Louisiana State University LSU Digital Commons

LSU Historical Dissertations and Theses Graduate School

1971 Process and Parameter Interaction in Rattlesnake Crevasse, Mississippi River Delta. David James Arndorfer Louisiana State University and Agricultural & Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses

Recommended Citation Arndorfer, David James, "Process and Parameter Interaction in Rattlesnake Crevasse, Mississippi River Delta." (1971). LSU Historical Dissertations and Theses. 2020. https://digitalcommons.lsu.edu/gradschool_disstheses/2020

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please contact [email protected]. 72-3M5L

ARNDOREER, David James, 1943- PROCESG AND PARAMETER INTERACTION IN RATTLL'PNAKE CREVALUE, MIEE IESIPPI RIVER DELTA.

The Louisiana Plate University and Agricultural and Mechanical College, Ph.D., 1971 Geography

U n iv e rs ity M icro film s, A XEROX C om pany , A n n A rb o r, M ic h ig a n

'■ P Il/N.P EXACTLY AL M T ’EIVIJ) PROCESS AND PARAMETER INTERACTION IN RATTLESNAKE CREVASSE, MISSISSIPPI RIVER DELTA

A Dissertation

Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College In partial fulfillment of the requirements for the degree of Doctor of Philosophy In The Department of Geography and Anthropology

by David James Amdorfer B.3., Portland State College, 19&7 August, 1971 PLEASE NOTE

Some Pages have indistinct print. Filmed as received.

UNIVERSITY MICROFILMS ACKNOWLEDGEMENT

The writer is indebted to Dr. William G. Mclntlre for his Invaluable advice, aselatance and support throughout the study. Statistical sampling designs were developed through extensive discussions with Dr. David S. McArthur. In addition. Dr s. L. D. Wright, Sherwood M. Gagllano and James M. Coleman contributed considerable advice and direction. Drs. Prentiss E. Sohilling and Kenneth L. Koonce of the Experimental Statistics Department provided assis­ tance and direction in all statistical aspects of the investigation. To these Individuals, the writer extends a special appreciation. Numerous logistloal difficulties frequently encountered during field work in a deltaic environment were minimized by employees of the Louisiana Wildlife and Fisheries Com­ mission. Pass a Loutre Camp, especially Messrs. Benny and Ellis Loga. The writer gratefully acknowledges the aid given by field assistants who volunteered their services during the field program. Donald W. Davis willingly provided valuable field assistance on numerous oocaslons. The investigation was supported by Coastal Studies Institute of Louisiana State University under contract with the Geography Programs of the Office of Naval Research.

11 Contract No, N00014-69A-0211-0003, Project No, NR 388 002. The United States Army Corps of Engineers, New Orleans District, Instrumented the study areas and collected valu­ able current velocity and water level records. Messrs. Jerome Baehr and Bert Kemp were particularly helpful. The remote-sensing Imagery was provided by the Spaoe Oceano­ graphic Program and National Aeronautics and Space Admini­ stration Earth Resource Aircraft Project.

Ill TABLE OP CONTENTS

Page Acknowledgement 11 List of Tables vl List of Figures vll Symbols and Abbreviations lx Abstract xl

INTRODUCTION 1 SALIENT CHARACTERISTICS OF CREVASSES 4 SALIENT CHARACTERISTICS OF THE STUDY REGION 7 DATA COLLECTION AND ANALYSES 13 STATISTICAL ANALYSES 14 Crevasse Channel Current Velocity 16 FACTORIAL ARRANGEMENT OF FACTORS 23 Discharge Analysis. Rattlesnake Crevasse 25 Suspended Load Analyses. Rattlesnake Crevasse 30 Sediment Accretion. Rattlesnake Crevasse 36 INTERPRETATION 43 RATTLESNAKE CREVASSE DISCHARGE 43 River Stage 43 Semi-monthly Tide 43 Diurnal Tide 45 Location 45 Interactions 46

lv Page DISCHARGE PATTERNS, RATTLESNAKE CREVASSE 46

LITTLE LAKE POND CREVASSE CIRCULATION 48 CREVASSE DISCHARGE, MISSISSIPPI RIVER DELTA 49 SUSPENDED LOAD TRANSPORT, RATTLESNAKE CREVASSE 51 Temperature 52 Discharge 54 River Stage 54

Semi-monthly Tide 56 Diurnal Tide 56 Location 57 Interactions 57 SEDIMENT ACCRETION, RATTLESNAKE CREVASSE 58 WIND 61

SUMMARY AND CONCLUSIONS 65 REFERENCES 70 Vita 74

v LIST OP TABLES

Page 1. Multiple Regression Analysis of Reversed and Null Current Duration on Diurnal Tide Range and River Stage, Rattlesnake Crevasse 18 2. Multiple Regression Analysis of Rattlesnake Crevasse Current Velocity on River Stage (Cubic) and Change in Water Level (Linear and Quadratic) 18 3. Multiple Regression Analysis of Little Lake Pond Crevasse Current Velocity on Various Independent Variables 21 2 4. Observations Obtained for the 2x3*4 Factorial Classification 28 5. Discharge Analyses of Variance, Rattlesnake Crevasse 28 6. Discharge Least-squares Means (cfs)i High and Low Stage 29 7. Discharge Least-squares Means (cfs)i Low Stage 29 8. Paired "t-test" of Vertloal Distribution of Suspended Load (mg/1), Rattlesnake Crevasse 31 9. Suspended Load Concentration Analyses of Variance 31 10. Susoended Load Least-squares Means (mg/l)i High and Low Stages 33 11. Suspended Load Least-squares Means (mg/l)i Low Stage 33 12. Mean Suspended Load Coneentration. Discharge, and Temoerature 13. Analyses of Variance for Mean Daily Sediment Accretion in Rattlesnake Crevasse 39 14. Sediment Accretion Least-squares Means *4-0 15. Annual Sediment Accretion Least-squares Means 41

vi LIST OP FIGURES

Page 1. Rattlesnake Crevasse (Black and White IR Imagery 2 2. Sealed Crevasse along Raphael Pass, Cubits Gan Subdelta 6 3. Location Map of the Mississippi River Delta 8 4. River Stage Hydrograph at Carrollton Gage (New Orleans), 1968-1970 9 5. River Stage Hydrographs at Rattlesnake Crevasse and Little Lake Pond Crevasse, I968 -I969 10 6. Diurnal Gradient Changes along South Pass 11

7. Crevasse Into Little Lake Pond 13 8. Circulation In Rattlesnake Crevasse at Equatorial Tide, Low River Stage 15 9. Circulation in Rattlesnake Crevasse at Tropic Tide, Low River Stage 15 10, Instrument Tower at Little Lake Pond Crevasse 16

11, Effect of River Stage and Lunar Tide on Current Velocity, Rattlesnake Crevasse Channel 19 12, Effect of River Stage and Lunar Tide on Current Velocity, Little Lake Pond Crevasse 22 13, Location of Profiles and Accretion Measurements Rattlesnake Crevasse 26 14, Accretion Rod on Crevasse Solay Natural Levee, Rattlesnake Crevasse 37

15, Accretion Leveler 37 16, Accretion Isopach Maps 42

17, Current Profiles at the Mouth of South Pass on October 1, 1969 50

vll Page IB. Diurnal Discharge Variation In the Channel of Southwest Pass at Mile 8 50 ID. Thermal Scanner (Reconofax TV) Imagery of Rattlesnake Crevasse Study Area 5 3 20. Comoarlson of Suspended Load ConcentratIon at Profile A for Flood and Low River Stages 55 21. Effect of Wind Surge on Low Stage Discharge Pattern (TroDlc Tide) 62 22. Effect of Wind on Current Patterns In Rattlesnake Crevasse Channel 63

v 111 SYMBOLS AND ABBREVIATIONS

A z vertical Interval between current velocity obser­ vations m micron X summation symbol F F-test statistic H0 null hypothesis M.S. mean square M-fc discharge R multiple correlation coefficient R3 coefficient of determination for multiple regression

X the mean of a sample b sample partial regression coefficient bQ Y Intercept C with cfs cubic feet per second 7 mean difference between means of paired samples d.f. degrees of freedom g gram J horizontal interval between current velocity obser­ vations

1 liter mg milligram n sample size r simple correlation coefficient lx s standard deviation t t-test statistic u current velocity In discharge formula ' feet

* significant at the 0 .0 5 level (significant) ** significant at the 0.01 level (highly significant)

x ABSTRACT

Rattlesnake Crevasse and to a less extent, Little Lake Pond Crevasse In the Mississippi River Delta were lnvesti­ trated to Identify and evaluate variables governing Its evo­ lution, as well as to explain the interactions and processes by which the variables exert their Influence on the develop­ ment of the ci'evaSBe, Methods for applying statistical procedures to evaluate the variables controlling Its evo­

lution are developed. With the exception of water temperature, the variables evaluated In this Investigation Interact to cause gradient changes between the distributary and the adjacent lnter- dlstrlbutary bay. Flow through the crevasse channel results from the gradient determined by river stage, semi-monthly tide, diurnal tide, stage by semi-monthly tide Interaction, and semi-monthly by diurnal tide interaction. Wind surges, occasionally present, are also an Important factoi' influ­ encing gradients. The function of Rattlesnake Crevasse Is to contribute to the balance of Mississippi River Delta hydrology. By providing an overflow channel during the diurnal fluctuation of water level and the diurnal discharge variation In dis­ tributaries, the crevasse drains large quantities of water which are either stored in the Interdistributary bays, or

xl discharged through the bays Into the Gulf. Rattlesnake Crevasse provides an overflow channel during the diurnal tide as well as during flood stage. Data collected during a wind surge Indicates that the Importance of the storm surge to crevasses processes Is less than that of lunar tide. Unless a wind surge Is In associ­ ation with a tropic storm. It appears to have no lasting effect on sedimentation In the crevasse splay. Sediment discharge at flood stage was 25 times greater than at low water. The variation of suspended load with the diurnal tide was found to be a function of diurnal water temperature changes, discharge reversal In the crevasse channel, and suspended load fluctuation in Pass a Loutre. Diurnal water temperature changes favor the deposition of material In the crevasse splay. Sediment accretion In Rattlesnake crevasse splay ex­ ceeded 0.35 foot of upward growth during the 1970 hydrologic year. Within the splay, maximum accretion during flood oc­ curs on the low, submerged natural levees. During low water, this zone undergoes greatest dehydration, compaction, and erosion, producing maximum net annual accretion slightly basInward, nearer the end of the channels. This Investigation has successfully demonstrated that analyses of variance procedures can be applied to identify and evaluate the significant vai'lables in a delta crevasse. In addition, the methods employed establish that comparison xll of discharge patterns of the two crevasses provides a means of analyzing the relative intensity of the variables af­ fecting these crevasses.

zill INTRODUCTION

This Investigation Identifies and evaluates variables governing the evolution of Rattlesnake Crevasse in the MisslsslpDl River Delta, and explains the Interactions and processes by which variables exert their influence on the development of the crevasse. Rattlesnake Crevasse (Fig, 1) demonstrates how these Interactions and processes change in Intensity during the hydrologle year. The Investigation develops methods for applying statistical procedures to evaluate the variables controlling crevasse evolution. Crevasse discharge and sediment transport, operating within the framework of both basin and crevasse-splay geo­ metry, govern accretlonary rates. Primary factors regula­ ting discharge and sediment transport are (1) hydrologic regime of the Mississippi Rlveri (2) local drainage patternsj

