<%BANNER%>

A Floodplain Mining And Channel Change Analysis Of The Tangipahoa River, Louisiana Using GIS

Permanent Link: http://ufdc.ufl.edu/UFE0041740/00001

Material Information

Title: A Floodplain Mining And Channel Change Analysis Of The Tangipahoa River, Louisiana Using GIS 1980 - 2004
Physical Description: 1 online resource (74 p.)
Language: english
Creator: Marks, Steven
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: channel, gravel, louisiana, mining, river, sand, spearman, tangipahoa
Geography -- Dissertations, Academic -- UF
Genre: Geography thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science A FLOODPLAIN MINING AND CHANNEL CHANGE ANALYSIS OF THE TANGIPAHOA RIVER, LOUISIANA USING GIS: 1980-2004 By Steven R. Marks May 2010 Chair: Joann Mossa Major: Geography Sand and gravel mining in floodplains is widespread in Louisiana and can often result in changes in channel planform and position. The purpose of this study was to determine whether or not sand and gravel mining and river channel change are statistically linked over two time periods: 1980 and 2004. The area of focus was the Tangipahoa River, located in southeast Louisiana. This particular river was chosen due to its extensive sand and gravel mining operations within its floodplain. The 2004 time step imagery was Digital Orthorectified Quarter Quadrangles (DOQQ) downloaded from Louisiana s ATLAS website. The imagery for the 1980 time step was USGS 1:24,000 DRG s which was downloaded from the aforementioned website. Using GIS, two different geodatabase s (GDB) were created for each time step. These GDB s contain feature classes that represent the Tangipahoa s centerline, main channel, point bars, cleared land within the floodplain, and sand and gravel pits. In ArcMap, each feature was digitized on top of the corresponding imagery. All features were digitized at a scale of 1:2,000. This allows for accurate digitizing without causing the imagery to become pixilated. Old and new feature classes can then be superimposed on each other to visualize the changes that have occurred near and on this river over the 24 year period. The resulting data can be used to assist in river rehabilitation. The data can also be used to show the effect that sand and gravel mining can have on a river and its major parts. To determine whether or not there is a statistical relationship between sand and gravel mining and channel change, Spearman s Rank Correlation Coefficient was utilized. Many mining and point bar variables were found to be positively correlated including both negative and positive lags. The new point bar area lags were strongly correlated with mining variables, while the new number of point bar lags showed weak or no correlation. The newer mining variables were typically more strongly correlated with channel change variables opposed to the older mining variables. Point bar area and number has increased over the 24 year period, indicating instability in the river system. An increase in mining and stream power compared to 1980 could be the cause, introducing more sediment into the river creating larger and more point bars. This research could lead to less invasive mining practices and a greater awareness regarding environmental restoration to the Tangipahoa River and its floodplain.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Steven Marks.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Mossa, Joann.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041740:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041740/00001

Material Information

Title: A Floodplain Mining And Channel Change Analysis Of The Tangipahoa River, Louisiana Using GIS 1980 - 2004
Physical Description: 1 online resource (74 p.)
Language: english
Creator: Marks, Steven
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: channel, gravel, louisiana, mining, river, sand, spearman, tangipahoa
Geography -- Dissertations, Academic -- UF
Genre: Geography thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science A FLOODPLAIN MINING AND CHANNEL CHANGE ANALYSIS OF THE TANGIPAHOA RIVER, LOUISIANA USING GIS: 1980-2004 By Steven R. Marks May 2010 Chair: Joann Mossa Major: Geography Sand and gravel mining in floodplains is widespread in Louisiana and can often result in changes in channel planform and position. The purpose of this study was to determine whether or not sand and gravel mining and river channel change are statistically linked over two time periods: 1980 and 2004. The area of focus was the Tangipahoa River, located in southeast Louisiana. This particular river was chosen due to its extensive sand and gravel mining operations within its floodplain. The 2004 time step imagery was Digital Orthorectified Quarter Quadrangles (DOQQ) downloaded from Louisiana s ATLAS website. The imagery for the 1980 time step was USGS 1:24,000 DRG s which was downloaded from the aforementioned website. Using GIS, two different geodatabase s (GDB) were created for each time step. These GDB s contain feature classes that represent the Tangipahoa s centerline, main channel, point bars, cleared land within the floodplain, and sand and gravel pits. In ArcMap, each feature was digitized on top of the corresponding imagery. All features were digitized at a scale of 1:2,000. This allows for accurate digitizing without causing the imagery to become pixilated. Old and new feature classes can then be superimposed on each other to visualize the changes that have occurred near and on this river over the 24 year period. The resulting data can be used to assist in river rehabilitation. The data can also be used to show the effect that sand and gravel mining can have on a river and its major parts. To determine whether or not there is a statistical relationship between sand and gravel mining and channel change, Spearman s Rank Correlation Coefficient was utilized. Many mining and point bar variables were found to be positively correlated including both negative and positive lags. The new point bar area lags were strongly correlated with mining variables, while the new number of point bar lags showed weak or no correlation. The newer mining variables were typically more strongly correlated with channel change variables opposed to the older mining variables. Point bar area and number has increased over the 24 year period, indicating instability in the river system. An increase in mining and stream power compared to 1980 could be the cause, introducing more sediment into the river creating larger and more point bars. This research could lead to less invasive mining practices and a greater awareness regarding environmental restoration to the Tangipahoa River and its floodplain.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Steven Marks.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Mossa, Joann.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041740:00001


This item has the following downloads:


Full Text

PAGE 1

1 A FLOODPLAIN MINING AND CHANNEL CHANGE ANALYSIS OF THE TANGIPAHOA RIVER, LOUISIANA USING GIS: 1980 2004 By STEVEN R. MARKS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF GEOGRAPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Steven R. Marks

PAGE 3

3 To my wife, Cristina

PAGE 4

4 ACKNOWLEDGMENTS I would first and foremost like to thank my graduate chair, Dr. Joann Mossa, for her extreme patience, guidance, and suppor t th roughout this thesis process. I would also like to thank my other committee memb ers, Drs. Binford and Bejleri for their assistance and for being a part of my committee. Ursula Garfield and Keith Yearwood were also influential thanks to their GIS knowledge. They not only made me more comfortable with GIS, but also helped me complete my thesis. I would also like to thank my father, Ron Marks, for his support and statistical expertise. I would like to thank my mother and sister for being t here for me throughout this process as well. And last but not least I profusely express my gratitude to my wife, Cristina, for her constant encouragement, love, and school teacher patience during my graduate school experience.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDG MENTS .................................................................................................. 4 LIST OF TABLES............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 8 ABSTRACT................................................................................................................... 10 CHA PTER 1 INTRODUCTION.................................................................................................... 12 Object ives............................................................................................................... 13 Study Site............................................................................................................... 14 2 LITERATURE REVIEW .......................................................................................... 18 3 METHODS AND MATERIALS ................................................................................ 24 Geographic Informat ion Syst ems ............................................................................ 25 Change I ndice s ....................................................................................................... 28 Average Lateral Migrat ion ....................................................................................... 29 Average Channe l Width .......................................................................................... 29 Point Ba r Area ........................................................................................................ 30 Mining Area ............................................................................................................. 30 Statistical Analysis .................................................................................................. 30 4 RESULT S............................................................................................................... 39 Change I ndice s ....................................................................................................... 39 Average Lateral Migrat ion ....................................................................................... 40 Average Channe l Width .......................................................................................... 40 Point Ba r Area ........................................................................................................ 41 Mining Area ............................................................................................................. 42 Statistical Analysis .................................................................................................. 44 5 DISCUSSION......................................................................................................... 54 6 CONCLUSIONS AND F URTHER RE SEARCH ...................................................... 62 Limita tions............................................................................................................... 63 Further Re search .................................................................................................... 64 APPENDIX: FIGURES SHOWING CHANNEL CHANGE AND PIT CAPTURE .......... 66

PAGE 6

6 LIST OF RE FERENCES ............................................................................................... 71 BIOGRAPHICAL SKETCH ............................................................................................ 74

PAGE 7

7 LIST OF TABLES Table page 3-1 Topolog y rules .................................................................................................... 383-2 Proportional area change ratios (Mossa and Mc Lean 1997) .............................. 384-1 Totals of variables compared fo r 1980 and 2004................................................ 524-2 Results of Spearmans Rank Test for Correlation. The bold values represent the statistically significant val ues at the 95% confidence level........................... 524-3 Results of Spearmans Rank Test fo r Correlation with relationships involving point bar variables. The bold values represent the statistically significant values at the 95% confidence level.................................................................... 524-4 Results of Spearmans Rank Test fo r Correlation with relationships involving point bar variables with negative lags. The bold values represent the statistically significant values at the 95% c onfidence level................................. 534-5 Distances from pit captures cent roids and closest point to the 1980 main channe l............................................................................................................... 534-6 Closest distance of the 1980 mine pi ts to the 1980 main channel (p=0.006)...... 534-7 Area of the 1980 mi ne pits (p=0.884)................................................................. 53

