Louisiana State University LSU Digital Commons

LSU Historical Dissertations and Theses Graduate School

1994 Crust Formation in of Mexico and Louisiana. Miguel Angel Martinez-gamino Louisiana State University and Agricultural & Mechanical College

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

Recommended Citation Martinez-gamino, Miguel Angel, "Crust Formation in Soils of Mexico and Louisiana." (1994). LSU Historical Dissertations and Theses. 5888. https://digitalcommons.lsu.edu/gradschool_disstheses/5888

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]. INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely afreet reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 313/ 761-4700 800/ 521-0600

CRUST FORMATION IN SOILS OF MEXICO AND LOUISIANA

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 Agronomy

by Miguel Angel Martinez-Gamino B.S., Autonomous University of San Luis Potosi, 1978 M.S., Postgraduate College, Montecillos, 1985 December 1994 UMI Number: 9524468

UMI Microform Edition 9524468 Copyright 1995, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 ACKNOWLEDGEMENTS

I would like to extend my heartfelt gratitude to my Major Professor,

Dr. Paul Mark Walthall, for his excellent supervision, patience, and friendship during my studies at Louisiana State University.

I give my sincere gratitude to Dr. Ray E. Ferrell, Jr., for giving me his time and the opportunity to use the scanning electron microscope laboratory.

I thank my other committee members: Dr Wayne H. Hudnall, Dr.

Hussein M. Selim, Dr. Donald Robinson, and Dr. Richard L. Bengston for their assistance in reviewing and editing my manuscript.

Specials thanks goes to Will Day, Gina Maciasz, Rick, Patrick, Robin

Migue, and Louis for their help with laboratory analyses.

I am grateful for the financial support provided by the Mexican government, CONACYT, INIFAP, Fulbrigth-IIE, Louisiana State University, and the Agronomy Department at L.S.U.

Sincere appreciation goes to my colleagues in the National Institute of Agricultural Research in Pabellon, Aguascalientes, M. Sc. Salvador

Martin del Campo and M. Sc. Salvador Osuna for their help in sampling the

Mexican soils used in this study.

ii I send a very special thanks to Laura, for her love, inspiration, and help, not only in the last four years but also for the 15 years we have been married.

Thanks, to my sons; Daniel Ixtetzin, Itzen-Kin Alejandro, and Miguel

Itzamna for all the moments we have shared. Go for it kids, you too

Samuel.

Thanks, to my mother Ma. del Socorro and Porfirio, for their love, care, and understanding.

To my friends at Louisiana State University: Joycelyn, Andrew,

Nick, Monica and all the Mexican and Latin students. TABLE OF CONTENTS

Page ACKNOWLEDGMENTS...... ii

LIST OF TABLES ...... viii

LIST OF FIGURES...... xi

ABSTRACT...... xvi

CHAPTER 1 INTRODUCTION AND REVIEW OF LITERATURE Introduction...... 1 Review of Literature ...... 3 Physical dispersion...... 3 Rainfall characteristics...... 3 texture ...... 4 Aggregate stability...... 5 Chemical dispersion ...... 5 mineralogy...... 6 Exchangeable ions and electrolyte concentration...... 6 Segregation of soil particles ...... 8 Drying period ...... 10 Cementing agents ...... 10 Amelioration of soils with crusting problems...... 11 Soil crust modeling...... 12 References for Chapter 1...... 15

2 PHYSICAL, CHEMICAL, AND MINERALOGICAL PROPERTIES AFFECTING SOIL CRUST FORMATION IN SOILS OF MEXICO AND LOUISIANA Introduction...... 19 Materials and Methods ...... 20 Soil s e rie s...... 20 Physical characteristics...... 24 Particle-size distribution ...... 24 content ...... 24 Chemical characteristics...... 24 Soil reaction...... 24 Organic carbon ...... 24 G ypsum ...... 25 Page Calcium carbonate ...... 25 Amorphous silica and aluminum ...... 25 Free iron oxides ...... 25 Saturation extract...... 25 Electrical co n d u ctiv ity...... 25 Water-dispersible clay index ...... 26 Sodium adsorption ratio...... 26 Mineral characteristics ...... 26 Pre-treatments for clay mineralogy...... 26 Clay mineral identification ...... 27 Random-powder diffractograms ...... 27 Data quality ...... 29 Discussion of Results ...... 29 Physical characteristics...... 29 Particle size distribution ...... 29 Water-dispersible clay index ...... 31 Chemical characteristics...... 32 Potential cementing agents ...... 32 Dispersion factors...... 35 Mineral characteristics ...... 38 Clay mineral co m p o sitio...... n 38 Powder X-ray diffraction ...... 40 Summary and Conclusions...... 40 References for Chapter 2 ...... 43

3 MORPHOLOGICAL CHANGES DURING CRUST FORMATION IN SOILS OF MEXICO AND LOUISIANA Introduction...... 46 Materials and Methods ...... 47 Soil s e r ie s...... 47 Soil p rep aratio n...... 48 Bulk density ...... 48 Rainfall sim ulator...... 48 Drying period ...... 49 Management practices ...... 49 Sampling zone...... 49 Thin section preparation...... 49 Scanning electron microscope and XRF analyses . . 50 Discussion of results...... 51 Ginger soil c r u s ts...... 52 Asooueros soil c r u s t...... s 61 Sandovales soil crusts ...... 65

v Page Cementing agents ...... 72 Summary and Conclusions...... 78 References for Chapter 3 ...... 87

4 GEOCHEMISTRY OF CEMENTING AGENTS IN CRUSTS FROM SOILS OF MEXICO AND LOUISIANA Introduction...... 90 Materials and Methods ...... 91 Soil s e rie s...... 91 Soil p reparation...... 92 Rainfall sim ulator...... 92 Drying period ...... 92 Management practices ...... 93 Sampling zone...... 93 Experimental design ...... 93 Chemical analysis ...... 93 Random-powder diffractograms...... 94 Simulation of solution-mineral eouilibrium using GEOCHEM ...... 95 Discussion of Results ...... 96 Free Fe oxides and amorphous Al and S i ...... 96 G y p s u m...... 99 Calcium carbonates ...... 100 Powder X-rav analysis ...... 101 Saturated paste extracts ...... 110 Gypsum and calcite precipitation m odels...... 117 Simulated concentration of the soil solution 122 Summary and Conclusions...... 129 References for Chapter 4 ...... 132

5 EFFECT OF MANAGEMENT PRACTICES AND RAINFALL DURATION IN SOIL PROPERTIES AS THEY AFFECT SEEDLING EMERGENCE OF CORN (Zea mays L.) AND SOYBEANS (Glycine max (L.) Merr) Introduction...... 134 Materials and Methods ...... 135 Soil s e rie s...... 135 Soil preparation...... 136 Rainfall sim ulator...... 136 Runoff measurement ...... 137 Drying period ...... 137 Crust strength measurement...... 137 Management practices ...... 139

vi Page Experimental design ...... 139 Discussion of Results ...... 140 rate ...... 140 Soil ero sio n...... 145 Crust strength ...... 149 Corn and soybean emergence ...... 158 Summary and Conclusions...... 164 References for Chapter 5 ...... 169

6 SUMMARY AND CONCLUSIONS ...... 172

APPENDIX

A RAINFALL SIMULATOR...... 180

VITA ...... 183

vii LIST OF TABLES

Page

2.1 Characteristic X-ray diffraction peaks for minerals in clay fractions...... 28

2.2 Particle size distribution for dispersing agent (DA) and distilled water (DW) and water dispersible clay index (WDCI) ...... 30

2.3 Quantification of possible cementing agents in soils of Mexico and Louisiana ...... 33

2.4 Cation and anion concentrations in saturated paste extracts of the soils of Mexico and Louisiana...... 36

2.5 Chemical soil properties of soils of Mexico and Louisiana ...... 37

2.6 Clay mineral composition of the soils of Mexico and Louisiana ...... 39

2.7 Minerals identified using powder X-ray diffraction in soils of Mexico and Louisiana ...... 41

3.1 Effect of management practices on in soils of Mexico and Louisiana ...... 56

3.2 Particle-size distribution of sediment loss in the soils of Mexico and Louisiana ...... 58

3.3 Water-dispersible clay index in the soils of Mexico and Louisiana ...... 62

4.1 Potential cementing agents present in the crust and in the soils of Louisiana...... 97

4.2 Potential cementing agents present in the crust and subsurface in the soils of M exico...... 98

4.3 Calcite estimates using three methods in the soils of Louisiana ...... 102

viii Page 4.4 Calcite estimates using three methods in the soils of M ex ico...... 104

4.5 Gypsum estimates using three methods in the soils of Louisiana ...... 106

4.6 Gypsum estimates using three methods in the soils of M ex ico...... 108

4.7 Cation concentrations from saturated pastes in the soils of Louisiana ...... 111

4.8 Cation concentrations from saturated pastes in the soils of Mexico ...... 112

4.9 Anion concentrations from saturated pastes in the soils of Louisiana ...... 115

4.10 Anion concentrations from saturated pastes in the soils of Mexico ...... 116

4.11 Calcium and sulfate activities in solution and sulfate activity needed to precipitate gypsum in the soils of Louisiana ...... 120

4.12 Calcium and sulfate activities in solution and sulfate activity needed to precipitate gypsum in the soils of M ex ico...... 121

5.1 Selected physical and chemical properties in the soils of Mexico and Louisiana...... 142

5.2 Particle-size distribution of sediment loss in the soils of Mexico and Louisiana ...... 146

5.3 Effect of management practices on erosion in the soils of Mexico and Louisiana...... 147

5.4 Effect of management practices and rainfall duration on the crust strength at the seventh day after rainfall simulation in the soils of Louisiana...... 152

ix Page 5.5 Effect of management practices and rainfall duration in the crust strength at the seventh day after rainfall simulation in the soils of M e x ic o...... 153

5.6 Stepwise multiple regression of factors influencing crust strength ...... 157

5.7 Effect of management practices and rainfall duration in corn and soybean emergence in the soils of Mexico...... 160

5.8 Effect of management practices and rainfall duration in corn and soybean emergence in the soils of Louisiana 161

5.9 Stepwise multiple regression of factors influencing corn emergence...... 162

5.10 Stepwise multiple regression of factors influencing soybean emergence...... 165

x LIST OF FIGURES

Page

2.1 Site locations of the soils of Mexico and Louisiana...... 22

3.1. Micrograph of a vertical thin section with plain light showing a cratered surface (A) caused by raindrop impact in the bare Gigger soil at ponding t i m e...... 53

3.2. An SEM micrograph, vertical section, showing the homogeneous distribution of particles in the surface (A) and subsurface (B) in the bare Gigger soil at ponding tim e ...... 53

3.3. An SEM micrograph, vertical section, showing a compacted 0.1-mm surface layer (A) formed of single, clean grains in the bare Gigger soil at the end of the drying period ...... 55

3.4. Micrograph of vertical thin section with plain light, showing a flat-undisturbed soil surface (A) in the protected treatment of the Gigger soil at ponding time ...... 55

3.5. An SEM micrograph, vertical section, showing soil aggregates in the surface (A) of the protected Gigger soil at ponding time . . 57

3.6. An SEM micrograph, vertical section, showing the homogeneous particle size distribution in the surface (A) and subsurface (B) in the protected Gigger soil at the end of the drying period .... 57

3.7. Micrograph of a vertical thin section with plain light, showing the effect of raindrop impact in the surface (A) of the gypsum- amended treatment of the Gigger soil at ponding time ...... 60

3.8. An SEM micrograph, vertical section, showing the aggregation effect (A) of gypsum in the gypsum-amended treatment of the Gigger soil at ponding t i m e...... 60

3.9. An SEM micrograph, vertical section, showing an upper 0.1 mm layer surface layer (A) formed of single, clean grains in the gypsum-amended Gigger soil at the end of the drying period . . . 63

3.10. Micrograph of a vertical thin section with plain light, showing a cratered surface (A) caused by raindrop impact in the bare Asogueros soil at ponding time ...... 63

xi Page 3.11. An SEM micrograph, vertical section, showing a homogeneous particle-size distribution in the surface (A) and subsurface (B) in the bare Asogueros soil atponding tim e 64

3.12. Micrograph of a vertical thin section with plain light, showing an unaltered surface (A) in the protected Asogueros soil at ponding t i m e...... 64

3.13. An SEM micrograph, vertical section, showing a homogeneous particle-size distribution in the surface (A) and Subsurface (B) in the protected Asogueros soil at pondingtime ...... 66

3.14. An SEM micrograph, vertical section, showing clay-size particles sealing the surface (A) in the protected Asogueros at the end of the drying period ...... 66

3.15. Micrograph of a vertical thin section with plain light, showing raindrop impact (A) on the surface of the gypsum-amended Asogueros soil at ponding time ...... 67

3.16. An SEM micrograph, vertical section, showing the aggregation effect in the surface (A) and subsurface (B) in the gypsum- amended Asogueros soil at ponding time ...... 67

3.17. Micrograph of a vertical thin section with plain light, showing the raindrop impact on the soil surface (A) in the bare Sandovales soil at ponding time ...... 68

3.18. An SEM micrograph, vertical section, showing clay-size particles sealing the surface (A) in the bare Sandovales soil at ponding t i m e...... 68

3.19. An SEM micrograph, vertical section, showing a compacted 0.2-mm surface layer (A) in the bare Sandovales soil at the end of the drying period ...... 70

3.20. Micrograph of vertical thin section with plain light, showing an undisturbed soil surface in the protected Sandovales soil at ponding tim e...... 70

3.21. An SEM micrograph, vertical section, showing accumulation of clay-size particles in the surface (A) of the protected Sandovales soil at ponding tim e...... 71

xii Page 3.22. An SEM micrograph, vertical section, showing a homogeneous distribution of clay-size particles in the surface (A) and subsurface (B) ...... 71

3.23. Micrograph of a vertical thin section with plain light, showing raindrop impact on the surface (A) in the gypsum-amended Sandovales soil at ponding tim e...... 73

3.24. An SEM micrograph, vertical section, showing the aggregation effect (A) of gypsum in the gypsum-amended treatment of the Sandovales soil at the end of the drying period ...... 73

3.25. An SEM micrograph showing natural gypsum (A) bridging soil particles (B) in the bare Asogueros so il...... 75

3.26. An SEM micrograph showing coatings composed of Fe (A) in the bare Sandovales soil ...... 75

3.27. Sulfur and Ca peaks detected in the Asogueros bare soil by XRF, indicate the presence of gypsum as a cementing agent in the crust ...... 76

3.28. Aluminum and Si peaks detected by XRF in soil particles bound by gypsum in the crust of the Asogueros bare so il 76

3.29. Iron peak detected by XRF in the Sandovales bare soil, suggests the presence of free Fe oxides as a cementing agent in the c r u s t ...... 77

3.30. Aluminum and Si peaks detected by XRF in soil particles coated by free Fe oxides in the crust of the Asogueros bare soil...... 77

3.31. An SEM micrograph showing massive coatings of Si in the bare Reforma so il...... 79

3.32. An SEM micrograph showing calcite coating (A) soil particles in the bare Gigger soil ...... 79

3.33. Silica peak detected in the Reforma bare soil, suggests the presence of amorphous Si as a cementing agent in the crust . . . 80

3.34. Aluminum and Si peaks detected by XRF in soil particles coated by amorphous Si in the crust of the Reforma so...... il 80

xiii Page 3.35. Calcium detected by XRF in the Gigger bare soil, suggests the presence of calcite as a cementing agent in the c ru st...... 81

3.36. Aluminum and Si peaks detected by XRF in soil particles coated by calcite in the crust of the Gigger bare so il...... 81

3.37. An SEM micrograph showing filaments composed of Fe (A) bridging particles (B) in the bare Olivier s o i l ...... 82

3.38. An SEM micrograph showing a filament composed of Si (A) bridging particles (B) in the bare treatment of the Coteau soil . . 82

3.39. Iron peak detected by XRF in the Olivier bare soil, suggests the presence of free Fe oxides as a cementing a g e n t...... 83

3.40. Aluminum and Si peaks detected by XRF in soil particles bound by free-oxide bridges in the crust of the Olivier bare s o il ...... 83

3.41. Silica peak detected by XRF in the Coteau bare soil, suggests the presence of amorphous Si as a cementing agent in the crust 84

3.42. Aluminum and Si peaks detected by XRF in soil particles bound by amorphous Si in the crust of the Coteau so...... il 84

4.1 Solubility diagrams for calcite and gypsum under different management practices in the soils of Mexico and Louisiana . . . 118

4.2 Simulated concentration of soil solutions during the drying process for the Gigger soil ...... 123

4.3 Simulated concentration of soil solutions during the drying process for the Olivier soil ...... 125

4.4 Simulated concentration of soil solutions during the drying process for the Sandovales so il...... 128

5.1 Modified penetrometer used to measure crust strength ...... 138

5.2 Infiltration rates of the soils of Mexico and Louisiana with bare, protected, and gypsum-amended soils ...... 141

5.3 Relationship between ponding time and water-dispersible clay in d e x ...... 143

xiv Page 5.4 Relationship between sediment loss and water-dispersible clay in d e x ...... 148

5.5 Effect of management practices and rainfall duration in crust strength in the soils of Mexico and Louisiana...... 150

5.6 Relationship between crust strength and soil moisture ...... 151

5.7 Relationship between crust strength and water-dispersible clay in d e x ...... 156

5.8 Relationship between crust strength and amorphous S i...... 156

5.9 Effect of management practices and rainfall duration on corn and soybean emergence...... 159

A.1 Rotating-disk rainfall sim ulator...... 181

A.2 Rotating plastic disk with an aperture of 3 0 ° ...... 181

xv ABSTRACT

Crust formation in three soils from Mexico (Nadurargid, Durustoll and

Calciorthid) and three soils from Louisiana (two Fragiudalfs, and a

Hapludalf) was investigated. The objectives were: 1) to characterize crust morphology; 2) to quantify type, distribution, and geochemistry of cementing agents; and 3) to evaluate the effect of management practices on infiltration, erosion, crust strength, and corn and soybean emergence.

Three management practices, bare, protected, and gypsum-amended were evaluated. Crust morphology in the soils of Louisiana indicated the development of a compacted layer in the upper 0.1 mm of the bare and gypsum-amended soils. Aggregation was improved with the gypsum amendment. The soil surface was sealed by clay-size particles in the protected soil. In the Mexican soils, fine particles were dispersed, sealing the soil surface. Calcite, free iron oxides, amorphous Si, and gypsum, were identified as cementing agents by SEM/XRF analysis. There was no accumulation of free Fe oxides or amorphous Si and Al in the crusts, relative to the subsurface zones. As a soil amendment, gypsum induced precipitation of calcite in cases where calcite was not detected in the soil initially. Infiltration rate and ponding time were improved in all six soils by the gypsum amendment. Erosion was decreased in the gypsum-amended

xvi soils by 54% and 98% in the protected soils, compared to that in the bare soils. Crust strength increased by 90% in the soils of Mexico and by 25% in the soils of Louisiana as rainfall duration increased from 30 to 60 min.

The higher values of crust strength in the soils of Mexico were related to higher water-dispersible clay indices and higher amorphous Si contents in these soils. Corn emergence was reduced by 75% in the soils of Mexico and 13% in the soils of Louisiana when rainfall intensity w as increased from 30 to 60 min. Soybean emergence was decreased by 77% in the soils of Louisiana. No soybean emergence was observed in the soils of

Mexico when rainfall increased from 30 to 60 min.

xvii CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

Introduction

Loessial soils of Louisiana and soils from the north-central part of

Mexico have a tendency to form crusts at the soil surface. Crust formation starts when soil aggregates are dispersed by physical and/or chemical mechanisms. Dispersed soil particles may seal the soil surface, infiltrate into the soil or move away in the runoff. After these processes have taken place, a subsequent hot and dry period causes cementing agents to form and bind soil particles together. Finally, crust formation negatively affects surface strength, infiltration, gas soil-atmosphere interchange, and emergence of plants while runoff and soil erosion are increased.

Two types of soil crusts are recognized: i) a structural crust and ii) a depositional crust. The processes involved in the development of a structural crust are presented above. The depositional crust results from the transportation of dispersed particles by runoff and their deposition at a new location on the soil surface, burying the underlaying material. This study addresses the processes involved in structural crust formation.

The mechanisms involved in the formation of structural crusts are not clear in the loessial soils of Louisiana and soils from north-central

Mexico. In both areas, soils are often subjected to high-intensity rainfall at

1 2 the beginning of the growing season when soils have no protection from the beating action of rain-drops. In addition to this physical dispersion of soil aggregates, specific chemical properties may also affect the loss of ; exchangeable sodium, salt content, and the presence of highly dispersive clays such as smectites are examples. A characterization of the physical, chemical, and mineralogical properties which have been related to soil crusting is presented in Chapter 2.

The micromorphology observed during crust formation is presented in Chapter 3. Understanding morphological changes during crust formation helps to explain alterations in infiltration, erosion, and soil dispersion and allows one to observe the structural arrangement of the particles. In this chapter, the elemental composition of cementing agents is presented as well.

The complexity of soil crust formation increases with the precipitation of cementing agents in the soil surface. Chapter 4 addresses cementing agents present in the crusts of soils of Mexico and Louisiana, and considers the accumulation of cementing agents in the crust during evaporation.

Chapter 5 deals with the effects of rainfall duration and soil m anagem ent practices on infiltration, soil erosion, crust strength, and seedling emergence of soybean(Glycine max (L.) Merr.) and corn (Zea corn 3

L.). Finally, the last chapter presents an overall summary and conclusions of this study.

Review of Literature

The remainder of this chapter reviews literature addressing the processes involved in the formation of structural crusts: the dispersion of soil aggregates, and the segregation and cementation of soil particles.

Physical dispersion

Physical dispersion is considered one of the initial processes in the formation of soil crusts. The destruction of aggregates is related to , aggregate stability, soil moisture, and rainfall characteristics

(Bradford and Huang, 1991). An analysis of how these factors affect soil dispersability during crust formation follows.

Rainfall characteristics

The most important rainfall characteristics affecting the break down of soil aggregates and consequently the formation of soil crusts are rainfall intensity, duration, frequency, and drop size (Awadhwal and Thierstein,

1986).

Agassi et al. (1985) reported that no seal was formed in applying low-energy rainfall (0.01 J/m m /m 2) on either a -textured Calcic

Haploxeralf or a clay-textured Typic Chromoxerert. However, with high- energy rain (23 J/mm/m2), both soils formed a surface crust. In turn, Singh 4

(1979) stated that the emergence of plants decreased as intensity and duration of rainfall increased.

Rainfall frequency is the major factor in crust formation which limits seedling emergence. A rain event just after planting or during the first days following planting can result in a dense crust which seriously limits seedling emergence. Aujla et al. (1986) pointed out that a heavy rainfall tw o days after sowing wheat (Triticum aestivum L.) caused a 31 % reduction in grain yield due to the formation of a crust.

Larger rain drops have more kinetic energy to break down soil aggregates and to compact the soil surface. Sivrapasad and Sarma (1987) reported that as drop size increased, crust formation was more evident, consequently, mean time for seedling emergence of chickpea(Cicer arietun

L.), pigeonpea (Cajunus cajan L.), and pearl millet (Pennisetum typhoides

L.) were negatively influenced.

Soil texture

Crusts are formed in soils of almost any textural class except those with extremely low and clay content (Lemos and Lutz, 1957). Soil texture was reported by Singer and Warrington (1991) to be as the main factor which influences the formation, strength, and stability of crusts.

Ferry and Olsen (1974) reported that sandy soils formed weaker crusts because the irregular shape of caused more random arrangement and less close packing of soil particles. In clayey soils, crust 5 strength increases because soil particles in suspension settle in an oriented,

plate-like fashion, increasing particle-particle attraction. The electrostatic charge of clays allows them to act as cementing agents.

The strength of crusts was reported to increase as clay content

increased (Ben-Hur et al., 1985). Soils having a clay content of 10-30% were reported as being the most susceptible soils to form hard crusts.

Aggregate stability

The role of texture in aggregate stability is a key component in soil crusting. The stronger the soil aggregation, the less probability that a soil will become crusted (West et al., 1991). These authors found that the stability of aggregates increased as clay content increased. In sandy and sandy-loam soils, soil aggregates were easily dispersed due to the physical effect of raindrop impact.

Aggregate size is an important function preventing soil crusting.

Aggregates larger than 19 mm in diameter delayed crust formation compared to smaller aggregates in silt loam (Typic Hapludolls) and clay loam (Typic Haplaquolls) soils (Bradford and Huang, 1991).

Chemical dispersion

In addition to the physical dispersion of soil aggregates, specific chemical properties may also affect the loss of soil structure; exchangeable sodium, salt content, and the presence of highly dispersive 6 clays such as smectites are examples. A review of how these factors affect the chemical dispersion during soil crust formation follows.

