sign in sign up
<<
Home , Brazos River, Lampasas River, Leon River, Little River (Oberon), Little River (Texas), Salado Creek

DEPARTMENT OF THE INTERIOR ALBERT B. FALL, Secretary GEOLOGICAL SURVEY GEORGE OTIS SMITH, Director

Water-Supply Paper 488

THE FLOODS IN CENTRAL IN SEPTEMBER, 1921

BY

C. E. ELLSWORTH

Prepared in cooperation with THE STATE OF TE?A8

WASHINGTON 0OYERNMENT PRINTING OFFICE 1923 ADDITIONAL COPIES OF THIS PUBLICATION MAT BE PROCURED FKOM ' THE SUPERINTENDENT OF DOCUMENTS GOVEKNMENT PRINTING OFFICE WASHINGTON, D. C. AT 15 CENTS PER COPY

PURCHASER AGREES NOT TO RESELL OR DISTRIBUTE 1ES3 COPY FOR PROFIT. P.UB. RES. 67, APPROVED MAT 11, 1923 V CONTENTS.

Page. Introduction______1 Acknowledgments______2 Area covered by the storm______3 General features ______3 Topography and geology-______3 Damage caused by flood ___ 4 Precipitation ______, ______7 Maximum discharge per square mile _ : _ _ 13 Flood flows______-__-_ - _ 14 Brazos flood______14 General features______14 Determination of flow ______16 and tributaries______17 General features______17 Determination of flow______18 Little River below junction .of Leon and Lampasas 19 Little River at Cameron______20 near Salado______: 22 San Gabriel River at Georgetown______24 Brushy Creek at Round Rock______27 flood______29 flood______31 The flood at San Antonio______33 General features______33 Determination of discharge______37 Previous floods______46 General features of flood of 1913______46 Brazos River______48 Colorado River______,___ 49 Guadalupe River- ______51 River ______53 Index______55 in IV CONTENTS.

ILLUSTRATIONS.

Pas* PIATE I. Map of , showing drainage basins and total rain­ fall -11, 1921______4 II. A., View of Little River valley below junction of Leon and Lampasas rivers; B, View of Little River at Gulf, Colorado & Santa Fe Railway crossing near Cameron, Tex., at crest of flood ______20 III. A, View across Little River valley near Cameron, Tex., from Marlow Bridge; J5, View up Salado Creek from Jones mill, near Salado, Tex______20 IV. A, View down San Gabriel River valley below Georgetown, Tex.; B, View up Brushy Creek near Round Rock, Tex______20 V. A, Bridge across San Antonio Ri3ver near Pioneer Mills after flood of -10, 1921; B, Bridge across at Romana Street, San Antonio, Tex., after flood of September 9-10, 1921______^__ 44 VI. A, Water-stage recorder at South Alamo Street, San Antonio, Tex.; B, South Alamo Street Bridge, San Antonio, Tex., after flood of September 9-10, 1921______44 VII. Map of San Antonio, Tex., showing area flooded, September 9-10, 1921 ______44 VIII. A, View of San Antonio River showing slope stretch used in determination of maximum discharge during flood of Sep­ tember 9-10, 1921; B, View from right bank near center of slope stretch shown in A______44 . Map of San Antonio River and tributaries in vicinity of San Antonio, Tex., showing drainage basins and total rainfall, -10, 1921______^__.___._ 8 2. Gage-height graph of Little River at Cameron, Tex., during flood of September, 1921______21 3. Gage-height graph of San Antonio River and San Pedro Greek at San Antonio Tex., during flood of September 9-10, 1921 40 4. Rating curves of San Antonio River at South Alamo Street Bridge, San Antonio, Tex______42 5. Mean cross section of San Antonio River used in determination of discharge at maximum stage of flood of September 9-10, 1921______.______-_ 44 IV THE FLOODS IN CENTKAL TEXAS IN SEPTEMBER 1921.

By C. E. ELLSWORTH.

INTRODUCTION. Heavy rainfall over a large area in south-central Texas from Sep­ tember 8 to , 1921, produced great floods which caused the loss of at least 224 lives and damage to property amounting to more than $10,000,000. The most destructive flood in Texas prior to 1921 occurred in December, 1913. That flood, which is described on pages 46-47, caused the loss of 177 lives and destroyed prop­ erty valued at more than $8,500,000. The most densely populated and most highly developed com­ munity affected by the flood of 1921 was the city of San Antonio, where 51 lives were lost and property worth more than $3,000,000 was destroyed. It was the disaster at San^Antonio that received the most widespread notice in the press, though the aggregate loss of both life and property in other areas far exceeded that at San Antonio. The total rainfall and the run-off per square mile of drainage area were much less in the basin of San Antonio River than in the basins of many other streams in the path of the storm. If the rainfall in the basin of San Antonio River had been as heavy as it was in much of the basin of Little River, in Bell, Milam, and Williamson counties, the destruction at San Antonio would have been so great as to make that actually suffered tKere seem insignificant. The United States Geological Survey, in cooperation with the Texas Board of "Water Engineers, made special determination of the flood flow of San Antonio River at San Antonio and of Little River at Cameron immediately after the flood. Several months later, determinations were made of the maximum discharge of Lit­ tle River near the junction of Leon and Lampasas rivers, of Salado Creek near Salado, of San Gabriel River near Georgetown, and of Brushy Creek at Round Rock. The amount available for the work was so small that measurements were made only on streams where the conditions were favorable for obtaining reasonably accurate 1 2 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. results. In addition to the-measurements made on the streams men­ tioned above determinations of the flood flow at several regular gaging stations were made by routine methods. In the area north and east of Taylor, where the total rainfall exceeded 36 inches, the flood was so wide and was so much obstructed in its course down the valleys of San Gabriel River and Brushy Creek that the maximum discharge of those streams could not be determined with sufficient accuracy to make the determination valuable. The average annual losses in Texas by floods amount to several million dollars, and the people of the State are beginning to realize that its future prosperity must depend in large measure upon the wisdom with which they can control and utilize the streams. Over 30,000,000 acre-feet of water annually passes unutilized from the streams of Texas to the Qulf of , much of it in floods that cause great destruction. Good business sense demands that the floods of Texas be controlled and that the flood water be stored, so far as practicable, for the many uses for which it is needed. The Board of Water Engineers of Texas has planned a compre­ hensive study of the streams of the State and of the best methods of controlling floods and utilizing flood water. The report here presented has been compiled for the purpose of assembling under one cover all the essential information now available concerning the flood of 1921 and of giving a brief summary of the information available concerning previous floods. The report is in many respects incomplete, but it should be a valuable contribution to the body of information on floods that must eventually be compiled.

ACKNOWLEDGMENTS. The steam-gaging work done in Texas by the United States Geo­ logical Survey is carried on in cooperation with the State Board of Water Engineers, consisting of John A. Norris, chairman; C. S. Clark, and A. H. Dunlap. This board allots a large part of the funds available for such work. Acknowledgments are due the United States Weather Bureau for practically all the data on rainfall, many gage heights, and much in­ formation about the storm and its effect; to Col. Edgar Jadwin and Maj. W. N. Vance, of the United States Engineer Corps at Fort Sam , for their kindness in giving the writer access to their excellent unpublished report of the San Antonio flood; to C. Terrell Bartlett, for information that has been quoted freely from his published statements in engineering periodicals and local news­ papers regarding the flood at San Antonio; and to many county officials, engineers, railroad companies, residents, and the public press for statements of losses and other information which could AREA COVERED BY THE STOBM. O not have been obtained without their cooperation. Individual ac­ knowledgments are given throughout the report. In determining discharge of streams during the flood the .writer was assisted by C. E. McCashin, D. A. Dudley, R. G. West, and Trigg Twitchell. The office computations of discharge were made principally by C. E. McCashin and Trigg Twitchell, and assistance in assembling the matter and preparing it for publication was ren­ dered by Kate Casparis.

AREA COVERED BY THE STORM.

GENEBAL FEATURES. % The rainfall that caused the flood covered an elliptical area that includes nearly 10,000 square miles, whose longer axis extends southwestward along the Balcones escarpment from Temple beyond San Antonio, a distance of more than 150 miles. (See PI. I.) The larger streams that cross this area, all flowing southeastward, are Brazos, Colorado, Guadalupe, and San Antonio rivers. The counties within the storm area are Bell, Milam, Williamson, Travis, Hays, Comal, and Bexar. These counties contain much of the richest agricultural land in Texas. Practically all the area lying southeast of a line extending from Temple through Belton, George­ town, Austin, San Marcos, New Braunfels, and San Antonio is under cultivation. As the soil is fertile and the conditions for farming are otherwise favorable, this area is one of the most densely populated and highly developed in the State. Much of the land northeast of that line is also under cultivation, but to a less extent than that to the southeast, because the soil is thin and less fertile and the coun­ try is more broken and hilly. A large percentage of the land in both areas that is not under cultivation is covered with a fairly thick growth of post oak and stunted cedar, such as the thin soil is capable of supporting.

TOPOGBAPHY AND GEOLOGY, The axis of the area of maximum rainfall followed closely the Balcones escarpment, which is the dividing line between the Plain, on the southeast, and the and Grand Prairie region, on the northwest. The Balcones escarpment is the result of an enormous geologic fault that extends across the entire State from Red River to the Rio Grande. It marks a sud­ den change of topography from the gently sloping Rio Grande Plain, which stands at an elevation of about 500 feet, to the eroded plateaus, which range in elevation from 1,000 to 2,000 feet. In passing from the plateau to the plains the streams have cut deep channels, which have steeper grades than those on either side of the 4 FLOODS IN CENTBAL TEXAS IN SEPTEMBER, 1921. escarpment. The prevailing rock is limestone of the Cretaceous period. It is overlain principally with clay soil. In the area northeast of the fault zone the soil is thin and considerable areas of rock are exposed, but in the area southeast of it the soil is deeper and rock is seen at few places except along the banks of some of the streams.

Except for the loss in the city of San Antonio, the greater part of the total loss occurred in the rural districts, where crops were de­ stroyed and roads and bridges were washed out. Owing to con­ flicting opinions concerning the amount and the value of the crops destroyed it is impossible to determine accurately the damage caused in the rural districts. One item of loss for which no estimate is here given, and which can not be determined with reasonable accuracy, is the loss of soil from farm lands by erosion or by being covered with deposits of sand and gravel. Such losses undoubtedly aggregate millions of dollars. Walter E. Davis, agricultural demonstration agent for Travis County, stated that the loss of soil was much greater, than the loss of crops but that he could not estimate the total loss for the county. The loss of soil in Travis County was probably much greater than in any other, because much of the cultivated land in this county has a steeper slope than the average in the rest of the area covered by the storm. The damage was due largely to erosion by the rains that fell directly upon the land. The water attained scouring velocities before it left the fields and entered the usual stream channels. The crops, principally cotton and corn, were washed away with the soil; the devastated area ranged from narrow strips in some places to wide areas in others. Mr. T. S. Etenderson, of Cameron, estimated that more than 75,000 acres of farm land in Milam County was flooded by the over­ flow from Little River and its tributary, San Gabriel River. More than half of the area overflowed had been planted in corn and cotton, which were practically a total loss. Cultivated land cover­ ing thousands of acres that lay above the area flooded by overflow from the river was also greatly damaged by the torrential rain falling directly upon it or by the scouring action of the water as it flowed over the fields. The total loss of crops was at least $1,000,000. The loss of live stock was $60,000, and the damage to houses, fences, and miscellaneous farm structures was $20,000. Mr. Giles L. Avriett, county auditor, estimates the loss and damage to bridges and highways at $100,000, about equally divided between the two items. U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 488 PLATE I

_ _ _.___]_ _ N EXPLANATION Line of equal rainfall in inches -i Boundary of

MAP OF CENTRAL TEXAS, SHOWING DRAINAGE BASINS AND TOTAL RAINFALL SEPTEMBER 7-11,1921.

DAMAGE CAUSED BY FLOOD. & Though numerous letters requesting information on the flood were written to county officials, engineergj and others, only % few replies were received. Much information was obtained by personal interviews and from local newspapers. The statements in this report are based on a careful study of all information available. A table giving in detail the damages to property in San Antonio was furnished .by the United States Engineer Corps of Fort . The figures in this table were obtained by personal ex­ amination of all flooded areas and are undoubtedly the most com­ plete and accurate that have been compiled. They show that the total damage in San Antonio amounted to $3,745,200, a sum which, though enormous, is less than that generally reported immediately after the flood. At that time the loss was reported to be between $5,000,000 and $10,000,000. The total number of lives lost will never be known, but the best estimates available indicate that at least 224 people were drowned, most of whom were Mexicans who lived in poorly constructed houses, built along the low banks of the streams. Undoubtedly many others were drowned who were never reported missing. Many bodies were carried miles and buried in sand, mu4, and debris along the river bottoms. Along Little and San Gabriel rivers bodies were .found six months or more after the flood. A total of 52 lives were lost in San Antonio and vicinity, as follows: On San Pedro Creek, near South Flores and Mitchell streets, 15; on San Pedro Creek between West Commerce Street ajid the mouth of Alazan Creek, 10; on Alazan Creek between West Commerce Street and the mouth of San JPedro Creek, 20; on Apache Creek between Elmendorf Lake and South Brazos Stinpt, near Tampico Street, 3; on San Antonio River at Newell Avenge, four blocks south of Breckenridge Park, 3; and at San Antonio Kiver crossing, south of San Jose Mission, in the southern suburbs of: the city, 1. The greatest loss of life in Williamson and Milam counties oc­ curred on Little Kiver and' its chief tributary, San Gabriel River and its branches. At least 159 bodies were recovered, and undoubt­ edly many others were never found. Maynard JRobinson, division .master mechanic, and F. W. Leatherbury, division engineer, of the Qulf, Colorado & Santa Fe Railway, were drowned near Cameron by the capsizing of a boat from which they were inspecting the- railroad bridge over Little River. At or near Tfyorndale and Rock- dale 45 bodies were recovered; at Lanesport, on San Gabriel River near the east line of Williamson County, 23; at San Gabriel, 10; on Alligator Creek, 5; at Robyland farm, near Taylor, 4; at Elm Grove, 7 miles southeast of Taylor, 2*9; in immediate vicinity of Tay- ' 19526 23 2 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. lor, 6; at Lawrence Creek, 2; at Redville gin, 2. Along the bottoms between Gause and Valley Junction 20 negroes were drowned, 7 of them in Robertson County. In Travis County 6 ne­ groes were drowned on the farm of George Begg near the mouth of Onion Creek. The following table summarizes all available information con­ cerning the loss of life and the damage to property in central Texas caused by the flood of September, 1921: Loss of life and property caused by floods in central Texas in September, 1921.

Lives Bridges Other County. lost. Crops. and roads. items. Total. s

$5,000 $5,000 6,000 $2,000 8,000 Bell...... 600,000 225,000 $85,000 910,000 Bexar°...... 52 3,245,700 2,000 2,000 250,000 50,000 250,000 550,000 750,000 75,000 20,000 845.000 70.000 2,000 77,000 15,000 1,000 1,000 17,000 2,500 1,000 3,500 150,000 1,000 151 000 20,000 2,000 22,000 20,000 50000 50,000 120; ooo Milam...... 66 1,000,00 100,000 100,000 1,200.000 7 200,000 1,000 5,000 206000 6 600,000 100,000 30,000 730,000 Waller...... 37.500 2,500 2,500 42,500 30,000 1,000 4,000 35,000 93 640,000 400,000 200,000 1,240,000 Wilson...... 15,000 15,000 Damage to railroads: 340,300 187,220 72,500 58,383 56,750 10,139,853

oPractieaHy the entire loss of life and damage to property in Bexar County occurred within the city limits of San Antonio. See following table for itemized statement. Losses in San Antonio, Tex., caused by flood of September, 1921" Railroads.

Bridges Roadway Traffic and ana and transpor­ Total. structures. track. tation. &

$600 $800 $100 $1,500 * 65,000 <*) 500 65,500 Missouri, Kansas & Texas...... "7,700 <*) 200 7,900 1.400 400 200 2,00»» KTacrHo-lViln 76,900

a This table was taken from report by U. S. Engineer Corps at . 6 Includes costs of detouring and delay to trains. « Revised on basis of statements furnished by railroad companies.

Losses in San Antonio, Tex., caused &j/ flood of September, 1921 Continued. Municipal property.

Square Cost of Kind. yards replace­ ment.

Wood-block pavement...... 80,000 $240,000 Asphalt pavement...... 150 525 Brick pavement...... 100 300 Macadam pavement...... 12,000 18,000 Bridges...... 125,000 Public buildings...... 10,000 Parks, including driveways, bridges, buildings, and all improvements...... 20,000 Fire department equipment...... 4,175 418,000

Public-service equipment.

Company. Damage. ^il^1116 Total.

Water company-...... $3,000 $4,000 $7,000 Southwestern Bell Telephone Co...... 63,000 Public Service Co.: Street railway...... 15,000 55,000 1 Electricity...... 44,000 30,000 ^ 150,000 Gas plant...... 2,000 4,000 J ...... 220,000

Real- estate, personal property, and merchandise. Residences and contents (75 per cent loss): 200 valued at $4,000 each, contents valued at $1,500 each...... $325,000 1,000 valued at $1,000 each, contents valued at $250 each...... 937,500 Motor cars, accessories, stock, etc., owned by dealers (allowance made for salvage)...... 75,000 Motor cars, privately owned (allowancemade for salvage)...... ~...... 50,000 Realty improvements, fences, yard ornaments, flowers, trees, etc., particularly residential prop­ erty...... 50,000 Merchandise...... 423,000 Realty losses (personal effects, money, etc., not included above)...... - 50,000 2,410,500 *" United States property.

Damaged. Destroyed. Total.

Camp Travis (Salado Creek)...... $100 $100 San Antonio Arsenal...... $20,000 100,000 120,000 Personal property...... 200 200 ...... 120,300

Recapitulation. Railroads in San Antonio...... $76,90 Municipal property...... 418,00 Public-service equipment...... 220,00 Real estate, personal property, and merchandise...... 2,410, 50 United States property...... 120,30 Total...... 3,245,700 PRECIPITATION. In the storm of September 9 and 10, which originated in an area of low pressure in the and passed inland to central Texas, the rainfall was the greatest recorded in the United States for an equal length of time. Plate I shows the distribution 8 FLOODS IF CESTTBAL TEXAS 1ST SEPTEMBER, 1921. of the rainfall as recorded at United States Weather Bureau sta­ tions except in the immediate vicinity of £>an Antonio, where records were obtained at several points other than at Weather Bureau sta­ tions. (See fig. 1.)