(3) diurnal tide* (^) semi-monthly tldei (5) water tempera­ ture! (6) wind* and (?) proximity of the crevasse to the distributary mouth. In this study, a crevasse Is defined as both the breach In the natural levee and Its associated deposit. It results from both erosion and depositIoni erosion results In a crevasse channel, deoosltlon In a crevasse splay. The crevasse channel erodes deepest where It Is constricted through the breached natural levee. From the oonstrlctlon,

1 2

Figure 1. Rattlesnake Crevasse (Black and White IR Imag ery). Pass a Loutre (top) flows from left to right. 3 the channel extends Into the lnterdlstrlbutary bay with an adverse bottom slope. The crevasse sp 1c*y is a lenticular deposit built from the load transDorted by the crevasse channel. Natural levees, formed along each channel, are highest near the orifice and gradually decrease In height basinward. Crevasslng Is an Integral mechanism of deltaic sedimen­ tation. Once delta distributary natural levees are estab­ lished, further clastic accretion takes place only through overtopping the levees. This occurs relatively Infrequently, but crevasslng provides a continually discharging channel Introducing clastic material throughout the hydrologic year. In a delta with many lnterdlstrlbutary bays, this is an important mechanism of deltaic sedimentation. In addition to being an Integral mechanism of deltaic sedimentation, another Important consideration Is that crevasses can be viewed as a natural model of deltaic sedi­ mentation. Gilbert's law (1884) statesi "The capacity and competence of a stream for the transportation of deti*ltus are Increased and diminished by the Increase and diminution of the velocity." This law, although oversimplified, pro­ vides the logical basis for studying a crevasse as a natural delta model. Coleman, GagUano, and Morgan (I9 6 9 ) have shown that the large subdeltalc masses can be used In pre­ cisely this manner. This oonoept can easily be extended to Include small crevasses breaching the natural levee and building deposits In an lnterdlstrlbutary bay. The compact 4 crevasse offers conceptual and logistic advantages. Thus, processes of sediment dispersal readily lend themselves to direct analogy with the larger delta. Knowledge of crevasse processes has direct practical application to engineering problems. Since 1922 artificial breaks In the natural levee have been used In the Mississip­ pi River Delta to maintain natural levees undergoing severe erosion because of heavy ship traffic (Dent, 1924), This practice has also been applied to Southwest pass,

SALIENT CHARACTERISTICS OF CREVASSES

Crevasslng, Inherent In deltaic sedimentation, was recognized by Gilbert (1884), although he did not employ the term. So long as the bank approximates closely to the level of the surface at flood stage the current across the bank is slower than the current of the stream and deposits slit Instead of excavating. But whenever an accidental cause so far lowers the bank at some point that the current across It dui'lng flood Is swifter than that of the main stream, there begins an erosion of the bank which lnoreases rapidly as the volume of escaping water Is augmented. Continuing, Gilbert points out that the success of the cre­ vasse depends upon the slope of the levee and whether or not the new channel provides a shorter route to an adjacent basin. Crevasses occur In both alluvial valleys and deltas. In configuration, crevasses In the alluvial valley resemble those in the delta, but they are quite distinct hydrologl- cally. Alluvial valley orevasses Initiate and reach peak 5 discharge during flood stage, accrete during falling stage, and are abandoned as flood waters drop below the level of the controlling thalweg of the crevasse channel. In the delta, tide and river stage Interact to maintain the crevasse channels throughout the hydrologic year. In reality, no sharo boundary exists between the two types, but a steady gradation occurs from valley to delta crevasses with proxi­ mity to the river mouth. Johnson (1891)■ Elliott (1932), .^ssell et al. (1938), Russell (195*0. Happ (19*^. 19**8), and Trloart (1955. 1956) discuss alluvial valley crevasses. Delta crevasse sizes range from less than a square mile to massive subdeltalc lobes. The latter may be over 20 feet

In thickness and In many cases cover more than 50 square miles. Pour subdeltalc crevasses are currently active In the Mississippi River Delta, the two most active being

Cubits Gao and Garden Island Bay (Russell, 1936i irfelder,

1955. 1959i and Coleman, Gagliano and Morgan, 1 9 6 9). Smaller crevasses, one of which Is the subject of this report, breach the distributary natural levee, depositing a solay b or 5 feet In thickness. These usually cover less than a square mile of the adjacent lnterdlstrlbutary bay, differing from their larger subdeltalc counterpart primarily In scale. The principal result of the smaller features Is to fill adjacent lnterdlstrlbutary bays by accreting local­ ly higher, wider natural levees (Pig. 2). The forces governing water and sediment dispersal In the delta crevasses are Inertia, friction, and viscosity. Figure 2. Sealed Crevasse along Raphael Pass, Cubits Gap Subdelta (Center Foreground).

Borishansky and Mikhailov (1966) report that Inertia and friction forces dominate water and sediment dispersal In a channel debouching into a basin. Bonder (1970) considers buoyancy (freshwater/saltwater density contrast) as most important for the Danube Delta, where density contrasts are sharp, Wright and Coleman (in press) Indicate lnei'tla, friction, viscosity, and buoyancy as the forces active at the mouth of South Pass, In the presence of a salt-water wedge. In the Interior of the Mississippi River Delta, freshwater lnterdlstrlbutary bays are common, eliminating the force of buoyancy. Discharge, tides, wind and trans­ ported sediments constitute Important Influences on the In­ tensity of each of the three remaining forces. 7

SALIENT CHARACTERISTICS OP THE STUDY REGION

The distributary system of the Mississippi River Delta

Is dominated by trlfurcatlon of the main channel at Head of Passes and by four subdeltalc crevasses. At Head of nasses (Fig. 3), the river divides Into Southwest Pass,

South Pass, and Pass a Loutre. Pass a Loutre carries ap­ proximately 37# of the discharge. Southwest Pass 29#, and South Pass 15# (Holle, 1951). The minor outlets of the four subdeltalc crevasses account for 19# of the discharge. Mississippi River stages for the period of this study are presented in Figure stages for two delta crevasses are shown In Figure 5. The crevasse hydrographs are 15-day run­ ning averages of mean diurnal water level, Holle (1951) calculated discharge of 300,000 cfs at New Orleans for a stage of 3 feet (low stage) and 1,000,000 cfs for 17 feet (flood stage), U. S, Army Corps of Engineers

(1959), NEDICO (1961), Santema (1966), Tlson (1966), Dronkers (1969), van der Made (1970), and Wright (1970) report dis­ charge variation In delta distributaries resulting from gradient changes during the diurnal tide (Fig, 6), The diurnal tide of the Gulf of Mexico results pri­ marily from the moon's declination (Marmer, 195^). The mean diurnal tidal range at Head of Passes is 0,9 feet, and mean tidal level Is 0,U feet. At the mouth of Pass a Loutre the mean diurnal tidal range Is 1,2 feet and mean tidal level

Is 0,6 feet (U, S, D. C.t 1969). For a more detailed dls- 8

is 89

GULF

1/ F, X I C O Mi

te v v ,«VXLt'' ’> **''■* / i », t

,wW > /;.-vi

1 ; - ' . / — 'ute :te..? J^\Si-‘ /c "tev te'te- ? ’ i ': vt $

‘i 'V> > Vte' ..•*$£> v R DEN ! i LAXD £ « r ^ ’’^ -'- BAY BAY -h S DU f fl — ■ 29*

S o v t h we * t P** * * 10 • * I $ > Hi I u I * iSrt i I a t __L30r 13 Figure 3* Location Map of the Mississippi River Delta STAGE AT NEW ORLEANS (FEET) 12 14 10 8 4 6 0 2 Figure j. ------4 . 1968 ie Sae yrgaha arltnGg (New 1968-1970.Orleans), Gage atCarrollton Hydrograph Stage River ------— ------•— ------969 6 '9 ------1 ------— ------1970 ------n O Figure Ta GE (FEET) 10 12 14 16 0 2 6 4 6 AUG 5 . River Stage Hydrographs at Rattlesnake and Little and Rattlesnake at Hydrographs Stage River . Lake Pond Crevasses Pond Lake 901969 I960 oc r V O N t FE 6 OtC A A P JUN APR MAR JAN t 1968-1969. MAY 10 11 Jonuory 33-37, 1969 Stag* a) N*» O iU oni - 3 0 fl. |Lo*r itogo)

J o n . 3 3 Jon. 36 Jun. 37

o Htod of Potto* * 3

> o

Pori Eod *

I 3 0 I 3 1 3 H our*

January 31- Ftbruory 3, 1969 Stag* ol N«w Orl*an* - 7.3 It. (Normal Hog*) F*b, 3

V„./

1 3 0 I 3 Hour*

A p ril 19-31, 1969 Stag* ol N*w Orltani = 117 fl. (High ilag*j Apr. 19 Apr. 30 Apr. 3)

x X. X X 0 13 0 13 0 13 Hau r t Figure 6. Diurnal Gradient Changes along South pass. (Froa Wright. 1970). 12 cuss Ion of tides In the Mississippi River Delta see Scruton

(1956). Strong, persistent winds, usually associated with regional frontal storms, cause surges which frequently re­ sult in tides In excess of 2 feet higher than normal. Surges associated with hurricanes attain heights 15 feet above mean Gulf level (Gagllano and van Beek, 19?0). The volume of sediments transported by the Mississippi River is estimated to be from 1 to 1.5 million tons per day (Fisk, et aJL., 195*M Gagllano and van Beek, 1970), The suspended load of the delta distributaries is composed of 5* fine sand, 31^ slit, and clay at low stage (39*0 000), and of 8* fine sand, slit, and 98* clay at high stage (859,900 cfs)(U.S. Army Corps of Engineers, 1939)- 8©d load In the delta Is composed of fine sand, estimated by Holle (1951) to be 20* and Fisk et al. (195*0 to be 10* of total trans rorted load. DATA COLLECTION AND ANALYSES

Rattlesnake Crevasse (Pig. 1), located 3*5 miles below Head of Passes on the right bank of Pass a Loutre. was selected for Investigation because tide gage and current velocity data were available for this actively prograding crevasse, In addition, similar data existed for Little Lake Pond Crevasse, which breaches the right bonk of Raphael Pass ^ miles below Cubits Gap (Fig. 7). The latter was included in this Investigation to Indicate the variability of Mississippi River Delta crevasses. Water flows through Rattlesnake Crevasse Into a small freshwater interdistributary bay bounded on the southeastern

Figure 7. Crevasse Into Little Lake Pond.