Clay mineralogy

Clay mineral composition is important in crust formation. Smectites and illites are recognized as being more dispersible than kaolinite (Arora and

Coleman, 1979). Dispersion of 2:1 clay minerals arises from repulsive forces originating in the electrical double layer (Sumner, 1991). Brown

(1984) stated that the swelling effect in smectites is caused by hydration of interlayer cations. When highly hydrated Na is the interlayer cation and conditions of high relative humidity occur, the 2:1 structures are dispersed.

When Ca is the interlayer cation, swelling is limited, and the 2:1 structures are more stable (Brown, 1984).

Although kaolinitic soils are known to be less dispersible than smectitic soils, Stern et al. (1991) reported that highly weathered kaolinitic soils in the southeast U.S. were dispersed when a small amount of smectite was present in the clay complex. The same authors pointed out that while smectitic soils were more dispersible than soils with only small amounts of smectite, soils without smectite were more stable and less susceptible to crust formation than soils with small amounts of smectite.

Exchangeable ions and electrolyte composition

Chemical dispersion of soil aggregates is often due to the presence of high levels of exchangeable Na and variable electrolyte concentration. 7

Monovalent ions such as Na do not effectively reduce the electronegativity or zeta potential of clays. This allows for a repulsive force between clay particles which may cause dispersion (Brady, 1990; Sumner, 1991). High electrolyte concentration reduces chemical dispersion while low electrolyte concentration combined with a relatively high exchangeable sodium percentage (ESP) tends to disperse soil aggregates (Shainberg and Letey,

1984). Aggassi et al. (1981) reported that clay dispersion decreased as electrolyte concentration in applied water increased from zero in distilled water to 5.6 ds/m in well water.

Clay dispersion and soil crusting can occur even with low ESP values, especially when smectitic clays are present (Stern et al., 1991).

Kasman et al. (1983) reported formation of a soil crust in smectitic soils having ESP values of 1.0 and 2.2. In these soils, the final infiltration rate was 7.0 and 2.4 mm/h, respectively, suggesting that the decrease in the final infiltration w as due to caused by dispersion of clays.

Even though mechanical and chemical dispersion are recognized as different processes, they can be complementary factors in soil dispersion and crust formation. Shainberg and Singer (1988), observed that soil dispersion increased as rainfall energy increased in soils with ESP of 0.

However, in soils with ESP of 5 and 10, soil dispersion occurred at lower rainfall energy than in soils with ESP of 0. Another example of an interaction, affecting soil dispersion and crust formation, involves rainfall energy and electrolyte concentration in the applied rainfall. Agassi et al. (1985) pointed out that using low energy rainfall (0.01 mm/m2) and distilled water in soils with ESP of 2.5, dispersed soil aggregates reduced the final infiltration by as much as 75% compared to when saline water was used.

Exchangeable ion composition and electrolyte concentration can be altered by chemical fertilizers and soil amendments. For example, Miller and Scifres (1988) reported that the use of NaN03 as a fertilizer encouraged soil dispersion, reducing final infiltration to 2-3 mm/h. The opposite effect was found when phospho-gypsum was applied, yielding a final infiltration of 23 mm/h.

Segregation of soil particles

Once soil aggregates have been dispersed, fine particles may move with the infiltrating water, runoff, or settle on the soil surface. When such dispersed particles move with the infiltration water, soil pores beneath the soil surface are clogged, leading to the formation of a clay-rich layer at the bottom of the crust, referred to by McIntyre (1958) as the "washed in" zone.

The formation of the "washed in" zone is usually related to clay dispersion in the soil surface. Gal et al. (1984) reported that surface crusts in soils with ESP > 1 and exposed to rain with distilled w ater consisted of a skin of naked sand grains, followed by a clay-rich layer. However, the same authors reported that when the ESP of the soil was 1, clay was not dispersed and the crust consisted only of a compacted skin layer in the soil surface.

The manner in which suspended soil particles settle in the soil surface depends on the electrolyte concentration of the soil solution. With high electrolyte concentrations, clay particles tend to flocculate. As a result, the crust consists of particles deposited randomly forming a structure with high permeability. When the electrolyte concentration is low, the crust consists of dispersed clay particles which settle with parallel orientation. Crusts with oriented clays have been reported as having low permeability (Shainberg and Singer, 1984).

The findings discussed above suggest that segregation of soil particles within the soil profile, depends on chemical dispersion. However,

Tarchitzky et al. (1984) report that the presence of natural aggregates below the sealed surface is clear evidence that soil aggregates at the soil surface were dispersed by rainfall impact. These contrasting results support the idea that crust formation is the result of an interactive effect of physical and chemical factors rather than the product of a single process or factor. 10

Drying period

Cementing agents

Less is known about the nature and properties of cementing agents during soil crusting than the dispersion process. Silica in semiarid zones, sesquioxides in subtropical zones, and organic matter, in both cases, are considered the main cementing agents in soils from these respective areas.

Research addressing the role of cementing agents in soil genesis has focussed more on stable structures such as duripans, hard-setting horizons, and saprolites rather than in temporal surface crusts (Chartres and

Fitzgerald, 1988; Chartres et al., 1990).

Silica as a cementing agent is more likely to occur in semiarid zones because of its tendency to accumulate rather than leach from the profile.

Silica can occur in different forms: soluble molecules (silicic acid), homogeneously dispersed colloids (hydrosol), nonrigid gels (hydrogels) and rigid gels (xerogels) (Hallmark et al., 1982).

Chartres et al. (1990) conducted a study to determine the role of chemical cementing agents in a hard-setting soil. Amorphous Si, an imogolite-like aluminosilicate, a feldspathoid mineral, and possibly Si-Fe complexes were the major cementing agents. Chartres and Fitzgerald

(1990) pointed out that cementation in occurred as a result of impregnation of clays and matrix materials with small amounts of amorphous Si. Observations using transmission and scanning microscopy 11 in E horizons showed amorphous Si and amorphous aluminosilicates bonding the soil particles. These authors added that amorphous Si and aluminosilicates acted as temporary cementing agents when they precipitated during soil drying.

Brown and Mahler (1988) suggested that amorphous Si concentration in soil is increased by using ammonium-based N and P fertilizers. They added that long-term use of these fertilizers causes a surface acidification. Silica acts as a cementing agent by sorbing at soil particle-surfaces when levels of silica concentrations approach the solubility of amorphous Si. These results could be important as the use of ammonium-based N and P fertilizers is a common practice in the semiarid zones of Mexico. The increase in acidity would solubilize unstable silicate minerals, increasing the concentration of soluble silica in the soil profile, thereby increasing its potential as a cementing agent.

Amelioration of soils with crusting problems

Crust formation is common in soils with poor aggregation. The use of gypsum, phospho-gypsum, organic compounds, polyvalent salts, or synthetic polymers enhance soil aggregation, and consequently the risk of crust formation is reduced. The main effects of these products include the improvement of seedling emergence, pore space, infiltration, drainage, water holding capacity, and hydraulic conductivity. As a result, runoff, 12 erosion, and evaporation are reduced (Terry and Nelson, 1986; Wallace and

Abouzamzam, 1986; Shaviv et al., 1987; and Ben-Hur and Letey, 1989).

The addition of gypsum to soils with an excessive ESP of > 10 has reduced crust strength and improved aggregate stability according to

Awahwal and Thiertein (1986). In addition, Ben-Hur et al. (1992) reported that gypsum increased the electrolyte concentration in the soil solution, preventing clay dispersion and increasing final infiltration rates from 10.0 to

35.0 mm/h when soils were amended with gypsum.

Polyacrylamide (PAM), an organic polymer of high molecular weight, has the ability of binding clay particles together forming water-stable aggregates. This polymer improved seedling emergence by reducing crust strength (Wallace and Abouzamzam, 1986). Terry and Nelson (1986) reported that final infiltration was doubled after applications of 5-20 kg/ha of PAM.

Soil crust modeling

An early qualitative model describing soil crust formation was published by McIntyre (1958). He pointed out that a structural crust was formed when soil aggregates were destroyed by the beating action of raindrops. McIntyre's model stated that after the destruction of the aggregates, fine material was "washed in" to the soil with the infiltration water or "washed out" with the runoff. After the "washed in" process, two new layers in the soil surface were usually observed as a part of the 13 soil crust in the soil surface: i) a compacted "skin" 0.1 mm in thickness and ii) a "washed in" 1.5 to 2.5 mm thick layer. The acceptance of this model since then has been very controversial. Some authors have observed similar processes during soil crust formation (Tarchitzky et al.,

1984, Onofiok and Singer, 1984), while other authors have not observed the two layers (Chen et al., 1980, Epstein and Grant, 1973). The main objection to McIntyre's model has been the "skin" layer. An obvious question arises regarding how such a "skin" layer could be maintained at the surface of the soil during a rainfall event, given the direct impact of raindrops.

Most quantitative models have been developed to describe how water infiltration is altered by the presence of a sealing layer in the soil surface. For example, Seginer and Morin (1970) showed that the initial hydraulic conductivity was reduced when the impact of rain drops on the soil surface sealed the soil surface. However, the continuous impact of rain drops on the soil surface caused infiltration to return to its initial value by opening the momentarily sealed surface. This model considered that water infiltration in the upper soil layer was the average of the sealed layer and the layer beneath it. Based on the results of Seginer and Morin

(1970), Morin and Benjamini (1977) modified the model to make it applicable to different rainfall events by including rainfall intensity as a variable. Another quantitative model was reported by Maulem et al. (1992).

The basic approach of this model assumes that soil aggregates are destroyed by rainfall and clay-size material is released into suspension.

This fine material moves downward until clogging soil pores. The continuous destruction of soil aggregates by rainfall involves settling, filtering, and compaction of soil particles until an equilibrium point between erosion and seal construction is reached. This model assumes that the reduction in permeability of the disturbed layer results from physico­ chemical factors such as a reduction of voids, electrical conductivity, cation valence, cation exchange capacity, and ESP. This model also refers to rainfall kinetic energy and macroscopic soil properties such as seal thickness, water retention, and hydraulic conductivity as the basis for predicting the thickness of the soil crust and its effect on hydraulic conductivity.

A controversial point in crusting models has been the use of either cumulative rainfall or kinetic energy as the rainfall variable responsible for seal formation. A possible solution to this controversial point was presented by Maulem et al. (1992) in a model which considered cumulative rainfall rather than kinetic energy as the variable dominating seal formation. 15

References for Chapter 1

Agassi, M., I. Shainberg, and J. Morin. 1981. Effect of electrolyte concentration and soil sodicity on infiltration rate and crust formation. Soil Sci. Soc. Am. J. 45:848-851.

Agassi, M., J. Morin, and I. Shainberg. 1985. Effect of raindrop impact energy and water salinity on rates of sodic soils. Soil Sci. Soc. Am. J. 49:186-190.

Arora, H.S. and J. Letey. 1979. The influence of electrolyte concentration on flocculation of clay suspensions. Soil Sci. 127:134-139.

Aujla, T.S., B. Singh, and B.S. Sandhu. 1986. Effect of pre-sowing and post-sowing irrigation treatment to wheat following rice in Punjab. J. Ecology. 13(21:250-255.

Awadhwal, N.K. and G.E. Thierstein. 1986. Soil crust and its impact on crop establishment: a review. Soil Till. Res. 5(3):289-302.

Ben-Hur, M., and J. Letey. 1989. Effect of polysaccarides, clay dispersion, and impact energy on water infiltration. Soil Sci. Am. J. 53:233-238.

Ben-Hur, M., I. Shaiberg, R. Keren, and M. Gal. 1985. Effect of water quality and drying on soil crust properties. Soil Sci. Am. J. 49:191-196.

Ben-Hur, R.Stern, A.J. van der Merwe, and I. Shainberg. 1992. Slope and gypsum effects on infiltration and erodibility of dispersive and nondispersive Soils. Soil Sci. Am. J. 56:1571-1575.

Bradford, J., M. and C. Huang. 1991. Mechanisms of crust formation: physical components. ]n M.E. Sumner and B.A. Stewart (eds). Soil Crusting-Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 55-72.

Brady, N,C. 1990. The nature and properties of soils. 10th ed. MacMillan. New York. 621 p.

Brown, G. 1984. Crystal structures of clay minerals and related phyllosilicates. Phil. Trans. R. Soc. Cond. 311:221-240. 16

Brown, T.H. and R.L. Mahler. 1988. Effects of phosphorous and acidity levels of silica extracted from a Palouse silt loam. Soil Sci. Am. J. 51:674-677.

Chartres, C.J., J.M. Kirby, and M. Raupach. 1990. Poorly ordered silica and aluminosilicates as temporary cementing agents in hard-setting soils. Soil Sci. Soc. Am. J. 54:1060-1067.

Chartres, C.J, and J.D. Fitzgerald. 1990. Properties of silicious cements in some australian soils and saprolites. Dev. Soil Sci. 19:199-205.

Chen, J., J. Tarchitzky, J. Morin, and A. Banin. 1980. Scanning electron microscope observations on soil crust and their formation. Soil Sci. 130:49-55.

Epstein, E., and W.J. Grant. 1973. Soil crust formation as affected by raindrop impact. Ecol. Studies, 4:195-201.

Ferry, D.M. and R.A. Olsen. 1974. Orientation of clay particles as it relates to crusting of soil. Soil Sci. 4:367-375.

Gal, M., L. Arcan, I. Shainberg, and R. Keren. 1984. The effect of exchangeable Na and phosphogypsum on the structure of soil crust-SEM observations. Soil Sci. Soc. Am. J. 48:872-878.

Hallmark, C.T., L.P. Wilding, and N.E. Smeck. 1982. Silicon. ]n. A.L. Page, (ed) Methods of Soil Analysis, Part 2. Chemical and Microbiological Properties. 2nd. ed. Agronomy 9:159-165. Am. Soc. Agron. Madison, Wi.

Kasman, S., I. Shainberg, and M. Gal. 1983. Effect of low levels of exchangeable Na and applied phosphogypsum on the infiltration rate of various soils. Soil. Sci. 35:184-192.

Lemos, P. and J.F. Lutz. 1957. Soil crusting and some factors affecting it. Soil Sci. Soc. Am. J. 21:485-491.

Maulem, Y., S. Assouline, and H. Rohdenburg. 1990a. Rainfall induced soil seal. (B) Application of a new model to saturated soils. CATENA. 17:2105-2108.

Maulem, Y., S. Assouline, and H. Rohdenburg. 1990b. Rainfall induced soil seal. (C) A dynamic model with kinetic energy instead of cumulative rainfall as independent variable. CATENA. 17:289-303. 17

McIntyre, D.S. 1958. Soil splash and the formation of surface crusts by raindrop impact. Soil Sci. 85:261-266.

Miller, W.P., and J. Scifres. 1987. Effect of sodium nitrate and gypsum on infiltration and erosion of a highly weathered soil. Soil Sci. 148:304- 309.

Morin, J. and Y. Benjamini. 1977. Rainfall infiltration into bare soils. Water Resour. Res. 13:813-817.

Onofiok, 0., and M.J. Singer. 1984. Scanning electron microscope studies of surface crusts formed by simulated rainfall. Soil Sci. Soc. Am J. 48:1137-1143.

Seginer, I. and J. Morin. 1970. A model of surface crusting and infiltration of bare soils. Water Resources Res. 6:629:633.

Shaviv, A., I. Ravina, and D. Zaslavsky. 1987. Application of soil conditioner solutions to soil columns to increase stability of aggregates. Soil Sci. Am. J. 51:431-436.

Shainberg, I., and J. Letey. 1984. Effect of electrolyte concentration on the hydraulic properties of depositional crust. Soil Sci. Soc. Am. J. 49:1260-1263.

Shainberg, I., and M. J. Singer. 1988. Drop impact energy-soil exchangeable sodium percentage interaction in seal formation. Soil Sci.Soc. Am. J. 52:1449-1452.

Singer, M. J., and D.N. Warrington. 1991. Crusting in the western United States, jn M.E. Sumner and B.A. Stewart (eds). Soil Crusting-Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 179-204.

Singh, B. 1979. Effect of rainfall characteristics on crusting seedling emergence of bajra (Pennisetum typhoydes L.) and cotton (Gossypium histutum L.) in sandy loam and silty clay loam soils. Thesis Abstracts. 5(2):124.

Stern, R., M. Ben-Hur, and I. Shainberg. 1991a. Clay mineralogy effect on rain infiltration, seal formation and soil losses. Soil Sci. 152:455-462. 18

Stern, R., M.C. Laker, A.J. van der Merwe and I. Shainberg. 1991b. Seal formation, runoff, and erosion in dispersive and stable kaolinite soils. Soil Sci. Am. J.

Sivrapasad, B. and K.S.S. Sarma. 1987. Seedling emergence of chickpea (Cicer arietum L.), pigeonpea (Cajunus cajan L.), and pearl millet (Pennisetum typhoides L.). Effect of differential soil crusting as induced by raindrop size, and depth of sowing. Indian Agric. Res. 104(2):263-268.

Sumner M.E. 1991. The electrical double layer and clay dispersion, in M.E. Sumner and B.A. Stewart (eds). Soil Crusting-Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 1-31.

Tarchitzky, J., A. Bannin, J. Morin, and Y. Chen. 1984. Nature, formation and effects of soil crusts formed by water drop impact. Geoderma 33:135-155.

Terry, E.R., and S.D. Nelson. 1986. Effects of polyacrylamide and irrigation method on soil physical properties. Soil Sci. 141 (5):317-320.

Wallace, A., and A.M. Abouzamzam. 1986. Interaction of soil conditioner with other limiting factors to achieve high crop yields. Soil Sci. 141 (5):343-345.

West, L.T., S.C. Chiang, and L.D. Norton. 1991. The Morphology of surface crusts, in M.E. Sumner and B.A. Stewart (eds). Soil Crusting- Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 73-92. CHAPTER 2

PHYSICAL, CHEMICAL, AND MINERALOGICAL PROPERTIES AFFECTING SOIL CRUST FORMATION IN SOILS OF MEXICO AND LOUISIANA

Introduction

The factors associated with soil genesis are climate, parent material, topography, time, and living organisms (Jenny, 1941; Simonson, 1959; and Bawer, 1990). The intensity and interaction of these factors have produced the variety of soil properties observed around the world. For example, the major factors responsible for the differences between the loessial soils of Louisiana and the soils from the north-central part of

Mexico are contrasting climate and parent material.

Loess soils of Louisiana are formed predominantly in silt-size materials deposited by wind from the Mississippi River flood plain during periods of glaciation (Miller et al., 1987). In contrast, soils from the north- central part of Mexico have formed on alluvial deposits and sedimentary rocks of volcanic origin (CETENAL, 1971a, 1971b).

Soils of Louisiana soils have been subjected to a more intense weathering environment driven by a mean annual rainfall of 1340 mm while in the north-central part of Mexico, the mean annual precipitation is 450 mm. In Louisiana, the abundant rainfall has encouraged the of exchangeable bases such as Ca and Mg while in Mexico, the scarce precipitation has encouraged not only the accumulation of exchangeable

19 20 bases, but also the formation of carbonates and in some cases gypsum.

Despite the contrasting soil environments, a common problem occurs in both areas: the tendency to form surface crusts which restrict seedling emergence and infiltration and increases soil erosion.

In general, the tendency of soils to form a crust in the soil surface has been attributed to the presence of high levels of exchangeable Na and electrolytes, highly dispersive clays such as smectites, and rainfall characteristics (Agassi et al., 1985; Stern et al., 1991; Shainberg and

Letey, 1984). A knowledge of soil properties of the soils from the north- central part of Mexico and soils of Louisiana is needed to understand the crust formation process in these soils.

The objective of this study was to characterize the physical, chemical, and mineralogical properties of six soils of Mexico and Louisiana as related to crust formation.

Materials and Methods

Soil series

The surface horizons of three loess soils from Louisiana and three soils from the north-central part of Mexico were used in this study. Soils from Louisiana were: Gigger (fine-silty, mixed, thermic, Typic Fragiudalf),

Olivier (fine-silty, mixed, thermic, Aquic Fragiudalf), and Coteau (fine-silty, mixed thermic, Glossaquic Hapludalf). Soils from Mexico were: Asogueros 21

(loamy, mixed, hypothermic, Vertic Nadurargid), Sandovales (loamy, mixed, hypothermic, Aridic Durustoll), and Reforma (loamy, mixed, hypothermic,

Lithic Calciorthid). The approximate site locations of soils of Mexico and

Louisiana are given in Figure 2.1.

The Gigger soil from Franklin Parish formed in loess. Drainage and permeability are moderate and slow, respectively. Runoff and infiltration are medium and slow respectively. The main use of this soil is agriculture.

Main crops are cotton, soybeans, corn, oats, sweet potatoes, grain sorghum, truck crops, and wheat. Because erosivity in this soil is high, erosion control is required. The mean annual air temperature is 19°C and the mean annual rainfall is 1285 mm (Martin et al., 1981).

The Olivier soil from East Baton Rouge Parish formed in loess material and is located on ridges, broad flat valleys, and in slight depressions. Drainage and permeability are poor and slow, respectively.

Olivier soils are almost level so that runoff is not severe. However, erosion control is needed if row crops are grown. The mean annual air temperature is 20°C and the mean annual rainfall is 1360 mm (Dance et al., 1968).

The Coteau soil from Saint Landry Parish formed in loess material and is located on terrace uplands having a slightly convex ridge landscape.

Slope ranges from 0 to 3 percent. Drainage is poor, and runoff and infiltration are slow. Shrink-swell potential is moderate. Soils are suitable for soybeans, corn, vegetables, rice, and sweet potatoes. The mean Site Location

1 Gigger 2 Olivier 3 Coteau 4 Asogueros 5 Sandovales 6 Reforma

Figure 2.1. Site locations of the soils of Mexico and Louisiana. N> N3 23

annual air temperature is 20°C and the mean annual rainfall is 1387 mm

(Murphy et al., 1986).

The Asogueros soil is located in alluvial areas in Salinas de Hidalgo,

San Luis Potosi, Mexico. Texture in these soils is loam. Drainage and

infiltration is poor. Shrink-swell potential is high. Soils are suitable for

corn, beans, and wheat. The mean annual air temperature is 18°C and the

mean annual rainfall is 435 mm (CETENAL, 1971b).

The Sandovales soil is located on flat areas in Aguascalientes,

Mexico. The slope is <2% , the texture is sandy loam, and a duripan is

present within 50 cm of the soil surface. The parent material is sedimentary rock. Drainage and infiltration are poor. The most limiting factor to agriculture is the low and erratic precipitation. Organic carbon content and fertility level are very low. The most important crops grown in these soils are corn and beans. The mean annual air temperature is 18°C and the mean annual precipitation is 450 mm (CETENAL, 1971a).

The Reforma soil is located in Salinas de Hidalgo, San Luis Potosi,

Mexico in alluvial areas with poor drainage and infiltration. Shrink-swell potential is high. These soils are suitable for corn and beans. The mean annual air temperature is 17°C and the mean annual rainfall is 478 mm

(CETENAL, 1971b). 24

Physical characteristics

Particle-size distribution

Particle-size distribution was determined after dispersing the samples with sodium hexametaphosphate solution. The samples were shaken on a horizontal reciprocating shaker for 16 h. Clay percentage was determined using the pipet method described by Gee and Bauder (1986). Sand fractions were sieved and weighed. Silt content was calculated by difference.

Soil moisture content

Soil moisture content of samples used in physical and chemical analyses was determined from oven-dried subsamples at 105°C for 24 hours.

Chemical characteristics

Soil reaction

Soil reaction was measured in 1:1 H20 and 1:1 1M KCI on a

Beckman Zeromatic pH meter following the method described by Soil

Survey Staff (1984).

Organic carbon

Organic carbon (OC) was estimated by the modified Walkley-Black procedure described by Prince (1955). 25

Gypsum

Gypsum was quantitatively determined by the electrical conductivity method described by the U. S. Salinity Lab. Staff (1969).

Calcium carbonate

The CaC03 equivalent was determined by acid neutralization using the method described by the U. S. Salinity Lab. Staff (1969).

Amorphous silica and aluminum

Amorphous Si and Al were estimated using hot 1.0 M NaOH as described by Alexiades and Jackson (1967).

Free iron oxides

Free Fe oxides were determined by the dithionite-citrate-bicarbonate method described by Mehra and Jackson (1960).

Saturation extract

Saturated pastes were equilibrated for 24 hours. The extracted solution was obtained using Buchner funnels and vacuum as described by the U. S. Salinity Lab. Staff (1969). Soluble Ca, Mg, K, and Na were determined for the saturation extracts using ICP. Sulfates, Cl, and N03 were determined by ion chromatography (IC). The HC03 content was estimated as the anion deficit.

Electrical conductivity

The electrical conductivity (EC) of saturated extracts was obtained using a

YSI model 35 conductance meter (U.S. Salinity Lab. Staff, 1969). 26

Water-dispersible clay index

A water-dispersible clay index (WDCI) was estimated by dividing the clay percentage obtained using distilled water (with no dispersing agent) by the clay percentage using sodium hexamethaphosphate as a dispersing agent (Stern et al. 1991).