O Standard rain gage + Improvised rain gage Numbers and letters refer totableintext o____1____Z____3 Miles

1. Map of San Antonio River and tributaries in vicinity of San Antonio, Tex., showing drainage basins and total rainfall, September 8-10, 1921. Plate I, however, does not tell the whole story, as several areas in which no exact measurements were made received much greater rainfall than is shown. In a considerable area north and east of Taylor, for example, the total rainfall undoubtedly exceeded 36 inches. McAuliffe 1 writes:

iMcAuliffe, J. P., Excessive rainfall and flood at Taylor, Tex.: U. S. Weather Bureau Monthly Weather Review, September, 1921, p. 497. PRECIPITATION. y1

To give some idea of the amount of rainfall in this vicinity not reported by the Weather Bureau, the statements of six intelligent and trustworthy farmers living north and east of Taylor should be cited. These farmers usually have- barrels for hauling water when droughty conditions prevail. During the- drought which preceded the downpour there was much hauling of water, con­ sequently clean barrels were numerous. It is stated by these gentlemen that barrels on farms at different places measuring 36 Inches high by 18 inches in diameter were filled to overflowing on the morning of the 10th. It is certain, that not more than 2 inches of water were in the barrels prior to the exces­ sive rain, as this is the amount registered at the Weather Bureau station at that time. Allowing for all errors, it seems assured that some 80 inches of rain fell at many places in about 15 hours. These farmers deetare that -the barrels were empty prior to the beginning of rainfall. Statements made to the writer by several farmers near Taylor agreed closely with those given by McAuliffe. The basin of Salado Creek no doubt received considerably mare rain than much of the surrounding area. Many residents aldng the- stream said that it rained almost without cessation for more than 24r hours and that during several short periods in .that time the rail* was torrential. The storm came from the southwest, crossed the principal drain­ age lines, and followed closely the Balcones escarpment. The extent of the effect of that escarpment on precipitation is not known,, but it unquestionably increases it. The observers of the United States Weather Bureau at Houston,. San Antonio, and Taylor have described the storm as follows:2 Cause of the rains, Evidence is strong that the precipitation wa& the result of the breaking up in Texas of the disturbance that moved westward toward the Mexican coast south of Tampico on September 7, 1921. .Although, the distribution of the pressure was such that the storm could not be charted,, the shifting winds, the progressive northeastward extension of .the rainfall area, and the profound agitation of the atmosphere as evidenced by violent squalls and thunderstorms over the stricken section can hardly be ascribed? to any other cause. The storm apparently moved in from Mexico over Webb County and passed in a northeasterly direction over Bexar, Comal, Haysr and Travis counties into Williamson, Bell, and Milam counties, where it abruptly dissipated. Milam County borders on the west bank of the Brazos- River, and there was but little precipitation along the east bank of that stream. An area of high pressure of apparently feeble energy backing in over eastern Texas was probably a contributory cause of the record-breaking rains au$ their abrupt termination near the Brazos River. While the rains were dis­ astrous to life and property over a large area, there were many localities- in southern Texas where they proved beneficial by relieving the drought, reviving ranges, and providing stock water. Rainfall in San Antonio, A drought of two months' duration was broken* when a shower of 0.53 inch fell between 6 and 7 a. m. September 8. Seven­ teen hours later, between 12 midnight and 1 a. m, on the 9th, steady rains? » Mfonthly Weather Review, September, 1921, pp. 491-497. 10 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. began and continued until shortly after 11 p. m. a period of about 23 hours. The crest of the flood came through the city two hours after the precipitation ended. The amounts of precipitation, as measured at the United States Weather Bureau, are as follows: 7 a. m. 7 p. m. September 8______0. 53 inch. 0.01 inch. September 9______3.48 inches. 1.90 inches. September 10______1.46 inches. Trace, Total, 7.38 inches. This shows a total amount of 6.84 inches for the 24 hours ending about 11 p. m. September 9. If the showers that occurred in the morning of the 8th are included, the total is 7.38 inches. Records of the rainfall at San Antonio since 1885 show that only on one occasion has the 24-hour amount of September 9, 1921, been equaled or ex­ ceeded. This occurred on October 1-2, 1913, when 7.08 inches in 24 hours were recorded, and a destructive flood followed. At the Weather Bureau station, near the center of the city, 1.46 inches of rain fell between 7 p. m. and 11 p. m. September 9. However, the rainfall increased rapidly north and west of this point until amounts of 3, 4, and 5 inches occurred during this same period of time at stations located from 2 to 5 miles distant. On a map of the San Antonio River and its tributaries accompanying this report (fig. 1) are located 12 stations at which measured amounts of rain preced­ ing the flood are /shown. At five of these stations the precipitation was measured in regular 8-inch gages. At the remaining stations, including 9 others just outside limits of the map, improvised gages, such as cans, barrels, and in one instance a wooden trough, were used. Allowance was made for sloping sides and rounded bottoms where these occurred. With but one exception each of the 21 gages was visited, inspected, and the measurements carefully checked. In this connection the cooperation of the Engineering Department of the United States Army is kindly acknowledged. Without their aid this report could not have been so complete. Reliability of rainfall measurements. At seven stations using regular rain gages the measured amounts are considered very reliable. The observers in some instances have kept rainfall records for many years. The nine stations using improvised rain gages are given only approximate values in this report. However, the value of the data can not be based on the measurements of rain at any one locality, but on the large number of measurements taken over a relatively small area. The combined areas of the drainage basins responsible for this flood is approximately 75 square miles. In and adjacent to this drainage basin there were 12 measurements available and, with few exceptions, these measurements are in close accord. It is to be regretted that only at a few stations could amounts at different periods of the rain be obtained. Heavy rains shown. Study of the accompanying map shows the rapid increase of rainfall north and west of the Weather Bureau station. Two miles north 9.50 inches were recorded, with 3 inches after 7 p. m. on the 9th, as compared to 1.46 inches at the Weather Bureau station. Further study of the map brings out the fact that approximately 3 miles north and west from the city more than 10 inches must have fallen. Between 5 and 6 miles north and west from the city's center the rainfall undoubtedly reached 14 inches over a considerable portion of the drainage areas. Nine stations show 35 inches or more. PBECIPITATIOlSr. 11

The heaviest rainfall probably occurred 8 to 10 miles northwest of San Antonio. Beyond this point the rainfall was not so heavy, as shown by the 8-inch gage located 17 miles northwest, where 10 inches were recorded. It is possible, but hardly probable, that some sections of these drainage basins received 20 inches of precipitation in the storm preceding the flood. Two stations show this amount, but much evidence would be necessary before those familiar with rainfall data could accept the figures." Excessive rainfall and flood at Taylor, Tex. There occurred at Taylor, Tex., and its vicinity during September 9-10, 1921, the greatest 24-hour rainfall OH- record for the United States, 23.11 inches, with 23.98 inches in 35 hours. Dur­ ing the night of the 9th to 10th, ending with the 7 a. m. observation of the 10th, 19.49 inches of rain fell. The period of heaviest rainfall was from 6.45 p. m. to 9.42 p. m. of the 9th, when 10.50 inches were recorded. At this point the tipping bucket was flooded, stopping the automatic register. Rainfall was not so rapid after 10 p. m., and after midnight changed to intermittent showers, which continued until about 3 a. m. of the 10th, when the rate of fall again became excessive, the rain continuing as a steady downpour until 7.28 a. m., after which it fell at a moderate rate. The times of beginnings and endings of rain and the total amounts are as follows: Inches. September 9: Rain began about 3.30 a. m.; total, mid­ night to midnight______* 16.11 September 10: Rain ended 2.30 p. m.; total, midnight to midnight ______8 7. 87

Total duration, 35 hours______23.98 The most remarkable feature of this storm was its duration, covering a period of 35 hours, with an excessive rate over a period of more than 10 hours. Although continuous rainfall at an excessive rate ended shortly after midnight, it did not cease entirely, and from 12 midnight cf the 9th to-3 a. m. of the 10th there were several showers at an excessive rate, with moderately heavy rain intervening. From 3 a. m. of the 10th until 7.28 a. m. rain fell at an excessive rate. * * * Hutto, 10 miles southwest of Taylor, reported a tornado which damaged two churches, the white Baptist and the colored Baptist, and blew several houses off their foundations. This storm was also reported at Weir, 10 miles northwest of Taylor; and Mr. R. F. Young, cooperative observer at Georgetown, reported the same storm southeast of Georgetown. The time of occurrence was about midnight of the 9th. While this storm has been reported from these places as a " cyclone," it is almost certain that it was a severe squall in connection with the violent thunderstorm then in progress. There were several sudden gusts at Taylor during the same night, the strongest one being 31 miles an hour from the south at 3.11 a. m. of the 10th. * * * This remarkable rainfall, with its attendant floods and loss of life and prop­ erty, was the result of two thunderstorms of unusual violence. The first thunder was heard in the south at 4.19 p. m. of the 9th, and came at the close of a day that had given continuous rainfall at a slow rate since 3.30 a. m. The thunder in the south at the afternoon hour was of the deep, rolling variety, growing gradually louder and more frequent. By 7 p. m. the storm clouds had reached Taylor and the thunder and lightning were incessant. At 9.45 p. m. thunder became unusually heavy. All during the storm there was a continuous roar caused by thunder, falling rain, and more or less wind. Neither thunder

* Partly estimated; tipping bucket stopped before midnight. 12 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. nor lightning ceased the entire night. Neverthetess there was a lull in the storm after midnight, and comparative quiet prevailed about 3 a. m. Shortly after this another thunderstorm as severe as the preceding one, giving sharper crashes of thunder, rolled up from the south, and at the hour of the morning observation lightning was incessant, and the attendant thunder was deafening. This storm was last heard in the west at 12 noon of the 10th. Many houses were struck by lightning and areas in cotton, fields 20 feet square burned by lightning, but the flood damage was so great that very little attention has been paid to damage by lightning. The electric lights in Taylor were put oat of commission, and all telephone and telegraph service ceased. It is a peculiar circumstance that while this downpour occurred from about San Antonio to above Cameron, a southwest-northeast distance of more than 200 miles, Elgin, just 18 miles south of Taylor, received less than 4 inches of rainfall. The rainfall lines shown on figure 1 were taken from figure 1 in the Monthly Weather Review, September, 1§21. The drainage areas were measured on the topographic map of the San Antonio quadrangle, except the area north of its limits, which was measured on the progressive military map prepared by the United States Engineer Corps at Fort Sam Houston, Tex.

Rainfall at and near San Antonio, September 7 to 10, 1921. [The locations given are those shown on fig. 1.] Loca­ tion. Inches. Period covered or details.

1 15.0 Sept. 8, midnight, to Sept. 9, midnight. 2 17.5 Sept. 8, midnight, to Sept. 9, midnight. 3 .20.0 Sept. 9, 7 a. m.. to Sept. 10,, 7 a. m. 4 18.0 Sept. 8, midnight, to Sept. 9, midnight, 17.0 inches. 5 o21.0 Sept. 9, 7 a. m., to Sept. 10, 7 a. m. 6 625.0 Sept. 8, 7 p. m., to Sept. 9, 6 p. m., 10 inches: Sept. 9, 6 p. m. , to Sept. 10, 7 a. m.,15 inches. 7 13.0 Sept. 9, 7 a. m., to Sept. 10, 7 a. m. ' 8 13.18 Sept. 7, midnight, to Sept. 8, 8 p. m., 0.07 inch; Sept. 8, 8 p. m ,to Sept. 9, 11.30 p. m., 13.11 inches. 9 12-15 Sept. 9, 5 p. m., to Sept. 10, 7 a. in. 10 8-55 Sept. 7, 6 a.m. to 7 a.m., 0.60 inch; Sept. 8, 9 p.m., to Sept. 9, 9 a.m., 340 inches; Sept. 9, 9 a. m. to 4 p. m., 1.30 inches; Sept. 9, 4 p. m. to 11.30 p. m., 3.05 inches. 11 9.5 12 6.84 a 12.0 Sept. 8, midnight, to Sept. 9, midnight, 6 10.0 ^ept. S, midnight, to Sept. 9, 6 p. in., 6 indies; Sept. 9, 6-9 p. m., 3 inches; Sept. 9, 9-12 p. m., 1 inch. c 15.0 Sept. 8, midnight, to Sept. 9, midnight. a 18.0 Sept. 8, midnight, to Sept. 9, midnight. e 10.0 Sept. 8, midnight, to Sept. 9, midnight. f 11.0 Sept. 8, midnight, to Sept. 9, midnight. - 7.0 Sept. 8, midnight, to Sept. 9, midnight. I 10.14 Sept. 8, noon, to Sept. 9, noon, 4.62 inches; Sept, 9, noon, to Sept. 10, a. m., 3.74 inches; Sept. 10, a. m., to noon, 1.05 inches: Sept. 10, ntfon to eveniitg, 0.73 inches.

a Not considered reliable by Weather Bureau. 6 Note that Weather Bureau gives 15 inches as total for this station and not 25 inches, as shown here. NOTE. Numbered gages used bv United States Weather Bureau in drawing rainfall lines shown oa fig. 1. Records at lettered stations and notes regarding distribution furnished by United States Engineer Corps, Fort Sam Houston, Tex. MAXIMUM DISCHARGE PEE SQUARE MILE.

Rainfall, in inches, September 7-11,. 1921. Brazo* »iv«r haute. ' '

Statical. County. Sept. 7. Sept.*. Sept. «. Sept. 10. S*pt.ll. Total.

Brazoria...... 0.83 ass 0.96 0.13 0.34 3.14 Bre^hwm ,...,... . .01 . 2.14 .as 2.38 Cameron Milam.T...... 30 .77 .85 12.45 1437 CaUnoai8tatian Trace. .22 1.12 .07 1.41 1.07 2.85 3.82 Gatesville do. .... 1.00 1.00 .08 .70 13.00 1.43 15.19 Hamilton .... Hamilton ..»...,.,., .00 Waller...... 50 .77 .08 .00 L33 Lampasas. Tjftnripasws. 1.50 .03 1.10 .70 3.38 Navasota...... GHmps , .55 .53 1.08 Rosenberg...... Fort Bead.... .40 2.00 1.22 3.62 Sealy...... Austin...... 09 »30 .93 1.48 SomervUle ...... Burteson . .. .30 .50 .80 25 1.96 .18 .27 2.65 .04 .28 16.11 7.87 .18 24.48 Ben...... 04 .35 9.00 2.55 11.94 .25 2.00 .32 2.57

Coterado Blrcr bavin.

Austin...... 15 .05 15.00 4.03 Trace. 19.23 Columbus...... Colorado.. .14 2.00 .30 2.44 Borland...... 05 3.40 5.30 Trace. &75 .02 .49 .59 .09 1.19 fSJfrui, .14 1.65 1.79 Marble "ll^aiis Trace. .57 5.50 11.00 17.07 Morris Ranch. GiBaspte...... 02 1.24 .24 .01 1.51 rteree...... Wharton...... 92 3.85 .06 483 Smith ville...... 58 .28 1.16 1.36 3.88

Guadalupe River basin.

Blanco...... 18 .15 1.72 5.80 7.85 Cuero...... DeWitt...... 22 1.37 ,4ft 2.05 Gonzales...... Gonzales . . . .12 / 1.63 1.24 2.99 KerrvUle...... Kerr...... 02 .01 .26 2.87 .04 3.20 Luling...... Caldwell...... 08 1.50 .90 2.48 .18 9.38 9.56 Son Marcos.. ... 2.00 8.00 1.50 11.50 Victoria...... Victoria...... 75 .65 1.02 2.42

San Antonio Hirer basin.

Boerne...... Kendafl. .. .23 6.45 .04 .6.75 Golfed.... Goliad...... 56 .49 2.20 .03 .36 8.64 Kernes City. . 44 4.51 4.9! .....do...... 84 2.65 3.4! San Antonio...... 54 6.84 Trace. 7.3!

MAXIMUM DISCHARGE PER SQUARE MILE. . The tables on pages 19 and 43 show the unusually high rate .of discharge that occurred on some of the streams. Little River, at Cameron, whose drainage area is 7,010 square miles, had a discharge of 92.3 second-feet per square mile. As far as the writer knows, this is tHe highest rate, of discharge that has ever been recorded for so large an area. Furthermore, over 2,000 square miles of the Little River basin the part receiving less than 1 inch of rain did not materially contribute to the flood. , at Belton, rose only about 9 feet, a rise representing a discharge of probably less than 19528 28 3 14 FLOODS IN CENTRAL TEXAS IN SEPTEMBEB, 1921. 5,000 second-feet. above Salado Creek had only a moderate rise. The discharge : from 5,000 square miles above Cam- eron was therefore over 125 second-feet per square mile. A discharge of 966 second-feet per square mile from Salado Creek, which drains an area of 148 square miles, is believed to be unequaled, except, pos­ sibly, by the flood of June 10, 1905, on , Iowa, which was originally determined te have been at the rate of 1,300 second- feet per square mile from an area of 143 square miles.4 That, how­ ever, has recently been recomputed 5 and the rate reduced to 600 second-feet per square mile. FLOOD FLOWS. ; BRAZOS RIVER FLOOD.

GENERAL FEATURES. Brazos River rises in eastern and flows southeastward across the State of Texas. It enters the Gulf of Mexico at Quin- tana, east of Freeport, in Brazoria County. The river is about 840 miles long, and it drains an area of 41,700 square miles. Practically the entire flood in Brazos River came from its chief tributary, Little River, which joins the Brazos from the west at Valley Junction, Tex. Little River drains an area of 7,560 square miles, and the combined drainage area above the junction of the two streams is about 34,000 square miles. Elevations and distances along Brasos River from, Waco to mouth.

Descent between Distance.** points. Elevation Point. above sea level. From Point to Waco. point. Total. Per mile.

