Sensitivity Study of Gonu Cyclonic Storm Intensity and Track to Surface Exchanges Parameterization: Advanced Hurricane WRF Model Application

Document Type : مقاله پژوهشی


University of Hormozgan


. Introduction
Tropical cyclones (TCs), all over the world, are recognized as the most common and most devastating of all the natural disasters. Nearly 80–90 TCs occur around the globe that about 40–50 attain intensity of 33 m/s (Frank and George, 2007). Approximately 7% of TCs originate in north of Indian Ocean and nearly 2% in Arabian Sea (WMO report, 2008).
The prediction of TC track has been on a clear path toward improvement for many years and nearly has been achieved (Elsberry, 2005). The global and regional simulations have proved that high resolution is not a requirement for improved track prediction (Goerss, 2006). The prediction of TC intensity is very complicated, because track prediction depends more on large-scale processes, and intensity depends on the inner-core dynamics and its relationship to the environment (Marks and Shay, 1998). It means that the intensity is a multiscale problem (Goerss, 2006). Only recently has the computational capability to address multiple scales of convection (cell scale, mesoscale, and synoptic scale) been achieved. The requirement to resolve the inner core, including the eyewall, the eye, and inner spiral rain bands near the eyewall, has led to the application of models with grid lengths of only a few kilometers (e.g., Krishnamurti, 2005; Braun et al., 2006; Chen, 2006; Davis et al., 2008).
Further the importance of resolution, other key issues for TCs prediction include the effect of mixing-induced ocean-surface cooling, treatment of momentum, heat and mixture fluxes at the air–sea interface, and improvement of the initial vortex structure through data assimilation. These Allegers will be each more important when we are moving toward finer spatial resolution in simulations.
In this study, sensitivity of Gonu simulations to air–sea fluxes parameterization and model resolution are studied as key issues for improving simulation accuracy. Herein, sensitivity to atmospheric physical parameterizations such as cloud physics and the boundary layer physics that affect Gonu intensity too, are not considered explicitly.
2. Study Area
The Indian Ocean is the third largest of the world's oceanic divisions, covering approximately 20% of the water on the Earth's surface. It is bounded by Asia—including India, after which the ocean is named—on the north, on the west by Africa, on the east by Australia, and on the south by the Southern Ocean (or, depending on definition, by Antarctica).
3. Material and Methods
Gonu was strong cyclone (categury 5) that was formed in the Arabian Sea during the periods of 1-7 June 2007 .At the peak intensity, cyclones Gonu was caused 3-min sustained wind about 240 km/h and minimum pressure about 920 mb. Financial losses of Gonu cyclone was reported around 4.2 billion USD and for human casualties at least 78 persons (IMD report, 2008). Herein, we have not planned to discuss about synoptical conditions of the Gonu cyclone formation, and only the effects of Air-Sea interactions on intensification is interpreted.
The performance of Advance Hurricane WRF (Advanced Research Weather Research and Forecasting) model (AHW model: Davis et al., 2008) is evaluated in explicit simulations of Guno super cyclone (2007). The data from National Centre of Environmental prediction (NCEP) global final analysis (FNL) on a 1.0°×1.0° grid is used to provide initial and boundary conditions for numerical simulations. Sea surface temperatures were derived from the high-resolution real-time global sea surface temperature (RTG_SST) at 1/12-degree resolution analyses from NCEP/MMAB. The information of reference track for Gonu cyclone has been received from Indian Meteorological Department (IMD). The mixed layer ocean model requires specification of an initial mixed layer depth h0, and a deep-layer lapse rate Γ. The ocean model was initialized with h0 and Γ equal to 50 m 0.14 in order to the NCEP Global Ocean Data Assimilation System (GODAS), when the negative feedback of wind-driven ocean mixing on hurricane intensity was planned to be considered (Davis et al., 2008).
In order to implement the Gonu simulations we considered two domain, the first domain was fixed with 27-km horizontal grid resolution (with 145×110 grid points) and second domain was nested movable domain with 9-km horizontal grid resolution (with 49×49 grid points) that was configured with a two-way nesting (fig. 1). All domains had 41 vertical layers with a terrain that followed sigma coordinates with the model top at 0.5 hPa.
The surface-flux parameterizations include of both momentum and enthalpy (heat and moisture) exchange. Two main schemes for momentum exchange are presented by Charnock and Donelan (Charnock, 1955; Donelan et al., 2005) that Charnock formulation is a default scheme in mesoscale meteorological models. In Charnock formulation, roughness length given as