13 14

side by a service canal spoil bank, A broad, shallow apei'- ture opens to the Gulf of Mexico through the main body of Rattlesnake Pond. Reversal of flow characterizes both open­ ings In the debouching basin (Pigs, 8 and 9). Little Lake Pond Crevasse discharges unldirectlonally into the Interdistributary bay. Openings to the Gulf through other Interdistributary bays provide unimpeded access to the Gulf of Mexico. Reversal of flow is apparently rare through this crevasse* Enuner (1968) reports a reversal of discharge direction during wind surges which pile up water in the northwest c o m e r of the basin. However, it was not shown to have occurred during several wind surges recorded by instruments,

STATISTICAL ANALYSES

The data collected to identify and evaluate the vari­ ables governing the evolution of Rattlesnake Crevasse are analyzed within a statistical framework. Multiple regres­ sion, analysis of variance, and analysis of covariance pro­ cedures were employed. Cochran and Cox (1957), Li (1964),

Krumbein and Graybill (I9 6 5), and Snedecor and Cochran (1967) present the theoretical explanation of the statis­ tical computations contained In this study. The statistical tables employed were compiled by Raid (1952), General purpose programs In the L. S. U. computer library were used in the statistical analyses. Multiple Regression (MRF49) 15 I 1 V ( l O t 11 ¥ [11/ t* C I v r t o t it v ; l ■■ /» * < | n o t 11 v t 1 l ' U l M v h |.« < *•«] IE EE (nhi SLNT [. ^ IE EE SLNT [%0) SALINITY LEVEL TIDE ^ [%.) SALINITY (inthei) LEVELTIDE o 10

< »- 1 m ^ o o o 3 ^ A ^ f >

0 > TT -r¥- ■^i dr Ti -tV r r -TT— T r time tim e

?UiPtNDfD I 0*0 (n>g/l| rB SMIN1H (*,.) tn E ILriPINCtOiAHHIH lO a O 1-,/U o • 0 ■■ o 10 l/l < > o

£ *«- o Z 0 LLJ

0* TIME TIME Figure 8, Circulation In Rattlesnake Crevasse at Equatorls Tide, Low River Stage. A. Crevasse channel, B, Outlet

VhOOTY jit/ i»i|

\ 70 I J

' J o o O I

TIME TIME

r 1 0 300 o SUSHNbtO 10 40 1*14/11 o < S A IIN If ¥ l*^.| < itlSflM j.'O L'JAD \m,y 11 i AMNlT f l%*f o Z n 0 1 H

z U i

TIME TTME

Figure 9. Circulation In Rattlesnake Crevasse at Tropic Tide, Low River Stage, A. Crevasse channel. B, Outlet. 16 with a backward deletion technique and Least-squares and Maximum Likelihood General Purpose Program (LSMLGP) were used on the L. S. U. IBM 360-65 computer.

Crevasse Channel Current Velocity

A continuous-recoi'dlng tide gage and a self-recording current meter were installed at the two crevasses by the U, S. Army Corps of Engineers, New Orleans District px*lor to my Involvement In the study. The current meters, con­ sisting of Roberts-tyoe bodies and Gurley-type heads, samDled the velocity for one minute on the hour. The meter was susoended from a tower (Pig. 10). Water level fluctu­ ated above the meter, and so, depending upon river stage and tide, depth of the sample varied from 1 foot to 3 feet.

Figure 10. Instrument Tower at Little Lake Pond Crevasse. 17 Because the sampling program provided matched obser­ vations of water level and current velocity, the data are best analyzed using a multiple regression. The data were collected before a model was developed and In a sense gene­ rated the model. The results of the multiple regression analyses were used to generate new hypotheses and additional data collection. Although wind frequently Influences currents, data on wind velocity and direction were not available. According­ ly, before the analyses In Tables 2 and 3» data for all days when wind noticeably deformed the tidal curve were dis­ carded. This removes a source of variation, as much as possible, that could not be properly handled statistically. The first multiple regression (Table 1) establishes the importance of tidal range and river stage In determining the duration of current reversal In Rattlesnake Crevasse. The dependent variable Is the number of hours currents were stonoed or reversed in a 2^-hour period. Mean water level constitutes river stage (linear)r tidal range (linear) Is the vertical variation of diurnal water level. The second multiple regression analysis (Table 2) assesses components of diurnal tide and river stage In determining current velocity in Rattlesnake crevasse channel (Pig, 11). Change In water level constitutes the tide Inde­ pendent variable. Both llnaar and quadratic responses were evaluated. The third independent variable is the cubic ef­ fect of river stage. River stage, in this analysis, is Table 1. ’:jItirle r^rrossion .i.naive is cf reversed arc .t Ji«. - ** t" +■ ^'"■'■-‘ -■*■■1 or. Oiurral ~iie •- * y f» ** ^ ^ •rri f T1 ^ * r ° r' . -T c ’r^v'-in . _^ , __ __._ n _ . ______.,. . „... ..___ „ **■ t. * ^ increase < T** <- V - -''- b„ b r H ^-■w * t'- * 1 ^ Z1 UT 'I ^ r. 7 ^ Z * C+ + ' f- v 0 r ^ + i:-^ rna.''o 6. r0 * % — / 4 • > it'. . . r ' > > s. Vt./r" _*"» Zl ^ r> n . J • _ • * • - * _ ^7.i6++ d # o Zj £, -r'p

Table- 2. !dultirle ?.~, ™ession .Or ulysis of Ha ctlesnake Crevasse Current Velocity on Hiver Stace ( Cc c 1 c) and Charge ir. Vater level (Linear and Ouadratlc j.

increase V* H 7 c .f . i n 7 si 0.C1 .' £ f *■ ■*■ I?6.0++ l&^-d —* ’. * ^• • v'cior level o r\ 0.06 Z-.'f** 317.3*+ U765 / ; .VJ ^ ■ ■ T- -*.-i ' 1" v.-.:r letel ■"i ■*. -1.06 - r .V o * * O'1r #...7-t* if. - • zee..-••k j £ . -> nicn 0.27? 0.?/.** 0.70 ?$9.d+* 3i7?" 19 4 0

'' — J. fit

JO

— — — M u r 4, 19 6 ? S ta g * 1 4 — * —■ jv*# 10, 194 9 f**|s 3 3

-L -10 -10 0 0 *10 * i o * 1 0 -VELOCITY It/ toe +

- — J-"* 30, 1969 tfapi? T.

I 0 0 — VELOCITY ft/,«< + Figure 11. Effect of River Stage and Lunar Tide on Current Velocity. Rattlesnake Crevasse channel. A. Effect of river stage at tropic tide. B. Ef­ fect of lunar tide at low river stage (2.2*)* C . Effect of lunar tide at high stage (2.9*). 20 defined as the 15-day running average of mean diurnal water level, The third regression analysis (Table 3) was performed on the data for the crevasse Into Little Lake Pond (Pig, 12), using the Identical model and variable definitions of the preceding analysis. This test Included linear, quad­ ratic, and cubic responses of river stage In addition to linear and quadratic interactions between tidal level and river stage. Justification for the use of nonlinear responses of the Independent variables Is provided by the hydraulics of open-channel flow. Discharge In a 90° branching channel varies with the square root of the water-surface slope (Chow, 1959* Goncharov, 1964). As water-surface slope Is directly related to the vertical velocity of the surfaces of the distributary and the Interdistributary bay, conditions favor a nonlinear response between current velocity and the Independent variables. A second cause of a nonlinear response could result from crevasse-dlstrlbutary geometry. Plow In the crevasse channel Is zero bounded (Fig. 11). During low river stage, both crevasses frequently dralr completely, and current cessation lasts for as long as 10 hours. An additional Inflection point may be caused by overtopping of the natural levee, especially during flood stage. Kondrat'ev (1959) reports that, for a river, maximum current velocity Is reached Immediately prior to levee overtopping. Table 5. Ivoltirle Ro.rrcssicr Analysis of Little L:C:e for.i Crevasse Csrror.t Velocity on Variess I r_ i r: e r i e n c V a r:c b 1 e . n = 677

Ir.de' er.dert 0 ^’ increase r\ Varies le br b p P py p d.f. in yy

os, critic ^l.Q ■ ■ . ; 4Q?. lt-e?- 4q .2 R i v ■.? r c ^ . •' ’ - - r- ^ ^ r V <4 ;ia « < . • “• < t* '•• > - ■■’ j -. • -c lk£2a 16.1

Z ' -.Cl 1 ;•• t- ■w" # 0 A?6 •»-» 19.9 Cnr. -; ir. w.c.- l.-vel - ^ - r >' r •* * *-* 4b . 1 '-• « - ?C.6?*+ 1AS 79 1.7 Char, re ir. water level ■* -r< -a *■ ^ «■* S7CC.1 V' J . > . . S. _ S.. ” * 7r no**

V c + v _r 9^**. ^q 1 ^ e n d ratic -660.5 •— ♦ i f ^ ' • * H S 29 .*. •

Tile x oca. -e j- s i T J > — fc- - 1 2 . 6 11.07** 11-579 r * n CO, " y .-> omtg s ? i on 136.^ 0 • :'C** W ^ ' s > 479.£?** 7&R29 22

4 0

* 0

J 0 0 0) 0 J 0 1 0 VELOCITY U/tmc

4 0

1 0

I 0 0 0 1 0 7 0 J 0 VELOCITY h/s«c

* o -c

J 0

1 0 3 00 0 I 0 J 0 3 00 VELOCITY lt/*»c Figure 12. Effeot of River Stage and Lunar Tide on Current Velocity, Little Lake Pond Crevasse. A. Effect of Stage at Tropic Tide. B. Effect of Lunar Tide at Low Stage (2,5')* C. Effect of Lunar Tide at High Stage 0.5'). 23 While these results suggest that river stage and tidal

components are important factors governing current velocity, a 2x 3xU^ factorial arrangement of factors was set up to

provide further Information about the effects of river stage and tides. For this phase of the study, Rattlesnake Crevasse was selected,

FACTORIAL ARRANGEMENT OF FACTORS

Principles of open-channel flow establish that dis­ charge through a channel of fixed cross section branching at about 90° Is primarily a function of gradient between the channel and the receiving basin. This principle governs crevasse channels In the Mississippi River Delta. Accord­ ingly, factors which give rise to gradient changes between the distributary and the lnterdlstrlbutary bay In the delta were Included In an Investigation of discharge variations in Rattlesnake Crevasse, The previous multiple regressions (Tables 1, 2, and 3) as well as the literature (U. S. Army Corps of Engineers, 1952, 1959* NEDICO, 1961* Santema, 1966*

Tlson, 19661 Dronkers, 1 9 6 9* and Wright, 1970) Identify three main factors which cause gradient changes in a delta. The first Is river stage, which constitutes one of the most important for any fluvial feature. As the flood wave ad­ vances downstream, the gradient Increases between the Gulf and the delta distributary. Then, as the flood wave passes, the gradient decreases. The second source of gradient change Is the rise and fall of sea level in the Gulf of 2b

Mexico. The tidal component can be broken down into semi­ monthly tide and diurnal tide, each of which can be con­ sidered as independent of the other. Diurnal tide, the rise and fall of sea level during a 2U-hour period, causes a diurnal variation in discharge. The semi-monthly tide com- nrises the 2-week cycle during which tide range increases and then decreases, causing a day-to-day variation In dis­ charge, This is the third Important factor. The effect of the semi-monthly tide on current velocity is shown in Figures 11 and 12. In addition, location, which is composed of the three profiles at which observations were collected, is a fourth Important factor. This is Included in the following statis­ tical analyses to determine if the two main crevasse chan­ nels in Rattlesnake Crevasse react differently under chang­ ing combinations of stage and tidal conditions. Each of these four main factors will be defined and explained in more detail in a later section. Variation in a dependent variable frequently results from interactions between the primary factors. This is especially true with fluvial and deltaic phenomena. A statistical technique capable of simultaneously evaluating the main factors and their interactions is the factorial analysis. By analyzing interactions, the factorial classi­ fication yields detailed information about the pattern of water and sediment discharge variation. The factorial analysis is explained in most basic 25 Inferential statistics textsj therefore, the underlying principles will not be outlined here. Because there was only one replication in the following analyses, the ex*ror term is actually the pooled higher order interact ions (for discussion of this procedure see Cochran and Cox, 1957). The factorial arrangement was used for two dependent variablesi water discharge and suspended load concentration. The variable definitions and methods of data collection are presented below for each of the dependent variable analyses.

