of October 3-4, 1949 at Galveston and Freeport, Texas
(Excerpts from a study performed for the U.S. Army Coastal Engineering Research Center under contract with the U.S. Nuclear Regulatory Commission)
The 1949 Atlantic hurricane season was fairly active. Thirteen storms reached tropical storm strength. Out of a total of seven hurricanes that year, three developed into major hurricanes. Most destructive were a Category 4 hurricane that hit near West Palm Beach in August and the Category 4 hurricane that hit Freeport and Galveston, Texas on October 3 and 4, 1949.
The Texas coast has been struck by many hurricanes before and after 1949. The worse hurricane was the one that hit Galveston in 1900 and resulted in 6,000 - 8,00 deaths and the almost complete destruction of a large part of the city. The hurricane that struck Galveston and the Texas coastline on 3-4 October 1949 did not result in as many deaths, but nonetheless was very destructive. Storm surges caused extensive damage in Greens Bayou and at Matagorda. Two persons died in Freeport. Total damage was estimated at $6.7 million (1949 dollars).
Formation, Path and Landfall
This hurricane (unamed in 1949) formed in the Bay of Campeche, Mexico, from a tropical depression which was observed 2 or 3 days before 1 October, over Yucatan, Honduras, and Guatemala. During the night of 30 September and 1 October, the center of the depression moved into the Gulf of Mexico near Carmen, Mexico and increased to Category 4 hurricane intensity by 1045 hours, Central Standard Time (CST), on 2 October.
Track of the 1949 hurricane
The hurricane moved from Yucatan almost directly northward across the Gulf and its center crossed the Texas coast 22 miles southwest of Freeport during the nights of 3 and 4 October. The storm passed Houston in the early morning of 4 October. Winds were estimated at 135 mph, 5 miles west of Freeport (Zoch, 1949), and at 90 mph at Houston).
Hurricane Data - Parameters of the 1949 Hurricane
Surface wind fields and pressure fields for this hurricane were obtained from the U.S. Weather Bureau (HUR 7-37, 1957). The report provided hourly wind fields from 1800 hours CST, 3 October to 0500 hours CST, 5 October 1949. Other hurricane parameters were the following:
Central pressure 28.45 (inches of mercury).
Peripheal pressure 29.95 (inches of mercury)
Radius to maximum wind 15.0 (nautical miles)
Maximum gradient wind 88.0 (miles per hour)
Wind Isovels from October 2 to October 5, 1949 -Track of the 1949 hurricane and flooding due to hurricane surge (source: Pararas-Carayannis,1975)
Flooding by the Hurricane's Surges
The storm system generated the following surges along the Texas coast: 11 feet at Velasco, 8 feet at Matagorda, 9 feet at Anahuac, and 11.4 feet at Harrisburg in the Houston Ship Channel (Zoch, 1949), and 10.5 feet at Greens Bayou.
The Hurricane Moves Inland
Te storm's center moved across southeastern Texas, northwestern Louisiana, eastern Arkansas, southeastern Missouri, western Illinois, southeastern Wisconsin, and northern lower Michigan. The intense winds and rainfall associated with this storm caused heavy damage to crops and property in Texas, Louisiana and Arkansas (Seamon, 1949).
Past Hurricanes - The Hurricane of September 8-9, 1900 at Galveston
The historical record shows that from 1818 to 1885 at least twenty-eight hurricanes struck Texas. From 1885 through 1964, sixty-six tropical storms were recorded. Two-thirds of these were of hurricane force (with winds of more than seventy-four miles an hour). Two major hurricanes occurred from 1965 to 1970.
The deadliest hurricane in U.S. history and the worse to hit Texas was the one that struck Galveston on September 8-9, 1900. An estimated 6,000- 8,000 people purportedly lost their lives - however estimates have ranged as high as 12,000 deaths.
The track of the hurricane that struck Galveston in 1900
A tropical storm system developed In late August and early September 1900 and begun to move across the Caribbean Sea. The tropical storm passed the Florida Straits north of Cuba and appeared as though it would turn to a northeast direction. Instead, the storm continued in a northwest direction across the Gulf of Texas and subsequently strengthened into a hurricane on a path directly heading toward Galveston, Texas. The lowest barometric pressure recorded for the storm was 28.44 inches. However the National Oceanic and Atmospheric Administration (NOAA) later estimated the pressure near the storm’s center to have been closer to 27.49 inches.