PAGE 8

8 LIST OF FIGURES Figure page 1-1 Tangipahoa Riv er site map ................................................................................. 16 1-2 Tangipahoa Riv er drai nage bas in ....................................................................... 17 2-1 Sand and gravel mining produc tion on the Tangipahoa, 1932-2009. ................. 23 3-1 Imagery fly dates in comparison to discharge at Robert, LA. ............................. 35 3-2 Tangipahoa site map with tr ansects ................................................................... 36 3-3 Change indices methods on the Tangi pahoa Riv er ............................................ 37 4-1 Mine pit area on the Tangipahoa River 1980-2004 ............................................. 46 4-2 Number of mine pits on the Tangipahoa River 1980-2004 ................................. 46 4-3 Bare Area on the Tangipahoa Ri ver 19802004 ................................................. 47 4-4 Change indices on the Tangipahoa Riv er 198 0-2004 ......................................... 47 4-5 Tangipahoa River lateral migration rates per year 1980-2004 ............................ 48 4-6 Example of pit capture influencin g channel width on the Tangipahoa (1980) ..... 49 4-7 Average channel width on the Tangipahoa River 1980-2004 ............................. 50 4-8 Point bar area on the Tangipahoa Riv er 198 0-2004 ........................................... 50 4-9 Number of point bars on the Tangipahoa River 1980-2004 ................................ 51 4-10 Cumulative point bar area on the Tangipahoa River 198 0-2004 ........................ 51 5-1 Area of pit capture in transect 9 ( 2004 imag ery) ................................................. 57 5-2 Area of pit capture in transect 12 ( 2004 imag ery) ............................................... 58 5-3 Area of pit capture in transect 13 ( 2004 imag ery) ............................................... 59 5-4 Area of pit capture in transect 23 ( 2004 imag ery) ............................................... 60 5-5 Area of pit capture in transects 24 and 25 (2004 im ager y)................................. 61 A-1 Unstable reach in transects 6-12 ........................................................................ 66 A-2 Unstable reach in transects 23-29 ...................................................................... 67

PAGE 9

9 A-3 Stable reach in transects 46-52 .......................................................................... 68 A-4 Stable reach in transects 50-56 .......................................................................... 69 A-5 Stable reach in transects 57-61 .......................................................................... 70

PAGE 10

10 Abstract of Thesis Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for t he Degree of Master of Science A FLOODPLAIN MINING AND CHANNE L CHANGE ANALYSIS OF THE TANGIPAHOA RIVER, LOUISIANA USING GIS: 1980-2004 By Steven R. Marks May 2010 Chair: Joann Mossa Major: Geography Sand and gravel mining in floodplains is widespread in Louisiana and can often result in changes in channel planform and posit ion. The purpose of this study was to determine whether or not sand and gravel mining and river channel change are statistically linked over two time periods : 1980 and 2004. The area of focus was the Tangipahoa River, located in southeast Louisiana. This particular river was chosen due to its extensive sand and gravel mining operations within its floodplain. The 2004 time step imagery was Digital Orthorectified Quarter Quadrangles (DOQQ) downloaded from Louisianas ATLAS webs ite. The imagery for the 1980 time step was USGS 1:24,000 DRGs which was downloaded from th e aforementioned website. Using GIS, two different geodatabases (GDB) were created for each time step. These GDBs contain feature classes that represent the T angipahoas centerline, main channel, point bars, cleared land within the floodplain, and sand and gravel pits. In ArcMap, each feature was di gitized on top of the correspondi ng imagery. All features were digitized at a scale of 1:2,000. This allows for accurate digitizing without causing the imagery to become pixilated. Old and new feature classes can then be superimposed on each other to visualize the changes that have occurred near and on

PAGE 11

11 this river over the 24 year period. The re sulting data can be used to assist in river rehabilitation. The data can also be used to s how the effect that sand and gravel mining can have on a river and its major parts. To determine whether or not there is a statistical relationship between sand and gravel mining and channel change, Spearm ans Rank Correlation Coefficient was utilized. Many mining and point bar variabl es were found to be positively correlated including both negative and positive lags. The new point bar area lags were strongly correlated with mining variables, while the new number of point bar lags showed weak or no correlation. The newer mining variables were typically more strongly correlated with channel change variables opposed to the older mining variables. Point bar area and number has increased over the 24 year period, indicating instability in the river system. An increase in mining a nd stream power compared to 1980 could be the cause, introducing more sedi ment into the river creating larger and more point bars. This research could le ad to less invasive mining practices and a greater awareness regarding envi ronmental restoration to the Tangipahoa River and its floodplain.

PAGE 12

12 CHAPTER 1 INTRODUCTION Extraction sites for aggregate resource s occur in uplands, river floodplains, and in-stream for most of the United States. The Tangipahoa River in southeast Louisiana is no exception, especially in its middle and upland reaches. The Tangipahoa basin is a site of both wet and dry mining. Dry mining (open pit) occurs in the upland areas of the river where the water table is located below the base of the pit. Wet mining takes place in the midland settings where dredges are us ed to mine the sediment from the rivers floodplain paleochannels (Mossa and Autin 1998). Sand and gravel mining has occurred on this river for many decades, potentially causing t he channel to meander and sandbars to change in size as sediment is removed or added to the system (Mossa and Autin, 1998). Sand and gravel mining can have negative effects on both ecological and geomorphic aspects of river systems. The reduction of sediment through in-stream mining can cause upstreamand downstream-pr ogressing river incision, lateral channel instability, and bed armoring. Degradation or even total loss of aquatic and riparian habitats can also occur from sand and gravel mining (Rinaldi, 2005). The most obvious and possibly largest changes to a river system are planform changes. These types of changes can great ly impact the frequency of floodplain inundations, lower the valley-floor water t ables, and cause instability or even destroy bridges and channelization stru ctures (Rinaldi, 2005). Pr ofile changes include channel migration, an increase or decrease in the number and size of point bars, and the erosion of channel banks (Mossa, 2004). As a result of these geomorphic changes the hydrology of the river system is also altered. Mining can remove bed material from

PAGE 13

13 rivers, thus causing degradation. This can create a phenomenon known as head cutting that can have an effect upstream fo r many kilometers (Rinaldi, 2005). An increase in sediment load can result in aggradation downstream, and channel capacity will decrease causing flooding to become more widespread (Mossa, 199 5). Objectives The main reason for this study was to determine what effects sand and gravel mining had on the Tangipahoas g eomorphology. The first objective of this study was to see if reaches that had heavy mining had greater rates of channel change. The next objective was to determine whether or not mining had an effect on sand bar frequency and size change. These objectives were accomplished by studying the Tangip ahoa over a 24 year period. This helped determine whether or not there was a relationship between sand and gravel mining and channel change. It is believed that ther e is a statistical relationship between sand and gravel mining and the morphological changes that the Tangipahoa River has undergone ov er the 24 year period. This relationship can be investigated through both spatial and temporal analyses. The findings of this study can serve a wide purpose for state governments and other management agencies. Results from th is study could possibly lead to better mining practices, such as developing po licy about the more egregious practices and exploration for future mining sites. There is a great economic benefit and need for sand and gravel mining, but more scientific info rmation is required to better understand the environmental effects and the economic benefits of mining.

PAGE 14

14 Study Site The Tangipahoa River stretches 177 kilomete rs total (127 km in Louisiana) from its headwaters in southeastern Mississippi a nd flows south towards Lake Pontchartrain (Figure 1-1) in southeastern Louisiana (Envir onmental Protection Agency, 2005), with a drainage basin of 878 km in Louisiana (F igure 1-2). The T angipahoa River runs through the Tangipahoa Parish, passing closely by its major city, Hammond. Elevations in the Tangipahoa Parish range from 82 m in Spring Creek to 0.9 m in Ponchatoula (USGS, 2009). The average annual precipitation for the area of study is approximately 1,676 mm with floods mainly caused by fronta l systems, convective storms, tropical storms and hurricanes (Louisiana Office of State Climatology, 2000). The Tangipahoa River floodplain is comprised mainly of Alluvium, which is highly suitable for sand and gravel mining operations. Areas of Alluvium have water tables near the surface where certain areas can contain abundant deposits of gravel (Mossa and Autin, 1998). This alluvium contains gravels reworked from the PliocenePleistocene Citronelle formation, which comprises much of the basins uplands. The area of focus for this study star ted at the Kentwood gage station to 7 km south of Interstate-22. From this point on mining is very scarce due to decreases in bed material size in modern and ancestral rive r deposits (Mossa and Autin, 1998). The United States Geological Survey (USGS) has three continuous gaging stations located in Kentwood, Amite, and Robert, Louisiana. The Robert gage station was used for discharge data since it was the only station that had data from the 1980s. Aggradation of 0.75 to 1 meter was found at this st ation between 1982-1994 (Wilder, 1998). The Tangipahoas River basin mainly consists of riparian forests and agricultural land, dotted with small urban areas.

PAGE 15

15 The Tangipahoa was once considered a polluted and impaired river due to high levels of animal and human fecal matter runo ff. Thanks to 20 years of river clean up and strict enforcement against dumping, the Tangipahoa has been taken off the list of impaired waters by the National Resour ces Conservation Service and now supports both primary and secondary contact recreational uses (NRCS, 2005).