Feet. Milts. Miles. Feet. Feet. Waco...... 378 Contour crossing...... 350 21 21 28 US Do...... 325 40 19 25 1.3 Do...... 300 - 47 7 25 S.-6 Do...... 275 66 19 25 1.3 Do...... v...... 250 81 15 25 1.7 238 86 5 12 2.4 225 97 11 13 1.2 Do...... 200 120 33- 25 1.1 195 125 5 5 1.0 Contour crossing...... _ ...... 175 149 24 20 .8 170 155 6 5 .8 158 170 15 12 .8 150 181 11 8 .7 134 216 35 16 .5 129 224 8 5 .6 95 272 48 34 .7 San Antonio & Ry. bridge...... 77 291 19 18 LO 53 316 25 24 r.o 50 324 8 3 .4 25 345 21 25 1.2 Mouth...... 0 424 79 25 .3

Measured along bends of stream. NOTE. Baaed on survey made in 1900 by U. S. Engineer Corps. * U. S. GeoL Survey Water-Supply Paper 162, pp. 24-29, 1906. BAm. Spe. Civil Eng. Proc., May, 1922, p. 1232. BRAZOS RIVER FLOOD. 15 The United States Weather Bureau maintains gages on Brazos Kiver at Valley Junction, Washington, Hempstead, Kosenberg, and other points farther upstream in order to issue flood warnings. The following is quoted from a description of the recent flood of Brazos River :6 Flood of the Brazos. This flood was remarkable from the feet that it was caused by tremendeus rains over a single tributary, the Little River, which empties into the Brazos just above Valley Junction, and that the large volume of water spread out at an exceedingly rapid rate as it rushed downstream. * * * The flood waters from Little River began to pour into the Brazos just above Valley Junction on Saturday, September 10. The gage at Valley Junction showed a stage of 3.5 feet at 7 a. m. of that date. At 4.30 p. m. the river was up to 25 feet, rising fast; and at 6.30 p. m. bankfull. The observer then warned all residents to leave. No gage readings were taken , 12, and 13, but measurements made from marks left by the flood showed that the maximum height was 58.2 feet, only 0.8 foot below the record flood of December, 1913, but 4.2 feet higher than- flood of the spring of 1915. The flooded area was ap­ proximately 4 miles wide. Cotton and corn were ruined and railroad tracks and bridges washed out for a distance of 3 miles, suspending travel for six days. There were no deaths. Flood stage at Valley Junction is at 44 feet. At Washington, near Navasota, the stream began to rise rapidly about 8 p. m. September 10. The initial 24-hour rise amounted to 21.8 feet te a stage of 27 feet at 7 a. m. September 11. During the next 24 hours there was an ad­ ditional rise of 15.2 feet, and the stream ultimately reached the peak of the flood on the morning of , with gage reading 50 feet. This was 5 feet above flood stage, but 2.9 feet below the flood of April, 1915. At this time, the stream varied from 1 to 3 miles in width. The damage is estimated at $150,000. There was no loss of live stock. At Hempstead a high-water gage only is maintained on aeeerant of the yield­ ing nature of the banks. The maximum stage reported was 40.2 feet, 0.2 foot above flood stage, at 7 a. m. . This is 6.3 feet below the flood of April, 1915. The lowest section of the river gage was washed away with the initial rise. The damage from the flood is estimated at $43,500, including $1,600 for loss of live stock. The money value of the property saved by warnings is estimated at $100,000. Wallis, Austin County, located below Hempstead, reports $5,000 damage to cotton and corn, and a saving of 1,500 head of cattle, valued at $37,500, through the warnings. Flood stage was not attained at Rosenberg, although the stream was bank- full to ove'rflowing near the coast, where the land is l«vel and the run-off correspondingly slow. No damage occurred, except that the Freeport Harbor entrance had shoaled as a result of the flood and had to be dredged to re­ lease a steamer. 8 Banaemeyer, B., The Texas flood of September, 1921: U. B. Weather Bureau Monthly Weather Eeview, September, 1921, pp. 493-494. 16 FLOODS IF CENTRAL TEXAS IN SEPTEMBER, 1921.

Gage height, in feet, of Brazos River September 9-23, 192L

Valley Wash­ Rosen- Valley Wash­ Rosen- Date. Waco. Junction. ington. berg. Date. Waco. Junction, ington. berg.

Sept. 9...... 6.2 2.4 5.2 1 Q Sept. 17. 5.8 8.3 48.0 369 10. 6.2 3.5 5-2 1.9 18. 7.0 8.0 31.9 37.5 11. 6.2 («) 27.0 1.9 19. 8.8 8.0 24.6 35.4 12. . 6.1 << ) 42.2 20. 9.2 7.5 20.4 27.7 13. . 6.0 46.5 27.6 21. 8-8 7.0 18.4 17.3 14. . 6.0 21 5 50.0 31.4 22. 8.4 8.3 17.1 14.0 15. 5.9 17.3 49.8 35.5 23. 8.1 8.0 16.7 11. § 16. 5.9 9.7 47.1 37.7

<> Gage was inaccessible Sept. 11-13; maximum stage, 58.2 feet. NOTE. Gage heights at Valley Junction, Washington, and Rosenberg furnished by U. S. Weather Bu­ reau. Observations made at 7 a. m.

DETERMINATION OF FLOW. A gaging station is maintained by the United States Geological Survey on Brazos River at Waco, 60 miles above Valley Junction. The area drained by the Brazos above Waco is 25,500 square miles. The discharge at Waco was about 200 second-feet and was not in­ creased by the storm. A gaging station is also maintained on Brazos River at Jones Bridge near College Station, about 20 miles below Valley Junction. The drainage area above Jones Bridge is 35,400 square miles. Be­ cause of excessive overflow the discharge at Jones Bridge could not be determined at maximum stage or at any stage above the point of overflow. The maximum stage, 53 feet, occurred between 1 and 3 a. m. on . At that time the flood plain along the right bank was overflowed for a width of 3 to 5 miles. The levees above the gage were overtopped, seriously broken, and eroded at numerous places for a total length of about 3,000 feet. The discharge back of the levees at maximum stage was undoubtedly several times greater than that in the main channel. This station was visited by an en-r gineer of the United States Geological Survey on , and the maximum discharge in the main channel between the high bank on the left and the levees on the right was determined by the slope method by using Kutter's formula. The following results were obtained: Discharge, 143,000 second-feet; area, 23,500 square feet; mean velocity, 6.07 feet per second; wetted perimeter, 1,282 feet; hydraulic radius, 18.33; slope, 0.00030; distance over which slope was measured, 2,400 feet; assumed value of coefficient of rough­ ness ft, 0.030. BRAZOS RIVER FLOOD. IT

Gage .height and discharge of Brazog River ttt Jones Bridge, near OMege Station, Tea?., September 8-17, 1921.

Mean Mean Date: Time. Gage daily dis­ Date. Time. , Gage daily dis­ height. charge. hsigHt. charge.

Feet. Seeon&ft. Feet. Second-ft. Sept. 8...... 5.1 340 49.6 8...... 5-» 13...... 48.6 9...... 8.6 4,200 14...... M. 1 9...... 12.2 14...... 38.1 10...... 17.5 21,500 15...... 32.2 40,300 10...... 23.2 15...... 26.1 11...... 44 0 16...... 20.2 18,100 11...... 48.3 16...... 16.1 12...... 53.0 17...... 15.2 10,600 12...... c| Q 17...... 14.1 12...... 6p.m. 49.6 ......

LITTLE RIVER AND TRIBUTARIES.

GENERAL FEATURES. Little River is formed by the junction of Leon and Lampasas rivers, near the town of Little River, 6 miles southeast of Belton. It joins Brazos River near Valley Junction. The main headwater stream is Leon River, which rises in Eastland County near Eastland and Cisco at an elevation of 1,500 feet. The principal tributaries of Leon River are Cowhouse Creek, which enters from the west 5 miles above Belton, and Nolan Creek, a small stream which passes through Belton and also enters from the west. Salado Creek enters Lampasas River from the west about 1 mile above its junction with Leon River. San Gabriel River, the principal tributary of Little River below the junction of the Leon and Lampasas, enters from the south about 25 miles above the mouth of Little River. It is formed by the union of the North and South forks at Georgetown, and its largest tributary is Brushy Creek, which enters from the southwest, 3 miles from its mouth. That part of the Little River basin lying west of the Balcones escarpment is known as the Grand Prairie region; that to the east as the Black Prairie region. The latter is practically all cultivated, except along the channels of the streams, which are generally fringed with a narrow strip of timber. The soil is principally clay and ex­ tremely fertile. Considerable areas are cultivated in the Grand Prairie region, both along the alluvial stream bottoms and on the upland plateau, which has been deeply eroded. As the side slopes of the valleys are steep and are covered with only a thin layer of soil, they are suitable principally for grazing. 18 % FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. About the only timber in the basin is a narrow fringe along the watercourses and in the headwater areas, which lie in the Western belt.7 The following table was compiled chiefly from topographic maps:

Elevations and distances along Little River {imclnding Leon River) from source to mouth.

Descent between Eleva­ Distance. « points. tion Point. above sea level. From Point to source. point. Total. Per mile.

Feet. Miles. Miles. Feet. Feet. 1,650 Junction of North and South forks,...... 1,870 14 14 280 20.fr 1,140 50 36 230 6,4r 950 82 32 190 5.9 Gatesvffle...... 705 125 43 245 5.7 475 185 60 2% 3.» 450 194 9 25 2.S San Gabriel River...... 305 230 36 145 4.0 280 237 7 25 3.ft Mouth...... 235 256 19 45 2.*

a Measured along largest meanders of stream.

DETERMINATION OF FLOW. The only gaging station in the basin of Little River is at Cameron, where a staff gage at the city's pumping plant, half a mile south of the town, is ordinarily read twice daily. The maximum discharge of Little River was determined by the slope method by using Kutter's formula at two points the upper just below the junction of Leon and Lampasas rivers and the lower near Cameron. The maximum discharge per square mile of San Gabriel River and Brushy Creek increased rapidly below the points of measure­ ment, but the condition of the channel was so unfavorable that a determination of flow was impossible. However, the maximum discharge of San Gabriel River below Brushy Creek probably equaled and perhaps exceeded the maximum at Cameron. The results of measurements of maximum discharge in the Little River basin are summarized in the following table:

7 Hill, R. T., Geography and geology of the Black and Grand prairies, Tex.: TJ. S. GeoL Survey Twenty-first Ann. Kept, pt. 7, 1901. BRAZOS BITCH FLOOD.

Maximum discharge of Little River and its chief tributaries during flood of September 9-10,1921.

Maximum discharge, HydraoUc factors usW in eoraipotaticja.

Time of I flood whichDistanceover slopemeasured.was Stream. Location. Permile.square crest Wettedperimeter. ofrough­Coefficient Septem­ radius.Hydraulic (n).ness ber 10. || 1 | ! Sec.- Sec.- ft. ft. Feet. fat, Fed. 5 FL Little River- Near junction 5,300a 331,000 62.5 10.30a.m. 4,600 a 00078 3,073 14.70 45,170 aoss 7.32 of Leon and Lampasas rivers. Do...... NearCameron. 7,010 647,000 92.3 2 JO p.m. 2,660 .000214 6,198 20.12 124,700 .035 5.19 Salado Creek. SfearSalado... 148 143,000 966 9 a.m. 1,513 .0030 1,MO 11,03 18.560 .040 8.® San Gabriel Near George­ 431 160,000 371 11 a.m. 3,531 .00157 1,833 11.91 21,930 .045 7.29 River. town. Brushy Creek At Round 74.7 34,500 462 Early (0) (.) W Rock. morning. 0 See table on p. 28.

LITTXE BITES BELOW JUNCTION OF LEON AND &AXPASAS BIVEBS. The maximum discharge of Little River at a point below the junction of Leon and Lampasas rivers was measured on March 1&, 1922, The slope was determined from two well-defined high-water marks on the right bank, 4,600 feet apart. The upstream mark was a line in an old blacksmith shop just below the bridge over Lampasas River on the Belton-Holland road and about 200 feet upstream from the low-water point of of Leon and Lampasas rivers. The downstream mark was a notch ciit in the base of an oak tree by Mr. J. W. D. Wright on the day of the flood. In this stretch of 4,600 feet the direction of flow at maximum stage was straight, the cross-sectional area was practically uniform, and the channel was relatively clean and unobstructed, except for a fringe of trees and brush along the low-water channel. Plate II, A, gives a general idea of the condition of the channel covered by the flood. This view was taken from the right bank, looking upstream from a point near the lower end of the slope stretch. The line of trees in the background at left marks the location of the ordinary flow channel. The results of this measurement are given in the above table. FLOODS IN CENTRA!, TEXAS USt SEPTEMBER, 1921.

Cross section used in determination of maximum discharge of Little River 1 low junction of Loon and Lampasas rivers.

Distance frwn Depth Distance from Depth Distance from Depth Distance from Depth right edge of of right edge of, of right edge of of right edge of of water. water. water. water. water. water. water. water.

Feet. Feet. Feet. Feet. Feet. Feet. Feet. Ftet. 0...... 0.0 250...... 33.3 683...... 19.4 1,814...... *"1& 2 133...... 9,1 348...... 21.0 759...... 19.5 2,256...... 9.2 138.. . . 20.0 382...... 19.5 854 14.6 2,464...... 13.5 ITS.....;...... 36.2 451...... 17.5 i fljfci 14.1 2,770...... 8.2 198...... 43.3 498...... 22.8 1,394...... 15.5 2,960 5.6 221 48.1 «54...... 16.8 1 <»BS 15.7 n neA .0 244...... 43.3 608...... 16.5 1 '744 16.8

Mr. Wright stated that at his ranch the river began to overflow its banks about daybreak on September 10 and reached iis maximum stage at 10.30 a. m. It soon began to fall slowly and by night had fallen to a level only slightly above bankfull stage. Mr. Me F aria ml, who lives on the left bank of the river about a mile below Mr. Wright, stated that the maximum stage there occurred at about 11 a. m. on September 10. LITTLE RIVER AT CAMERON. Little Eiver began to rise at Cameron at about 1 a. m. on Septem­ ber 10; at 8 a. m. it had risen to a stage of about 44.4 feet, and at 2.30 p. m. it had reached its maximum stage of 49.5 feet. At 3 p. m. on September 11 it had fallen to a stage of 38.3 feet, and at 5 p. m. it had fallen to 37.6 feet but did not return to bankfull stage of 30 feet until the night of . The gage was not read from September 14 to . The stage for that period has been estimated from a study of gage heights on Brazos Eiver at Valley Junction and College Station, from statements of residents, and from a general knowledge of the characteristics of the stream. Figure 2 shows the gage-heigh^ graph from September1 9 to Septem­ ber 21. The maximum stage was determined by leveling from marks in the pump house near the gage. Other records of stage on Sep­ tember 10 and September 11 were determined by leveling from stakes set at the edge of the water during the flood. The rating for this gage has been developed to 30 feet, above which the banks are overflowed for a width of more than a mile. The rating, however, includes only the discharge in the main chan­ nel, because at a stage of 18 feet (discharge, 6,950 second-feet) water enters an old channel about a mile above the gage and returns to the river below the gage. The discharge through the main channel at a stage of 30 feet is 12,700 second-feet, and it is estimated that 9,000 second-feet passes through the side channel, making a total discharge of 21,700 second-feet at bankfull stage. U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 488 PLATE II

A. VIEW OF LITTLE RIVER VALLEY RELOW JUNCTION OF LEON AND LAMPASAS RIVERS.

B. VIEW OF LITTLE RIVER AT GULF, COLORADO & SANTA FE RAILWAY CROSSING NEAR CAMERON, TEX., AT CREST OF FLOOD. U. S. GEOLOGICAL SURVEY WATEE-SUPPLY PAPEE 488 PLATE III

A. VIEW ACROSS LITTLE RIVER VALLEY NEAR CAMERON, TEX., FROM MARLOW BRIDGE.

B. VIEW TJP SALADO CREEK FROM JONES MILL, NEAR SALADO, TEX. I*. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 488 PLATE IV

SEy^^""...~i . .- *, ,|^^^y||lil*^'* , v1 '",... 'te^-S.^^

A. VIEW DOWN SAN GABRIEL RIVER VALLEY BELOW GEORGETOWN, TEX.

B. VIEW UP BRUSHY CREEK NEAR ROUND ROCK, TEX.

BEAZOS RIVER FLOOD. 21 The maximum discharge near Cameron was determined at Mar- low Bridge, 4,000 feet below the crossing of the Gulf, Colorado & Santa Fe Railway, 2 miles northeast of Cameron and 2 miles down­ stream from the gage. Plate II, #, shows the high velocity and resulting destruction when the flood was at its crest. Plate III, J.,

50 2. Wp.m. ft. >/

4O Estii nated '.ota/ n

Id Ul Bar k-fuil stage L.30 Dhchargt 22,OOi pa "tly es Vmate 3

20 rge 6,8 *O sec>ft. by cut rent-m etermt 'asuret tents

10

Discha -ge {ZOt i sec.-n,

9 - IO II 12 13 14 15 16 17 18 19 20 21 SEPTEMBER PiotTfiH 2. Gage-height graph of Little River at Cameron, Tex., daring flood of Septem­ ber, 1921. is a view from the left end of the main span of Marlow Bridge showing the flood channel. The slope was determined from fairly well defined high-water marks on the left bank. A stretch of chan­ nel was measured from a point 1,560 feet above the road crossing to a point 1,100 feet below. As the channel in that 2,660-foot stretch was uniform only one cross section along the center of the road was measured. 19526 23 4 22 FLOODS IN CENTBAL TEXAS IN SEPTEMBER, 1921. The following table shows the elements of this section:

Cross section used in determination of maximum discharge of Little River at Cameron, Tex.

Distance from Depth Distance from Depth Distance from Depth Distance from Depth left edge of of left edge of of left edge of of left edge of Of water. water. water. water. water. water. water. water.

Fed. Feet. Feet. Feet. Feet. Feet. Feet. Feet. 0...... 0.0 1,700...... 1

NOTE. Smooth cultivated or grass land from 0 to 5,800 feet. From 5,800 to 6,168 feet there is a dense growth of trees and brush, except for a width of about 100 feet, which is that stretch occupied by the ordinary flow of the river. High-water marks sufficiently definite to determine the slope on the right bank could not be found, although the survey was made on , only 18 days after the flood. Practically the entire overflowed area within the slope stretch was smooth, cultivated land and the only large obstructions were the trees and brush, along the main channel. These obstructions, however, covered a wiclth of less than 400 feet of the total width of 6,168 feet. The value of n is assumed to be 0.035, which, if in error, is too large rather than too small. The results of this measurement are given in the table on page 19. The total run-off past Cameron has been computed from the meager data available as 900,000 acre-feet on September 10 and 11, about equally divided between the two days. The run-off for September 12 to 21 was about 170,000 acre-feet, making a total for the flood of more than 1,000,000 acre-feet. (See fig. 2.)

