Abstract

A positive impact of adding directional information to observations from the Cyclone Global Navigation Satellite System (CYNGSS) constellation of microsatellites is observed in simulation using a high-resolution nature run of an Atlantic hurricane for a 4-day period. Directional information is added using a two-dimensional variational analysis method (VAM) for near-surface vector winds that blends simulated CYGNSS wind speeds with an a priori background vector wind field at 6-h analysis times. The resulting wind vectors at CYGNSS data locations are more geophysically self-consistent when using high-resolution 6-h forecast backgrounds from a Hurricane Weather Research and Forecast Model (HWRF) control observing system simulation experiment (OSSE) compared to low-resolution 6-h forecasts from an associated Global Forecast System (GFS) model control OSSE. An important contributing factor is the large displacement error in the center of circulation in the GFS background wind fields that produces asymmetric circulations in the associated VAM analyses. Results of a limited OSSE indicate that CYGNSS winds reduce forecast error in hurricane intensity in 0–48-h forecasts compared to using no CYGNSS data. Assimilation of VAM-CYGNSS vector winds reduces maximum wind speed error by 2–5 kt (1 kt = 0.51 m s⁻¹) and reduces minimum central pressure error by 2–5 hPa. The improvement in forecast intensity is notably larger and more consistent than the reduction in track error. The assimilation of VAM-CYGNSS wind vectors constrains analyses of surface wind field structures during OSSE more effectively than wind speeds alone. Because of incomplete sampling and the limitations of the data assimilation system used, CYGNSS scalar winds produce unwanted wind/pressure imbalances and asymmetries more often than the assimilation of VAM-CYGNSS data.

Abstract

The NASA Cyclone Global Navigation Satellite System (CYGNSS) was launched in late 2016. It will make available frequent ocean surface wind speed observations throughout the life cycle of tropical storms and hurricanes. In this study, the impact of CYGNSS ocean surface winds on numerical simulations of a hurricane case is assessed with a research version of the Hurricane Weather Research and Forecasting Model and a Gridpoint Statistical Interpolation analysis system in a regional observing system simulation experiment framework. Two different methods for reducing the CYGNSS data volume were tested: one in which the winds were thinned and one in which the winds were superobbed.

The results suggest that assimilation of the CYGNSS winds has great potential to improve hurricane track and intensity simulations through improved representations of the surface wind fields, hurricane inner-core structures, and surface fluxes. The assimilation of the superobbed CYGNSS data seems to be more effective in improving hurricane track forecasts than thinning the data.

Abstract

The record-breaking U.S. tornado outbreaks in the spring of 2011 prompt the need to identify long-term climate signals that could potentially provide seasonal predictability for U.S. tornado outbreaks. This study uses both observations and model experiments to show that a positive phase TransNiño may be one such climate signal. Among the top 10 extreme outbreak years during 1950–2010, seven years including the top three are identified with a strongly positive phase TransNiño. The number of intense tornadoes in April–May is nearly doubled during the top 10 positive TransNiño years from that during 10 neutral years. TransNiño represents the evolution of tropical Pacific sea surface temperatures (SSTs) during the onset or decay phase of the El Niño–Southern Oscillation. A positive phase TransNiño is characterized by colder than normal SSTs in the central tropical Pacific and warmer than normal SSTs in the eastern tropical Pacific. Modeling experiments suggest that warmer than normal SSTs in the eastern tropical Pacific work constructively with colder than normal SSTs in the central tropical Pacific to force a strong and persistent teleconnection pattern that increases both the upper-level westerly and lower-level southwesterly over the central and eastern United States. These anomalous winds advect more cold and dry upper-level air from the high latitudes and more warm and moist lower-level air from the Gulf of Mexico converging into the east of the Rockies, and also increase both the lower-tropospheric (0–6 km) and lower-level (0–1 km) vertical wind shear values therein, thus providing large-scale atmospheric conditions conducive to intense tornado outbreaks over the United States.

Abstract

The ocean surface latent heat flux (LHF) plays an essential role in global energy and water cycle variability. In this study, monthly LHF over global oceans during 1992–93 are compared among Goddard Satellite-Based Surface Turbulent Fluxes, version 2 (GSSTF2), Hamburg Ocean–Atmosphere Parameters and Fluxes from Satellite Data (HOAPS), NCEP–NCAR reanalysis (NCEP), and da Silva et al. (da Silva). To find the causes for discrepancies of LHF, monthly 10-m wind speed (U _10m), 10-m specific humidity (Q _10m), and sea–air humidity difference (Q _S − Q _10m) are also compared during the same period. The mean differences, standard deviations of differences, and temporal correlation of these monthly variables over global oceans during 1992–93 between GSSTF2 and each of the other three datasets are analyzed. The large-scale patterns of the 2-yr-mean fields for these variables are similar among these four datasets, but significant quantitative differences are found.