When it made landfall at Galveston on Sept. 9, the hurricane had winds of 131-155 mph - and was later assigned a Category 4 designation. Flooding caused most of the damage. The hurricane surge turned into a wall of water over 15 feet high and struck the low lying island which had its highest elevation at 8.7 feet above sea level. The island was completely inundated. After the hurricane passed Galveston layed in total desolation. Property losses amounted to about $40 million (1900 dollars).
After landfall at Galveston, the storm system moved north through Texas and then into Oklahoma and Kansas. The remnants of the storm made it northeastward across the Great Lakes and into Canada before passing north of Halifax on Sept. 12 and disappearing into the North Atlantic.
Following the disaster an extensive levee system and a six-mile long seawall were constructed at Galveston to mitigate future impact of storms. Also, new codes were implemented for elevated building and for road foundations. These protective measures helped considerably when a hurricane hit Galveston again on August 16-19, 1915. Damage amounted to $50 million, but only 275 lives were lost because of the protective coastal works.
Mathematical Modeling Study of the 1949 Hurricane at Freeport and Galveston
A study of the hurricane of 3-4 October 1949 was undertaken by Pararas-Carayannis (1975) for the purpose of verifying a mathematical model for computing maximum hurricane surge flooding as it compared with the maximum surge height observed at Galveston and Freeport. The following hurricane parameters were used for the numerical calculations.
Central pressure - 28.45 (inches of mercury)
Peripheral pressure - 29.95 (inches of mercury)
Radius to maximum wind - 15.0 (nautical miles)
Maximum gradient wind -88.0 (miles per hour)
Track of the Hurricane of 3-4 October 1949. Storm surge (lower panel) and tide observations chart (hourly values only). Insert map for Galveston, Texas (source: Harris, 1963).
Surface wind fields and pressure fields for this hurricane were obtained from the U.S. Weather Bureau (HUR 7-37, 1957). The report provided hourly wind fields from 1800 hours CST, 3 October to 0500 hours CST, 4 October 1949. These were the synoptic weather charts used in the modeling study. The map below shows the traverses that were used for the calculation of the hurricane's surge. Other parameters of the Hurricane of 3 October 1949 that were used for the computations and the numerical model verification were the hydrographs of surges recorded by the Brazos River and the Galveston tide gauges. For the Freeport traverse the nearest tidal station which recorded the hurricane surge was at Brazos River Gates. The Brazos River station is located 1.2 nautical miles inland from the Gulf of Mexico, near Freeport, where the Brazos River and the Intracoastal Waterway intersect.
For the Galveston traverse, the nearest tidal station was located at Galveston's Pier 21, adjacent to the Galveston Channel. This station provided the best recorded data.
Other Computational Parameters
It is outside the scope of this report to present the details of the computations. However there were a number of corrections applied for water level datum and yjr phase of the predicted astronomical tide - in reference to MLW. The astronomical tide at the time of maximum surge inundation at the Brazos River tide gauge station was determined to be 1.42 feet above MLW and 1.02 feet at Pier 21 of the Galveston Channel.
Other hydrographic data of major importance in the calculation of the maximum storm surge involved an estimate of the initial rise in water level preceding the arrival of the hurricane. Such water level deviations can be as much as 2 feet.
Galveston and Freeport Traverses used in the computation of the 1949 hurricanee surge (source: Pararas-Carayannis, 1975).
Computational Traverses - Computational traverses, used for the surge calculations were chosen on the basis of their proximity to the hurricane track, to the recording tide gauges on relatively open coasts, and to points on the coast where maximum, well documented observations of surge water levels were available. The selected traverses were also situated, in both cases, to the right of the hurricane track - which also accommodated a limitation of the bathystrophic numerical model. For engineering design purposes, when evaluating maximum probable hurricane surge using a synthetic hurricane, the distance of about 1 to 3 times the radius to maximum winds away from the path of the hurricane center, is used in the computation in order to intercept the maximum hurricane effect. However, for the Hurricane of 3 October 1949 traverses were selected at Galveston and Freeport for which tide gauge records were available. The bearings of the traverses were established by a perpendicular orientation to the bathymetric contours between the tide gage stations on the open coast (the shore-intercept sites), or nearby points, and the 600-foot depth contour on the Continental Shelf. The bathymetry for the two profiles was taken from detailed nautical charts and integrated as input.