Sodium adsorption ratio

The sodium adsorption ratio was obtained using the following equation (1) (U. S. Salinity Lab. Staff, 1969):

Na SAR = ------((Ca + Mg)/2)1/2 where ion concentrations were expressed in cmol/L.

Mineral characteristics

Pre-treatments for clay mineralogy

Soils were treated to remove organic matter, carbonates, iron oxides, and amorphous materials. Organic matter was removed using hydrogen peroxide as described by Kunze and Dixon (1986). Carbonates were removed using the acetic acid (HOAc) method cited by Jackson (1985).

Iron oxides were removed using dithionite-citrate-bicarbonate (DCB) according to Mehra and Jackson (1960).

The whole clay fraction (<2.0//m) was separated by sedimentation.

Fine clay (<0.2//m) and coarse clay (2.0 to 0.2fjm) fractions were 27

separated by centrifugation (Jackson, 1985). After separation of coarse

and fine clays, amorphous silica and aluminum were removed using boiling

1M NaOH as described by Alexiades and Jackson (1967).

Weigth percentages of coarse and fine clays were determined

gravimetrically. Coarse and fine clays were saturated with 0.33 N KCI and

0.33 N MgCI2 prior to XRD analysis. Treatments with KCI were rinsed with

distilled water. Magnesium chloride treated clays were solvated with

ethylene-glycol and glycerol.

Clay mineral identification

Oriented slide mounts were run on a Philips X-ray diffractometer

using Cu-Ko radiation. The Mg-saturated ethylene-glycol solvated treated

slides were run from 2° to 30° 20. All other treatments were run from2° to 15° 20 using 40 KV and 25 ma. The characteristic d-values

(angstroms) for mineral identification are shown in Table 2.1. Semi-

quantitative estimates of mineral composition were based on areas of

characteristic diffraction peaks.

Random-powder diffractograms

Identification of other soil minerals was also done by XRD. Samples

from bulk soils were finely ground and uniformly packed in a dry-powder

sample holder. Caution was used to avoid orientation of particles by

pressure effects during the packing procedure. Samples were run from 2° to 50° 20 using 40 KV and 25 ma. 28

Table 2.1. Characteristic X-ray diffractio peaks1 for minerals in clay fractions. Mineral Treatment d Value hkl

o (A) Smectite MgEG 17.0 001 8.5 002 K550 10.0 001 5.0 002 3.3 003 Vermiculite MgEG 14.2 001 7.1 002 K550 10.0 001 5.0 002 lllite MgEG 10.0 001 5.0 002 K550 10.0 001 5.0 002 Kaolinite MgEG 7.1 001 3.6 002 K300 7.1 001 3.6 002 K550 none none Quartz All 4.3 100 3.3 101 Feldespar All 3.1 002 3.2 040 Ti.'Jackson____i______t (1975).m i t » 29

Data quality

Physical and chemical analyses were performed on three replicates.

Data are reported as the mean of these replicates.

Discussion of Results

Physical characteristics

Particle size distribution

Particle size distribution influences the formation and strength of soil crusts. Soil crusts can be formed in soil of almost any texture except in sandy soils with extremely low silt and clay content (Singer and

Warrington, 1991; Lemos and Lutz, 1957). Particle size distribution results, with and without a dispersing agent, are presented in Table 2.2.

Silt was the dominant particle size fraction using sodium hexametaphosphate, for the Gigger, Olivier, and Coteau, ranging from 72 to 87%. In the soils of Mexico, sand was the dominant particle size fraction, ranging from 42 to 52%. Clay content in all six soils was similar, ranging from 11 to 24%.

The differences in soil texture between the soils of Mexico and

Louisiana can be attributed to contrasting parent materials. High silt content is typical of loessial soils of Louisiana as a consequence of the uniform particle size deposited by wind (Miller, 1984). The coarser texture Table 2.2. Particle size distribution for dispersing agent (DA) and distilled water (DW) and water dispersible clay index (WDCI).

------DA------Textural DW ------Soil Sand Silt Clay Class Sand Silt Clay WDCI ------% ------for DA % ------

Gigger 12 72 16 Silt loam 13 79 8 0.50 Olivier 2 87 11 Silt 8 90 2 0.18

Coteau 4 82 14 Silt loam 10 87 3 0.21 Asogueros 46 31 23 Loam 45 36 19 0.83

Sandovales 52 29 19 Sandy loam 57 28 15 0.79

Reforma 42 34 24 Loam 43 38 19 0.79

oo o 31 of the soils of Mexico is a result of the high sand content in their respective parent materials.

It is important to point out that the sand content of 12% obtained in the Gigger soil surface is not typical of loessial soils. However, Miller et al.

(1988) point out the presence of a basal mixing zone in the loessial soils of

Louisiana. This basal mixing zone is defined as a zone where a thin layer of loess is contaminated by underlying alluvial materials.

Water-dispersible clay index

When particle size distribution was determined using only distilled water, the clay content decreased about 80% in the Olivier and Coteau soils, and 50% in the Gigger soil. In the soils of Mexico, the decrease of clay content with distilled water was approximately 5.5% (Table 2.2).

Higher water-dispersible clay indices (WDCI) were observed in the soils of

Mexico (0.79 to 0.83) than in soils of Louisiana (0.18 to 0.50). These results clearly suggest that soils of Mexico are much more dispersible than soils of Louisiana. The high WDCI found in the soils of Mexico suggests this factor may have a substantial role in crust development. The low

WDCI in soils of Louisiana suggests that dispersed clays may not play a critical role in crust development.

One reason that the soils of Mexico are more dispersible is the higher

SAR in these soils compared to the soils of Louisiana. The higher WDCI in

Gigger with respect to that in Olivier and Coteau is most likely related to 32 the low OC content in this soil. A more detailed discussion addressing SAR

and OC will be presented later in this chapter.

Chemical characteristics

Potential cementing agents

The amounts of free Fe oxides, amorphous Si and Al, CaC03,

gypsum, and OC available to act as potential cementing agents are shown

in Table 2.3.

Iron oxides and amorphous Si and Al tend to accumulate in the soil

rather than being leached. The percent of free Fe oxides was higher in soils of Louisiana (0.19 to 0.49%) than in soils of Mexico (0.02 to 0.08%).

In contrast, amorphous Si and Al were 3 to 4 times higher in soils of

Mexico than in soils of Louisiana. The differences in the amount of these potential cementing agents suggests that the warm and humid climate in

Louisiana tends to concentrate free Fe oxides through hydrolysis and oxidation of Fe and that the semi-arid climate in the north-central part of

Mexico fails to remove the weathering products of amorphous Si and Al from the soil profile. Another important factor which may have influenced the contrasting amounts of amorphous Si and Al was the parent material.

The soils of Mexico were derived from parent material of volcanic origin which has a high probability of containing substantial amounts of amorphous glass. Table 2.3. Quantification of possible cementing agents in soils of Mexico and Louisiana.

Soils Fe203 Al203 Si02 CaC03 Gypsum OC"

% ------

Gigger 0.49 0.57 1.09 - - 0.27

Olivier 0.28 0.49 1.09 -- 1.02

Coteau 0.19 0.50 1.18 -- 1.19

Asogueros 0.07 1.46 4.38 3.90 - 1.15 Sandovales 0.08 1.42 4.51 - 0.33 0.33

Reforma 0.02 1.150 4.02 11.40 • 1.76 * = Organic carbon

co co 34

The presence of CaC03 and gypsum is another consequence of the contrasting climate between soils of Mexico and Louisiana. High Ca concentrations in the soils of Mexico favors the formation of calcite and gypsum. Calcium carbonates were present only in Asogueros and Reforma soils with 3.9 and 11.4% respectively, while gypsum was detected only in

Sandovales with 0.33% (Table 2.3).

Oades (1985) reported that increasing the OC from <1% up to

1.5% caused an increase in aggregate stability. Considering the importance of OC in the formation of stable aggregates, a relationship was observed between the OC content in the soils of Louisiana and the WDCI.

As presented early in this chapter, the WDCI in the Gigger soil was 0.5, while in the Olivier and Coteau, the WDCI was 0.18 and 0.21, respectively. The higher WDCI found in the Gigger soil coincides with its low OC content (0.27), indicating that the absence of OC could be responsible for the increased WDCI.

In the case of the soils of Mexico, OC was 1.15, 0.33, and 1.76% in Asogueros, Sandovales, and Reforma respectively. Although the soils of

Mexico have similar OC values to that in the soils of Louisiana, there was no relationship between OC and WDCI. This suggests that some other factor, perhaps the high SAR content of the soils of Mexico relative to soils of Louisiana has a stronger influence on the dispersion of these soils than any cementing effect linked to OC. 35

Dispersion factors

The concentration of cations and anions in saturated paste extracts exemplify the contrast in weathering conditions and parent material between soils of Mexico and Louisiana (Table 2.4). Soils from the north- central part of Mexico are subject to insufficient rainfall to leach base cations (Ca, Mg, Na, and K). In turn, the soils of Louisiana are subjected to a more abundant rainfall and consequently to more effective leaching of bases, while substantial accumulation of salts occurs in the soils of

Mexico. As a result of accumulation of Na in the soil solution and on exchange sites, the SAR values in the soils of Mexico (3.81 to 4.57) are higher than those in the soils of Louisiana (0.29 to 0.73) (Table 2.5).

Comparing SAR with the WDCI of each soil, a positive relationship is apparent. Low SAR and WDCI values characterize the soils of Louisiana, while high SAR and WDCI values represent the soils of Mexico. A similar relationship between EC and WDCI was observed.

Another consequence of the contrasting weathering conditions prevailing in soils of Mexico and Louisiana and related to soil dispersion is pH (Table 2.5). The pH in the soils of Louisiana and the Sandovales soil are more acid than in the Asogueros and Reforma soils. The acidity in soils of Louisiana results most likely from the loss of base cations through leaching. In Sandovales the acidity could be a consequence of continuous application of NH4-N fertilizer. An acidifying effect in soils after several Table 2.4. Cation and anion concentrations in saturated paste extracts of soils of Mexico and Louisiana.

Soil Na Mg Ca K Cl U) o n o 3 H c c y

lllillUI/L Gigger 0.96 1.00 2.30 0.83 1.68 0.63 0.73 4.72

Olivier 1.00 0.51 1.41 1.00 1.04 0.83 0.04 3.08 Coteau 0.69 0.84 4.96 0.62 0.66 0.45 0.05 11.30

Asogueros 14.47 1.70 9.86 2.72 2.77 10.53 2.61 13.84

Sandovales 17.65 3.04 8.83 3.73 3.37 19.55 0.05 2.44

Reforma 11.90 0.71 8.36 1.10 4.40 8.71 0.21 9.52 t Calculated as difference between (Ca + Mg + Na + K) - (Cl + S04 + N03). 37

Table 2.5. Chemical soil properties of soils of Mexico and Louisiana.

Soil pHk PHW EC' SAR” dS/m

Gigger 5.48 6.05 0.74 0.53 Olivier 5.30 5.27 0.39 0.73 Coteau 6.75 7.25 0.79 0.29

Asogueros 7.29 7.89 2.83 3.98 Sandovales 5.00 5.74 3.06 4.57 Reforma 7.27 8.11 2.23 3.81 * = Electrical conductivity (EC) ** = Sodium adsorption ratio (SAR) pHk =1:1, soil:1M KCI pHw = 1:1, soihdistilled water 38 years of applying NH4-N fertilizer was reported by Brown and Mahler

(1988). However, the pH below the surface horizon in Sandovales is reported to be 6.8 (CETENAL 1971a).

The relatively high SAR and alkaline pH favor chemically dispersive conditions in these soils.

Mineral characteristics

Clay mineral composition

Kaolinite, smectite, and illite were identified in the fine and coarse fractions in all six soils (Table 2.6). Vermiculite was present in all of the soils in the coarse fraction and in the fine fraction of the Gigger and Olivier.

Interstratified clays were identified in all of the soils by a broad peak between 17 and 27 A. Diffraction peaks at 27 and 24 A corresponding to illite/smectite and illite/vermiculite interstratified clays, respectively

(Sawnhey 1989), were clearly identified in the fine clay fraction of the

Reforma soil.

Clay composition has been pointed out as a major factor affecting crust formation, especially the presence of smectite (Stern et al., 1991).

Smectite was present in all of the soils (Table 6); however, its relative quantity was not correlated to the WDCI of each soil. For example, the

Reforma soil had a WDCI of 0.79 which was similar to that in Sandovales, however the amount of smectite in Sandovales was approximately one- third that in the Reforma. There was no correlation between the range of Table 2.6. Clay mineral composition of soils of Mexico and Louisiana.

Soil Kaolinite lllite Vermiculite Smectite Interstratified

UUdloU uldy

Gigger XXX XXX X XX X

Olivier XXX XXX X XX X

Coteau XXX XXX X XX X

Asogueros X x x x x XXX

Sandovales XXX XX X XX X

Reforma X XXX X XX X

rme uidy

Gigger XX XX X XXX X

Olivier XX XX X XXX X

Coteau XX XX - x x x x X

Asogueros X XX - XXX XX Sandovales xxxx X - XX Reforma X XXX - xxxx XX Relative quantities: XXXX>40%, XXX 25-40%, XX 10-25%, X<10%. Estimates were derived from relation peak areas percentage. 40

WDCI in soils of Mexico and Louisiana and the amount of smectite, given that smectite was present in similar quantities in all of the soils. These results suggest that other chemical characteristics are responsible for the differences in WDCI between soils of Mexico and Louisiana.

Powder X-ray diffraction

The minerals identified using powder X-ray diffraction are shown in

Table 2.7. Calcite was detected in Asogueros and Reforma soils while gypsum was present only in Sandovales. These results correspond with the chemical identification of calcite and gypsum as potential cementing agents. Cristobalite was detected in all of the soils of Mexico. Quartz, K- feldspar, and Na-rich plagioclase were identified in all six soils. The cristobalite identified in the soils of Mexico is the most apparent link to the higher amorphous Si contents found in these soils. Furthermore, the solubility of cristobalite is greater than quartz (Lindsay, 1979), which would allow it to play a more active role as a cementing agent in the soils of Mexico.

Summary and Conclusions

The objective of this study was to characterize the physical, chemical, and mineralogical properties of six soils of Mexico and Louisiana as related to crust formation. The contrasting parent materials determined the soil texture in the soils of Mexico and Louisiana. Although smectite Table 2.7. Minerals identified using powder X-ray diffraction in soils of Mexico and Louisiana.

Soils Quartz Oligoclase Orthoclase Calcite Gypsum Cristobalite

Gigger X X X - -- Olivier X X X- --

Coteau XXX- -- Asogueros X X X X -X Sandovales X XX-XX

Reforma XXX X -X 42 was a major component in all six soils, the higher WDCI in the soils of

Mexico (0.79 to 0.83) compared to that in the soils of Louisiana (0.18 to

0.5) is attributed to the relatively higher SAR values of the soils of Mexico.

In the soils of Louisiana, the higher WDCI of the Gigger (0.5) with respect to that in the Olivier and Coteau (0.18 and 0.21, respectively), is attributed to its low OC content (0.27%). The higher OC content in the Olivier and

Coteau (1.02 and 1.19%, respectively) is believed to be responsible for a higher degree of particle aggregation in these soils.

The differences in parent material between the soils of Mexico and

Louisiana are believed to be responsible for the contrasting amount of free

Fe oxides and amorphous Si and Al which could serve as potential cementing agents in these soils. The free Fe oxide contents were higher in the soils of Louisiana (0.19 to 0.49%) than in the soils of Mexico (0.02 to

0.08%). Amorphous Si and Al contents were 3 and 4 times higher in the soils of Mexico than in soils of Louisiana. Calcite was present only in the

Asogueros and Reforma, 3.9 and 11.4% respectively. Gypsum was only detected in the Sandovales at 0.33%.

Based on these results, the soils of Mexico have a greater probability of developing dense crusts than the soils of Louisiana. The amount and type of potential cementing agents were related to the contrast in weathering conditions and parent material existing between the arid north- central part of Mexico and the humid climate of Louisiana. Calcite, gypsum 43 and amorphous Si and Al may play an important role in the cementation of the soils of Mexico soils while free Fe oxides may be a potential cementing agent in the soils of Louisiana crusts.

References for Chapter 2

Agassi, M., J. Morin, and I. Shainberg. 1985. Effect of raindrop impact energy and water salinity on rates of sodic soils. Soil Sci. Soc. Am. J. 49:186-190.

Alexiades, C.A., and M.L. Jackson. 1967. Quantitative clay mineralogical analysis of soils and sediments. Clays and Clay Minerals. 14:35-52.

Baver, L.D., W.H. Gardner, and W.R. Gardner. 1972. . Jhon Wiley and Sons, Inc., New York.

Bhon, H., B. McNeal, and G. 0 Connor. 1985. . 2nd. ed. Wiley-lnterscience Publications, NY, 341 p.

Brown, T.H. and R.L. Mahler. 1988. Effects of phosphorous and acidity levels of silica extracted from a Palouse Silt Loam. Soil Sci. Am. J. 51:674-677.

CETENAL. 1971a. Carta edafologica 1:50 000, Aguascalientes, F13-D- 19. Mexico, D.F. Mapa.

CETENAL. 1971b. Carta edafologica 1:50 000, Salinas de Hidalgo, F14- A-61. Mexico, D.F. Mapa.

Ferry, D.M. and R.A. Olsen. 1974. Orientation of clay particles as it relates to crusting of soil. Soil Sci. 4:367-375.

Gee, G.M., and J.W. Bauder. 1986. Particle size analysis, in. A. Klute (ed.) Methods of Soil Analysis, Part 1. Physical and Mineralogical Properties. 2nd. ed. Agronomy 9:383-411. Am. Soc. Agron. Madison, Wl.

Jackson, M.L. 1975. Soil Chemical Analysis. Advanced Course. 2nd ed., 10th printing. Published by the author, Madison, Wl. 895 p. 44

Jenny, H. 1941. Factors of soil formation. A System of Quantitative . McGraw-Hill, NY, 281 p.

Kunze, G.W. and J.B. Dixon. 1986. Pretreatment for mineralogical analysis, in. A. Klute (ed.) Methods of Soil Analysis, Part 1. Physical and Mineralogical Properties. 2nd. ed. Agronomy 9:383-411. Am. Soc. Agron. Madison, Wl.

Lemos, P. and J.F. Lutz. 1957. Soil and some factors affecting it. Soil Sci. Soc. Am. J. 21:485-491.

Lupercio-Huerta, F.J. 1987. Identification de factores en la formacion de costras en suelos del altiplano Potosino-Zacatecano. (Identification of factors in soil crust formation in the Potosino-Zacatecano plateau soils). Tesis Profesional. ITTA No. 20. Aguascalientes, Mexico. 94pp.

Mehra, O.P., and M.L. Jackson. 1960. Iron oxide removal from soils and clays by dithionite-citrate system with sodium bicarbonate. Clays and Clay Minerals. 7:317-327.

Miller, B.J., G.C. Lewis, J.J. Alford, and W.J. Day. 1984. in Louisiana and at Vicksburg, Mississippi. Guidebook for the friends of the Pleistocene field trip. Dept, of Agronomy, Louisiana State Univ., Baton Rouge, LA., 121 p.

Miller, B.J., B.A. Schumacher, G.C. Lewis, J.A. Rehage, and B.E. Spicer. 1988. Basal mixing zones in loesses of Louisiana and Idaho: II. Formation, spatial distribution, and stratigraphic implications. Soil Sci. Soc. Am. J. 52:759-764.

Oades, J.M. 1989. An introduction to organic matter in mineral Soils. ]n. J.B. Dixon and S.B. Weed. (eds). Minerals in soil environments. 2nd ed. SSSA, Madison, Wl.

Prince, A. L. 1955. Methods in soil analysis, im F.E. Bear (ed.). Chemistry of the Soil. Reinhrold Pub. Co., New York.

Shainberg, I., and J. Letey. 1984. Effect of electrolyte concentration on the hydraulic properties of depositional crust. Soil Sci. Soc. Am. J. 49:1260-1263.

Simonson, R.W. 1959. Outline of a generalized theory of soil genesis. Soil Sci. Soc. Am. Proc., 23:152-156. 45

Singer, M. J., and D.N. Warrington. 1991. Crusting in the western United States. ]n M.E. Sumner and B.A. Stewart (eds). Soil Crusting-Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 179-204.

Dance, E., B.J. Griffis, B.B. Nutt, and A.G. White. 1968. of east Baton Rouge Parish, Louisiana. U.S. Govt. Printing Office, Washington, D.C.

Soil Survey Staff. 1981. Soil survey of Franklin Parish, Louisiana. U.S. Govt. Printing Office, Washington, D.C.

Murphy, K.E., J. K. Vidrine, and D. R. McDaniel.. 1986. Soil survey of St. Landry Parish, Louisiana. U.S. Govt. Printing Office, Washington, D.C.

Martin E. Ch., L.J. Trahan, and T. Midkiff. 1984. Procedure for collecting soil samples and methods of analysis for soil survey. Soil Survey Investigations Report No. 1. U.S. Gov. Print. Office, Washington, D.C.

Stern, R., M. Ben-Hur, and I. Shainberg. 1991. Clay mineralogy effect on rain infiltration, seal formation and soil losses. Soil Sci. 152:455-462.

U.S. Salinity Laboratory Staff. 1969. Saline and alkali soils. Agriculture Handbook No. 60. U.S. Gov. Print. Office, Washington, D.C. CHAPTER 3

MORPHOLOGICAL CHANGES DURING CRUST FORMATION IN SOILS OF MEXICO AND LOUISIANA

Introduction

The dispersion of soil aggregates by physical and or chemical factors during soil crust formation results in an alteration of the soil surface morphology. Porosity, bulk density, particle size distribution, and particle orientation are reported as the morphological features of surface soil most affected during crust formation (Chen et al., 1980; Onofiok and Singer,

1984; Gal et al., 1984; Remley and Bradford, 1988; and Bresson and

Boiffin, 1990).

Crust formation occurs as a consequence of dispersion of the surface layer (McIntyre, 1958; Morin et al., 1981; and Stern et al., 1991).

The use of soil conditioners such as gypsum, phosphogypsum, and organic polymers has been tested in an attempt to stabilize soil aggregates and prevent crust formation (Gal et al., 1984, Shainberg et al., 1989; and Ben-

Hur, et al., 1990). The use of vegetative covers or organic mulches has also been reported as an effective way to prevent soil dispersion caused by raindrop impact (Wilson, 1982; McVay et al., 1989; and Bruce et al.,

1990). The effectiveness of these practices has been evaluated in terms of soil physical parameters such as infiltration (Morin and Benjamini, 1977;

Miller, 1987; and Smith et al., 1990), crust strength (Bradford, 1982 and

46 47 le Bissonnais, 1989), and seedling emergence (Goyal, 1981 and Aujla et al., 1986).

A gap exists in the understanding of crust formation in terms of morphological changes as related to the use of different practices. Another gap which has been left in investigations addressing crust formation is the identification of cementing agents. The use of the scanning electron microscope (SEM) to describe crust morphology has been described as a more effective technique than the petrographic microscope by Chen et al. (1980) and Onofiok and Singer (1984). One option to detect the presence and chemical composition of cementing agents is the combined use of SEM and XRF. These combined techniques have been reported as one of the best approaches for obtaining the qualitative chemical composition of unknown materials (Sawhney, 1986).

The objectives of this study were: i) to identify changes in the morphology of the soil surface during crust formation and ii) to determine the elemental composition of cementing agents in the crusts of soils from

Louisiana and Mexico using SEM/XRF.

Materials and Methods

Soil series

The surface horizons from three soils from Louisiana: Gigger (fine- silty, mixed, Typic Fragiudalf), Olivier (fine-silty, mixed, thermic, Aquic 48

Fragiudalf), and Coteau (fine-silty, mixed, thermic, Glossaquic Hapludalf);

and three soils from the north-central part of Mexico: Asogueros (loamy,

mixed, hypothermic, Vertic Nadurargid), Sandovales (loamy, mixed,

hypothermic, Abruptic Aridic Durustoll), and Reforma (loamy, mixed,

hypothermic, Lithic Calciorthid) were used in this study.

Soil preparation

Soil samples were collected from the plow layer (0 to 20 cm) of each

soil, air dried, and sieved through a 2 mm sieve. Soils were packed in

wooden boxes (30 x 30 x 7 cm). The bottom of the boxes were

perforated and covered with a 1 cm layer of coarse sand to improve soil drainage. Dry soil was uniformly packed into the boxes and lightly agitated to attain reproducible bulk densities. The wooden boxes were positioned at a slope of 0.2% to allow during rainfall simulation.

Bulk density

Bulk density was determined for each soil on uniformly packed boxes. The volume and weight were measured for each soil and a density value calculated (g/cm3).