SALADO CREEK NEAB BALADO. Salado Creek heads in the northern part of Williamson County. It flows northeastward for 34 miles and enters Lampasas River in Bell County, a mile above the mouth of Leon River. It drains a plateau region through which the streams have cut deep, steep-sloped valleys. The flow is ordinarily derived from springs near Salado, which have a total discharge of less than 25 second-feet. Above Salado the thin covering of stony clay soil is underlain with lime­ stone. A considerable part of the basin is covered with small trees, BRAZOS RIVER FLOOD. principally post oak and cedars. The cultivated area, which em­ braces probably less than 10 per cent of the total drainage area, con­ sists of small areas along the flood plain of the main stream and of larger areas on the flat-topped plateau between the branches of the creek. The following table has been compiled from topographic maps: ' Elevations and distances along Salado Creek from source to mouth.

Descent between Distance. ° points. Elevation Point. above sea level. From Point to source. point. Total. Per mile.

Feet. Miles. Miles. Feet. Feet. Contour crossing...... 1.000 2 2 100 50.0 DO...... ;. 900 6 4 100 25. » Do...... 800 11.5 5.5 100 18.2 Do...... 700 16.5 5 100 20.0 Do...... 600 23.5 7 100 14.3 Salado...... 585 24.5 1 15 15.0 550 27 2.5 35 140 515 28.5 1.5 20 13.3 500 29.5 1 15 15.0 Mouth...... 450 34 4.5 50 11.1

a Measured along largest meanders of stream. According to statements of residents the flood on Salado Creek reached the highest stage ever known. The average rainfall in the basin was undoubtedly considerably greater than is indicated on Plate I. The heaviest precipitation occurred in the central and lower part of the basin. At Florence, near the head of the stream, the rainfall was considerably less than at Salado. The superintendent at Jones mill reported that two distinct peaks occurred on the morning of September 10. The first came about 7 a. m., after which the stream fell about a foot. The second rise then began and reached a stage 8.4 feet higher than the first peak in less than an hour. The time of the maximum stage is not definitely known, but it was probably between 9 and 10 a. m. The stream re­ mained high for several hours but began to fall rapidly about 2 p. m. The maximum discharge was determined at Jones mill, 3 miles below Salado, by the slope method by using Kutter's formula. The slope was determined from two well-defined high-water marks on the left bank, 1,533 feet apart. The upstream mark was in a dwelling house just above road crossing, and the downstream mark was in Jones mill. Both marks were well-defined lines on the inside walls of the buildings. As the channel was uniform throughout the slope stretch only a single cross-section, one about midway of the stretch, was measured. The elements of this section are given in the fol­ lowing table: 24 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921.

Cross section used in determination of maximum discharge of Salttdo Creek at Jones mttl, near Salado, Tex.

Distance from Distance from Distance from Distance from left edge of Depth left edge of Depth left edge of Depth left edge of Depth water. of water. water. of water. water. of water. water. of water. x Feet. Feet. Feet. Feet. Feet. Fed. Feet. Feet. 0...... 0.0 222...... 18.8 475...... 26.2 1,016...... 5.3 52...... 4.1 250...... 16.4 519...... 21.8 1,151...... 3.5 73...... 6.7 344...... 21.0 572...... 17.3 1,431...... 1.1 96...... 11.2 345...... 21.8 IV7Q 16.9 1,491...... 0 155...... 17.1 464...... 19.3 791...... 14.5 195...... 13.6 474...... 20.2 925...... 10.2

The main channel and the left bank of the stream were consider­ ably obstructed by trees, rocks, and irregularities of the ground sur­ face. The ground to the right of the main channel was smooth and unobstructed. About one-third of the wetted perimeter was rough, and the rest was smooth. The results of this measurement are given in the table on page 19. Plate III, 5, is a view of the valley of fealado Creek looking upstream from a point on the left bank just above Jones mill. This picture exaggerates the roughness because the trees and brush in the foreground obscure the clean smooth part of the channel.

SAN GABRIEL RIVER AT GEORGETOWN. San Gabriel River is formed by the junction of North and South forks at Georgetown. They head in Burnet County on the Colorado River divide at an elevation of 1,500 feet. The basin above the forks lies on the upland part of the Grand Prairie region and consists of flat-topped hills broken by deeply cut stream channels with rather steep slopes. The thin covering of clay soil is underlain with lime­ stone. The crosses San Gabriel River at Georgetown. It is at that point that the river leaves the Grand Prairie region and enters the more level Black Prairie region, finally joining Little River, near the western border of the Coastal Plain. BRAZOS EIVEE FLOOD. The following tables have been compiled from topographic maps: Elevations and distances along San Gabriel River, including North Fork, from source to month.

Descent between Eleva­ . Distaace.o points. tion Point. above sea From Point to Per level. source. point. Total. mile.

Feet: Miles. Miles. Feet. Feet. 1,550 1,300 4 4 250 62.5 Do...... 1,200 7 3 100 33.3 Do...... 1,100 11.5 ,4.5 100 22.2 Do...... 1.000 15.5 4 100 20.0 Do...... 900 23.5 8 100 12.5 Do...... 800 32 8.5 100 11.8 Do...... 700 41 9 100 11.1 South Pork...... :...... 680 42 1 20 20.0 650 43.5 1.5 30 20.0 DO...... :.:...... 600 47.5 4 50 12.5 Do...... 500 57.5 10 100 10.0 DO...... 400 72.5 15 100 6.7 Brushy Creek...... 325 84.5 12 75 6.2 305 87.5 3 20 6.7

o Measured along largest meanders of stream. Elevations and distances along South Fork of San Gabriel River from source to mouth.

Descent between Eleva­ Distance." points. tion Point. above sea From Point to Per level. source. point. Total. mile.

Feet. Miles. Miles. Feet. Feet. 1,500 Contour crossing...... 1,300 3.5 3.5 200 57.1 Do...... 1,200 6 2.5 100 40.0 Do...... 1,100 9 3 100 33.3 Do...... 1,000 13.5 4.5 100 22.2 Do...... 900 20.5 7 100 143 Do...... 800 29 8.5 100 11.8 Do...... 700 35 6 100 16.7 Mouth...... 680 36 1 20 20.0

0 Measured along largest meanders of stream. The rainfall in the San Gabriel basin ranged from about 5 inches at the headwaters to 15 inches at Georgetown. The maximum rainfall, which was probably more than 35 inches, occurred near San Gabriel, 30 miles east of Georgetown. Mr. J. M. Shell, who lives on the right bank of North Fork of San Gabriel River about half a mile above the mouth of South Fork, stated that the crest of the flood at his house occurred about 10 a. m. on September 10. Residents at the junction of the fork stated that the crest occurred there about 10.30 a. m., and Mr. M. M. Moore, who lives 3 miles below the forks,, stated that the maximum stage at that point was reached at 11 a. m. Residents of San Gabriel reported two distinct rises; the first, which was caused by torrential rains in the vicinity, reached its crest at about 5 a. m. on September 10; the second reached its crest. FLOODS IN CENTRAL TEXAS IN' SEPTEMBER, 1921. at 5 p. m. and was 1 foot higher than the first. Those living at the Felcher ranch, half a mile above San Gabriel, said that the maxi­ mum stage occurred at 9 a. m. The maximum discharge at a point about 3 miles below the junction of the North and South forks was determined on March 16, 1922, by the slope method by using Kutter's formula. The slope was de­ termined by two distinct high-water marks on the left bank, 3,531 feet apart. The upstream mark was a clear line in a barn on the C. C. Cody farm, and the downstream mark was a notch cut in a tree near the ranch house of M. M. Moore. Between the marks the channel was straight and fairly uniform in cross- sectional area. Cross sections were measured near both ends of the stretch. They differed in area by less than 2 per cent. There was a dense growth of trees along the low-water channel, which was also somewhat obstructed by a few large rocks. More than half of the channel, however, was smooth plowed land. Plate IV, A, is a view of the valley of San Glabriel River. The pic­ ture was taken from the high-water line on left bank, looking down­ stream from upper end of slope stretch. The results of this meas­ urement are given in the table on page 19. The elements of the cross sections are given in the following table: Cross sections used in determinations of maximum disehnrge of &mi Gabriel Rwer near Georffetotcn, Tex., during flood of September, 1921. ' Upper section.

Distance Distance Distance Distance fromleftedge Depth from left edge Depth from left edge Depth fromleftedge Depth of water. of water. of water. of water. of water. of water. of water. ol water.

Feet. Feet. Feet. Feet. Feet. Feet. Feet. Feet. 0...... 0.0 430 20.6 800...... 24.1 1,440...... 2.6 a)...... 6 450...... 22.7 817...... 23.8 1,500...... 2.6 100...... 1.4 480...... 25.3 823...... 27,2 1,600...... 2;S 150...... 2.4 530...... 27.8 830...... 24.6 1,680...... 3.4 200...... 3.8 580...... 29.3 840...... 17.5 1,730...... 4.6 240...... 4.4 620...... 29.7 850...... 14.4 1,820...... 8.5 280...... 7.0 680...... 29.1 860...... 12.2 1,915...... 10.2 820...... 10.2 720...... 30.3 890...... 12.4 1,920....'..... 11.9 330...... 13.4 740...... 31.4 960...... 8.8 1,930...... 12.2 340...... 12.3 760...... ;. 32.3 1,000...... 7.3 1,950...... 17.8 350...... 14.0 768...... 31.5 1,100...... 6.0 1,955...... 21.7 360...... 16.5 780...... 29.7 1,209...... 4.5 1,980...... 21.7 400...... 18.4 790...... 26.4 1,320...... 3.3 2,000...... 0

Lower section.

Distance Distance Distance Distance fromleftedge Depth from left edge Depth fromleftedge Depth from left edge Depth of water. of water. of water. of water. of water. of water. of water. of water.

Feet, Feet. Feet. Feet. Feet. Feet. Feet. Feet. 0...... 0.0 720...... 11.8 1,080...... 16.1 1,430...... 29.6 20...... :... 5.0 820...... 11.5 1,165...... 15.7 1,434...... 28.5 50...... 7.4 870...... 12.5 1,185...... 30.2 1,440...... 27.2 100...... 7.6 880...... 13.1 1,200...... 29.3 1,443...... 23.1 160...... 7.5 890...... 15.8 1,220...... 25.4 1 iSfl 25.1 240...... 6.9 900...... 15.8 1,250...... 24 7 1,495...... 8.6 360...... 8.1 910...... 10 4 1,260...... 27.8 1,550...... 3.0 480...... 10.0 940...... 17.1 1,330...... 28.5 1,623...... 0 570...... 10.1 960...... 17.5 1,380...... 29.9 BRAZOS RIVER FLOOD.

BRITSHY CREEK AT ROUND BOCK. Brushy Creek heads in Williamson County near the Travis County liiief It joins San Qabriel River from the south in Milam County, 3. miles from the mouth of Little Eiver. The topography and geology of the basin above Eound Rock is similar to that of San Gabriel Eiver above Georgetown and of Salado Creek above Salado. (See pp. 22-24.) The Balcones fault crosses Brushy Creek at Round Rock. The following table was compiled from topographic maps: ,

Elevations and distances along Brushy Creek from source to mouth.

Descent between Eleva­ Distance." points. tion Point. above sea From Point to Per level. source. point. Total. mite.

Fed. Miles. Miles. Feet. * Feet. 1,000 900 2.5 2.5 100 40.6 Do...... 800 6.5 4 100 25.0 Junction of Running Brushy and Dry Brushy creeks. . 750 10.5 4 50 12.5 Round Rock ...... 700 14.5 4 50 12.5 Contour crossing...... 650 17 2.5 50 20.0 DO...... :...... 600 23 6 50 8.3 Do...... 550 26 3 50 1«.T Do...... 500 31.5 5 50 10.0 Do...... 450 36.5 5 50 10.0 Do...... 400 45.5 9 50 5.6 Do...... 350 53 7.5 50 6.7 Mouth...... 325 56 3 25 8.3

a Measured along largest meanders of stream. The maximum discharge of Brushy Creek was determined at a point one-fourth of a mile below the crossing of the Austin-George­ town Highway and the International & Great Northern Railway. The slope was measured from an exceptionally well defined deposit of small pieces of drift along the high-water line on the left bank for a distance of 1,246 feet. No marks sufficiently distinct to de­ termine the slope could be found on the right bank. Plate IV, B, shows the condition of the channel in the upper section, which is not materially different from that in the lower. From a point about 100 feet above the middle section to a point about 100 feet below the channel was considerably rougher than in the upper sec­ tion. In this stretch of 1,246 feet, the condition of the channel was not uniform, as shown in the following table, which gives the hydraulic factors for each of the three cross sections measured. 28 FLOODS IN CENTRAL TEXAS OT SEFTEMBEB, 1921.

Summary of determination of ma&imwn, ^dtecV&rje of Brushy Creek at Round Rock, Tex.

Between Between Upper Middle Lower upper and middle Hydraulic factor. section. section. section. middle and lower sections. sections.

3,888 3,680 2,695 485 386 330 8.36 9.47 8.17 5^6 660 Slope...... 0.0030 O.OOD5 0.040 0.050 a 91 8.82 9.15 10.92 34,535 34,518 Independent computations of discharge were therefore made for the stretch between the upper and middle sections and for the stretch between the middle and lower sections. The coefficient of roughness (n) was determined at the time of the survey, before any computations were made. The relative correctness of the determina­ tion is shown by the close agreement in the computed discharge for the two stretches, which in turn increases confidence in the ac­ curacy of the final results. Elements of the cross sections are given in the following table: Cross sections used in determination of maasinwm discharge of Brushy Creek at Round Rock, Tex. Upper section. Distance from Distance from Distance from Distance from left edge of Depth of left edge of Depth of left edge of Depth of left edge of Depth of water. water. water. water. water. water. water. water.

Feet. Feet. Feet. Feet. Feet. Feet. Feet. Feet. 0...... 0 100...... 13.81 160...... 14.61 216...... 13.81 29 . . 7.81 107...... 14.21 182...... 13.81 237...... 9.58 47...... 7 97 115...... 19 01 185...... 13.31 318...... 5.80 47...... 9.47 120...... 12.60 188...... 13.81 410...... 3.39 70...... 10.97 125...... 10 Qt 202.. 14.30 0 70...... 12.47

Middle section. Distance from Distance from Distance from Distance from left edge of Depth of left edge of Depth of left edge of Depth of left edge of Depth of water. water. water. water. water. water. water. water.

Feet. Fed. Feet. Feet. Feet. Feet. Feet. Feet. 0...... 0 151...... 6.28 293...... 17.09 353...... 13.74 47...... 3.86 isn 6.34 307..:...... 18.30 365...... 12.92 87...... 8.67 212...... 8.01 341 18.37 ^65...... 8.93 111...... 8.88 217...... 10.40 347...... 17.09 377...... 0 121...... 6.01 240...... 14.09

Lower section.

Distance from Distance from Distance from Distance from left edge of Depth of left edge of Depth of left edge of Depth of left edge of Depth of water. water. water. water. water. water. water. water.

-Feet Feet. Feet. Feet. Feet. Feet. Feet. Feet. 0...... 0 155...... 11.00 205.. 14.95 297...... 11.32 43...... 1.42 157...... 12.25 230...... 13.95 307...... 8.10 97...... 4.68 157...... H ot 280...... 13.45 314...... 2.88 137...... 6.33 181...... 14.45 280...... 12.45 322.. .. . 0 FLOODS 12? CENTRAL TEXAS IN SEPTEMBER,

COLORADO EIVEB FLOOD. Colorado River heads in eastern New Mexico near the Texas line, at an elevation of less than 4,000 feet. It flows southeastward across the State of Texas and enters the Gulf of Mexico through near Matagorda. Its total length is about 600 miles, and it drains an area of 37,800 square miles. The head of the basin is in the semiarid plains, from which there is little surface run-off except during rains. Above Ballinger the river has a drainage area of 6,460 square miles, and at that place it enters a more mountainous country, through which it flows until it reaches Austin, where it enters the Coastal Plain. Below Austin the basin becomes narrower, its maximum width being only about 3$ miles. . The United States Geological Survey maintains gaging stations on Colorado River at Ballinger, Chadwick, Marble Falls, Austin, Co­ lumbus, and Wharton. The gages at Austin, Columbus, and Whar- ton are automatic recorders. No rise of consequence occurred at or above Marble Falls. Elevations and distances along Colorado River below Marble Falls, compiled chiefly from, topographic maps, are given in the following table: Elevations and distances t^ong Colorado River from Marble FaMs to mouth.

Descent between Distance. points. Elevation Point. above sea level. From Point Per source. to point. Total. mile.