The chosen traverse intercept at Freeport had the following geographical coordinates: 28.55.5 N and 95.17.25 W. The azimuthal orientation of this traverse was S28E. The chosen traverse intercept at Galveston was at 29. 15.4 N and 94.78.45 W. The azimuthal orientation was S25E.
The initial rise used as input to the numerical calculation was the average difference between the predicted astronomical and the observed tide at or closest to the shore-intercept points of the traverse before the influencing effects of hurricane winds and pressures. This initial rise was estimated to be 2.0 feet at Freeport and at Galveston. In the numerical calculations this value was treated as a constant and added to the total water level. This may have been an oversimplification, since the cause for such initial rise or its exact magnitude during the passage of the storm is not known with certainty. This could also account for the small difference found between observed and computed maximum surge flooding.
Finally, values of relative wind stress k, and bottom friction, K, coefficients were applied to the numerical equations for surge computations. The constants K1 and K2 that were used account for pressure, density, precipitation, temperature and other factors of the hurricane system, which are not easily measured, and can only be obtained empirically.
Wind Data Reduction - As mentioned, prerequisite of accurate storm surge calculation is the availability of good wind field data along a selected computational traverse. Such good wind data did not exist in 1949. Most wind field data was provided in 3- and 6-hour weather charts; however, critical hurricane surge effects may be experienced at a coastline between time intervals for which wind field charts are available - so that was a limitation. Also, drastic changes in windspeed and direction can occur during these widely separated time intervals which may also render storm surge calculations invalid. Frequent wind information is desirable, at least along traverses where storm surge calculations are made. In th e case of the 1949 hurricane, as with other hurricanes for which ithere is not adequate date, an accurate interpolation procedure must he used to generate additional temporal and spacial wind information.
Unfortunately, in 1949, there were no synoptic measurements of hurricane winds or satellite imagery - as readily available at present - so this was a limitation in getting accurate results. Furthermore, as mentioned, the hurricane's wind-field charts were available for only 6-hour intervals and the highest surge occurred sometime between the times shown in these charts. Additionally, the interpolated wind field for that period that was used in the calculation probably did not accurately represent the actual wind conditions which produced maximum surge.
The data reduction was done graphically and by computer. However, both schemes also had their limitations. When a storm is moving slowly, the wind field data may be voluminous and graphical interpolation time consuming. When the storm is moving rapidly, linear interpolation may give inaccurate wind field data, especially in situations where abrupt wind direction reversals occur, and when the hurricane center crosses the coastline. In interpolating, the abrupt wind direction changes often require a subjective evaluation and interpretation, or very sophisticated computer programs. In this study, a graphical, nonlinear, data reduction method was used to interpolate windspeeds, W, valid wind vector directions, 0, for time intervals of 1 hour or less along each traverse. A computer subroutine was introduced in the numerical scheme to obtain the distances of the storm center, r, to points on the traverse. The methodology used for reduction of wind data is ommitted. The interested reader is referred to the full report (Pararas-Carayannis, 1975).
Windspeeds and Wind Vector Directions.
the following methodology was used for wind data reduction
and wind estimates along the traverses of
(1) Computational lines (traverses) for the 1949 hurricane were selected and drawn on a chart of the same scale as the wind charts in the same geographical area.
(2) Each traverse was plotted, or a transparency showing the traverse was superimposed, on all appropriate weather charts of the same scale which covered the chosen time period for which calculations of surge were to he made, for this particular hurricane of 1949. Weather charts with contours of windspeeds in knots (30-foot surface isotachs or isove1s), and vectors of wind directions, were used.
(3) Points along the traverse were selected for calculations at different time stages of the storm. Wind vector directions and speeds were obtained from each weather chart for a time interval starting at the seaward point of the traverse and at a subsequent point on the traverse. This was done at the earliest phase of the storm for which data was available, when the windspeeds at the seaward point exceeded the critical value of 14 knots.
W, were obtained by interpolation between isovels and converted
from knots to miles per hour (1 knot = l.I5 miles per hour).
Wind direction angles, 8, (the angles between traverse line
and wind direction) were obtained by averaging two or more
wind vectors on either side of the traverse and these measured
counterclockwise from the traverse line from 0 to 360 degrees.