Rainfall simulation

Soils were placed under a modified, rotating-disk rainfall simulator described by Morin et al. (1966). The terminal velocity for drops with a medium diameter of 2.0 mm was 6.4 m/s, using a relationship between terminal velocity (m/s) and drop diameter (mm) reported by Laws (1941). 49

The rainfall intensity was 54 mm/h and the kinetic energy was 26.97 J/m 2- mm, obtained using the method reported by Wischmeier and Smith (1978).

A detailed description of the rainfall simulator can be found in Appendix A.

Prying period

Following the rainfall simulation, soil drying was achieved by placing infrared light bulbs 100 cm above the soil surface. A diurnal effect was simulated by alternating on-off periods of 12 hr for 7 days.

Management practices

The management practices evaluated were: bare, protected and gypsum amended soils. The protected soil was covered with a 2-mm mesh. In the gypsum-amended soil an application corresponding to 5 ton/ha of gypsum was spread on the surface. The amount of gypsum used was based on a recommendation given by Ben-Hur et al. ( 1992).

Sampling zone

Soil samples (2.5 x 2.5 cm) were taken from the upper 2-cm of each simulated crust for SEM observations and thin section preparations.

Samples were collected at two stages to provide evidence of morphological changes occurring during crust formation: 1) just before runoff started or ponding time, and 2) at the end of the drying period.

Thin section preparation

Thin sections were prepared to obtain small scale observations of changes in the soil surface caused by raindrop impact. Surface samples 50

were taken at the two stages of crust formation described above, air dried,

and impregnated with a mixture of Araldite resin and hardening solution

(7:1 ratio). The impregnation of samples was performed under a 300 watt

light bulb in order to heat the sample to 60°C decreased the viscosity of the resin. After the sample was covered with resin, vacuum was applied to expel air from the soil sample and to allow better impregnation. Additional

resin was added to completely cover the impregnated sample. Vacuum was applied after each addition of resin to the sample.

Cured samples were sliced using a diamond blade to obtain 1- to 2- cm thick sections having a vertical orientation. One side of the sample was then polished and glued to a glass slide. Samples were then sliced again

and mechanically ground to obtain a 2-mm thick soil section. Manual

grinding was used to obtain thin sections approximately 30 //m in thickness. Finally, thin sections were mechanically polished and thinned to the point where polarized light passing through quartz grains turned from

black to white.

Scanning electron microscope and XRF analyses

Soil samples were coated with carbon and gold. A Denton Vaccum

DV-502A and a Sputter coater were used to coat the soil samples with carbon and gold respectively. A JEOL T-300 SEM was used in scanning the morphological features of crusts. Chemical composition of cementing agents was assessed by energy dispersive XRF. Micrographs were taken 51 of filaments bridging particles and coating features. The following criteria were used for interpreting mineral composition of cementing agents: i) the presence of amorphous Si was assumed when only Si was detected in masses having an amorphous appearance, ii) the presence of free Fe oxides was assumed when only Fe was detected, iii) calcite was assumed to be present when only Ca was detected (C or CO3 can not be detected due to the low energy release during excitation), and iv) gypsum was assumed to be present when Ca and S coincided at the same point analyzed.

Discussion of Results

The effect of management practices on morphological features at two stages of crust formation is presented first. A discussion of the elemental composition of cementing agents identified as bridging soil particles together in the crust follows. Because of the similarity in the morphological features present in the crust of soils of Louisiana, only the morphology of the Gigger crusts will be described. In the soils of Mexico,

Asogueros and Reforma produced quite similar soil crusts, but they were different from those in the Sandovales. For this reason, Asogueros and

Sandovales soil crusts will be discussed in connection with the soils of

Mexico. Bulk densities of 1.30, 1.35, 1.30, 1.25, 1.30, and 1.20 g/cm3 52

were obtained for Gigger, Olivier, Coteau, Asogueros, Sandovales, and

Reforma, respectively.

Gigger soil crusts

During the early stage of crust formation at ponding time, the main

morphological feature of the Gigger bare soil was a cratered o pitted

surface (Figure 3.1). This surface alteration is an example of what happens

in the field, especially at the beginning of the growing season where

unprotected soil is subject to direct raindrop impact. Raindrops act as small bombs, splashing soil particles, and forming crater-like structures

(Figure 3.1). Because of the continuous surface alteration, there was no accumulation of fine particles in the surface. The soil surface was not compacted, consequently, porosity was similar in both the surface and subsurface. No orientation of particles nor an upper layer formed by clean

grains was differentiated at this stage in the soil surface (Figure 3.2).

At the final stage of crust formation or end of the drying period, the soil surface did not have the uneven microtopography observed at ponding time (Figure 3.3). A compacted 0.1 mm thick layer, composed of clean

grains of uniform size with a single grain structure and moderate particle orientation, formed at the surface. This layer was the major morphological feature observed at the end of the drying period in contrast with the

morphology at ponding time. The formation of this layer in the crust was

reported by Mclntere (1958), Tarchitzky et al. (1984), and Onofiok and 53

Figure 3.1. Micrograph of a vertical thin section with plain light showing a cratered surface (A) caused by raindrop impact in the bare Gigger soil at ponding time.

Figure 3.2. An SEM micrograph, vertical section, showing the homogeneous distribution of particles in the surface (A) and subsurface (B) in the bare Gigger soil at ponding time. 54

Singer (1984). This layer is believed to result from the continuous beating

of raindrops on the soil surface and the removal of clay-sized particles by

runoff or infiltration. In the subsurface (Figure 3.3), clay-sized particles are

more abundant than in the upper layer. Because of surface compaction

due to rainfall impact and particle orientation, macroporosity in the upper

0.1 mm layer is reduced, compared to the large voids and macropores

present below this layer (Figure 3.3).

The morphological features in the protected soil at ponding time

reveal a flat, undisturbed soil surface (Figure 3.4). Soil aggregation can be

observed in the surface and subsurface layers. Large voids are apparent

below the 0.2-mm upper layer (Figure 3.5).

At the end of the drying period, the protected soil of the Gigger did

not develop a compact 0.1-mm upper layer composed of single-grain

particles as in the bare soil. The greater soil erosion observed in the bare

soil (172 g/m2), compared to that in the protected soil (74.17 g/m2), (Table

3.1), supports the argument that the compacted, single-grain particle layer

in the surface of the bare soil formed as a result of raindrop impact with the removal of clay-sized particles by runoff (Figure 3.6). The amount of clay-sized particles lost in runoff in the bare Gigger soil was 34% (Table

3.2). Protection of the soil surface avoided removal of clay-sized particles from the surface by runoff. The amount of clay-sized particles lost in runoff was only 9%. Instead, sealing of the soil surface appears to have Figure 3.3. An SEM micrograph, vertical section, showing a compacted 0.1 -mm surface layer (A) formed of single, clean grains in the bare Gigger soil at the end of the drying period.

Figure 3.4. Micrograph of vertical thin section with plain light, showing a flat-undisturbed soil surface (A) in the protected treatment of the Gigger soil at ponding time. Table 3.1. Effect of management practices on soil erosion in soils of Mexico and Louisiana. Soil Management Sediment Series Practices Loss (g/m2) Gigger Bare 172.34 Gypsum 105.00 Protected 74.17

Olivier Bare 174.44 Gypsum 113.65 Protected 74.37

Coteau Bare 204.13 Gypsum 127.26 Protected 84.66

Asogueros Bare 233.06 Gypsum 172.26 Protected 112.26

Sandovales Bare 376.82 Gypsum 223.38 Protected 123.28

Reforma Bare 246.63 Gypsum 181.97 Protected 146.67 Figure 3.5. An SEM micrograph, vertical section, showing soil aggregates in the surface (A) of the protected Gigger soil at ponding time.

100 um

Figure 3.6. An SEM micrograph, vertical section, showing the homogeneous particle size distribution in the surface (A) and subsurface (B) in the protected Gigger soil at the end of the drying period. 58

Table 3.2. Particle size distribution of sediment loss in the soils of Mexico and Louisiana. Soil Surface Sand Silt Clay Series Management ------% ------Gigger Bare 1 65 34 Gypsum 1 68 31 Protected 2 89 9

Olivier Bare 4 81 15 Gypsum 6 81 13 Protected 6 92 2

Coteau Bare 2 92 6 Gypsum 3 91 6 Protected 1 96 3

Asogueros Bare 5 65 30 Gypsum 3 69 28 Protected 3 67 20

Sandovales Bare 55 30 15 Gypsum 51 35 14 Protected 38 50 12

Reforma Bare 17 49 34 Gypsum 12 61 27 Protected 5 70 25 59 occurred. These observations suggest that chemical dispersion is important in the crust formation of the protected soil as the soil surface was not dispersed by rainfall. Dispersion most likely occurred because of chemical and mineralogical soil properties.

The morphology of the soil surface amended with gypsum shows an increase in aggregation at ponding time (Figure 3.8). This is attributed to the flocculating effect of Ca coming from the dissolution of gypsum. Even though the surface was exposed to the beating action of raindrops, soil aggregation occurred, reducing the disruptive effect of raindrops. As a result, the surface in this soil does not show the dispersion observed in the protected soil (Figure 3.7). Porosity and particle size are uniform in the upper 0.4 mm of the crust. Aggregation was improved compared to that in the bare and protected soils (Figure 3.8).

At the end of the drying period of the gypsum-amended soil, a compact 0.1 mm thick layer, composed of clean grains of uniform size with a single grain structure and moderate particle orientation, was formed

(Figure 3.9). Below this washed out layer, clay-sized material coating large particles and forming aggregates can be observed. As discussed above, it appears that clay-sized particles in the upper 0.1 mm layer of the immediate surface, were dispersed by the beating action of raindrops and removed by runoff (Figure 3.9). Erosion in this soil was 105 g/m2 which is intermediate between the erosion observed in the bare and protected soils 60

tW t& 'i

Figure 3.7. Micrograph of a vertical thin section with plain light, showing the effect of raindrop impact in the surface (A) of the gypsum- amended treatment of the Gigger soil at ponding time.

Figure 3.8. An SEM micrograph, vertical section, showing the aggregation effect (A) of gypsum in the gypsum-amended treatment of the Gigger soil at ponding time. 61

(Table 3.1). Clay-sized particles lost in runoff in the gypsum-amended soils was 3% (Table 3.2).

Asoqueros soil crusts

The soil surface of the Asogueros soil was strongly altered by the beating effect of raindrops at ponding time (Figure 3.10). The soil surface was completely sealed by clay-sized which were coating larger particles.

Even though this soil had a coarser texture than that in the soils of

Louisiana, no clean sand grains could be observed as dispersed clay-sized particles formed a massive structure. Below the upper 0.1 mm layer, macropores are more abundant (Figure 3.11).

At the end of the drying period, the crust morphology was very similar to that described at ponding time. No clean grains were observed in the soil surface because of the high dispersivity of this soil. The highly dispersive nature of the soils of Mexico, in general is reflected by their relatively high WDCI as compared to those of the soils of Louisiana (Table

3.3).

In the protected Asogueros soil, there was no alteration of the soil surface by the impact of raindrops at ponding time (Figure 3.12).

Dispersed clay-sized material seal the soil surface, clogging soil pores. No macropores are present in the upper 0.1 mm of the crust (Figure 3.13). At the end of the drying period, clay-sized particles sealed the surface (Figure

3.14). No clean grains are present in or below the soil surface. The 62

Table 3.3. Water-dispersible clay index in the soils of Mexico and Louisiana. -dispersible clay index Series

Gigger 0.50 Olivier 0.18 Coteau 0.21 Asogueros 0.83 Sandovales 0.79 Reforma 0.79 Figure 3.9. An SEM micrograph, vertical section, showing an upper 0.1 mm layer surface layer (A) formed of single, clean grains in the gypsum-amended Gigger soil at the end of the drying period.

Figure 3.10. Micrograph of a vertical thin section with plain light, showing a cratered surface (A) caused by raindrop impact in the bare Asogueros soil at ponding time. Figure 3.11. An SEM micrograph, vertical section, showing a homogeneous particle-size distribution in the surface (A) and subsurface (B) in the bare Asogueros soil at ponding time.

Figure 3.12. Micrograph of a vertical thin section with plain light, showing an unaltered surface (A) in the protected Asogueros soil at ponding time. 65 dispersed soil particles formed a massive structure (Figure 3.14). As discussed in the protected soil of the Gigger soil, clay-sized particles accumulated in the soil surface. This effect was also reflected in erosion data of the Asogueros where the bare soil yielded a sediment loss of 233 g/m2 compared to only 113 g/m2 in the protected soil.

The soil surface of the gypsum-amended Asogueros soil was altered by the beating action of raindrops at ponding time (Figure 3.15). There is evidence in the surface and subsurface of soil aggregation in the gypsum- amended soil. Clay-sized coated larger particles at ponding time and at the end of the drying period. Porosity was higher in the subsurface than in the surface at both stages (Figure 3.16).

Sandovales soil crusts

In the Sandovales bare soil, the surface was altered by the impact of raindrops at ponding time (Figure 3.17). Clay-sized material filled soil pores in the surface and partially coated large grains. The presence of pores in the upper 0.2-mm suggests that the soil surface was not compacted by raindrops at this stage (Figure 3.18).

A compact layer was formed in the upper 0.2 mm at the end of the drying period in the bare soil. The dominant particle size fraction in this layer were fine sand and coarse and fine silt (as estimated from SEM micrograph). Clay-sized material is more abundant in the subsurface and occurs in aggregates. Porosity is higher in the subsurface than in the upper 66

Figure 3.13. An SEM micrograph, vertical section, showing a homogeneous particle-size distribution in the surface (A) and Subsurface (B) in the protected Asogueros soil at ponding time.

Figure 3.14. An SEM micrograph, vertical section, showing clay-size particles sealing the surface (A) in the protected Asogueros at the end of the drying period. 67

i& r

Figure 3.15. Micrograph of a vertical thin section with plain light, showing raindrop impact (A) on the surface of the gypsum-amended Asogueros soil at ponding time.

100 uml!F

Figure 3.16. An SEM micrograph, vertical section, showing the aggregation effect in the surface (A) and subsurface (B) in the gypsum-amended Asogueros soil at ponding time. Figure 3.17. Micrograph of a vertical thin section with plain light, snowing the raindrop impact on the soil surface (A) in the bare Sandovales soil at ponding time.

Figure 3.18. An SEM micrograph, vertical section, showing clay-size particles sealing the surface (A) in the bare Sandovales soil at ponding time. 69 layer, which was apparently compacted by the impact of raindrops on the surface (Figure 3.19).

At ponding time in the protected Sandovales soil, no disturbance of the soil surface by raindrops was observed, similar to the bare soil (Figure

3.20). Clay-sized particles accumulated in the soil surface, clogging soil pores and sealing the soil surface. The soil surface was not compacted, as a result porosity is similar in the surface and subsurface (Figure 3.21).

At the end of the drying period of the Sandovales protected soil, dispersed clay-sized material can be observed to coat large particles without evidence of an upper layer of clean grains. Clay-sized particles are more concentrated in the crust at this stage, compared to that at ponding time, further clogging and sealing the crust. Porosity is reduced in the upper 0.2-mm layer compared to the subsurface (Figure 3.22). The accumulation of clay in the immediate surface is again reflected in the erosion of this soil which was reduced 67% compared to the erosion in the bare soil.

In the case of the gypsum-amended Sandovales soil, even though the inmediate soil surface does not show cratering, close inspection reveals cratering in the subarface zone that has subsequently been buries by particles previously suspended by rainfall impact (Figure 3.23). Soil aggregation was improved in the soil surface and subsurface of this soil Figure 3.19. An SEM micrograph, vertical section, showing a compacted 0.2-mm surface layer (A) in the bare Sandovales soil at the end of the drying period.

Figure 3.20. Micrograph of vertical thin section with plain light, showing an undisturbed soil surface in the protected Sandovales soil at ponding time. Figure 3.21. An SEM micrograph, vertical section, showing accumulation of clay-size particles in the surface (A) of the protected Sandovales soil at ponding time.

Figure 3.22. An SEM micrograph, vertical section, showing a homogeneous distribution of clay-size particles in the surface (A) and subsurface (B). 72

(Figure 3.24), compared to that in the bare soil, and porosity was similar in both the upper 0.2-mm and subsurface.

At the end of the drying period of the gypsum-amended soil, a contrast was observed between this soil treatment and the same treatment in the Gigger soil. The soil surface was not compacted as in the Gigger and there was no accumulation of clean grains at the surface (Figure 3.24).

Aggregation was maintained in the surface and subsurface. Porosity is similar in the immediate surface and subsurface layers (Figure 3.24).

Cementing agents

Calcium and S were detected by XRF as the only chemical components in bridges bonding soil particles in the bare soil of the

Asogueros soil (Figure 3.25). The sharp Ca and S peaks and the absence of Si and Al peaks in the XRF analysis (Figure 3.27), suggests the presence of gypsum as a cementing agent. Gypsum bridges were also detected in the protected soil of this soil. Particles bridged by gypsum did have Si and

Al peaks, characteristic of aluminosilicate minerals (Figure 3.28). In this soil, no gypsum was detected in the original soil by chemical analysis indicating that the gypsum bridges detected in the crust formed as a result of the accumulation of Ca and S 04 in the crust during evaporation.

Gypsum was also found as a cementing agent in the bare and protected soils of the Reforma soil. As in the Asogueros, no gypsum was found in 73

Figure 3.23. Micrograph of a vertical thin section with plain light, showing raindrop impact on the surface (A) in the gypsum-amended Sandovales soil at ponding time.

lOOunrVt

Figure 3.24. An SEM micrograph, vertical section, showing the aggregation effect (A) of gypsum in the gypsum-amended treatment of the Sandovales soil at the end of the drying period. 74 the initial chemical analysis. It is believed that in both cases gypsum formed due to an increase in Ca and S04 in the crust during evaporation.

In the bare soil of the Sandovales, the XRF analysis detected coatings formed only of Fe (Figure 3.29). Soil grains were covered and bridged by Fe coatings, reducing soil porosity compared to areas without coatings (Figure 3.26). The XRF analysis of soil particles in areas of the crust without coatings revealed the presence of Si and Al, with no Fe being detected (Figure 3.30). The presence of free Fe oxides as a cementing agent in the Sandovales was anticipated because of its acidic pH. In addition to the Fe bridges, gypsum was also detected as a cementing agent in the bare and protected soils of this soil with a similar morphology to those already described in the Asogueros soil. Native gypsum was present in the initial soil analysis so that this cementing agent was detected in almost all of the soil crust samples analyzed by SEM/XRF.

In the bare soil of the Reforma, only a Si peak was detected by XRF analysis in massive structures coating soil particles (Figures 3.31 and

3.33). The absence of any other chemical element in the XRF analysis suggests the presence of amorphous Si as a cementing agent. Silica and

Al were detected by XRF analysis in the soil particles surrounding the spot with only Si (Figure 3.34). To detect Si was no easy task because clay­ sized soil particles were usually coating or covering soil particles in the Figure 3.25. An SEM micrograph showing natural gypsum (A) bridging soil particles (B) in the bare Asogueros soil.

100Turrl % ' Figure 3.26. An SEM micrograph showing coatings composed of Fe (A) in the bare Sandovales soil. 76

S Co

Au

keV Figure 3.27. Sulfur and Ca peaks detected in the Asogueros bare soil by XRF, indicate the presence of gypsum as a cementing agent in the crust.

4

Ca Au

Fe

keV Figure 3.28. Aluminum and Si peaks detected by XRF in soil particles bound by gypsum in the crust of the Asogueros bare soil. Figure 3.29. Iron peak detected by XRF in the Sandovales bare soil, suggests the presence of free Fe oxides as a cementing agent in the crust.

2

Au

Au

Fe Ca

keV Figure 3.30. Aluminum and Si peaks detected by XRF in soil particles coated by free Fe oxides in the crust of the Asogueros bare soil. 78 crust, contaminating the sample with Al and Si peaks from fine aluminosilicates and clay-sized soil particles.

In the case of the soils of Louisiana soils, calcite was observed in the bare soil of the Gigger soil, coating and bridging soil particles in the crust

(Figure 3.32). Presence of calcite was recognized when the XRF analysis detected only a Ca peak without any other element (Figure 3.35). Discrete particles of Si and Al were also identified present in the XRF analysis of the crust (Figure 3.36). The presence of Ca, however, was detected only in small areas of the crust surface (Figure 3.32).

In the bare soil of the Olivier, only Fe was detected as the chemical component in filaments bridging soil particles in the bare soil (Figures 3.37 and 3.39). Silica and Al were detected by XRF where these filaments joined to soil particles (Figure 3.40). In the bare soil of the Coteau, Si was detected by XRF analysis as the chemical component of a filament bridging soil particles (Figures 3.38 and 3.41). Silica and Al were present in particles joined by this filament (Figure 3.42), while only Si was present in the filament. These bridges are suspected of being due to the fibrous growth habit not being characteristic of a crystalline phase.

Summary and Conclusions

The objectives in this study were i) to identify changes in the morphology of the soil surface during crust formation and ii) to determine Figure 3.31. An SEM micrograph showing massive coatings of Si (A) in the bare Reforma soil.

Figure 3.32. An SEM micrograph showing calcite coating (A) soil particles in the bare Gigger soil. 80 i a

Ca

koV Figure 3.33. Siiice peak detected in the Reforma bare soil, suggests the presence of amorphous Si as a cementing agent in the crust.

4

Au

Au

keV Figure 3.34. Aluminum and Si peaks detected by XRF in soil particles coated by amorphous Si in the crust of the Reforma soil. 81 1357

Au

■V * U -* ■ * n V ■*-»> L

Figure 3.35. Calcium detected by XRF in the Gigger bare soil, suggests the presence of calcite as a cementing agent in the crust.

Figure 3.36. Aluminum and Si peaks detected by XRF in soil particles caoted by calcite in the crust of the Gigger bare soil. Figure 3.37. An SEM micrograph showing filaments composed of he (A) bridging particles (B) in the bare Olivier soil.

Figure 3.38. An SEM micrograph showing a filament composed of Si (A) bridging particles (B) in the bare treatment of the Coteau soil. 83

Au

2.00 keV Figure 3.39. Iron peak detected by XRF in the Olivier bare soil, suggests the presence of free Fe oxides as a cementing agent.

t

Au

koV Figure 3.40. Aluminum and Si peaks detected by XRF in soil particles bound by free-oxide bridges in the crust of the Olivier bare soil. 84 7

Au

Au

6.00

Figure 3.41. Silice peak detected by XRF in the Coteau bare soil, suggests the presence of amorphous Si as a cementing agent in the crust.

2

Au

Au

Ca

keV Figure 3.42. Aluminum and Si peaks detected by XRF in soil particles bound by amorphous Si in the crust of the Coteau soil. 85 the elemental composition of cementing agents in the crust. The soil surface was not altered in either the soils of Mexico or Louisiana when it was protected from the direct impact of raindrops. The presence of a compacted, 0.1-mm upper layer, consisting of single clean grains was the major morphological feature observed at the end of the drying period in the bare and gypsum-amended Louisiana soils. In contrast, the high dispersivity of the Asogueros and Reforma contributed to seal the soil surface and to form a massive structure in all of their treatments. A compacted 0.2-mm upper layer of clean silt grains underlain by a homogeneous mmixture of sand, silt, and clay, was the major morphological feature in the bare Sandovales soil. An improvement in aggregation was observed in this soil for the gypsum-amended soil.

Accumulation of clay-sized material in the surface layer without a compact upper layer consisting of single, clean grains, was observed in the protected soils of all six soils. Soil erosion in the soils of Mexico was reduced 52% and 32% in the protected and gypsum-amended soils, respectively, compared to the erosion in the bare soils. Soil erosion in the soils of Louisiana decreased 60% and 39% in the protected and gypsum- amended soils, respectively, compared to the erosion in the bare soils. The presence of calcite was identified in the SEM micrograph of the Coteau crust by XRF. Iron bridges between particles were detected in the crust of the Olivier. Silica was detected as filaments between particles in the crust 86 of Gigger. Gypsum was observed bridging particles in the non-gypsum- amended Reforma and Asogueros soils. Massive coatings of Si and calcite were observed in the Reforma. In addition to pedogenic gypsum, massive

Fe coatings were detected in the Sandovales crust.

As previously stated the purpose of this work was to identify mechanisms and pathways of crust formation by studying crust formation.

In the protected soils, the kinetic energy of raindrops was reduced when the soil surface was protected. This protection prevented aggregate dispersion, particle suspension, and particle segregation at the soil surface.

This interpretation is based on a general crust morphology consisting of an undisturbed surface and a soil matrix dominated by a homogeneous particle-size distribution.

Aggregates in the soil surface were destroyed by raindrop impact in unprotected, non-calcareous soils. Dispersed clay was either lost through runoff or transported to underlying depths. Consequently, an upper 0.1 to

0.2 mm layer formed of clean silt grains underlain by a porous, homogeneous mixture of sand, silt, and clay.

