MEASURING AND MODELING THE EFFECTS OF SEA LEVEL RISE

ON NEAR-COASTAL RIVERINE REGIONS: A GEOSPATIAL

COMPARISON OF THE SHATT AL-ARAB RIVER IN SOUTHERN IRAQ

WITH THE MISSISSIPPI RIVER DELTA IN SOUTHERN LOUISIANA,

USA.

By

Ameen Awad Kadhim

A DISSERTATION

Submitted to

Michigan State University

in partial fulfillment of the requirements

for the degree of

Geography - Doctor of Philosophy

2018

ABSTRACT

MEASURING AND MODELING THE EFFECTS OF SEA LEVEL RISE ON

NEAR-COASTAL RIVERINE REGIONS: A GEOSPATIAL COMPARISON OF THE
SHATT AL-ARAB RIVER IN SOUTHERN IRAQ WITH THE MISSISSIPPI RIVER

DELTA IN SOUTHERN LOUISIANA, USA.

By

Ameen Awad Kadhim

There is a growing debate among scientists on how sea level rise (SLR) will impact

coastal environments, particularly in countries where economic activities are sustained

along these coasts. An important factor in this debate is how best to characterize coastal

environmental impacts over time. This study investigates the measurement and model-

ing of SLR and effects on near-coastal riverine regions. The study uses a variety of data

sources, including satellite imagery from 1975 to 2017, digital elevation data and previous

studies. This research is focusing on two of these important regions: southern Iraq along

the Shatt Al-Arab River (SAR) and the southern United States in Louisiana along the Mis-

sissippi River Delta (MRD). These sites are important for both their extensive low-lying

land and for their significant coastal economic activities. The dissertation consists of six

chapters. Chapter one introduces the topic. Chapter two compares and contrasts both

regions and evaluates escalating SLR risk.

Chapter three develops a coupled human and natural system (CHANS) perspective

for SARR to reveal multiple sources of environmental degradation in this region. I use the

CHANS perspective to identify the problems, and which ones (human or natural systems)

are especially responsible for environmental degradation in SARR. I use several measures

of ecological, economic, and social systems to outline the problems identified through

the CHANS framework. I modeled and analyzed the regions land cover between 1975

and 2017 to understand how the environment has been affected, and found that climate

change is responsible for what happened in this region based on other factors.

Chapter four constructs and applies an error propagation model to elevation data in

the Mississippi River Delta region (MRDR). This modeling both reduces and accounts for

the effects of digital elevation model (DEM) error on a bathtub inundation model used to

predict the SLR risk in the region. Digital elevation data is essential to estimate coastal

vulnerability to flooding due to sea level rise. Shuttle Radar Topography Mission (SRTM)

1 Arc-Second Global is considered the best free global digital elevation data available. In

this study, I conduct an error assessment and develop statistical error modeling for SRTM

to improve the quality of elevation data in these at-risk regions.

Chapter five applies MRDR-based model from chapter four to enhance the SRTM 1

Arc-Second Global DEM data in SARR. As such, it is the first study to account for data

uncertainty in the evaluation of SLR risk in this sensitive region. This study transfers an

error propagation model from MRDR to the Shatt al-Arab river region to understand the

impact of DEM error on an inundation model in this sensitive region. The error propaga-

tion model involves three stages. First, a multiple regression model, parameterized from

MRDR, is used to generate an expected DEM error surface for SARR. Second, residuals

from this model are simulated for SARR: these are mean-zero and spatially autocorrelated

with a Gaussian covariance model matching that observed in MRDR by convolution fil-

tering of random noise. More than 50 realizations of error were simulated to make sure

a stable result was realized. These realizations were subtracted from the adjusted SRTM

to produce DEM realizations capturing potential variation. Third, the DEM realizations

are each used in bathtub modeling to estimate flooding area in the region with 1 m of sea

level rise.

I dedicate this work to my mom, who passed away in June 2013. It was early days of my

Ph.D. program. My mom was everything in my life. She supported me to pursue the

To My Mom...

Ph.D. degree and supported me at every moment during the program, and my life. I feel
that she is happy and proud of me as the dream, that she shares with me, becomes true.

iv

ACKNOWLEDGMENTS

This dissertation would not have been possible without the support of my advisor Dr.

Ashton Shortridge and the financial support of the Higher Committee For Education De-

velopment in Iraq (HCED-Iraq ), the Department of Geography, Environment, and Spatial

Sciences and college of science at Michigan State University. I am especially indebted to

my advisor Dr. Ashton Shortridge who is supporting my PhD program goals and how

he is working actively to provide me with the academy knowledge with the time to pur-

sue those goals. I am very grateful to my family and my committee (Dr. Lifeng Luo, Dr.

Jiaguo Qi, and Dr. Hayder Radha, Dr. Warren W Wood) and the large group of people

to whom I owe a lot of gratitude especially my friends (Wahab Al-Barakat, Khaldoun Al-

Yasiri, Falah AL-Lahaybey and Brad Peter). I would like to thank the Center for Global

Change and Earth Observations (CGCEO) at MSU for their hospitality. Thanks for my

big family, the geography department at MSU.

v

TABLE OF CONTENTS

LIST OF TABLES .

.

LIST OF FIGURES .

.

.

.

.

.

.

.

.

.

.

KEY TO ABBREVIATIONS .

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

x

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. xiii

Chapter 1

.

.

.
.

1.1 Background . .

Introduction .
.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.

1
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
1.1.1
Shatt Al-Arab River . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
1.1.2 Mississippi River Delta . . . . . . . . . . . . . . . . . . . . . . . . . . .
8
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

1.2 Study Goals
.
1.3 Study Objectives .

.
.

.
.

.
.

.

.

Chapter 2

.

.

.

.

.

.

2.2.1.1
2.2.1.2

Introduction .

A Comparison Between Shatt Al-Arab River Region in Southern
Iraq with Mississippi River Region in Southern Louisiana, USA. . . 13
2.1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 The Geographical Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.1 Topography Characteristics . . . . . . . . . . . . . . . . . . . . . . . . 18
19

Topographic Characteristics of Shatt Al-Arab River Region
Topographic Characteristics of Mississippi River Delta Re-
gion .

2.2.2 Geology Characteristics

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
. . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2.2.1 Mississippi River Delta Geology and Ecology . . . . . . . . 22
2.2.2.2
Shatt Al-Arab Geology and Ecology . . . . . . . . . . . . . . 23
2.2.3 Hydrology Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.3.1 Mississippi River Flow . . . . . . . . . . . . . . . . . . . . . 24
2.2.3.2
Shatt Al-Arab River Flow . . . . . . . . . . . . . . . . . . . . 25
2.2.4 Climate Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.5 Human Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

2.3 Conclusion .

.

.

.

.

.

.

.

Chapter 3

.
.

Introduction .

.
3.1
.
3.2 Data and Methods .
Study Area .

Causes and Consequences of Environmental Degradation Along
Shatt Al–Arab River: A Coupled Human and Natural System Per-
spective .
.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2.1
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2.2 Coupled Human and Natural System . . . . . . . . . . . . . . . . . . 34
. . . . . . . . . . . . . . . . . . . . . . . . . 35
Environmental Impacts . . . . . . . . . . . . . . . . . . . . . 37
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.2.2.1 Human Impacts
3.2.2.2

3.2.3 Methodology .

.
.
.
.
.
.
. .

.

vi

3.3 Results .
3.4 Discussion . .

. .

.
.

.
.

.
.

. .

.
. .

3.4.1 Human Impacts .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
. 42
3.4.1.1 War and Government Policy . . . . . . . . . . . . . . . . . . 42
3.4.1.2 Upstream Water Demand . . . . . . . . . . . . . . . . . . . . 44
3.4.1.3 Hydroengineering in SARR . . . . . . . . . . . . . . . . . . 45
3.4.1.4
Pollution and Degradation . . . . . . . . . . . . . . . . . . . 46
3.4.2 Environmental Impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
. 47
3.4.2.1 Meteorological Factors Associated With Drought
3.4.2.2
. 47
3.4.2.3 Groundwater Salinity . . . . . . . . . . . . . . . . . . . . . . 48
. 49

. . . . .
Sea Level Rise . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.4.3 Conclusions .

.

.

Chapter 4

.

Introduction .

4.1
.
4.2 Data and Methods .
Study Area .
. .

4.2.1
4.2.2 Data .

. .

.

.
.
.
.
. .
. .

.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.

Flooded With Error: Handling Uncertainty in SRTM For The As-
sessment of Sea Level Rise in Mississippi River Delta . . . . . . . . 51
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
. 54
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.2.2.1 National Elevation Dataset (NED) . . . . . . . . . . . . . . . 55
4.2.2.2
SRTM 1 Arc-Second Global Coverage ( 30 meters) . . . . . . 56
4.2.2.3 Vegetation Data . . . . . . . . . . . . . . . . . . . . . . . . . . 56
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.2.3.1 Regression Kriging . . . . . . . . . . . . . . . . . . . . . . . . 58
4.2.3.2
Inundation Modeling . . . . . . . . . . . . . . . . . . . . . . 59
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.2.4.1
Parameter Estimation For SRTM 1 Error Regression Model . 60
4.2.4.2 Regression and Variogram Results . . . . . . . . . . . . . . . 62
4.2.4.3
Inundation Modeling . . . . . . . . . . . . . . . . . . . . . . 63
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

. .

.
.

.
.

.

.

4.2.3 Methodology .

4.2.4 Results . .

4.2.5 Discussion .
4.2.6 Conclusion .

Chapter 5

.

.

.
.
.
.
. .

Introduction .

.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.

.
5.1
5.2 Data and Methods .
Study Area .
SRTM 1 Arc-Second Global Coverage ( 30 meters)

Accounting for DEM Error in Sea Level Rise Assessment for Shatt
Al-Arab River Region . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
. 72
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5.2.1
5.2.2
. . . . . . . . . . . 74
5.2.3 Vegetation Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2.4 Methods .
5.2.4.1 Monte Carlo Error Propagation . . . . . . . . . . . . . . . . . 75
5.2.4.2 Unconditional Simulation with Convolution Filtering . . . . 76
Inundation Modeling . . . . . . . . . . . . . . . . . . . . . . 77
5.2.4.3
.
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.3.1 Regression Kriging-Based Error Modeling . . . . . . . . . . . . . . . . 79

5.3 Results .

. .

. .

.

.

.

.

.

vii

5.3.2 Realizations .
5.3.3

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. 80
Inundation Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

5.4 Limitations . .
5.5 Conclusion .
.

. .
.
.

.
.

.
.

.
.

.

.

Chapter 6

.

.
.

.
.

. .

. .

6.1 Methods

Conclusion .
.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
6.1.1 Coupled human and natural systems (CHANS)
. . . . . . . . . . . . 89
6.1.2 The inundation model (bathtub modeling) and error propagation . . 90
6.2 Dissertation Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.3 Future Research .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

.

.

.

viii

LIST OF TABLES

Table 2.1: Comparison of the study regions.

. . . . . . . . . . . . . . . . . . . . . . . 18

Table 3.1: Satellite sensors utilized for the study used three sensors to cover the

time (1975, 1985, 1995, 2005 and 2017).

. . . . . . . . . . . . . . . . . . . . 38

Table 4.1: Regression coefficients of SRTM error model with NED and VCF. . . . . . 63

Table 5.1: Regression coefficients for regression model variables

. . . . . . . . . . . 81

ix

LIST OF FIGURES

Figure 1.1: Modeling the effects of sea level rise on near-coastal riverine regions based on

SRTM 30 m data preprocessing for the research work.

. . . . . . . . . . . . .

6

Figure 1.2: Linking dissertation objectives.

. . . . . . . . . . . . . . . . . . . . . . . . . 12

Figure 2.1: The natural system which causes the sea level to change, adopted from Brandi-

marte et al. (2015).

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Figure 2.2: The change of the mean temperature and the carbon dioxide in the atmo-
sphere for the global in the past and future, adopted from (Clark et al. (2016)).

16

Figure 2.3: Total sea level change since 1993 based on TOPEX/Poseidon.

. . . . . . . . . 17

Figure 2.4: Both SARR and MRDR have high rising sea levels, adopted from (Williams

(2009)).

.

.

.

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Figure 2.5: Shatt Al-Arab river region in southeastern Iraq. . . . . . . . . . . . . . . . . . 20

Figure 2.6: Mississippi River Delta region in Louisiana, USA. . . . . . . . . . . . . . . . . 21

Figure 2.7: World climate regions according to Koppen-Geiger classification, adopted from

(Peel et al. (2007)).

.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Figure 2.8: Projects by Louisiana’s Coastal Protection and Restoration Authority since

2007, adopted from (Louisiana (2012)).

. . . . . . . . . . . . . . . . . . . . . 29

Figure 3.1: Total study area of Shatt Al-Arab river region.

. . . . . . . . . . . . . . . . . 35

Figure 3.2: Total conceptualization of coupled human and natural systems (CHANS) in

Shatt Al-Arab river region. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Figure 3.3: Temporal variations of NDVI (A-1975, B-1985, C-1995, D-2005 and E-2017

NDVI greenness density plot) for the Shatt Al-Arab river region. . . . . . . . . 39

Figure 3.4: Regressing trends of temperature, precipitation, humidity, and sea levels for

the periods (1975, 1985, 1995, 2005, and 2017) associated with the NDVI trends. 41

Figure 3.5: Comparative between 1975 and 2017 NDVI in Shatt Al Arab Region (a,b)which
shows the difference in the density of vegetation distribution between up and
down the junction of Karun river in Iraq and Iran with Shatt Al Arab river.

. . 42

x

Figure 3.6:

Increase in salinity (ppm) in Shatt Al-Arab river from 1980 to 2012.

. . . . . . 49

Figure 4.1:

(a) study area (Mississippi River Delta region in Louisiana state); (b) NED
1/3 arc-second (approximately 10 meters); (c) VCF Landsat 5 TM Collection 1
products 30 m and (d) SRTM 1 Arc-Second Global coverage ( 30 meters).

. . . 55

Figure 4.2: Error modeling and bathtub modeling for Mississippi River Delta Region.

. . 58

Figure 4.3: SRTM error: (a) Regression original SRTM error and raw SRTM error versus:

(b) VCF; (c) SRTM elevation; (d) SRTM slope.