PAGE 16

16 Figure 1-1. Tangipahoa River site map

PAGE 17

17 Figure 1-2. Tangipaho a River drainage basin

PAGE 18

18 CHAPTER 2 LITERATURE REVIEW Sand and gravel is used in many everyday materials and applications, ranging from roads to building material s. This type of mining is very common near alluvial systems. For millions of years, vertical and lateral erosion within a rivers floodplain has reworked gravel-bearing deposits such that modern alluvium contains an abundance of sand and gravel (Rinaldi, 2005). These resour ces are dominantly located in or near river channels, which then become sites for mining. Sand and gravel mining could have a negative effect on an alluvial system. These effects range from a change in the sediment budget, lateral migration of a rive r channel, or effect s on the aquatic and vegetative habitats (Rinaldi, 2005). Some st ates, such as Arkansas and Missouri, have placed restrictive laws on sand and gravel mini ng. In Arkansas in particular, mining permits must be obtained from the Surface Mi ning and Reclamation Div ision, part of the Arkansas Department of Pollution Control and Ecology (ADPCE). Permitted mining can be conducted in upland areas where in bank sand and gravel deposits occur below high-water marks. Six requirements must be me t in order to attain a permit: Proof of right to mine land, a mining plan, and a re clamation plan to name a few (Rogue Valley Council of Governments, Oregon, 1997). Laws in Louisiana for sand and gravel mining are loosely enforced compared to the two af orementioned states. Projects like this could help push Louisiana and other state gove rnments to increase regulations on the sand and gravel mining industry. With sand and gravel mining continuing to occur in or near rivers, questions about its physical and environmental effects have increased. In the 1950s and 1960s, sand and gravel mining occurred almost ex clusively from the point bars and channel

PAGE 19

19 bottoms within the river itse lf. More recently, mining has occurred on adjacent floodplain surfaces and low terraces, where practices have included removal of vegetation and fine-grained overburden, and digging of pits and ponds (Mossa and Autin, 1998). These types of practices can gr eatly affect a river physically, ecologically, and environmentally (Rinaldi, 2005). Sand and gravel mining within a river system is very popular for two major reasons; first, in-stream gravel acquired from an active channel are especially important because river transport removes weak materials by abrasion and attrition. The remaining aggregate are a higher quality: durable, rounded, well -sorted, and relatively free of interstitial fine sediment (Kondolf, 1994) Active channel sediments can be easily quarried (deep quarrying is not necessary), r equire very little processing, and are periodically replaced from upstream during high flow events (Rinaldi et al., 2005). If mining equipment can stay in one place for a longer period, it saves time and money. These areas of mining are also often located closer to markets that use the aggregate or are close to transportation routes, again saving money (Kondolf, 1994). However, the costs to the environment from in-str eam mining are often not factored into production costs, making in-stream aggregate mo re economically attractive compared to other alternatives (dry terrace, quarries or distant sources) (Kondolf, 1994). This makes aggregate mining a very cost effectiv e way to acquire resources for many different types of comm ercial projects. The forms of sand and gravel mining with the most potential to affect a river system include floodplain and in-stream mining. Floodplain mining occurs in pits, which are areas where floodplain vegetation (us ually bottomland forest) and an overburden

PAGE 20

20 consisting of more organic and fine materials than beneath have been removed. The sand and gravel are removed from these pits t hat essentially leave large holes next to the river. In-stream mining is the extraction of sediment from an active channels bed, often by heavy machinery (Kondolf, 1997). Both of these practices have detrimental effects on the river itself and the fl oodplain that borders it. Floodplain mining has occurred fo r decades along the Tangipahoa River. Records kept beginning in 1932 indicate that gravel has been the most popular aggregate mined from the Tangipahoa floodplai n. The largest amount of gravel removed occurred in 1967, with 1,419,433 tons ex tracted. The largest amount of sand was removed in 1974, with 921,230 tons extrac ted. The majority of sand and gravel mining took place between the years of 1953-1983 with a steep decrease in production bottoming out in 1992 with almost the same amount of production for both types of aggregate (Figure 2-1). There is a dat a gap between 1992 and 2004 due to the repeal of a sand and gravel tax. The tax was reinst ated for sand in 2005, but not for gravel (Louisiana Department of Revenue). This pr ocess occurs by extracting sand and gravel from pits along active floodplains or adjacent river terraces. For the Tangipahoa the mining pits are wet which occurs when the pi t is below the water table (Kondolf, 1994). These pits tend to be located right next to the river. During fl ood events or lateral channel migration, these pits can be breache d and can become part of the river channel (avulsion or pit capture). As a result these pits become in-stream and can cause similar issues as in-stream mining, such as headcutting (Kondolf, 1997). These pits, lying close to the river, can often intersect t he water table essentia lly contaminating the aquifer (Rinaldi et al, 2005), although this is not the case for the Tangipahoa since the

PAGE 21

21 aquifer is significantly lower than the depths of the pits. In areas that have heavy floodplain mining, it has been seen that channel length can be shortened (Mossa, 1983). Other geomorphic changes to the river can occur both upstream and downstream. Some of these changes are riverbank erosion, cutoffs, avulsions, aggradation and degradation (Mossa, 1995). In areas of aggradat ion, channel capacity is decreased due to an increase of sediment from erosion which can lead to large scale flooding and increased point bar area and num bers. Sediment transport is a slow process and can be deposited on the channel beds, banks and bars, and can remain in place for highly variable time periods unt il they are remobilized and moved further downstream (Jacobson et al, 1999). Degradat ing reaches have an increased channel capacity leading to erosion that can undermine bridges, pipelines, and ot her structures. Implications to humans can range anywher e from property loss, land disputes, and increased taxation. These issues can become expensive and a risk to public safety (Mossa and Autin, 1998). In-stream gravel mining is done with heavy machinery, such as dredges, and involves the removal of sand and gravel from active river channels and streams. This removal can occur using two different methods; excavating trenches or pits in the gravel bed, or by gravel bar skimming. Bar skimmi ng withdraws sediment from the transport system altering the supply to downstream re aches. Although volumes of sediment are typically smaller than pit mining, profound effects can still occur to aquatic habitats downstream (Kondolf, 1997). Active channel deposits are favorable for construction projects because they are typically durable, well-sorted, and are normally near markets

PAGE 22

22 or transportation routes, thus lowering costs (Kondolf, 1994). In-s tream mining can directly alter the channel geometry and bed elevation and may involve extensive clearing, diversion of flow, stockpiling of sediment, and excavation of deep pits (Sandecki, 1989). The removal or introduction of sediment to a system due to sand and gravel mining can greatly alter a river s geomorphology. W hen sediment has been removed from an active channel due to in-str eam mining, a rivers transport capacity will increase causing bed erosion, a process referred to as headcutting or knickpoint migration (Kondolf, 1994). These affects can be seen for miles upstream. Incision can also occur downstream causing massive er osion which could affect the frequency of floodplain inundation along the ri ver courses, lowering valley floor water tables, and frequently leading to the destruction of brid ges and channelization structures (Rinaldi, 2005). In-stream mining also introduces major environmental issues to a river system. With the removal of sediment, erosion c an become widespread and cause loss of land and wooded areas. The loss of aquatic and ripar ian habitats can also result from instream mining (Sandecki, 1989).

PAGE 23

23 0 200000 400000 600000 800000 1000000 1200000 1400000 1600000 1920193019401950196019701980199020002010 YearMining Production Gravel Sand Figure 2-1. Sand and gravel mining production on the Tangipahoa, 1932-2009.

PAGE 24

24 CHAPTER 3 METHODS AND MATERIALS Certain methods for this thesis were adopt ed from the theses of Ursula Garfield and April Hendrix-Davis thanks to simila rities between these three projects. Geographical Information Systems (GIS) was an integral part of this project. A 1998 time step was also considered, but ther e was a close similarity between the1998 and 2004 main channel, mine pits, and sand bars. As a result, the 1998 time step was not used. Vector GIS was used for the s patial analysis thanks to its accuracy when delineating centerlines, main channels, mine pits, and point bars. Vector GIS is also advantageous when calculating abnormal polygons, i.e. mi ne pits. These vector features from each time period were ov erlaid and analyzed to determine the changes that the Tangipahoa has incurred. Excel and the Sigma Stat 3.1 were other types of software that were used for quantitative analysis. Excel was used to create spreadsheets and graphical displays. Sigma Stat 3.1 was utilized to carry out the statistical analysis. The imagery was downloaded from Louisi anas ATLAS websit e: The Louisiana Statewide GIS (LSU, 2009). Digital Ort hophoto Quarter Quadrangles (DOQQs) were downloaded for the 2004 time st ep. Imagery acquisition dates for the 2004 imagery varied from May 11th to May 23rd, 2004. Each image covers an area of 4 miles wide by 4.5 miles in the north/south direction. Images have been acquired using color infrared with a horizontal accuracy of approximately three (3) feet (ATLAS) and a scale of 1:12,000 (1 inch = 1,000 feet) with an inheren t error of 10m for length and 100m for area (USGS). As for the 1980 time step, no DOQQs were available. Instead, USGS Digital Raster Graphics were used (ATLAS) DRGs are scanned images from USGS