The temporal correlation is higher in the northern extratropics than in the south for all variables, with the contrast being especially large for da Silva as a result of more missing ship observations in the south. The da Silva dataset has extremely low temporal correlation and large differences with GSSTF2 for all variables in the southern extratropics, indicating that da Silva hardly produces a realistic variability in these variables. The NCEP has extremely low temporal correlation (0.27) and large spatial variations of differences with GSSTF2 for Q _S − Q _10m in the Tropics, which causes the low correlation for LHF. Over the Tropics, the HOAPS mean LHF is significantly smaller than GSSTF2 by ∼31% (37 W m⁻²), whereas the other two datasets are comparable to GSSTF2. This is because the HOAPS has systematically smaller LHF than GSSTF2 in space, while the other two datasets have very large spatial variations of large positive and negative LHF differences with GSSTF2 to cancel and to produce smaller regional-mean differences. Based on comparison with high-quality flux observations, we conclude that the GSSTF2 latent heat flux, surface air humidity, and winds are likely to be more realistic than the other three flux datasets examined, although those of GSSTF2 are still subject to regional biases.

Abstract

Aircraft reconnaissance missions remain the primary means of collecting direct measurements of marine atmospheric conditions affecting tropical cyclone formation and evolution. The National Hurricane Center tasks the NOAA G-IV aircraft to sample environmental conditions that may impact the development of a tropical cyclone threatening to make landfall in the United States or its territories. These aircraft data are assimilated into deterministic models and used to produce real-time analyses and forecasts for a given tropical cyclone. Existing targeting techniques aim to optimize the use of reconnaissance observations and partially rely on regions of highest uncertainty in the Global Ensemble Forecast System. Evaluating the potential impact of various trade-offs in the targeting process is valuable for determining the ideal aircraft flight track for a prospective mission. AOML’s Hurricane Research Division has developed a system for performing regional observing system simulation experiments (OSSEs) to assess the potential impact of proposed observing systems on hurricane track and intensity forecasting. This study focuses on improving existing targeting methods by investigating the impact of proposed aircraft observing system designs through various sensitivity studies. G-IV dropsonde retrievals were simulated from a regional nature run, covering the life cycle of a rapidly intensifying Atlantic hurricane. Results from sensitivity studies provide insight into improvements for real-time operational synoptic surveillance targeting for hurricanes and tropical storms, where dropsondes released closer to the vortex–environment interface provide the largest impact on the track forecast. All dropsonde configurations provide a positive 2-day impact on intensity forecasts by improving the environmental conditions known to impact tropical cyclone intensity.

Abstract

In this study, a 72-h cloud-permitting numerical prediction of Hurricane Wilma (2005), covering its initial 18-h spinup, an 18-h rapid intensification (RI), and the subsequent 36-h weakening stage, is performed using the Weather Research Forecast Model (WRF) with the finest grid length of 1 km. The model prediction uses the initial and lateral boundary conditions, including the bogus vortex, that are identical to the Geophysical Fluid Dynamics Laboratory’s then-operational data, except for the time-independent sea surface temperature field. Results show that the WRF prediction compares favorably in many aspects to the best-track analysis, as well as satellite and reconnaissance flight-level observations. In particular, the model predicts an RI rate of more than 4 hPa h⁻¹ for an 18-h period, with the minimum central pressure of less than 889 hPa. Of significance is that the model captures a sequence of important inner-core structural variations associated with Wilma’s intensity changes, namely, from a partial eyewall open to the west prior to RI to a full eyewall at the onset of RI, rapid eyewall contraction during the initial spinup, the formation of double eyewalls with a wide moat area in between during the most intense stage, and the subsequent eyewall replacement leading to the weakening of Wilma. In addition, the model reproduces the boundary layer growth up to 750 hPa with an intense inversion layer above in the eye. Recognizing that a single case does not provide a rigorous test of the model predictability due to the stochastic nature of deep convection, results presented herein suggest that it is possible to improve forecasts of hurricane intensity and intensity changes, and especially RI, if the inner-core structural changes and storm size could be reasonably predicted in an operational setting using high-resolution cloud-permitting models with realistic initial conditions and model physical parameterizations.

Abstract

Monthly averages of daily latent heat fluxes over the oceans for February and August 1988 are estimated using a stability-dependent bulk scheme. Daily fluxes are computed from daily SSM/I (Special Sensor Microwave/Imager) wind speeds and EOF-retrieved SSM/I surface humidity, National Meteorological Center sea surface temperatures, and the European Centre for Medium-Range Weather Forecasts analyzed 2-m temperatures. Daily surface specific humidity (Q) is estimated from SSM/I precipitable water of total (W) and a 500-m bottom layer (W _B ) using an EOF (empirical orthogonal function) method. This method has six W-based categories of EOFs (independent of geographical locations) and is developed using 23 177 FGGE IIb humidity soundings over the global oceans. For 1200 FGGE IIb humidity soundings, the accuracy of EOF-retrieved Q is 0.75 g kg⁻¹ for the case without errors in W and W _B , and increases to 1.16 g kg⁻¹ for the case with errors in W and W _B . Compared to 342 collocated radiosonde observations, the EOF-retrieved SSM/I Q has an accuracy of 1.7 g kg⁻¹. The method improves upon the humidity retrieval of Liu and is competitive with that of Schulz et al.