Where and m and frictional velocity defined as

Where τ is Reynolds stress (shearing stress) and ρ is air density. The relation between roughness length and frictional velocity is recursive, but the model formulations use values from the previous time step and adjusted quickly.
In atmospheric boundary layer, drag coefficient is defined as

Where V10is the wind speed at 10-m height. The Charnock relation gives a 10-m drag coefficient that generally increases from about 0.001 to 0.003 at normal TC wind strengths, and it would recent to 0.005 for category 5 storms (wind >70 m s-1). However, this estimation for category 5 storms does not match with observational evidence (e.g., Black et al. 2007) that suggests drag coefficient remains near 0.003 for high wind speeds.
Donelan is introduced a drag formulation based on the high wind speed wind-tunnel experiments (Donelan et al. 2004) that results are showed Cdlower than those from the Charnock relation for low winds with a linear increase up to a maximum near 0.0024 at about 35 m s-1. In Donelan formulation, roughness length given as

With a lower and upper limit on z0 of 0.125×10-6 and 2.85 ×10-3 m, respectively. The Donelan formulation, with less drag than the Charnock formulation shows yourself effects with higher wind speeds but also higher central pressures and a larger eyewall radius in strong TCs simulations. This subject needs to validate with TCs observations over all the tropical.
The exchange coefficients of heat (Cɵ), moisture (Cq) and enthalpy (Ck) formulate by friction velocity, scaling temperature ɵ* and scaling moisture q*. These scaling parameters define the similarity theory profile. The available schemes for estimation of molecular viscosity sub-layer roughness length confirm an inverse relationship between roughness length and wind speed, and has the effect of a resistance to the eddy scalar transports that increases with wind speed. It means that the q* contribution to the surface moisture flux u*q* tends to oppose the effect of u* increasing with wind speed (similarly for heat and enthalpy). The coefficients of Cɵ, Cq and Ck formulate as

Where Δθ and Δq are difference in water vapor mixing ratio and potential temperature between the surface and 10-m. CP and Lv are heat capacity at constant pressure and specific latent heat of vaporization, respectively.
For example, estimated exchange coefficient of enthalpy (Ck)during the simulations are related to selected scheme for estimation of ɵ* and q*. The Ck will increase slowly in Carlson and Boland, stay steady in Large and Pond and decrease in Garratt parameterizations with increasing of wind speed (Carlson and Boland, 1978; Large and Pond, 1981; Garratt, 1992).
According above description, considering different schemes for different roughness lengths velocity, water vapor mixing ratio and potential temperature scaling in atmospheric surface layer over water would significantly affect estimated Ckand CD and also TC intensity. For example, the ratio of Ckto CDis introduced as an important factor in TC intensity (Emanuel, 1995; Braun and Tao, 2000; Davis et al. 2008).
4. Results and Discussion
In the first, the results of simulated intensity (minimum pressure and maximum wind speed) for coarse domain (Fig. 2) are analyzed. It is clearly showed that the sensitivity of simulations to the different three air–sea fluxes parameterization for simulations of Gonu at 27-km grid resolution at 0000 UTC 02 June 2007 (144-h model forecast). The Donelan-Large formulation, with less drag and steady Ckthan the Charnock-Carlson formulation that have higher drag and increasing Ckwith wind speed, results in higher wind speeds also lower central pressures. This happens after 48-h simulation where that wind speed increase until 78-knot. After this condition, high drag in the Charnock-Carlson formulation tends to oppose increasing wind speed. Comparison between Donelan-Large formulation and Donelan- Garratt simulation results with the same drag but different Ckin this interval shows Ck decreasing with wind speed that it tends to oppose wind speed increasing. Also Fig.2 shows steadily wind speed increase and decrease during first 48-h and last 48-h simulation for all three parameterizations for air–sea fluxes. The comparison of the sensitivity of simulated Gonu intensity with air–sea exchange parameterizations clearly showed that the Donelan-Large configuration has achieved significantly better simulated results at 27-km grid resolution but with 30-knot wind speed less than reported speed by IMD.
In order to study sensitivity of simulated Gonu intensity to horizontal grid resolution, it was conducted three experiments at 9-km grid resolution similar to last three experiments. The results were showed similar sensitivities at 27-km grid resolution but it was significantly improved the wind speed at the peak of cyclone intensity for these three air–sea exchange parameterizations (Fig. 3). Also these results showed that the central pressure simulated by Charnock-Carlson formulation, has been better compared to the 27-km resolution (as Donelan-Large formulation).
The results of simulated tracks were not significantly sensitive to these surface exchange parameterizations (not shown) but it showed more sensitivity to horizontal grid resolution (Fig.4). As can be seen in Fig. 4, the simulated track for 27-km resolution is better than the 9-km resolution that it consistent with the results of Goerss (2006). This proved that high resolution is not a requirement for improvement of track prediction.