Feet. Mile*. Mile*. Feet. Feet. Marble Falls...... 700 590 25 25 iio 4.4 Contour crossing...... 600 48 23 90 3.9 425 72 24 75 3.1 380 90 18 45 2.5 BftSfjTOJ) 310 124 34 70 2.1 SmithvUle...... o270 145 21 40 1.9 o200 210 65 70 1.1 Wharton «65 240 59 135 2.7 Mouth...... 0 305 45 65 1.4

* From report by Victor L. Lieb, assistant engineer to U. S. Engineer Corps, published in H. Doc,. 304, 6eth Cong., 1st sess. At Austin the river began to rise about noon on September 9 and rose from a stage of 0.7 foot (discharge, 215 second-feet) to a maximum of 19.4 feet (discharge, 75,700 second-feet) at midnight of September 10. Between Marble Falls and Austin the drainage area increases from 32,200 to 34,200 square miles. Of the intervening area of 2,000 square miles, 1,300 square miles is drained by , which enters the Colorado from the west about midway be­ tween Marble Falls and Austin. The basin of the Pedernales lies 30 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. entirely in the Edwards Plateau. From Johnson City to its mouth the river has cut a deep valley in the limestone formation and falls 510 feet in 40 miles. No reliable information is available regarding the flood in Pedernales River, but according to statements.in local newspapers it reached the highest stage ever known. The; peak discharge of Pedernales River at its mouth probably exceeded that of Colorado River at Austin, where the crest was reduced and the flood period prolonged by the retarding effect of and the much wider channel of the Colorado, which more than offset the inflow from the remaining 700 square miles between Marble Falls and Austin. The United States Weather Bureau maintains a gage on Colo­ rado River at Smithville, about midway between Austin and Co­ lumbus. The rise at that point has been described by Bunnemeyer * as follows: The flood wave moved rapidly downstream, and at Smithville~ Bastrop County, the stream rose from 1.4 feet on the evening of September 9 to 22.1 feet by 7 p. m. September 10 (flood stage 24 feet). The crest of the flood passed at noon, September 11, with stage 26 feet. The subsequent fall was equally rapid, and by 9 a. m. September 12 the water was within banks, and 10 hours later down to 12 feet. Smithville is located on a bluff forming the right or south bank of the river, but the left bank is low and the lowlands on that side were submerged for a distance varying from one-half to two miles at 4he peak of the flood. The damage to crops is estimated at $6,000, including loss of farm animals, and damage to roads and bridges at $2,000. That the crop loss was not greater is due to the fact that the first picking of the cotton crop had been made and that about 65 per cent of the corn had been gathered. All tributaries of the Colorado between Austin and Smithville were in flood. Onion Creek, which drains an area of about 310 square miles and enters from the west about 15 miles below Austin, rose rapidly to a height of 40 feet at the George Begg farm near its mouth. That is the highest stage recorded for this creek since the flood of 1869. . Below Smithville the rainfall was light and little water was re­ ceived from the tributaries. The maximum discharge decreased rapidly, and by the time the crest reached Columbus it was not high enough to cause much damage. At Wharton the river did not overflow its banks. Records of stage and discharge are summarized in the following table:

8 Monthly Weather Review, September, 1921, p. 493. GrtTADALUPE MVEB FLOOD.

Mean-daily gage height, in feet, and discharge, .in second-feet, of Colorado River during flood of September, 1921. *

Marble Falls. Austin. Columbus. Wharton. Date. Dis­ Gage Dis­ Gage Dis- Dis­ height. charge. charge. height. charge. height. charge.

Sept. 8...... 1.1 137 0.5 222 5.3 271 5.0 290 3.8 2,500 2.4 6,320 5.4 310 5.1 325 10...... 4.2 3,450 12.9 45,800 5.4 313 5.2 360 11...... 3.0 1,300 10.1 31,500 23.0 30,000 9.5 3,3|0 12...... 2.9 725 2.9 4,850 31.4 50,900 26.2 ,,27300 13...... 2.9 725 2.0 2,300 30.4 48,300 29.2 32,100 14...... 2.6 552 1.7 1,506 17.0 16,400 31.0 35,000 15...... 2.4 388 1.5 1,170 12.8 8,390 17.8 16...... 2.2 324 1.4 972 11.2 5,800 13.8 ^JB 17...... 2.1 286 1.3 840 10.2 4,410 11.6 5?270 18...... 2.0 252 1.2 720 9.6 3,550 10.5 4,160 / 4.5 4,196 / 19.4 75,700 / 33.8 57,300 / 31.6 34,409 MfMrimnTn ...... \Sept. 9, £ .30 p. m. \Sept. 10, 11 p. m. \Sept. 13, 7 a. m. \Sept. 14, 11.56a.m.

GUADALUPE EIVEB FLOOD. Guadalupe Kiver rises in the Edwards Plateau in the western part of Kerr County, flows southeastward for 283 miles, and enters the Gulf of Mexico through . Its drainage basin lies between the Colorado River basin on the northeast and the San Antonio River basin on the southwest and comprises 6,000 square miles. Comal and San Marcos rivers are the principal tributaries of the Guadalupe. Comal River has its source in fissure springs that have a practically constant flow of about BOO second-feet. It enters Guadalupe River at New Braunfels, a mile below the springs. drains an area of 1,380 square miles and enters Guadalupe River from the east, 1£ miles southwest of Gonzales. It has its principal source in fissure springs that rise near the city of San Marcos and ordinarily have a flow of about 150'second-feet* At New Braunfels Guadalupe River emerges from the Balcones escarpment and enters the Coastal Plain. Between Comfort and New Braunfels the stream has cut a deep canyon several hundred feet below the level of the plateau. The damage from overflow is slight in this stretch because little land below flood level is suitable far cultivation. Below New Braunfels large areas of bottom land are cultivated, and practically every year the crops are damaged by overflow. Some damage from overflow is also experienced near Comfort. FLOODS IN CENTRAL TEXAS IN, SEPTEMBER, 1921.

Elevations and distances along Guadalupe River from source to mouth, ^

Descent between Eleva- Distance. *> points. tion Point. above sea From Point to Per level, o source. point. Total. mile.

Feet. Mites. Miles. Feet. Met. 2.300 Kerrvffle...... 1.610 36 36 690 19.2 1,400 54 18 210 11.7 1,300 62 8 100 12.5 1.010 90 28 290 10.4 580 128 38 430 11.3 450 147 19 130 , 6.* Gonzales...... 270 183 36 180 5.0 Victoria...... 35 255 72 235 3.3 Mouth...... 0 283 28 35 1.2

a Estimated. * Measured along largest bends of stream. Gaging stations are maintained by the United States Geological Survey on Guadalupe River near Comfort, at New Braunfels, near Gonzales, and below Cuero and on San Marcos River at Ottine, Staff gages are used at Comfort, Gonzales, and Ottine and automatic recorders at New Braunfels and Cuero. At Comfort the flood was not unusually high. The maximum stage recorded was 32.0 feet, on September 9. No discharge measurements have been made at the station above a stage of 3.0 feet, but a study of the cross-section and low-stage measurements shows that the dis­ charge at a gage height of 12 feet could not have been greater than 5,000 second-feet. At New Braunfels the river began to rise rapidly from a stage of 3 feet at noon on September 9 and reached a maximum of 28.5 feet (discharge, 56,000 second-feet) at 2 a. m. on September 10. It be­ gan to fall immediately at nearly as rapid a rate as it rose and reached a stage of 12.6 feet (discharge, 12,800 second-feet) at 4 p. m. After that time the rate of fall decreased, and the river did not return to a stage of 3 feet until about noon on September 14. Below New Braunfels the maximum discharge decreased rapidly, and by the time the, crest reached Gonzales the maximum stage was only 31.4 feet, which, though sufficient to cause some damage to crops by overflow, was not an unusually high stage. By the time the crest had reached Cuero it was well below the danger line^ The following table shows gage heights and discharge at the gaging stations ^on Guadalupe and San Marcos rivers during this flood: THE FLOOD AT SAN ASTTOHIO. 33 Mean daily ffaffe height, in feet, and discharge, in, second-feet, of GuadaJnpe and San Marcos rivers during flood of September, 1Q21.

Guadalupe Guadalupe Guadalupe Guadalupe San Marcos Elver near River at New River near River below River at Comfort. Braunfels. Gonzales. Cuero. Ottiae. Date. Gage Dis­ Gage Dis­ Gage Dis­ Dis­ Gage Dis­ height. charge. height. charge. height. charge. «& charge. height. charge.

Sept. 8...... 1.66 40 1.83 407 1.2 510 1.36 513 2.05 151 9...... 11.4 04,700 6.6 6.580 1.2 510 1.50 595 2.5 201 10...... 5.0 ol.lOO 17.9 25,600 1.3 550 1.46 571 19.4 4,890 11...... 63.0 372 5.1 3,060 23.2 c!4, 100 2.35 1,140 31.2 9,640 12...... 2.04 78 3.25 1,150 30.7 «29,300 9.6 6,780 16.0 3,770 13...... 1.88 54 3.3 1.180 24.7 d5,900 13.0 9,600 8.0 1,320 14...;...... 1.84 49 3.1 1,060 7.1 3' 170 14.8 11,100 6.2 874 15...... 1.82 47 2.90 940 5.3 2,320 15.9 12,100 4.2 475 16...... 1.80 44 2.80 880 4.4 1,900 6.8 4,540 4.6 547 17...... 1.78 42 2.81 886 3.9 1,660 3.9 2,280 4.4 511 18...... 2.68 814 3.6 1,530 3.55 2.010 4.2 475 19...... 1.78 42 2.63 786 3.6 1,530 3.30 1,820 4.1 457 20...... 1.76 40 2.55 742 3.3 1,400 3.15 1,710 3.9 422 Tw»jrfmnT»i / 12.0 05,000 / 28.6 56,600 / 30.7 29,300 / 16.4 12,500 f 35.5 al9,400 \Sept.9, a.m. \Sept. 1C ,2a.m. \8ept. IS ,7a.m. \Sept. 1 5, 12.15 \Sept. 1 1, 8.30 a. m. a.m.

o Determined from extension of rating curve and subject to considerable error. 6 Estimated. * At a stage of about 22 feet (discharge, 11,400 second-feet) water begins to enter an old channel locally known as "Cross Timbers" on right bank, 1 mile above gage, and returns to main channel below gage. Consequently, records of flow greater than 11,400 second-feet do not represent the total flow of the stream but only that of the main channel.

THE FLOOD AT SAN ANTONIO.

GENERAL FEATURES. The city of San Antonio was founded by the Spaniards in 1718. The population in 1920 was 161,379. It is an important railroad center and distributing point for a vast agricultural area in south­ ern Texas. The railroads that enter the city are the Gftilf, Harris- burg & San Antonio (Southern Pacific); Missouri, Kansas & Texas; International & Great Northern; San Antonio & Aransas Pass; and San Antonio, Uvalde & Gulf. Fort Sam. Houston, headquarters for the Southern Military Department, is just east of the city limits. San Antonio lies on the Coastal Plain at the edge of the Balcones escarpment. San Antonio Eiver and San Pedro, Alazan,9 and Apache 10 creeks flow through the city limits and cause damage in times of flood. The location of these streams and of their drain­ age basins is shown on figure 1 (p. 8). San Antonio River ordinarily has its source in fissure springs in the Balcones fault zone near tfye northern limits of San Antonio. The continuation of the river channel above the springs is known as Olmos Creek, which is the principal source of the disastrous floods that overwhelm the business part of the city. Olmos Creek heads 11

Shown as Martincz Creek on San Antonio topographic map but locally known as Alazan Creek. Martinez Creek is tributary to Alazan Creek from the north. 10 Shown as Salsamora Creek on San Antonio topographic map but locally knows as Apache Creek. 34 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. miles northeast of the springs and has practically no flow except immediately after rains. The flow of the springs ranges from less than 10 second-feet during long periods of drought to more than 200 second-feet in periods of heavy rainfall. San Antonio River enters the Gulf of Mexico through San Antonio Bay, 120 miles southeast of the city. The topography of the Olmos drainage basin and of the upper part of the Alazan and Apache drainage basins is rolling and the watercourses have been cut deep. As the soil con­ sists of a thin layer of rather impervious clay underlain with lime­ stone the run-off of the region and the rate of discharge of the streams are probably rather high. About 25 per cent of the area has been cleared of trees and brush and much of the cleared land has been devoted to agriculture. Practically the entire drainage basin of San Pedro Creek is covered with residences, business houses, roads, and paved streets. San Pedro Creek has an ordinary flow of about 10 second-feet, derived from fissure springs in San Pedro Park, 1-| miles northwest of the center of the city. Alazan and Apache creeks have no flow except after rains. The following table has been compiled from the topographic map of the San Antonio quadrangle except that part of the streams lying north of the quadrangle, which is based on the progressive military map prepared by the United States Engineering Corps at Fort Sam Houston.

Mevations and distances along San Antonio Hirer and tributaries.

' San Antonio Hirer.

Descent between Distance." points. Elevation Point. above sea level. From Point to source. point. Total. Per mile.

Feet. Miles. Mites. Feet. Feet. bl,000 740 7.2 7.2 260 36.1 675 11.5 4.3 65 15.1 670 11.9 .4 5 12.5 Houston Street ...... 635 16.9 5.0 35 7.0 615 18.9 2.0 20 10. J San Pedro Creek...... 595 21.5 2.6 20 7.7 580 22.9 1.4 15 10.7

San Pedro Creek.

730 660 1.2 1.2 70 58.3 R44 3.0 1.8 16 8.9 Alazan Creek...... 623 4.2 1.2 21 17.5 Mouth...... 595 6.0 1.8 28 15.6

a Distances measured along bends of streams. 6 Approximate. THE FLOOD AT SAF ASTTONIO. as-

Elevations distances alonff Son Antonio River and tributaries Continued. Alazan Creek.

Descent between Distance. points. Elevation Point. above sea level. From Point to source. point. Total. Per mite.

Feet. Miles. Miles. Feet. Feet. Source...... 61,000 Contourcrossing...... 850 2.7 2.7 150 55.6 Do...... 750 4.5 1.8 100 55.« West End Lake...... 680 6.3 1.8 70 38.9 Martinez Creek...... 648 8.1 1.8 32 17.8 Mouth...... 623 10.1 2.0 25 1Z5

Apache Creek.

61,000 Contour crossing...... 850 1.8 1.8 150 83.3 Do...... 800 3.3 1.5 50 S3. 3 Do...... 7W1 5.1 1.8 50 27.8 Do...... :...... 700 7.3 2.2 50 22.7 Elmendorf Lake...... 665 10.3 3.0 35 11.7 623 13.4 3.1 42 13.5

6 Approximate. The following is quoted from an excellent concise description of the flood:" Olmos Creek, north of the city, overflowed its banks and swept into the city about 10 p. m., reaching its crest within an hour. At the center of the city the river, not more than 2 feet above normal at 6 p. in., rose with increasing rapidity, overflowing its banks immediately north of the business district at midnight. Within an hour a large part of the business quarter was inun­ dated, and-the crest was reached about 2 a. m. [See PI. VII, this report] At this time, six of the principal north and south streets were channels carrying the swift flood waters across a great bend of the river. The water In these channels ranged in depth from 1 to 8 feet where they crossed Houston Street, the principal east and west street, and at places in the business area the water was 12 feet deep. North of Houston Street the water spilled through five streets over a low divide to the west into the upper San Pedro basin. In the central district the river rose from 2 to 8 feet above all bridge floors, and at each bridge a great mass of dSbris was caught by the piers and railings. [See PL V, A and B, this report] In the northern part of the city the rush of water in the meantime had inundated Breckenridge Park along the river and all the adjoining low resi­ dential and industrial areas. Hundreds of homes were flooded to depths vary­ ing from a few inches to 8 and 10 feet. In this section a few poorly built frame dwellings were washed away. A scum of heavy fuel oil from some of the industries coated everything it touched and incidentally everywhere below left clear and unmistakable high-water marks. In the business district all the basements were filled, and most of the first floors were flooded to depths from a few inches to 12 feet The severity of the flood crest in the business district was undoubtedly somewhat mitigated by the pondage of water and its retention by houses and " Harriett, c. T., The -flood of September, 1921, at San Antonio, Text: -Am. Sqe Civil Bug. Proc., Nor., 1921, pp. 443-454. 36 FLOODS IN CENTRAL TEXAS IN SEPTEMBEB, 1921.

other obstructions in the valley above. In like manner the obstructions formed by the business district, the storage of water in basements, and probably, also, a considerable absorption from the river channel into the gravel beds which underlie many parts of the valley reduced the flood markedly below. This effect was partly counteracted, as far as the flooded residence district just below the business section was concerned, due to two low dams and the blocking of two railway trestles with drift, all of which were situated about 1 mile below the center of the city. Below these obstructions the flood was carried by the river channel and natural overflow bottoms, which have not been encroached on by retaining walls and buildings to the same extent as in the central upper portions of the valley. Notwithstanding many narrow es­ capes, there was practically no loss of life in the valley of San Antonio River. Owing to the comparatively light rainfall in the city, the small upper San Pedro basin did not flood badly, and its crest passed before the overflow from San Antonio River. The latter inundated a number of buildings in the western edge of the business district, the water in it and the San Pedro forming a continuous sheet. The lower San Pedro traverses the western part of the business district and, together with Alazan and Apache creeks, runs through some of the cheaper and more thickly settled residence districts, especially the Mexican quarter. i Alazan Creek, with comparatively steep grade from the high ground north­ west of the city, and its two branches flooding simultaneously, swept suddenly through the thickly settled low areas along its banks, playing havoc with bridges and tearing loose and destroying many cheap frame houses. It was in this section that the principal loss of life occurred. Conditions on Apache Creek were similar, but there was less development along its banks and the loss was smaller. Below the confluence of Alazan and Apache creeks, the San Pedro again passes through a thickly settled area, where there was consider­ able damage and loss of life. The following notes on the flood are based on a report prepared by the United States Engineer Corps at Fort Sam Houston, Tex.: Olmos Creek began to rise rapidly on the night of September 9 and at 9 p. m. had overflowed its banks at the dam site for a reservoir 1 mile above the springs. It reached its crest there at 11 p. m. San Antonio River overflowed its banks in Breckenridge Park at 10.30 p. m. and at 10.50 p. m. was flowing 1 foot deep across River Avenue. There the water rose at the rate of 1 foot in 5 minutes until it reached its crest at 11.50 p. m. It remained at crest stage for about 30 minutes and then fell at the rate of 1 foot in 20 minutes. On Josephine Street water reached its maximum height at 12.28 a. m. September 10. The depth of water at that time was 5 feet at a point 200 feet west of the ordinary channel and 7 feet at a point 800 feet east. At 12.30 a. m. the maximum height was reached at the intersection of Grayson Street and the river; the depth of water was 7 feet over the road. At the Eleventh Street Bridge (Grand Avenue) wate^r rose to a height of 3 feet over the floor of the bridge between 11.15 p. m. and 12.25 a. m., reached a maximum of 8 feet at 1 a. m., and began to fall at 1.30 a. m. THE EbOaD AT SAN ANTONIO* 87 At the intersection of JFourth Street and Lexington Avenue the water was 2 leet above the banks at 11.30 p. ia., and at 12.25 a. ml it reached its maximum stage of 6.8 feet. At 1.30 a. m. it had fallen 1.4 feet; at 2 a. m., 1.6 feet; at 3.30 a. m., 1.8 feet; and at 3.40 a. m., 2 feet. At 8.30 p. m. water extended across the main channel from wall to wall at the Romana Street Bridge, and at 11 p. m. it was 10.5 feet below top of wall. At midnight water began to overtop the banks and flow down Sk Mary's Street past the Gunter Hotel. At 12.07 the water was too deep to cross at the intersection of St; Mary's and Travis streets, and at 12.12 a. m. it was 6.2 feet deep. At 11.30 p. m. the water was 11.6 feet below top of the rail of the Travis Street Bridge, and at 12.05 a. m. it was 6.65 feet below. At 1.30 a. m. water was 10 feet deep at the intersection of Navarrp and Martin streets. At 11.37 p. m. water was 8.25 feet below top of rail of the Houston Street, Bridge; at 12.01 a. m., 6.35 feet below the rail; and at 12.15 a. m., 5.09 feet below the rail. At 12.45 a. m. water was 1 foot deep at the intersection of Jeffer^ son and Houston streets. At 1.40 a. m. water was on sidewalk at the intersection of Avenue C and Houston Street. ' That was practically the crest. At 3 a. m. water was still in the street at that point. At 12.30 a. m. water was 2.8 feet below top of the rail of the Commerce Street Bridge. The operation of the city's pumping plant at 106 Market Street ceased at 12.30 a. m. because the fires were put out by water. Pumping was resumed at 3 p. m. September 11. At 1.30 a. m. water reached the top of the walls of the Villita Street boiler house of the Public Service Co.'s electric light plant, and at 1*45 a. m. it was necessary to pull switches. At 11.30 p. m. telephone service was discontinued.

