Research

New York State hurricane hazard: history and future projections

This study addresses hurricane hazard to the state of New York in past, present, and future, using synthetic storms generated by the Columbia HAZard model (CHAZ) and climate inputs from the fifth coupled model intercomparison project (CMIP5), in conjunction with historical observations. The projected influence of anthropogenic climate change on future hazard is quantified by the normalized differences in statistics of hurricane hazard between the recent historical period (1951-2005) and the two future periods under the Representative Concentration Pathway 8.5 warming scenario: the near-future (2006-2040) and the late 21st century (2070-2099). Changes in return periods of storms affecting the state at given intensities are computed, as are wind hazards for individual counties. Other storm characteristics examined include hurricane intensity, forward speed, heading and rate of change of the heading.

SUB-SEASONAL TROPICAL CYCLONE TRACK PREDICTIONS IN THE S2S DATASET

Probabilistic tropical cyclone (TC) occurrence, at lead times of week 1–4, in the Subseasonal to Seasonal (S2S) dataset are examined here using the Brier skill score with reference to climatological reference forecasts. In this study, I examined hree calibration techniques, removing the mean genesis and occurrence forecast biases, and a linear regression method, are explored here. The linear regression method performs the best and guarantees a higher skill score when applied to the in-sample dataset. After applying the linear regression method, the sklills of the global weather prediction systems improved. Calibarted ECMWF forecasts is skillful in most of the TC basins up to week 2 lead time.

STATISTICAL-DYNAMICAL DOWNSCALING PRJECTION OF TROPICAL CYCLONE ACTIVITY IN A WARMING CLIMATE: THE TWO DIVERGING GENESIS SCENARIO

TC climatology and the consequent TC hazard under future climate projections using a statistical-dynamical downscaling approach. Specifically, six CMIP5 models are downscaled using the Columbia HAZard model (CHAZ). We found that projections of global and basin TC frequency depend sensitively on the choice of moisture variable used in the genesis component of CHAZ. Simulations using column relative humidity (TCGI_CRH) show an increasing trend in the future, while those using saturation deficit (TCGI_SD) show a declining trend, though both give similar results in the historical period. This figures shows that the bification in the projected annual TC frequency is because of the changes in the seeding rate (numbers of precursors) while ratio ratio of storms that intensified from these precursors declines. TC frequency directly affect the projected TCs' tracks and the frequencies of strong storms, and thus this bification lead to large uncertainty in assessing regional and local storm hazards. While the cause may be different, the two diverging genesis scenarios we found is consistent the state of knowledge -- some studies show increasing TC frequency while others shows no changes or declining trend. We argue that such uncertainty is fundamental and epistemic in nature; we have no basis for assigning probabilities that one or the other simulation is correct. Changes in other aspects of TC climatology, such as increases in the frequency of major TCs, rapid intensification rate, and decreases in forward speed are insensitive to TC frequency. Regardless of the changes in frequency, however, our results are consistent with other studies in indicating that those TC events which do occur will, on average, be more destructive in the future because of the robustly projected increases in intensity.

STATISTICAL-DYNAMICAL DOWNSCALING MODEL for TROPICAL CYCLONG HAZARD ASSESSMENT

Although a TC is a highly nonlinear and chaotic system, we have learned that some aspects of it can be represented using linear models, for example, its intensification and inter-annual variability. Therefore, I have studied to what extent a low-order model, i.e., a statistical or hybrid statistical-dynamical model (with linear assumptions) can be used to understand the dependence of TCs on climate. I found that to capture TC's genesis, track, and intensity climatology, we need only a few essential environmental parameters. They are the potential intensity, 850 and 250 hPa steering flow, deep-layer mean vertical shear, low-level vorticity, and midlevel relative humidity. We then develop a new statistical-dynamical downscaling TC hazard model, called Columbia Hazard Model (CHAZ) that generates ensembles of synthetic storms whose property is based on the above parameters. This model, using 400 realizations of a 32-year period (approximately 3000 storms per realization) with environment conditions from ERA-Interim, captures many aspects of TC statistics, such as genesis and track density distribution. The figure here shows tracks color-coded with intensity from observations, and from one of the 400 realizations (labeled with GTI, the name of the experiment).

RAPID INTENSIFICATION AND TROPICAL CYCLONE LIFETIME MAXIMUM INTENSITY

The severity of a tropical cyclone (TC) is often summarized by its lifetime maximum intensity (LMI), and the climatological LMI distribution is a fundamental feature of the climate system. The distinctive bimodality of the LMI distribution means that major storms (LMI larger than 96 kt) are not very rare compared with less intense storms. Rapid intensification (RI) is the dramatic strengthening of a TC in a short time, and is notoriously difficult to forecast or simulate. Here we show that the bimodality of the LMI distribution reflects two types of storms: those that undergo RI during their lifetime (RI storms) and those that do not (non-RI storms). The vast majority (79%) of major storms are RI storms. Few non-RI storms (6%) become major storms. While the importance of RI has been recognized in weather forecasting, our results demonstrate that RI also plays a crucial role in the TC climatology.

SUB-SEASONAL TROPICAL CYCLONE GENESIS PREDICTION AND MJO

A seamless prediction approach for intraseasonal forecasts of TCs shows promise. For such forecasts, atmospheric models need to adequately simulate both TCs and the primary intraseasonal forcing - the Madden-Julian Oscillation (MJO) - as well as the link between them. As efforts to develop dynamical intraseasonal TC forecasts are still in the early stages, it is important to have a better understanding of how the TC-MJO relationship depends on model characteristics and lead time. The multi-model S2S dataset is ideal for this task, and is used to investigate Subseasonal probabilistic prediction of tropical cyclone (TC) genesis. Forecasts are produced for basin-wide TC occurrence at weekly temporal resolution. Forecast skill is measured using the Brier skill score relative to a no-skill climatological forecast: a seasonal climatology that varies monthly through the TC season. Skill depends on model characteristics, lead time, and ensemble prediction design. Most forecasts show skill for week one (days one to seven), when initialization is important. Among the six S2S models examined here, the European Centre for Medium-Range Weather Forecasts (ECMWF) model has the best performance, with skill in the Atlantic, western North Pacific, eastern North Pacific, and South Pacific at week two. The MJO modulates observed TC genesis, and there is a clear relationship, across models and lead times, between models' skill scores and their ability to represent the MJO--TC relation accurately. The figure above is a Candy plot for MJO–TC relationship in observations ECMWF from week 2 forecasts. This is a new way of looking at MJO modulation on the TCs. Color of each candy indicates the probability density function (PDF in percentage) in the corresponding MJO phase in the basin. Sum of the circles across the MJO phases in each basin is 100%. In the title of each subplot from simulations we label the r2, which represents the fraction of the observed pattern explained by the model simulation.

AIR-SEA INTERACTION IN TROPICAL CYCLONE