PAGE 25

25 topographical maps and georefer enced to the earths surface and fit to the UTM projection (USGS). These DRGs are 7.5 q uadrangles at a scale of 1:24,000 (1 inch = 2,000 feet) acquired from aerial photography taken between the dates of January 12th to January 27th, 1980. The DRGs have an inherent error of 12.2m for length and 148.84m for area (USGS). The DOQQs, DRGs, and feature classes share NAD 1983 UTM Zone 15 North as their spatial reference. The images were used to determine the location of mining and channel change variables and to determine how much the Tangipahoa has moved over the 24 year period. Discharge data was acquired from the USGS for thr ee gage stations: Kentwood, Amite, and Robert, Louisiana (USG S, 2009). Kentwood is t he northernmost station but has no discharge data recorded. The Amite gage station is in the center of the study area and also has no discharge data. The Robert gage station is the southernmost station on the Tangipahoa, and was the only station that had discharge data starting in 1938. Discharge data was included for t he Robert gaging stat ion from 1938-2004 with the flood stage of 5750 ft/sec (Figure 3-1). Geographic Information Systems All of the vector-based data creation and analysis occurred in the ESRI ArcGIS software suite. Two different geodatabases we re created for each separate time step. Each geodatabase contained feature classes to represent the variables that were digitized. The feature classes that were cr eated were the river c enterline, main channel, point bars, sand and gravel mining pits, and bare area. All relevant information (shape length, shape area, reach block ID for exampl e) were saved in attribute tables that correspond with each feature class. The nu mber of gravel pits, for example, was determined by using these attribute tables. Gravel pits from each time step can be

PAGE 26

26 compared to determine whether the amount of gravel pits have increased, decreased, or stayed the same. The main channel of the Tangipahoa was di gitized separately for each timestep. This gave the rivers width which allowed for analysis based on area. The scale at which the main channel was digitized was 1:2,000 for DOQQs. This scale provides detail of the channel without bei ng zoomed out too far where important features could be missed. DRGs were used to digitize the main channel for the 1980 timestep. Next the rivers centerline was digitized separately for each time step using the midpoint tool. With snapping turned on, a ve rtex would be placed on one side of the river polygon followed by a vertex on the op posite side. A vertex would be dropped in the center of the two other vertices. This process w ould be continued for the entire river linking each center vertex, thus creating the c enterline. The centerline is important in creation of the widths of each transect and lateral migration rates. Point bars were the next feature to be di gitized. They were represented on the DRGs by white areas outlined by brown dots. These were digitized as polygons and snapped to the main river channel. On the D OQQs, the point bars we re often found in the river bends and were identified by sandy areas. These digitized polygons were also snapped to the river channel to ensure no over lap or gaps between polygons. After the point bars were digitized, pol ygons were created to capture mine pits and bare area. These areas on the DRGs were normally demarcated by blue and white areas representing mine pits and cleared land respectively. In some cases, a mining symbol (crossed shovels) was presen t. Areas without the symbol were mapped by interpretation, estimating where the pits and cleared land were. Mining pits that

PAGE 27

27 occurred in the floodplain were delineated by contour lines and accompanied by the words sand and gravel pits. On the Tangip ahoa the majority of these pits were filled with water. On the DOQQs, aerial photo interpretation guided the digitizing of bare area and mining pits. The bare area appeared to be cleared, sandy land next to the mine pits with high reflectance. The boundaries of the sandy areas were digitized right up to the areas of vegetation. The mine pits, which in most cases bordered the bare area, were digitized to the ex tent of the water that filled them. All areas for mine pits and point bars were rounded to the nearest hundr edth to account for the inherent error that occurs in the DRGs and DOQQs. As previously mentioned feature cla sses were created for both sand and gravel pits and bare area. These were digitized to determine if avulsions or captures had occurred. Avulsions occur when a newer channel has intersected an earlier sand and gravel pit. Avulsions are discovered by simp ly comparing older sand and gravel pits to the newer channels. If they intersect at any point an avulsion had occurred and a point was placed on that particular spot in ArcM ap. Corresponding attributes were created with each point describing if the avulsion happened inside or out side the meander bend. Avulsions can be caused by lateral migrat ion, channel widening due to in-stream mining, or flood events. To compare the digitized shapefiles from the two time steps, they were converted into feature classes in a personal database. This process must be done in a personal geodatabase in order to creat e relationships between feat ure classes, generate and run topology rules (Table 3-1), and to calculate the area of each polygon. Topology rules

PAGE 28

28 were used to ensure that there were no gaps or overlaps between particular feature classes. Once the previous actions have been completed, a valley centerline was generated based off the flow direction of th e main channels. This centerline was needed to create transects that run perpendicular to the rivers flow. These transects were spaced at a distance of 1 kilometer for the entire length of the river. The width of each transect was determined by using equation 3-1: Width = Area/Channel Length Centerline (eq. 3-1) The river channel and centerline was intersec ted by the transect feature class to calculate area and channel length by transect. This feature class was created using the measure tool. A point was added on the cent erline for every kilometer until the end of the river was reached. Next, lines were ma nually added at each point to create reach blocks. If a transect crossed the river centerline at more than one point, the transect was moved slightly up or down along the cent erline. These transects allow for the summarization of certain areas and indices by reach (Figure 3-2). Change Indices To analyze the Tangipahoas channel chang e four indices were created. First, the main channels from each time period were combined in ArcMap using the union tool. This tool computes a geometric inte rsection between the two main channels. The no gaps check box located in the tool menu must be checked in order to create the between area index. Once the channels were combined, the fo ur indices were created. This was done by loading the unioned main channel feature class into the single to multi-part tool. This explodes the unioned main channel feature into single polygons. The features and attributes of the main channels were stor ed in the new output feature

PAGE 29

29 class. These indices represent how the ri ver channels have changed over time. If the channel crosses at the same point during both time steps, the index was labeled as unchanged (U). The new channel area index va lue was labeled as erosion (E). The third index value represented the old channel, which was l abeled as deposition (D). The fourth and final index will represented the area that is between the old and the new channels, and was labeled as area in between (B). Any area that fell outside the unioned main channel was labeled as other (O) (Mossa, 2006) (Figure 3-3 and Table 32). Average Lateral Migration The average lateral migration was calc ulated in ArcGIS using the centerline shapefiles of the two time per iods. The feature to polygon tool in the Data Management toolbox was used to create new polygons between the two centerlines. New fields were added and calculated for shape le ngth and area of the polygons. This new polygon shapefile was then intersected by the transec t shapefile to split the polygons by reach block. This data were then exported and so rted in Excel. Using equation 3-2, The migration rates were then calculated by dividing the area of the polygon by the perimeter of the two centerline lengths, and then divided by the difference in years between the two time steps: Lateral Migration = (A/L)/24 (eq. 3-2) The results were then graphed in Excel. Thes e methods were drawn from Larsen et al. 2006. Average Channel Width The average channel width was calculat ed by dividing the area of the main channel by the length of the centerline. Both of these features were already digitiz ed,

PAGE 30

30 but were exported to new shapefiles to keep the integrity of the origin al shapefiles. Both shapefiles were then intersec ted with the transect shapefile to divide the main channel and centerline by reach. The calculate geo metry feature was again used to calculate both the length of the center line and the area of the main channel. These attributes were then exported, sorted, calcul ated, and graphed using Excel. Point Bar Area Point bar area (m) was calculated usi ng the calculate geometry feature in ArcGIS located in the attribute table. T he point bar shapefile was then intersected with the transect shapefile in order to identify whic h point bars occur in which reach. This data was then exported to Excel and sorted by reach then by area. Point bar area, number of point bars, and cumula tive point bar area were all calculated and graphed vs. reach block number. Mining Area Mining area was the sum of mine pits and bare area. The methods used to digitize and calculate mine pits and bare area we re identical to point bar area. Statistical Analysis In order to determine if channel change and mi ning activity are statistically linked, a measure of correlation must be used. Since the data wasnt normally distributed, Spearmans Rank Correlation was chosen bec ause it is a non-parametric statistic (Earickson and Harlin, 1994). Spearmans Rank Correlation equation 3-3: (eq. 3-3) In order to run this statisti c, all data from ArcMap were imported into Microsoft Excel then sorted. Next, the data wa s imported into Sigma Stat 3.1 software which was used

PAGE 31

31 to calculate the Spearman Rank Correlati on. Since tied ranks can occur in the Spearman Rank Correlation, an adjustment for ties was impl emented. Values in the Spearman Correlation can range from (-1) to (+1) Rank values that are close to (-1) indicate a complete discordance. Rank val ues close to a (+1) indicate a complete concordance. Rank values that are close to (0) indicate little or no relationship between the two variables that are being compared. The rank correlation was run on all variables that may be linked by physical proce sses. Below are the variables that were compared: New Pit area vs. Between area New Pit Area vs. Old Channel Area New Pit Area vs. New Channel Area New Pit Area vs. Overlapping Old and New Channels New Number of Pits vs. Between Area New Number of Pits vs. Old Channel Area New Number of Pits vs. New Channel Area New Number of Pits vs. Ov erlapping Old and New Channels New Mining Area vs. Between Area New Mining Area vs. Old Channel Area New Mining Area vs. New Channel Area New Mining Area vs. Over lapping Old and New Channels Old Mining Area vs. New Point Bar Area Old Mining Area vs. New Number of Point Bars Old Mining Area vs. New Point Bar Area