Discharge Analysis. Rattlesnake Crevasse

Three profiles across the channels of Rattlesnake Cre­ vasse (Fig. 13) represent the location main effect. Five segments of equal width divide each profile. These inter­ vals are (A) 1^.5 feet( (b) kz feeti and {c ) 12 feet. On each line, velocity measurements were obtained at 2-foot intervals beginning immediately below the surface. The bottom measurement was taken within 1 foot of the channel bed. The current meter used was a Narine Advisors Roberts- type model B-3C, which produces a readout of both velocity and direction. Discharge was determined following the procedures out­ lined by Wright (1970). Discharge is estimated from

Mt = £Uj(dz)j J where u is the velocity, Az Is the vertical Interval be­ tween observations, and J refers to the horizontal interval. 26

PASS LOUTRt: CHs£> I -

200 400

Figure 13# Location of Profiles and Accretion Measurements, Rattlesnake Crevasse, 27

All values are expressed In cubic feet per second (cfs). Two field trips were made to cover low anlhigh river stages and form the second main effect. The first level was at flood (10.1 feet to 12.9 feet at New Orleans), from May 29 to June 5» 1970, Current meter problems developed, and several observations were missed (Table 4). The second field trip, made at low stage (^.1 feet to 5.2 feet), was conducted from October 17 to November 1, 1970. Four days of repeated observations during the semi­ monthly tide cycle constitute the third main effect. The

selected days represent (1) equatorial tidei (2) midway to tropic tide* (3) tropic tide* and (4) midway to equatorial tide. Likewise, repeating the observations four times during the diurnal tide comprises the diurnal tide main effect. These are (1) low tldet (2) flooding tide (midway to high tide* (3) high tidei and (**) ebbing tide (midway to low tide). Two analyses are presented for the dlschax'ge variables (Table 5). The first combines the results of the June and October field trip, which was made to determine the signi­ ficant variables affecting crevasse discharge at low river stage. Tables 6 and 7 present the least-squares means of

the two analyses. 28

T , v .9- >1 •1 Ar.'. C:4' Vn r■ ? : r r t ]; ■ 1 1 t 1 •. - r ■ 4. c ;VcV'.

J\.r ■ A • () ^ (V to'' r , — — ' s 4 / fi . JV :\V. v ri .‘f . ~ 7 y “

Tot :• ' <;3 4 9 S t-/ i- 1 87.4 5’'*' . ■+ V 1'+ r 1 -i *3 f ■ / * ni »• * t_V" b.Go S *•(■.•* 7 1 VU ' 1 37.90* 1 21.77** 7 VO 4 1 41.73* 1 77.95** l . 5 v r 2, 4 1 96.67* > 1 7 4.1 7** B iu rn ul Title 3 74 • 71' 3 4.3.91** ] v r- 5 1 10.97 ] 60.7 5** P Vi: 4 1 1.57 1 17 .94 * 1. 5 v n ? , 4 1 39.6 7** 1 ?4. 17 * * L o c u tio n ? 707.4 6 ■ ’ 2 47.07* * B vc C 1 1 3 7 . ' ^ * * 1 2 0 .50*> '> S ta rr X St.-r.i-:: on . i 1 7 .0 7 *' S t a r t * X B i u rn.u 1 3 3.47 S t a . v X I, /: ~ . t . ' n 7 6 9 .0 “ * * Soirji~i;:o. X jjl u r ; ; ,'i i 9 17.94* 9 19.99** Ser.i-r.n, a l o . - n t i o n 6 e .74 6 6.07 Diu rii.il X Loo.'!t'i on 6 17.64 6 10.17** E rr or 70 ...... 9-9/> 16 7.5 5 29

T; M - i— 1 (v

l ■, ; yj ' i _ \ y ... ■ ■ 'T "Jrj _ /t_ __

‘ < !) b > - 7 > :■ ?~fC m m L ' ? / r,t , f 77 7 AC'A P10 * .■ .* ( V" 6‘*9 c;*' 776 t : .... ■i / ^ t / ■- r <"1 J y / i - A*‘J> r | 1 ; ( n r l V’fvJ 'K 7 -r^ ] V- j i (-'vT - ! '''- 7 6:lp 67 n 7. • . r 7 n 7,: .'■ A(A) i i j V-n ‘A_ _Vii. :n ____ } r.r i i s, ji ’ _vv/___ < :S S O 1, r/,7 —VIA f rn 77 ■: / ?a

:sj n: riry (cf: ): Ir

126 30 Suspended Load Analysis. Rattlesnake Crevasse

Sediment discharge is as Important as water dlschai'ge to processes of crevasse evolution. During preliminary field work, water samples were collected at each point where velo­ city observations were taken. To limit the number of neces­ sary samples, a "paired t-test" was conducted, pairing the mean concentration of suspended load for each line and the surface sample (Table 8), The null hypothesis was accepted. There Is no difference between the surface sample and the average for the line with respect to suspended load In Rattlesnake Crevasse. Accordingly, for the factorial anal­ ysis, water samples were collected 6 Inches below the sur­ face of each line, Determination of suspended load was obtained by a Mllllpore 100-ml pressure filtering apparatus and Millipore

47 -mm filters with a 0.45m pore. Filters were oven dried at 85°C and weighted to the nearest 0.0001 g on a Mettier Analytic Balance. In most cases, 100 ml of sample was passed through the filter, although at low stage 200 ml was frequently used. All values are expressed as milligrams per liter (mg/l). Samples which deviated sharply from others taken at the same time were analyzed a second time; the results were then averaged. The results of the statis­ tical analyses are presented In Table 9. Definitions of river stage, semi-monthly tide, diurnal tide, and location remain the same as those in the water discharge analyses. 31

Tablr P. T - i :-f»d "I,-'1', t" cf V.-rti r.'H liistr: but ion cT bunmr.dcd bond ’ i, j -J • 1 'I'D. * n - Mi d - Ob:'-.- 15 i ^ (' , ■ ■ ■ . ■’

ri - : - - 3 .V-jO r. 7'1 d .f.

*;■ x- •■■,■■■■, ■i a thin "tost w ‘in.1 i'rc 7'. J> x x : ,bt *r , .l-.riuary, an! A m i l l"i '!.! ; j’i : v,.

T1'lb’1 ■'■ '‘. ! ’ 1 i,ip r ^ ■: .' ’ Vx: Cv‘!, itX 'ty o n An n't;' : rb V;;-' . • •. 7' ' .

.Tunc- >i O rta bn r 0 r i r. V f ■ r J cj'.i rc - d . f . ":.T r7 V T n '' d 7 f7 ’*.!•. x ] 0

T o ta l 6? 43 :•* •„•••• 1 2464.10*» .j^ir.i-rrojib1. ly ;iu t‘ 5 HPtHy’* 3 IP-}. 16*' 1 vs 3 1 1067-4?** 1 3 4 .2 ! ** 2 V'; ^ 1 5 8 .1 0 ‘ 1 204. 00-* ¥ 1, 3 vt: ?. 4 1 ?0? P .30"* 1 14.1? lHurnal Ti r 3 210.36** 3 3 /,4 . n0* * 1 v : 5 ^ 1 73.4?*-* 1 204.00* » ? V!' 4 l G3.1?** 1 3.01 J.1 * ^X V w i 'J. t 4 1 ?R9o.50*» 1 14.1? Locnt iot: 2 0.08 2 4.31 B vf. C 1 1 33 .74 11 * 1 6.23 A vs B, C 1 ?06. 00* ■* 1 0.04 Gtayo X • -ro r.tb jy 3 101. 01** B ta /x X Lu'.yr, ■a Tide* 3 2 1.0 7** b t X X X LOf.':4,i on 2 0.33 t'ami-r.o. X hi x rn a l Q 4 1.7 0** 9 33.?6** Gori-rro. X Ixr a tio n 6 30.70* 6 3.13 D iu n in l X ba - r t ion 6 3.30 6 3 .? ^ I:i^xr.. v OY bi a par 1 6.33* 1 0.73 Tor.]:. CuV eai* 1 102? ,7 3 ** 1 921.07* * K r r o r 22 4 .2? 1 ? _____ JL-5? . 32 Because of the Importance of temperature and discharge

on suspended load concentration (Carey, 19651 Colby and

Scott, 19651 and Burke, 1966) these two factors were in­ cluded as covariables in the analyses of variance. The least-squares means (Tables 10 and 11) are, therefore, ad­ justed for the average effects of these two covariables. For the analyses, water temperatures, wind observations, and salinity detenninations were collected along with suspended load measurements. Surface water temperature was measured on each line, using a Celsius thermometer accurate to 0,1°C. Salinity was determined by titration and the

formula in Strickland and Parsons (1 9 6 5). Wind direction was measured with a Brunton compass 1 wind velocity, with a nortable Casella anemometer. Additional information on water temperature was avail­ able in thermal scanner (Reeonnofax IV) imagery of Rattle­ snake Crevasse, The Imagery was obtained on October 22 and 2 3 during four flights over the study area. Each flight was flown at altitudes of 3,000 feet and 6,000 feet. Ther­ mal scanner (Reconnofax IV) imagery provided data on the diurnal pattern of temperature change in Rattlesnake Pond which is presented and discussed in the next chapter. The diurnal pattern of temperature change was confirmed by field measurements (Table 12). 33

Tahir 10. fhirprn:!''d LoiH ncjunvt-;; .M.-ani; T:.' Mirh Low '• • . ______

1?0 L_' ••LV 1 ? 1 Li' 'i ' ,r OIL 2'jl /.■’?! 40? 104 ;vo oil ;nb 146 iL't loo m ■ 1 ?19 1 2 1 'i ... j 67 r n 076 040 31.7 V)'j v “ f 1 * '• pT' if*7 704 : 6r pco .1r': • 1 } .■'i'- )'-i v.>-< 2^2 ' > y/r* 1f 7 0 3 0

10 jr>o L04 ?6b |

Table 1 I. Orr rrrH n -d Load T.e ■rofi !'• >r.nr ( r ; y l ) : _ J.ow 0 1 ’;<

LOc.vno;;

_ .w ___ ?4 3 or 106 f n 70 06 Table 12. 1' *** ? n S115 ""°rded Load Ccr. cent ra ti Qym X^i-P "*3" ^ " ^ /• V

HIGH STAG?:

■£^7 ’’'t x X1 ?li Ac cat lor. A X £ r T1 lor 2 *i * — n ' >-i • rt 1 r- c Siseh. .._* * ! ■rc L _ . . - _. k.* ■ . • O Ti jG ■* * 2 1 3 V y U 24.4 2?3 765 ta .4 3 2 7 179 24,4 ?: oodir.f" ------__ _ _ ** _ _ _ ~*z z i ri ^ ?'■? 1 440 24.4 240 C * £ 24.4 . 2 \T • -. ■^ • /. * j*>

, iov 'cly to .T'sT-ie . T , *7 .L^ * 527 ca;w .• J 24.4 Z ^ 74.4 521 1 ^ 2 74.4 3 7^, T?”*. — w^ w nk X.-i ^* r* 3-1 1?37 24.3 342 1 2 5 0 24.? s . " i ■ 24.2 j >“. - rt'l S'"-'? .ce 1T02 .' w « r ------X-J 7 .•' '> ■j - "C ' j • . L ---- 10' 6 24.4 4 ' ■ 74 .4 2~^ • *.