. . . . . . . . . . . . . . . . . . 62

Figure 4.4: Error assessment for variogram model: (a) Regression model performance;
(b) histogram of validation error of the regression model; (c) variogram and
model of regression model residuals over short distances; (d) variogram and
model of regression model residuals over long distances; (e) map of model-
predicted SRTM elevation errors and (f) histograms of predicted SRTM errors). 64

Figure 4.5:

Inundation models (predicting the spatial extent of a 1m sea level rise in the
MRD region: (a) initial SRTM and (b) adjusted SRTM DEM after the error
modeling. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

.

.

.

.

.

Figure 4.6:

Inundation model predicting the spatial extent of a 1m sea level rise in the
MRDR: (a) NED 10m DEM and (b) adjusted SRTM DEM.

. . . . . . . . . . . 66

Figure 4.7: Marsh classes in Mississippi river delta region: a. Spatial temporal change for

vegetation classes. b. Marsh classes changing based on mean NED DEM.

. . . 67

Figure 5.1:

(a) Landsat imagery for Shatt Al-Arab River study region in Iraq; (b) SRTM 1
Arc-Second Global coverage ( 30 meters); (c) SRTM-derived slope; (d) Vegeta-
tion cover fraction (VCF) derived from Landsat 5 TM Collection 1. . . . . . . . 73

Figure 5.2: Transferring the error propagation model and bathtub modeling from Missis-

sippi River Delta Region (A), to Shatt Al-Arab river region (B). . . . . . . . . . 77

Figure 5.3: The autocorrelated Gaussian covariance process and its associated convolu-
tion filter. This filter is passed over spatially random Gaussian noise to pro-
duce rasters with the desired spatial structure.

. . . . . . . . . . . . . . . . . 78

Figure 5.4: Error assessment for variogram model: (a) MRDR error model residual vari-
ogram and model (line); (b) SARR variogram model (line) and variogram of
simulation output (circles); (c) MRDR map of model-predicted SRTM eleva-
tion errors; (d) MRDR histogram of predicted SRTM errors; (e) SARR map
of model-predicted SRTM elevation errors; (f) SARR histogram of predicted
SRTM errors.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

.

.

.

.

xi

Figure 5.5: SRTM realizations: (a) random Gaussian noise for the SARR and spatially
autocorrelated noise following convolution filtering; (b) QQ plots of random
and convolution-filtered Gaussian noise; (c) DEM realization number 10; (d)
bathtub modeling SRTM realization number 10. . . . . . . . . . . . . . . . . . 82

Figure 5.6:

(a) SARR map of regression-adjusted SRTM elevation; (b) histograms of ad-
justed SRTM ; (c) SARR probability map for land remaining dry following
bathtub modeling of 1m inundation (d) histogram of flooding probability; (e)
SARR bathtub model using raw SRTM data; (f) Proportion of SARR predicted
to remain dry following 1m inundation: raw DEM is dashed line, histogram
are areas across 50 realizations.

. . . . . . . . . . . . . . . . . . . . . . . . . 83

xii

KEY TO ABBREVIATIONS

CHANS Coupled Human And Natural Systems

DEM Digital Elevation Model

GIS Geographic Information System

IFSAR Interferometric Synthetic Aperture Radar

km Kilometer

mm Millimeter

MRDR Mississippi River Delta Region

MSL Mean Sea Level

NDVI Normalized Difference Vegetation Index

NED National Elevation Data

SARR Shatt Al-Arab River Region

SLR Sea Level Rise

SRTM Shuttle Radar Topography Mission

VCF Vegetation Cover Fraction

USGS The United States Geological Survey

xiii

Chapter 1

Introduction

During the past century, global sea level has risen about 0.17 ± 0.05 meter due to the

increase in global average temperature (Hansen (2007); Li et al. (2009); Rahmstorf (2007)).

Historically, the Earth has warmed between 1900 and 1940, cooled slightly after 1940 and

then warmed markedly from 1970 onward. While global warming in the past century was

estimated to be 0.8°C, the rise in temperature in the past three decades alone was 0.6°C

(a rate of 0.2°C per decade) as greenhouse gases became the dominant climate forcing in

recent decades (Nageswara Rao et al. (2008)). Importantly, thermal expansion of ocean

water due to increasing the global warming has been responsible for about half of the sea

level rise in the last century (Meyssignac and Cazenave (2012)). In addition, ice is melting

in the significant ice repositories of the world.

The Greenland ice sheet is melting at a rate of 23,923 km3 per year. The extent of Arctic

sea ice has been decreasing at almost 8% per decade since the middle of the last century

(Nageswara Rao et al. (2008); Rahmstorf (2007)). The effect of SLR on low-lying coastal

areas is tremendous. This dissertation focuses on two vulnerable areas: SARR in southern

Iraq and MRDR in Louisiana, USA. SARR is estimated to have lost more than 90 % of

its cultivable land between 1970 and 2002, in addition to loss of other human economic

activity and environmental degradation (Singh et al. (2005); Brandimarte et al. (2015)).

No comprehensive study has evaluated the role that SLR may have played in these recent

1

changes in the SARR. The MRDR in Louisiana has high known rates of local relative sea-

level due in large part to the sinking of land, including the loss of around 4,920 square

kilometers of coastal wetlands, more than one-third of its coastal plain between 1932 and

2010 (Couvillion et al. (2011)).

This study aims to investigate the influence of SLR within these sensitive and impor-

tant coastal regions in two ways. First, it uses geospatial methods like bathtub modeling

to estimate the position and extent of areas in these regions, which were affected by SLR

from 1975 to 2017 and predict areas, which will be affected in the near future. Further,

the impact of this change on land cover, u and ecosystem will be estimated. Second, since

DEM data for the SARR is not very accurate, an error propagation model will be con-

structed to improve flooding estimates from bathtub modeling. This study will apply

bathtub modeling and analyze different imagery and data sources for different time peri-

ods and will explain impacts of SLR on both regions. This study improves understanding

of SLR impacts on vegetation, ecosystems and groundwater around these rivers using

geospatial analysis. The temporal and spatial impact of SLR on these factors was very

clear through the integration of remotely sensed land change with data on water salinity,

SLR rate and rate of land subsidence in the river regions.

1.1 Background

The threat of SLR has arisen due to climate change and the growth of greenhouse gas

emissions leading to global warming. The global average temperature is increasing at

the rate of 0.2°C per decade, which will accelerate future SLR (Dasgupta et al. (2009); Li

et al. (2009)). Global SLR is due to thermal expansion of ocean water and the breakup

2

of the Greenland and West Antarctic ice sheets; together these might produce 1- 5m SLR

by 2100 (Rahmstorf (2007); Hansen (2007); Mcleod et al. (2010)). Many research studies

have attempted to assess the impact of sea level rise in coastal environments, particularly

in areas with important economic activities along these coasts. An essential factor in this

research is how best to characterize coastal environmental impacts over time, through

both future-looking site-specific numerical modeling and past and present site-specific

observational studies (Werner and Simmons (2009)). Most hydrologic studies focus on the

freshwater zone density in the coastal area, which is difficult to estimate because of the

lack of data about properties of water. Usually, these studies use the Ghyben-Herzberg

relation equation (W. BadonGhijben (1888, 1889) and A. Herzberg (1901), Verruijt (1968)).

It is a relatively uncomplicated equation, and it has been used successfully in various

areas to estimate the intensity of salt water and the density of freshwater in an aquifer

(Barlow (2003)). In the following equation, the density of the freshwater area above sea

level is h and that below sea level is z. Both densities are related by:

z = PF/(Ps − PF) ∗ h

(1.1)

Where PF represents the density of freshwater (1.000 grams per cubic centimeter (g/cm3)

at 20°C), and Ps represents the density of saltwater (1.025 grams per cubic centimeter
(g/cm3)) (Verruijt (1968)). The equation may be simplified to z = 40 × h (Antonellini et al.

(2008); Verruijt (1968)).

Most related studies focus only on the characteristics of particular study regions, so it

is hard to draw universally applicable conclusions about the impact of SLR (Werner and

Simmons (2009)). For example, Sherif and Singh (1999) used numerical simulations to

3

model SLR-driven saltwater intrusion in two different coastal aquifers. The first was in

the Nile Delta, where SLR was predicted to cause intrusion of 9.0 km. The second was

in the Madras aquifer in India, where SLR in the Bay of Bengal was predicted to cause

intrusion of 0.4 km in this delta (Sherif and Singh (1999)). These regions exhibited widely

different conclusions in physical or more hypothesized terms. These variations do not

describe a systematic comparison between these places (Werner and Simmons (2009)).

Also, a study by Bobba (2002) focused on sea level rise intrusion in Godavari Delta

in India and developed numerical simulations for that delta. In this study, simulations

were applied to describe an obvious risk of mixing fresh water with salt water because

of sea-level rise, but it is very complicated to describe the influence of vertical intrusion

of seawater into the aquifer (Bobba (2002)). The complexity of the numerical model did

not allow an easy investigation of the latter influence, and the study does not express any

quantitative results about the amount of seawater intrusion produced by sea-level rise

(Barlow (2003); Bobba (2002)).

More uncertain too is that although many of these studies use sophisticated quanti-

tative modeling techniques, most do not provide quantitative results about the impact

of sea-level rise, probably the result of an original inability to explain the different ways

contributing to both salinity effect measurement and groundwater and surface water hy-

draulic head (Barlow (2003)). SLR risk is high in coastal river regions with canals and

drainage networks enabling saltwater intrusion from the sea to inland areas (Good and

Wilson (1995)). So, good digital elevation models are essential to hydrologic processes

and modeling the SLR (Poulter and Halpin (2008)). Digital elevation data with different

vertical and horizontal resolutions can produce different estimates of coastal vulnerabil-

ity to flooding for the same regions (Poulter and Halpin (2008)). In many coastal regions,

4

hydrologic complexity is increased by the presence of canals and drainage networks, fea-

tures small enough to be challenging to identify. Because of that, models require DEM

data with low error and high spatial resolution (Lindsay (2006)).

Finally, inundation under various SLR scenarios is commonly characterized using the

bathtub model. Given the lack of hydrologic and subsurface data for more complex hy-

drological modeling in the SARR, the bathtub model is the best method to estimate im-

pacts of SLR in the proposed research. SLR is treated like rising water in a bathtub, where

regions with an elevation less than a targeted ood level will be predicted to be inundated

(Yunus et al. (2016)). The main steps and a schematic diagram of the processing scheme

followed to provide the foundation for the research from obtaining modeling the effects

of SLR on near-coastal riverine regions research grade dataset through the preprocessing

steps that have been followed before. The main research components of this dissertation

are summarized in Figure 1.1. The conceptual model shows stages of the study process to

present the foundation for the research to model the effects of SLR on near-coastal river-

ine regions. In chapter four and five, I use SRTM DEMs and error propagation modeling

to enhance the SRTM DEMs. SRTM 30 m is the best global DEM data for bathtub model-

ing. Bathtub modeling is a valuable GIS-based approach for detecting and characterizing

impacts from SLR on both regions. The use of bathtub modeling to identify human and

environmental SLR impacts in these regions is important and to predict SLR over the next

several decades.

1.1.1 Shatt Al-Arab River

Shatt Al-Arab River (SAR) has received little attention, despite its potential susceptibility

to SLR impacts, which are not well known. Regional SLR studies for Persian Gulf did

5

Figure 1.1: Modeling the effects of sea level rise on near-coastal riverine regions based on SRTM
30 m data preprocessing for the research work.

not discuss Iraq’s coast, arguing that this coast is a narrow area (El Raey (2010)). This

study did not take into consideration Shatt Al-Arab river, which flows 193 km from the

junction of Tigris and Euphrates rivers to the Gulf. The river’s width is 790 m at the coast

(Al-Saaidy (2008)). A number of regional studies have used tidal gauge data in Persian

Gulf to estimate relative sea level change for locations near the outlet of Shatt Al-Arab

river.

Alothman (2014) constructed a time-series record from seven coastal tide gauges in
this area. A relative sea level rise of 2.2 ± 0.5 mm/year was observed from this record

over the period 1979-2007. This result contrasts with the findings of Hassanzadeh (2007),

who reported a rate of 2.8 mm/year (1990-1999) for the northern part of Persian Gulf.

This study used data from tide gauges at Bandar Abbas, and Bushehr combined with

6

atmospheric data at both stations to investigate the mean sea level (MSL) response to me-

teorological forcing functions in this region. An earlier study by Sultan (1995) estimated
a constant rate of 2.1 ± 0.1 mm/year (1980–1990) using two-tide gauges. In an update,

Sultan (2000) identified a sea level rise of 1.7 mm/year (1980-1994) using nine tide gauges.
Hossein (2007) reported a mean rate of 2.34 ± 0.07 mm/year (1990–1999).

The variability of these estimates of mean sea level change may be due to incomplete

and inaccurate records. In addition, it can be related to the significant seasonal variation

of sea level in Persian Gulf due to air pressure: water level in the Gulf is up to 26 cm

higher in summer than in the winter (Sultan et al. (1995b); Hassanzadeh et al. (2007)).

Nevertheless, all studies agreed that regional sea level rise is 2–2.5 mm per year from the

1980s to the 2000s.

The influence of SLR may be impactful in this region, for large areas of low-lying land

have important human activities like agriculture. Lining the river is the largest date palm

forest in the world. Stretching back from the riverbanks towards the desert, date planta-

tions extend for distances varying from a few hundred meters to almost six kilometers. In

the mid-1970s, the region counted some 17-18 million date palms or the fifth rank of the

world’s 90 million palm trees. By 2002, more than 14 million, or 80 percent, of the palms

were wiped out due to salinization in Shatt Al-Arab region, which began emerging in the

late 1960s (Singh et al. (2005); Brandimarte et al. (2015)). Of the 52,000 hectares of date

farms fringing Shatt Al-Arab in 1975, only 11,000 hectares, or 21% remained in 2002. In

total, war, salt and pests destroyed approximately 14 million palms: around 9 million in

Iraq and 5 million in Iran. Moreover, of the remaining 3-4 million palms, many are in poor

condition (Singh et al. (2005); Brandimarte et al. (2015)). Many studies have focused on

environmental problems (Jabbar and Zhou (2011)), land change (Al-Bahili et al. (2010))

7

increased concentration of salt in Shatt Al-Arab (Al-Mahmood (2009); Essa (2012)) and

increased concentration of salt in the land around the river (Mahmood et al. (2013)). All

of these studies agree that the tide is the source of increasing salinity and its effects. How-

ever, no one mentioned the role of SLR impact on environmental problems. However,

this region is affected twice daily by the tide, a source of salinity in the region which con-

tribute to environmental problems and reduced vegetation around the river due to both

increased water salinity and increased concentration of salt in soils.