The definitive morphology of the calcareous, unprotected soils is a nonsegregated, dispersed, massive system. There is no layer of clean silt grains at the surface. It is suspected that a very rapid reduction in infiltration rate (discussed in Chapter 5) occurred due to chemical dispersion. This sealing effect then promoted continual runoff and erosion 87 of the dispersed system. The only morphological remnant is the massive, dispersed sealing layer.

The morphology of material cementing particles together occurred in two forms. Bridges between particles consisted of amorphous Si, Fe oxides, and gypsum. Massive coatings cementing particles together consisted of amorphous Si, Fe oxides, and calcite.

References for Chapter 3

Aujla, T.S., B. Singh, and B.S. Sandhu. 1986. Effect of pre-sowing and post-sowing irrigation treatment to wheat following rice in Punjab. J. Ecology. 13(2):250-255.

Ben-Hur, M., J. Letey and I. Shainberg. 1990. Effect of crust formation and polysaccharide on soil erodibility. Soil Sci. Soc. Am. J. 54:1092- 1095.

Ben-Hur, M., R. Stern, A. J. van der Merwe, and I Shainberg. 1992. Slope and gypsum effect on infiltration and erodibility of dispersive and nondispersive soils. Soil Sci. Am. J. 56:1571-1576.

Bradford, J.M. and R.B. Grossman. 1982. In situ measurement of near­ surface soil strength by the fall-cone device. Soil Sci. Soc. Am. J. 46:685- 688 .

Bresson, M.L., and J. Boiffin. 1990. Morphological characterization of soil crust development on an experimental field. Geoderma, 47:301-325.

Chen, J., J. Tarchitzky, J. Morin, and A. Banin. 1980. Scanning electron microscope observations on soil crust and their formation. Soil Sci. 130:49-55.

Gal, M., L. Arcan, I. Shainberg, and R. Keren. 1984. The effect of exchangeable Na and phosphogypsum on the structure of soil crust-SEM observations. Soil Sci. Soc. Am. J. 48:872-878. 88

Goyal, M.R., G.L. Nelson and L.O Drew. 1981. Moisture and soybean seedling emergence. Trans. Am. Soc. Agric. Eng. 24:1432-1435.

Laws, J.O. 1941. Measurements of fall-velocity of water-drops and raindrops. Trans. Am. Geog. Union 22:709. le Bissonnais, Y., A. Bruand and M. Jamagne. 1989. Laboratory experimental study of soil crusting: relation between aggregate breakdown mechanisms and crust structure. Catena. 16:377-392.

McIntyre, D.S. 1958. Soil splash and the formation of surface crusts by raindrop impact. Soil Sci. 85:261-266.

Miller, W.P., and J. Scifres. 1987. Effect of sodium nitrate and gypsum on infiltration and erosion of a highly weathered soil. Soil Sci. 148:304- 309.

Morin, J., S. Goldberg and I. Seniger. 1966. A rainfall simulator with a rotating disk. Trans. Am. Soc. Agric. Eng. 10:74-79.

Morin, J. and Y. Benjamini. 1977. Rainfall infiltration into bare soils. Water Resour. Res. 13:813-817.

Morin, J., Y. Benjamini and A. Michaeli. 1981. The dynamics of soil crusting by rainfall impact and the water movement in the soil profile. J. 52:321-335.

Onofiok, O., and M.J. Singer. 1984. Scanning electron microscope studies of surface crusts formed by simulated rainfall. Soil Sci. Soc. Am J. 48:1137-1143.

Ramley, P.A., and J.M. Bradford. 1989. Relationship of soil crust morphology to inter-rill erosion parameters. Soil Sci. Soc. Am. J. 53:1215- 1221.

Sawnhey, B.L. 1986. Electron microscope analysis. Jn. A. Klute (ed.) Methods of Soil Analysis, Part 1. Physical and Mineralogical Properties. 2nd. ed. Agronomy 9:383-411. Am. Soc. Agron. Madison, Wl.

Shainberg, I., M.E. Sumner, W.P. Miller, M.P.W. Farina, M.A. Pavan and M. V. Fey. 1989. Use of gypsum on soils: a review. Adv. Soil. Sci. 9:1- 111. 89

Smith, H.J.C., G.J. Levy and I.Shainberg. 1990. Water-droplet energy and soil amendments: effect on infiltration and erosion. Soil Sci. Am.J. 54:1084-1087.

Stern, R., M. Ben-Hur, and I. Shainberg. 1991. Clay mineralogy effect on rain infiltration, seal formation and soil losses. Soil Sci. 152:455-462.

Tarchitzky, J., A. Bannin, J. Morin, and Y. Chen. 1984. Nature, formation and effects of soil crusts formed by water drop impact. Geoderma 33:135-155.

Wischmeier, W.H. and D.D. Smith. 1978. Predicting rainfall erosion losses. USDA Agricultural Hanbook. 537p. CHAPTER 4

GEOCHEMISTRY OF CEMENTING AGENTS IN CRUSTS FROM SOILS OF MEXICO AND LOUISIANA

Introduction

Crystallization of inorganic cementing agents can be accomplished when the activities of ions meet or exceed the solubility of a given mineral.

When ion activities are below the equilibrium point or undersaturated, the crystallization of a given mineral is not possible (Lindsay, 1979; Rai and

Kittrick, 1989).

Soil solution composition depends on the moderation of ions by the soil (dissolution, exchange, desorption, etc), additions by either water or wind, or both, and translocation of solutions from above and below the soil profile. In the case of crust formation, translocation of ions from the subsoil to the soil surface by evaporation is an important process that concentrates the soil solution. This ion concentration by surface evaporation may be one of the mechanisms responsible for cementing soil particles in the crust during hot, dry periods immediately following intense rainfall.

Calcite, gypsum, and amorphous Al and Si are reported to be the most probable cementing agents in soils from the north-central part of

Mexico (Detenal, 1971; Lupercio, 1986). Iron oxides, and amorphous Al

90 91 and Si are the most probable cementing agents in loess soils of Louisiana

(Bawer, 1990).

The objectives of this study were: i) to identify the cementing agents present in loessial soils of Louisiana and soils from the north-central part of Mexico, ii) to assess the distribution of such cementing agents in the crust and subsurface, iii) to determine the effect of different management practices on the cementation process during the soil crust formation, and iv) to determine possible solubility relationships between gypsum and calcite that may affect surface crusting depending on specific management practices.

Materials and Methods

Soil series

The surface horizons of three soils from Louisiana and three soils from the north-central part of Mexico were used in this study. Soils from

Louisiana were: Gigger (fine-silty, mixed, thermic, Typic Fragiudalf), Olivier

(fine-silty, mixed, thermic, Aquic Fragiudalf), and Coteau (fine-silty, mixed thermic, Glossaquic Hapludalf). Soils from Mexico were: Asogueros

(loamy, mixed, hypothermic, Vertic Nadurargid), Sandovales (loamy, mixed, hypothermic, Aridic Durustoll), and Reforma (loamy, mixed, hypothermic,

Lithic Calciorthid). 92

Soil preparation

Soil samples were collected from the plow layer (0 to 20 cm), air dried, and sieved through a 2 mm sieve. Soils were packed in wooden boxes (30 x 30 x 7 cm). The bottom of the boxes were perforated and covered with a 1 cm layer of coarse sand to improve soil drainage. Dry soil was uniformly packed into the boxes and lightly agitated to attain reproducible bulk densities. The wooden boxes were positioned at a slope of 0.2% slope to allow surface runoff during rainfall simulation.

Rainfall simulation

Soils were placed under a modified, rotating-disk rainfall simulator described by Morin et al, (1966). The terminal velocity for drops with a medium diameter of 2.0 mm was 6.4 m/s, using a relationship between terminal velocity (m/s) and drop diameter (mm) reported by Laws (1941).

The rainfall intensity was 54 mm/h and the kinetic energy was 26.97 J/m2- mm, obtained using the method reported by Wischmeier and Smith (1978).

A detailed description of the rainfall simulator can be found in Appendix A.

Drying period

Following the rainfall simulation, soil drying was achieved by placing infrared light bulbs 100 cm above the soil surface. A diurnal effect was simulated by alternating on-off periods of 12 hr for 7 days. 93

Management practices

The management practices evaluated were: i) bare, ii) gypsum- amended, and iii) protected. In the protected soil, a 2-mm mesh screen was set 3 cm above the soil surface to simulate a vegetative cover. In the gypsum amended soil, an application rate corresponding to 5 ton/ha of gypsum was spread on the surface. The amount of gypsum used was based on na recommendation given by Ben-Hur et al. (1992).

Sampling zone

Soil samples from the crust and subsurface were taken for chemical and mineralogical analyses. The crust samples were taken from the upper

0.5 cm and the subsurface samples from a 0.5 cm thick zone at a depth of

3.0 cm.

Experimental design

A split plot with randomized block design with two replications was employed. The management practices were placed in the whole plot and sampling zone was a fixed factor in the subplot. Analyses of variance and means separation tests were performed using the Statistical Analyses

Program (SAS Institute, Cary, North Carolina, USA, 1993).

Chemical analysis

Gypsum was quantitatively determined by the electrical conductivity method described by the U. S. Salinity Lab. Staff (1969). 94

The CaC03 equivalent was determined by acid neutralization using

the method described by the U. S. Salinity Lab. Staff (1969).

Amorphous Si and Al were estimated using hot 1.0 M NaOH as

described by Alexiades and Jackson (1967).

Free Fe oxides were determined by the dithionite-citrate-bicarbonate

method described by Mehra and Jackson (1960).

Saturated pastes were equilibrated for 24 hours. The extract was obtained using Buechner funnels and vacuum as described by the U. S.

Salinity Lab. Staff (1969). Soil reaction was measured in the saturated

paste using an Orion EA 940 microprocessor controlled pH/ISE unit following the method described by the Soil Survey Staff (1984). Soluble

Ca, Mg, K, and Na were determined in the saturated paste extracts using

inductively coupled plasma spectrometry (ICP). The anions S 04, Cl, and

N03 were determined by ion chromatography. The HC03 content was

calculated as the anion deficit between the cations and anions. The

electrical conductivity of the saturated extract was obtained using a YSI

model 35 conductivity meter.

Random-powder diffractoarams

Identification of soil minerals by X-ray diffractograms was made on a

Philips diffractometer. Samples from bulk soils were finely ground and

uniformly packed in a dry-powder sample holder. Samples were run from 95

2° to 50° 20 using 40 KV and 25 ma. Caution was used to avoid orientation of particles by pressure effects during the packing procedure.

Simulation of solution-mineral equilibrium using GEOCHEM

The multi-purpose chemical speciation program, GEOCHEM, developed by Sposito and Mattigod (1979) was used to obtain ion activities and simulate the precipitation and dissolution of gypsum and

CaC03. Inputs to this program included: i) total molar concentrations of

Ca, Na, Mg, K, Cl, S04, N03, and pH from the saturated paste extracts, and ii) assumed C02 partial pressures of 10'3 52 and 10‘2 52 atm, with 10'3 52 atm representing atmospheric conditions and 10'2 52 atm representing a higher C02 level expected in subsurface environments of the soil.

A problem could arise when using the water content of a saturated paste to represent equilibrium conditions in the drying environment of a soil crust. Cementing agents present in minor amounts in the crust could be completely dissolved at the water content of a saturated paste. On the other hand, the undersaturated condition might shift to one of supersaturation and precision as concentration of the soil solution occurs during drying. An alternative to this explanation of undersaturation could be that dissolution of some minerals by the saturated paste may be kinetically limited. This lack of equilibrium suggest that a 24 hour equilibrium was not long enough to establish equilibrium conditions. 96

To reverse the dilution or kinetic-limitation effect, solution concentrations of ions were concentrated, using the saturated paste as a

100% saturated condition. A proportional accumulation of ions in the crust was assumed as the soil moisture was decreased. To illustrate, a concentration of 25 ppm Ca in the saturated paste would be doubled to 50 ppm Ca in a simulation for solution conditions at 50% saturation. The levels of soil moisture used were: 100, 90, 80, 70, 60, 50, 25, 12.5, and

6.25%. The pH was calculated by GEOCHEM for each soil moisture content, as it would not be possible to measure by conventional methods at these low soil moisture contents.

Discussion of Results

Identifying cementing agents in soil crusts and the processes by which they form is necessary to understand crust formation. Discovering the origin of cementing agents is an important key to formulating viable solutions to the crust formation problem.

Free Fe oxides and amorphous Al and Si

The effect of management practices (bare, gypsum-amended, and protected soils) on the accumulation of free Fe oxides or amorphous Si and

Al, indicted no statistical difference (p>0.05) between the crust and subsurface in any of the soils (Tables 4.1 and 4.2). However, the soils of

Louisiana had approximately twice the free Fe oxides content of the soils of 97

Table 4.1. Potential cementing agents present in the crust and subsoil in the soils of Louisiana.

Soil Management Sampling Fe20 3 AI2O3 S i02 CaS04- C aC 03 Series Practices Zone 2H20

%

Gigger Bare Crust 0.40at 0.63a 1.22a 0 .00 b 0 .00 b

Subsurface 0.49a 0.63a 1.21a 0 .00 b 0 .00 b

Gypsum Crust 0.60a 0.30a 1.06a 5.95a 1.37a

Subsurface 0.71a 0.31a 1.22a 0 .00 b 0 .00 b

Protected Crust 0.46a 0.76a 1.36a 0 .00 b 0 .00b Subsurface 0.46a 0.83a 1.46a 0 .00 b 0 .00b

Olivier Bare Crust 0.43a 0.41a 1.21a 0 .00 b 0 .00b Subsurface 0.43a 0.44a 1.29a 0 .00 b 0 .00b

Gypsum Crust 0.36a 0.32a 1.25a 5.34a 0.40a Subsurface 0.42a 0.42a 1.35a 0 .00 b 0 .00b

Protected Crust 0.48a 0.42a 1.25a 0 .00 b 0 .00b

Subsurface 0.50a 0.39a 1.24a 0 .00b 0 .00b

Coteau Bare Crust 0.24a 0.43a 1.35a 0 .00 b 0 .00 b

Subsurface 0.24a 0.53a 1.55a 0 .00 b 0 .00b

Gypsum Crust 0.31a 0.37a 1.34a 5.32a 2.92a Subsurface 0.33a 0.37a 1.37a 0 .00b 1.48a

Protected Crust 0.26a 0.46a 1.35a 0 .00 b 0 .00b

Subsurface 0.25a 0.42a 1.39a 0 .00 b 0 .00 b t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 98

Table 4.2. Potential cementing agents present in the crust and subsurface in the soils of Mexico.

Soil Management Sampling Fe20 3 a i2o3 Si02 C aS04 CaC03 Series Practices Zone •2H20

%

Asogueros Bare Crust 0 .2 4 a t 1.28a 4.96a 1.72b 5.06a

Subsurface 0.27a 1.24a 5.00a 0.53c 4.81a Gypsum Crust 0.23a 1.12a 4.95a 6.47a 4.02a

Subsurface 0.26a 1.23a 5.31a 0.45c 4.36a Protected Crust 0.27a 1.00a 4.24a 1.55b 4.23a Subsurface 0.28a 1. 11a 4.90a 0.30c 3.98a

Sandovales Bare Crust 0.27a 1.26a 5.56a 1.37b 1.77a Subsurface 0.28a 1.20a 5.39a 0.58c 0.98a

Gypsum Crust 0.23a 0.93a 5.00a 5.31a 2 .22a Subsurface 0.23a 1.26a 5.50a 0.56c 0.87a

Protected Crust 0.26a 1. 11a 4.92a 2.05b 1.72a

Subsurface 0.27a 1.29a 5.82a 0.34c 0.39a

Reforma Bare Crust 0.24a 0.90a 4.65a 4.99a 9.65a

Subsurface 0.26a 0.89a 4.78a 0.50b 9.33a

Gypsum Crust 0.18a 1.19a 5.62a 8.33a 11.38a

Subsurface 0.18a 1.50a 6.79a 0.93b 11.41a Protected Crust 0 .22a 0 .88a 5.28a 6 .01a 9.19a

Subsurface 0.26a 1.02a 5.26a 0.37b 9.29a t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 99

Mexico. Amorphous Si was approximately 2 times higher and Al approximately 4 times higher in the soils of Mexico compared to the soils of Louisiana.

Free Fe oxides and amorphous Al contents in the crusts were similar to those in the subsurface samples as these elements have a low mobility once they precipitate as hydroxides or oxyhydroxides (Bohn et al, 1985).

The low solubility of Fe and Al oxyhydroxides, compared to that of gypsum, calcite and even Si, apparently allows them to remain unaltered in the soil rather than being solubilized and transported in solution to the soil surface during evaporation.

In the case of Si, whose mobility should be higher than that of Fe and Al under the conditions investigated, the NaOH extract may not have been sensitive enough to detect trace amounts that may have moved in solution towards the soil surface and played an effective role in the crust cementation.

Gypsum

The analyses of variance of gypsum revealed a significant (p<0.05) management practice x sampling zone interaction effect in all of the soils

(Tables 4.1 and 4.2). In the case of soils of Louisiana, gypsum was found only in the crusts of the gypsum-amended soil as a result of the gypsum amendment added in this soil. However, in the soils of Mexico, pedogenic gypsum was precipitated not only in the crust of the gypsum-amended soil, 100 but also in the crust and subsurface of the bare and protected soils. The highest increase of pedogenic gypsum was 6.0% in the crust of the bare

Reforma soil where no gypsum was detected in the initial soil.

The higher precipitation of pedogenic gypsum in the crust than in the subsurface supports the argument that Ca and S04 are transported in solution towards the soil surface during evapotranspiration. The lesser amounts of gypsum in the subsurface of the soils of Mexico suggests a gradual increase in the amount of Ca and S 04 transported by capillary rise from the subsurface to the soil surface during evaporation.

Calcium carbonates

The analyses of variance of CaC03 indicated a significant (p<0.05) management practice x sampling zone interaction effect only in soils of

Louisiana (Tables 4.1 and 4.2). Precipitation of CaC03, however, did occur only in the crusts of the Gigger and Olivier and in the crust and subsurface of the Coteau in the gypsum-amended soil. Considering that no CaC03 was detected in the initial soil properties in the soils of Louisiana, the highest increase in CaC03 was detected in the crust of the Coteau at

2.92%. In the soils of Mexico, the Asogueros and Reforma did not have an increase in the amount of CaC03 present in the crust and subsurface.

Sandovales with no CaC03 in the initial soil analysis, indicated precipitation of CaC03 in the crust and subsurface in all of the soil treatments. The 101

highest increase was detected in the crust of the gypsum-amended soil at

2.22%.

Precipitation of CaC03 in the Louisiana and Sandovales soils in the

gypsum-amended soil was caused by an obvious concentration of Ca in the crust, coming from the dissolution of gypsum added to the surface. In contrast, precipitation of CaC03 in the bare and protected soils of the

Sandovales soil was due to an accumulation of Ca in the crust, originating from the dissolution of gypsum initially present in the soil. A more detailed discussion addressing precipitation of gypsum and CaC03 will be presented later in this chapter.

Powder X-rav analysis

Powder XRD analysis was performed on the soil crust and subsurface samples to confirm the minor amounts of cementing agents found by chemical means were accurate and the mineral form actually present. Gypsum and calcite were the only cementing agents identified by

XRD. These two minerals were present in the crusts of the Gigger,

Coteau, and each of the soils from Mexico amended with gypsum (Tables

4.3, 4.4, 4.5, and 4.6). Pedogenic gypsum and calcite were detected in the crusts of all treatments of the soils of Mexico. Calcite with no gypsum was detected in the subsurface of Coteau, Asogueros, and Reforma.

These results confirm the presence of gypsum and calcite in almost all of the soil treatments where these minerals were detected by chemical 102

Table 4.3. Calcite estimates using three methods in the soils of Louisiana.

Soil Treatment Sampling ■Calcite Estimates— Series Zone Chemical XRD GEOCHEM

<%) (%> Gigger Bare Crust ND NDt 0.02 Subsurface ND ND 0.02 Initial ND ND Gypsum Crust 1.4 Present 0.1 Subsurface ND ND 0.1 Initial ND ND Protected Crust ND ND 0.02 Subsurface ND ND 0.02 Initial ND ND

Olivier Bare Crust ND ND 0.02 Subsurface ND ND 0.02 Initial ND ND Gypsum Crust 0.4 ND 0.02 Subsurface ND ND 0.02 Initial ND ND Protected Crust ND ND 0.02 Subsurface ND ND 0.02 Initial ND ND

(table con'd) 103

Coteau Bare Crust ND ND 0.2 Subsurface ND ND 0.2 Initial ND ND Gypsum Crust 2.9 Present 0.2 Subsurface ND ND 0.2 Initial ND ND Protected Crust ND ND 0.2 Subsurface ND ND 0.2 Initial ND ND t Amount of calcite calculated to precipitate from soil extract at 6.25% moisture content, t ND - None detected. 104

Table 4.4. Calcite estimates using three methods in the soils of Mexico. Soil Surface Sampling Calcite Estimates— Series Management Zone Chemical XRD GEOCHEIV

(%) (%) Asogueros Bare Crust 5.1 Present 0.2 Subsurface 4.8 Present 0.2 Initial 3.9 Present Gypsum Crust 4.02 Present 0.2 Subsurface 4.36 Present 0.2 Initial 3.9 Present Protected Crust 4.23 Present 0.2 Subsurface 3.98 Present 0.2 Initial 3.9 Present

Sandovales Bare Crust 1.77 Present 0.1 Subsurface 1.0 NDt 0.1 Initial ND ND Gypsum Crust 2.2 Present 0.1 Subsurface 0.9 ND 0.1 Initial ND ND Protected Crust 1.7 Present 0.1 Subsurface 0.4 ND 0.1 Initial ND ND

(table con'd) 105

Reforma Bare Crust 9.65 Present 0.2 Subsurface 9.33 Present 0.2 Initial 11.4 Present Gypsum Crust 11.4 Present 0.2 Subsurface 11.4 Present 0.2 Initial 11.4 Present Protected Crust 9.2 Present 0.2 Subsurface 9.3 Present 0.2 Initial 11.4 Present t Amount of calcite calculated to precipitate from soil extract at 6.25% moisture content, t ND - None detected. 106

Table 4.5. Gypsum estimates using three methods in the soils of Louisiana.

Soil Surface Sampling Gypsum Estimates— Series Management Zone Chemical XRD GEOCHEMt

(%) (%) Gigger Bare Crust ND NDt 0.5 Subsurface ND ND ND Initial ND ND Gypsum Crust 6.0 Present 1.2 Subsurface ND ND 0.2 Initial ND ND Protected Crust ND ND 0.2 Subsurface ND ND ND Initial ND ND

Olivier Bare Crust ND ND 0.3 Subsurface ND ND ND Initial ND ND Gypsum Crust 5.3 Present 1.7 Subsurface ND ND 0.2 Initial ND ND Protected Crust ND ND 0.2 Subsurface ND ND ND Initial ND ND

(table con'd) 107

Coteau Bare Crust ND ND 0.1 Subsurface ND ND ND Initial ND ND Gypsum Crust 5.3 Present 1.3 Subsurface ND ND 0.2 Initial ND ND Protected Crust ND ND 1.2 Subsurface ND ND ND Initial ND ND t Amount of gypsum calculated to precipitate from soil extract at 6.25% moisture content. t ND - None detected. 108

Table 4.6 Gypsum estimates using three methods in the soils of Mexico.

Soil Surface Sampling —Gypsum Estimates-— Series Management Zone Chemical XRD GEOCHEM t

(%) (%) Asogueros Bare Crust 1.7 Present 1.5 Subsurface 0.5 ND* 1.0 Initial ND ND Gypsum Crust 6.5 Present 1.9 Subsurface 0.5 ND 0.7 Initial ND ND Protected Crust 1.6 Present 1.6 Subsurface 0.3 ND 0.7 Initial ND ND

Sandovales Bare Crust 1.4 Present 0.9 Subsurface 0.6 ND 0.2 Initial 0.3 Present Gypsum Crust 5.3 Present 1.2 Subsurface 0.6 ND 0.2 Initial 0.3 Present Protected Crust 2.1 Present 0.9 Subsurface 0.4 ND 0.2 Initial 0.3 Present

(table con'd) 109

Reforma Bare Crust 5.0 Present 1.7 Subsurface 0.5 ND 0.4 Initial ND ND Gypsum Crust 8.3 Present 1.8 Subsurface 0.9 ND 0.8 Initial ND ND Protected Crust 6.0 Present 1.5 Subsurface 0.4 ND 0.3 Initial ND ND t Amount of gypsum calculated to precipitate from soil extract at 6.25% moiture content, t ND - None detected. 110 analysis. However, while found by chemicals analysis, calcite was not detected by XRD in the crust of the gypsum-amended soil of the Olivier, nor in the subsurface of the bare and protected soils of the Sandovales soil.

A factor common to the soil treatments where CaC03 was not detected by

XRD, was an accompanying CaC03 content of < 1 % in the chemical analysis.