PAGE 32

32 Old Number of Mining Pits vs. New Point Bar Area Old Number of Mining Pits vs. New Number of Point Bars New Pit Area vs. Larsen Lateral Migration New Mining Area vs. Lar sen Lateral Migration New Number of Mine Pits vs. Larsen Lateral Migration Old Mining Area vs. Larsen Lateral Migration Old Number of Pits vs. Larsen Lateral Migration Old Mining Area vs. New Channel Width Old Number of Mining Pi ts vs. New Channel Width New Channel Length/Reach Block vs. New Number of Point Bars Old Number of Mining Pits vs. New Number of Point Bars Lag 1 Old Mining Area vs. New Number of Point Bars Lag 1 Old Number of Mining Pits vs. New Point Bar Area Lag 1 Old Mining Area vs. New Point Bar Area Lag 1 New Number of Mine Pits vs. New Number of Point Bar Lag 1 New Mining Area vs. New Nu mber of Point Bars Lag 1 New Number of Pits vs. New Point Bar Are Lag 1 New Mining Area vs. New Point Bar Area Lag 1 Old Number of Mining Pits vs. New Number of Point Bars Lag 2 Old Mining Area vs. New Number of Point Bars Lag 2 Old Number of Mining Pits vs. New Point Bar Area Lag 2 Old Mining Area vs. New Point Bar Area Lag 2 New Number of Mine Pits vs. New Number of Point Bar Lag 2

PAGE 33

33 New Mining Area vs. New Nu mber of Point Bars Lag 2 New Number of Pits vs. New Point Bar Are Lag 2 New Mining Area vs. New Point Bar Area Lag 2 Old Number of Mining Pits vs. New Number of Point Bars Lag (-1) Old Mining Area vs. New Number of Point Bars Lag (-1) Old Number of Mining Pits vs. New Point Bar Area Lag (-1) Old Mining Area vs. New Point Bar Area Lag (-1) New Number of Mine Pits vs. New Number of Point Bar Lag (-1) New Mining Area vs. New Numb er of Point Bars Lag (-1) New Number of Pits vs. New Point Bar Are Lag (-1) New Mining Area vs. New Point Bar Area Lag (-1) Old Number of Mining Pits vs. New Number of Point Bars Lag (-2) Old Mining Area vs. New Number of Point Bars Lag (-2) Old Number of Mining Pits vs. New Point Bar Area Lag (-2) Old Mining Area vs. New Point Bar Area Lag (-2) New Number of Mine Pits vs. New Number of Point Bar Lag (-2) New Mining Area vs. New Numb er of Point Bars Lag (-2) New Number of Pits vs. New Point Bar Are Lag (-2) New Mining Area vs. New Point Bar Area Lag (-2) These associations were used due to the likelihood of geomorphic interactions. The variables for mining areas were used as independent variables to determine the chance of both mined areas and pits being link ed to channel change. Associations of dependent and independent variables were used bas ed off of the results of a previous

PAGE 34

34 study on the Amite River (McLean, 1995). Spat ial lags were created with the variables new number of point bars and poi nt bar area. These lags are used to decide if the mining that occurred in one reach impacted the number of point bars or point bar area within a reach or two downstream. The first lag is spatially one reach block downstream, while the second lag is two r each blocks downstream. Negative lags were also considered to determine if head cutti ng occurred upstream. Spearmans rank values were tested for statistical significance utilizing the t-distribution. The equation for t-value was altered by isolating the r, whic h is the critical value of Spearmans r for statistical significance. This was ca lculated using equation 3-4 (Rogerson, 2001) (eq. 3-4) The ts in the equation represent the values for the t-distribut ion for 67 degrees of freedom and n is the sample size. Because ti es existed in the data, Pearsons Moment Correlation Coefficient was used to adjus t the ranked data (Myers and Well, 2003). The above combinations were all run using the Pearsons Correlation Coefficient on ranked data. The Mann-Whitney U-Te st was run to determine whether the distance and area of the 1980 mine pits and avulsions that occurred on the 2004 main channel were statistically significant.

PAGE 35

35 100 1000 10000 100000 1940195019601970198019902000 YearQ (ft^3/sec) Photo Dates Discharge Flood Stage: 5750 ft/s Figure 3-1. Imagery fly dates in com parison to discharge at Robert, LA.

PAGE 36

36 Figure 3-2. Tangipahoa si te map with transects

PAGE 37

37 Figure 3-3. Change indices me thods on the Tangipahoa River

PAGE 38

38 Table 3-1. Topology rules Feature Rule Feature Class Point_Bars04 Must not overlap with Mainchannels_04 Mainchannels_04 Must not overlap with Point_Bars04 Meander_Islands04 Must not ov erlap with Mainchannels_04 Table 3-2. Proportional area ch ange ratios (Mossa and McLean 1997) Indices Description Comment U/I Ratio of U to I Shows proportion of initial channel area in same position D/I Ratio of D to I Shows proportion of initial channel area abandoned B/I Ratio of B to I Shows proportion of initial channel area between channels E/I Ratio of E to I Shows proportion of initial channel area eroded or created I D + U Initial area

PAGE 39

39 CHAPTER 4 RESULTS Multiple var iables were tested to determi ne the stability, or lack thereof, of the Tangipahoa River. Point bar area, mini ng pit and bare area, average channel change, average lateral migration, and change indices were all analyzed to determine if mining played a role in the spatial and tempor al aspects of the Tangipahoa. Spearmans Rank Correlation was used to determine if there was a statistical relationship between the aforementioned variables. Different combinations of mining and channel change variables were used to determine whether or not these variables were correlated. Change Indices Change ratios were used as measure to determine the stability of a river system. The B/I rati o should not be very high when com paring short time periods. But this ratio will increase if meander cutoffs, avulsions, or very rapid late ral migration occur in the river system. The D/I ratio represents the area of the river that was formerly water in an earlier time period, while the E/I ratio repres ents area in the more recent time period that is now currently water or the new river channel. Lastly, the U/I ratio represents the area of the river that is unchanged, or in its initial position (Mossa and McLean 1997). The B/I ratio is generally low in t he reach blocks 1-8 and 32-61 where mining doesnt take place, reaching 0.5 twice. However, the B/I ratio increases significantly in reach blocks 9-32. Three spikes occur where the B/I index is 1.9, 1.4, and 1.8, at reach blocks 9, 12, 24 respectively. The area betw een channels in those reach blocks is an average of more than 170% of th e initial area. Large values of the B/I ratio indicates channel instability. Reach blocks 15, 17, 18, 21, 22, 29, and 30 hav e extremely low B/I

PAGE 40

40 values with an average of 10% change from the initial channel area. As a result, the U/I value is generally lower where the B/I value is higher, and vice versa. The U/I value is much higher in the reaches where mining does not occur topping out at 87% and 89% at reach blocks 2 and 60 respectively, indicating stability. The average U/I index for the whole river system is 42% versus the B/I index of 37%. The D /I index ratio shows variability between reach blocks 8 and 42, with va lues ranging from 0.93 to 0.27, with an average change of 72%. This means t hat 72% of the channel has been abandoned between these reaches or has found a different position. The E/I index value spikes 4 times between reach blocks 14 and 31, with an average index valu e of 1.2 or 120% channel enlargement from the in itial area (Figure 4-4). Average Lateral Migration Lateral migration rates (figure 4-5) fo r the Tangipahoa v aried greatly throughout the entire river system. The highest rates occurred in the mining reaches, where most disturbances occurred. The three highest ra tes were in reach blocks 9, 12, and 24 with lateral migration rates of 2.3, 2.1, and 2.15 meters/yr respectively. The lowest lateral migration rate that occurred in the mined reaches was in block 17, with 0.31 m/yr, highlighting the amount of variability. T he migration rates quickly taper off farther downstream outside the mined reaches. The ov erall average lateral migration rate for the entire Tangipahoa River was 0.73 m/yr, indicating a measure of stability over the study period. Average Channel Width The average channel width for the 1980 timest ep fluctuates greatly, from 26 to 82 meters. This is due to three spikes that occur throughout the stretc h of the river, with the greatest one occurring in a mined reach. T he cause of this spike is pit capture from

PAGE 41

41 earlier mining pits. Previous floods could ha ve forced to main channel to capture these mine pits, thus causing the channel to bec ome much wider (Figure 4-6). The 2004 timestep channel width is more consistent, ranging from 25 to 50 meters. The channel gradually widens downstream in the mined ar ea and peaks at 50 meters in reach 29. The width begins to decrease further downstr eam where it again peaks at 50 meters in reach 41. From there the width greatly tapers off to meet its low at 23 meters in reach 58 (Figure 4-7). Although pit capture occu rs between the 2004 main channel and the 1980 mine pits, this plays little to no role in the 2004 main channels width. More recent biological and physical changes might have affected the 2004 main channel width, such as incision, lower water tables, or both. Point Bar Area When considering point bar variables, it is important to take into account the discharge rates for the days that the im agery was acquired. Unfortunately, the Tangipahoa has three gage stat ions with only one that has dischar ge rates from the 1980s. This gage station, located in R obert, LA., is near the end of the Tangipahoa where point bars rarely occur. This makes it impossible to get a true indication of the discharge rates in the mined reaches of the Tangipahoa. A large discharge rate can both hide point bars and decrease their area. A low discharge rate can reveal hidden point bars and allow for an increase in point bar area. The discharge rate was 2540 ft/s for the 1980 fly dates and 5210 ft /s for the 2004 fly dates. For a true comparison, the actual months being investigated should be t he same with similar discharge rates. In the mined reaches, the number of point bars was relatively similar between the 1980 and the 2004 time steps (59 to 66). Furt her downstream in reach blocks 32 to 61, the number of point bars increases greatly for the 2004 timestep with 72, compared to