1.1.2 Mississippi River Delta

The MRDR has a highest relative sea-level rise in the U.S. due to the combination of global

sea-level rise and local subsidence (sinking land). Land-surface altitude data collected in

a region indicated mean annual subsidence of 5 mm per year during five survey epochs

between 1951 and 1995 (Burkett et al. (2003)). Subsidence rates in Mississippi river delta

are further increased due to a large amount of fluid withdrawal in the northern Gulf of

Mexico. Oil and gas extraction leads to increased soil compaction over a short amount of

time (Blum and Roberts (2009); Penland and Ramsey (1990)).

Many previous remote sensing studies have estimated MRDR flooding and how much

this delta lost land because of SLR. (Couvillion et al. (2011)) based their estimate on the

survey and aerial data and raised concerns about comparability of the 1932 and 1956

datasets. In contrast, (Blum and Roberts (2009)) based their estimates on aerial and satel-

lite imagery from 1932 to 2010. Both of the studies reveal that MRDR lost more than 25%

of its land area with a net change of about -1,883 square miles (mi2) in land area from 1932

to 2010. About 95 percent of the decrease in land area accounts for persistent losses. Trend

analysis investigations from 1985 to 2010 designate a wetland loss rate of 16.57 mi2 per

8

year (Dasgupta et al. (2009); Barras et al. (2008); Couvillion et al. (2011); Blum and Roberts

(2009)). The mean sea level rise at Grand Isle, Louisiana (NOAA gauge station 8761724)

is 9.03 millimeters/year based on monthly mean sea level data from 1975 to 2017 (Tides

and Currents, noaa.gov).

1.2 Study Goals

The broad goals of this dissertation are:

1. Reveal the historic and potential future impact of SLR in SARR using geospatial

data and methods.

2. Conduct an error assessment and develop statistical error modeling for Shuttle Radar

Topography Mission (SRTM) 1 Arc-Second DEM. SRTM is considered the best free

global digital elevation data available. This model improves the quality of eleva-

tion data in the MRDR. Inundation modeling is used on SRTM to estimate coastal

vulnerability to flooding due to SLR.

3. Use the bathtub model on SRTM Global DEM 30 to measure and assess the impact

of SLR on SARR and MRDR.

4. Appropriate the right error modeling to enhance the SRTM 30 m DEMs. Then apply

bathtub modeling to detect and predict its future flooding over the next several

decades.

9

1.3 Study Objectives

The dissertation has four objectives and these objectives are linked to each other as shown

in Figure 1.1. Each objective of this study has many geospatial methods.

I am using

several software packages to complete each objective: GIS software (ArcGIS, ERDAS),

and the R statistical software. The dissertation objectives are:

Objective one. Compare and contrast MRDR, SARR, and identify whether these regions

have environmental degradation because SLR is based on the theoretical framework and

previous studies. To achieve this objective, the following steps are used:

1. Characterize the geomorphology and climate in both regions and show how that is

correlated with SLR in both regions.

2. Identify human activity and economic importance in both regions and also identify

how the human system accelerates SLR risk in both regions.

3. Assess the different ways that SLR may affect both regions through rising salinity

in groundwater, pressure from subsidence erosion and regional flooding.

Objective two. Use a CHANS perspective to characterize and delineate a clear conceptual

model for factors impacting SARR and identify how human and natural systems are asso-

ciated with SLR risk in this region. For achieving this objective, these steps are followed:

1. Identify primary human and environmental drivers in SARR system.

2. Characterize NDVI as a proxy for vegetation quality and quantity and measure it

over a five decade period.

3. Identify spatial and temporal changes in NDVI throughout the SARR.

10

4. Link observed changes to dominant drivers of change in the SARR.

Objective three. Enhance SRTM Global DEM to better model SLR and characterize land

change in MRDR. For this objective to be accomplished, certain are followed:

1. Determine how SRTM error correlates with canopy cover (derived from Landsat 7).

2. Identify the extent to which SRTM error is reduced using canopy cover and other

globally available covariates.

3. Assess the effectiveness of geostatistical models in reproducing inundation model

results using high-accuracy USGS NED in this region.

Objective four. Develop error propagation models to enhance SRTM 30 for modeling sea

level rise and characterizing the land change in SARR. This is achieved via the following

steps:

1. Apply the error model from objective three, developed for MRDR, to SARR.

2. Determine and map land flooding due to SLR in SARR and identify the land which

will become inundated based on the bathtub model.

3. Apply a valid geostatistical model to account for SRTM error and asses uncertainty

in SLR inundation model.

11

Figure 1.2: Linking dissertation objectives.

12

Chapter 2

A Comparison Between Shatt Al-Arab

River Region in Southern Iraq with

Mississippi River Region in Southern

Louisiana, USA.

2.1

Introduction

SLR is a natural process greatly accelerated by humans in direct and indirect ways. For

example, human activity creates greenhouse gases which cause thermal expansion of sea-

water and melt glaciers, ice caps, and ice sheets, directly leading to sea level increase (Pf-

effer et al. (2008)). Humans also indirectly affect SLR through actions like dam building

which reduces sediment supply to river deltas as in MRDR (Blum and Roberts (2012)).

Dam construction also affects deltas due to reduced fresh water which encourages salt-

water intrusion as in SARR (Al-Mahmood (2009); Essa (2012)). Human actives in the

coastal area like oil extraction, irrigation, and groundwater mining can also be related

to SLR and saltwater intrusion in these regions (Blum and Roberts (2012); Al-Mahmood

(2009); Essa (2012)). According to (Nicholls and Cazenave (2010)) and Figure 2.1, there

13

are a range of factors contributing to global SLR:

1. Warming water in oceans which causes thermal expansion for the water and rising

sea level because of global warming.

2. Increased water input to oceans because of melting glaciers in the tops of mountains

due to global warming.

3. Melting ice in Greenland and Antarctica at a faster rate than ice accretion because of

global warming.

4. The Earths tectonic activity which causes land subsidence in some coastal areas (rel-

ative SLR).

The Fourth Assessment Report (AR4) of the Intergovernmental Panel on Climate Change

(IPCC (2013)) projected that global SLR will exceed 1m by 2100 in response to warming

oceans and melting glaciers (Rahmstorf (2007)). That means the SLR is expected to in-

crease over time. The sea level will be rising because of ongoing temperature increase.

Each decade temperatures rise 0.5 Celsius degree (Cabanes et al. (2001)). This increase in

temperature will continue with the continuation of industrial pollution and deforestation

by people and misuse of water sources.

According to (Clark et al. (2016)), the temperature has changed since the glacial maxi-

mum (21,000 years ago) and the mid-Holocene (6,000 years ago). Looking forward, global

temperature estimates for the late 21st century are projected to be increasing more rapidly

than current trends according to the IPCC scenario AR5V, as shown in Figure 2.2(c and e)

(Meehl et al. (2005); Clark et al. (2016)). Anthropogenic increases in atmospheric carbon

dioxide (Figure 2.2 (a and b) is a principal driver of these changes (Clark et al. (2016)).

14

Figure 2.1: The natural system which causes the sea level to change, adopted from Brandimarte
et al. (2015).

SLR is not uniform across the earth (Rahmstorf (2007); Vermeer and Rahmstorf (2009)),

according to Figure 2.3:

1. The areas near the equator have higher sea level than other regions of the world due

in part to greater thermal expansion in warm regions.

2. The areas which are geologically unstable along the boundaries between Earth’s

tectonic plates or due to isostatic rebound have higher SLR than other more stable

places in the world. For example, MRDR has experienced 9 mm SLR during the last

decade because the land surface is dropping.

3. The areas experiencing human activity like oil extraction and groundwater extrac-

tion can have higher relative SLR. Also, elevations of delta areas may be influenced

by upstream human-built dams which store the sediment that would otherwise be

deposited in these regions, leading to relative SLR (IPCC (2013)). Human manage-

15

Figure 2.2: The change of the mean temperature and the carbon dioxide in the atmosphere for
the global in the past and future, adopted from (Clark et al. (2016)).

ment of coastal areas is very important in protecting coastal areas from saltwater

intrusion. Irrigation and drainage should be controlled and projects in these regions

should be continuously maintained as well (IPCC (2013)).

4. Weather activity like hurricanes and storms can affect the coastal areas. Also, earth-

quakes, tsunamis, and the El Nino phenomenon can have greater effects from rising

SLR and further damage coastal regions.

The SARR and the MRDR have different rates of SLR. MDR has higher SLR but the

16

Figure 2.3: Total sea level change since 1993 based on TOPEX/Poseidon.

geographic characteristics in SARR may enhance the influence of SLR in this region.

2.2 The Geographical Characteristics

The study of geographical characteristics is important to clarify the risks of SLR in both re-

gions. Table 1 compares and contrasts these general characteristics for MRDR and SARR.

Both regions have important similarities and differences. A review of the characteristics

in both regions can help to identify a robust and general model to detect and characterize

SLR impact on these near coastal riverine areas. This chapter compares and contrasts the

SLR-related social and environmental characteristics in SARR and MRDR.

17

Characteristics in both of
the regions
Geomorphological

Environmental

Climatic characteristics

Economic importance

Population

Cities

Mississippi River Delta re-
gion
Land created by sediment
from Mississippi River
mouths in the Gulf of
Mexico
Most region is marshes

Mild winters and hot and
humid summers
Shipping and commercial
fisheries. Oil and gas de-
posits
More than two million
people
New Orleans

Shatt Al-Arab River region

Land created by sediment
from Tigris, Euphrates and
Karun Rivers in Persian
(Arabian) Gulf
Most region is agricultural
fields
Hot desert climate

Shipping and commercial
fisheries. Oil and gas de-
posits
More than three million
people
Basra

Table 2.1: Comparison of the study regions.

2.2.1 Topography Characteristics

Most parts of both regions are less than one meter above sea level as shown in Figure

(2.4). According to SLR scenario (Williams (2009)), land with absolute elevation below

one meter is very vulnerable to SLR. Many studies have presented the problem of SLR

in MRDR from a variety of disciplinary perspectives, including engineering, biology and

geography (Burkett et al. (2003)). There is no published study specifically investigating

the impact of SLR on the SARR because of the lack of quality digital elevation data for

this region. So one of the objectives of this research is to enhance the digital elevation

data to show vulnerability to SLR in this region (El Raey (2010)). However, both regions

are threatened by SLR and both will be flooding due to SLR inundating low elevation

areas. The topographic characteristics of both regions are considered in the following

subsections.

18

Figure 2.4: Both SARR and MRDR have high rising sea levels, adopted from (Williams (2009)).

2.2.1.1 Topographic Characteristics of Shatt Al-Arab River Region

This region includes land on both sides of Shatt Al-Arab south of the city of Basra to Per-

sian (Arabian) Gulf, as shown in Figure 2.5. Elevations in Shatt Al-Arab region descend

slowly from 2m in the north to 0.5 m and less in the south. Land elevation also decreases

from the banks of Shatt Al-Arab 5.2 m to less than 1 millimeter in the lowlands away

from the river (Al-Saaidy (2008)). The total area of Shatt Al-Arab study region is 2189 km

2. River width increases from about 0.5 m at Basra to 0.8 km at its mouth. Along the set-

tled banks there are date-palm groves, which are naturally irrigated by tidal action (Saad

(1978)).

2.2.1.2 Topographic Characteristics of Mississippi River Delta Region

This region ranges along Mississippi from its junction with the Bayou Lafourche down-

river to the Gulf of Mexico and covers 11,894 km 2, as shown in Figure 2.6. This study

19

Figure 2.5: Shatt Al-Arab river region in southeastern Iraq.

focuses on an area around the river that is part of Louisiana coastal plain. This important

region, which is mostly coastal wetlands, has a distinctive ecology and landscape and is

rich in resources (Steyer and Llewellyn (2000)). The width of the alluvial region along

MRDR from 15 to 100 km. Mississippi River flows along a ridge formed by its own nat-

ural deposits (known as a levee), from which the land elevation declines at an average

fall of 33 m/km. The alluvial lands along other streams have similar terrain (Zellmer and

Klein (2007)).

20

Figure 2.6: Mississippi River Delta region in Louisiana, USA.

2.2.2 Geology Characteristics

The coastal areas with unstable geology may experience a higher relative rate of SLR (Ver-

meer and Rahmstorf (2009); Rahmstorf (2007)). It is, therefore, important to understand

the influence of the geologic characteristics in both regions. The following subsections

report the geology of each region in this study.

21

2.2.2.1 Mississippi River Delta Geology and Ecology

The MRDR formed 100 million years ago in the late cretaceous Period (Galloway et al.

(2011)). Since the end of Cretaceous times, the central portion of North America has

drained to the south into the Gulf of Mexico. Mississippi and other rivers drained of

the Appalachian Mountains to the east and the Rocky Mountains at the western margin

and delivered large volumes of sediment to the northern rim of the Gulf of Mexico. The

ancestral Mississippi and other rivers created Progradation in sedimentary geology and

geomorphology. The term progradation refers to the growth of a river delta farther out

into the sea over time (Galloway and Hobday (1983)). MRDR created six delta lobes

during the Holocene. These lobes are as follows: the Maringouin delta formed 7,500 to

5,500 years ago, the Teche delta created 5,500 to 3,500 years ago, the St. Bernard delta

created 4,000 to 2,000 years ago, and the Lafourche delta created 2,500 to 500 years ago.

The modern delta, called the Plaquemines-Balize delta or Birds Foot Delta, began forming

1,500 years ago, and the Atchafalaya and Wax Lake Deltas began forming 500 years ago

(Blum and Roberts (2009)).

The dominance of this river system produces a complex branching small river width

to length ratio for the delta plain. The current annual sediment discharge of 145 million

metric tons marks a substantial reduction from the 400 million metric tons per year of

material that built most of the delta before large-scale human activities modified the river

basin with dams, locks, meander cutoffs, and other engineering structures (Giosan and

Freeman (2014)). There are different kinds of swamp and marsh habitats in the delta

plain: the upper delta plain has freshwater habitats, while the lower delta plain features

brackish to saline marshes because of tidal inundation (Giosan and Freeman (2014)).