2106 1C17 id* ay to ""■itorial

*AI1 discr.ar/T? values are exrrrsred ir cfn. Tabl° 12 (2ori'^lrl.V— .or.cor.zT^- r;*on, > 1

L C W S?ASE

;ua ia l cation A Loc

77? r s, 1059 212

19.5 1?.4

de 2C.9 166 117 ?c.e 2G " 2G

Lrt 36

Sediment Accretion. Rattlesnake Crevasse

Sediment accretion was recorded by Installing 3/8-Inch aluminum rods mounted on a 1 ft3 piece of 1/8-lnch thick aluminum plate. These rods were positioned during field trips In November and December, 1969» on a grid of de­ creasing spacing along the crevasse channels (Fig. 13). This grid was used to provide greater sensitivity near the end of the channels, where splay growth was anticipated to be greater. The sampling pattern was established by measuring 600 feet from the edge of the right bank of Pass a Loutre olus an additional 7 feet (determined from a random numbers table). The Interval of 600 feet was selected to avoid Influence of overbank flow from Pass a Loutre during flood and storm surges. Further profiles were emplaced at 1,500 feet, 1,950 feet, and 2,175 feet. The last profile was located beyond the end of the channels In two of the three cases. Three of the four channels In the crevasse splay received profiles. Each profile consisted of four to six rods. The first stake (0 foot) on each side of the channel was positioned at the edge of the channel. Subsequent stakes were located 100 feet and 250 feet from the 0 foot stake. To minimize scoui*, the plates were burled 4 to 6 Inches under the sur­ face (Fig. 1U).

Determination of acoretlon was accomplished by mea­ suring the length of rod extending above the surface 37

Plf^ure ]>. Accretion Rod on Crevasse Splay Natural Levee, Rattlesnake crevasse.

Figure 15* Accretion Leveler 38 (Fig. 15). The difference from measurement to measurement constituted the net accretlonary rate. Mean dally accretion was used In the analyses presented In Table 13, Tables 1*4- and 15 present the least-squares means. To relate the analyses of discharge and suspended load concentration to growth of the crevasse splay, sediment accretion was monitored for one hydrologic year. The ei- nerlraental plan used for the sediment accretion phase of the Investigation was based on three sources of variation. The three sources were hypothesized from the geometry of the crevasse splay. Statistical evaluation was performed on the basis of channel, profile, and side. First, the splay is divided Into units by the channels formed through bifurca­ tion. Thus, each channel gives rise to expected variation by forming the network through which sediment Is Introduced Into the splay. Second, since the splay decreases In height with distance from the orifice, a second division was per­ formed and termed "profile," The third, called "side," results from considering each side of a profile separately. Accordingly, on the basis of these thi*ee expected sources of variation, a least-squares analysis of variance for un­ equal subclass numbers was employed. Density and water content observations were not col­ lected, Accordingly, no comparisons were made across measuring periods by combining results. Thus, the net ac­ cretion represents upward growth of the splay and not volume of the sediment. Consequently, loss of sediment, particularly 39

r - r-* {\j o t-i r— <11 CM OJ »-H t ) On H

O CM N ‘n fO CM Xf

•H

0-J OJ C) t )

>

H ^ CNj OJrJ

IT* vJ t'i CJ * 0 *- t u o + > ot hO ■*;?CN in O o\ * * * • • vf (3: ■\'.c

t: o rJ no OJ rH 4 ^

O O -t; r- r— a) U > CM ♦ ’ ■ * KQ tb CM Ur I rti C.'J

o r j uo ao r " ^13 hO \Q n *r; r i vj- • * * • O' m r n i-H uioi C o h o C7\ hi \ cCL) > ✓ vt3 ; l rI a \ i " ■r< O trf ^ - s' CO

J ° *■, o , 1 fj? r- i r H #"3 CL> II i*»W * r < C O cl

CM N > f—i r i 0 ) r i ,*2 ,i £ ' C/J t.1 CJ O [rj UO

T/il. I U v ■ ■1 til A - ■ 1 An !'•■

jaat;/V-l sf KAHCH rTV.!;1’9\ c:\\' ’ *> ' J > i y J-

c, T J i ■ 11* j c 1 A h 6 . 0 r * -'j i r £ ■1 i 7 15 ■r * c 8 7 1 1 7 9 " C •: - o j ■}; -116 --V. _7 u O 3 8 7 6 0 6 0 7 3 4 *. I ' T I n JL. 1 ’ <■ n [ 4 ) f O i ?Q Cl 8 6 j 1:3 -14 -117 -}Q l 8 1 1 9 7 f f t : • ■/ i -‘1 176 6 7 -ir- -m _ 7 _ ___ 4 ? _ _ V' L. -101 -71 ~V- . X 6 6 8 7 6 3 7 1

A K! 1i. jut;

CllVJ’ili*.______nLMiir-’i, 1 1 c '> V "T "36 ~ 63' Q! 1 /j o 7 4 6; ~ 7 T r ^ f * ' J 104 I1/, 1/17 f 61b 461 9° 7 T 1 7 r/' ■> i n 7 ■"l -T ^ 1 ; L- . -. J. 1 o 7 < P J* 633 403 * _] 4 ^ ■: ? 7 1 ! 4 p. 1 ')2 6 79 *4 » 1 ■ _4_. - ^ 2L- c* Cj Cl i 17 94 137 6 9 T.1 1 >97 473 741 977 f-) £' 1 1 0 10J 1(7 111 CJ i ; ’ 7 9 700 1 070 69^

A A 9 9 — ____■ __ 81 177 71 7 986 f 8 9 909

OCTOl^T? 41

__r} TahJ <■' 1 Q . Annin 1 Podjnont Accre tio n T-p-int-_fU M rrr Vi-anD _{ x 10 ).

P ro file Channel Pi dr 1 77 103 03 2 122 10] 133 3 112 133 A .. 14.4. ... Grand Moan 114

zreat during falling and low stage period, Is a reflection of dehydration, compaction, and erosion. Figure 16 presents the lsopach maps which are Interpreted and discussed In the following chapter. -p r\>

Figure 16. Accretion Isopach Maps. A. January. B. March. c. April. D. June. E. October. F. Net Annual Accretion. INTERPRETATION

■1A.TTLES\TAKE CREVASSE DISCHARGE

River Stage

The significance of river stage In determining dis­ charge through the crevasse channel Is revealed In Tables 5 and 6 , The combined factorial analysis shows a highly sig­ nificant difference In crevasse-channel discharge between high and low river stages. Discharge at flood normally is over twice that at low stage as a result of higher current velocities and less reversal of current direction.

Semi-monthly Tide

The results of the factorial analysis cannot be ac­ cepted without reservations concerning the semi-monthly tidal effect. This Is because prior to the October field oeriod a storm elevated the water level In the delta over 1.5 feet higher than normal. The high stage caused by the wind surge was falling through the first week of the data collection neriod, partially confounding the effects of the semi-monthly tide and wind surge. The least-squares means (Table 7) Illustrate the effect of the wind surge. Holding other effects constant, days 2 and U have the same expected discharge, the reason being

43 i+4

that both days have the same expected tide range. Table ? shows that this Is not the case for the October field trip. The highly significant differences between day 2 and day U In Table 5 are attributed to the wind surge effect. The discharges for day 3 (tropic tide) and day k are nearly equal even though tropic tide measurements were taken the

first day and day *+ measurements the second data collection day. Because Pass a Loutre's natural levees were Inundated at high tide on these two days the wind surge presented a strong Influence, The semi-monthly tidal effect can still be analyzed, with resei-vat ion, In this study. The wind surge caused flood-stage conditions in the delta. The expected variation In discharge caused by semi-monthly tide is less at flood stage than at low stage (Pig. 11). However, the orthogonal comparison of day 1 versus day 3 (equatorial versus tropic tide) Is highly significant, presenting strong evidence that semi-monthly tide Is significant during low stage. Under hlgh-stage conditions, the difference between these two days Is decreased. Although day 1 and day 2 were unaffected by the wind surge, the least-squares means show that equatorial tide has over three times greater discharge (515 cfs versus 1^5 cfs). Further evidence of a significant effect of the semi­ monthly tide is shown In Figure 11 and Table 1. Tidal range explains most of the variation (6?.0#)f which primarily re­ flects the semi-monthly tide. Evidence, therefore, supports 4 5 a highly significant semi-monthly tidal effect during low water. But the nonsignificant difference In the combined analyses (Table 5 ) Indicates that the semi-monthly tidal effect is restricted to low river stage.

Diurnal Tide

Both water discharge analyses (Table 5 ) show a signi­ ficant difference In discharge caused by the diurnal tide. Only one orthogonal comparison is highly significant, that of high and low tide versus flooding and ebbing tide. The least-squares means (Table 6) Indicate that the highest discharge takes ulace around high tide (724 cfs) and that lowest discharge occurs near low water (314 cfs),

Location

Location was Included In the analysis primarily for the Interactions. Since location represents the different pro­ files across the crevasse channel, the highly significant difference found In both analyses (Table 5 ) reflects channel bifurcation (Pig. 13)« Accordingly, the only orthogonal comparison with meaning Is locations B versus Ci channel 2 carries significantly more discharge. In October, their resDectlve mean discharges were 340 cfs for location B and 160 cfs for C. 46

Interact Ions

Stage by semi-monthly tidal Interaction Is significant.

During low river stage, the crevasse channel discharges less on day 2 than on day 1, But at high river stage, day 2 lncx*eases over day 1 (Table 6), reflecting the stronger semi-monthly tidal effect during low water. The nonsignifi­ cant stage by diurnal tidal Interaction Indicates that there Is no significant change In the diurnal tidal effect with changing river stage. Stage by location interaction shows channel 2 carrying 96t of the flow during flood stage and only ?1% at low stage (Table 6). The channel 2-channel 3 route carries most of the discharge of flood because this routing provides the most direct path to the opening In the spoil bank. Semi-monthly by diurnal tidal Interaction Is signifi­ cant In both the combined and the October analyses. The Interaction is accepted, with reservation, on the basis of the previous discussion concerning semi-monthly tide.

