An Integrated Model for Tropical Cyclone Track Forecasting Based on Deep Learning Techniques

Since 1970, tropical cyclone (TC) track forecasts have improved steadily in the Atlantic basin. This improvement has been linked primarily to advances in numerical weather prediction (NWP) models. Concurrently, with few exceptions, the development and operational use of statistical track prediction schemes have experienced a relative decline. Statistical schemes provided the most accurate TC track forecasts until approximately the late 1980s. In this note, it is shown that increased reliance on the global NWP models does not always guarantee the best forecast. Here, Hurricane Ivan is used from the 2004 Atlantic TC season as a classical example, and reminder, of how strong climatological signals still can add substantial value to TC track forecasts, in the form of improved accuracy and increased timeliness at minimal computational cost. In an 8-day period in early September 2004, Hurricane Ivan was repeatedly, and incorrectly, forecast by 12 operational NWP models to move with a significant northward (poleward) component. It was found that the mean 24-h trajectory forecasts of a consensus of five commonly used NWP track prediction aids had a statistically significant right-of-track bias. Furthermore, the official track forecasts, which relied heavily on erroneous numerical guidance over this period, were also found to have significant poleward trajectory errors. At the same time, a climatology-based prediction technique, drawn entirely from the historical record of motion characteristics of TCs in geographical locations similar to Ivan, correctly and consistently indicated a more westward motion component, had a small directional spread, and was supported by a large number of archived cases. This climatological signal was in conflict with the deterministic NWP model output, and it is suggested that the large errors in the official track forecast for TC Ivan could have been reduced considerably by taking into greater account such a strong climatological signal. The potential impact of such an error reduction is a saving of lives and billions of dollars in both actual damage and unnecessary evacuations costs, for just this one hurricane. We also suggest that this simple strategy of examining the strength of the climatological signal be considered for all TCs to identify cases where the NWP and official forecasts differ significantly from strong, persistent climatological signals.

This study compares the 5-day tropical cyclone (TC) track forecasts from the Global Ensemble Forecast System (GEFS) with a machine learning approach applied to this ensemble. A convolutional neural network (CNN) model was developed using northern Atlantic TC training data from 2008 to 2022, while an independent set of storms was used to evaluate the CNN model. The CNN for tropical cyclone track forecasting was trained using fivefold cross validation on a shuffled dataset split into 90% training/validation and 10% evaluation. PyTorch was used to develop a CNN model with a custom loss function, and performance was assessed using the haversine function and error metrics, comparing the CNN’s TC track forecasts to GEFS mean track forecasts. The CNN has a 58%–86% better track prediction against GEFS mean for forecast hours 0–96, which decreases to 35%–53% for hours 108–120. Track differences between those cases that the CNN improved the forecast versus did not improve are also explored, which shows that most of the CNN improvement is from a decrease in the along-track error (ATE). This is consistent with past studies of this ensemble, which showed that the largest bias exists in the along-track direction. Significance Statement This research focuses on enhancing the accuracy of tropical cyclone track forecasts by leveraging a convolutional neural network (CNN). Traditional forecasting models have relatively large errors in predicting the track of tropical cyclones, especially at forecast lead times of 3–5 days. These errors can result in inadequate warnings and misallocation of resources during storm preparations, potentially endangering lives and property. Our CNN model addresses these issues by correcting systematic biases, resulting in more accurate forecasts.

In the application of Tropical Cyclone Track (TCT) forecast in South China Sea (SCS), pure linear neural network (PLNN) is used as the expert in the committee machine model, and it partly determines the model output. Data normalization is one of the most important factors, which affect the performance of the individual expert net. This paper aims to find how much data normalization affects the convergence process and forecast accuracy of the independent samples, and whether data normalization could bring reasonably accuracy and diversity that the committee machine model requires.

A New Verification Measure for Tropical Cyclone Track Forecasts and its Experimental Application

In this study, the general impact of high-resolution moving nesting domains on tropical cyclone (TC) intensity and track forecasts was verified, for a total of 107 forecast cases of 33 TCs, using the Weather Research and Forecasting (WRF) model. The experiment, with a coarse resolution of 12 km, could not significantly capture the intensification process, especially for maximum intensities (>60 m s−1). The intense TCs were better predicted by experiments using a moving nesting domain with a horizontal resolution of 4 km. The forecast errors for maximum wind speed and minimum sea-level pressure decreased in the experiment with higher resolution; the forecast of lifetime maximum intensity was improved. For the track forecast, the experiment with a coarser resolution tended to simulate TC tracks deviating rightward to the TC motions in the best-track data; this erroneous deflection was reduced in the experiment with a higher resolution. In particular, the track forecast in the experiment with a higher resolution improved more frequently for intense TCs that were generally distributed at relatively lower latitudes among the test cases. The sensitivity of the track forecast to the model resolution was relatively significant for lower-latitude TCs. On the other hand, the track forecasts of TCs moving to the mid-latitudes, which were primarily influenced by large-scale features, were not sensitive to the resolution.