Also, the initial distance values, rLM' of the storm center to the traverse shore-intercept were obtained in nautical miles for each position of the storm given on the weather charts. Subsequent values of distances r to all other points on the traverse were developed internally by the computer program, according to the method described in the full report (Pararas-Carayannis, 1975). This data reduction procedure was repeated for each time step in the development of the storm (using a 6·, 3-, or 1-hour weather charts). The process of interpolation was subjective, but attention was given in developing the data in a reasonable fashion.
Hydrographic Data Reduction - For the purpose of this study, corrections and adjustments were applied to the hydrographic data to account for tide gage locations, tide phases, water level datums of tide gages, and initial rise preceding the arrival of the hurricane.
Sea Level Datum Corrections - Some recorded or observed surge data associated with the hurricane of 1949 were referenced to MSL. Other tide gage data referred to MLW, or SLD. Therefore, corrections were applied for differences in the water level datum. Also, corrections were applied for tide gage datums.
Rise - Other
hydrographic data of major importance in the calculation
of the storm surge with the numerical model involved
the initial rise in water level preceding the arrival of
the 1949 hurricane. At such times, water levels along a
coast have been observed to be different than the predicted
tide. This difference in water levels between observed
and predicted tides appears to be in many cases, independent
of the storm or the astronomical forces. Examination
of prestorm tide gage records (from which initial rise
was obtained) indicates that a considerable range of values
may be selected. Harris (1963) has shown that such water
level deviations can be as much as 2 feet..
The initial rise used as input to the numerical calculation of storm surge is usually the average difference between predicted astronomical and observed tides at a tide gage station at or closest to the shore-intercept (open coast) of the traverse before the influencing effects of hurricane winds and pressures. In the numerical calculation, this value was treated as a constant and added to the total water level. This may be an oversimplification, since the cause for such rise or its exact magnitude during the passage of the storm is not known with certainty. For the purpose of calibrating the numerical model with the 1949 hurricane data, the initial rise was defined as "the average water level variation above the predicted astronomical tide at a station during the 2 days preceding the occurrence of a 15 to 20 miles per hour isovel of the advancing hurricane across the Continental Shelf." Based on this definition, the initial rise in water levels used in the calculation of surges at Galveston and Freeport, were estimated using the predicted tides, taken from the tide tables and the observed or recorded hydrographs of water levels at the nearest tide gauge stations. However, because of these limitations, the initial rises in water levels used in this study represented rough approximations.
Observed and computed surge hydrographs at Freeport, Texas, for the hurricane of 1949 with datum corrections and optimal bottom frictional coefficient and surface wind stress (Pararas-Carayannis, 1975).
The predicted tides for the various locations at or near the traverses used in the calibration were obtained from tide tables which are referenced to MLW. The tide tables give maximum and minimum heights of tides and the times of respective occurrences. Height values were plotted and smooth curves fitted to obtain continuous values of astronomical tides. These tide curves represented the projected astronomical tides used for estimating the initial rise and for computing the surge hydrographs. All reported water level values refer to MLW datum.
Results of the Computation
In spite of the above stated data limitations of the 1949 hurricane, there was adequate correlation between the computed hurricane surge hydrographs and the observed high water elevations. The adjacent figure illustrates a fairly good correlation in time and height of the 1949 hurricane surge at Freeport. Similarly good correlation was obtained for the Galveston traverse.
With good synoptic wind data, satellite imagery, better determination of hurricane parameters and modern fast computers, present mathematical models can estimate the maximum hurricane surge with greater accuracy.
REFERENCES AND ADDITIONAL READING
BODINE, B. R., "Storm Surge on the Open Coast: Fundamentals and Simplified Prediction," TM-35, U.S. Army, Corps of Engineers, Coastal Engineering Research Center, Washington, D.C., May 1971.
BRETSCHNEIDER, C. L., and COLLINS, J. I., "Prediction of Hurricane Surge; An Investigation for Corpus Christi, Texas and Vicinity," NESCO Technical Report No. SN-120, prepared by National Engineering Science Company for U.S. Army Engineer District, Galveston, Tex., 1963.
FREEMAN, J. C., Jr., BAER, L., and JUNG, C. H., 'The Bathystrophic Storm Tide," Journal of Marine Research, Vol. 16, N9. 1, 1957, pp. 12-22.