DISCHARGE PATTERNS, RATTLESNAKE CREVASSE

Discharge in Rattlesnake Crevasse follows two patterns, one dominant during low stage, the second during high stage. Between these two extremes there Is a gradation from one to the other. These patterns result from interactions between river stage, semi-monthly tide, diurnal tide, and Impeded local drainage. k7 The low stage pattern results from river stage, semi­ monthly tide, diurnal tide, and semi-monthly by diurnal tidal Interaction, The mechanism by which these variables control discharge In the crevasse channel Is changes In gradient between Pass a Loutre and Rattlesnake Pond. As the tide wave passes In the Gulf of Mexico, water In the delta altematlngly rises and falls. During flooding tide, water backs up In the delta In response to the higher sea level offshore, A gradient favorable for flow from T’ass a Loutre Into Rattlesnake Pond exists, as water ac­ cumulates In the channel more rapidly than In the lnter- dlstrlbutary bays. The extant gradient Is a function of the change In water level In Pass a Loutre and the Interdistri­ butary bay. The change In water level of Pass a Louti'e is greatest midway during flooding tide, resulting In the greatest slope and the highest discharge of the diurnal period. When the airadlent decreases around high tide, discharge ceases as Pass a Loutre and Rattlesnake Pond achieve the same level. Ebbing tide Is a period of flow from Rattlesnake Pond to Pass a Loutre, The surface of the foi*mer, with the lm- neded drainage, has a change In water level less than that of Pass a Loutre, Accordingly, some water* drains through both openings In the bay, and discharge Increases as the gradient Increases, Reversed discharge continues until low tide or until Pass a Loutre drops below the base of the crevasse-ehannel network. If the latter occurs, no discharge ^8

takes t>lace until Pass a Loutre again rises above the base of the solay. The maximum effect of the semi-monthly tide appears during low stage. The greater the tide range, the greater the discharge In both directions {Pigs. 11B and 11C). Reversal is absent during low stage only at equatorial tide. The second pattern Is that of flood stage, when the river stage dominates crevasse discharge. Flood stage In the delta represents water level Increase of 1.5 feet to 2.0 feet. The effect of diurnal tide remains, but the semi­ monthly tidal effect is overwhelmed. Flow through the crevasse channel shows a similar diurnal fluctuation, but reversal Is absent at highest stages. Discharge from Pass a Loutre begins Increasing during late ebbing tide, reaches a peak during flooding tide (Fig, 11A), and then rapidly decreases. Semi-monthly tide gradually becomes more Important as river stage drops. When reversal reappears, It first occurs late during ebbing tide. Water inundates much of the delta for over half the diurnal tidal cycle, flowing almost uncon­ fined to the Gulf. Later, toward low tide, the level drops to where drainage becomes more restricted to interdistri­ butary bay tidal channels, and reversal appears.

LITTLE LAKE POND CREVASSE CIRCULATION

The Increased dominance of river stage upriver Is shown by the data on Little Lake Pond Crevasse. The multiple 49 regression results (Table 3) Indicate that the three com­ ponents of river stage account for 76.2^ of the current velocity variation i 4-9,1# Is attributed to the quadratic effect of river stage alone. Tidal effects and interactions account for only 3,8%, Illustrating the increased dominance of the river. This crevasse maintains an average stage 0.5 foot higher than that of Rattlesnake (Fig. 5)* which Is suffi­ cient for river stage to dominate all year. Local drainas-e to the Gulf allows unidirectional flow. During flood, river stage overwhelms diurnal and semi-monthly tidal effects (Fig. 12C). At low water, currents show the characteristic tidal fluctuation.

CREVASSE DISCHARGE, MISSISSIPPI RIVER DELTA

Discharge in two crevasse channels has been shown to be the result of gradient changes initiated by the differ­ ential surface velocity of both the interdistributary bay and the distributary. If these two crevasses are represen­ tative of all Mississippi River Delta crevasses, then the following discussion explains the role of crevasses In balancing the discharge of the Mississippi River. Water surface gradients In delta distributaries in­ crease and decrease with the diurnal tide (U. S. Army Corps of Engineers, 1952, 1959* NEDICO, 1961* Santema, 1966)

Tison, 1966) Dronkers, 1969* and Wright, 1970). Figure 17, reproduced from Wright (1970), illustrates the case for the iue ? Current Profiles Mouth the at Passof South Figure on 1?. iue 8 Diurnal Discharge InVariation the ChannelofFigure 18, Discharge I in thousand* ] cfs a fi a. Iso c u «> 30 60 0 9 0 }n 2D 20 40 60 80 0 0 6 0 O C 4 0 0 0 3 0 0 0 4 2 0 0 3 2 0 0 0 2 1800 1600 1400 1200 I 0 0 October 1969(Modified1, Wright,from 1970). Englneers, 19 59). Southwest Pass (U. Mile at8 S.Corps ofArmy 30 4-0 30 2 0 0 1 M IE AM TIME PM -C c V 24 10 0 _ L. 1 0 3 0 2 0 0 4 50

1 U 3 0 2 0

Water level ( ft above msi ) 51 mouth of* South Pass at ebbing and flooding tide. A similar diagram, using discharge Instead of current profiles (Pig.

1R), shows the condition for Southwest Pass (U. S. Army Corps of Engineers, 1959).

Comparison of Figures 17 and 18 with Figures 8, 9, and 11 Indicates the Importance of crevasses to the hydrology of the Mississippi River Delta. During the low gradients accompanying flooding tide, the channels of South and South­ west passes drop to minimum dischargei the two crevasses attain maximum discharge. Then, during ebbing tide the pat­ tern reverses, This represents an attempt to balance the discharge of the river when seaward flow in the distributary channels is alternately enhanced and Inhibited. The discharge characteristics of Pass a Loutre and the various passes of the several active subdeltale crevasses are not known. Although these distributaries may also serve to balance the discharge of the river during the diurnal tide, certainly the many crevasses of the delta contribute to maintaining the discharge of the Mississippi River. Crevasse discharge will not be fully understood until the entire hydrology of the Mississippi River Delta is studied,

SUSPENDED LOAD TRANSPORT, RATTLESNAKE CREVASSE

Material Is Introduced Into the crevasse splay by the water discharge patterns discussed previously. Suspended load transport dominates sediment movement In a crevasse channel tapping only the upper layer of flow In a 52 distributary. In the vicinity of Rattlesnake Crevasse, uass a Loutre Is 85 feet deep; bed load transport is well

below the 4-8 feet of water tapped by Rattlesnake crevasse channel. Fisk (1952) and Linder (1952) reported that since fine materials disperse comparatively uniformly throughout the river cross section, diversion of sediment Into a crevasse channel Is approximately in direct proportion to the dis­ persion of water quantities. The vex*tlcal distribution of suspended load (Table 8) In Rattlesnake Crevasse lends cre­ dence to their findings.

Temperature

Because cold water Is more dense and more viscous than warm water, sediment picked up by cold-water flow is carried in suspension longer than sediment carried by warm-water flow (Carey, 19651 Colby and Scott, 19651 and Burke, 1966). In Rattlesnake Crevasse, the Inflowing water Is colder than water carried through the crevasse channel during discharge reversal. The sharp diurnal temperature contrast, as much as 5.3°C, favors deposition and retention of clastic sus­ pended sediment In the splay. During the low water period, the lowest diurnal water temperature appears during maximum discharge from Pass a Loutre (Table 12). This condition is Illustrated in Figure 19A, which shows a cold (dark) water mass overlying the crevasse splay. The cold Inflowing water mixes with and B

Figure 19* Thermal Scanner (Reconnofax IV) Imagery of Rattlesnake Crevasse Study Area. A. Early Flooding Tide. B. Flooding Tide Discharge Pattern Well Established. C. Ebbing Tide with Reversal and Warmer Plume of Water Ex­ tending into Pass a Loutre, displaces the warmer, ambient InterdIstributary bay water, resultin* In a nearly uniform water body over the splay. In Figure 19B Pass a Loutre water and Rattlesnake Pond water have nearly the same tone, reflecting the uniform water body During discharge reversal, water warmed by heat transfer and solar radiation occupies Rattlesnake Pond (Fig. 19C). The warmer water is shown as a lighter toned plume extending into Pass a Loutre from the orifice of the crevasse. The water of Pass a Loutre Is shown In Figure 19c as colder (darker) than the interdisti-lbutary bay water body. The water temperature observations in Table 12 substantiate the thermal scanner Imagery. In this study, water temperature varied less between flood stage and low stage than during the diumal tide at low stage (Table 12). This is not the case for the entix-e hydrologic year (Burke, 1966).

Discharge

Results Indicate that discharge Is significant in deter mining suspended load concentration. The absence of a sig­ nificant value at low stage reveals that the effect results from the increased discharge occurring between low and high stages,

River Stage

The effect of river stage on suspended load concen­ tration is well established. In the case of Rattlesnake Crevasse, suspended load concentration Is ovex* 5 times greater dux'lng flood than during low water. Even more Im­ portant, suspended load transport during flood Is over 2 5 times greater than that during low water. Data collected Indicate that for day 2 of the semi-monthly tide, 2,18 tons of sediment was Introduced Into the crevasse during flood stage and 0,18 tons during low river stage. The latter value reflects reversed discharge, low total discharge, and low suspended load concentration (Fig, 20), Obviously, not all of the sediment Is retained by the crevasse splayi some Is lost through the outlet in the spoil bank (Figs, 8 and 9) Comparison of suspended load concentrations at the crevasse orifice (profile A) and the outlet for low-water tropic and equatorial tides Indicates that **-5-50# of the suspended material introduced Into the crevasse Is retained by the

o> £ 300 "O — _ „ tu.ii. 1, 19 7 0 o o — 1 Oiiobf j 31, 19 7 0 T) 200 4) T> C 4) a vt3 tOC CO

_L. I i>. td. flood, ft ^ T j d t High T i d o t b 111 'i 'j lid# TIME Figure 20, Comparison of Suspended Load Concenti'atlon at Profile A for Flood and Low River Stages (Day 2) 06 splay {Fig. 8, and Table 12). Microscopic examination of water samples from both sites Indicates that virtually all

sand and nearly all silt is retained by the splay.

Semi-monthly Tide

The wind surge present during the low stage data col­ lection period produces uncertain results. The decrease in suspended load undoubtedly accounts for a large portion of the variation (Table 12). Unlike the discharge analyses, the wind surge cannot be Isolated as a measured source of variation. Days J and k were the period of wind surge in­ fluence. Table 12 indicates that day 3 has a very high suspended load* the other days have little absolute dif­ ference. Likewise, the orthogonal comparisons are not ac­ ceptable as representlng components of the semi-monthly source of variation. The principal conclusion from this test Is that a wind surge has a highly significant effect on the concentration of suspended load. That this is sig­ nificant is shown by the orthogonal comparison of day 2 versus day 4.

Diurnal Tide

Even after the analysis of covariance, the results In­ dicate a significant variation in suspended load with di­ urnal tide. Diurnal reversal of discharge explains only nart of the variation. Even on days when wind surge and reversal did not occur, suspended load variation was still 57 evident, It Is apparently related to discharge fluctuations

In lass a Loutre i Wright (1970) found a similar variation at the mouth of South Pass In the presence of salt-water intrusion. Suspended load variations In Pass a Loutre a n ­ neal* to have a similar diurnal pattern.