The orthogonal conditional nonlinear optimal perturbations (O-CNOPs) approach for measuring initial uncertainties is applied to the Weather Research and Forecasting (WRF) Model to provide skillful forecasts of tropical cyclone (TC) tracks. The hindcasts for 10 TCs selected from 2005 to 2020 show that the ensembles generated by the O-CNOPs have a greater probability of capturing the true TC tracks, and the corresponding ensemble forecasts significantly outperform the forecasts made by the singular vectors, bred vectors, and random perturbations in terms of both deterministic and probabilistic skills. In particular, for two unusual TCs, Megi (2010) and Tembin (2012), the ensembles generated by the O-CNOPs successfully reproduce the sharp northward-turning track in the former and the counterclockwise loop track in the latter, while the ensembles generated by the other methods fail to do so. Moreover, additional attempts are performed on the real-time forecasts of TCs In-Fa (2021) and Hinnamnor (2022), and it is shown that O-CNOPs are very useful for improving the accuracy of real-time TC track forecasts. Therefore, O-CNOPs, together with the WRF Model, could provide a new platform for the ensemble forecasting of TC tracks with much higher skill.

High-altitude, long-endurance unmanned aircraft systems (HALE UAS) are capable of extended flights for atmospheric sampling. A case study was conducted to evaluate the potential impact of dropwindsonde observations from HALE UAS on tropical cyclone track prediction; tropical cyclone intensity was not addressed. This study employs a global observing system simulation experiment (OSSE) developed at the National Oceanic and Atmospheric Administration/Earth System Research Laboratory (NOAA/ESRL) that is based on the NOAA/National Centers for Environmental Prediction gridpoint statistical interpolation (GSI) data assimilation system and Global Forecast System (GFS) model. Different strategies for dropwindsonde deployment and UAS flight paths were compared. The introduction of UAS-deployed dropwindsondes was found to consistently improve the track forecast skill during the early forecast up to 96 h, with the caveat that the experiments omitted both vortex relocation and dropwindsondes from manned flights in the tropical cyclone region. The more effective UAS dropwindsonde deployment patterns sampled both the environment and the body of the tropical cyclone.

<p>The accuracy of tropical cyclone (TC) track forecast has been improved year by years, on the other hand, the forecast of tropical cyclone intensity still has a difficulty of improvement. Recently the relationship between lightning activity and tropical cyclone intensity has been investigated. Lightning tends to increase during the rapid intensification of the TC. Therefore, monitoring the lightning activity becomes important for a TC intensity forecast. Lightning observation network are deployed over the western north Pacific by five very long frequency events trigger measurements called V-POTEKA at Palau, Guam, Manila, Okinawa Japan and Serpong Indonesia under the ULAT (Understanding Lightning and Thunderstorm) of SATREPS (Science and Technology Research Partnership for Sustainable Development) in the Philippines.</p><p>Tropical storm (TS) Bavi and Maysak were generated over the Philippine Sea on August 2020. We found that lightning activity drastically increased in the TC during the TC genesis stage when the TC reached TS criteria. Numbers of lightning reached the maximum during the life cycle of the TCs. Lightning was concentrated in the convective clouds about 100 to 200 km size. They are located around 400 to 500 km and 100 to 200 km from the TC center respectively and lasted about few hours. We called this phenomenon as “lightning burst”. On the other hand, when TS Haishen and Dolphin were generated over the Philippine Sea in September 2020, they did not observe lightning burst. About half of the TCs observed lightning burst in the Philippine Sea in 2020. We will investigate further what kind of structure occurred during the lightning burst and what kind of mechanism responsible for the lightning burst.</p>