22

The MRDR is experiencing the highest relative SLR in the U.S. due to the combination

of global SLR and local subsidence (sinking land) (U.S. Global Change Research Program,

2009). Great subsidence is possible in river deltas where rapid sediment accumulation

traps a significant amount of water. Over time, as layers of new sediment are deposited,

water is squeezed out of the underlying deposits, causing compaction and the land sur-

face to sink. Subsidence rates in the MRDR have increased due to the presence of levees

and diversions which have reduced river flooding of the marshes behind the levees, starv-

ing them of sediment. On another hand, the soil compaction form oil and gas extraction

in the region cause coastal land loss (Blum and Roberts (2012)). While levees can limit

flooding damage, their presence also reduces sediments that build the delta. Taken to-

gether, these factors have caused the recent loss of about 4,920 km 2 of coastal wetlands

in the state, more than one-third of its coastal plain (Blum and Roberts (2012)). The analy-

ses of landscape change presented in (Couvillion et al. (2011)) report the use of historical

surveys, aerial data, and satellite data to track landscape changes. Summary data are pre-

sented from 1932 to 2010; trend data are presented from 1985-2010 with analysis based

on satellite and aerial imagery. These findings show that Louisiana coast has undergone

a net change in land area of about -3030.3948 (km 2) from 1932 to 2010 (Couvillion et al.

(2011)).

2.2.2.2 Shatt Al-Arab Geology and Ecology

This region includes land on both sides of Shatt Al-Arab south of the city of Basra to Per-

sian Gulf. Shatt Al-Arab region descends slowly from 5m a.s.l. in the north to 0 in the

south. Land elevation also decreases from 5m along the riverbanks to less than 1 millime-

ter in the lowlands away from the river (Al-Saaidy (2008)). Total area of Shatt Al-Arab

23

study region is 2,188.615 sq km. Its width increases from about 37 m at Basra to 0.8 km at

its mouth. The water level, in at least the lower portion of the system, is considerably af-

fected by the tides of the Gulf (Saad (1978)). SARR is located on the unstable tectonic plate

(Buday et al. (1987)). SARR is part of the north of Persian Gulf. Much of the deformation

and seismicity responsible for the high-relief topography of the Zagros Mountains and

northeastern margin of Persian Gulf can be explained within the context of active tecton-

ics associated with plate convergence. Known to be a tectonically quiescent region is the

seismic low-relief coastal areas along the southern sideline of the Gulf. As a result, sea-

level changes have been a focus in previous studies as a description of the evolution of

the Pleistocene and Holocene coastline site. We hypothesize that the tectonic movement

was more significant in the late-Pleistocene/Holocene shore development as opposed to

the sea-level variations which also influenced this region (Wood et al. (2012)).

2.2.3 Hydrology Characteristics

The flow of water through a river basin is of course important for conditions in its delta.

Discharge is affected by the size of the basin, the balance of precipitation, evapotran-

spiration within that basin and also by water storage and diversion upstream. Human

influence can play a major role because of dam construction and irrigation, as well as

climate change. In this section, the hydrology of both regions in this study are discussed.

2.2.3.1 Mississippi River Flow

Mississippi River is one of the world’s greatest in size, environmental diversity and nat-

ural fertility. It carries the greatest volume of water in the US, flowing 3,705 km from

its origin at Lake Itasca in Minnesota, through the mid-continental United States, and to

24

the Gulf of Mexico Riverine Plain and its subtropical Louisiana Delta (Zellmer and Klein

(2007); Blum and Roberts (2012)). Within the state of Louisiana, Mississippi River flows

from north to south for a length of approximately 1,000 km and drains into the Gulf of

Mexico; the Red River; the Ouachita River and its derivatives; and additional secondary

streams. The extent of the alluvial area along Mississippi is of 16 to 97 km.

The study area in MRDR region ranges along Mississippi from its junction with the

Bayou Lafourche downriver to the Gulf of Mexico 11894.27 sq km (Couvillion et al. (2011)).

We are focusing on an area around the river that is part of the Louisiana coastal plain (Bar-

ras (2006)). This important region, which is mostly coastal wetlands, have a distinctive

ecology and landscape and is rich in resources (Steyer and Llewellyn (2000)). The Missis-

sippi River flows along a ridge formed by its natural deposits (known as a levee), from

which the lands decline at an average fall of 3 m/km. The alluvial lands along other

streams present similar features (Zellmer and Klein (2007)).

2.2.3.2 Shatt Al-Arab River Flow

SARR is in southeastern Iraq and extends from the north to the center of Basra city until

the southeastern part of it to form the international border between Iran and Iraq before

flowing into Persian Gulf. SARR is formed by the junction of Tigris and Euphrates rivers

at Al-Qurnah city. It continues southeastward for 193 km and flows past the Iraqi port

of Basra and the Iranian port of Abadan before draining into Persian Gulf. For about the

last half of its route, the river forms the boundary between Iraq and Iran; it receives a

tributary, the Karun river, from the eastern (Iranian) side. Its width increases from about

40 m at Basra to 0.8 km at its mouth. Along the river banks, there are date-palm groves,

which are typically irrigated by tidal action (Saad (1978); Buday et al. (1987)). However,

25

the water depth relatively decreases at the confluence of the Karun and Shatt al-Arab

rivers, due to the continuous accumulation of large amounts of sediments transported by

Karun to this locality. The water level is considerably affected by the tides of the Gulf. The

water of Shatt Al-Arab may reach Hor Hammar, a large marsh lying at the northwestern

sides of Sindbad Island, during the high tide, and the brackish water of this Hor may

subsequently enter into Shatt al-Arab and affect its salinity. The level of Shatt al-Arab

rises in spring mainly due to the increase in the levels of Tigris and Euphrates during this

season. The increase in river flow follows the melting of snow in Turkey. The fresh water

from this flood decreases the salinity of Shatt Al-Arab during spring (Al-Saaidy (2008);

Saad (1978)).

2.2.4 Climate Characteristics

There are some distinct climatological issues in both study regions that directly affect the

SLR problem. For MRDR, it is a region affected seasonally by powerful tropical storms.

These storms can exacerbate SLR issues in several ways: first, during the storm itself, by

pushing massive amounts of seawater into areas not normally inundated; second, after

the storm, by damaging barrier islands and other features which act to reduce the intru-

sion of the sea into marshlands. SARR, in contrast, is affected by high seasonal variation

of the height of Persian Gulf, which may bring impacts of higher sea level much earlier

to the region. Also, water demand in SARR must be satisfied by rainfall in distant areas,

which has consequences for the viability of human activity in the area.

Surface water is not the only water affected by SLR according to the Ghyben Herzberg

relation equation (Herzberg, 1901; Badon-Ghijben, 1888; Badon-Ghijben, 1889): aquifer

water will have higher concentrations of salts in regions with higher temperature. So, the

26

influence of the SLR on groundwater in this region will be high (Barlow (2003)). MRDR

region is a Cfa climate based on the Koppen climate classification as depicted in Figure

(2.7). It is a humid subtropical climate defined by mild winters and hot and humid sum-

mers which exhibit mean temperatures between 0°C (32°F) and 18°C (64°F) in the coldest

month and mean temperatures between 22°C (72°F) or higher in the warmest month (Peel

et al. (2007)). In contrast, SARR is a hot desert climate (BWh) according to the Koppen

climate classification, as shown in Figure 2.7. BWh is an arid climate which supports little

or no vegetation (Peel et al. (2007)).

Figure 2.7: World climate regions according to Koppen-Geiger classification, adopted from (Peel
et al. (2007)).

27

2.2.5 Human Activity

There are many studies concerned with the effect of SLR on MRDR including the rise of

saline groundwater levels, storm waves, and flooding. Both regions are affected more

because they are connected directly with Mexican Gulf and Persian Gulf (Steyer and

Llewellyn (2000)). Many studies on land change and inundation in the MRDR have iden-

tified relative SLR and land subsidence as primary factors (Blum and Roberts (2012)). In

contrast, studies on environmental degradation in SARR have not diagnosed SLR as the

source of the problems (Al-Saaidy (2008); Jabbar and Zhou (2011); Al Isawi (2011)). Both

regions have been affected by human activities like oil extraction and irrigation which

accelerate SLR effects (Blum and Roberts (2012)). There are also large differences between

the regions. For example, MRDR has a lot of projects to conserve the region. Most no-

tably, Louisiana’s Coastal Protection and Restoration Authority (CPRA) was developed

in 2007 to implement projects to restore and protect the region. Federal efforts like CPRA

invested more than 20 billion dollars on a range of activities, including levee improve-

ment, fill and construction of berm and barrier islands (Louisiana (2012)). Figure 2.8 is an

illustration from CPRA detailing their efforts. On another hand, there is nothing spent to

protect SARR.

2.3 Conclusion

This chapter has outlined the Geography, Geology, Geomorphology, Climate, and Human

activities in both regions and how these characteristics accelerated SLR:

1. Geologic development of both the Gulf of Mexico and Persian (Arabian) Gulf and

how this relates to the general hydrology, sediment transport and solute chemistry

28

Figure 2.8: Projects by Louisiana’s Coastal Protection and Restoration Authority since 2007,
adopted from (Louisiana (2012)).

of the two river systems.

2. Hydrology of Mississippi, Tigris and Euphrates.

3. Changes in hydrology, sediment transport and solute chemistry of both regional

river systems caused by various forms of anthropogenic stress (climate and engi-

neering structures).

Both regions are similar in geomorphological, environmental, economic and climatic

characteristics. Both are geologically recent creations from the deposits of several of the

world’s greatest rivers (Tigris, Euphrates, and Mississippi) flowing into gulfs (Persian

Gulf and the Gulf of Mexico). Both regions are important regional economic centers for

agricultural production, oil production, fishing and marine transport. Both rivers have

natural deltas that have formed where the mouth of each river meets its gulf. Indeed,

in recent (1000s’ of years) times, the sea extended hundreds of miles inland, and land in

these regions are based on alluvium. Also, land areas on either side of these rivers are

29

highest along the edges of the channels and decrease in elevation farther from the river.

Delta building for both rivers is an integrated and long-term process that has given both

regions a unique mix of ecosystems, habitats and landforms. Both regions face increasing

pressure from subsidence, erosion, sea level rise and rising salinity in groundwater, lead-

ing to coastal retreat in these regions because of the climate change and human activities

there.

30

Chapter 3

Causes and Consequences of

Environmental Degradation Along Shatt

Al–Arab River: A Coupled Human and

Natural System Perspective

3.1

Introduction

Coupled human and natural systems (CHANS) research concepts explicitly acknowledge

the feedbacks between human drivers and natural processes through interactions or flows

(Marina et al. (2011); Phillips (2016); Tu (2011)). Research that utilizes a CHANS ap-

proach focuses on understanding feedbacks, nonlinearities, thresholds and legacy effects

across multiple spatial, temporal and organizational scales (Geri et al. (2010)). CHANS

perspectives may be especially useful for the analysis of marginal geographic areas and

the people who live on them, as such areas may be particularly vulnerable to social and

environmental hazards. Few geographies are more marginal than the land along Shatt

Al-Arab River, which begins at the marshy confluence of Euphrates and Tigris rivers in

southeastern Iraq and ends 120 km away at the head of Persian Gulf. The lower portion

31

of the river serves as the international boundary between Iraq and Iran, dividing the re-

gion politically. Environmentally, the region is a transition zone between land and sea:

most of the area is close to sea level, and in historic times, the waters of Persian Gulf have

overlain the region along the river (Meehl et al. (2005)).

The environmental degradation in this region has influenced large areas around Shatt

Al-Arab River that provide important human services like agriculture. Extending away

from the riverbanks toward the desert, the palm tree plantation extents increase from

the south to the north of the region. In the midtwentieth century, this was one of the

worlds most important date palm forests, with more than 17 million trees in 1975. The

region now has around 2 million palm trees. More than 14 million trees died because

salinization developed starting in the late 1960s (Singh et al. (2005)). Of more than 52,000

hectares of SARR date palm plantations in 1975, (Singh et al. (2005)) found that only 11,000

hectares or 21% remained in 2002. More than 14 million palms were destroyed due to salt

and pests: around 9 million in Iraq and 5 million in Iran. Moreover, of the remaining 34

million palms, many are in bad condition (Singh et al. (2005); Brandimarte et al. (2015)).

A number of studies have focused on factors leading to this degradation, including land

change increasing concentration of salt in the river (Al-Mahmood (2009); Jabbar and Zhou

(2011); Essa (2012)), and salinization of land around the river (Mahmood et al. (2013)). All

studies agree that naturally occurring tidal action is the source of salinity in the river as

well as the salinity in the region. In addition, studies have implicated human activity

as having played an essential role in the environmental degradation of this transitional

region (Jabbar and Zhou (2011); Al Isawi (2011)), including the construction of upstream

reservoirs and a high frequency of devastating war within the area.

Remote sensing-based analysis and geospatial methods can be essential to understand

32

CHANS at multiple spatial, temporal and organizational scales. Long-term records of

satellite image data can be used to identify land cover change and key human and envi-

ronmental processes driving these change (Emel et al. (2014); Liu et al. (2003); Fox et al.

(2003)). To date, however, no study of SARR has comprehensively itemized the human

and environmental factors that have led to degradation over the past five decades or

evaluated their spatial impact. To address this gap, we develop a CHANS framework to

identify conflict, water engineering, economic development policy and climate change to

environmental degradation and food security in the SARR. Further, we quantify change

within the region since the mid-1970s using a mix of ground and remotely sensed data

to assess change in normalized difference vegetation index (NDVI). The objectives of this

chapter are to:

1. Identify primary human and environmental drivers in SARR system.

2. Characterize NDVI as a proxy for vegetation quality and quantity and measure it

over a five-decade period.

3. Identify spatial and temporal changes in NDVI throughout SARR.

4. Link observed changes to dominant drivers of change in SAAR.

3.2 Data and Methods

3.2.1 Study Area

SARR is in southeastern Iraq and southwestern Iran and includes land on both sides of

Shatt al-Arab River from north of Basra City to the Persian Gulf as depicted in Figure (3.1).

