dust indices used in the algorithm, type of dust detection techniques followed and the specifications of remote sensing sensors used to retrieve the dust image. The integrated dust detection algorithm developed in this study has the capability to eliminate the problem in variations of predicted dust loadings in atmosphere. Validation of the developed algorithm to detect dust pixels showed good agreement with independent observations on aerosol optical depth (AOD), wind speed profile data and ground visibility data. The method adopted can be helpful to implement an operational system for detection and monitoring of dust storms over the Thar Desert region.

Thunderstorm days over Argentina: Integration between human observations of thunder and the world wide lightning location network lightning data

AbstractStorms are one of nature's most dangerous phenomena; therefore, knowing their spatial distribution and evolution over time is of great interest for the protection of society, as well as for climate change adaptation strategies. The measurement of Thunderstorm days (Td) was one of the first tools used to monitor storms. The advent of automatic detection networks on the surface has allowed us to advance in the understanding and characterization of the electrical activity in the atmosphere, locating in real‐time electrical discharges and providing information over previously unrecorded regions. This work focuses on the integration of human observations at conventional meteorological stations and the data provided by the WWLLN surface discharge detection network in Argentina. The calibration methodology applied determined a mean human thunderstorm detection radius of 21 km which allowed the elaboration of isokeraunic maps for the period 2008–2017 for the region of interest. The spatial distribution of storms yielded the highest values of Td in the Argentine Northwest region with values above 100 Td·year−1 followed by a relative maximum in the Argentine Northeast with 80 Td·year−1 and the Sierras de Córdoba with 50 Td·year−1.

Pengembangan Sistem Peringatan Dini Cuaca Ekstrem Terintegrasi Berbasis Y-Model Webgis Development Methodology (Y-WDM)

The incidence of hydrometeorological disasters in Indonesia is in the largest percentage for the last 10 years. The development of an integrated extreme weather early warning system is one of the initial processes in a series of hydrometeorological disaster mitigation efforts. Data models, observation and remote sensing become the database in early warning of extreme weather, the officers who carry out early detection of hazards so that they have sufficient waiting time to prepare an effective emergency response plan. In this study, an early warning system was built by implementing the Y-Model WebGIS Development Methodology (YWDM) which is a combination of geographic information systems and web-based application development context. Spatial data processing is carried out using the python programming language, including radar and weather satellite data, surface observation, lightning detection, storm detection and identification, and nowcasting models. Web GIS visualization is built by using JavaScript and php programming language, database management managed using PostgreSQL. The results of the implementation are in the form of a SPACE (Spatial Based Extreme Weather Warning System) framework which Sidarma-Nowcast is a system within the framework. The development of this system shows an increase and the level of improvement in the process of producing and disseminating early warning information.

A novel hybrid sand and dust storm detection method using MODIS data on GEE platform

ABSTRACT Accurate sand and dust storm (SDS) detection is important for assessing SDS disaster risk. Machine learning (ML) based SDS detection approaches have been widely used in recent years due to their higher accuracy and better detection results. However, this approach usually requires manual annotation of numerous training samples that are, in practice, laborious and time-consuming. To overcome this challenge, we propose a novel hybrid SDS detection method that combines the support vector machine (SVM) algorithm implemented on the Google Earth Engine (GEE) cloud computing platform with a spectral index to aid the automatic labelling of training samples. Based on 8 SDS events captured by MODIS over Arid Central Asia (ACA), the effectiveness and accuracy of this method were assessed and compared to traditional approaches. The experimental results indicate that the proposed method can distinguish between mixed pixels (thin cloud and land surface) and SDS pixels and that it minimizes misdetection more effectively. This method achieved more than 98% training accuracy and validation accuracy in SDS detection.

Dust storm detection of a convolutional neural network and a physical algorithm based on FY-4A satellite data

Quantitative identification of a dust storm weather is key to forecasting and early warning of dust storm disasters. However, traditional visibility ground-based measurements cannot be extended to regional observations. Remote sensing of dust storms is associated with large uncertainties in dust thresholds. For accurate quantification of the dust storm region, this study proposes a dust storm mask algorithm to identify dust storm in the Tarim Basin. The dust storm mask includes two algorithms to identify the dust storm outbreak and the spatial extent by using the Advanced Geostationary Radiation Imager (AGRI) on board the FY-4A satellite. A deep learning convolutional neural network (CNN) is employed for the dust storm mask, and the AGRI bands 1–3, 5–6, and 11–13 are used as model parameters. A physical algorithm (PA) is adopted to construct a dust storm mask using three physical dust indices: the Normalized Difference Dust Index (NDDI2.25μm-0.47μm/2.25μm+0.47μm), the Dust Ratio Index (DRI7.10μm/3.75μm), and the Brightness Temperature Difference (BTD3.75μm-13.50μm). The results show that the CNN algorithm has a higher classification accuracy on dust storm detection compared to the PA. This advantage suggests that the CNN can effectively monitor large-scale dust storms. The dust storm identification results were compared and analyzed with the AGRI true color, Aerosol Optical Depth products, and Ultra Violet Aerosol Index products.