DETERMINATION OF DISCHARGE. A gaging station has been maintained on San Antonio Kiver at San Antonio by the United States Geological Survey since January 26, 1915. A vertical staff gage attached to the downstream side of middle pier of Commerce Street Bridge was used from January 26, 1915, to February 28, 1916 ; and a vertical staff gage attached to the upstream side of second bent of Presa Street Bridge was used from February 28, 1916, to April 8, 1920. On April 8, 1920, a Gurley water-stage recorder was installed at the downstream side of the right abutment of South Alamo Street Bridge. (See PI. VI, A.} jM»iflee&i of js^fikaent .height to cause damage occurred between the installation of the first gage in 1915 and the flood x>f September, 1921. 38 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. The following table shows the maximum, minimum, and mean monthly discharge from January 16, 1915, to , Monthly discharge Of San Antonio Rfoer at San Antonio, Te®., for the ending Sept. SO, 1915 to 1921.

Discharge in second-feet. Run-off in Month, acre-feet. Maximum. Minimum. Mean.

1915. February...... 104 100 103 5,720 March...... 117 104 106 6,520 April...... 916 112 219 13.060 May...... 266 139 179 11,000 Jtrae...... 167 122 140 4,330 July...... 217 122 171 10,500 Angust...... 246 . 197 222 . 13,600 Septembts.^...... 209 117 136 8,090 76,800 1915-16. October...... 124 106 116 7,130 November...... 108 100 102 6,070 December...... 111 92 94.4 5,800 January ...... 96 90 92.7 5,700 February...... 92 82 85.0 4,890 March... 81 72 75.3 4,630 April...... 520 92 111 6,600 May... 206 81 98.9 6,080 Jane...... 106 77 86.2 5,130 July...... 130 75 81.6 5,020 371 82 95.9 5,900 850 78 110 6,550 The year...... 850 72 95.7 69,500 1916-17. October...... 90 81 82.8 5,090 November...... 81 81 81.0 4,820 December...... 81 81 Sl.O 4,980 January...... 81 78 78.9 4,850 February...... 78 78 78.0 4,330 March...... ;...... 147 38 68.9 4,240 April...... 41 28 34.1 2,030 May...... :;...... 81 26 41.1 2,530 June...... 35 23 25.9 1,540 July...... 70 23 26.8 1,650 August...... 26 24 24.4 1,500 September ...... 27 23 23.6 1,400 The year...... 147 23 53.8 39,000 1917-18. October...... 27 23 23.9 1,470 30 23 23.4 1,390 25 23 23.8 1,460 24 23 23.4 1,440 25 22 22.9 1,270 March .... 49 14 22.2 1,360 AjHiL...... 309 18 48.3 2,870 May...... 1,860 12 82.9 5,100 46 12 16.5 982 July...... 22 * 12 14.8 910 513 12 30.9 1,900 21 11 14.0 333 The year...... 1,860 11 29.0 21^000 1918-19. 70 8.0 15.1 928 17 7.0 9.52 566 124 7.0 17.7 1,096 260 9.5 33.9 2,060 76 33 38.7 2,161 March...... 63 24 31.5 i,m April...... 159 31 46.6 2,770 May...... 145 3036 ' 53.6 3j3 0 601 66.3 3,950 July...... 634 70 116 7,130 August...... 112 80 91.9 5,650 1,820 81 186 11,100 The year...... <...... 1,820 7,0 58.8 43,700 THE FLOOD AT SAN ANTONIO. 39

Monthly discharge of San Antonio River at San Antonio, re»., e*c. CanttaBfid,

Discharge in second-feet. Boo-Off In Month. acre-feet. Maximum. Minimum. Mew.

1919-20. Oetafeer...... 1,040 152 295 18,100 398 2131 _ 233 13,900 5J88 197 208 'SJSK January...... 478 201. 258 15,900 222 213 216 12,400 March ...... 213 191 204 12,500 April...... 238 168 183 10,900 May...... 182 162 168 10,300 -June ...... 175 150 158 9,400 July...... 162 132 143 8,790 146 128 134 8,240 129 119 123 7,320 The year...... 1,040 119 194 141,000 1920-21. -October...... 126 111 119 7,320 182 107 118 7,020 114 107 111 6,820 _Jftnr(ftry , . . , . , , 114 100 109 6,700 105 94 98.4 5,460 Hareh...... 414 112 144 8,850 April...... 309 103 123 7,320 Mav 118 85 99.7 6,130 250 77 107 6,370 July...... 95 72 84.7 5,210 74 64 69.2 4,260 3,190 65 250 14,900 3,190 64 119 86,400

Discharge measurements of San Antonio River at San Antonio, Tex., in 1921.

No. of Dis­ meas­ Gage charge in ure­ Date. height second- Method. Location. ment. in feet. feet.

81 .Mar. 11 2.51 391 Current meter.... At gage. 87 Sept. 8 1.38 76 .....do...... Do. 88 9 3.31 622 .....do...... Do. 89 olO 20.14 15,300 2,000 feet below gage. 90 10 8.94 2,200 Current meter.... Nueva Street Bridge, 1.3 miles above gage. 91 10 5.57 866 .....do...... Railroad bridge, 400 feet above gage. 92 21 3.13 153 .....do...... At gage. 93 Oct. 4 2.90 117 .....do...... DO.

* Measurement of cross section and slope were made September 15-47. NOTE. Numerous discharge measurements made at stages less than 2 feet are not listed above. A gaging station has been maintained by the United States Geo­ logical Survey on San Pedro Creek at San Antonio since July 20, 1916. Prior to March 14, 1921, a vertical staff gage was attached to wall of building at 713 Commerce Street on the ^upstream side of bridge. On March 14, 1921, a Gurley water-stage recorder was installed at the lower end of the covered part of the concrete-lined channel that extends from Durant to Arsenal Street. The South Alaino Street Bridge failed during the flood, but the right abutment to which the recorder was attached was not damaged. (See PL VI, B.) 40 FLOODS IN CE&TBAL TEXAS IK SEPTEMBER, 1921. Figure 3 shows the gage heights on San Antonio River and San Pedro Creek during the flood of September, 1921. The water-stage recorder on San Antonio Kiver did not operate from 1.45 to 5.45 a. m. on September 10, when the stage was above 15 feet. The esti­ mated part of the graph, however, can not be greatly in error, be­ cause the maximum height of 20.14 feet is clearly marked in the gage house, and the time it occurred has been definitely established as 3 a. m. by statements of persons living near the gage, who also have stated that the crest stage lasted about 30 minutes.

'* -Ma .! ,fJ A. t m nt KG r' >90i e ho us e 1 . ; \

i f. t! f" t( t r6 W < ! Sf set ^ x n

/ \ (*' J in At it in o It 'e / \ / \ i/* \ \ ;* - - - - - f\ «a k a JS id *, Sc In tic IV IF it' :r s \ RJ t'l< fo f> w f Oil er m r>$ n te c ha "' tl /i V ' ',ri \ fj p c, \li 1 \ s * e ** r i* & l « )l ^ *"|f * / il k» HH In * "*, 'fine ff>Iti a '/< :fif \ 1 > V S .J } t o 'v^ V i (\r ,i

*- A.M. -H«- P.M. -»«- A.M. -» -P.M.- -,---»! _J FIGUBB 3. Gage-Beight graph of San Antonio River and San Pedro Creek at San Antonio, Tex., during flood of September 9-10, 1921. The following table, which gives the mean hourly gage height, dis­ charge, and run-off in acre-feet from noon September 9, several hours before the flood began, to 12 midnight September 10, when the river had fallen to nearly normal stage, shows that the most disastrous flood in the history of San Antonio was caused by less than 8,000 acre-feet of water. THE FLOOD AT SAN ANTONIO. 44.

height and discharge of San Antonio River at South Afonw Street Bridge, San Antonio, Tea?., during flood of September 9-10, 1921.

September 9'. September 10. Hour. Gage Dis­ IMs- height. charge. Run-off. 4E. charge. Run-off.

A.M. Second- Acre- Sectmt- Acre* Feet. feet. **. Feet. feet. feet. ia-1...... 1.30 68 11.93 4,690 388 1-2...... 1.45 80 14.30 6,040 499 W...... 1.92 176 15 18.50 12,100 1,000 3-4...... 2.52 369 30 19.40 13,800 1,140 4-6...... 2.98 534 44 17.49 10,300 851 5-«...... 3.13 588 49 15.40 7,300 603 6-7...... 3.29 645 53 13.38 5,190 429 7-8...... 3.44 701 58 11.61 3,760 311 8-0...... 3.18 606 50 9.95 2,740 226 9-10...... 2.92 512 42 8.72 2,090 173 10-11...... 2.67 422 35 7.67 1,640 136 11-12...... 2.68 426 35 6.40 1,140 94 P.M. > is-i...... 2.74 447 37 5.53 849 70 1-2...... 2.60 397 33 5.17 741 61 2-3...... 2.93 516 43 4.92 666 65 3-4...... 3.23 624 52 4.59 567 47 4-6...... 3.07 566 47 4.36 498 41 5-6...... 3.10 577 48 4.15 435 36 6-7...... 3.08 570 47 3.95 375 31 7-R 2.96 527 44 3.83 339 28 S^t...... 3.87 860 71 3.78 324 27 9-10...... 5.67 1,530 126 3.72 306 25 10-11...... 7.78 2,430 201 3.67 29i 24 11-12...... 10.02 3,510 290 3.63 279 . 23 3.44 737 1,463 8.59 3,190 6,318

A study of the overflowed area and of the condition of the channel, especially where debris collected at bridge piers and elsewhere, shows that the relation between stage and discharge at any section above the South Alamo Street Bridge would have been decidedly un­ certain at any time during the flood and that the flow of water below the gage was relatively unobstructed and permanent. Even after the bridge had failed and partly collapsed (PL VI, B) the capacity of the channel was not reduced so much as it was at many other bridges which did not fail but at which greater quantities of de"bris had col­ lected (PI. V, A and B). Practically the only effect the failure of the bridge had on the rating was caused by a few small blocks of concrete, which broke loose and lodged in the channel below the gage. That effect has been determined definitely by comparison of measurements made before and after the flood at relatively low stages. (See fig. 4 and discharge measurements on pp. 38-39.) The difference in discharge shown by the curves, considered as a per­ centage of total discharge, would be less for high stages and would be negligible at the crest; therefore, in the absence of definite infor­ mation as to whether the blocks of concrete found in the channel after the flood were dropped there before or after the time of the maximum stage, both curves have been drawn through measurement 89, which represents the discharge at the maximum stage. 42 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. The graph of the gage heights of San Pedro Creek is shown irf" connection with that of San Antonio Kiver because of its possible value in determining the volume of water that passed over the divide from San Antonio River into the San Pedro channel. The riser shown in figure 3 as coming from San Antonio Kiver, obviously did not come from any other source, as the floodwater from the San Pedro basin had passed the gage several.hours earlier. It is impos-

-Curve ust'd sfier^ fet'/ure at"bridge

Curye used before _fsi/ure ofbridge, tvb/c/> occurredat 3 a.m.Sept. JO

J-

tOOO 2,000 3OOO 4,000 S.OOO 6.000 ?OOO &OOO SOOO IQ/OOO HOOO E000 OflOO MIP00 GflOO DISCHARGE IN SECOND-FEET FIODBB 4. Rating curves of San Antonio River at South Alamo Street Bridge, San An­ tonio, Test. sible to compute with reasonable accuracy the discharge of San Creek during the flood, because of an indeterminate volume of water from Alazan £5reek. The following rating table for the San Pedro gage is submitted for the benefit of anyone who may be sufr ficiently interested to attempt to compute the volume of water that passed from the San Antonio Kiver basin down the San channel: Rating table for San Pedro Creek at San Antonio, Tea;.

Dis­ Dis­ Dis­ Dis­ Gage height. Gage height. charge*. Gage height. charge. Gage height. charge. charge.

Feet. Sec.-ft. Feet. Sec.-ft. Feet. Sec.-ft. Feet. Sec.-ft. O.5...... 8 3.0...... 297 5.0...... 728 7.0...... 1,220 1.0...... ?s 3.5...... vu /5.5...... 848 7.5...... 1,345 1. 5...... an 4.0...... 500 6.0...... 970 8.0...... 1,470 o n 137 4.5...... 610 6.5...... 1,095 2.5...... 210

This rating table is directly applicable to San Pedro Creek only when backwater does not affect the gage heights. When backwater THE KM)OD AT SAN ANTQJTJO. began and ended is not definitely known, but it probably began about 10 p. m. September 9 and lasted until sometime in the early morning of September 10. The rating table for San Pedro Oteek is baaed.on current-jneter measurements for discharges of less than 200 seconfl-feet and is extended for more than that rate by using Kutter's formula. Above a 6-foot stage the lined sectk>Eujof,tbe chan­ nel overflows and the rating is subject to large errors. Numerous determinations of the maximum discharge of San An-- tonio River and its tributaries were made by several engineers im* mediately after the flood, and all available results of these measure­ ments are summarized in the following table: Measurements of maximum discharge of San Antonio River and its tributaries near San Antonio, Test., during the flood of September 9-10, 1921.

Ve­ Dis­ Dis­ Drain­ charge Stream. Location of measurement. Date and time of locity age pa- crest. per charge. area. square second. mite.

feet. Sec. ft. Sq.mi. See.-fl. 6 0 25,000 26.4 oas Do...... do.*...... do...... 31,000 26.4 1,174 San Antonio River. Houston Street 6- ...... Sept. 10, 1.45-2.15 ...... 23,700 34.3 601 a.m. Do...... do...... 23,500 34.3 685 Do...... Sept. 10, 2.45-3.15 9 7 12,541 35.6 352 a.m. Do...... 2,500 feet below South Ala­ 7.35 13,340 36.5 366 mo Street Bridges Do...... doA...... 7.51 15,300 36.5 419 Do...... 7 IM 42,427 85.0 499 San Pedro Creek.. Above Alazan Creek a...... 1.5 2,025 2.7 750 Do...... 5.3d 32,443 46.5 698 above mouth.o * Alazan Creek...... Below Martinez Creek « ..... Sept. 9, llp.m..... 5.02 25,900 17.1 1,515 Do...... 24,000 17.7 1,356 Creek.a Do...... 28,000 18.7 1,497 Apache Creek..... 4.000 feet above mouth a..... Sept. 10, 12-2.45 a.m. 4.85 22,560 23.8 948 Do...... 19,000 24.0 792

, oMadebyC. K. McDonald, top apMcal draftsman (formerly hydrographer, U. S. Reclamation Serv- ice)i office of Engineers, Eighth C< sarea. 6 Bartlett, C^T-TTheflood of S« mber, 1921, at San Antonio, Tex.: Am. Soc. Civil Bag. Proa, March, , j by R. G. HempHU, irrigation engineer, Bureau p/ Public Roads, U. S. Dspt, Agr. 'Sea pp. 44-45 for detailed description of tMs measurement. d Made by engineers of U. S. Geol. Survey and Texas Board of Water Engineers. See pp. i3-44 for detailed description of this measurement. The reduction in maximum discharge of Olmos Creek between the dam site, above the city limits, and the point of measurement 2,500 feet below the South Alamo Street Bridge was due principally to temporary detention of the. water by overflow between the two points. The overflowed area was about 4 miles long, and its average width was more than one-third of a mile. (See PI. VII.) The quantity of water thus temporarily held at the peak of the floxxl no doubt amounted to several thousand acre-feet. The measurement of the maximum discharge of San Antonio River by the United States Geological Survey was made about 2,500 feet below the South Alamo Street Bridge, where the river was con- FLOODS IN CENTRAL TEXAS IJT SEPTEMBER, 1921. fined within narrow limits and the channel was favorable fo* a relatively accurate determination by the slope method by using Kutter's formula. The high-water slope was well marked by floating oil. A stretch of straight channel, 800 feet long, was selected for the measurement. The center of this stretch is near the intersection of Temple Street and the river. Cross sections were measured every 100 feet, and the slope was carefully measured on both banks. The slope on the left bank for the upper 375 feet and the lower 50 feet of the stretch selected was irregular and differed from that on the right bank. The left bank, which is bordered by buildings and other obstruc­ tions, was overflowed, but the right bank, which is high, was not overflowed. For the intervening 3T5 feet the slope of the two banks was practically the same and the mean slope was used, the area being the mean of the area at stations 400, 500, 600, and 700 feet below the upper end of the stretch. Figure 5 shows the mean cross

Maximum stage Sept 10, Bfl ~-~- - i - | -.Part f 7 -^x a

x A >/*/4 / 50iooio"° NOU.VA313J.33JNl /« king do vnstrea HYDRAULIC FACTORS Part A ' fart 3 x A...... /.Tea. . . 248 ^ ^"X / S...... OOISS...... eoiss X ^« 'ina/y s, #-...... 035...... oso ate I C...... +.64.82...... A '.SO- r V...... O.Z3., ...... \ . '. Q.~.-...flf,aee...... 46, 1 Total Q * IS,3OO second- feet V.^ 1 i 0 80 100 GO HO ISO 180 ZOO 220 2W ISO 280 300 3ZO 34 DISTANCE IN FEET FROM BASE LINE FIGUBE 5. Mean cross section of San Antonio River used in determination of discharge at maximnm stage of flood of September 9-10, 1921. section and a tabulation of the hydraulic factors. In making the computations th cross section was divided into two parts because of differences in roughness and in mean depth. The channel in the slope stretch is straight and relatively clean and unobstructed as compared with other parts of the river. Plate VIII, J., is a view of the channel looking upstream from a point on the right bank just below the lower end of the slope stretch, and Plate VIII, B^ is a view from the right bank near the center of the slope stretch, showing the condition of the left bank and the position of the flood line. The following detailed description of the maximum discharge ol San Antonio River at Houston Street, in San Antonio, during the flood of September, 1921, was prepared by Mr. Hemphill. The meas­ urements made by him are summarized in the table on page 43. At the corner of Jeffersan Street and the alley between Jefferson and Travis streets the elevation of the water surface was 98.80 feet (datum assumed). U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 488 PLATE V

A. BRIDGE ACROSS SAN ANTONIO RIVER NEAR PIONEER MILLS AFTER FLOOD OF SEPTEMBER 9-10, 1921.

B. BRIDGE ACROSS SAN ANTONIO RIVER AT ROMANA STREET, SAN ANTONIO, TEX., AFTER FLOOD OF SEPTEMBER 9-10, 1921. U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 488 PLATE VI

A. WATER-STAGE RECORDER AT SOUTH ALAMO STREET, SAN ANTONIO, TEX.

B. SOUTH ALAMO STREET BRIDGE, SAN ANTONIO, TEX., AFTER FLOOD OF SEPTEMBER 9-10, 1921. U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 488 PLATE VII

98°3O'

98°3O'

3 ABIes MAP OF SAN ANTONIO, TEX., SHOWING AREA FLOODED SEPT. 9-10, 1921

U. S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 488 PLATE VIII

A. VIEW OF SAN ANTONIO RIVER SHOWING SLOPE STRETCH USED IN DETERMI­ NATION OF MAXIMUM DISCHARGE DURING FLOOD OF SEPTEMBER 9-10,1921.