33

Its total area is 2,189 km2 (Al-Saaidy (2008)). Land elevation along the river descends

gradually from five meters above sea level near Al Qurnah in the north to zero in the

south. The banks of the river generally have higher elevations than interior areas. The

Shatt al-Arab River flows through Iraq for about 80 km to just above the confluence with

the Karun River from the east; from that point to the Gulf the river forms the border

between Iran and Iraq. Cities and towns, which have more three million persons as shown

in Fig. 1. interspersed with agricultural land and marshes, occupy its banks. The date-

palm groves along the river are naturally irrigated by tidal action (Saad (1978)). Most

studies treat the Shatt Al-Arab as a unique river because it has a tide two times during

the day. Tidal movement from the Gulf controls the hydrology of the river, especially

after the decline in freshwater coming from Euphrates, Tigris, and Karun rivers (Abdullah

et al. (2015)) in the period evaluated in this study (1975 to 2017). Tides enter the river from

Persian Gulf and reach at least as far as Basra City (Cressey (1958)).

3.2.2 Coupled Human and Natural System

To understand the challenges SARR has faced in recent decades, this study develops a

framework for both human and natural systems to highlight the dominant factors re-

sponsible for environmental degradation in the region. This CHANS framework is repre-

sented in the conceptual model shown in Figure (3.2). This model shows how human and

natural systems affect the region and how they are linked to environmental degradation

in SARR during the periods 1975, 1985, 1995, 2005, and 2017.

The concept Figure (3.2) shows how human and natural systems affect the region and

how they are linked to environmental degradation in the SARR during the periods (1975,

1985, 1995, 2005 and 2017).

34

Figure 3.1: Total study area of Shatt Al-Arab river region.

3.2.2.1 Human Impacts

The human carbon footprint has spiked more in the last 50 years than in any other period

in history, and continues to increase by 2% per year (Sohl et al. (2012); Smiley et al. (2016)).

Ecosystems are negatively impacted with less biodiversity because of human interaction.

Water ecosystems are responding negatively to human activities, such as pollution, habi-

tat destruction, and overfishing and are becoming more unpredictable and less resilient

35

Figure 3.2: Total conceptualization of coupled human and natural systems (CHANS) in Shatt
Al-Arab river region.

to external pressures (Jackson et al. (2001); Liu et al. (2007)). Human activities are re-

sponsible for global climate change and its effects on plants and animals (Liu et al. (2007);

Wohlfart et al. (2017)). These influences can be seen in SARR. The main human systems

directly influencing this area are:

• War

• Government land policy

• Upstream hydroengineering

• Pollution and degradation

36

3.2.2.2 Environmental Impacts

While factors like war and environmental pollution provide direct local impacts on SARR,

and factors like upstream dams are a regional source of challenges in the area, climate

change is the main global-scale challenge facing SARR. Climate change drives sea level

rise and meteorological factors leading to regional drought, which in turn negatively im-

pacts vegetation, agricultural productivity, and the livelihood of the people in this region

(Jabbar and Zhou (2011); Al-Bahili et al. (2010)).

I identify and assess three specific environmental consequences:

• Meteorological factors associated with drought

• Sea level rise

• Groundwater salinity

3.2.3 Methodology

CHANS drivers and effects are complex, multi-factored, and evolving. To provide as

comprehensive an assessment as possible, we collected a variety of datasets to evaluate

spatial and temporal characteristics of a range of environmental properties for SARR.

Long-term (19752017) monthly meteorological data (Minimum and Maximum Tempera-

ture, Relative Humidity, Rain Quantity) were obtained from researchers at the University

of Basrah (Yassein et al. (2016)) to evaluate annual patterns and identify changes over

the duration of the study period data on salinity of water from the Shatt Al-Arab River

from 1980 to 2012 was obtained from Hamid (2014) and Abdullah (2015). I also utilized

mosaicked pairs of Landsat images that cover the study area in each of the years eval-

37

uated (1975, 1985, 1995, 2005 and 2017). Table 3.1 lists the characteristics of each image

set. All Landsat images were systematically radiometrically and geometrically corrected

and transformed to a common analysis framework with standard output map projection

UTM zone, image orientation, and 30 m pixel size. We calculated the normalized dif-

ference vegetation index (NDVI) at each time period. NDVI is a standard measure of

vegetation intensity and a signal of land use and land cover (Fuller (1998); Justice et al.

(1985); Prince (1991); Goward and Prince (1995)). It is measured on a per-pixel basis as

the normalized difference between the red (RED) and near infrared (NIR) bands of an

image: NDVI = (NIR - RED) / (NIR + RED). The analysis was restricted to those areas

along the Shatt Al-Arab River with a 1975 NDVI value above 0.1. Regional NDVI changes

over time were evaluated, and correlations between NDVI, climatological variables, and

salinity were assessed. Finally, NDVI change from 1975 to 2017 was examined within the

study area. Three subareas were identified: the northwestern region above the confluence

of Shatt Al-Arab and Karun Rivers, which falls almost entirely in Iraq (Upriver), the right

(southern) bank below the confluence, which falls in Iraq (Downriver Iraq), and the left

(northern) bank below the confluence which falls in Iran (Downriver Iran). NDVI change

within each subarea is assessed to understand how degradation processes differentially

affect these areas.

Sensors

Satellite

MSS
TM
ETM+

Landsat 1-2
Landsat 4-5
Landsat 7-8

Overpass / Or-
bit Frequency
18 days
16 days
16 days

Data-Record
(years)
1975 and 1985
1995 and 2005
Landsat 7-8

Spatial Resolu-
tions
57 m
30 m
30 m

Table 3.1: Satellite sensors utilized for the study used three sensors to cover the time (1975,
1985, 1995, 2005 and 2017).

38

3.3 Results

1. The chapters second objective is addressed by using NDVI to evaluate regional re-

sponse over time in vegetation due to the drivers highlighted in the previous sections.Figure

(3.3)(a)(e) present maps of SARR NDVI derived from Landsat imagery in 1975, 1985,

1995, 2005, and 2017, while 3f provides NDVI density plots for each time period. The

maps qualitatively illustrate substantial changes over time in intensity and distribution

of vegetated areas, which are largely date palm plantations, other agriculture, and marsh.

Modal vegetation density was highest in 1975 and lowest in 1995, while the upper tails

were thickest (indicating more high-NDVI pixels) in 2005. The substantial spatial and

temporal variation in vegetation characterized by these findings are responses to human

and environmental drivers identified in this study.

Figure 3.3: Temporal variations of NDVI (A-1975, B-1985, C-1995, D-2005 and E-2017 NDVI
greenness density plot) for the Shatt Al-Arab river region.

The third objective links climate indicators to NDVI trends by regressing mean NDVI

39

for each year (1975, 1985, 1995, 2005, and 2017) with mean temperature, precipitation, and

sea level in the same years. Figure (3.4) graphically illustrates the trends in sea level, tem-

perature, precipitation, humidity and NDVI. This modeling indicates that: (i) there is a

significant positive relationship between the NDVI trend (as an independent factor) with

the SLR and precipitation variables (as dependent factors). Meanwhile, the correlation

was considerably negative with the mean air temperature; (ii) these observed relation-

ships would result in a declining trend of vegetation greenness (NDVI values), by affect-

ing the vegetation directly (e.g., inhibiting growth due to inauspicious daily meteorologi-

cal conditions) and indirectly by degrading soil characteristics (e.g. through groundwater

level rise and increasing soil salinity) at such coastal/estuarine regions.

2. The fourth objective is to understand intra-regional NDVI change between 1975

and 2017. Figure 3.5 (a) maps NDVI change by pixel within the study area. In general,

the region witnessed declines in NDVI, but spatial heterogeneity is apparent. Some very

large decreases are evident along the riverbanks in the Upriver subarea, and in places

elsewhere along the right-hand, Iraqi bank. Few substantial increases in NDVI are appar-

ent anywhere in Iraq. In contrast, much of Lower Iran has experienced modest increases

in NDVI values. The southernmost portion closest to Persian Gulf has experienced little

or negative change on both sides of the river. Figure 3.5 (b) presents box plots of NDVI

change distributions within each of the three subareas. The Upriver and Lower Iraq sub-

regions experienced general declines in vegetation cover with similar, largely negative

distributions. In contrast, Lower Iran experienced overall increases in NDVI change dur-

ing this time, with a mean increase of 0.09. The observed changes have large magnitudes

relative to the overall distribution of NDVI values, which average 0.081 and 0.13 in 1975

and 2017, respectively.

40

Figure 3.4: Regressing trends of temperature, precipitation, humidity, and sea levels for the peri-
ods (1975, 1985, 1995, 2005, and 2017) associated with the NDVI trends.

3.4 Discussion

The changes observed from 1975-2017 in vegetation cover in the SARR derive from the

reciprocal interactions between natural and human components. The remote sensing and

geospatial analysis reported in the previous section is valuable for understanding impacts

of coupled human and natural systems, enabling the development of testable hypotheses

and better policy decision making for climate change adaptation at multiple spatial, tem-

41

Figure 3.5: Comparative between 1975 and 2017 NDVI in Shatt Al Arab Region (a,b)which shows
the difference in the density of vegetation distribution between up and down the junction of Karun
river in Iraq and Iran with Shatt Al Arab river.

poral and organizational scales. In this section we consider a range of factors associated

with both direct human impacts and direct environmental impacts, as presented in the

conceptual model in Figure 3.2 that could explain these changes.

3.4.1 Human Impacts

Human activity has been influencing the region in different ways as detailed below.

3.4.1.1 War and Government Policy

SARR was known for palm trees and crop cultivation; much of this agricultural resource

has been lost during the time evaluated in this study because of government policies

(Singh et al. (2005)). In the 1930s, the population in Iraq started relocating from rural to

urban areas due to oil industry development, and consequently the agricultural sector be-

42

gan to suffer and become stagnant (Gibson et al. (2017)). Once Saddam Hussein became

president in 1979, an effort was made to put agriculture back into the private sector. Dur-

ing the IranIraq War (1980 to 1988) Iraq was able to increase its production and expand

crops in spite of labor resources being diverted to the war. At the same time, the war

resulted in a significant economic impact causing delays in development projects and

defaults to foreign contractors and loans (Springborg (1986); Gibson et al. (2017); Tripp

(2007)). Much of the war was fought within SARR: the eightyear period of the IranIraq

war centered on military operations in this borderland. The war left behind many types

of military waste, including mine fields, endangered the lives of the regional population,

and rendered vast areas (29,000 acres) of SARR unfit for human use up to the present time

(Mohammed (2008)). The inability to return the population to these areas is a source of

ongoing human casualties and a major obstacle for regional agricultural and economic re-

habilitation (Mohammed (2008)). In 1990, Saddam Hussein invaded Kuwait and the UN

Security Council imposed sanctions against Iraq. When the sanctions failed, the UN used

military force, attacking Iraqs leadership, production, infrastructure, and population. The

conflict destroyed Iraqs economic infrastructure, causing extensive damage, even greater

than that from the years of war with Iran (Gibson et al. (2017); Tripp (2007); Reynolds

(1995)).

The comprehensive sanctions continued after military conflict ended and became stricter

as Saddam Hussein remained defiant. As foreign companies were prohibited from invest-

ing in Iraq with oil exports, revenues needed for Iraq to purchase food and agricultural

products were reduced. There was a ban on agricultural dual-use products including

fertilizers, machinery, pesticides, and chemicals, and many countries stopped exporting

food to Iraq (Gibson et al. (2017)). Consequently, increased domestic agricultural produc-

43

tion became a high priority, and the government created both incentives for farmers to

expand their crop areas and punishments, including death, if they did not meet specified

production quotas (Foote et al. (2004); Gibson et al. (2015); Gibson et al. (2017)). At the

same time, a food rationing system was set up at local food stores for everyone in the

country to received distributed rations. When the effort failed to meet basic needs rec-

ommended by the World Health Organization, the UN set up the Oil-for-Food Program

in 1995. Iraq was allowed to sell oil on the world market in exchange for food, medicine,

and humanitarian needs. The sanctions continued until the U.S. invasion of Iraq in 2003.

The effects of war and sanctions in Iraq had a devastating effect on the agricultural sec-

tor. Soil fertility was depleted, irrigation infrastructure was neglected or destroyed, and

cropland suffered from salinization (Gibson et al. (2017); UN (2003) ). Once operation

Iraqi Freedom began in 2003, the Hussein regime fell without much resistance. The U.S.

led operations to rebuild the country and reinstate political stability. Once the sanctions

were lifted, the food imports began again. When aid organizations assisted with the food

rationing systems, local farmers found it difficult to compete with the subsidized food

imports (Foote et al. (2004); Gibson et al. (2017); Gibson et al. (2015)). In contrast to the

1990s, when food imports stopped, and the size of cultivated areas increased, the past

fifteen years have seen a reduction in the area of cultivated land in Iraq as food imports

resumed (Gibson et al. (2012); Gibson et al. (2017)).

3.4.1.2 Upstream Water Demand

The available water in the Shatt AlArab River is greatly impacted by increased water de-

mands upstream. Its four major tributaries, the Tigris, Euphrates, Karkheh and Karun

Rivers, rise in the mountains of Turkey and Iran. The entire basin is shared by Turkey,

44

Iran, Syria, Saudi Arabia, and Jordan, which are all neighbors of Iraq. The surface area is

about 938,300 km2 (with Iraq providing just 15.81% of the total water volume of the basin)

(Al-Asadi (2017)) making it the largest river basin in southwest Asia. Climate change,

dam construction, population growth, and increased demands for irrigation is putting

added pressure on water withdrawal while the arid climate is keeping the regeneration

rate low. Dam building upstream is decreasing the amount of available freshwater down-

stream and is significantly affecting the quality of water from Basra to the mouth of the

river. International water management policies between Turkey, Iran, and Iraq are needed

to avoid any further water quantity issues (Beaumont (1996); Abdullah et al. (2015)).

3.4.1.3 Hydroengineering in SARR

Figure 3.5 shows the NDVI in the southwest of the region has a clear decline of the green-

ness in the land cover, more than in the southeast and It appears that loss of NDVI has

been very similar in the northern, Upriver subregion and the southwestern, downriver

Iraq portion. The southwest close to the Persian Gulf is affected by sea level rise, and it

does not have more fresh water like the northern part of the region. The southeast region

has higher greenness because the Iranian Government redirection of Karun river water

and used the water for irrigation in this part of the region. That might explain the differ-

ent land degradation trajectories of these subareas. This study suggests that two related

factors may be at work: first, a recent diversion of the systems main source of fresh wa-

ter to benefit primarily fields in the Lower Iran subarea, and second, different degrees of

investment by Iraq and Iran in these areas. In 2008, the Iranian government redirected

the flow of Karun River to the Kashmir channel, which helped revive the vegetation and

agriculture fields in this area (Hamid (2014)). On the other hand, redirecting Karun River

45

water caused a humanitarian catastrophe because of the high concentration of Salinity

River water in the city of Fao to the south on Shatt Al-Arab River (Hamid (2014); Abdul-

lah et al. (2015)). Given that those portions of the region reliant on Shatt Al-Arab river

water experienced a decline in vegetation cover while the one area with contributions

from Karun saw an increase, we maintain that decreasing water quality (largely due to

rising salinity) from Shatt Al-Arab is in large part responsible for the regional vegeta-

tion decline. Other possible factors may be differing national policies in research and

investment in date palm agriculture in this region (Hajian and Hamidi-Esfahani (2015);

Hazzouri et al. (2015)).