Detection of Martian dust storms using mask regional convolutional neural networks

Martian dust plays a crucial role in the meteorology and climate of the Martian atmosphere. It heats the atmosphere, enhances the atmospheric general circulation, and affects spacecraft instruments and operations. Compliant with that, studying dust is also essential for future human exploration. In this work, we present a method for the deep-learning-based detection of the areal extent of dust storms in Mars satellite imagery. We use a mask regional convolutional neural network, consisting of a regional-proposal network and a mask network. We apply the detection method to Mars daily global maps of the Mars global surveyor, Mars orbiter camera. We use center coordinates of dust storms from the eight-year Mars dust activity database as ground-truth to train and validate the method. The performance of the regional network is evaluated by the average precision score with 50% overlap (mAP_{50}), which is around 62.1%.

A characterisation of Alpine mesocyclone occurrence

Abstract. This work presents a characterisation of mesocyclone occurrence and frequency in the Alpine region, as observed from the Swiss operational radar network; 5 years of radar data are processed with a thunderstorm detection and tracking algorithm and subsequently with a new mesocyclone detection algorithm. A quality assessment of the radar domain provides additional information on the reliability of the tracking algorithms throughout the domain. The resulting data set provides the first insight into the spatiotemporal distribution of mesocyclones in the Swiss domain, with a more detailed focus on the influence of synoptic weather, diurnal cycle and terrain. Both on the northern and southern side of the Alps mesocyclonic signatures in thunderstorms occur regularly. The regions with the highest occurrence are predominantly the Southern Prealps and to a lesser degree the Northern Prealps. The parallels to hail research over the same region are discussed.

ESTIMATION OF NITROGEN DIOXIDE (NO2) CONCENTRATION USING FENGYUN-4A DUST STORM DETECTION (FY-4A DSD) PRODUCT DURING THE COVID-19 LOCKDOWN IN METRO MANILA, PHILIPPINES

Abstract. The coronavirus disease was discovered in 2019 (COVID-19) and was eventually declared by the World Health Organization (WHO) as a global pandemic on March 11, 2020. This study aims to determine the correlation of Dust Storm Detection (DSD) index from the Fengyun-4A (FY-4A) data, NO2 ground observations, and daily positivity rate of COVID-19 cases in Metro Manila and predict the confirmed daily cases of COVID-19 using the established correlations. FY-4A DSD products were used to obtain different Dust Storm Indices (DSI) and daily COVID-19 confirmed cases were tallied during the period November 01–30, 2020, and March 01–31, 2021. Ground-observed NO2 levels from Environmental Management Bureau (EMB) monitoring stations were gathered for validation and regression analysis. Results of linear regression analysis between the DSI and NO2 exhibited a weak correlation (0.24) with the available observations at the specified period. Ground-observed NO2 levels exhibit the same trend with the daily positivity rate of COVID-19 considering only a smaller area and short period of observations. Moreover, results showed a weak correlation (0.07) between the positivity rate of COVID-19 case as a function of the DSD Index and ground-observed NO2 levels. Uncertainty of results from this study may be attributed to the fact that it has focused on a relatively small area due to limited available ground observations. Therefore, it is recommended to apply the same analysis on different periods of observations using as many NO2 level ground observations as available and determine if variables follow the same trend and correlations as reported in this study.

Coastal storm waves detection system design using Beaufort scale standardization and Sugianto wave forecasting method in Timbulsloko, Demak, Central Java, Indonesia

Storm is defined as a disturbance of the atmosphere marked by winds and usually by rain. Coastal storms must comprise a maritime component, such as waves, currents and/or water levels. Coastal storm detection is necessary so the number of casualties and losses caused by these events can be reduced. The method used in this system is the Sugianto wave forecasting method with standardization of coastal storms using the Beaufort scale. The purpose of this study is to built up an internet of things based system to observe coastal storm information and wave forecasting data from wind speed data that obtained in Timbulsloko, Demak, Central Java, Indonesia. The tidal data is processed using the Admiralty method. This system was built using Arduino Uno equipped with anemometer JL-FS2 to measure wind and waves parameters. The power source from 100 wp solar panels stored in a 40 Ah accumulator. Data from field instrument is stored to the IoT MAPID database using NodeMCU ESP8266. This system is placed in Timbulsloko, Demak. The results of field observation then validated using BMKG. This method could be applied in other location along the north coast of Java. The results of field observation showed an average wind speed 3.9848 m/s; significant wave height 0.4632 m; significant wave period 3.8641 s; wave energy 493.90 J/m2; wind energy 116.74 W/m2.