Saturated paste extracts

The analyses of variance of Ca, Mg, Na, and K revealed a significant

(p<0.05) management practice x sampling zone interaction effect in all of the soils (Tables 4.7 and 4.8). Obviously, concentrations of those cations in the crust were higher than those in the subsurface as a result of the mobility of Na, Ca, K, Mg, Cl, and S04 compared to Fe, Si, and Al (Bohn et al., 1985 and Brady, 1990).

Calcium was the dominant cation in solution in all of the soils, except in Sandovales where Na was the dominant cation. The highest concentration of Ca in solution was found in the crust of the Reforma and

Asogueros, ranging from 29.03 to 68.04 mmol/L. The higher concentrations of Ca in the soils of Mexico are explained by the differences in rainfall and leaching conditions between the Mexican and Louisiana sites.

Conditions for a high leaching intensity exist for the soils of Louisiana where the average annual precipitation is 1350 mm. In the soils of 111

Table 4.7. Cation concentrations from saturated pastes in the soils of Louisiana.

Soil Management Sampling Ca Na Mg K Series Practices Zone

------mmol/L

Gigger Bare Crust 18.55bt 3.06a 9.04b 1.83a

Subsurface 1.10e 1.49c 0.41 e 0.35d

Gypsum Crust 31.11a 2.58b 11.75a 1.00c Subsurface 3.56d 2 .66 b 1.39d 0.37d Protected Crust 13.65c 2.57b 6.07c 1.46b

Subsurface 0.99e 0 .86 d 0.38e 0.31 d

Olivier Bare Crust 6.84c 1.94c 2.47b 1.69a

Subsurface 0.80d 0.52d 0.26c 0.28b

Gypsum Crust 19.82a 7.00a 3.63a 1.52a

Subsurface 2.75d 3.18b 0.60c 0.46b Protected Crust 10.23b 1.53c 3.59a 1.64a

Subsurface 1.27d 0.82cd 0.40c 0.33b

Coteau Bare Crust 16.49c 1.91a 2.32c 1.39d

Subsurface 4.14e 0.51a 0.48e 0.35e

Gypsum Crust 42.54a 1.67a 8.24a 4.74a

Subsurface 6.87d 6.69a 1.48b 2.13b Protected Crust 28.28b 2.92a 4.24c 1.62c Subsurface 3.31e 1.05a 0 .1 8f 0 . 12f t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 112

Table 4.8. Cation concentrations from saturated pastes in the soils of Mexico.

Soil Surface Sampling Ca Na Mg K Series Management Zone

------mmol/L------

Asogueros Bare Crust 29.03bct 51.03a 8.11bc 6.39a Subsurface 18.72cd 22.37b 3.36cd 3.38b

Gypsum Crust 58.81a 51.00a 13.54a 6.78a

Subsurface 13.31 d 17.27b 2.36d 2.85b

Protected Crust 36.77bc 41.31a 7.57bc 6 .20a Subsurface 15.16cd 20 .20b 2.72d 3.24b

Sandovales Bare Crust 16.08a 63.25b 17.96a 8.37b

Subsurface 3.89b 17.95c 1.44b 2.64d

Gypsum Crust 17.56a 20.09c 5.73b 4.41c

Subsurface 3.86b 19.05c 1.28b 2.50d

Protected Crust 18.12a 76.95a 16.80a 10.08a

Subsurface 3.74b 13.80c 1.35b 2.49d

Reforma Bare Crust 68.04a 73.97a 10.00a 4.25a

Subsurface 9.08b 11.46a 0.87b 1.06b

Gypsum Crust 45.14b 40.75b 6.75a 3.37a

Subsurface 13.47c 13.95c 1.22b 1.17b Protected Crust 49.35b 67.77a 5.94a 3.82a

Subsurface 7.82c 11.30c 0.73b 1.05b t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 113

Mexico, an annual precipitation of 450 mm fails to leach the Ca from the soil profile.

It is also clear that the addition of gypsum to the soil surface significantly increased the amount of Ca in the crust in all of the soils, except for the Sandovales and Reforma (Table 4.8). The lack of increase of Ca in these soils with the gypsum-amended soil indicates that Ca in solution is controlled by the solubility of gypsum. Once the solution is saturated with respect to Ca and S04, gypsum can not dissolve.

Another effect of adding gypsum to the soil is the replacement of Na by Ca in solution and on exchange sites. This effect is clearly shown in

Sandovales and Reforma where the concentrations of Na in the crust of the gypsum-amended soil decreased 43 and 33 mmol/L respectively, compared to that in the bare soil. Excess Na was apparently lost through runoff or drainage.

In the soils of Louisiana, the level of Na in the crust (0.51 to 7.00 mmol/L) was very low compared to that in the soils of Mexico (11.3 to

76.95 mmol/L). The highest increase of Na was observed in the gypsum- amended soil of the Olivier, where the content of Na was 7.00 mmol/L, compared to 1.94 mmol/L in the crust of the bare soil. These results suggest that the increase of Na in the crust was possibly due to the effect of displacement of Na by Ca on the exchange complex, leaving more Na in solution. The increased Na in solution was then available to accumulate in 114 the crust by capillary rise. It is not obvious why this increase in Na occurs in the Olivier in response to the gypsum amendment. Apparently the high silt content prevents leaching and removal by runoff.

In the case of Mg, K and Na, their relative importance as cementing agents depends on the evaporite minerals that may be formed, such as epsomite (MgS04.7H20), sylvite (KCI), and halite (NaCI). However, these minerals are very soluble and unstable, so that they are easily leached from the soil. For example, in Sandovales and Reforma, the Mg content decreased 12 and 4 mmol/L, respectively, in the crust of the gypsum- amended soil compared to that in bare soil. These results suggest that Mg present on the exchange complex or as soluble salts was displaced by Ca or dissolved and leached out of the soil during the rainfall simulation.

The analyses of variance of Cl, S 04, and N03 revealed a significant

(p<0.05) management practice x sampling zone interaction effect in all of the soils (Tables 4.9 and 4.10). The higher concentration of these anions in the crust relative to that in the subsurface, indicates again the high solubility and mobility of these anions in the soil system. As in the case of the cations, the anion contents also reflect the effect of contrasting parent materials and climatic conditions among the soils used in this study. For example, the soils of Mexico provide a huge difference in the content of Cl and S 0 4, compared to that in soils of Louisiana. 115

Table 4.9. Anion concentrations from saturated pastes in the soils of Louisana.

Soil Management Sampling Cl S 0 4 n o3 Series Practices Zone

------mmol/L —

Gigger Bare Crust 2 6 .9 5 a t 7.15b 15.69b Subsurface 1.28d 1. 12d 0.52c Gypsum Crust 9.59c 19.59a 29.84a

Subsurface 2.45d 5.03c 0.34c

Protected Crust 15.29b 4.81c 14.60b

Subsurface 0.95d 1.00d 0 .21c

Olivier Bare Crust 5.36a 4.11 bed 7.45b

Subsurface 0.98b 0.55e 0.55d

Gypsum Crust 7.58a 30.22a 0.70d

Subsurface 1.96b 5.51 be 0.30d

Protected Crust 2.94a 2.62cd 12.76a

Subsurface 0.74b 0.38e 3.08c

Coteau Bare Crust 21.80b 2.07d 12.76c Subsurface 0.72cd 0.29e 2.58d

Gypsum Crust 48.40a 15.05a 19.19b

Subsurface 4.00e 5.91b 9.58c Protected Crust 23.41b 2.82c 22.16a

Subsurface 0 .86 d 0.30e 1.24d t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 116

Table 4.10. Anion concentrations from saturated pastes in the soils of Mexico.

Soil Management Sampling Cl SO„ n o3 Series Practices Zone

------mmol/L

Asogueros Bare Crust 1 8 .0 5 at 9 .0 3 1! 4 0 .4 4c

Subsurface 9.78b 4.89c 19.96d Gypsum Crust 34.19a 17.10a 86.63a

Subsurface 3.94b 1.97c 5.96e Protected Crust 16.36a 8.18b 56.23b

Subsurface 7.33b 3.67c 1 6.75d

Sandovales Bare Crust 19.25b 50.81a 12.32

Subsurface 5.70c 14.81cde 1.63d

Gypsum Crust 8.81c 27.61b 9.01c Subsurface 2.94c 16.79cd 0.40d

Protected Crust 21.30a 51.63a 27.50a

Subsurface 2.99c 11.47de 1.60d

Reforma Bare Crust 132.10a 20.16abc 40.49

Subsurface 8.65d 8.65d 3.03d

Gypsum Crust 90.08c 22.31ab 12.48c Subsurface 4.50d 18.08bc 0.90d Protected Crust 110.91b 19.81bc 23.68b

Subsurface 7.20d 7.88d 2.51d t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 117

A system dominated by Cl in the crust was identified in Reforma, ranging from 90 to 132 mmol/L. The Cl system in Reforma was higher than that found in Coteau, where Cl in the crust varied from 22 to 48 mmol/L. Sandovales was characterized by a crust system dominated by

S 0 4 in a range from 28 to 52 mmol/L, while Olivier had a S 04 content in the crust, ranging from only 3 to 30 mmol/L.

Gypsum and calcite precipitation models

Solubility diagrams for calcite and gypsum in the soils of Mexico and

Louisiana are presented in Figure 4.1. Two groups can be distinguished based on Ca2+ activities computed by GEOCHEM in each soil. In the first group, consisting of the Asogueros, Reforma, and Coteau soils, all of the soil treatments indicate saturation with respect to calcite (except in the gypsum-amended soil of the Coteau subsurface). In these soils, their respective Ca2+ activities and pH fall on the calcite solubility line at a C 02 partial pressure of 10'3 53 atm. In the second group, consisting of the

Sandovales, Gigger, and Olivier soils, an udersaturated state relative to both gypsum and calcite can be observed as the Ca2+ activities in all of the treatments and their respective pH values fall below the calcite and gypsum solubility lines.

In the first group, the probable precipitation of calcite in the crust system can be characterized by a two-step model. In the first step, the capillary rise of Ca2+ from the subsurface to the soil surface promotes the CM + iue41 Slblt igasfrclieadgpu ne ifrn aaeetpatcsi ol fMxc ^ ^ insoilspracticesMexico of differentmanagement under Solubilitygypsum forcalciteand diagrams Figure 4.1. log (Ca -2 -2 -3 -3 1 and Louisiana.and log SO^ log Gypsum Gypsum Asogueros Coteau C =3.52 pCO C =3.52 pCO act \ \ Calcite act \ Calcite ysm rtce B= Bare = B protected = p Gypsum = G PB ■a □ G

—-3 Gypsum rs □ Subsoil □ Crust 1 Gigger * .p B* Po° PH C =3.52 pCO pCO = 3.52 = pCO act \ \ Calcite act \ \ Calcite advlsReforma Sandovales —-3 !??3S04 Gypsum Gypsum °B Olivier • B pCO = 3.52 = pCO C =3.52 pCO act \ \ Calcite act \ \ Calcite op oo 119 concentration of Ca2+ in the crusts of the bare and protected soils. In the gypsum-amended soil, the increase of Ca2+ is due to the Ca2+ coming from the dissolution of gypsum added to the soil and from the subsurface. In the second step, the concentrated Ca2+ in solution and atmospheric C02 react to precipitate CaC03. Protons are generated in this reaction, lowering pH. This reaction can be written as:

Ca2+ + C 02 + H20 CaC03 + 2H+ (1)

In the second group, all of the treatments show an undersaturated system with respect to calcite and in most cases gypsum. Concentration of Ca2+ in the crust, either by capillary rise in the bare and protected soils during drying or by capillary rise plus addition of Ca2+ in the gypsum- amended soil would be expected.

The pH in all of the soil treatments in the Olivier increase relative to the subsoil, as a result of the concentration of alkalinity (OH ) in the crust

(Figure 4.1). A similar trend to increase pH was observed in the

Sandovales and Gigger with the gypsum-amended soils. The decrease in pH in the bare and protected soils in the Gigger can not be explained.

The activities of Ca2+ and S04‘ speciated for all of the soil treatments are shown in Tables 4.11 and 4.12. It is apparent that a supersaturated condition is reached with respect to gypsum in the crusts of 120

Table 4.11. Calcium and sulfate activities in solution and sulfate activity needed to precipitate gypsum in the soils of Louisiana.

Soil Surface Sampling pCa2+t pSO„2t pSO„2f Saturation Series Management Zone

Gigger Bare Crust 2.19 2.79 2.45 Undersaturated

Subsurface 3.15 3.16 1.49 Undersaturated Gypsum Crust 2.06 2.41 2.58 Supersaturated

Subsurface 2.79 2.64 1.85 Undersaturated Protected Crust 2.28 2.89 2.36 Undersaturated

Subsurface 3.19 3.19 1.45 Undersaturated

Olivier Bare Crust 2.51 2.83 2.13 Undersaturated Subsurface 3.24 3.44 1.40 Undersaturated

Gypsum Crust 2.32 2.08 2.32 Supersaturated

Subsurface 2.91 2.56 1.73 Undersaturated Protected Crust 2.34 3.09 2.30 Undersaturated Subsurface 3.06 3.64 1.58 Undersaturated

Coteau Bare Crust 2.14 3.26 2.50 Undersaturated

Subsurface 2.58 3.88 2.06 Undersaturated

Gypsum Crust 1.91 2.58 2.73 Supersaturated

Subsurface 2.54 2.72 2.10 Undersaturated Protected Crust 1.99 3.24 2.65 Undersaturated Subsurface 2.68 3.83 1.96 Undersaturated t Calcium and sulfate activities from extract solutions. t Sulfate activity needed to be in equilibrium with gypsum. 121

Table 4.12. Calcium and sulfate activities in solution and sulfate activity needed to precipitate gypsum in the soils of Mexico.

Soil Management Sampling pCa2+t pS042 t pSO„2t Saturation Series Practices Zone

Asogueros Bare Crust 2.18 2.16 2.46 Supersaturated Subsurface 2.26 2.36 2.38 Supersaturated Gypsum Crust 1.84 2.41 2.80 Supersaturated

Subsurface 2.41 2.29 2.23 Undersaturated Protected Crust 2.03 2.34 2.61 Supersaturated

Subsurface 2.31 2.41 2.33 Undersaturated

Sandovales Bare Crust 2.51 1.95 2.13 Supersaturated

Subsurface 2.93 2.24 1.71 Undersaturated

Gypsum Crust 2.36 2.14 2.28 Supersaturated Subsurface 2.95 2.19 1.69 Undersaturated

Protected Crust 2.46 1.97 2.18 Supersaturated

Subsurface 2.90 2.33 1.74 Undersaturated

Reforma Bare Crust 1.77 2.57 2.87 Supersaturated

Subsurface 2.47 2.52 2.17 Undersaturated

Gypsum Crust 1.94 2.42 2.70 Supersaturated Subsurface 2.90 2.24 2.24 Supersaturated Protected Crust 1.90 2.50 2.74 Saturated

Subsurface 2.53 2.53 2.11 Undersaturated t Calcium and sulfate activities from extract solutions. t Sulfate activity needed to be in equilibrium with gypsum. 122 the gypsum-amended soils. Soils of Louisiana do not show any saturated system in the crust or subsurface of the bare or protected soils, nor in the subsurface of the gypsum-amended soils. The soils of Mexico, however, reveal a saturated system with respect to gypsum, indicating the presence of pedogenic gypsum in the crust of the bare and protected soils.

Supersaturated conditions with respect to gypsum were reached only in the subsurface of all of the management practices for the Asogueros and in the subsurface of the gypsum-amended soil of the Reforma.

Based on the speciation models of the saturated pastes, pedogenic gypsum and calcite should coexist in the crust of the Asogueros and

Reforma soils. However, in the chemical and X-ray analyses already discussed, calcite was detected not only in the Asogueros, Reforma, and

Coteau, but also in Sandovales, Gigger, and Olivier. In these soils, calcite was also found in the subsurface in all of the soil treatments. A common characteristic in these soils was the low CaC03 detected by HCI titration, ranging from 0.44 to 2.22%.

Simulated concentration of the soil solution

The results obtained for the simulated concentration of the Gigger soil solution are presented in Figure 4.2. A two-phase model was differentiated during the drying process in both the crust and subsurface for all three management practices. The first phase was characterized by an accumulation of alkalinity (OH ) that produced an increase in pH from Figure 4.2. Simulated concentration of soil solutions during the drying process for the Giggerforsoil. the soilof drying process during solutionsthe concentration Simulated Figure 4.2.

, , log (Ca -2 5 L-2 -L log SO log ysm ( Gypsum Gypsum 6 ( 100 62% (6.25%) (6.25%) %) 7 pCO = 3.5: = pCO Calcite Calcite > Crust > Crust Step 2 Step 8 5 log SO log Gypsum Bare d Subsurface (%) Moisture Content (%) Moisture Subsurface 6 pH 7 C =3.5: pCO Calcite Calcite (70%) Step 2 Step 8 5 log SO log Gypsum Protected ( 100 %) (6.25%) pCO = 3.5: = pCO Calcite Calcite Step 2 Step (70%) 124

6.4 to 7.0. No CaC03 precipitated during this phase in any of the soil treatments. Precipitation of gypsum occurred only in the gypsum-amended soil as Ca2+ and S 042' activities were initially at saturated conditions with respect to gypsum at 100% soil moisture content. In the second phase, the increase of Ca2+ activities forced CaC03 to precipitate at approximately

70% of soil moisture in the crust and subsurface of each soil treatment.

From this point, the pH decreases as protons are released during CaC03 precipitation. Gypsum precipitates in the crusts of the bare and protected soils at 50 and 25% of soil moisture, respectively. No gypsum precipitated in the subsurface of these treatments at the lowest water content simulated (6.25%).

Two different pathways of CaC03 precipitation were detected in the

Olivier (Figure 4.3). The first pathway was observed in the crust of the gypsum soil. The initial step was characterized by a decrease in the activity of Ca2+ caused by precipitation of gypsum and an increase in pH from 6.4 to 7.0. In the second step, calcite started to precipitate at a soil moisture of 70% saturation. The pH during this stage increased from 7.0 to 7.8 as the acidity produced during the precipitation of CaC03 was apparently negated by the concentration of alkalinity in the crust. The

Ca2+ activities were almost constant during the second step as gypsum and calcite precipitated simultaneously. Finally, in the third step, a drastic decrease in Ca2+ activity was accompanied by an increase in pH from 7.9 Gypsum Bare Protected

(6.25%) Step Step -2 Gypsum Gypsum Gypsum (70%) - r Step 1 (70%) Step Step 1 (100%) (100%) (70%) Step 1 O) (100%) log SO log SO log SO -3

Calcite Calcite Calcite pCO. = 3.52 pCO =3.52 pCO =3.52

5 67 85 67 8 5 67 8 pH

■ Crust □ Subsurface (%) Moisture Content

Figure 4.3. Simulated concentration of soil solutions during the drying process for the Olivier soil.

CJl 126 to 8.45. Even though CaC03 precipitated in this soil treatment from the soil moisture of 70% saturation, the acidity produced during the precipitation of CaC03 was not enough to override the alkalinization effect brought about by concentrating the solution. In the subsurface of this treatment, the pH decreased slightly as CaC03 started to precipitate at

70% of soil moisture. The main limitation for CaC03 precipitation was the low Ca2+ activity in the subsurface, which was suppressed at the same level during the drying simulation by the competing precipitation of gypsum.

In the second pathway observed for the Olivier in the bare and protected soils, calcite precipitation began at approximately 70% of soil moisture as a consequence of a shift in pH to the calcite solubility line. In the second step, an increase in the activity of Ca2+ encouraged CaC03 precipitation, which apparently dominated the system as pH decreased during the remainder of the drying process. Gypsum started to precipitate at 25% of soil moisture and continued to 6.25% of soil moisture as Ca2+ activity increased sharply in this range of moisture contents. A pathway similar to the one described in the crusts of the bare and protected soils was observed in the subsurface. However, no precipitation of gypsum occurred in the subsurface as Ca2+ and S042' activities were undersaturated with respect to the solubility of gypsum. 127

In Sandovales, a decrease in the activity of Ca2+ generated a two- step model in the gypsum-amended, bare, and protected soils (Figure 4.4).

In the first step of the gypsum-amended soil model, the activity of Ca2+ was controlled by the precipitation of gypsum. During this step, concentration of alkalinity (OH ) in the soil solution increased pH. In the second step, calcite started to precipitate at approximately 70% of moisture saturation moisture. A decrease in Ca2+ activity was evident as

CaC03 and gypsum precipitated simultaneously. In this step, pH increased as the concentration of alkalinity outweighed the acidity generated from the precipitation of CaC03.

The two-step model observed in the bare and protected soils involved a similar CaC0 3 precipitation model. Calcium activity was controlled by the precipitation of gypsum initially. Calcite started to precipitate at 70% of soil moisture, and pH increased due to a concentration of alkalinity in excess of acidity generated from the precipitation of CaC03. In this phase, Ca2+ decreased as CaC03 and gypsum precipitated simultaneously.

In all three soil treatments in the subsurface of the Sandovales, the activity of Ca2+ was controlled by gypsum in the initial drying stage. The pH increased substantially due to concentration of alkalinity before CaC03 began to precipitate at 70% of soil moisture. The pH continued to Gypsum Bare Protected

t -3

-2 Gypsum Gypsum Gypsum

(100%). ,ste.p 1 (70%) (100%) SteP 1 \(70%) Step 2 (100%) Step 1 \(70%) \ Step 2 Step 2 -2 e O) log SO lag SO log SO -3 (6.25%) (6.25 (6.25%V ■ Calcite' Calcite' Calcite' pCO =3.52 pCO =3.52 pCO =3.52

5 6 7 8 56 7 8 5 67 8 pH

■ Crust □ Subsurface (%) Moisture Content

Figure 4.4. Simulated concentration of soil solutions during the drying process for the Sandovales soil. 129 increase during simulated drying as alkalinization exceeded acidity generated during the precipitation of CaC03.

Summary and Conclusions

The identification and distribution of cementing agents in the crust and subsurface, as well as the effect of different management practices on the cementing process were assessed. Calcite and gypsum were used to identify pathways by which cementing agents could form in the soil crusts.

The drying process was simulated during soil crust formation by concentrating solution analysis obtained from saturated pastes. The computer model, GEOCHEM, allowed simulation of solution chemistry while gypsum and calcite were quantitatively precipitated from solution.

There was no detectable accumulation of Fe iron oxides or amorphous Si and Al in the crusts, relative to the subsurface zones in any of the soil treatments in the soils of Mexico and Louisiana. However, amorphous Si and Al were 2 and 4 times higher in the soils of Mexico than in the soils of Louisiana. Free Fe oxides were twice as high in the soils of

Louisiana as in the soils of Mexico. Gypsum precipitated in the crusts of the non-gypsum amended Reforma and Asogueros soils, ranging from 1.6 to 6.0%. No gypsum could be detected in the initial state of these two soils. Calcite precipitated in all crusts when soils were amended with gypsum and in the non-gypsum amended soils of the Sandovales. Calcium 130 was the dominant cation in the crust of both the soils of Mexico (29 to68 mmol/l) and the soils of Louisiana (7 to 43 mmol/l).

Asogueros, Reforma, and Coteau were initially saturated with respect to calcite. Gigger, Olivier, and Sandovales were initially undersaturated with respect to calcite. Concentration of soil solutions to near dryness illustrated the moisture content (70%) at which precipitation of calcite occurred in Gigger, Olivier, and Sandovales. Gypsum also precipitated in the crustal zone during simulating drying in Gigger and

Olivier. The alkalinity produced during concentration of the soil solution and the acidity generated during the precipitation of calcite were identified as two competing mechanisms. Alkalinity dominated the drying period of the soils of Mexico as pH continued to increase. Precipitation of calcite dominated the drying period of the soils of Louisiana as pH decreased during the late stages of crust formation.

Mineral solubility is a major factor affecting the accumulation of cementing agents in soil crusts during evaporation. The higher concentrations of gypsum in the crust relative to the subsurface, indicate the soluble nature of this mineral and its apparent mobility. On the other hand, the lower solubility of Fe oxides and amorphous Al and Si, is responsible for the absence of a detectable accumulation of these cementing agents in the crust relative to the subsoil. However, SEM micrographs and XRF analyses (Chapter 3) revealed that these solid phases 131 do participate in the cementing process, even though their mobility may be limited.

Definite limitations were encountered in trying to examine extracts from saturated pastes. Although calcite could be identified by XRD in soil crusts, solutions were found to be undersaturated with respect to this mineral. Most likely this undersaturation is either i) a kinetic limitation preventing calcite from reaching equilibrium in the saturated paste, or ii) calcite is formed in only trace amounts which are completely dissolved by the paste extract, thereby forcing undersaturation upon further dilution. In either case, artificial concentration of the solution from the extracted pastes to near dryness allows one to examine whether or not it is possible for a mineral to precipitate.

Two competing reactions were recognized from the simulation work.