We used a two-dimensional quasi-geostrophic barotropic model simulation to study effects of an initial brows-like meso-scale vortex on tropical cyclone (TC) track. Our results show that the impact of each of the three foundational factors (the environmental current, the asymmetric structure and the asymmetric convection system) on TC track varies with time and the importance of each of the factors is different for the different TC motion time period. They show two kinds of the effects. One is a direct way. The asymmetric outer wind structure and the positive longitudinal wind speed averaged in radial-band (100–300) km in the period of (0–11) h are caused by the introduction of the initial brows-like meso-scale vortex, which results in TC track to turn to the north from the northwest directly. The other is an indirect influence. First, initial TC axisymmetric circulation becomes a non-axisymmetric circulation after the addition of the meso-scale vortex. The initial non-axisymmetric circulation experiences an axisymmetrizational process in the period of (0–11) h. Second, axisymmetrizationed TC horizontal size is enlarged after t=12 h. Third, both the TC asymmetric structure and the TC energy dispersion induced-anticyclone are intensified, which quickens the TC motion and results in the track to turn to the north indirectly. The TC motion is characterized by the unusual track under the direct and the indirect effect. The formation of the unusual track should be attributed to the common effects of three factors, including the environmental flow, the TC asymmetric structure and the asymmetric convection system.

The impacts of special Geostationary Operational Environmental Satellite (GOES) rapid-scan (RS) wind observations on numerical model 24–120-h track forecasts of Hurricane Katrina are examined in a series of data assimilation and forecast experiments. The RS wind vectors are derived from geostationary satellites by tracking cloud motions through successive 5-min images. In these experiments, RS wind observations are added over the area 15°–60°N, 60°–110°W, and they supplement the observations used in operational forecasts. The inclusion of RS wind observations reduces errors in numerical forecasts of the Katrina landfall position at 1200 UTC 29 August 2005 by an average of 12% compared to control cases that include “targeted” dropsonde observations in the Katrina environment. The largest average improvements are made to the 84- to 120-h Katrina track forecasts, rather than to the short-range track forecasts. These results suggest that RS wind observations can potentially be used in future cases to improve track forecasts of tropical cyclones.

An accurate tropical cyclone track and intensity forecast is very important for disaster management. Specialized numerical prediction models have been recently used to provide high-resolution temporal and special forecasts. Hurricane Weather Research and Forecast (HWRF) model is one of the emerging numerical models for tropical cyclone forecasting. This study evaluates the performance of HWRF model during the post monsoon tropical cyclone Nilofar on the north Indian Ocean basin. The evaluation uses the best track data provided by the Indian Meteorological Department (IMD) and the Joint Typhoon Warning Centre (JTWC). Cyclone track, central pressure, and wind speed are covered on this evaluation. Generally, HWRF was able to predict the Nilofar track with track error less than 230 km within the first 66 h of forecast time span. HWRF predicted more intense tropical cyclone. It predicted the lowest central pressure to be 922 hPa while it reached 950 hPa according to IMD and 937 hPa according to JTWC. Wind forecast was better as it predicted maximum wind speed of 122 kt while it reached 110 and 115 kt according to IMD and JTWC, respectively.

Forced interannual-to-decadal variability of annual tropical cyclone (TC) track density in the western North Pacific between 1979 and 2008 is studied using TC tracks from observations and simulations by a 25-km-resolution version of the GFDL High-Resolution Atmospheric Model (HiRAM) that is forced by observed sea surface temperatures (SSTs). Two modes dominate the decadal variability: a nearly basinwide mode, and a dipole mode between the subtropics and lower latitudes. The former mode links to variations in TC number and is forced by SST variations over the off-equatorial tropical central North Pacific, whereas the latter might be associated with the Atlantic multidecadal oscillation. The interannual variability is also controlled by two modes: a basinwide mode driven by SST anomalies of opposite signs located in the tropical central Pacific and eastern Indian Ocean, and a southeast–northwest dipole mode connected to the conventional eastern Pacific ENSO. The seasonal evolution of the ENSO effect on TC activity is further explored via a joint empirical orthogonal function analysis using TC track density of consecutive seasons, and the analysis reveals that two types of ENSO are at work. Internal variability in TC track density is then examined using ensemble simulations from both HiRAM and a regional atmospheric model. It exhibits prominent spatial and seasonal patterns, and it is particularly strong in the South China Sea and along the coast of East Asia. This makes an accurate prediction and projection of TC landfall extremely challenging in these regions. In contrast, basin-integrated metrics (e.g., total TC counts and TC days) are more predictable.