5. Conclusions
In this study, it is seen a significant sensitivity of simulated intensity of one TC to air–sea exchange parameterization and horizontal grid resolution. For selected TC, the experiment with Donelan parameterization for momentum exchange and Large - Pond parameterization for heat and enthalpy are found more efficient. Also the results showed that simulated track could be more sensitive to horizontal grid resolution than applied air–sea fluxes parameterization. The difference in drag between the two drag formulations (Donelan and Charnock) is relatively smaller than reality. It must be pointed out that the real drag force on surface winds is determined by the time-evolving ocean wave spectrum, prediction of which requires a wave model (e.g., Chen et al. 2007). Therefore the drag parameterizations discussed above must be considered as crude representations of the bulk effects of waves in TCs. This sensitivity of differences in the three air–sea fluxes parameterization is less than would be inferred from the dependence of maximum wind on (Ck/Cd)1/2, derived by Emanuel (1995), While the proper wind speed dependence of Ck remains a topic of active research.


سالاری، س.، فراهانی، م.، آزادی، م. (1390). بررسی تغییرات هلیسیتی وشار هلیسیتی درتوفان استوایی گونو، مجله ژئوفیزیک ایران، شماره 115،صفحه97-115.
لشکری، ح.، کیخسروی، ق.، (1389).تحلیل سینوپتیکی توفان گنو و اثرات آن بر جنوب شرق ایران، مجله جغرافیا و برنامه ریزی محیطی، شماره 3، صفحه20-1.
محمدی، ع.، فرهادیپور، س.، شیخ بهایی، ع.، (1392). توابع پایداری و تاثیرآن برمشخصه‏های لایه مرزی اقیانوسی، موسسه ملی اقیانوس شناسی، شماره 5، صفحه 20-30.
مشهدی، ل.، حاجی زاده ذاکر، ن.، سلطانپور، م.، مقیمی، س.، (1392). شبیه سازی عددی امواج و مد ناشی از توفان گنو در خلیج چابهار، فصلنامه مهندسی دریا، شماره 17، صفحه 37-50.
ملکوتی، ح.، علیمحمدی، م.، (1391). شبیه سازی طوفان حاره ای گونو با استفاده از مدل Advanced Hurricane WRF:حساسیت به طراحی محدوده ها، آشیانه سازی، تفکیک افقی و زمان شروع، مجله علوم و فنون دریایی خرمشهر، شماره 15، صفحه 25-35.
صلاحی، ب.، عالی جهان، م.، (1392). تحلیل سینوپتیک مخاطرات اقلیمی شهرستان یاسوج، مطالعه موردی: بارش سنگین 20 اسفند 1389، مجله جغرافیا و مخاطرات محیطی، شماره 5، صفحه 73-89.
Black, P. G., D’Asaro, E.A., Drennan, W. M., French, J. R., Niiler, P. P., Sanford, T. B., Terrill, E. J., Walsh, E. J., Zhang, J. A., 2007. Air-sea exchange in hurricanes: Synthesis of observations from the Coupled Boundary Layer Air-Sea Transfer Experiment. Bulletin of the American Meteorological Society 88, 357–374.
Carlson, T. N., Boland, F. E., 1978. Analysis of urban-rural canopy using a surface heat flux/temperature model. Journal of Applied Meteorology 17, 998–1013.
Charnock, H., 1955. Wind stress on a water surface. Quarterly Journal of the Royal Meteorological Society 81, 639–640.
Chen, S. S., Price, J. F., Zhao, W., Donelan, M. A., Walsh, E. J., 2007. The CBLAST-Hurricane Program and the next-generation fully coupled atmosphere–wave–ocean models for hurricane research and prediction. Bulletin of the American Meteorological Society 88, 311–317.
Davis, C., Wang, W., Chen, S. S., Chen, Y., Corbosiero, K., DeMaria, M., Dudhia, J., Holland, G., Klemp, J., Michalakes, J., Reeves, H., Rotunno, R., Snyder, C., Xiao, Q., 2008. Prediction of landfalling hurricanes with the Advanced Hurricane WRF model. Monthly Weather Review 136, 1990–2005.
Donelan, M. A., Haus, B. K., Reul, N., Plant, W. J., Stiassnie, M., Graber, H. C., Brown, O. B., Saltzman, E. S., 2004. On the limiting aerodynamic roughness of the ocean in very strong winds. Geophysical Research Letters 31, L18306, doi:10.1029/ 2004GL019460.
Dudhia, J., Davis, C., Wang, W., Torn, R., 2011. HFIP Physics in AHW. Hurricane Tutorial NCAR/ NESL/ MMM, Apr 27 – 29, 2011.
Dudhia, J., 1989. Numerical study of convection observed during the winter monsoon experiment using a mesoscale two-dimensional model. Journal of the Atmospheric Sciences 46, 3077–3107.