PAGE 42

42 the 1980 timestep with 26 point bars. T he highest amount of point bars for each timestep was 9, located in reach block 20 for 1980 and 37 for 2004 (Figure 4-9). The total number of point bars for 1980 is 91 compared to 158 for 2004 (Table 4-1). Point bar area is almost non-existent in reach blocks 1 through 6 and 43 through 61. This is an indication of higher flow rates and lack of mining, allowing sediment to easily pass through the river system. This tr end changes greatly in the reach blocks 7 to 23 where heavy mining occurred. For the 1980 timestep, point bar area spikes twice with an average area of 117,500 m. For the 2004 timestep, point bar area spikes three times for the same reach blocks, with an average area of 143,400 m (Figure 4-8). Throughout the entire Tangipaho a, point bar area increases in every reach block, except for three, from t he 1980 to the 2004 timestep. The cumulative point bar area for the 1980 timestep is 911,300 m. The cumulative point bar area for the 2004 time step is 1,570,400 m, a difference of 659,102 m or a 58% increase in area (Figure 4-10). Mining Area As expected, the greatest number of mine pits and mined area occurred between reaches 9 and 31, which was where mining took place on the Tangipahoa. It is clear from the data that mining has greatly increased between 1980 and 2004. The total area of mine pits has increased by over 2 milli on square meters, from 2,271,700 m i n 1980 to 4,272,900 m in 2004. The largest mine pi t area that was measur ed in 1980 was in reach 11 with 461,100 m. As for 2004, the largest amount of mine pit area measured was in reach 11 with 1,064,000 m (Figure 4-1) The average distan ce from the pit captures centroid from the 1980 main channel was 199 meters. The average distance from the pit captures closest point to the 1980 main channel was only 62 meters (Table

PAGE 43

43 4-5). Through the use of t he Mann-Whitney U-Test, the di stance of the 1980 mine pits and avulsions is statistically significant, while the area of the mine pi ts is not (Table 4-6 and 4-7). This means that the distance to the main channel plays a much greater role in pit capture compared to the size of the mine pits. The number of mine pits also increased gr eatly, from 64 in 1980 to 145 in 2004. The largest number of mined pits for 1980 was located in reaches 9 and 28 with 7 pits. As for 2004, the largest amount of mine pits was located in reach 22 with 15 pits. Despite the fact these reaches have the mo st mine pits, their mine pit area is quite small. In 1980, reach 9 had a mine pit area of 221,600 m. Reach 28 had a mine pit area of 105,000 m, which is extremely low cons idering its amount of pits. Results are similar for 2004, where the mine pit area for reach 22 is only 152,700 m. But the second and third most amount s of mine pits that occurred in 2004 are located in reaches 11 and 26 with 11 and 12 pits respective ly. These reaches also hold the third and second most amount of mine pit area, which is not true for 1980 (Figure 4-2). Bare area is the only variable that has decreased between 1980 and 2004. The total area for this variable in 1980 was 3,925,100 m and 1,880,600 m for 2004. This is a decrease of more the 2 million square meters or 48% of the previous area. For 1980 the greatest amount of bare area was located in reach block 11 with 1,171,900 m. Bare area is very little for 2004, where t he largest area was loca ted in reach 28 with 243,400 m (Figure 4-3). Although there has been a large increase in mine pit area and number of point bars between 1980 and 2004, the total mi ning area between the two time steps was almost identical with 6, 079,300 m in 1980 and 6,153,500 m in 2004.

PAGE 44

44 Statistical Analysis Spearmans Rank Correlation, with an adjustment for ties, was chosen to determine if there was a statistical link bet ween mining related variables and variables representing channel change because the data was not normally distributed. Out of 54 comparisons, 42 were found to be statistically significant at a 95% confidence level (Table 4-2, Table 4-3, and Table 4-4). There was a weak to moderate statisti cal significance between mining related variables and change indices. Negative corre lations occur in the unchanged channel area exclusively. The only two variables that showed no significance were new pit area and new number of pits vs. new channel area. A weak correlation was found between new channel width and mining related variables. This is of no surprise since pit capture can have an extreme effect on channel width. Lateral migration also shows a positive moderate correlation especially in relationshi p to the new mining variables. This is expected since floodplain mining can add sedim ent to the system forcing the channel to shift laterally. Floods can also cause late ral migration when the channel and mine pits become one. The mining related variables that were compared to point bar variables (including positive and negative lags) also showed mixed correlations. These relationships show fewer statistically significant results co mpared to the change index relationships. Positive and negative lags were used in the poi nt bar variables to determine if there was a relationship between mining variables and point bars dow nstream one or two transects in a positive lag or one or two transects upstream in a negative lag. Negative lags were used to determine whether or not headcutting occurred upstream.

PAGE 45

45 New point bar area showed a moderate to high level of significance, with correlation coefficients ranging from 0.583 to 0.767. This indicated that both new and old mining had a great influence on the area of the new point bar s. This can be explained by the increase in sediment that is introduced into a river system from mining. The new number of point bars shared a we ak positive correlation with the new mining variables. New number of point bars is also weakly correlated to the channel length per reach block. This could indicate that the longer the river channel is in a particular reach block, the more point bars will be present. The new point bar area lag 1 and 2 also shared a moderate significance with old and new mining variables, but was again more strongly influenced by the new mining variable s. The new number of point bars lag 1 only shared significance with new mining area. On the contrary, the new number of point bars lag 2 showed a weak positive correla tion with old and new mining variables. The new point bar area lags -1 and -2 we re both moderately correlated to the old and new mining variables, with correlation co efficients ranging from 0.516 to 0.692. The new number of point bars lag -1, similar to t he positive lag 1, was weakly correlated with the new mining variables. The new number of point bars -2 lag was almost the opposite of the positive lag 2, sharing a weak positive correlation with old mining area. Since a majority of the relationships test ed were statistically significant, it can be concluded that mining played a major role in the change of the T angipahoa River in and just below the mined reaches. Based on t hese results, both ol d and new mining variables have an effect on the change of the Tangipahoas channel and point bars, with newer variables having a slightly greater effect.

PAGE 46

46 0 200000 400000 600000 800000 1000000 1200000 15913172125293337414549535761 Distance Downstream ( km ) Mine Pit Area (m^2) 2004 1980 Mined Area Figure 4-1. Mine pit area on the Tangipahoa River 1980-2004 0 2 4 6 8 10 12 14 16 15913172125293337414549535761 Distance Downstream (km)Number of Mine Pits 2004 1980 Mined Area Figure 4-2. Number of mine pits on the Tangipahoa River 1980-2004

PAGE 47

47 0 200000 400000 600000 800000 1000000 1200000 1400000 15913172125293337414549535761 Distance Downstream (km)Bare Area (m^2) 2004 1980 Mined Area Figure 4-3. Bare Area on the Tangipahoa River 1980-2004 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 15913172125293337414549535761 Distance Downstream (km)(I) B/I D/I E/I U/I Mined Area Figure 4-4. Change indices on the Tangipahoa River 1980-2004

PAGE 48

48 0 0.5 1 1.5 2 2.5 15913172125293337414549535761 Distance Downstream (km)Distance (m/Year) 1980-2004 Mined Area Figure 4-5. Tangipahoa River late ral migration rates per year 1980-2004 Avulsions

PAGE 49

49 Figure 4-6. Example of pit capture infl uencing channel width on the Tangipahoa (1980)

PAGE 50

50 0 10 20 30 40 50 60 70 80 90 15913172125293337414549535761 Distance Downstream (km)Width (m) 2004 1980 Mined Area Figure 4-7. Average channel width on the Tangipahoa River 1980-2004 0 20000 40000 60000 80000 100000 120000 140000 160000 180000 15913172125293337414549535761 Distance Downstream (km)Area (m^2) 2004 1980 Mined Area LA-440 LA-10 LA-16 LA-40 LA-442 LA-443 US-190 I-12 LA-22 Figure 4-8. Point bar area on the Tangipaho a River 1980-2004 Avulsions

PAGE 51

51 0 1 2 3 4 5 6 7 8 9 10 15913172125293337414549535761 Distance Downstream (km)# of Pointbars 2004 1980 Mined Area LA-440 LA-10 LA-16 LA-40 LA-442 LA-443 US-190 I-12 LA-22 Figure 4-9. Number of point bars on the Tangipahoa River 1980-2004 0 200000 400000 600000 800000 1000000 1200000 1400000 1600000 1800000 15913172125293337414549535761 Distance Downstream (km)Area (m^2) 2004 1980 Mined Area LA-440 LA-10 LA-16 LA-40 LA-442 LA-443 US-190 I-12 LA-22 Figure 4-10. Cumulative point bar area on the Tangipahoa River 1980-2004

