Predictions of seasonal to near-future tropical cyclone activities over the western North Pacific and the North Atlantic

Abstract

Every summertime, tropical cyclone (TC) activity over the worldwide tropical ocean has been receiving large attention due to its destructive impacts on heavily populated countries. To reduce and prepare the potential damages from the TC approach/landfall, development of skillful TC prediction model has been one of the most essential missions for meteorological agency. In this dissertation, the detailed physical relationships between TC activity and environmental fields are investigated. On the basis of these understandings, a track-pattern-based model is developed to predict seasonal to near-future TC activity over the western North Pacific (WNP) and the North Atlantic (NA) basins. This model employs a hybrid statistical–dynamical method and is the first approach to predicts spatial distribution of TC track density covering the entire basin. Thus, it would be a milestone for the prediction of long-term TC track distribution without simulating the climate model. There are three major steps to operate the track-pattern-based model. First, climatological basin-wide TC tracks during the TC season are identified into several patterns using the fuzzy c-means method. Second, the TC counts for each cluster are predicted by using a hybrid statistical–dynamical method. The hybrid prediction for each pattern is based on the statistical relationships (interannual correlation in this thesis) between the seasonal TC frequency of the pattern and the seasonal-mean key predictors dynamically forecast by the National Centers for Environmental Prediction (NCEP) Climate Forecast System version 2 (CFSv2). Third, the final forecast map of track density is constructed by merging the spatial probabilities of the all clusters and applying necessary bias corrections. The leave-one-out cross validation shows good skill of the WNP TC prediction model, with the correlation coefficients between the hindcasts and the observations ranging from 0.71 to 0.81. The hindcasts of the WNP seasonal TC track density exhibit significant predictability in reproducing the observed pattern. As a real forecast, this model fairly forecast the anomalous spatial distribution of WNP TC track density for the 2010 typhoon season, representing the lowest count since 1951. A higher-than-normal track density was successfully forecast near the East China Sea, Korea, and Japan. The total seasonal TC genesis frequency integrated over the seven patterns is well below normal (about 16.4) close to the observations. The skillful performance in 2010 using the seasonal TC prediction model is attributed to the skillful forecast of the ENSO transition by the NCEP CFS, cooperated with the validity of the prediction model itself. In addition to the WNP basin, a seasonal prediction model of the NA TC activities for the period August–October has been developed on the basis of representative TC track patterns. Using the fuzzy c-means method, a total of 432 TCs are categorized into the following four groups: 1) TCs off the East Coast of the United States, 2) TCs over the Gulf of Mexico, 3) TCs that recurve into the open oceans of the central NA, and 4) TCs that move westward in the southern NA. The model is applied to predict the four TC groups separately in conjunction with global climate forecasts from the NCEP CFSv2. By adding the distributions of the four TC track patterns with pre-calculated TC genesis frequencies, this seasonal TC forecast model provides the spatial distribution of TC activities over the entire NA basin. Multiple forecasts initialized in six consecutive months from February to July are generated at monthly intervals to examine the applicability of this model in operational TC forecasting. Cross-validations of individual forecasts show that the model can reasonably predict the observed TC frequencies over NA at the 99% confidence level. The model shows a stable spatial prediction skill, proving its advantage for forecasting regional TC activities several months in advance. In particular, the model can generate reliable information on regional TC counts in the near-coastal regions as well as in entire NA basin. Among the TC activity, intense TCs accompanying torrential rain and powerful wind gusts often cause substantial socio-economic losses in the regions around their landfall than weak TCs. Thus, we develop the prediction model targeting only intense TCs in the WNP and the NA basins. Different intensity criteria are used to define intense TCs for these two basins, category 3 and above for WNP and category 1 and above for NA, because the number of TCs in the NA basin is much smaller than that in the WNP basin. Using a fuzzy clustering method, intense TC tracks in the WNP and the NA basins are classified into three and two representative patterns, respectively. On the basis of the clustering results, a track-pattern-based model is then developed for forecasting the seasonal activities of intense TCs in the two basins. Generally, the WNP intense TC patterns have predictors of dynamical factor (vertical wind shear or low-level relative vorticity) because of thermally mature state over the WNP to develop the TC whereas the NA intense TCs have thermodynamical factor (sea surface temperature) to the predictor due to the thermally insufficient condition to generate TC over the NA. Cross-validation of the model skill for entire training period as well as verification of a forecast for the 2014 TC season suggest that our intense TC model is applicable to operational uses. Although many studies have attempted to predict TC activities on various time scales, very few focused on near-future predictions. Here we show a decrease in seasonal TC activity over the NA for 2016–2030 using the track-pattern-based TC prediction model. The prediction model is forced by long-term coupled simulations, CFSv2 free runs, initialized using reanalysis data. Unfavorable conditions for TC development including strengthened vertical wind shear, enhanced low-level anticyclonic flow, and cooled sea surface temperature over the tropical NA are found in the simulations. Most of the environmental changes are attributable to cooling of the NA basin-wide sea surface temperature (NASST) and more frequent El Niño episodes in the near future. Consistent NASST warming trend in the Coupled Model Intercomparison Project phase 5 projections suggests that natural variability is still dominant than anthropogenic forcing over the NA in the near-future period.