A new approach is proposed to assess the possible impacts of the global climate change on tropical cyclone (TC) tracks in the western North Pacific (WNP) basin. The idea is based on the premise that the future change of TC track characteristics is primarily determined by changes in large-scale environmental steering flows and in formation locations. It is demonstrated that the main characteristics of the current climatology of TC tracks can be derived from the climatological mean velocity field of TC motion by using a trajectory model. The climatological mean velocity of TC motion, composed of the large-scale steering and beta drift, is determined on each grid of the basin. The mean large-scale steering flow is computed from the NCEP–NCAR reanalysis for the current climate state. The mean beta drift is estimated from the best-track data by removing the steering flow. The derived mean beta drift agrees well with the results of previous observational and numerical studies in terms of its direction and magnitude. The approach is applied to assessing the potential impacts of global warming on TC tracks in the WNP. The possible changes in the large-scale steering flows are taken from the output wind fields of two Geophysical Fluid Dynamics Laboratory (GFDL) global warming experiments and possible changes in the TC formation locations are considered by shifting the formation locations as a whole. The GFDL experiments suggested that the changes in the future large-scale steering flows are dominated by the easterly anomalies in the Tropics and westerly anomalies in the midlatitudes with the enhanced northward component during the period of 2030–59. Based on the assessments using two different ways to reduce climate model biases, the prevailing TC tracks shift slightly southwestward during the period of 2000–29, but northeastward during the period of 2030–59. More TCs will take a recurving track and move northeastward during the period of 2030–59. The El Niño–like climate change predicted in many climate models can significantly enhance the track changes if the TC formation locations in the WNP shift eastward as a whole.

The impact of different vertical structures of a nearby monsoon gyre (MG) on a tropical cyclone (TC) track is investigated using idealized numerical simulations. In the experiment with a relatively deeper MG, the TC experiences a sharp northward turn at a critical point when its zonal westward-moving speed slows down to zero. At the same time, the total vorticity tendency for the TC wavenumber-1 component nearly vanishes as the vorticity advection by the MG cancels the vorticity advection by the TC. At this point, the TC motion is dominated by the beta effect, as in a no-mean-flow environment, and takes a sharp northward turn. In contrast, the TC does not exhibit a sharp northward turn with a shallower MG nearby. In the case with a deeper MG, a greater relative vorticity gradient of the MG promotes a quicker attraction between the TC and MG through the vorticity segregation process. In addition, a larger outer size of the TC also favors a faster westward propagation from its initial position, thus having more potential to collocate with the MG. Once the coalescence is in place, the Rossby wave energy dispersion associated with the TC and MG together is enhanced and rapidly strengthens the southwesterly flow on the eastern flank of both systems. The steering flow from both the beta gyre and the Rossby wave dispersion leads the TC to take a sharp northward track when the total vorticity tendency is at its minimum. This study indicates the importance of good representations of the TC structure and its nearby environmental flows in order to accurately predict TC motions.

Recent studies have shown that assimilating enhanced satellite-derived atmospheric motion vectors (AMVs) has improved mesoscale forecast of tropical cyclones (TC) track and intensity. The authors conduct data-denial experiments to understand where the TC analyses and forecasts benefit the most from the enhanced AMV information using an ensemble Kalman filter and the Weather Research and Forecasting Model. The Cooperative Institute for Meteorological Satellite Studies at the University of Wisconsin provides enhanced AMV datasets with higher density and temporal resolution using shorter-interval image triplets for the duration of Typhoon Sinlaku and Hurricane Ike (both 2008). These AMV datasets are then spatially and vertically subsetted to create six parallel cycled assimilation-forecast experiments for each TC: all AMVs; AMVs withheld between 100 and 350 hPa (upper layer), between 350 and 700 hPa (middle layer), and between 700 and 999 hPa (lower layer); and only AMVs within (interior) and outside (exterior) 1000-km radius of the TC center. All AMV subsets are found to be useful in some capacity. The interior and upper-layer AMVs are particularly crucial for improving initial TC position, intensity, and the three-dimensional wind structure along with their forecasts. Compared with denying interior or exterior AMVs, withholding AMVs in different tropospheric layers had less impact on TC intensity and size forecasts. The ensemble forecast is less certain (larger spread) in providing accurate TC track, intensity, and size when upper-layer AMVs or interior AMVs are withheld. This information could be useful to potential targeting scenarios, such as activating and focusing satellite rapid-scan operations, and decisions regarding observing system assessments and deployments.