Location

Location was nonslgriifleant In both analyses, Indi­ cating essentially no erosion 01* deposition between the pro­ files. This result reflects the form and composition of the channel bed between the orifice and the middle ground.

During flood stage, the bed was found to be a cleanly scoured surface almost as far as the middle ground. From that point to profiles B and C the bed was composed of a thin layer of fine sand and slit overlying the eroslonal surface. During low water, the sand and silt were found slightly closer to the orifice.

Interact Ions

Of the Interactions presented In Table 9# only those not Involving the semi-monthly tide can be accepted as valid because of the wind surge. Of those, stage by diurnal tide is highly significant. 58 SEDIMENT ACCRETION, RATTLESNAKE CREVASSE

Sediment accretion in the crevasse splay results from the pattern of sediment discharge and the processes which act to redistribute the sediment in the crevasse splay. This asnect of the study centered on the pattern of sediment discharge and Its relation to sediment accretion. During the first three measuring periods only one sig­ nificant value was found (Table 13). Table 14 shows that channel 1 had net deposition, whereas the other channels had net loss, of sediment. The amounts of loss and gain were not great. The grand mean indicates that 0,05 foot of ac­ cretion took place in the Interval between February 1 and

April 5. 1970. The period between April and June, which constitutes the highest portion of flood, was marked by deposition of a tremendous quantity of sediment {Fig, 16D), The channel effect was highly significant, reflecting channel 3 as the preferred channel (Table 1^). Channel 3 is shown by the orthogonal comparisons and in Table 1^ to be dominant. Deposition along channel 3 exceeded that of the other chan­ nels by a factor of 2i or, O .36 foot of net accretion was distributed over the splay, Channel was again significant in October, after the river had fallen from high to low stage, Channel 1 con­ tinued to accrete during this period. The difference 59

between profiles 3 and 4 on tne Individual channels accounts for the significant channel by profile Interaction (Tables 13 and 14). In the evaluation of annual accretion, no significant differences are found. The entire splay Is characterized by rapid upward growth, and there Is little significant areal variation. During the hydrologic year, accretion is concentrated during flood stage, and only small amounts take place during the rest of the hydrologic year (Table 14). The success of the sediment accretion phase of the In­ vestigation was below expectations because unforeseen prob­ lems developed. Of the 46 reference points Installed, only 19 remained at the end of the 1970 hydrologic year. Float­ ing debris, especially logs, broke off rods, making relo­ cation difficult and usually impossible. Unanticipated horizontal growth of the splay caused a second serious problem, BasInward extension of the defined channels docu­

ments horizontal growth of the splay. In November, 1 9 6 9 ,

profiles 3 and 4 on channel 2 and profile 4 on channel 1 were emplaced beyond the end of the defined channels. By November 1, 1970, these two channels had progi'aded approxi­ mately 400 feet and 200 feet respectively. The result of the basInward extension was that the accretion grid did not adequately cover the splay by the end of the hydrologic year. The annual accretion lsopach map (Fig. 16f) reveals that accretion Is progressively thicker basInward, It Is possible that the zone of maximum accretion was not measured, 60 although, for reasons given below, this may not be the case.

In general, the areas receiving greatest deposition during flood underwent the greatest loss of thickness be­ tween June and October. This suggests that dehydration and compaction may well be more Important than erosion and re­ distribution of sediment. This does not preclude the lat­ ter two nrocesses from being important. The writer ob­ served three annual cycles of deposition and consolldation of sediment In Rattlesnake Crevasse, As the river drops from peak flood, the splay is altematingly Inundated and exposed by the diurnal tide. Each year the newly deposited material is progressively dehydrated and compacted as a result of the diurnal exposure, Evidence for this Is pro­ vided by field observations, With the lone exception of equatorial tides, during low water the splay Is exposed as far basInward as the 2,175-foot profile. During this period, a crust covered by an algal mat develops and becomes firmer and thlckei' as low river stage progresses.

This suggests that the zone of maximum accretion may well have been measured. But, because of dehydration and compaction on the higher portions of the crevasse splay, maximum thickness may appear to shift basInward, where con­ solidation Is less. Accordingly, a value of total volume of new material can be calculated, subject to these limi­ tations . The vertical growth, according to the grand mean In

'"able 15, is 0,*+2 foot ± 0.08 foot. The sampling grid, M however, weighted the mean toward the areas of greatest deposition. By weighting the means of the profiles with

the distance between them, net accretion becomes 0.35 foot for the year. During the 1970 hydrologic year, 1.061 X 10° cubic feet of sediment was deposited within the area of the ^amoline- grid. The total area was determined by plani- metering the surveyed map of Rattlesnake Crevasse, While this value is not as accurate as desired, it most likely represents an underestimatlon of the total volume inasmuch as the entire splay was not represented in the sampling pattern.

'a?T MD

Wind modifies water and sediment discharge patterns in Rattlesnake Crevasse two ways. The first is the wind surge, which represents a temporary reproduction of flood stage conditions in the delta, causing levee overtopping, increas­ ing suspended load transport, and modifying the discharge nattem (Fig, 21), The importance of the wind surge to crevasse accretion should not be overemphasized. Under the high stage water and sediment discharge pattern described above, 0,175 foot of sediment accumulated on the left levee of the crevasse at the survey pipe during the 1970 hydrologic year. This accretion resulted from Inundation of the levee for no less than 50 consecutive days during flood. When the wind surge was present in October, no observable, nor measurable h2

I 000

o>

-800 l--- 1______1 low nrf* Hooding Tidn High Tido tbbing T i d * TIME

Figure 21. Effect of Wind Surge on Low Stage Discharge Pattern (Tropic Tide),

accretion took place In the crevasse. Comparison of tide

ime-e and climatic data (U. S. D. C., 19691 1970a, 1970b) reveals that durlno: the 11-month Derlod when the tide gage functioned on Rattlesnake Crevasse, wind surges occurred five times for a total of nine days. To cause significant accretion, wind surges would have to be much more common. The second manner In which wind Influences crevasse processes Is by disturbing the circulation In the crevasse channels. Figure 22B Indicates the effect of a strong, per­ sistent wind blowing directly up channel 2, Surface flow was nil on four of the five lines, producing the character­ istic disturbed current velocity profile. Figure 22A pre­ sents the same channel when wind Is not a factor. The disturbed pattern, coupled with wind-generated waves. 63 to/i 1 //o ? tso

A A jimullii lo i (uricnli on each line of Profile B

N N N N N

* 5 S ------* I - ■ ~ * i no wino

I0/I///0 1600

B A iim u lh i lor ru n e n li on Both line ol Profile B

N N N N N

W ' NiD w —- e A J m p h

W ater Surface

Figure 22. Effect of Wind on Current Patterns In Rattle­ snake Crevasse Channel. A. No Wind. B. Wind Present. 64 red 1stributes sediment within the splay. This is confirmed hy the frequency of starved rlonles observed on the crevasse snlay and by Emmer (1968), SUMMARY AND CONCLUSIONS

The function of the two delta crevasses Investigated

In this reoort Is to balance Mississippi River Delta hydro­ logy, Py providing overflow channels during the diurnal fluctuation of water level and discharge In the distribu­ taries, these crevasses allow large quantities of water to be stored In their Interdistributary bays. They function, therefore, to provide overflow channels as much during the diurnal tide as during flood stage,

With the exception of water temperature, the variables evaluated and explained In this investigatlon Interact to cause gradient changes between the distributary and the ad­ jacent Interdistributary bay. Plow through the crevasse channel results from the gradient determined by river stage, diurnal tide, semi-monthly tide, wind, and local drainage cond1tIons,

Local drainage conditions provide the framework* the quantity of water that can be stored or discharged to the

Culf, Little Lake Pond Crevasse, where drainage Is uncon- flned, breaches Raphael Pass' natural levee. In Rattle­ snake Crevasse, where drainage is impeded, the discharge pattern depends upon the degree of obstruction. Reversal occurs when Pass a Loutre Is subject to rapid water level changes, changes too rapid for the Interdistributary bay to

6 5 66

equal. Consequently, Impounded water drains back Into the

dIstributary•

Within the framework provided by the local drainage

natterns, river stage dominates the flow of water through

these crevasses throughout the hydrologic year. This Is

evident from the discharge oatterns of the two crevasses

during flood stage and the low stage discharge oattern at

equatorial tide. During flood stage, flow through Rattle­

snake Crevasse Is unidirectional.

The changing tidal level to which the river must adjust

-rives rise to diurnal gradient changes In the delta. Water

cannot escape to the Gulf j instead, It fills the inter-

distributary bays through crevasse channels. in this manner,

the diurnal discharge patterns of the distributary and these

two crevasses are comolementary.

Diurnal tide gives rise to a daily pattern of discharge

In the crevasses, Maximum discharge into the interdlstrl-

-utary bay occurs during flooding tide, when seaward flow

Is not favored in the distributary. Minimum crevasse dis­

charge occurs during ebbing tide, taking the form of re­

versed flow (negative) or low oosltlve velocity. This oat­

tern characterizes Rattlesnake Crevasse even during flood

but Is absent at high stage at Little Lake Pond Crevasse,

The semi-monthly tidal effect causes significant

changes in discharge. The change arises from the day-to- day tidal range variation. Tropic tide is the period of

maximum gradient change and maximum change In water level of the dlstrlbutary-lnterdlstrlbutary bay surfaces. Pro­

gressing1 from equatorial tide to tropic tide, the maximum

diurnal gradient between the distributary and Interdlstri­

butary hay becomes steeper; larger discharge and longer

duration of discharge reversal (when oresent) results. At

low stage, these latter results are frequently balanced by

cessatlon of dIscharge for several hours around low tide,

Wind Is imnortant to crevasse evolution pi* 1 marl ly In

Its Influence on the material Introduced Into the crevasse

snlay. Wind surges cause flood stage conditions during low

river stage, overtopping levees and Increasing the trans-

oorted load. In view of the results of this Investigation,

it aDDears that In previous literature the Importance of

wind surge effects to delta processes has been as much over­

emphasized as lunar-tlde effects have been neglected. Wind

surges, unless driven by a tropical storm, appear to have

no lasting measurable effect. Far more lmoortant to sedi­

ment accretion Is the 2 5 -tlmes-greater sediment discharge

of flood over low stage. Discharge, both water and sediment

is determined to a much greater extent by the periodic lunar

tide and river stage effects.

Data for these two crevasses Indicate that the Inten­

sity of the above variables changes with proximity to the

distributary mouth. If the two crevasses are representative of delta crevasses, then upstream the river dominates com­

pletely; the form resulting Is the alluvial valley crevasse.