Precipitation of calcite generates protons, which generates acidity.

Concentration of hydroxyl ions in the soil solution generates alkalinity.

Two precipitation pathways were proposed for calcite from the extract solutions. One pathway involves an initial increase in solution pH during the concentration process, prior to the precipitation of calcite. Upon the initiation of calcite precipitation, the change in pH reverses and decreases with further precipitation, indicating a dominance of the calcite precipitation reaction over the concentration of hydroxyl ions. The second pathway differs in that, at the initiation of calcite precipitation, pH continues to 132 increase, indicating that the concentration of hydroxyl ions is greater than the amount of protons generated from the precipitation of calcite.

References for Chapter 4

Alexiades, C.A., and M.L. Jackson. 1967. Quantitative clay mineralogical analysis of soils and sediments. Clays and Clay Minerals. 14:35-52.

Baver, L.D., W. H. Gadner, and W.R. Gadner. 1972. Soil Physics. John Wiley and Sons, Inc., New York.

Ben-Hur, M., R. Stern, A. J. van der Merwe, and I Shainberg. 1992. Slope and gypsum effect on infiltration and erodibility of dispersive and nondispersive soils. Soil Sci. Am. J. 56:1571-1576.

Bhon, H., B. McNeal, and G. O’Connor. 1985. Soil chemistry. 2nd. ed. Wiley-lnterscience Publications, NY, 341 p.

CETENAL. 1971. Carta edafologica 1:50 000, Aguascalientes, F13-D-19. Mexico, D.F. Mapa.

Laws, J.O. 1941. Measurements of fall-velocity of water-drops and raindrops. Trans. Am. Geog. Union 22:709.

Lindsay, W. L. 1979. Chemical equilibria in soils. Wiley-lnterscience, New York.

Lupercio-Huerta, F.J. 1987. Identification de factores en la formacion de costras en suelos del altiplano Potosino-Zacatecano. (Identification of factors in soil crust formation in Potosino-Zacatecano plateau soils). Tesis Profesional. ITTA No. 20. Aguascalientes, Mexico. 94pp.

Mehra, O.P., and M.L. Jackson. 1960. Iron oxide removal from soils and clays by dithionite-citrate system with sodium bicarbonate. Clays and Clay Minerals. 7:317-327.

Morin, J., S. Goldberg, and I. Seniger. 1966. A rainfall simulator with a rotating disk. Trans. Am. Soc. Agric. Eng. 10:74-79. 133

Rai, D., and J.A. Kittrick. Mineral equilibria and the soil system, im J. B. Dixon and S.B. Weed (eds). Minerals in Soil Environments. 2nd. ed. Society of America. Madison, Wl. 161-197.

SAS Institute, Inc. 1982. SAS user's guide: statistics. SAS Inst., Inc., Cary, NC.

Soil Survey Staff. 1984. Procedure for collecting soil samples and methods of analysis for soil survey. Soil Survey Investigations Report No. 1. U.S. Gov. Print. Office, Washington, D.C.

Sposito, G. and S.V. Mattigod. 1980. GEOCHEM: A computer program for the calculation of chemical equilibria in soil solutions and other natural water systems. Department of Soil and Environmental Sciences, Univ. of California, Riverside, California.

U.S. Salinity Laboratory Staff. 1969. Saline and alkali as soils. Agriculture Handbook No. 60. U.S. Gov. Print. Office, Washington, D.C.

Wischmeier, W.H. and D.D. Smith. 1978. Predicting rainfall erosion losses. USDA Agricultural Hanbook. 537p. CHAPTER 5

EFFECT OF MANAGEMENT PRACTICES AND RAINFALL DURATION IN SOIL PROPERTIES AS THEY AFFECT SEEDLING EMERGENCE OF CORN (Zea mays L.) AND SOYBEANS (Glycine max (L.)Merr.)

Introduction

The direct impact of raindrops on the soil surface, the presence of highly dispersive clays such as smectites and high levels of exchangeable sodium may cause dispersion of soil aggregates, thereby sealing the soil surface, reducing infiltration, and increasing the hazard of soil erosion

(Hillel, 1980; Miller and Radcliffe, 1992; Shainberg, 1992; and Singer and

Warrington, 1992).

The use of mulches on soil surfaces has been reported as an effective method of preventing soil dispersion caused by raindrop impact

(Wilson et al., 1982; McVay et al., 1989; and Bruce et al., 1990). Soil conditioners such as gypsum, phosphogypsum, and organic polymers have also been tested to stabilize soil aggregates in order to avoid the formation of soil crusts (Gal et al., 1984; Shainberg et al., 1989; and Ben-Hur et al.,

1990).

A major limitation to crop production is the effect crusting has on seedling emergence. A decrease in seedling emergence of wheat (Triticum aestivum L.), grain sorgum (Sorgum bicolor (L.) Moench), and soybeans

134 135

(Glycine max (L.) Merr) with an increase in crust strength was reported by

Hanks and Thorp (1957).

Rainfall duration is another factor influencing crust formation. A decrease in seedling emergence was reported as rainfall duration increased

(Agassi et al. 1985; Awadhwal and Thierstein, 1986).

Crust formation is a common problem in soils from the north-central part of Mexico and the loessial soils of Louisiana. Information identifying factors involved in the formation of soil crusts is very limited, as is information addressing seedling emergence of corn and soybeans in these regions.

Based on the need for practical solutions to these problems, the objectives of this study were: to evaluate the effect of different management practices on infiltration, soil erosion, crust strength, and the effect of different management practices and rainfall duration on seedling emergence of corn (Zea mays L.) and soybeans (Glycine max (L.) Merr.).

Materials and Methods

Soil series

The surface horizons of three soils from Louisiana and three soils from the north-central part of Mexico were used in this study. Soils from

Louisiana were: Gigger (fine-silty, mixed, thermic, Typic Fragiudalf), Olivier

(fine-silty, mixed, thermic, Typic Fragiudalf), and Coteau (fine-silty, mixed 136 thermic, Glossaquic Hapludalf). Soils from Mexico were: Asogueros

(loamy, mixed, hypothermic, Vertic Nadurargid), Sandovales (loamy, mixed, hypothermic, Aridic Durustoll), and Reforma (loamy, mixed, hypothermic,

Lithic Calciorthid).

Soil preparation

Soil samples were collected from the plow layer (0 to 20 cm) of each soil, air dried, and sieved through a 2 mm sieve. Soils were packed in wooden boxes (30 x 30 x 7 cm). The bottom of the boxes were perforated and covered with a 1 cm layer of coarse sand to improve soil drainage. Dry soil was uniformly packed into the boxes and lightly agitated to attain reproducible bulk densities. The wooden boxes were positioned at a slope of0 .2% to allow surface runoff during rainfall simulation.

Rainfall simulator

Soils were placed under a modified, rotating-disk rainfall simulator described by Morin et al. (1966). The terminal velocity for drops with a medium diameter of 2.0 mm was 6.4 m/s, using a relationship between terminal velocity (m/s) and drop diameter (mm) reported by Law's (1941).

The rainfall intensity was 54 mm/h and the kinetic energy was 26.97 J/m2- mm, obtained using the method reported by Wischmeier and Smith (1978).

A detailed description of the rainfall simulator can be found in Appendix A. 137

Runoff measurement

Runoff was collected through a plastic drain pipe attached at ground level to the low end of the wooden boxes. Runoff was quantified by taking a subsample every 5 min after the initiation of runoff. Volume of runoff and sediment load were determined.

Drying period

Following the rainfall simulation, a hot, dry period was simulated for

7 days. Infrared lamps were placed 100 cm above the soil surface with alternating on-off periods of12 hours.

Crust strength measurement

Crust strength was measured every other day with a modified penetrometer similar to the one described by Holder and Brown (1974),

Figure 5.1. The penetrometer consisted of a probe mounted on the stage of a balance. A water receptacle was connected to the end of the balance arm. The spherical probe was introduced through one of the perforations in the bottom of the boxes. Water was poured into the receptacle at a rate of 50 mm/min. The volume of the water needed to cause the rupture of the crust was expressed in KPa. The main objective of this measurement was to obtain an estimation of the resistance that a seed would encounter when penetrating the crust. 138

:ii •• ■* - - r d

Figure 5.1. Modified penetrometer used to measure crust strength. 139

Management practices

Three management practices bare, gypsum-amended, protected were evaluated at two rainfall durations of 30 and 60 min were evaluated. The surface of the protected soil was covered with a 2-mm mesh screen set 3 cm above the soil surface to simulate a vegetative cover. In the gypsum- amended soil, an application rate corresponding to 5 ton/ha of gypsum was spread on the surface. The amount of gypsum was based on a recommendation given by Ben-Hur et al. (1992). Two crops were used: soybean and corn. Planting depth was 2.5 cm, in rows having a width of

3.5 cm. Soybean and corn varieties used were Buckshot 723 and Delta

Pineland 4581, respectively.

Experimental design

A split-split plot with randomized block design with two repetitions was employed. Each of the two rainfall durations was applied to a set of three boxes. Each set was considered as the whole plot in each soil. Each one of the three management practices was in each one of the boxes of the whole plot. The middle plot was one box. Corn was planted in one half of each box and soybeans in the other half. Analysis of variance and mean separation tests were performed using the Statistical Analyses

Program (SAS Institute, Cary, North Carolina, USA, 1993). 140

Discussion of Results

Infiltration rate

The infiltration rates for the soils of Mexico and Louisiana with bare, protected, and gypsum-amended soils are shown in Figure 5.2. In each of the six soils, the initial infiltration rate was controlled by the rainfall intensity. The infiltration rate decreased very rapidly as rainfall proceeded.

This decrease in infiltration rate was controlled by either saturation of the soil profile or alteration of the soil surface due to soil dispersion. This trend to decrease infiltration rate can result from a gradual deterioration of soil structure and formation of a surface crust (Morin and Benjamini, 1977;

Hillel, 1980; and Kazman et al., 1983).

In the case of Asogueros, soil dispersion was the main factor in reducing the infiltration rate in all three soil treatments. It is important to point out that a high WDCI was found in the initial soil characterization of this soil (Table 5.1). The elapsed time at the beginning of runoff (ponding time) was only 5 min in the Asogueros soil, compared to 20 to 25 min in the Coteau and Olivier, respectively. The correspondence between ponding time and WDCI is illustrated in Figure 5.3. In general, the high WDCI values in the soils of Mexico corresponded with the early ponding times in these soils. These results agree with the findings of Miller and Bharuddin

(1986), who observed a reduction in the infiltration rate because of sealing of the soil surface in soils with WDCI values above 0.5. Figure 5.2. Infiltration rates of the soils of Mexico and Louisiana with bare, protected, and gypsum-amended gypsum-amended and protected, Infiltration Louisiana bare,soilsMexico andof with the of rates Figure 5.2. Infiltration Rate (mm/h) 20 40 60 soils. 0 0 0 50 40 30 20 sgeo Sandovales Asogueros oeuGigger Coteau Bare 0 0 0 50 40 30 20 Time (min) Time Protected Gypsum 20 Reforma 40

Olivier 141 Table 5.1. Selected physical and chemical soil properties from the soils of Mexico and Louisiana.

Soil WDCIt Si02 ai2o3 Fe20 3 CaC03 Gypsum Series 0/O/ Gigger 0.50 1.09 0.57 0.49 - - Olivier 0.18 1.09 0.49 0.28 -- Coteau 0.21 1.18 0.50 0.19 - - Asogueros 0.83 4.38 1.46 0.07 3.9 - Sandovales 0.79 4.51 1.42 0.08 - 0.0 Reforma 0.79 4.02 1.15 0.02 11.4 0.33 t Water-Dispersible Clay Index Figure 5.3. Relationship between ponding time and water-dispersible water-dispersible and time ponding Relationship between Figure 5.3. Ponding Time (min) clay index.clay 24 - 24 20 - 20 - 28 12 - 12 16 - 16 n - 4 8 - 8 . 03 . 07 0.9 0.7 0.5 0.3 0.1 _i ------Olivier ■ ■ Coteau 1 ------■ Water-Dispersible Clay Index (WDCI) Index Clay Water-Dispersible 1 ------1 ------Gigger ■ 1 ------1 ------1 ------Sandovales Reforma Asogueros and ■ 1 ------■ 143 144

A positive effect on infiltration rate was found in all six soils when

gypsum was added to the surface. Ponding time was delayed

approximately 5 min, except in Asogueros. The two most marked

examples of increasing the infiltration rate were detected in the Asogueros

and Olivier, in the Asogueros, the infiltration rate ranged from 40 to 15

mm/h in the gypsum-amended soil, compared to20 to 15 mm/h in the bare soil. In general, infiltration rate increased in all six soils. It is assumed that

Ca coming from the dissolution of gypsum added to the surface brought about flocculation of dispersed colloids in the soil surface.

The infiltration rate curves for the protected and bare soils indicate that ponding time occurred at the same time for both, except in the Gigger and Olivier. However, a noticeable reduction in the infiltration rate was measured in the protected soil throughout the rainfall event in all six soils

(Figure 5.2). For example, the final infiltration rate in the protected soil of the Asogueros was 6 mm/h compared to 15 mm/h for the bare soil. This infiltration rate reduction indicates that the soil surface was more severely sealed in the protected soil than in the bare or gypsum-amended soils.

An increase in infiltration rate caused by the continuous pounding of the surface by raindrops was pointed out by Seginer and Morin (1970) and

Maulen et al (1990). This increase in infiltration rate resulted from opening up the bombarded surface crust. In this study the principal reason for the reduction in infiltration rate in the protected soil was the protection of the 145 soil surface against the direct impact of raindrops. Dispersed clay-size particles were not removed and washed out, but left in place to clog and seal the soil surface. These results are supported by the lower amount of clay in sediment loss in the protected soils 12 ( %) compared to the bare soils (22%) (Table 5.2). A thin layer formed and was not altered by raindrops as in the case of the bare and gypsum-amended soils where raindrops acted as small bombs continuously destroying this thin surface layer and avoiding the clogging of soil pores. Dispersed soil particles were removed by runoff in the bare soil and gypsum-amended soils.

Soil erosion

The continuous disruption of the soil surface by raindrop impact in the bare and gypsum-amended soils resulted in an increase in sediment loss through runoff. The beneficial effect of surface cover to prevent soil erosion is apparent in studies that compared soil loss in covered and bare surfaces (Lai, 1990; Miller and Radcliffe, 1992). In this study, reduction of sediment loss fell between 41 to 67% in all six soils when the soil surface was protected compared to the bare soil, and a reduction in sediment loss of 19 to 45% was obtained when soils were amended with gypsum compared to the erosion in the bare soil (Table 5.3). In general, a correspondence was found between sediment loss and the WDCI where the soils of Mexico had the highest values of sediment loss and WDCI

(Figure 5.4). 146

Table 5.2. Particle size distribution of sediment loss in the soils of Mexico and Louisiana. Soil Surface Sand Silt Clay Series Management ------% ------Gigger Bare 1 65 34 Gypsum 1 68 31 Protected 2 89 9

Olivier Bare 4 81 15 Gypsum 6 81 13 Protected 6 92 2

Coteau Bare 2 92 6 Gypsum 3 91 6 Protected 1 96 3

Asogueros Bare 5 65 30 Gypsum 3 69 28 Protected 3 67 20

Sandovales Bare 55 30 15 Gypsum 51 35 14 Protected 38 50 12

Reforma Bare 17 49 34 Gypsum 12 61 27 Protected 5 70 25 147

Table 5.3. Effect of management practices on soil erosion in the soils of Mexico and Louisiana. Soil Surface Sediment Reduction Series Management in Erosion

(g/m2) (%) Gigger Bare 172.34 Gypsum 105.00 39 Protected 74.17 57

Olivier Bare 174.44 Gypsum 113.65 35 Protected 74.37 57

Coteau Bare 204.13 Gypsum 127.26 38 Protected 84.66 59

Asogueros Bare 233.06 Gypsum 172.26 26 Protected 112.26 52

Sandovales Bare 376.82 Gypsum 223.38 41 Protected 123.28 67

Reforma Bare 246.63 Gypsum 181.97 26 Protected 146.67 41 Figure 5.4. Relationship between sediment loss and water-dispersible water-dispersible and loss sediment Relationship between Figure5.4. Sediment Loss (g/m ) 100 200 120 120 140 220 240 260 160 160 180 clay index.clay . 03 . 07 0.9 0.7 0.5 0.3 0.1 Olivier Coteau ■ t I i r i i I I I ■ Water-Dispersible Clay Index (WDCI) Index Clay Water-Dispersible Gigger ■ Sandovales Reforma Asogueros ■ ■ ■ 148 149 Crust strength

The effect of management practices and rainfall duration on crust strength in soils of Mexico and Louisiana is illustrated in Figure 5.5. A more impenetrable crust developed after 60 min of rainfall, compared to a

30-min rainfall event. Soil moisture played an important role in determining crust strength during the drying period. The increase in crust strength during this period was negatively correlated to a quadratic expression of soil moisture (Figure 5.6). It is widely accepted that a relationship exists between soil strength and soil moisture (Hussein et al., 1985; Joshi, 1987; and Rot, 1992). This relationship illustrates how crust strength increases as soil moisture decreases.

The highest value of crust strength occurred on the seventh day of drying in five of the six soils. The one exception was the Asogueros, where the highest value of crust strength was obtained on the fifth day.

Analyses of variance of crust strength on the seventh day of drying

(fifth day for Asogueros) after rainfall simulation revealed a significant effect (p<0.05) related to management practice x rainfall-duration for five of the soils. The exception in this case was the Reforma (Tables 5.4 and

5.5). When rainfall duration was 60 min, crust strength was 90% higher in the soils of Mexico compared to that in soils of Louisiana. When rainfall duration was 30 min, crust strength was 25% higher in the soils of Mexico than in soils of Louisiana. 24 Asogueros Sandovales Reforma 20 16 12

o 24 .<2 sz Coteau Gigger Olivier <0 h- 20

0 24 6 0 Days after rainfall ■ b B-60 + -+ G-60 ° ° P-60 a ----- A B-30 x---- x v----- vP-30 B - Bare G - Gypsum p . Protected

Figure 5.5. Effect of management practices and rainfall duration in crust strength in the soils of Mexico and Louisiana. cn o Figure 5.6. Relationship between crust strength and soil moisture. and strength crust Relationship between Figure 5.6. Crust Strength (KPa) (Thousands) 24 20 16 12 4 0 8

1 2 3 4 50 40 30 20 10 0 Soil Moisture (%) Moisture Soil 151 152

Table 5.4. Effect of management practices and rainfall duration in the crust strength at the seventh day after rainfall simulation in the soils of Louisiana. Soil Surface Rainfall Crust Strength Series Management Duration (min) (KPa) Gigger Bare 30 5,722cet 60 12,607ab Gypsum 30 10,922b 60 12,780a Protected 30 4,037de 60 6,910cd

Olivier Bare 30 3,913b 60 4,111b Gypsum 30 3,962b 60 8,123a Protected 30 4,195b 60 4,210b

Coteau Bare 30 4,954d 60 6,687bc Gypsum 30 5,895c 60 6,935abc Protected 30 4,954d 60 7,480d t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 153

Table 5.5. Effect of management practices and rainfall duration in the crust strength at the seventh day after rainfall simulation in the soils of Mexico. Soil Surface Rainfall Crust Strength Series Management Duration (min) (KPa) Asogueros Bare 30 5,746dt 60 9,263c Gypsum 30 11,071b 60 15,108a Protected 30 5,250d 60 11,665b

Sandovales Bare 30 12,037d 60 17,535bc Gypsum 30 8,124e 60 20,978bc Protected 30 14,415cd 60 25,478a

Reforma Bare 30 10,823a 60 10,848a Gypsum 30 10,848a 60 13,424a Protected 30 8,124a 60 10,452a t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 154

A varied response in crust strength was observed when gypsum was used as a soil amendment. Crust strength increased 74% and 51% in the

Asogueros and Olivier, respectively, with respect to crust strength in the bare soil. No significant increase in crust strength was detected in

Sandovales. These results suggest that the gypsum added to these soils maintained Ca2+ and S0 42 activities at a level that allowed gypsum to precipitate during the drying process. As a result, the precipitated gypsum acted as a cementing agent in the crust as it dried, increasing crust strength.

The lack of an increase in crust strength observed in the Sandovales soil is apparently due to the fact that this soil was the only one having gypsum initially present in its native state (). The increase of crust strength detected in the other soils due to the gypsum-amended soil was eliminated by the presence of native gypsum acting as a cementing agent in the bare and protected soils of the Sandovales.

The crust strength in the protected soil of the Gigger was 29% lower than that in the bare soil. The crust strength of the Reforma protected soil was 17% lower than that in the bare soil. These reductions in crust strength were related to the higher moisture content of these soils on the seventh day, reflecting the importance of finding alternative management practices which allow the soil to conserve moisture for longer periods of time. 155

The soils of Mexico had higher values of WDCI and amorphous Si than the soils of Louisiana (Table 5.1). These two properties are correlated with crust strength (Figures 5.7 and 5.8). The higher crust strength of the soils of Mexico is possibly due to the higher binding effect of dispersed clay among soil particles. In the same manner, the higher amorphous Si content of the soils of Mexico possibly allows for stronger cementation of soil particles. Furthermore, the combined effect of having clay in a highly dispersed state combined with an elevated amorphous Si content would favor conditions for crust development through particle bridging or cementation (Chartres and Fitzgerald, 1990).

The use of the stepwise multivariate model revealed that silt content, rainfall duration, organic carbon (OC), gypsum, and sodium adsorption ratio

(SAR) were the major soil characteristics having significant effects

(p<0.15) that influenced crust strength (Table 5.6). These results indicate that crust strength is related more to physical than chemical properties as the partial r2 for silt content and rainfall-duration is 0.52. Organic carbon, gypsum, and SAR produce a combined r2 of only 0.23. The inclusion of

OC content in the soils as a factor which reduces crust strength is explained by the improvement of soil aggregation in soils with higher OC content. Gypsum, as discussed earlier in this chapter, increased crust strength by cementing particles together, and high SAR values encouraged dispersion in the crust. 156

24 Sandovales ■ _ 20 C. X. ^ tn i s £TJ 10 *± c 2c * « OT Asogueros £ 3 12 . f c o x< Gigger . ■ ■ Reforma wTo F t 3 Coteau oft- 8 ■ ■ Olivier

0.1 0.3 0.5 0.7 0.9 Water Dispersible Clay Index (WDCI)

Figure 5.7. Relationship between crust strength and water-dispersible clay index.

24 Sandovales ■ ~ 20 ra o. * _ w 16 J C T O ■t-i r - o> ra m OT Asogueros Gigger £ § 12 a . ■ Coteau Olivier

Amorphous SiO^r/o

Figure 5.8. Relationship between crust strength and amorphous Si. 157

Table 5.6. Stepwise multiple regression of factors influencing crust strength. Variable Constant Partial r2 P> F Silt -40.61 0.4037 0.0001 Rainfall-Duration 108.90 0.1520 0.0001 Organic Carbon -3779.60 0.1037 0.0001 Gypsum 518.50 0.0599 0.0003 SARt 472.07 0.0354 0.0030 Intercept 6659.37 I r2 0.7547

t Sodium Adsorption Ratio 158

Corn and soybean emergence

The effect of management practices and rainfall-duration on corn and soybean emergence is shown in Figure 5.9. The analysis of variance for corn revealed a significant (p<0.05) management practice x rainfall- duration interaction effect in Olivier, Asogueros, and Reforma (Tables 5.7 and 5.8). The interaction of the gypsum-amended soil x 60 min rainfall duration decreased the emergence of corn for the Asogueros (98%),

Sandovales (41 %), and Reforma (300%) compared to emergence in the gypsum-amended soil x 30 min rainfall-duration.

Seedling emergence was reduced 50% for corn in the Olivier, for the interaction bare soil x 60 min rainfall-duration compared to the bare soil x

30 min rainfall-duration. In general, there was a trend to decrease emergence of corn when rainfall duration increased from 30 to 60 minutes.

This reduction was more pronounced in the soils of Mexico (55 to 95%) than in soils of Louisiana (5 to 20%). Corn emergence increased 21% only in the protected Asogueros soil compared to that in the bare soil.

A stepwise multivariate model indicated that amorphous Si, rainfall duration, gypsum, soil moisture at the fifth day, and OC were the factors that significantly (p<0.15) influenced the emergence of corn (Table 5.9).

The presence of rainfall duration in this model coincides with the model for crust strength in which rainfall duration was one of the most dominant factors. Figure 5.9. Effect of management practices and rainfall duration on corn and soybean emergence. rainfallandsoybean andpractices corn onduration management Effectof Figure5.9. Seedling Emergence (%) 100 100 20 40 20 40 60 80 60 80 -0 G-30 B-30 -30 Asogueros -0 G-60 3-60 Coteau = Bare = B R X X X X X 6 B3 G3 P3 B6 G6 P6 B3 G3 P3 B6 G6 P-60 G-60 B-60 P-30 G-30 B-30 P-60 G-60 B-60 P-30 G-30 B-30 -60

Management Practices-Rainfall Duration (min) Duration Practices-Rainfall Management Corn = Gypsum = G Gigg Sandovales a Soybeans Soybeans Protected = P Olivier Reforma CD cn 160

Table 5.7. Effect of management practices and rainfall duration in corn and soybean emergence in the soils of Mexico.