Dudhia, J., Chen, S., 2004. A revised approach to ice microphysics process for the bulk parameterization of clouds and precipitation. Monthly Weather Review 132, 103–120.
Elsberry, R. L., 2005. Achievement of USWRP hurricane landfall research goal. Bulletin of the American Meteorological Society 86, 643–645.
Emanuel, K. A., 1995. Sensitivity of tropical cyclones to surface exchange coefficients and a revised steady-state model incorporating eye dynamics. Journal of the Atmospheric Sciences 52, 3969–3976.
Frank, W. M., George, S. Y., 2007. The interannual variability of tropical cyclones. Monthly Weather Review 135, 3587–3598.
Garratt, J. R., 1992. The Atmospheric Boundary Layer. Cambridge University Press, 316 pp.
Goerss, J. S., 2006. Prediction of tropical cyclone track forecast error for Hurricanes Katrina, Rita, and Wilma.Preprints, 27th Conf. on Hurricanes and Tropical Meteorology, Monterey, CA, American Meteorological Society 11A.1., 1-12.
Hong, S. Y., Dudhia, J., Chen, S. H., 2004. A revised approach to ice microphysical processes for the bulk parameterization of clouds and precipitation. Monthly Weather Review 132, 103–120.
Indian Meteorological Department report 2008. Report on Cyclonic Disturbances over North Indian Ocean during 2007. RSMC-Tropical Cyclone, New Delhi, 1-93.
Large, W. G., Pond, S., 1981. Open ocean momentum flux measurements in moderate to strong winds. Journal of Physical Oceanography 11, 324–336.
Lashkari, H., Kikhosravi, GH., 2010. Synoptic Analysis of Gonu storm and its effects on the East South of Iran. Journal of Geography and Environmental Planning 3, 1-20.
Malakooti, H., Alimohammadi, M., 2012. Simulation of Cyclone Gonu using the Advanced Hurricane WRF: Sensitivity to Domain Coverage, Nesting, Resolution and Starting Time. Journal of Marine Science and Technology, Khorramshahr 15, 25-35.
Marks, F. D., Shay, L. K. 1998. Landfalling tropical cyclones: Forecast problems and associated research opportunities. Bulletin of the American Meteorological Society, 79, 305–323.
Mashhadi, L., HajizadehZaker, N., Soltanpour, M., Moghimi, S., 1392. Numerical simulation of waves and tide induced Gonu storm in Chabahar Gulf. Journal of Marine Engineering 17, 37-50.
Mlawer, E. J., Taubman, S. J., Brown, P. D., Iacono, M. J., Clough, S. A., 1997. Radiative transfer for inhomogeneous atmosphere: RRTM, a validated correlated-k model for the long-wave. Journal of Geophysical Research 102, 16663–16682.
Mohammadi, A., Farhadipour, S., Shekhbahaei, A., 2013. Stability functions and its impact on the oceanic boundary layer characteristics. National Institute of Oceanography 5, 20-30.
Noh, Y., Cheon, W. G., Hong, S. Y., Raasch, S., 2003. Improvement of the k-profile model for the planetary boundary layer based on large eddy simulation data. Boundary-Layer Meteorology 107, 401–427.
Kain, J. S., 2004. The Kain–Fritsch Convective Parameterization: An Update. Journal of Applied Meteorology 43, 170–181.
Kanase, R. D., Salvekar, P. S., 2011. Numerical simulation of severe cyclonic storm LAILA (2010): Sensitivity to initial condition & cumulus parameterization scheme. Disaster Risk Vulnerablity Conference. Germany, March 12–14, 2011.
Salari, S., Farahani, M., Azadi, M., 2011. A study of helicity and helicity flux in the Gonu tropical storm. Iranian Journal of Geophysics 2, 97-115.
Salahi, B., Alyjhan, M., 2013. Synoptic analysis of climate risks in Yasouj city, Case Study: Heavy rain on 20th Esfand 1389, Journal of Geography and Environmental Hazards 5, 73-89.
World Meteorological Organization technical document, 2008. Tropical cyclone operational plan for the Bay of Bengal and the Arabian Sea. Document No. WMO/TDNo.84, 1-89.
54- Webb, R. H., Rathburn, S. L., 1988. Paleoflood hydrologic research in the soutwstern united states. Transportation Research Record 1201, 9-21.