3.4.1.4 Pollution and Degradation

A major source of environmental degradation in SARR is due to pollutants directly or

indirectly discharged into water bodies without adequate treatment to remove harmful

compounds. According to several environmental assessment studies, SARR has not made

any progress in addressing its environmental issues or improving environmental quality

(Nomas (2006); Al Isawi (2011)). Further, there are no regional management policies for

the mitigation of human impacts on the environment, such as maintaining irrigation and

drainage systems in the region or increasing natural vegetation problems in the rivers

(Jabbar and Zhou (2011); Nomas (2006)). The political division of the region between Iraq

and Iran complicates coordinated effort on this problem.

3.4.2 Environmental Impacts

Figure (3.4) shows that NDVI has decreased over time, as has precipitation and humid-

ity. There have been increases in temperature and sea level from 1975 to 2017) that are

46

associated with the negative NDVI trends. The physical components associated with en-

vironmental degradation in the region are discussed in this section.

3.4.2.1 Meteorological Factors Associated With Drought

The study area is situated in the Subtropical Desert Koppen climate classification (BWh),

characterized by warm, dry summers (Peel et al. (2007)). The mean maximum tempera-

ture can reach over 48°C in the warmest months, while the winter is cold, with the mean

minimum temperature dropping to 5°C. Annual precipitation in the region is less than

100 mm and 90% of the annual rains occur between November and April (Muslih (2014)).

Monthly data obtained for the Basra weather station (Basra Weather Station 2017) provide

clear evidence of trends across many meteorological factors in the past five decades. Fig-

ure (3.4) illustrates consistently rising average temperature, constant precipitation, and

decreasing relative humidity since 1975 to 2017. Changes in these climatic elements have

direct impact on the environment in SARR.

3.4.2.2 Sea Level Rise

Sea level has risen globally in recent decades. A number of regional studies have used

tidal gauge data in Persian Gulf to estimate relative sea level change for locations near

the outlet of Shatt Al-Arab River. Alothman (Alothman et al. (2014)) constructed a time-

series record from seven coastal tide gauges in this area. A relative sea level rise of 2.2
± 0.5 mm/year was observed from this record over the period 19792007. This result

contrasts with the findings of Hassanzadeh (Hassanzadeh et al. (2007)), who reported a

rate of 2.8 mm/year (19901999) for the northern part of Persian Gulf. This study used

data from tide gauges at Bandar Abbas, and Bushehr combined with atmospheric data at

47

both stations to investigate the Mean Sea-Level (MSL) response to meteorological forcing

functions in this region.

An earlier study by Sultan (1995) estimated a constant rate of sea level rise of 2.1 ±

0.1 mm/year (1980–1990) using two tide gauges. In an update, Sultan (2000) identified a

sea level rise of 1.7 mm/year (19801994) using nine tide gauges. Hossein (2007) reported
a mean rate of 2.34 ± 0.07 mm/year (1990–1999). The variability of these estimates of

mean sea level change may be due to incomplete and inaccurate records but can also be

related to the large seasonal variation of sea level in Persian Gulf due to air pressure:

water level in the Gulf is up to 26 cm higher in summer than in the winter (Sultan et al.

(1995a); Hassanzadeh et al. (2007)). Nevertheless, all studies are generally in agreement

that regional sea level rise is 2-2.5 mm per year from the 1980s into the 2000s.

3.4.2.3 Groundwater Salinity

The tide in Shatt Al-Arab happens two times during the day, which is considering the

source of salinity in SARR and the salinity of the groundwater as well. The salinity of

SARR has increased because of increased salt intrusion from the Gulf occasioned by re-

duced inflows from the upstream rivers which allow the seawater wedge to move further

upstream. Data collected and reported by Hamid (2014) and Abdullah (2015) provide in-

formation on temporal changes in salinity, measured in ppm, for the Shatt Al-Arab River

between Basra City and the Persian Gulf during 1980 to 2012, as illustrated in Figure

(3.6). This figure shows a strongly positive and increasing trend. The cultivable areas

in SARR are affected with varying degrees of soil salinity depending on the elevation of

land. Recent monitoring indicates that the regions soil salinity has increased. Soil salinity

is believed to be mainly responsible for low land use intensity in the area (Brandimarte

48

et al. (2015)).

Figure 3.6: Increase in salinity (ppm) in Shatt Al-Arab river from 1980 to 2012.

3.4.3 Conclusions

Coupled human and natural systems are complex, presenting an interdisciplinary chal-

lenge for understanding the range of factors that shape the trajectories of such systems.

This study developed a CHANS framework to understand the factors driving the past

five decades of environmental degradation along Shatt Al-Arab River. Through process-

ing of Landsat imagery across that span, the geographic distribution of vegetation quan-

tity (largely in the form of date palm agriculture) was assessed and linked to regional en-

vironmental data. We identified a range of factors that have contributed to degradation of

vegetation in this dynamic, marginal area. The most important factors are directly or indi-

49

rectly driven by global climate change. We believe that climate change is the greatest and

most challenging problem now facing SARR: dealing with this is of course a tremendous

challenge because climate change is a global problem.

Better water, land, food and energy management will reduce costs and increase ben-

efits for both humans and the environment in SARR and everywhere on Earth, but im-

portant scientific and policy questions remain. If developed and implemented correctly,

complex systems models, tools and analysis should advance the human development

and environmental sustainability agenda but achieving this will require concerted action

from all sectors and actors ranging from academia to the government to the private sector

(Bizikova et al. (2013)).

From the geographers point of view, the CHANS framework is groundbreaking for

its potential to understand and suggest solutions for geographic problems like environ-

mental degradation that operate over multiple spatial and temporal scales. CHANS can

account for the reciprocal interactions between human and natural components in the

SARR and identify the most relevant variables that have an essential role in this system.

The present study has shown that these variables (meteorological conditions, sea level

rise and human conduct) are consequences of climate change as in Shatt Al-Arab river

region. Of course, humans are primarily responsible for climate change, but the people

in the study area are also victims of climate change. They have played little role in the

regional and global factors driving environmental degradation in SARR. As sea level rise

accelerates in this century, threats to this low-lying coastal region will only become more

acute.

50

Chapter 4

Flooded With Error: Handling

Uncertainty in SRTM For The

Assessment of Sea Level Rise in

Mississippi River Delta

4.1

Introduction

SLR is one of the biggest threats to coastal habitats and communities worldwide (Mcleod

et al. (2010)) posing critical scientific and management challenges (Gesch (2009)). Estuar-

ine wetlands like MRD are characterized by complex interactions between surface water

fluxes, vegetation type and poor water movement. The tidal wetlands are very sensitive

to very small changes in topography within this tidal regime (Wang et al. (2007)). To un-

derstand the ecology of these environments, it is important to understand the hydrology,

and to recognize feedbacks from the ecology to the hydrological balance as well. Because

of the intimate relationship between estuarine wetland ecology and hydrology, knowl-

edge of wetland hydrology is critical if I am to predict and manage changes in wetland

environments. Long-term drivers such as climate change and sea level rise, as well as

51

human interferences of hydraulic modification of water flow are critical for these systems

as well (Hughes et al. (1998)).

Mississippi river delta located north of the Gulf of Mexico in southern Louisiana is

most noted for its extensive low-lying topography and its economic activities. MRDR has

the highest relative sea-level rise in the U.S. due to the combination of global sea-level

rise and local subsidence (sinking land). Based on analysis roughly corresponds with

measured water level changes: the mean sea level rise at Grand Isle, Louisiana (NOAA

gauge station 8761724) is 9.03 millimeters/year based on monthly mean sea level data

from 1975 to 2017 (Tides and Currents. noaa.gov).

Land-surface altitude data collected in this region indicated mean annual subsidence

of 5 mm per year during five survey epochs between 1951 and 1995 (Burkett et al. (2003)).

Subsidence rates in Mississippi River Delta are even further increased due to the large

amount of fluid withdrawal in the northern Gulf of Mexico. Oil and gas extraction leads

to increased soil compaction over a short amount of time (Blum and Roberts (2009); Pen-

land and Ramsey (1990)).

In addition, river dominated deltas such as MRDR have a different morphology than

wave or tidally influenced deltas such as the Nile and Ganges deltas. Mississippi delta

reflects primarily on the sediment input from rivers, which can be distinguished from

wave or tidal dominated deltas such as the Nile and Ganges deltas (W. van de Lageweg

et al. (2017)). Many previous remote sensing studies estimates MRDR flooding and how

much this delta lost land because of SLR. Couvillion et al. (2011) was based on survey and

aerial data and addressed concerns about comparability of the 1932 and 1956 datasets.

Blum and Harry (2009) was based on aerial and satellite imagery taken from 1932 to 2010.

Both studies reveal MRDR lost more than 25% of its land with a net change of about 4,000

52

km2 in land area from 1932 to 2010. About 95% of the decrease in land area accounts

for persistent losses, while the remainder of the area that has converted to water is not

classified as loss because it does not exhibit the persistence necessary to be classified as

such. Trend analysis investigations of wetland loss from 1985 to 2010 identify a rate of 43

km2 per year (Barras et al. (2008); Couvillion et al. (2011); Blum and Roberts (2009)).

There are different kinds of hydrologic models that could be used for inundation. The

advantage of GIS inundation models is inexpensive to run and doesn’t need a lot of time

either. In addition, these inundation models are based on elevation datasets which are

available in a different kind resolution. Some of these data are free. There are various of

inundation model studies for local, regional and global scale appropriate to mapping of

a coastal region based on digital elevation modeling and sea level rise modeling (Mcleod

et al. (2010); Titus and Richman (2001)). The inundation model which used high resolu-

tion as lidar data is suitable to determine the risks of SLR (Kulp and Strauss (2016); Titus

and Richman (2001)). On another hand, the lidar data doesn’t available to cover all the

earth because SRTM is the best digital elevation modeling for the low and flat topography

(Tara et al. (2010); Kulp and Strauss (2016)).

Many studies focused on error propagation modeling to enhance (SRTM in different

types of land cover (Tara et al. (2010); Shortridge (2006); Hofton et al. (2006); Bhang et al.

(2007); Rodriguez et al. (2006); Gamba et al. (2002); Kulp and Strauss (2016)). Our study

is creating correlated surfaces in near coastal riverine regions to achieve the main aims

of our study, which are: first, using propagation error model to reduced (SRTM) 1 Arc-

Second Global elevation error as global DEM available correlate the error with global land

use cover vegetation cover fraction (VCF) in near coastal riverine regions. Two, applying

this modeling (reduce (SRTM) 1 Arc-Second Global error) in the global near coastal river-

53

ine regions which are potentially vulnerable to inundation. This chapter concerns with

the accurate modeling of (SRTM) 1 Arc-Second Global in near coastal riverine regions.

The modeling in this chapter possesses excellent efficiency for Mississippi River delta re-

gion in southern Louisiana and it can apply in near coastal riverine regions. I address

several key questions for developing globally applicable error models for inundation as-

sessment:

1. How does SRTM error correlate with global canopy cover (Landsat 5)?

2. To what extent is SRTM error reduced using Landsat 5 and other global covariates?

3. How effective are geostatistical models at reproducing inundation model results

using high-accuracy USGS NED in this region?

4.2 Data and Methods

4.2.1 Study Area

The study focuses on the Louisiana coastal plain around Mississippi River from its junc-

tion with the Bayou Lafourche downriver to the Gulf of Mexico and covers 11,894 km2 as

shown in Figure (4.1)(a). This important region, which is mostly coastal wetlands, has a

distinctive ecology and landscape and is rich in resources (Steyer and Llewellyn (2000)).

The alluvial generally ranges from 15 to 100 km in width, which is supported by the find-

ings of the region along Mississippi. Mississippi River flows along a ridge formed by its

own natural deposits (known as a levee), from which the lands decline at an average fall

of 3 m/km (Zellmer and Klein (2007)).

54

(a)

(b)

(c)

(d)

Figure 4.1: (a) study area (Mississippi River Delta region in Louisiana state); (b) NED 1/3 arc-
second (approximately 10 meters); (c) VCF Landsat 5 TM Collection 1 products 30 m and (d) SRTM
1 Arc-Second Global coverage ( 30 meters).

4.2.2 Data

4.2.2.1 National Elevation Dataset (NED)

In this study, I used the National Elevation Dataset (NED) 1/3 arc-second (approximately

10 meters) DEM, which is the best publicly available high-resolution elevation DEM for

most of the US at this time as shown in Figure (4.1),(b). The 1/3 series is produced and

distributed by the U.S. Geological Survey (USGS) and is one of many seamless datasets.

These data are available through the Earth Rxplorer portal (https://earthexplorer.usgs.gov/),

55

(Gesch et al. (2002)). I use NED 1/3 arc-second data because it has been editing and to

remove structures and vegetation to produce a bare-surface DEM suitable for modeling

the water plane. I downloaded two one-degree tiles to cover the study area.These tiles

were mosaicked and then projected UTM zone 16N, WGS84 with a cell size of 10 m, and

finally resampled to 30 m resolution to match the SRTM 1 DEM.

4.2.2.2 SRTM 1 Arc-Second Global Coverage ( 30 meters)

The Shuttle Radar Topography Mission (SRTM), collected interferometric synthetic aper-

ture radar (IFSAR) data in February of 2000 covering 80% of the earth’s land surface. The

importance of this data stems from the vast areas of the world it covers, especially regions

with little or no free terrain data at medium to high resolution. The advantages of this

data for scientific and research applications for such regions are substantial. Accuracy

is one important issue with SRTM DEMs (Shortridge (2006)). 1 arc second SRTM was

downloaded from Earth Explorer: three tiles were required to cover the study site. I did

subsite and mosaic of these images to get MRDR SRTM, and then projected to UTM zone

16 N referenced to the WGS 1984 datum; final cell size was 30 meters as shown in Figure

(4.1),(d).