H. E., and HUDSON, G. N., "Surface Winds Near the
Center of Hurricanes (and other Cyclones)," NHRP
Report No. 39, U.S. Weather Bureau, Washington, D.C.,
GRAHAM, H. E., and NUNN, D. E., "Meteorological Considerations Pertinent to Standard Project Hurricane, Atlantic and Gulf Coast of the United States," NHRP Report No. 33, U.S. Weather Bureau, Washington, D.C., Nov. 1959.
HARRIS, D. L., "Characteristics of the Hurricane Storm Surge," Technical Paper No. 48, U.S. Weather Bureau, Washington, D.C., 1963.
HARRIS, D. L., "A Critical Survey of the Storm Surge Protection Problem," The 11th Pacific Science Congress Symposium on Tsunami and Storm Surges, Mar. 1967, pp. 47-65.
JELESNIANSKI, C. P., "Numerical Computations of Storm Surges with Bottom Stress," Monthly Weather Review, Vol. 95, No. 11, 1967, pp. 740~756.
JELESNIANSKI, C. P., "Bottom Stress Time-History in Linearized Equations of Motion of Storm Surges," Monthly Weather Review, Vol. 98, No.6, 1970, pp. 462-478.
MARINOS, G., and WOODWARD, J. W., "Estimation of Hurricane Surge Hydrographs," Journal of the Waterways and Harbors Division, ASCE, Vol. 94, No. WW2, May 1968, pp. 189-216.
PARARAS-CARAYANNIS, G. 1975, Verification Study of a Bathystrophic Storm Surge Model. U.S. Army, Corps of Engineers Coastal Engineering Research Center, Washington, D.C., Technical Memorandum No. 50, May 1975.
PERDIKES, H. S., "Hurricane Flood Protection in the United States," Journal of the Waterwa)'s, Harbors and Coastal Engineering Division, ASeE, Paper No. 5081, Vol. 93, No. WWl,Feb. 1967,p.2.
REDFIELD, A. C., and MILLER, A. R., "Water Levels Accompanying Atlantic Coast Hurricanes," Meteorological Monographs, American Meteorological Society, Vol. 2, No. 10, June 1957, pp. 1-23.
REID, R. 0., "Short Course on Storm Surge," Lectures, Texas A&M University, College Station, Tex., 1964.
REID, R. 0., and BODINE, B. R., "Numerical Model for Storm Surges Galveston Bay," Journal of the Waterways and Harbors Division, ASCE, Vol. 94, No. WWl, Procedures Paper 5805, 1968, pp. 33-57.
SEAMON, L. H., "The Weather of 1949 in the United States," Monthly Weather Review, Vol. 77, No. 12, Dec. 1949, pp. 325-333.
U.S. ARMY, CORPS OF ENGINEERS, COASTAL ENGINEERING RESEARCH CENTER, Shore Protection Manual, Vols. I, IT, and III, Stock No. 0822-00077, U.S. Government Printing Office, Washington, D.C., 1973, 1,160 pp
U.S. WEATHER BUREAU, "Wind Speeds and Directions Over the Gulf of Mexico During the Hurricane of 3 October, 1949," Memorandum HUR 7-37, Washington, D.C., Aug. 1957.
U.S. WEATHER BUREAU, "Interim Report-Meteorological Characteristics of the Probable Maximum Hurricane, Atlantic and Gulf Coasts of the United States,"
Hydrometeorological Memorandum HUR 7-97, Washington, D.C., May 1968.
WELANDER, 0." "Numerical Prediction of Storm Surges," Advances in Geophysics, Vol. 8, 1961, pp. 316-379.
ZOCH, R. T., "North Atlantic Hurricanes and Tropical Disturbances of 1949," Monthly Weather Review, Vol. 77, No. 12, Dec. 1949, pp. 339-341.
Web Design by Dr. Carolyn Carayannis © Copyright 2008 / all rights reserved. © Copyright 1963-2007 George Pararas-Carayannis / all rights reserved / Information on this site is for viewing and personal information only - protected by copyright. Any unauthorized use or reproduction of material from this site without written permission is prohibited. Material included at the website links above is for informative and educational purposes and for disaster preparedness only. Any predictions of large earthquakes, destructive tsunamis, or any other natural disasters presented in these pages are based primarily on statistical determinations of the historical recurrence frequencies of such events. Such historical/statistical approaches are used only for long-term predictions. There is no intent by the author to predict or forecast any type of natural disaster or to frighten people.