B. VIEW FROM RIGHT BANK NEAR CENTER OF SLOPE STRETCH SHOWN IN A.

~\ THE FLOOD AT SAN ANTONIO. 46

At the corner of Jefferson and Houston the elevation of the surface was, 98.08. The distance between these two points is 151 feet, which shows that the slope of the surface was 0.0053. The width of Houston Street at Jefferson is 60 feet. The stream coming through Jefferson Street carried floating objects across Houston Street in 5 to 6 seconds, showing a surface velocity of 10 to 12 feet per second. Applying a factor of 0.8 to the surface velocity gives a mean velocity of 8 to 9.6 feet per second. To be conservative, assume the true mean velocity to be 8 feet per second. Then substitute in the Kutter-Chezy formula A=148 8= 0.0053 j»= 53 v=* 8 . r= 2.8 which will give n a value of approximately 0.027. This appears to be high but is probably accounted for by the rough condition of the bottom of the channel after the wood block pavement had been carried away; breaks In the sides of the channel due to doorways, windows, and protruding ledges; iron poles at the curb on the west side of the street: This value of n is used in computing the discharge of the other channels through which the water passed except the main channel of the river above the Houston Street Bridge, for which a value of 0.03 was used to compensate for the bends in the channel. Jefferson, Street. A=148; t?=8; discharge, 1,184 second-feet. Nav&rro Street. Elevation of water surface at Travis and Navarro streets on St. Anthony wall, 98,85. Elevation at steps across street on church building, 98.84. Average elevation at Travis Street, 98.84. Elevation at Houston Street at northeast corner, 97.37. Elevation at northwest corner, 97.18. Average elevation at Houston Street, 97.27. This drop of 1.57 feet in 267 gives a sur. face slope of 0.0059. A=200; p=68.8; r=4.22. s .0059. w=.029; v=10.3; discharge, 2,987 second-feet. &#. Mary's Street. Elevation of water surface at southeast comer of Travis and St. Mary's streets, 98.56. Elevation at southwest earner, 98.84. Average, 98.70. Elevation at Houston and St. Mary's streets, 97.24. This drop of 1.46 feet in 352 gives a surface slope of 0.0041. A=449; p=74.4; r=6.X)3; 5=0.0041; «=0.029; v=10.9; discharge, 4,939 second-feet. Mam, channel (above Houston Street Bridge). Elevation of water surface at bridge near St. Mary's and Fannin streets, 99.75. Elevation at bridge near St. Mary's and College streets, 96.74. This drop of 3 feet in 2,500 gives a slope of 0.0012. A=l,243; p=107; r=11.36; s=0.0012; w=0.03; v=8.8; discharge, 10,930 second-feet. Soledad Street. Average elevation of water surface at Soledad and Houston streets, 97.27. Average elevation at Soledad and Travis streets, 98.15. The drop of 0.88 foot in 320 gives a slope of approximately 0.0028. A=253; p=68.6; r=3.69; s=0.0028; w=0.029; t>=6.5; discharge, 1,644 second-feet. Streams flowing west from Soledad Street. Second-feet. Travis Street: 55 wide, 4 deep, 4 feet per second- 880 Salinas Street: 28 wide, 2.5 deep, 4 feet per second. ___ 280 Alley: 20 wide, 2 deep, 3 feet per second _ __ _ '. 120 New Street: 20 wide, 2 deep, 3 feet per second- ___ __ 120 Alley: 24 wide, 3 deep, 3 feet per second______.____ 216 Giraud Street: 34 wide, 3.5 deep, 2 feet per second_____ 238 Romana Street: 28 wide, 1.5 deep, 2 feet per second 84 '*- 'Total______:____ 1.938 46 FLOODS IF CENTRAL TEXAS IF SEPTEMBEE, 1921.

Summary. Second-feet. Jefferson Street ______1,184 Navarro Street ______2,987 St. Mary's Street______4,939 River channel above Houston Street Bridge____ 10,938 Soledad Street______1,644 Streams flowing west from Soledad Street_____ 1, 938

Total______23,630 Or, in round numbers______:_ 23,500 It is estimated that of this 23,500 second-feet, 1,500 second-feet flowed across Soledad Street and reached the channel of San Pedro Creek. The remaining 22,000 second-feet passed down the river channel. ' PREVIOUS FLOODS. GENERAL FEATURES OF FLOOD OF 1913. The flood of December, 1913, exceeded any other recorded flood in Texas, although there is reliable information that greater floods occurred over a considerable area in central Texas in 1869 and pos­ sibly in 1843, 1853, and 1870. The flood of 1913 covered a greater area, and the large streams reached higher stages than during the flood of September, 1921. The flood of 1921, however, caused higher stages on many tributary streams than were ever known before. The most complete information in published form of the flood of 1913 is that by B. Bunnemeyer,12 of the United States Weather Bureau at Houston. His description of the causes of that flood is quoted in full below because of its value in connection with the numerous studies that are being made by State departments, munici­ palities, and other organizations in Texas for the purpose of pro­ viding means of controlling floods. While the rains of December 1-6 were unusually heavy, and in some cases torrential in nature, they would not have resulted in such unprecedented floods had they been preceded by a spell of fair weather; but the ground was thor­ oughly saturated by heavy rains during the last 10 days of November, and the run-off, therefore, was at its maximum speed as well as in amount. Tfiese rains swelled the streams to abnormal heights, with freshets in the Colorado River at Columbus and in the Guadalupe at Victoria conditions which de­ cidedly aggravated the subsequent floods. Another aggravating factor was the breaking in many sections of levees. Many were thus caught unawares. The total rainfall during the last 10 day* of November averaged 4.21 inches for the Guadalupe Valley; 3.74, Colorado; 3.53, Brazos; 2.98, Trinity; and 4.05, San Antonio. These rains laid the foundation J!or the greatest floods of which there is authentic record; they themselves would have caused serious floods if they had fallen within a day or two instead of extending over a period of 10 days. The rains during the first few days of December, which were the immediate precursor of the floods, were more or less continuous but were heaviest on De-

14 Bunnemeyer, B., The December flood of Texas: Eng. News, vol. 71, No. 21, pp. 1136- 1121, May 21, 1914. PREVIOUS FLOODS. 47 cember 2, 3, and 4. The total rainfall averaged 4.78 inches for the Guadalupe watershed; 3.95, Colorado; 5.37, Brazos; 5.30, Trinity; and 2.94 for the San Antonio. About 85 per cent of the amounts were reported on December 2, 3, and. 4. The heaviest amounts in the Colorado watershed were reported on the 2d; in the Brazos and San Antonio on the 3d; and in the Trinity and Guada­ lupe on the 4th. Weather conditions. The cause of these excessive rains is not far to seek. The weather maps of December 1-4 showed an area of low barometric pressure hovering over the Southwest, which was attended by unusually warm weather in Texas. The abnormal warmth naturally induced strong ascending currents which, expanding aloft and cooling to below the dewpoint, resulted in copious precipitation. The ascending currents were replaced by warm, moistures-laden air from the Gulf of Mexico, which served as feeder and kept up the process of condensation. As the area of low pressure changed but little from day to day, condensation continued until the depression began to move eastward on December 5. On that day clearing and colder weather set in; the 6th was clear and cool. Effect of precipitation. In studying the effect of precipitation upon the flow of rivers, probably the two most important factors to be considered are evap­ oration and condition of soil. This is evidenced from the fact that the floods of December, 1913, while exceeding any previous floods on record, were caused by less precipitation than were some of the former floods. For instance, the flood of the Brazos in 1899 was caused by copious rains during the last five days of June of that year, which averaged heavier than those that caused tiie recent flood. Among the heavier total amounts (in inches) reported for these five days were: Waeo, 7.30; Hewitt, 14.95; Temple, 9.62; Brenham, 20.08; Sugarland, 12.30; Columbia, 9.17; and Brazoria, 8.68. The redeeming features were « more active evaporation and a period of 10 fair days preceding the precipita­ tion. The total rainfall for the State for June, 1899, averaged 7.07 inches against 5.03 inches for December, 1913. The big flood of the Guadalupe of October, 1913, was also caused by heavier rains than the bigger flood of December, 1913. The October flood gave high- water mark to date and caused more damage than the recent flood because the crops had not been gathered. The average rainfall in the Guadalupe Valley for the four days causing the October flood was 6.76 inches against only 4.78 inches for the four days in December. Other instances could be cited of heavier rains with less severe floods. In glancing over the State records and without going into details, the records show that there were numerous months in which the precipitation for Texas was much greater than it was in December, 1913; in fact it was greater in all months except March. Since 1888, it was greater once in February and Octo­ ber; twice in January, August, and November; three times in May, July, and September; four times in April; and five times in June. These statistics are of interest in showing not only the frequency of exces­ sively wet months but also the probability of floods, whose severity naturally depends in a large measure upon the combinations favorable for a large run-off. Floods of more or less extent have occurred in Texas since records were kept, in all months of the year except in January. They were more frequent in April, May, and June than in any other season of the year. Bunnemeyer estimates that 177 deaths and losses of property exceeding $8,500,000 resulted from the flood of 1913. 48 FLOODS IF CENTRAL TEXAS IF SEPTEMBER, 1921.

BRAZOS RIVER. The flood in Brazos Eiver in December, 1913, exceeded any for that stream of which there is knowledge. The following table gives all available data on the annual maximum stage and discharge of Brazos Eiver from Waco to its mouth. It is interesting to note that the flood of 1921, which came almost entirely from Little River, reached a maximum stage of 58.2 feet at Valley Junction, or only 0.8 foot lower than the flood of 1913; though at Kosenberg the max­ imum was 8.7 feet lower in 1921 than in 1913.

Annual maximum gage height, in feet, and discharge, in second-feet, of Brazos River.

Waco. Valley Junction. College Station. Year. Date. Gage Dis- Gage Gage Dis­ height.* charge.6 Date. height.* Date. height. charge.

1884...... May 27 31.7 104,000 1885...... May 28 34.2 119.000 1887...... Aug. 31 31.8 104.000 1890...... Apr. 22 31.6 ' 99,609 1895...... Mar. 28 30.1 94,300 1897...... Mar. 29 32.6 109,000 1SQO June 20 20.7 44,800 July 5 39.6 lorn Sept. 28 25.7 69,800 1901...... May 19 14.8 22,700 1902...... July 26 32.1 106,000 100* Oct. 1 14. r 22,400 1904...... June 30 12.8 17,000 1905...... May 14 28.6 85,800 May 2 40.6 HK»...... : June 6 19.8 40,900 June 7 24.3 1907...... Dec. 22 14.8 22,600 Dee. 23 20.7 1908...... May 25 36.7 134,000 May 27 50.9 1909...... Dec. 2 16.8 29,200 Dec. 4 16.4 1910...... May 22 15.8 25,700 May 23 20.0 1911 July 18 18.5 35,400 July 19 22.6 1912...... Aug. 7 16.2 27,100 Mar. 6 23.5 1913...... Dec. 3 39.7 <*) Dec. 4 55.0 1914...... May 31 27.8 81,300 June 1 39.0 1915...... Apr. 26 26.0 104,000 Apr. 24 «50.0 1916...... Apr. 2 33.8 113,000 Apr. 4 28.6 1Q17 Apr. 30 9.8 4,550 May 13 12.0 lOlfi Nov. 9 36.4 125,000 Nov. 10 34.8 Nov. 11 41.0 69.300 1Q1Q Oct. 23 27.9 78,100 Oct. 24 36.5 Oct. 25 41.5 70,500 1920...... May 11 22.4 53,100 Sept. 9 29.5 May 19 34.0 53,200 1921...... June 11 18.0 31,100 Sept. 11 58.2 Sept. 12 58.0

a Gage heights, 1884-1897, from high-water marks kept by Crow Bros., proprietors of Waco Steam Laundry. Gage height, 1912-1914, furnished by United States Weather Bureau. «> Discharge rating curve for 1884-1905 based on measurements made from 1898-1905. Stage-discharge relation not permanent. All discharge figures liable to large errors. e Zero of gaga lowered 4.0 feet on Mar. 23,1918. d Levees overtopped and broken. East Waco flooded. Determination of discharge impossible. « Corresponding rise at Waco reached a maximum of 21.2 feet oa Apr. 21. PREVIOUS I*LOODS. 49 Annual maximum gage height, in feet, an& discharge, in second-feet, of River Continued.

Washington. Hempstead. Bosenberg./ Richmond. Year. Gage Gage Gage Gage Dis­ Date. height. Date. height. Date. height .0 Date. height. charge.

1899 July 50.0 July 8 44.0 1902...... Aug. 8 38.0 1903...... Mar. 7 38.7 Ma*. 7 33.4 68,600 1904...... May 9 35.3 May 10 30.2 May 10-11 25.5 47,600 1905...... May 5 40.5 May 20 37.5 May 6 33. & 65606 1906...... June 9 29.2 June 9 21.4 June 9 21.2 37,309 1907...... Nov. 24 39.0 Nov. 24 35.3 1908...... June 2 42.6 June 6-7 42.3 1909...... ftJune 20 16.0 June 21 15.5 1910...... May 24 21.5 May 23 20. ft 1911...... July 21 22.1 July 25 20.4 1912...... Mar. 7 23.6 Jan. 1 10.7 1913...... Dec. 6 62.0 Dec. 9 <52.8 Dec. 9 52.5 1914...... May 9 4L5 1915...... Apr. 28 52.9 Apr. 29 46.5 May 1 46.4, 1916...... Apr. 5 40.5 Apr. 5 /SO. 7 May 6 26.8 1917 ...... I...... Sept. 8 13.7 Sept. 9 6.8 Apr. 20 2.9 1918 ...... Nov. 11 41.9 Nov. 12 31.3 Nov. 13 28,7 1919...... Oct. 25 45.8 Oct. 26 35.7 Oct. 28 347 1920- ...... Jan. 26 40.9 Jan. 25 32.7 Jan. 26 36.4 1921...... Sept. 14 50.0 Sept. 16 40.2 Sept. 16 37.7

/ Gage at Booth, 10 miles below Bosenberg from 1902-1912. ff Zero of gage lowered 3.94 feet on Mar. 1, 1918. ft Also on Dec. 4. » A higher stage was probably reached during night preceding this reading. / Corresponding rise at Rosenberg reached a maximum of 26.8 feet on Apr. 6. Little River and many of its tributaries reached higher stages in 1921 than ever before. The highest stage at Cameron prior to 1921 was reached in December, 1913, when the river stood at a stage of 44.9 feet, referred to the United States Geological Survey gage, The next greater flood occurred in 1916 and reached a maximum of 37.4 feet. The flood height of 1921 was 49.5 feet. The relative heights of several great floods on San Gabriel River have been recorded by notches cut in a tree on Fulcher's ranch near the mouth of Pecan Creek on the right bank of the river half a mile above San Gabriel. From these notches it has been determined that the crest of the flood of December 4, 1913, was 6.5 feet lower than during the flood of 1921 and that the crest of the flood of April 7, 1900, was 8.1 feet lower than the flood of 1921. Mr. Fulcher stated that the flood of 1869 was about half a foot higher than the ilood of 1913, which makes it 6 feet lower than the flood of 1921. San Gabriel River and practically all its tributaries reached higher stages in September, 1921, than at any other time since that region has been inhabited by white people. COLORADO RIVER. The flood of July, 1869, on Colorado River, was by far the highest for which there is any information. The capacity of the river chan­ nel below Austin is considerably greater at present than it was in 1869. The bottom land has been cleared for farming, and those who have lived along the river for many years believe that the chan- 50 x FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. nel has been lowered considerably. Therefore the maximum dis­ charge at Austin during the flood of 1869 was not so great as the ratings made since 1900 indicate^ although it probably greatly ex­ ceeded any flood since that time. No information is available regarding the floods of 1843,1852, and 1870, except maximum stages as given by Professor Taylor.13 He has described the flood of 1869 as follows: On July 3, 1869, began the longest and most uninterrupted rain ever known in Austin. It rained without stopping for about 64 hours, and the river rose to a height of 43 feet. The lower part of the city was overflowed and several people were drowned. The town of Webberville, 16 miles below Austin, was also overflowed, and Bastrop, 30 miles below, was inundated. Eecords of the annual maximum gage height and discharge of Colorado River at Marble Falls, Austin, and Columbus are sum­ marized in the following table:

u Taylor, T. TJ., The Austin dam: U. S. Geol. Survey Water-Supply Paper 40, p. 34, 1900. PREVIOUS FLOODS. 51 Annual maximum gage height, in feet, and discharge, in second-feet, of Colo­ rado River.