A new threshold free dust storm detection index based on MODIS reflectance and thermal bands

ABSTRACT Wind and soil confrontation under specific atmospheric conditions leads to the horrendous phenomenon of dust storms, particularly in arid lands. The negative impact of storms on the health and life of living beings is evident, resulting in many losses in the economic, social, and environmental sectors. Dust storm detection is the most important topic in dust studies, where many algorithms and indices have already been proposed using satellite imageries and remote sensing techniques. Despite the advantage of dust detection, these algorithms present drawbacks. This includes the need to determine different thresholds for different regional events or simply better respond to oceanic dust storms. The objective of this study is to develop a new Dust Storm Detection Index, called DSDI, without the need to determine different thresholds for each event. To establish the new index, it was necessary to distinguish the properties of dust, clouds, and land surfaces in the images of the MODIS sensors. These properties have been characterized by analyzing their spectral and thermal profiles at different wavelengths during dust storms from 2004 to 2009 in Yazd province, Central Iran, which is constantly coping with severe storms. Spectral and thermal ranges of 0.46, 0.56, 3.9, 1.4, 11, 12, and 13.6 μm were the most suitable discriminating bands of dust from the cloud and land surfaces. The brightness temperature difference of the thermal bands and the spectral ratio of reflectance bands, i.e. DSDI, has developed an appropriate relationship for separating land surfaces and clouds of dust particles. There was a significant correlation between DSDI and horizontal visibility (P-value= 0.05 & 0.01). It confirmed the success of this algorithm to detect storms in the study area, as an arid land.

A Variational Method for Filling in Missing Data in Doppler Velocity Fields

AbstractSingle-Doppler radar data are often missing in important regions of a severe storm due to low return power, low signal-to-noise ratio, ground clutter associated with normal and anomalous propagation, and missing radials associated with partial or total beam blockage. Missing data impact the ability of WSR-88D algorithms to detect severe weather. To aid the algorithms, we develop a variational technique that fills in Doppler velocity data voids smoothly by minimizing Doppler velocity gradients while not modifying good data. This method provides estimates of the analysed variable in data voids without creating extrema.Actual single-Doppler radar data of four tornadoes are used to demonstrate the variational algorithm. In two cases, data are missing in the original data, and in the other two, data are voided artificially. The filled-in data match the voided data well in smoothly varying Doppler velocity fields. Near singularities such as tornadic vortex signatures, the match is poor as anticipated. The algorithm does not create any velocity peaks in the former data voids, thus preventing false triggering of tornado warnings. Doppler circulation is used herein as a far-field tornado detection and advance-warning parameter. In almost all cases, the measured circulation is quite insensitive to the data that have been voided and then filled. The tornado threat is still apparent.

Improving the effectiveness of monitoring of dangerous meteorological phenomena

Three approaches to improving the effectiveness of monitoring of dangerous meteorological phenomena are considered. The first one is based on the modernization of incoherent weather radar on the example of MRL-5 for the possibility of estimating the width of the doppler spectrum by implementing algorithms for pseudo-coherent frequency signal processing with subsampling. Mathematical modeling is carried out, confirming the technical feasibility of the algorithms. On the basis of mathematical analysis, expressions are obtained and graphs of the suppression coefficient of the moving target selection system are presented. The second approach offers a way to improve the efficiency of monitoring for turbulence. An algorithm for detecting zones of increased turbulence and estimating their parameters is proposed. Equations are obtained, on the basis of which algorithms for detecting and estimating the normalized width of the doppler spectrum of zones of increased turbulence, as well as determining the power level of the reflected signals, are synthesized. The third approach describes a way to improve the efficiency of thunderstorm detection and the accuracy of their localization by supplementing the MRL-5 with magnetometric equipment. An example of the implementation and brief technical specifications are given.

Detection and Tracking of Tropical Convective Storms Based on Globally Gridded Precipitation Measurements: Algorithm and Survey over the Tropics

AbstractThis paper is the first attempt to document a simple convection-tracking method based on the IMERG precipitation product to generate an IMERG-based Convection Tracking (IMERG-CT) dataset. Up to now, precipitation datasets have been Eulerian accumulations. Now with IMERG-CT, we can estimate total rainfall based on Lagrangian accumulations, which is a very important step in diagnosing cloud-precipitation process following the evolution of air masses. Convection-tracking algorithms have traditionally been developed on the basis of brightness temperature (Tb) from satellite infrared (IR) retrievals. However, vigorous rainfall can be produced by warm-topped systems in a moist environment; this situation cannot be captured by traditional IR-based tracking but is observed in IMERG-CT. Therefore, an advantage of IMERG-CT is its ability to include the previously missing information of shallow clouds that grow into convective storms, which provides us more-complete life cycle records of convective storms than traditional IR-based tracking does. This study also demonstrates the utility of IMERG-CT through investigating various properties of convective systems in terms of the evolution before and after peak precipitation rate and amount. For example, composite analysis reveals a link between evolution of precipitation and convective development: the signature of stratiform anvils remaining after the storm has produced the maximum rainfall, as average Tb stays almost constant for 5 h after the peak of precipitation. Our study highlights the importance of joint analysis of cloud and precipitation data in time sequence, which helps to elucidate the underlying dynamic processes producing tropical rainfall and its resultant effects on the atmospheric thermodynamics.

A Broadcast Storm Detection and Treatment Method Based on Situational Awareness

A Broadcast Storm Detection and Treatment Method Based on Situational Awareness