Soil Surface Rainfall Series Management Duration Corn Soybeans (min) - % ------Asogueros Bare 30 80bt Oa 60 60b Oa Gypsum 30 99a 9a 60 50c Oa Protected 30 100a Oa 60 70b Oa

Sandovales Bare 30 79a 8 c 60 42a Od Gypsum 30 100a 17b 60 71a Od Protected 30 83a 54a 60 46a Od

Reforma Bare 30 100a 8 b 60 75a Oc Gypsum 30 100a 8 b 60 25c Oc Protected 30 100a 63a 60 54b 8 b t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. Table 5.8. Effect of management practices and rainfall duration corn and soybean emergence in the soils of Louisiana.

Soil Surface Rainfall Corn Soybeans Series Management Duration (min) - % ------Gigger Bare 30 98at 46a 60 100a 8 a Gypsum 30 100a 21a 60 100a Oa Protected 30 100a 54a 60 99a 37a

Olivier Bare 30 100a 21a 60 50b 17a Gypsum 30 100a 83a 60 100a 50a Protected 30 100a 98a 60 100a 85a

Coteau Bare 30 100a 50a 60 95a 17b Gypsum 30 100a Oc 60 95a Oc Protected 30 100a 50a 60 90a 17b t Means followed by the same letter are not significantly different at the 0.05 level of probability according to the LSD test. 162

Table 5.9. Stepwise multiple regression of factors influencing corn emergence.

Variable Constant Partial r2 P>F

Si02 -0.1542 0.3146 0.0001 Rainfall-Duration -0.0122 0.0925 0.0001 Gypsum 0.0529 0.0564 0.0019 Moisture (5th day) 0.0152 0.0360 0.0316 Intercept 1.8657

0.4995 163

In the case of soybeans, the analysis of variance showed a significant (p<0.05) management practice x rainfall-duration interaction effect in Coteau, Sandovales, and Reforma soils (Tables 5.7 and 5.8).

Seedling emergence of soybeans in the Coteau soil was reduced by approximately 60% when the rainfall duration increased from 30 to 60 minutes in the bare and gypsum-amended soils. When gypsum was added to the soil surface, no emergence of soybeans was observed in the Coteau soil. In the Sandovales, no emergence of soybeans was observed when rainfall duration increased from 30 to 60 min. A similar trend was obtained for the Reforma soil, except for an emergence of8 % in the protected soil x

60 minutes rainfall-duration. In general, there was a trend to decrease the emergence of soybeans when the rainfall duration increased from 30 to 60 minutes. A very low emergence rate of soybeans was observed in the soils of Mexico (10%) and soils of Louisiana (36%).

This effect is explained by the different physiological and morphological characteristics of corn and soybeans. The emergence of corn occurred during the fourth or fifth day after planting, while emergence of soybeans occurred in the sixth and seventh days. As discussed earlier, maximum crust strength was reached on the seventh day so that corn seedlings were not subjected to the more impenetrable crusts that the soybeans were. Another factor which favored a higher emergence of corn was that the sprout emerged with almost no interference of the first 164 plumular leaf. In the case of soybeans, the cotyledons are needed to develop a strong coleoptile before emerge. If the crust is too hard, the coleoptile is easily damaged, leading to no or reduced emergence.

The use of the stepwise multivariate model revealed that soil moisture content at the fifth day of the drying period, rainfall duration and total clay were the major soil characteristics having a significant effect

(p<0.15) which influenced soybean emergence (Table 5.10).

Rainfall duration and soil moisture content on the fifth day were two common factors which determined the emergence of corn and soybeans.

From these two factors, rainfall duration was also detected as an important factor influencing crust strength as discussed earlier in this chapter. These results indicate that rainfall duration is the principal factor affecting crust strength and seedling emergence of corn and soybeans at the rainfall intensities used in this study.

Summary and Conclusions

Early ponding times in the soils of Mexico were related to high values of WDCI. An increase of 5 min in ponding time was detected in all six soils when they were amended with gypsum. A high infiltration rate (40 to 20 mm/h) was measured in soils amended with gypsum during the rainfall simulation compared to a low infiltration rate (20 to 15 mm/h) of the bare soil. The ponding time in all the protected soils was the same as in the Table 5.10. Stepwise multiple regression of factors influencing soybean emergence. Variable Constant Partial r2 P>F Mosture (5th day) 0.0249 0.3238 0.0001 Rainfall-Duration -0.0105 0.1600 0.0001 Total clay -0.0350 0.1467 0.0001 Intercept 1.1985 1 r2 0.6305 166 bare soils. A reduction in infiltration rate from 15 mm/h in the bare soil to

6 mm/h in the protected soil was detected in the Asogueros and was attributed to the formation of a surface seal in the protected soil.

Soil erosion was reduced 41 to 67% when soil surfaces were protected, compared to erosion measured in bare soils. When soils were amended with gypsum, erosion decreased by 129 to 67% compared to erosion in bare soils.

Crust strength increased 90 and 25% in the soils of Mexico and

Louisiana, respectively, when rainfall duration increased from 30 to 60 min.

When gypsum was added to soil surfaces, crust strength increased by

74% in the Asogueros and 21% in the Olivier, compared to the bare soil.

No differences were observed in Sandovales which has a native gypsum content of 0.33%. Crust strength decreased only in the protected soil of the Gigger (29%) and Reforma (17%), relative to the bare soil. The higher values of crust strength in the soils of Mexico were related to high values of WDCI and amorphous Si. Crust strength increased by 100% in Olivier when gypsum was used as a soil amendment compared to that in the bare soil. Factors related to crust strength were silt content, rainfall duration, gypsum, and SAR. From these factors, silt and rainfall duration explained

52% of the variability of crust strength while the other factors explained only 23%. 167

There was a decrease in corn emergence in the soils of Mexico (55

to 95%) and soils of Louisiana (5 to 20%) when rainfall duration increased

from 30 to 60 min. Emergence of corn was increased by 21 % in the

protected soil of the Asogueros. Factors related to the emergence of corn

were amorphous Si, rainfall duration, gypsum, and soil moisture in the 5th

day. Emergence of soybeans was more severely affected than that of

corn. The percent of soybean emergence was only 10% in the soils of

Mexico, compared to 36% in the soils of Louisiana. Factors related to

emergence of soybeans were rainfall duration and soil moisture on the 5th day after rainfall simulation.

Based on these results, the following conclusions are made.

Infiltration rate into dry soil is initially governed by rainfall intensity. As

rainfall proceeds and the soil matrix becomes saturated, the infiltration rate

is governed by hydraulic properties of the soil. However, alteration of the soil surface may influence the infiltration rate, even in unsaturated soils.

Sealing of the surface by chemical dispersion and the sheltering effect provided in the protected soils reduced infiltration rate. The opposite effect, an increase in infiltration rate occurred in the unprotected soils where the soil surface was continuously attacked by raindrop impact.

The continual disruption of the soil surface by raindrops increased physical dispersion of aggregates and, consequently, soil erosion. The increase of electrolytes in soil solution from the gypsum amendment 168 improved soil aggregation. This effect on soil aggregation was reflected in higher infiltration rates in the gypsum amended soils compared to the bare soils. The increase in soil aggregation by gypsum decreased soil erosion compared to the bare soil. However, the highest reduction in soil erosion was obtained in the protected soils.

Increased crust strength was related to 1) the degree of compaction resulting from rainfall impact on the surface and 2) precipitation of cementing agents in the crust. Soils with high levels of amorphous Si and high WDCI, combined with long rainfall duration developed the hardest crusts. In this study, soil texture, rainfall duration, gypsum, and SAR were identified as the main factors affecting crust strength.

The major impact of soil crusting in agriculture is its effect on seedling emergence. Corn and soybean emergence was negatively affected when rainfall duration of the same intensity increased. The most obvious explanation of the successful emergence of corn is that it sprouts and emerges during moist stages before maximum crust strength is developed.

Soybeans were slower in sprouting, and emergence, when it occurred, took place after more extreme conditions of crust hardness had developed in the later stages of drying. 169

References for Chapter 5

Agassi, M., J. Morin, and I. Shainberg. 1985. Effect of raindrop impact energy and water salinity on rates of sodic soils. Soil Sci. Soc. Am. J. 49:186-190.

Awadhwal, N.K. and G.E. Thierstein. 1986. Soil crust and its impact on crop establishment: a review. Soil Till. Res. 5(3):289-302.

Ben-Hur, M., R. Stern, A. J. van der Merwe, and I. Shainberg. 1992. Slope and gypsum effect on infiltration and erodibility of dispersive and nondispersive soils. Soil Sci. Am. J. 56:1571-1576.

Bradford, J., M. and C. Huang. 1991. Mechanisms of crust formation: physical components, in M.E. Sumner and B.A. Stewart (eds). Soil Crusting-Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 55-72.

Bruce, R.R., G.W. Langdale and L.T. West. 1990. Modifications of soil characteristics of degraded soil surfaces by biomas input and tillage affecting soil water regime. Trans. 14th Intern. Congr. Soil Science, Kyoto, Japan, 1990.

Gal, M., L. Arcan, I. Shainberg, and R. Keren. 1984. The effect of exchangeable Na and phosphogypsum on the structure of soil crust-SEM observations. Soil Sci. Soc. Am. J. 48:872-878.

Hanks, R.J. and F.C. Thorp. 1957. Seedling emergence of wheat, grain sorgum, and soybeans as influenced by soil crust strength and moisture content. Soil Sci. Soc. Am. J. 21:357-359.

Hillel, D. 1980. Applications of Soil Physics. Academic Press, NY., 285p.

Holder, C.B., and K.W. Brown. 1974. Evaluation of a simulated seedling emergence through rainfall crusts. Soil Sci. Soc. Am. Proc. 38:705-710.

Hussain, S.M., G.W. Smillie, and J.F. Collins. 1985. Laboratory Iraqi studies of crust development in Irish and Iraqi soils. II. Effects of some physico-chemical constituents on crust strength and seedling emergence. Soil and Tillage Res. 6:123:138.

Joshi, N. L. 1987. Seedling emergence and yield of pearl millet on natural crusted soils in relation to sowing and cultural methods. Soil and Tillage Res. 10:103-112. 170

Kazman, S., I. Shainberg and M. Gal. 1983. Effect of low levels of exchangeable Na and applied phosphogypsum on the infiltration rate of various soils. Soil Sci. 35:184-192.

Lai, R. 1990. Soil erosion and land degradation: the global risk. Adv. Soil Sci. 11; 129-172.

Laws, J.O. 1941. Measurements of fall-velocity of water-drops and raindrops. Trans. Am. Geog. Union 22:709.

Maulem, Y., S. Assouline, and H. Rohdenburg. 1990a. Rainfall induced soil seal. (B) Application of a new model to saturated soils. CATENA. 17:2105-2108.

McVay, K.A., D.E. Radcliffe, and W. L. Hargrove. 1989. Winter legume effects on soil properties and nitrogen fertilizer requirements. Soil Sci. Soc. Am.J. 53:1856-1862.

Miller, W.P., and D.E. Radcliffe. 1992. Soil crusting in the southern U.S. in M.E. Sumner and B.A. Stewart (eds). Soil Crusting-Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 233-266.

Miller, W.P. and M.K. Baharuddin. 1987. Relationship of soil dispersability to infiltration and erosion of southeastern soils. Soil Sci. 142:253-240.

Morin, J., S. Goldberg and I. Seniger. 1966. A Rainfall simulator with a rotating disk. Trans. Am. Soc. Agric. Eng. 10:74-79.

Morin, J. and Y. Benjamini. 1977. Rainfall infiltration into bare soils. Water Resour. Res. 13:813-817.

Roth, C.H. 1990. Soil sealing and crusting in tropical south America. Jn M.E. Sumner and B.A. Stewart (eds). Soil Crusting-Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 55-72.

Seginer, I. and J. Morin. 1970. A model of surface crusting and infiltration of bare soils. Water Resources Res. 6:629:633.

Shainberg, I. 1990. Chemical and mineralogical components of crusting, in M.E. Sumner and B.A. Stewart (eds). Soil Crusting-Chemical and Physical Processes. Lewis Publishers, Boca Raton, FL, pp. 55-72. 171

Wilson, G.F., R. Lai and B.N. Okigbo. 1982. Effects of cover crops on soil structure and on yield of subsequent arable crops grown under strip tillage on an eroded . Soil Tillage Res. 2:233-250.

Wischmeier, W.H. and D.D. Smith. 1978. Predicting rainfall erosion losses. USDA Agricultural Handbook. 537p. CHAPTER 6

SUMMARY AND CONCLUSIONS

The higher values of WDCI in the soils of Mexico (0.79 to 0.83) compared to those in the soils of Louisiana (0.18 to 0.50), contribute to the massive structure and sealing of the soil surface observed in the soils of Mexico. The morphology of the crust at the end of crust formation in the soils of Louisiana soils revealed a compacted, 0.1-mm surface layer, consisting of single clean grains in the bare and gypsum soils. Fine material accumulated in the surface of the protected soils of Mexico and

Louisiana.

The formation of a thin layer of clay size material in the soil surface played an important role in soil erosion and infiltration. Soil erosion was reduced 52 and 39% in the protected soils of Mexico and Louisiana, respectively. A reduction in infiltration rate of 9 mm/h was attributed to the formation of a seal in the protected soil compared to the bare soil in the

Asogueros. In general, the final infiltration rate was lower in the protected than that in the bare soils for all six soils.

The addition of gypsum influenced soil aggregation in the surface, especially in the soils of Louisiana and Sandovales. This aggregation effect reduced soil loss by 32 and 39% in the soils of Mexico and soils of

Louisiana, respectively. A high infiltration rate (20 to 40 mm/h) was

172 173

measured in soils amended with gypsum during rainfall simulation

compared to a lower infiltration (15 to20 mm/h) in the bare soil.

Regarding the presence and distribution of cementing agents, no

significant difference was detected in the accumulation of free Fe oxides or

amorphous Si and Al in the crusts, relative to the subsurface zones in any

of the treatments or soils. Bridges of the following cementing agents were

detected in crusts of bare soils by the SEM/XRF: amorphous Si in the

Reforma and Coteau, free Fe oxides in the Sandovales and Olivier, gypsum

in the Asogueros, and calcite in the Gigger. Calcite was detected by chemical analysis in all crusts when soils were amended with gypsum and in the non-gypsum amended soils of the Sandovales. Gypsum contents ranged from 0.4 to 2.9%. Gypsum was also detected by chemical analysis in the crusts of the non-gypsum amended Reforma and Asogueros soils, and ranged from 1.6 to 6%. No gypsum could be detected in the initial chemical analysis of these two soils.

The solubility diagrams for calcite revealed a system in equilibrium with calcite in the Asogueros, Reforma, and Coteau. In the cases of the

Gigger, Olivier, and Sandovales, concentration of soil solution to near dryness illustrated the moisture content (70%) at which precipitation of calcite probably occurred. Two competing reactions were identified during the concentration of soil solutions to near dryness. Alkalinity dominated the drying period of the soils of Mexico as pH continued to increase. 174

Precipitation of calcite dominated the drying period of the soils of Louisiana as pH decreased during the late stages of crust formation.

The higher values of crust strength in the soils of Mexico (5,250 to

25,478 KPa) were related to high values of WDCI and amorphous Si.

Crust strength increased 90% and 25% when rainfall duration increased from 30 to 60 min in the soils of Mexico and Louisiana, respectively.

When gypsum was added to the soil, crust strength in the gypsum- amended soil increased 74% in the Asogueros and 21% in the Olivier with respect to the bare soil. Crust strength decreased 29% and 17% only in the bare soil of the Gigger and Reforma respectively. Factors related to crust strength were silt content, rainfall duration, gypsum, and SAR.

There was a decrease in corn emergence in the soils of Mexico (55 to 95%) and Louisiana (5 to 20%) when rainfall duration increased from 30 to 60 min. Emergence of corn was increased by 21 % in the protected soil of the Asogueros. Factors related to the emergence of corn were amorphous Si, rainfall duration, gypsum, and soil moisture in the 5th day.

Emergence of soybeans was more severely affected by crust formation than that of corn. The percent of soybean emergence was only 10% in the soils of Mexico, compared to 36% in the soils of Louisiana. Factors related to emergence of soybeans were rainfall duration and soil moisture on the

5th day after rainfall simulation. 175

1) The soils of Mexico have a greater probability of developing dense crusts than the soils of Louisiana. The amount and type of potential cementing agents were related to the contrast in weathering conditions and parent material existing between the arid north-central part of Mexico and the humid climate of Louisiana. Calcite, gypsum and amorphous Si and Al may play an important role in the cementation of the soils of Mexico soils while free Fe oxides may be a potential cementing agent in the soils of

Louisiana crusts.

2) In the protected soils, the kinetic energy of raindrops was reduced when the soil surface was protected. This protection prevented aggregate dispersion, particle suspension, and particle segregation at the soil surface.

This interpretation is based on a general crust morphology consisting of an undisturbed surface and a soil matrix dominated by a homogeneous particle-size distribution.

3) Aggregates in the soil surface were destroyed by raindrop impact in unprotected, non-calcareous soils. Dispersed clay was either lost through runoff or transported to underlying depths. Consequently, an upper 0.1 to

0.2 mm layer formed of clean silt grains underlain by a porous, homogeneous mixture of sand, silt, and clay.

4) The definitive morphology of the calcareous, unprotected soils is a nonsegregated, dispersed, massive system. There is no layer of clean silt grains at the surface. It is suspected that a very rapid reduction in 176 infiltration rate (discussed in Chapter 5) occurred due to chemical dispersion. This sealing effect then promoted continual runoff and erosion of the dispersed system. The only morphological remnant is the massive, dispersed sealing layer.

5) The morphology of material cementing particles together occurred in two forms. Bridges between particles consisted of amorphous Si, Fe oxides, and gypsum. Massive coatings cementing particles together consisted of amorphous Si, Fe oxides, and calcite.

6) Mineral solubility is a major factor affecting the accumulation of cementing agents in soil crusts during evaporation. The higher concentrations of gypsum in the crust relative to the subsurface, indicate the soluble nature of this mineral and its apparent mobility. On the other hand, the lower solubility of Fe oxides and amorphous Al and Si, is responsible for the absence of a detectable accumulation of these cementing agents in the crust relative to the subsoil. However, SEM micrographs and XRF analyses (Chapter 3) revealed that these solid phases do participate in the cementing process, even though their mobility may be limited.

7) Definite limitations were encountered in trying to examine extracts from saturated pastes. Although calcite could be identified by XRD in soil crusts, solutions were found to be undersaturated with respect to this mineral. Most likely this undersaturation is either i) a kinetic limitation 177

preventing calcite from reaching equilibrium in the saturated paste, or ii)

calcite is formed in only trace amounts which are completely dissolved by the paste extract, thereby forcing undersaturation upon further dilution. In either case, artificial concentration of the solution from the extracted

pastes to near dryness allows one to examine whether or not it is possible for a mineral to precipitate.

8 ) Two competing reactions were recognized from the simulation work.

Precipitation of calcite generates protons, which generates acidity.

Concentration of hydroxyl ions in the soil solution generates alkalinity.

Two precipitation pathways were proposed for calcite from the extract solutions. One pathway involves an initial increase in solution pH during the concentration process, prior to the precipitation of calcite. Upon the initiation of calcite precipitation, the change in pH reverses and decreases with further precipitation, indicating a dominance of the calcite precipitation reaction over the concentration of hydroxyl ions. The second pathway differs in that, at the initiation of calcite precipitation, pH continues to increase, indicating that the concentration of hydroxyl ions is greater than the amount of protons generated from the precipitation of calcite.

9) Infiltration rate into dry soil is initially governed by rainfall intensity.

As rainfall proceeds and the soil matrix becomes saturated, the infiltration rate is governed by hydraulic properties of the soil. However, alteration of 178 the soil surface may influence the infiltration rate, even in unsaturated soils.

10) Sealing of the surface by chemical dispersion and the sheltering effect provided in the protected soils reduced infiltration rate. The opposite effect, an increase in infiltration rate occurred in the unprotected soils where the soil surface was continuously attacked by raindrop impact.

11) The continual disruption of the soil surface by raindrops increased physical dispersion of aggregates and, consequently, soil erosion. The increase of electrolytes in soil solution from the gypsum amendment improved soil aggregation. This effect on soil aggregation was reflected in higher infiltration rates in the gypsum amended soils compared to the bare soils. The increase in soil aggregation by gypsum decreased soil erosion compared to the bare soil. However, the highest reduction in soil erosion was obtained in the protected soils.

12) Increased crust strength was related to 1) the degree of compaction resulting from rainfall impact on the surface and 2) precipitation of cementing agents in the crust. Soils with high levels of amorphous Si and high WDCI, combined with long rainfall duration developed the hardest crusts. In this study, soil texture, rainfall duration, gypsum, and SAR were identified as the main factors affecting crust strength.

13) The major impact of soil crusting in agriculture is its effect on seedling emergence. Corn and soybean emergence was negatively affected 179 when rainfall duration of the same intensity increased. The most obvious explanation of the successful emergence of corn is that it sprouts and emerges during moist stages before maximum crust strength is developed.

Soybeans were slower in sprouting, and emergence, when it occurred, took place after more extreme conditions of crust hardness had developed in the later stages of drying. APENDIX A

RAINFALL SIMULATOR

The rainfall simulator (Figure A.1) consisted of a 50 L reservoir from which distilled water was pumped using a 1/5 HP electrical pump and a plastic hose with a 1.32 cm inner diameter. An 8100 veejet nozzle was placed 2.05 m above the soil surface. This nozzle provided drops with a medium diameter of 2.0 mm when 6 psi were applied according to a calibration reported by Morin et al (1966). Using a relationship between terminal velocity in m/s and drop diameter in mm cited by Laws (1941), the terminal velocity for drops with a medium diameter of 2.0 mm was 6.4 m/s. The rainfall intensity was 54 mm/h. The kinetic energy for this intensity was estimated using the following equation cited by Wischmeier and Smith (1978):

E = 11.9 + 8.7 log(l)

Where:

E = Kinetic energy in J/m2-mm

I = Rainfall intensity in mm/h

The resulting kinetic energy was 26.97 J/m2-mm. A rotating plastic disk (Figure 2) with an aperture of 30° was placed under the nozzle. This plastic disk was attached to a ceiling fan motor which allowed the disk to

180 Figure A.1. Rotating-disk rainfall simulator.

Figure A.2. Rotating plastic disk with an aperture of 30' 182 rotate at variable speeds. The water passed through the disk aperture only when the nozzle was above it. The water which did not pass through the disk aperture was removed by a plastic panlike configuration which drained to the water reservoir. In this way, the excess water was recycled. A plastic shelter was placed under the panlike collector to avoid any water splash in the testing area. To prevent any wind effect, plastic walls were attached to the wooden structure in which the rainfall simulator was mounted.

References for Apendix A

Laws, J.O. 1941. Measurements of Fall-Velocity of Water- Drops and Raindrops. Trans. Am. Geog. Union 22:709.

Morin, J., S. Goldberg and I. Seniger. 1966. A Rainfall Simulator with a Rotating Disk. Trans. Am. Soc. Agric. Eng. 10:74-79.

Wischmeier, W.H. and D.D. Smith. 1978. Predicting Rainfall Erosion Losses. USDA Agricultural Handbook. 537p. VITA

Miguel Angel Martinez-Gamino was born in San Luis Potosi, S.L.P.,

Mexico on September 30, 1956. He graduated from Preparatoria # 3 in

San Luis Potosi in June, 1974. He obtained his degree of Agronomy

Engineer in December, 1978. He began his professional career working as a soil and water conservation researcher in the National Institute of

Agricultural Research in Aguascalientes, Mexico in January, 1979. He attended the Postgraduate College in Montecillos, Mexico from August,

1983, to June, 1985, and completed a Master of Science degree in

Edaphology. After obtaining his Master of Science degree, he returned to

Aguascalientes to continue his professional career in soil and water conservation. In August, 1990, he received a COANCYT/IIE-FULBRIGHT scholarship to attend the Agronomy Department at Louisiana State

University where he is presently a candidate for the degree of Doctor of

Philosophy with a major in Agronomy.

183 DOCTORAL EXAMINATION AND DISSERTATION REPORT

Candidate: Miguel Angel Martinez-Gamino

Major Field: Agronomy

Title of Dissertation: Crust Formation in Soils of Mexico and Louisiana

Approved: 7?kd .TLJU*# Major Professor and Chairman

Dean of ate Scnool

EXAMINING COMMITTEE:

Date of Examination:

September 15. 1994