PAGE 52

52 Table 4-1. Totals of va riables compared for 1980 and 2004 1980 2004 Number of Point Bars 91 158 Point Bar Area (m) 911,300 1,582,400 Number of Mine Pits 64 145 Mine Pit Area (m) 2,271,700 4,272,900 Bare Area (m) 3,925,100 1,880,600 Mining Area (m) 6,079,300 6,153,500 Table 4-2. Results of Spearmans Rank Test for Correlation. The bold values represent the statistically significant val ues at the 95% confidence level. Spearman's Rank Values with Adjustment for Ties Independent Variables Dependent Variables Between Area Old Channel Area New Channel Area Overlapping Old and New Channels New Channel Width Larsen Lateral Migration New Pit Area 0.457 0.326 0.241 -0.561 ~ 0.640 New Number of Pits 0.428 0.306 0.221 -0.513 ~ 0.606 New Mining Area 0.497 0.3590.299-0.535 ~ 0.674 Old Mining Area ~ ~~~ 0.281 0.547 Old Number of Pits ~ ~~~ 0.295 0.590 Table 4-3. Results of Spearmans Rank Test for Correlation with relationships involving point bar variables. The bold values represent the statistically significant values at the 95% confidence level. Spearman's Rank Values wit h Adjustment for Ties Independent Variables Dependent Variables New Point Bar Area New Number of Point Bars New Point Bar Area Lag 1 New Number of Point Bars Lag 1 New Point Bar Area Lag 2 New Number of Point Bars Lag 2 Old Mining Area 0.583 0.200 0.597 0.149 0.597 0.271 Old Number of Pits 0.596 0.162 0.654 0.206 0.602 0.332 New Mining Area 0.767 0.328 0.7520.2950.694 0.317 New Number of Pits 0.737 0.296 0.655 0.231 0.697 0.336 New Channel Length/Reach Block ~ 0.283 ~~~ ~

PAGE 53

53 Table 4-4. Results of Spearmans Rank Test for Correlation with relationships involving point bar variables with negative lags The bold values represent the statistically significant values at the 95% confidence level. Spearman's Rank Values with Adjustment for Ties Independent Variables Dependent Variables New Point Bar Area Lag -1 New Number of Point Bars Lag -1 New Point Bar Area Lag -2 New Number of Point Bars Lag -2 Old Mining Area 0.562 0.214 0.516 0.214 Old Number of Pits 0.594 0.195 0.566 0.135 New Mining Area 0.6920.3120.576 0.162 New Number of Pits 0.6860.2730.585 0.208 Table 4-5. Distances from pit captures c entroids and closest point to the 1980 main channel. Table 4-6. Closest distance of the 1980 mi ne pits to the 1980 main channel (p=0.006) Mann-Whitney U-Test Type N Median (m) 25% (m) 75% (m) Captured 6 19 13 123 No Capture 50 196 56 393 Table 4-7. Area of the 1980 mine pits (p=0.884) Mann-Whitney U-Test Type N Median (m) 25% (m) 75% (m) Captured 6 6643 3080 10808 No Capture 50 8107 2074 29493 Avulsion Lat Long Area (m) Distance From Centroid to 1980 Channel (m) Closest Distance from 1980 Channel (m) 1 30.863627 -90.496676 127,100 330 20 2 30.834594 -90.503035 10,800 220 150 3 30.828472 -90.502441 900 140 120 4 30.731020 -90.487567 4,500 80 30 5 30.719284 -90.488439 83,000 430 70 6 30.710063 -90.489833 8,600 70 10

PAGE 54

54 CHAPTER 5 DISCUSSION GIS and high resolution aerial photography were pivotal in the assessment of channel change on the Tangipah oa River. They allow for accurate detection, digitization, and analysis of a number of physical and biological variables that occurred in or near the river. Although maps aren t preferable, they ar e still important when studying river systems. The Tangipahoa has shown a significant amount of change between the two time periods particularly in the mined reaches. Both mining and point bar variables have greatly increased since 1980. The change indi ces indicate relative channel stability upstream and downstream of the mined reaches, but the unchanged ratio in the mined reaches is quite low, while the between ratio is high. The average U/I ratio for the entire Tangipahoa is 0.43, indicating th at less than 50% of the river is in its initial position, indicating a level of instability. This means that the channel is more unstable in the mined reaches. Mixed results occurred with other mined rivers in similar studies. Garfield (2008) found that the Bogue Homo and Bowie River in Mississippi had high U/I ratios while Thompson Creek had an average U /I ratio of about 50%, indicating some instability. Davis (200 9) found that the Amite River in Louisiana had changed significantly. The average U/I ratio was 0.35 over an 18 year period, indicating that 35% of the river remained in its init ial position. In some cases, the B/I ratio was a 2, meaning that the Amite has shifted by over 200% of its init ial position. Lateral migration rates also indicate instability in the mined reaches Rates of more than 2 m/yr occur in the mined reaches or 48 meters for the entire 24 year period. Garfield (2008) and Davis (2009) both found that lateral migration rates were hi gher in the mined reaches

PAGE 55

55 compared to the areas with no mining, but with varying results. Over 22 years, Davis (2009) found that the Amite River migrated about 200 meters in the mined reaches. Garfield (2008) found that rate s on the Leaf River were more in line with rates of the Tangipahoa ranging from 0.5 to 3.5 meters per year in the mined reaches between 1982-1996. Lateral migration rates upstream and downstream from the mined reaches are significantly less. Average width has decreased, especially in the mined reaches in the 2004 timestep. Incision could be occurri ng upstream of the mined area. Decreased flow could have also caused a decrease in channel width. The 2004 main channel captured seven mine pits from the 1980 time peri od (Figure 5-1 to 5-5) All of these pit captures occurred in the mined reaches of the Tangipahoa. Although the 2004 channel has captured these mine pits, it is difficult to tell how much of an effect these captures have had on the 2004 channel width and lateral mi gration rates. Further investigation into the years between 1980 and 2004 could ta ke place to get a better idea of how these pits influenced channel with and lateral migration. Spearmans Rank Correlation showed that multiple relationships between mining and point bar variables were correlated. All of the change indices were found to be statistically significant, exce pt for new channel area vs. new pit area and new number of pits. This could be due to the lack of extr eme floods inducing pit ca pture. Most of the point bar variables were significant as well, except for old mining areas vs. number of point bars and lag 1, -1. Newer mining introduc ed more sediment into the system, thus creating more and larger point bars.

PAGE 56

56 Similar results to this thesis were al so discovered in studies not only done in Louisiana, but also in Italy and California on mined rivers. Rinaldi (2005) found that the Tagliamento and Brenta Rivers experienced incision and channel narrowing. James (2005) also found that mined tr ibutaries in the Yuba Basin, CA. were incised and narrowed compared to pre-mining conditions Studies conducted in Louisiana on the Amite River concluded that weak to m oderate correlations occurred between mining variables and channel changes (Davis, 2009 and McLean, 1995). Very little floodplain mining studies on the Tangipaho a River have occurred, but W ilder (1998) found that the Tangipahoa River aggraded 0.75 to 1 meter at the Robert gaging station between 1982 and 1994. Aggradation of a rive r can result from upstream floodplain mining. These studies have all concluded that mining can have a temporal and spatial effect on a rivers geomorphology.

PAGE 57

57 Figure 5-1. Area of pit captur e in transect 9 (2004 imagery)

PAGE 58

58 Figure 5-2. Area of pit capture in transect 12 (2004 imagery)

PAGE 59

59 Figure 5-3. Area of pit capture in transect 13 (2004 imagery)

PAGE 60

60 Figure 5-4. Area of pit capture in transect 23 (2004 imagery)

PAGE 61

61 Figure 5-5. Area of pit capture in transects 24 and 25 (2004 imagery)

PAGE 62

62 CHAPTER 6 CONCLUSIONS AND FURTHER RESEARCH This study was undertaken to determine if fl oodplain sand and gravel mining and channel change on the Tangipahoa River were st atistically linked between the years of 1980 and 2004. GIS and Excel were invaluable tools for data creation, collection, and processing. Aerial photographs and DRGs were also important elements of this study to establish the main channels, centerlines mine pits and point bars. Reach blocks were created every kilometer to determi ne the degree of channel change both up and downstream. Change indices we re created in each reach block to identify new, old, unchanged, and between channel areas. Thes e indices were also used to help determine channel change both s patially and temporally. Spearmans rank correlation coeffici ent was utilized to determine statistical significance. It was discovered that most of the relationships had a positive moderate to high levels of significance, with very few hav ing no significance at all. This indicated that mining variables played a role in causing channel change, especially in the mined reaches. There is a stronger statistica l link between the newer mining variables and channel change variables suggesting that t he newer mining variables have a stronger influence on channel change. Relationships wit h weak or no statistical links could be the result of a large gap in ti me between the two studied periods. The analysis of the afor ementioned variables provid es evidence that sand and gravel mining in the floodplain does play a role in channel change on the Tangipahoa River. There has been an increase in latera l migration rates, s and bar area, and the number of sand bars, which co incides with the increase of mining area from 1980 to 2004. Although some stability exists on the Tangipahoa, the mined reaches and area

PAGE 63

63 near have become unstable due to the mining practices that take place in the Tangipahoas floodplain. Limitations The first and most glaring limitation to th is project was the use of two different data sources: DRGs and DOQQs. The scale is different between these two sources of data, which could lead to the incorrect ar eas of point bars and mine pits. Multiple interpretations are involved when looking at these two data sources as well. One persons idea of the extent of a mining pit, point bar or any other variable m ay differ slightly from another. The lack of consistent stages could also cause error in the interpretation of the digitized variables. A high river stage could hide or decrease the area of point bars due to a wider channel. If the stage is low, more point bars and a larger area would be digitized, including a smaller channel width. To get the most accurate results, it is important to examine a river when the stages are as similar as possible. The elapsed time between 1980 and 2004 coul d also cause error. There is no question that many things happened in terms of mining and flooding on the Tangipahoa that had an effect on its channel change. To remedy this, a more in depth study could take place to look at the intervening years to get a more comple te idea of the changes that have occurred. Other factors that were not taken into a ccount could also play an imperative role on channel change. Deforestation and urbanizati on can play a key role in geomorphic change. Vegetation cover controls bank er osion during weather events. If this vegetative cover is removed, it will accelerate soil erosion and introduce an increase in sediment into the river system (Knighton, 1998). Urbanization can introduce more water