The SSM/I surface humidity and latent heat fluxes of these two months agree reasonably well with those of COADS (Comprehensive Ocean–Atmosphere Data Set). Compared to the COADS, the sea–air humidity difference of SSM/I has a positive bias of approximately 1–3 g kg⁻¹ (an overestimation of flux) over the wintertime trade wind belts and wintertime extratropical oceans. In the summertime extratropical Pacific and summertime eastern equatorial Pacific Ocean, it has a negative bias of about 1–2 g kg⁻¹ (an underestimation of flux). The results further suggest that the two monthly flux estimates, computed from daily and monthly mean data, do not differ significantly over the oceans.

Abstract

Forecasting intensity changes in tropical cyclones (TCs) is a complex and challenging multiscale problem. While cloud-resolving numerical models using a horizontal grid resolution of 1–3 km are starting to show some skill in predicting the intensity changes in individual cases, it is not clear at this time what may be a reasonable horizontal resolution for forecasting TC intensity changes on a day-to-day-basis. The Experimental Hurricane Weather Research and Forecasting System (HWRFX) was used within an idealized framework to gain a fundamental understanding of the influence of horizontal grid resolution on the dynamics of TC vortex intensification in three dimensions. HWFRX is a version of the National Centers for Environmental Prediction (NCEP) Hurricane Weather Research and Forecasting (HWRF) model specifically adopted and developed jointly at NOAA’s Atlantic Oceanographic and Meteorological Laboratory (AOML) and Earth System Research Laboratory (ESRL) for studying the intensity change problem at a model grid resolution of about 3 km. Based on a series of numerical experiments at the current operating resolution of about 9 km and at a finer resolution of about 3 km, it was found that improved resolution had very little impact on the initial spinup of the vortex. An initial axisymmetric vortex with a maximum wind speed of 20 m s⁻¹ rapidly intensified to 50 m s⁻¹ within about 24 h in either case. During the spinup process, buoyancy appears to have had a pivotal influence on the formation of the warm core and the subsequent rapid intensification of the modeled vortex. The high-resolution simulation at 3 km produced updrafts as large as 48 m s⁻¹. However, these extreme events were rare, and this study indicated that these events may not contribute significantly to rapid deepening. Additionally, although the structure of the buoyant plumes may differ at 9- and 3-km resolution, interestingly, the axisymmetric structure of the simulated TCs exhibited major similarities. Specifically, the similarities included a deep inflow layer extending up to about 2 km in height with a tangentially averaged maximum inflow velocity of about 12–15 m s⁻¹, vertical updrafts with an average velocity of about 2 m s⁻¹, and a very strong outflow produced at both resolutions for a mature storm. It was also found in either case that the spinup of the primary circulation occurred not only due to the weak inflow above the boundary layer but also due to the convergence of vorticity within the boundary layer. Nevertheless, the mature phase of the storm’s evolution exhibited significantly different patterns of behavior at 9 and 3 km. While the minimum pressure at the end of 96 h was 934 hPa for the 9-km simulation, it was about 910 hPa for the 3-km run. The maximum tangential wind at that time showed a difference of about 10 m s⁻¹. Several sensitivity experiments related to the initial vortex intensity, initial radius of the maximum wind, and physics were performed. Based on ensembles of simulations, it appears that radial advection of the tangential wind and, consequently, radial flux of vorticity become important forcing terms in the momentum budget of the mature storm. Stronger convergence in the boundary layer leads to a larger transport of moisture fluxes and, subsequently, a stronger storm at higher resolution.

Abstract

In preparation for the launch of the NASA Cyclone Global Navigation Satellite System (CYGNSS), a variety of observing system simulation experiments (OSSEs) were conducted to develop, tune, and assess methods of assimilating these novel observations of ocean surface winds. From a highly detailed and realistic hurricane nature run (NR), CYGNSS winds were simulated with error characteristics that are expected to occur in reality. The OSSE system makes use of NOAA’s HWRF Model and GSI data assimilation system in a configuration that was operational in 2012. CYGNSS winds were assimilated as scalar wind speeds and as wind vectors determined by a variational analysis method (VAM). Both forms of wind information had positive impacts on the short-term HWRF forecasts, as shown by key storm and domain metrics. Data assimilation cycle intervals of 1, 3, and 6 h were tested, and the 3-h impacts were consistently best. One-day forecasts from CYGNSS VAM vector winds were the most dynamically consistent with the NR. The OSSEs have a number of limitations; the most noteworthy is that this is a case study, and static background error covariances were used.