4.2.2.3 Vegetation Data

Since SRTM error is positively associated with vegetation cover (Shortridge (2006)), I used

vegetation cover fraction (VCF) as a secondary variable to correct the SRTM error. VCF

has been used in many global land process models as earth surface change and climate

change in regional and global studies (Barlage and Zeng (2003); Zhang et al. (2013); Jia-

paer et al. (2011); Baret et al. (2013)). I produce VCF Landsat 5 TM Collection 1 products

56

are available for download (Earth Explorer). I used four images (Path 21, 22,23 Row 39,40)

to cover the study area. All these images acquisition date: 19 / February / 02, and these

images are 30-m resolution, the same as the SRTM resolution. The observation of the

earth was conducted at the same time too as shown in Figure (4.1),(c). VCF mean the

percentage of green vegetation spatial distribution to the total statistical region (Jing et al.

(2011); Barlage and Zeng (2003); Zhang et al. (2013)). I created the normalized difference

vegetation index (NDVI) image first and I determined two values from this NDVI which

bare soil and NDVI value full canopy. The VCF is of two components: vegetation and

soil. I estimate VCF percentage by the following equation:

VCF = (NDVI − NDVIbaresoil)/(NDVI f ullcanopy − NDVIbaresoil)

(4.1)

where NDVI is normalized difference vegetation index, NDVIc is NDVI for a full

canopy pixel, and NDVIb is bare soil NDVI value pixel (Wang et al. (2017)) (Estimating

Fractional Vegetation Cover from Landsat 5 TM Collection 1 products reflectance data in

Mississippi river delta).

4.2.3 Methodology

SRTM error can be characterized by models that combine correlated environmental vari-

ables and spatial autocorrelation. In this study, I investigate the utility of vegetation index

from Landsat along with elevation and slope to account for SRTM error in MRDR. SRTM

error can be characterized by models that combine correlated environmental variables

and spatial autocorrelation. I use error model based on a set of 10K points, validated and

then used to adjust SRTM, and that gets applied to a bathtub model.

57

In the first step, 10,000 random point samples from reference NED are extracted. These

were used to develop measures of SRTM error across the region. Also, I apply regression

kriging to fitting an OLS regression model with covariates to SRTM error, and then mod-

eling the spatial autocorrelation of the residuals via simple kriging and combining those

into a single raster of SRTM error. The steps for SRTM error modeling and estimation of

inundation are summarized in Figure 4.2.

Figure 4.2: Error modeling and bathtub modeling for Mississippi River Delta Region.

4.2.3.1 Regression Kriging

I used regression-kriging models to generate estimates of SRTM error for every cell in

MRDR. Regression-kriging is the best linear interpolation calculating a universal model

of spatial difference, and the best linear predictive unbiased predictor of spatial data was

performing. The result of this model has high accursed simulations, and the topogra-

phy elevation values represent more details about spatial distribution. See (Leon et al.

58

(2014); Hengl et al. (2008); Christensen (2001)) for more details about the regression-

kriging model. Regression-kriging prediction is calculated as follows (Leon et al. (2014)):

ˆz(s0) = ˆm(s0) + ˆe(s0) =

p∑

k

ˆβk × Xk(s0) +

n∑

i=1

λi.e(si)

(4.2)

where ˆz(s0) is the predicted SRTM error, ˆm(s0) is the fitted linear regression of SRTM

error using VCF, SRTM elevation, and slope as independent variables, and ˆe(s0) is the
simple kriging interpolated residual at the prediction location s0. The ˆβk are the estimated

regression coefficients and Xk are explanatory variables. The λi are kriging weights for

the residuals, where e(si) is the residual at measurement location si, that is, the difference
between the elevation error and the fitted trend at si. The regression coefficients ˆβk are

estimated using generalized least squares (Leon et al. (2014)).

Spatial structure of the residuals from the regression model was fit using a combina-

tion of two Gaussian variogram models to fit both short range and long-range structure.

Parameters of the first variogram model were: partial sill = 0.1, range = 800, and nugget

= 0.04, while those for the second variogram model were partial sill = 0.12, range = 4000,

and nugget = 0. The regression model was validated by drawing a new set of 10,000

random locations and predicting SRTM error for those locations.

4.2.3.2

Inundation Modeling

I used inundation modeling which can be filled with hydrological connectivity or with-

out hydrological connectivity. The vertical and horizontal resolution of digital elevation

data used to estimate coastal vulnerability will impact estimates of flooding. Despite,

the high resolution of digital elevation models (DEM) used in environmental evaluation

59

can be important to the exposure of topographic characteristics and the measurement of

hydrological processes (Poulter and Halpin (2008)). I apply inundation modeling to de-

termine which area from the region will suffer from flooding due to SLR. The Function to

implement the bathtub inundation model is:



Floodedj =

2 if DEMj > f
1 if f ≥ DEMj ≥ 0
0 if DEMj ≤ 0

(4.3)

where Floodedj is a binary variable indicating the flooding status of cell j, DEMj is the

elevation of cell j, and f is the flood elevation above sea level.

In this experiment, the original SRTM DEM, the NED DEM, and an adjusted SRTM

DEM based on the error model were all used as inputs to an inundation model set to flood

all land less than 1m in absolute elevation above sea level. The geostatistical modeling of

the SRTM elevation error and the inundation model were calculated using the R statistical

package and the following libraries: gstat, raster, and grdevices (R Core Team (2017);

Bivand (2017) ; Robert J. Hijmans (2017)).

4.2.4 Results

Results will be linked to the three research questions the paper intends to address.

4.2.4.1 Parameter Estimation For SRTM 1 Error Regression Model

To address question one, I evaluated the relationship between Landsat-based VCF and

SRTM error at a random sample of 10,000 points across the study region while also ac-

60

counting for relationships between error and SRTM slope and raw elevation. Figure 4.3

shows a map of SRTM error across the MRDR along with scatterplots of relationships with

VCF, SRTM elevation itself, and SRTM slope. Table 4.1 summarizes the regression coef-

ficients: SRTM error model with (VCF, SRTM elevation, and SRTM slop) by variogram

modeling is positive regression coefficients. The estimate of SRTM error with SRTM el-

evation was 0.800835, and the estimate of SRTM error with VCF was 0.0054241, indicat-

ing that mean parameter estimation SRTM error is positive. The T value was positive

for SRTM elevation (159.976) and VCF (7.991). Figure 4.4 provides a range of model vi-

sualizations. Regression model performance took advantage of the strong relationship

between SRTM elevation and the error. Histogram of model predictions was around one

meter. A variogram of regression model residuals was fit. The spatial structure for first

variogram model of SRTM errors conferred (partial sill = 0.1, range = 800 and nugget =

0.04) and the second variogram model was (partial sill = 0.12, range = 4000 and nugget =

0).

Adjusted SRTM elevation predictions for the entire study region ranged between 20

m and 25 m. The regression model R-square is 0.8342, indicating the model effectively

predicts SRTM error in this region. The SRTM1 error model was used to predict error

across the entire MRDR region as shown in (4.3),(a) and visual investigation proposed

that error was not spread randomly in this region. However, the spatial characteristics of

SRTM1 error and geographical variability exhibit low correlations with the spatial error

distribution in Figure 4.3. Scatterplots of raw SRTM error with SRTM slope and VCF

exhibit low correlations (Figure (4.3),(b), (d)), but raw SRTM error and SRTM elevation

are highly positively correlated (Figure 4.3 (c). Visual analyses between SRTM1 and VCF

in Figures (4.3), (a) and Figure (4.1),(b), (d) show VCF segmented the results of accuracy

61

(a)

(c)

(b)

(d)

Figure 4.3: SRTM error: (a) Regression original SRTM error and raw SRTM error versus: (b) VCF;
(c) SRTM elevation; (d) SRTM slope.

assessment.

4.2.4.2 Regression and Variogram Results

The regression kriging model was validated on a second sample of points to determine

model capability for the composition of the SRTM errors and the variogram of regression

model residuals. The efficient model included the following key predictors: VCF, SRTM,

SRTM slope, and intercept (Figure 4.3(a), (b), and Table 4.1).

The result of the variogram modeling is the adjustment of SRTM elevation errors.

Figure (4.3)( e ), (f) shows two meters adjustment, and the modified errors distribution

was between 20 and 25 meters.

62

Estimate
Parameter
Intercept
-0.1381996
SRTM elevation 0.8008351
VCF
0.0054241
-0.0803025
SRTM Slope

Std. Error
0.0188362
0.0050060
0.0006788
0.0103973

t value
-7.337
159.976
7.991
-7.723

Pr(> |t|)
2.36e-13 ***
< 2e-16 ***
1.49e-15 ***
1.24e-14 **

Table 4.1: Regression coefficients of SRTM error model with NED and VCF.

4.2.4.3

Inundation Modeling

The regression kriging predicted SRTM error surface for MRDR was subtracted from the

original SRTM data to produce an adjusted SRTM elevation to develop a DEM with lower

error. This adjusted SRTM DEM, along with NED (as the reference DEM) and the original

SRTM, was used as input to the inundation modeling experiment. Figure 4.5 compares

the spatial extent of 1m sea level rise in MRDR using the original SRTM DEM before

error modeling and the adjusted SRTM DEM. Each map shows three classes: water, areas

which were at or below 0m of elevation; land, areas which are above 1m; flooded areas

between 0 and 1m which are inundated by this amount of flooding. Differences are stark:

only 8.8% of the land surface is flooded using the original SRTM, while nearly 52% is

flooded in the adjusted SRTM.

In contrast, Figure 4.6 shows the inundation of the reference NED DEM in comparison

to the adjusted SRTM. The DEMs produced similar inundation model results meaning

that the error model performed very well. Inundation area estimates are relatively similar

between the NED (42.1%) and the SRTM (51.8%).

4.2.5 Discussion

Topography is a key factor that influences the processes of coastal change, so up-to-date,

high–resolution, high–accuracy elevation data is required to model the coastal environ-

63

Figure 4.4: Error assessment for variogram model: (a) Regression model performance; (b) his-
togram of validation error of the regression model; (c) variogram and model of regression model
residuals over short distances; (d) variogram and model of regression model residuals over long
distances; (e) map of model-predicted SRTM elevation errors and (f) histograms of predicted
SRTM errors).

ment (Gesch (2009)). Lower elevation, marshes, irrigation canals and drainage networks

allow saltwater intrusion from the gulf to near-coastal riverine regions (River-dominated

deltas) (Good and Wilson (1995)). The high resolution of digital elevation models (DEM)

is used in environmental evaluation and can significantly influence the detail of topo-

graphic characteristics, creating varying results in inundation and the measurement of hy-

drological processes (Poulter and Halpin (2008)). Topographic detail increases in coastal

regions with the presence of canals, marshes and drainage networks (Lindsay (2006)),

64

Figure 4.5: Inundation models (predicting the spatial extent of a 1m sea level rise in the MRD
region: (a) initial SRTM and (b) adjusted SRTM DEM after the error modeling.

therefore, DEM is a valuable resource to provide horizontal resolution to measure hydro-

logical processes. (Poulter and Halpin (2008)).

Chabreck describes the major habitats in the coastal Louisiana region based on a veg-

etation survey of more than 5000 stations (Chabreck (1972)). Chabreck (1970, 1972) de-

lineated and mapped four vegetation zones (fresh, intermediate, brackish, and saline)

mainly based on Penfound and Hathaway (1938) descriptions of the major vegetation

types of the MRDR (Visser et al. (1998)). The topography elevation was used to deter-

mine the spatial distribution of these marshes (low elevation is open water) and Saline

(Saltwater) marsh and higher elevation is Intermediate marsh and brackish (Figure (4.6)),

( 1). Zonal statistics were used as the table to summarize the values of the NED 10m

within the zones of the marshes dataset (USGS, BRD, NWRC, LDWF), and I reported the

results to a table.

These analyses show that Mississippi River Delta Region has undergone a net change

in the land area of spatial and temporal from 1978 to 2016. This net change is illustrated

65

Figure 4.6: Inundation model predicting the spatial extent of a 1m sea level rise in the MRDR: (a)
NED 10m DEM and (b) adjusted SRTM DEM.

and mapped in five vegetation zones on NED (fresh (–20.57 %), intermediate (10.97 %),

brackish (–13.09%), open water (20.93 %) and saline (–26.19 %) and other which is inside

levees (–20.29 %)) Figure (4.6) , (2). However, The DEM data which has high accuracy as

NED DEM 10m is so important to explain the land change as near coastal riverine regions,

but unfortunate NED is not available just in the US. Because of that, I developed a glob-

ally applicable SRTM error model for inundation assessment. Also, I have presented the

analysis of the topographic vulnerability of the MRDR based on SRTM DEM 30m, which

covers most of the world, adjusted to account for high sea level rise. This adjustment

carries significant consequences of estimating the region flooding due to sea level rise.

4.2.6 Conclusion

In this study, I created an error assessment and developed a regression kriging error

model for SRTM1. Inundation modeling using SRTM and the regression kriging-adjusted

SRTM was employed to estimate the regional vulnerability to flooding due to SLR. The

66

Figure 4.7: Marsh classes in Mississippi river delta region: a. Spatial temporal change for vege-
tation classes. b. Marsh classes changing based on mean NED DEM.

error assessment found that SRTM error in the MRDR is correlated with Landsat-derived

canopy cover, SRTM slope and SRTM elevations. The regression model showed that these

correlations can be used to develop reliable, high accuracy predictions of SRTM error. Re-

gression kriging took further advantage of the spatial autocorrelation of the regression

residuals to closely reproduce the reference NED elevation surface. Key findings and

implications of the chapter include:

1. SLR inundation can be accurately modeled and forecast for coming decades on a

global scale with SRTM 30m with good geostatistical models and error propagation.

2. Using SRTM is helpful in inundation modeling (bathtub modeling), as it is a freely

available global dataset. It saves time and money when it is applied to other global

regions without high-resolution, high accuracy DEMs.

67

3. There are promising results from adjusted SRTM bathtub modeling that correlate

well with NED data. In contrast, raw SRTM greatly underestimated the inundation

threat.

4. The method developed to remove SRTM error works well in our study area and it

will be possible to apply the same model to other regions with similar geographical

features.