PAGE 64

64 into the system from efficient drainage syst ems. Urbanization can increase sediment load from construction sites, or decrease t he amount of sediment into the system with the stability in the urban landscape. Flood events have also been shown to have an effect on geomorphic changes. The increa sed stream power can cause channels to widen by increasing erosion. Pit captures induced by flooding can cause the channel to migrate laterally (Knighton, 1998). These ar e all factors that could have some effects on channel change and should be accounted for. Further Research There are several factors that can be studied in the future to understand more clearly what effects mining has had on t he Tangipahoa. New time steps could be studied to fill in the gaps between 1980 and 2004. Time steps could be digitized at more frequent intervals. This would give a more complete story about how mining can influence channel change. From this pit c aptures could be more closely examined and if they have caused any channel changes such as lateral migration and channel widening to the Tangipahoa. Since stream power plays such a large role in channel change, it would be interesting to study the Tangipahoa befor e and after a large flood event. Floods can induce pit captures resulting in the reducti on of point bar areas and numbers and large scale channel change. This type of resear ch could be important for future development and the protection of wildlife habitats si nce a better understanding of the Tangipahoa will react during a flood event will be known. Other variables could be adjusted in future projects such as the distance of the transects, increase lags both up and downstream to see if there is a statistical link

PAGE 65

65 between mining area and channel change in diffe rent reaches, and explore more time periods. Factors that were not included in this project that could induce channel change should be studied as well. These include urbanization, deforestation, dams and bridges, all have geomorphic effects on river systems.

PAGE 66

66 APPENDIX FIGURES SHOWING CHANNEL CHANGE AND PIT CAPTURE Appendix A-1. Unstable reach in transects 6-12

PAGE 67

67 Appendix A-2. Unstable reach in transects 23-29

PAGE 68

68 Appendix A-3. Stable r each in transects 46-52

PAGE 69

69 Appendix A-4. Stable r each in transects 50-56

PAGE 70

70 Appendix A-5. Stable r each in transects 57-61

PAGE 71

71 LIST OF REFERENCES Cluer, B., Dexter, L., 1994. An evaluation of the effects of the interim flows from Glen Canyon Dam on the daily change of beach area in Grand Canyon, AZ. Glen Canyon Environmental Studies, Report PH Y 0 109. Northern Arizona Universit y, Flagstaff, AZ Earickson, R.J., Harlin, J.M., 1994. Geographic Measurement and Quantitative Analysis. Macmillan College Publishing Company, New York, 350pp. Environmental Protection Agency. Targeted Watersheds Grants: Tangipahoa River. 2005. Web. 20 Oct. 2009. < http://www.epa.gov/watershed/initia tive/2005/pdf/tangipahoa_river.pdf >. Hartfield, P.D., 1993. Headcut s and their effect on freshwater mussels. In K.S. Cummings, A.C., Buchanan & L.M. Koch (eds), Conservation and Management of Freshwater Mussels, Proceedings of the Upper Mississippi River Conservation Committee Symposium, October 1992, St. Louis, Missouri, 131-141. Davis, A. H., 2009. Floodplain mining and channel planform change along the Amite River, Louisiana: 1976-1998. Masters T hesis, University of Florida. Garfield, U. A., 2008. Channel planform anal ysis of the Leaf River and tributaries in Mississippi: A decade after an in-stream mining moratorium. Masters Thesis, University of Florida. Jacobson, R.B. and K. Bobbitt Gran, 1999, Gr avel sediment routing from widespread, lowintensity landscape disturbance, Current River Basin, Missouri, Earth Surface Processes and Landforms, v. 24(10), pp 897-917. James, L.A., 2005. Sediment from hydrau lic mining detained by Englebright and small dams in the Yuba basin. Geomorphology 71, 202-226. James, L.A., 2006. The human role is changi ng fluvial systems: re trospect, inventory, and prospect. Geomor phology 79, 152-171. Jones, C., Rich, J., Novak, I.D., 2006. M apping sand and gravel mining within the Presumpscot River Watershed, Maine, us ing remote sensing and other digital data in a G.I.S. environment. University of Southern Maine, Gorham, Maine. Knighton, D., 1998. Fluvial forms and processe s. Oxford University Press Inc., New York. 383pp. Kondolf, G.M., 1993. The reclamation concep t in regulation of gravel mining in California. Journal of Environmenta l Planning and Management 36 (3), 395-406.

PAGE 72

72 Kondolf, G.M., 1994. Geomorphic and environmental effects of in stream gravel mining. Landscape and Urban Planning 28, 225-243. Kondolf, G.M., 1997. Hungry water: effect s of dams and gravel mining on river channels. Environmental M anagement 21 (4), 533-551. Larsen, E.W., Fremier, A.K., Girvetz, E.H., 2006. Modeling t he effects of variable annual flow on river channel meander migration pat terns, Sacramento River, California, USA. Journal of the American Water Resources Association 42 (4), 1063-1075. Louisiana Department of Revenue. Severance Tax Collectio ns and Distributions Fiscal Year Ending July 06-June 2007. 2007. Web. 12 February 2010. < http://www.rev.state.la.us/forms/s tatisticalreports/FiscalYear0607Distributions.pdf >. Louisiana State University. Atlas: The Louisia na Statewide GIS. 2009. Web. 2 June 2009. < http://atlas.lsu.edu >. McLean, M.B., 1995. A geographic informa tion systems analysis of floodplain land cover and channel position changes along t he Amite River, Louisiana. Masters thesis, University of Florida. Mitchell-Tapping, A.M., 1998. A GIS and statistical analysis of channel change on a mined river floodplain, Bogue Chitto River, Loui siana. Masters thesis, University of Florida. Mossa, J., 1983. Morphologic changes in a segment of the Amite River, Louisiana. Abstracts with Programs, 96th Annual Meeting of the Geological Society of America 15 (6), 648. Mossa, J., 1995. Sand and gravel mining in t he Amite River floodplain. Guidebook of Geological Excursions, Geological So ciety of America. 235-360; 1995 New Orleans Meeting Mossa J., Autin, W.J., 1998. Geologic an d geographic aspects of sand and gravel production in Louisiana. Aggregate Resour ces: A Global Perspective, 439-460. Mossa, J., Coley, D., 2004. Planform c hanges of Pascagoula River Tributaries, Mississippi: Year 2 Interim Report. Subm itted to the U.S. Army Corps of Engineers, Pat Harrison Waterway Distr ict, Mississippi Nature Conservancy and the U.S. Geological Survey, 280 pp. Mossa, J., McLean, M., 1997. Channel planform and land cover changes on a mined river floodplain: Amite River, Louisiana, US A. Applied Geography 17 (1), 43-54.

PAGE 73

73 Murphy, J.L., Hawkins, C.P., Anderson, N.H., 1981. Effects of canopy modifications and accumulated sediment on stream comm unities. Transitions of the American Fisheries Society 110, 469-478. Myers, J.L., Well, A.D., 2003. Research Design and Statistical Analysis. Taylor & Francis, Inc., New York. 760 pp. Natural Resources Conservation Services. Louisiana Watershed Receives National Watershed Award. 2005. Web. 15 August 2009. < http://wmc.ar.nrcs.usda.gov/news/lawatershed.html >. Rinaldi, M., Wyzga, B., Surian, N., 2005. Sediment mining in alluvial channels: physical effects and management perspectives. River Research and Applications 21, 805828. Roell, M.J., 2003. Abundance, distribution, and c haracteristics of gravel mining sites in streams of the Salem Plateau, Missour i. http://dnr.missouri.gov/env/lrp. Rogerson, P.A., 2001. Statistical Methods for Geography. Sage Publications, London. 236 pp. Rogue Valley Council of Govern ments. Sand and Gravel Mini ng Restrictions. 1997. Web. 09 June 2009. < http://www.rvcog.org/ftp/Applegate%20Aggregate%20Project /General%20Grav el%20Mining%20Background%20Informa tion/fisheries_gravel_mining_ issues.pdf >. Sandecki, M., 1989. Aggregate mining in river systems. California Geology, 42 (4), 8894. Sanders, R., 1978. Climates of th e states: Louisiana: climates of the states with current tables of normals 1941-1970 and means and extremes to 1975. v.1, Gale Research Company, Detroit, MI., 409-424. Topozone. Louisiana Topology. 2009. Web. 24 October 2009. < http://www.topozone.com/states/Louisi ana.asp?county=Tangipahoa&feature= Civil >. United States Geological Survey. Daily Data for Kentwood, Hammond, and Robert, Louisiana. 2009. Web. 14 October 2009. < http://waterdata.usgs.gov/nwis/dv/?site_no=07375300&am p;referred_module>. Wilder, M. B., 1998. Aggradat ion and the degradation of rive rs in Louisiana. Masters Thesis, University of Florida. 83 p.

PAGE 74

74 BIOGRAPHICAL SKETCH Steven R. Marks was born in 1983 in Ga inesville, Florida. He has lived in Gainesville his entire life and received his Bachelors Degree in Geography from the University of Florida with a strong focus on GIS and Urban Planni ng. He started working at Jones Edmunds and Associates in May 2008 as a GIS technician. He also currently runs a small car detailing business that he started in 2002.