Marble Falls. Austin. Columbus. Year. Gage Dis­ Gage Dis­ .Gage Dis­ Date. height. charge. Date. height. charge. Date. height. charge.

1843...... February 36 1852...... March... 36 1869...... July..... 43 ...... July..... 41.2 ...... 1870...... October. 36 1898...... June 16 10.7 30,200 1899...... June 8 21.0 91,800 1900...... Apr. 7 23.9 Apr. 7 («) 151,000 1901...... July 13 10.2 28,700 1902...... July 28 11.85 35,900 1903...... Feb. 27 11.3 33,700 Mar. 1 35.2 58,600 1904...... June 8 10.4 31,500 May 5 28.5 642,400 1905...... Apr. 30 16.1 57,600 May 2 33.0 652,700 1906...... Aug. 12 19.5 78,500 Aug. 15 33.5 654,000 1907.;..... May 29 10.2 28,100 May 31 35.0 658,100 1908...... Apr.22-23 20.0 103,000 Apr. 23 22.0 82,300 Apr. 27 35.8 66tt300 1909...... July 23 12.9 44,200 July 24 11.1 31,700 June 5 20.9 23,000 1910...... Sept. 6 11.6 36,200 Sept. 9 10.3 28,500 May 23 25.0 32,600 1911...... Feb. 22 - 10.6 30,500 Sept. 7 iao 27,400 Sept. 9 22.2 26,000 1912...... Oct. 13 9.1 22,400 Oct. 18 5.8 Feb. 24 18.6 17,900 1913...... Dec. 3 21.0 112.000 Dec 5 c27.0 d 164, 000 Dec. 6 44.1 1914...... June 19 15.5 61,400 June 20 «16.5 61,500 May 28 35.8 60,300 1915...... Sept. 17 20.0 103,000 Sept. 17 /247 135,000 Apr. 28 36.2 61,300 1916...... May 2 9.0 21.900 May 22 13.1 46,000 May 23 24,7 36,800 1917...... Sept. 5 8.0 17.100 Sept. 7 4.3 8,420 May 7 12.1 4,520 1918...... ffNov. 11 17.6 76,600 Nov. 9 19.5 74,800 Nov. 14 33.9. 55,100 1919...... Sept. 24 18.7 85,000 Sept. 25 18.8 72,700 Oct. 14 34.6 55,800 1920...... ftSept. 10 10.4 29,400 Aug. 8 10.25 33,000 Jan. 24 33.4 53,800 1921...... June 12 las 28,800 Sept. 10 19.4 75,700 Sept. 13 33.8 57,300

o The Austin dam broke during the flood, which reached its peak just bsfore the failure with a depth of 11.07 feet over the crest of the dam and a discharge of 151,000 second-feet (revised in 1922). After the failure of the dam the maximum height at the gage, 34 miles below, was 33.5 feet. From statements by people living near the gage, it has-been determined by leveling that the water rose 6.1 feet at the gage as a result of the failure of the dam. Therefore the gage height corresponding to a discharge of 151,000 second-feet was 27.4 feet. 6 Revised. c Partly estimated. A Determined by extending rating curve for 1915 parallel to curve developed in 1922. The curve for 1915 was fairly well defined below 60,000 second-feet, and the curve for 1922 was well defined below 112,000 second- feet (gage height, 22 feet). A comparison of measurements made from 18_97 to 1922 indicates a progressive change in the stage-discharge relation, probably due principally to clearing of flood plain for agricultural use. « Corresponding rise at Columbus occurred on June 22 and reached a matin-mm stage of 32.2 feet (dis­ charge, 50,600 second-feet). / Corresponding rise at Columbus occurred on Sept. 20 and reached a maximum stage of 32.5 feet (dis­ charge, 51,400 second-feet). g Corresponding rise at Austin occurred on Nov. 11 and reached a maximum stage of 17.8 feet (discharge, 67.700 second-feet). » Corresponding rise at Austin occurred on Sept. 10 and reached a maximum stage of 8.4 feet (discharge,. 24,800 second-feet). NOTE. The accuracy of the records at Austin prior to 1898 is doubtful. Gage heights at Marble Fall s prior to 1917, at Austin from 1912 to 1914, and at Columbus from 1912 to 1915 were furnished by U. S. Weather Bureau. Discharge records at Austin prior to 1916 revised in 1922. The greatest floods on Colorado Kiver for which there are authen­ tic records occurred in 1900 and 1913. A study of all available information shows that those floods were of about the same magni­ tude. The floods of 1900 caused the failure of the Austin dam, which released a large volume of stored water and made it im­ possible to determine accurately the maximum natural flow stage.

GUADALUPK RIVER. The flood of 1921 was the greatest flood on Guadalupe River at New Braunfels since the station was established in 1915. Though FLOODS IN CENTRAL TEXAS IN SEPTEMBEK, 1921. residents report that much higher floods have occurred at that point, possibly as high as 40 feet, no authentic records are available. The greatest recorded flood on this river occurred at Gonzales, just be­ low the mouth of San Marcos River, on December 4, 1913. This flood came principally from San Marcos River. The annual maxi­ mum gage heights and discharge of Guadalupe River and its chief tributary, San Marcos River, are given in the following table: Annual maximum gage heif/ht, in feet, and discharge in second-feet of Gua­ dalupe and San Marcos rivers.

Guadalupe River Guadalupe River at Guadalupe River. near Comfort. New Braunfels. at Gonzales.a Year. Gage Dis­ Date. heigjtt. Dat«. height. charge. Date. &.

1900...... Apr. 8 36.3 1903...... 1904...... 1905...... Apr. 26 24.0 1906...... Feb. 14 14.0 1907...... i Nov. 22 31.0 1908...... May 27 20.5 1909...... July 26 12.5 1910...... Apr. 12 11.2 1911...... Mar. 20 r'l 1912...... Feb. 25 23.2 1913...... Dec. 4 38.1 1914...... May 22 31.8 1915...... Sept. 17 27.2 51,500 Apr. 25 30.0 1916...... May 22 20.8 30,800 May 25 22.3 1917...... Sept. 7 4.6 2,360 May 8 13.8 1918...... Apr. 5 17.5 Dec. 24 9.5 8,350 Apr. 30 18.0 1919...... Aug. 21 41.0 Aug. 23 25.7 46,300 Oct. 18 341 1920...... 1...... Aug. 9 CO May 16 33.4 1921...... June 12 25.0 Sept. 10 28.6 56,600 Sept. 11 81.4

Guadalupe River near Guadalupe River San Marcos River at Cuero.c at Victoria.** Ottine. Year. Dis­ Gage Dis­ Date. £B. charge. Date. height. Date, height. charge.

1900...... , lorn Mar. 1 43.0 71,300 July 22.4 1904...... May 7 20.8 6,890 1905...... Apr. 29 24.9 10,600 Apr. 30 22.1 1908...... Feb. 15 17.0 4,650 Feb. 16 148 1907...... Nov. 25 23.3 1908...... May 29 21.1 1909...... July 25 14.6 1910...... May 24 19.0 1911...... May 15 12.9 1912...... Feb. 28 19.9 1913...... Dec. 7 248 1914...... May 25 23.8 1915...... <*Sept. 22 29.5 52,000 Apr. 28 245 1916...... May 27 19.5 15,100 May 27 18.2 May 22 22.2 5,800 1917...... May 9 9.6 6,930 May 10 12.9 May 7 23.5 7,480 1918...... Oct. 29 142 10.600 Oct. 30 19.2 "Apr. 6 29.5 «8,750 1919...... 30 8-55,000 Oct. 21 25.1 July 23 346 18200 1920...... May 20 24.6 May 16 37.5 22,000 1921...... Sept. 15 16.4 12,500 Sept. 16 20.5 Apr. 8 35.5 19,500

a United States Weather Bureau gage. 6 F ecorder out of order. Maximum stage probably did not exceed-10 feet. c Gage used from 1903 to 1906 and from Aug. 19,1915, to Aug. 6,1916, at Schleicher Bridge, 4 miles above present gage. Discharge at the two points is practically the same but gage heights are not comparable. d Gage installed Aug. 19,1915. « Revised. / Recorder out of order. Maximum stage of about 30 feet occurred between July 23 and 28. g Determined from extension of rating curve and subject to considerable error. PREVIOUS FLOODS. ' 53

SAN ANTONIO BIVEB. The flood of September 9-10, 1921, at San Antonio greatly ex­ ceeded all other floods there of which there is any information except the flood~of July 5,1819, which probably equaled if not exceeded it. Information regarding the flood of 1819 has been obtained from old Spanish records by Rev. Eugene Sugraves, C. M. F.,1* who has writ­ ten as follows: The worst flood that ever swept the city of San Antonio was the one that befell the sleepy old pueblo, the " Villa de San Fernando " and " Real Presidio de San Antonio de Bejar," on July 5, 1819. The 1819 flood stands out now as an established ^incontrovertible and his­ torical fact. After a long and painstaking research through the dust-laden archives of the old San Fernando Cathedral, the writer had the good fortune of finding the record of this frightful flood, the memory of which has been handed down to us from generation to generation. The document is written in Spanish and appears in the death records ("Li- bro de Difuntos") of the parish, Nos. 956 and 957. According to this record, the flood victims were 16 six white grown persons ("gentes de raz6n") and ten Indian children. * * * During the flood of 1819 the waters of the San Antonio River and those of. San Pedro Creek came together on Military Plaza and Soledad Street. A number of houses around the present Main and Military plazas were washed away in that flood of 1819. Bear in mind that these houses were substantial constructions similar to that of the old portion of the San Fernando Cathedral, which has stood against the buffets of storm and stress for nearly 200 yefcrs. The cause of the flood is given as a " culebra de agua " in the Olmos Creek; a " serpent" or snake of water, which means in plain English a cloudburst. Among other things the flox>d of 1819 wrecked and ruined the following buildings: One at the corner of Main Avenue, then known as Acequia Street, and Main Plaza; several houses on Curbelo Street, now Soledad Street; one on the east side of Mam Plaza, about halfway between Market and Commerce streets; still another one near the southeast corner of Dwyer Avenue, then known as Quinta Street, and Dolorosa Street; also several on the east and south sides of Military Plaza, opposite the present City Hall. The first San Fernando Cemetery, which occupied the east portion of the old San Fernando Church grounds, and now a part of the new portion of the Cathedral, was covered with water 5 feet deep. In what is now Main Plaza, then Plaza de las Islas, or Plaza Mayor, on the south side of Market Street, extending to Galan and running southward to Dolorosa Street, there was a row of stone buildings. Two of these buildings were washed away. In Military Plaza, then Plaza de las Armas, there were four rows of stone buildings. One was located on the north side from Flores Street to Amargura Street and south of Presidio, now Commerce Street. This is where the main military barracks were located. Theve was a second row of stone buildings from the row mentioned above, 'southward along the west side of Flores to Dolorosa Street. This is where the first public school house of this town was located. The school, however, was not like our modern public schools. It was public in the sense that it was supported by the community. But, like the early New England

11 San Antonio Express, Sept. 18, 1921. 54 FLOODS IN CENTRAL TEXAS IN SEPTEMBER, 1921. colonial schools, it was built under the shadow of the church. Religion was taught there regularly by the padre. Another row of stone buildings was situated on the south side of said Mili­ tary Plaza from Flores to Arnargura Street, arid another row was situated on the west side of said plaza from Dolorosa Street to Presidio, now Commerce Street. These were occupied as dwellings by the officers of the garrison. The sclioolhouse and one other building in the east row of the buildings and several other buildings in the row on the south side were demolished by the great flood of 1819. As hereinbefore stated the waters of the San Antonio River left their western bank above Presidio Street now Commerce Street, and flowed into Curbelo, now Soledad Street, down into Main Plaza, across said plaza in a south and west­ ward direction. Then they ran through Presidio, Trevino, and Galan streets into Military Plaza, then known aa Plaza de las Armas, where the waters of the river met those of San Pedro Creek. Other destructive floods at San Antonio have occurred as follows: 1S

Estimated Year. discharge. Remarks.

Second-feet. 1852...... 4,000 Considerable damage. 1865...... 7; ooo Heavy loss of life and property. 1866...... (°) 1868...... («) 1880...... 8,000 1899...... 7,000 October, 1913...... 7,200 Considerable damage. December, 1913.... 8,000 Do. October, 1914...... 5,500 Estimated maximum discharge of San Pedro Creek, 2,000 second-feet. " Floods resulting from a rainfall of 5 inches in 3 hours caused the death by drowning of about 20 persons in San Antonio, Tax., Oct. 23. The property loss is estimated at $150,000." 6

o Not estimated. 6 Eng. News., vol. 72, No. 18, p. 899, Oct. 29,1914. «From unpublished report by United States Engineer Corps, Fort Sam Houston, Tex. INDEX.

Page. Fas*. Acknowledgments for aid 2-3 Gaging stations, location of_ _. 37, 39 Alazan Creek, basin and flow of 33, 34 Guadalupe River, basin .and tribu­ discharge o£_ 43 taries of _ _ 31 elevations and distances on - 35 elevations on and distances flood on___ _ 36 along 32 Apache Creek, basin and flow of 33, 34 floods on , 51-52 discharge of 43 gage height and discharge of 33, 52 elevations and distances on 35 high water on 32 Area covered by rainfall that caused Guadalupe River basin, measure­ the flood _ 3 ments of rainfall in_ 13

Bartlett, C. Terrell, acknowledg­ Hemphill, R. G., cited_____ 43,44-45 ment to 2 Hutto, tornado at______11 cited ______35-36, 43 Board of Water Engineers, Texas, Jadwin, Col. Edgar, acknowledgment plan and work of 2 to ______2 Brazos River, distances between Jones Bridge, gage height and dis­ points of measured charge at 16-17 elevation on 14 elevations on______14 Life, loss of- _ 5, 6 flood of 1921 on______15-17 Little River, crest of flood on, plate gage height and discharge of 16, showing 20 17, 48-49. determination of flow of 18 previous floods on _ 48-49 discharge of, at Cameron 19, 20-22 Brazos River basin, measurements below junction of Leon and of rainfall in _ _ 13 Lampasas rivers 19-20 Brushy Creek, cross sections used distances between points' of in determining dis­ measured elevation on_ 18 charge of 28 elevations on______18 discharge of______27-28 heights of floods on 49 plate showing. _ 20 tributaries and basin of 17-18 Bunnemeyer, B., cited __ 15, 30,46-47 valley of, plates showing- 20 Live stock, loss of, in Milam County. 4 Colorado River, course and basin of- 29 Losses caused by the flood--_ 4-7 elevations on and distances along 29 McAuliffe, J. P., cited______8-9 flood of 1921 on______29-31 McDonald, C. K., cited______43 gage height and discharge of 31, 51 Map of central Texas showing drain­ previous floods on_ _ _ 49-51 age basins and total Colorado River basin, measurements rainfall September 7- of rainfall in______13 11, 1921______4 Crops, losses of_ __ _ 4, 6 San Antonio showing area flooded September 9- Damage caused by the flood 4-7 10, 1921______- 44 Discharge per square mile, high Martinez Creek. See Alazan Creek. records of______13-14 Milam County, losses of property in_ 4, 6

Elgin, small rainfall at. 12 Olmos Creek, basin and flow of 33-34 Extent of the floods____ 1 discharge of- 43 flood on____ _ 35, 36 Flood of December, 1913, causes of_ 46-47 Onion Creek, flood on 30 Floods, need for control of ___ 2 previous 46 Pedernales River, flood on. 30 55 56 INDEX. Page. Page. Railroads, losses sustained by_ . 6 San Antonio Raver Continued. Rain, flood, cause of______9 view of left bank of, plate show­ flood, measurements and esti­ ing 44 mates of______7-13 San Gabriel River, basin of 24-25 Rosenberg, report of high water at 15 flood of 1921 on______25-26 heights of floods on 49 valley of, plate showing 20 Salado Creek, basin of______22-23 cross section used in determin- San Gabriel River and forks of, dis­ ing maximum dis­ tances between points charge of ______24 of measured elevation flood on______23-24 on _ 25 plate showing __ _ 20 elevations on 25 Salsamofa Creek. See Apache Creek. San Marcos River, gage height and San Antonio, damages to property discharge of _____- 33, 52 in ______5, 6-7 San Pedro Creek, basin and flow of- 33, 34 flood in ______35-37 discharge of _ 43 flood of 1819 in______53-54 elevations and distances on 34 map of, showing area flooded flood on ______36 September 9-10, 1921- 44 gage heights on 40 measnrements of rainfall at- 9-10,12 rating table for______42-43 municipal property damaged in_ 7 Soil of the flooded area 3 public service corporations in, damage to 4 losses sustained by__ 7 situation of______33 Streams, stations on, selected for South Alamo Street bridge, measurements of flow- 1-2 after flood, plate show­ Sugraves, Rev. Eugene, cited- 53-54 ing _ ___ _ 44 streams flowing through____ 33-34 Taylor, T. W., cited- 50 water-stage recorder at South Taylor, thunderstorms at and near- 11-12 Alamo Street, plate rainfall at and near 8-9,11-12 showing______44 Thunderstorms, effects of 11-12 San Antonio River basin, measure­ Topography of the area of flood rain- 3-4 ments of rainfall in__ 13 San Antonio River, bridge across, United States, damage to prop­ at Romana Street, plate erty of 7 showing______44 United States Weather Bureau, ac­ bridge across, near Pioneer Mills knowledgment to____ 2 after flood, plate show­ report of, on flood on Brazos ing _ ___;_____ 44 River ______15 course of 33-34 on flood on Colorado River- 30 discharge measurements on 37-46 on the storm 9-13 distances between points of measured elevation on_ 34 'Valley Junction, report of flood at- 15 elevations on _,___ 34 Vance, Maj., W. N., acknowledg­ gage heights on______40, 41 ment to______2 previous floods on . 53-54 Vegetation of the flooded area______3 view of, showing slope stretch used during flood, plate Wallis, report of flood at______15 showing 44 Washington, report of flood at. 15 o