Farther downstream, tide and wind effects become significant f'fi

The closer the proximity to the mouth of the river, the greater the tidal and wind effects. Ultimately, In the vicinity of the distributary mouth, marine processes become Important, and another set of variables governs crevasses. The water discharge pattern exerts a strong effect on sediment transport and crevasse-splay accretion. The sedi­ ment Introduced Into the crevasse Is primarily suspended load of which ^5-50^ is retained by the crevasse during low water. ^ost sediment Is Introduced during flood stage. In Rattlesnake Crevasse, diurnal sediment discharge at flood Is over 25 times greater than that at low water. This value reflects higher water discharge and higher suspended load concentration at flood stage. It is not surprising, then, that sediment accretion In the crevasse splay Is concen­ trated during peak flood. Suspended load fluctuates during the diurnal tide In crevasse channels In the Mississippi River Delta. The variation extends primarily from the diurnal water tempera­ ture variation. In addition, seemingly present Is a fluc­ tuation In suspended load concentration In the water Intro­ duced from Pass a Loutre. Sediment accretion Is rapid In the Rattlesnake crevasse splay, exceeding 0.35 foot of upward growth during the 1970 hydrologic year for a total of 1,061 x 10^ cubic feet. Within the crevasse splay, maximum sediment accretion during flood stage prevails near the end of the defined channels, on the low, submerged natural levees. During low water, 69 this zone Is alternately Inundated and exposed, leading; to compaction and consolidation. Along with sediment redistri­ bution, this produces a pattern of maximum net annual ac­ cretion slightly basInward, nearer the end of the channels.

Sediment accretion analyses were limited by the method of measuring sediment accretion. Metal accretion rods have the advantage of providing precise, relocatable reference points but are easily broken off by floating debris. Some modification or new method for measuring sediment accretion should be employed In future studies.

This investlgatIon has collected and analyzed detailed measurements of the variables controlling the evolution of

Rattlesnake Crevasse and has successfully demonstrated that analyses of variance and multiple regression procedures can be applied to Identify and evaluate the significant vari­ ables In a delta crevasse. In addition, the methods estab­ lish, through comparison of discharge patterns of Rattle­ snake and Little Lake Pond crevasses, that such an approach provides a useful means of analyzing the relative Intensity of the variables affecting the two crevasses. The next step Is to apply these methods to other delta crevasses to determine the relationship of crevasses to Mississippi River

Delta morphology. REFERENCES nondar, C., 1970, Theoretical study on the spreading of a light-liquid current In a basin with a heavier liquid. In Hydrology of deltas. UNESCO, pp. 246-257. noi'ichansky, L. S., and Mikhailov, V. N.f 1966, Interactions of river and sea water In the absence of tides. Li Scientific problems of the humid tropical zone deltas and their Implications, UNESCO, pp. 175-180.

Burke, P., 1966, Effect of water temperature on discharge and bed configuration, Mississippi River at Red River Landing, Louisiana. U. S. Army Corps of Engineers, Committee on channel stabilization. Tech. Rept. No. 3* 25 pp.

Carey, W. C., 1965, Effect of temperature on riverbed configuration. In Proceedings of the Federal Inter- Agency sedimentaFTon conference, pp. 237-271.

Chow, V. T.f 1959* Open-channel hydraulics. New York (McGraw-Hill Book Co., Inc.), 680 pp.

Cochran, W. G.# and Cox, G. M., 1957. Experimental designs. New York (John Wiley A Sons, Inc.), 617 p p .

Colby, P. R. and Scott, C. H. , 19&5. Effects of water temperature on the discharge of bed material. U. S. G, S. Prof. Paper No. 462-G, 2 5 PP.

Coleman, J. M., Gagllano, S. M., and Morgan, J. P., 1969, Mississippi River Subdeltas 1 natural models of deltaic sedimentation. Coastal Studies Bulletin No. 3, pp. 2 3-2 7. Dont, E. J., 1924, The mouths of the Mississippi River. Trans, A. S. C. E., 87i pp. 997-1006. Dronkers, J. J., 1969, Tidal computations for rivers, coastal areas and seas. Proc. A. S. C. E., J. Hydraulics Div., 95t634l, pp. 29-77. Elliott, D. 0., 1932, The improvement of the Lower Mississippi River for flood control and navigation, u. S. Waterways Experiment station, 3 vol.

70 71 Emmer, R. E., 1968, Crevasses of the Lower Mississippi River Delta, Unpublished Masters thesis, Louisiana State University, 54 pp, Fisk, H, N.t 1952, Geological investigations of the Atchafalaya Basin and the problem of Mississippi River diversion, U, S, Army Corps of Engineers, Waterways Experiment Station, 2 vol.

Fisk, H. N.f MeFarlan, E., Jr., Kolb, C. R,, and Wilbert, L, J., 195i+» Sedimentary framework of the modern Mississippi delta, J, Sed, Pet,, 241 pp, 76 -99. Gagllano, S, M, , and van Beek, J, L. , 1970, Geologic and geomorphic aspects of deltaic processes, Mississippi delta system. Part I, vol. 1 of Hydrologic and geologic studies of coastal Louisiana. Louisiana State University Press, 140 pp, Gilbert, G. K., 1884, The topographical features of lake shores, Ann. Rept. U. S. Geol. Surv. 5 . PP. 104-108. Goncharov, V, N., 1962, Dynamics of channel flow. (Trans 1, from Russian), U. S. Dept, of Commerce 0T5 64-11003, 317 pn. Hald, A., 1952, Statistical tables and formulas. New York (John Wiley & Sons, Inc.), 9? pp.

Had p , S, C., 1948, Sedimentation In the middle Rio Grande Valley, New Mexico. G. S. A. Bulletin, 59* p p . 1191- 1216. ______, 1944, Significance of texture and density of alluvial deposits in the middle Rio Grande Valley. J. Sed. Pet., l4il pp. 3-19*

Holle, C. C., 1951, Sedimentation at the mouth of the Mississippi River. Proc. Second Conf, on Coastal Engln,, Council on Wave Res., Pt. 2, pp. 111-129.

Johnson, L. C., I89 I, The Nita Crevasse. G. S. A. Bulletin, 2*1, pp. 20-25.

Kondrat'ev, N. E. , 1959, River flow and l'lver channel for­ mation. (Transl. from Russian) U. S. D. C., 0TS-61- 114 32, 172pp.

Krumbeln, W. C., and Graybill, F. A., 1965, An introduction to statistical models in geology. New York (McGraw- Hill), 475 PP. 72 II, J. C. R., 1964, Statistical inference I. Ann Arbor, Michigan (Edwards Brothers, Inc. ), 658 pp. Linder, C, P., 1952, Diversions from alluvial streams, Trans. A, S. C. E.f 118* pp. 245-288. Mariner, H. A., 19 54, Tides and sea level in the Gulf of Mexico. U. S. D. I., Fish and Wildlife Serv. Bull., 8 9 * nn. 101-118. M\DTC0, 1961, The waters of the Niger Delta. The Hague (Netherlands Engineering Consultants), 317 PP* :fussell, R. J. , 1954, Alluvial morphology of Anatolian rivers. Annals, A. A. G., 44* pp. 363-391* ______, 1936, Physiography of the Lower Mississippi River Delta. 2H Lower Mississippi Delta* Reports on the geology of Plaquemines and St. Bernard Parishes. Louisiana Dept. Cons. Geol. Bull. No. 8, pp, 3-199. ______, et^ al,, 1938, Reports on the geology of Iberville and Ascension Parishes. Louisiana Dept. Cons. Geol. Bull. No. 13. 223 pp.

Santema, P., 1966, The effect of tides, coastal currents, waves and storm surges on the natural conditions pre­ vailing in deltas. In Scientific problems of the humid tropical zone deltas and their implications. UNESCO, pp. 109-114. Scruton, p. C., 1956, Oceanography of Mississippi Delta sedimentary environments. Bull. A. A. P. G., 40il2, p p . 2864-2952. Gnedecor, G. W., and Cochran, W. G., 1967, Statistical methods, 6th ed,, Ames, Iowa (The Iowa State Univ. Press), 593 pp. Strickland, J. D. H., and Parsons, T. R., 1965, A manual of sea water analyses. Fisheries Res. Board of Canada, Bull. No. 125* Tlson, L. J., 1966, The problem of sedlmentation in deltas and their estuaries. (In French) In Scientific Prob­ lems of the humid tropical zone deltas and their im­ plications, pp. 41-58, Tricart, J., 1956, Aspects geomorohologlques du delta du Senegal. Revue de geomorphologle dynamlque. No. 5-6, pp. §5-86. 71

Trlcart, J,, 1955. Aspects sedlmentologiques du delta du Senegal. Geol. Rundschau, 3*4-1 Dp. 38^-397.

United States Army Corps of Engineers, 1959, Investigations and data collection for model study of Southwest Pass, Mississippi River. U. S. A. C. E., New Orleans Dis­ trict, 2 vol.

______, 1952, Report of Improvements of the passes of the Mississippi River. Unpublished. Plate 26 In 6 sheets.

______, 1939, Study of materials in suspension, Mississippi River, Waterways Experiment Station, Tech. Memo 122- 1 . 27 pp. United States Department of Commerce, ESSA, 1969* Tide Tables, East coast of North and South Amei'ica. Washington D. C.

______, NOAA, EDS, 1969, Local cllmatologlcal data, New Orleans (Molsant), La. Asheville, N, C.

______, NOAA, EDS, 19?0a, Local Cllmatologlcal data. New Orleans (Molsant), La. Asheville, N, C.

______, NOAA, EDS, 1970b, Dally weather maps, weekly series. Washington D. C. Van Der Made, J. W., 1970, Design levels in the transition zone between the tidal and the river regime reach. In Hydrology of deltas. UNESCO, pp. 2 57-272.

Welder, F. A., 1955. Processes of deltaic sedimentation in the Lower Mississippi River. Unpublished Ph. D. Dissertation, L. S. U., 119 pp.

______, 1959, Processes of deltaic sedimentation In the Lower Mississippi River. Coastal Studies Inst. Tech. Rent. No. 12, 90 pp. (Abbreviated form). Wright, L. D., 1970, Circulation, effluent diffusion, and sediment transport, mouth of South Pass, Mississippi River Delta. Coastal Studies Inst, Tech. Rept. No. 8 U, 56 pp. . and Coleman, J. M., in press. Effluent expansion ancT lnterfaclal mixing in the presence of a salt wedge, Mississippi River Delta. 21 pp. VITA

David James Arndorfer* was born In Portland, Oregon, July 17, 19^3. He received his elementary education In Portland and his secondary education in Portland and St. Benedict, Oregon, Mr. Arndorfer majored in Geography at Portland State College where he received the Bachelor of

Science In June, 1967 . In September, 1967, he entered graduate school at Louisiana State University, Baton Houge, under an NDEA-IV fellowship. Mr. Arndorfer majored in Geography and minored In Marine Science, He married Kathryn Dewar on December 31, 1966. They have one child, Heidi, bora December 7» 1970, EXAMINATION AND THESIS RETORT

candidate: David James Arndorfer

Major Field: Geography

Title of Thesis: Process and Parameter Interaction in Rattlesnake Crevasse, Mississippi River Delta

A p p ro ved :

M a jiir I*i (i[t‘s.'-iir ami C h .m m a n

------/* 1. L.- —— 1 lean nfihi 1 (ir .kIUlItf Scluxil

KXAMINIM; C< tMMl'i'TKE