68

Chapter 5

Accounting for DEM Error in Sea Level

Rise Assessment for Shatt Al-Arab River

Region

5.1

Introduction

Sea level rise varies regionally around the world: in Persian (Arabian) Gulf the rate is
similar that in south Asia which is 2.2 ± 0.5 mm/year (Alothman et al. (2014); Hassan-

zadeh et al. (2007). As the Persian Gulf experiences high temperatures in the summer,

the mean sea level is lower in winter and higher in summer with an inter-annual range

of 26 cm (Hassanzadeh et al. (2007); Sultan et al. (1995a)). As a consequence, tidal wet-

lands and nearby low elevation areas are extremely sensitive to even slight changes in

altitude (Wang et al. (2007)) or sea level. Hydrology plays a key role in understanding the

ecology of these environments: predicting and managing changes in near coastal envi-

ronments requires an understanding of the linkages between estuarine wetland ecology

and hydrology. Also critical for these systems are the long-term drivers of climate change,

sea level rise, and human interferences of hydraulic modification of water flow (Hughes

et al. (1998)). In the Shatt al-Arab River Region (SARR), low-lying areas support impor-

69

tant human activities such as settlements and agriculture, the influence of SLR may have

substantial human impact in this region. The SARR has supported the largest cultivated

date palm forest in the world, which stretches from the riverbanks towards the desert

for distances varying from a few hundred meters to almost six kilometers. A fifth of the

world’s palm trees were cultivated in this region in the mid-1970s, amounting to 17-18

million date palms. By 2002, salinization wiped out more than 80 percent of the SARR

palm forest, a total of 14 million palms (Singh et al. (2005); Brandimarte et al. (2015)).

Environmental problems have been a major focus of several studies in this region, in par-

ticular land change due to an increased salinity in both the river water and the soil (Jabbar

and Zhou (2011); Al-Bahili et al. (2010); Al-Bahili et al. (2010); Essa (2012); Mahmood et al.

(2013).

Decreasing levels of freshwater in this river as a primary driver of many problems in

the region (Abdullah et al. (2015)). However, there are two main reasons for the lack of

fresh water. The first involves the concentration of Shatt Al-Arab salinization, which is

increasing from the south due to being so close to the source of salinization, which comes

from the Gulf. The northern part of Shatt Al-Arab river is less affected by salinization due

to the fact that it is moving away from its source (the Gulf). The second problem involves

the vegetation area around Shatt Al-Arab river, which is notably disappearing from the

south, yet there is still some of vegetation in the northern part of the river. SLR plays

a major role in the environment as well as the land change problems as it measures the

impact of the previous years on this region as well as estimates their impact on the future

of SARR.

In this study, I used a digital elevation model (DEM), the Shuttle Radar Topography

Mission (SRTM). This DEM was produced from data collected during a US Space Shuttle

70

mission in February of 2000 that covered 80% of the global land surface (Farr et al. (2007)).

It is the first global DEM with 30m resolution, but this DEM data has large vertical errors

in densely vegetated and urban areas (Shortridge (2006); Rabus et al. (2003); Shortridge

and Messina (2011)). To apply geospatial inundation which is widely used to model SLR

(e.g., Titus and Richman (2001); Kulp and Strauss (2016); Tara et al. (2010); Mcleod et al.

(2010)) while accounting for these errors, I must apply error propagation modeling.

I

use geostatistical simulations and the Monte Carlo method to understand the effect of

SRTM error on inundation likelihood. More specifically, kriging is used to develop linear

estimates of parameters for the SRTM error distribution for each cell in DEM, then sam-

ple from that distribution to generate error, and hence true elevation, realizations. I do

not have actual elevations for this region, so I will use parameters from an error model

constructed on the SRTM 1” data in the MRDR.The study will create correlated, statisti-

cal DEM error realizations in the study regions to achieve the main aims of my research,

which are:

1. Develop an error propagation model to enhance (SRTM) 1 Arc-Second Global ele-

vation model for the region.

2. Transfer parameters like vegetation cover and slope for error assessment of SRTM.

3. Apply this model to identify areas which are potentially vulnerable to inundation

in SARR.

4. The error model will be calibrated and validated using MRDR and applied in SARR.

5. Understand the impact of error on inundation models in low-lying coastal regions.

71

5.2 Data and Methods

5.2.1 Study Area

SARR is in southeastern Iraq and southwestern Iran and includes land on both sides of

Shatt al-Arab River from north of Basra City to the Persian Gulf as depicted in Figure 5.1

Its total area is 2,189 km 2 (Al-Saaidy (2008)). Land elevation along the river descends

gradually from five meters above sea level near Al Qurnah in the north to zero in the

south. The banks of the river generally have higher elevations than interior areas. The

Shatt al-Arab River flows through Iraq for about 80 km to just above the confluence with

the Karun River from the east; from that point to the Gulf the river forms the border

between Iran and Iraq. Cities and towns, which together are home to more than three

million persons as shown in Figure 5.1 interspersed with agricultural land and marshes,

occupy its banks. The date-palm groves along the river are naturally irrigated by tidal

action (Saad, 1978). Most studies treat the Shatt Al-Arab as a unique river because it has

a tide two times during the day. Tidal movement from the Gulf controls the hydrology of

the river, especially after the decline in freshwater coming from the Euphrates, Tigris, and

Karun rivers (Abdullah et al. (2015)) in the period evaluated in this study (1975 to 2017).

Tides enter the river from Persian Gulf and reach at least as far as Basra City (Cressey

(1958)). The region includes a portion of the Governate of Basra in southeast Iraq, and

of Abadan County, located in western Iran as shown in Figure 5.1 (a). This region is

rich in economically important agricultural resources, including crops, vegetables and

palm trees (Jabbar and Zhou (2011); Al-Bahili et al. (2010)). Elevation typically ranges

from 5 to 0 meters in the lowlands away from the river. The size of the total area is

2,189 km 2 as shown Figure 5.1(a) and (b). Various studies found that Shatt Al-Arab is

72

distinctive in that it has two tide sequences during the day such that the water level of

Shatt Al-Arab is directly affected by the water in the Gulf (Abdullah et al. (2015)). It is the

tidal movement that regulates the hydrology of the river, especially after the reduction in

freshwater approaching from Euphrates, Tigris and Karun rivers (Abdullah et al. (2015))

in the phase evaluated in this study (1975 to 2017). Tides are found to come into the river

from Persian Gulf and will spread at least as far as Basra city (Cressey (1958)).

Figure 5.1: (a) Landsat imagery for Shatt Al-Arab River study region in Iraq; (b) SRTM 1 Arc-
Second Global coverage ( 30 meters); (c) SRTM-derived slope; (d) Vegetation cover fraction (VCF)
derived from Landsat 5 TM Collection 1.

73

5.2.2 SRTM 1 Arc-Second Global Coverage ( 30 meters)

In February of 2000, interferometric synthetic aperture radar (IFSAR) data was collected

by the Shuttle Radar Topography Mission (SRTM) covering 80% of the earth’s land sur-

face (Rabus et al. (2003)). The data collected brings substantial significance with it, espe-

cially in regions with little or no free terrain data at medium to high resolution. SRTM

DEMs are not without issues, of which accuracy is particularly important here (Short-

ridge (2006)). SRTM one arc second data were downloaded from USGS Earth Explorer,

then clipped and mosaicked in ArcGIS to get a DEM for SARR. This DEM was then pro-

jected to UTM zone 15 N referenced to WGS 1984 datum; final cell size was 30 meters as

shown in Figure 5.1, (b).

5.2.3 Vegetation Data

In regional and global studies, vegetation cover fraction (VCF) has been used as a variable

in global land process models for earth surface change and climate change assessment

(Zhang et al. (2013); Barlage and Zeng (2003); Jing et al. (2011)). This study also required

the use of VCF as a variable since SRTM error is associated with vegetation cover (Short-

ridge (2006); Barlage and Zeng (2003); Zhang et al. (2013); Jiapaer et al. (2011); Baret et al.

(2013)). Since VCF Landsat 5 TM Collection 1 products are available for download from

USGS Earth Explorer, four images, which were acquired on February 19, 2002, were used

in this study (Path 165,166 and Row 39). Since the SRTM resolution was 30 m, the down-

loaded images were also projected and resampled to 30m resolution. All of the images

also had the same time observation of the earth as shown in Figure 5.1, (d). It is important

to note that the percentage of green vegetation spatial distribution to the total statistical

74

region is the definition of VCF (Jing et al. (2011); Barlage and Zeng (2003); Zhang et al.

(2013)). As part of this study, the normalized difference vegetation index (NDVI) im-

age was created first, and it was determined to have two values, which are bare soil and

NDVI value full canopy. The NDVI allowed for a conclusion that VCF is made up of two

components, vegetation and soil. The VCF percentage was estimated by the following

equation:

VCF = (NDVI − NDVIs)/(NDVIc − NDVIb)

(5.1)

where NDVI is normalized difference vegetation index, NDVIc is NDVI for a full

canopy pixel, and NDVIb is bare soil NDVI value pixel (Wang et al. (2017)).

5.2.4 Methods

5.2.4.1 Monte Carlo Error Propagation

I use Monte Carlo simulation which is one of the main approaches to error propagation

modeling as a technique to determine how spatially autocorrelated error affects applica-

tions using spatial data. The Monte Carlo simulation operation is repeated many times

to generate a sample for the input probability distribution (Leon et al. (2014); Hengl et al.

(2008)). SRTM error for SARR is expected to have a mean component, that is associated

with available covariates (SRTM elevation, VCF and SRTM slope) and a spatially autocor-

related, mean-zero residual component. I developed the error model (raw error ~ SRTM

elevation + VCF + SRTM slope) in a separate study for the Mississippi River Delta Region

(MRDR). I then did validation which is fitted error estimates versus actual SRTM error

and finally I made a variogram of these residuals. I use regression kriging to fit the model

75

and handle its spatially autocorrelated residuals (Hengl et al. (2003)).

ˆzRK(s0) =

p∑

k=0

ˆβk × qk(s0) +

n∑

i=1

wi(s0).e(si)

(5.2)

Where: ˆβk is estimated drift model coefficients, wi is the semi-variance function deter-

mine weights and e is the regression residuals

Predictions for the residual error surface in SARR must be unconditional, meaning

that residual realizations match spatial properties of MRDR model but don’t account for

ground truth, since I do not have reference elevation data for SARR. The error realizations

are used to enhance SARR SRTM 30 m DEM, and I apply SLR inundation model to detect

and characterize the flooding impact of SLR for the study region, as shown in Figure 5.2.

5.2.4.2 Unconditional Simulation with Convolution Filtering

This study has a large amount of data, too computationally intensive for standard simu-

lation methods. I employed process convolution (Oksanen and Sarjakoski (2005)) to de-

velop the residual simulations in an efficient manner. The convolution filter reproduces a

spatial Gaussian covariance function:

ρ(h) = exp[−3(

h2
r2 )]

(5.3)

where h is the correlation coefficient and r is the specified range of the DEM error.

The convolution filter method works as follows. I generate a vector of locations for each

SRTM pixel to simulate values z : (si), i = 1, ..., n in the following steps, itemized here and

as shown in Figure 5.3 (Oksanen and Sarjakoski (2005)):

76

Figure 5.2: Transferring the error propagation model and bathtub modeling from Mississippi
River Delta Region (A), to Shatt Al-Arab river region (B).

5.2.4.3

Inundation Modeling

Many previous studies on geospatial inundation modeling (bathtub modeling) used ei-

ther high resolution digital elevation data like lidar data or medium resolution SRTM 30

and used error modeling to remove an error from SRTM based on the truth data that has

high resolution (Yassein et al. (2016); W. van de Lageweg et al. (2017);Leon et al. (2014);

Kulp and Strauss (2016); Sande et al. (2012)). In SARR, I do not have reference elevation

data, so this study uses the error propagation model developed for the MRDR and applies

it to the SARR.In this study, a basic inundation model (without considering surface hydro-

logical connectivity) was coded by hand in the R statistical package using the following

77

Figure 5.3: The autocorrelated Gaussian covariance process and its associated convolution filter.
This filter is passed over spatially random Gaussian noise to produce rasters with the desired
spatial structure.

libraries: sp, gstat, raster, grdevices and gstat (Bivand (2017); R Core Team (2017);Robert

J. Hijmans (2017)). A sea level rise of 1 m, a value often cited in the literature as possible

by the end of the 21st century (IPCC (2013)), is projected, and inundation modeling is

applied to the exhibition area from the region which will be flooding in 1m of SLR. The

inundation model function is:

Floodedj =



2 if DEMj > f
1 if f ≥ DEMj ≥ 0
0 if DEMj ≤ 0

(5.4)

where Floodedj represents cell jin the DEM, the cell elevation value for j is represented

with DEMj, and f is the inundation elevation. The inundation model will be applied to

78

every DEM realization in SARR, resulting in dozens of output maps. The distribution of

inundated area across these maps is calculated and portrayed in a histogram. Finally, the

probability of inundation for each cell is computed and a map of inundation probability

for SARR is produced.

5.3 Results

5.3.1 Regression Kriging-Based Error Modeling

The error model was developed and validated using data from MRDR and transferred

to SARR. Validation in the MRDR was positive: the regression model shows high sig-

nificance and explanatory power, and the residuals were spatially autocorrelated. The

variogram model of SRTM errors was a Gaussian model with range = 800, partial sill

= 0.1 and nugget = 0.04, as shown in Figure 5.4(a) and Table 5.1. I transferred the er-

ror model to SARR to model SRTM error for this region and produce realizations of the

potential DEM surface, accounting for uncertainty in SRTM associated with vegetation,

slope, elevation values and spatially autocorrelated residuals. Figure 5.4(b) illustrates the

predicted SRTM error surface from the regression model for MRDR, while Figure 5.4(d)

shows the predicted error surface for SARR. Histograms of both are shown in Figure 5.4(c)

and (e), illustrating that these are centered just above zero with a bell-shaped distribution

with long tails.

79

(a)

(b)

(c)

(e)

(d)

(f)

Figure 5.4: Error assessment for variogram model: (a) MRDR error model residual variogram and
model (line); (b) SARR variogram model (line) and variogram of simulation output (circles); (c)
MRDR map of model-predicted SRTM elevation errors; (d) MRDR histogram of predicted SRTM
errors; (e) SARR map of model-predicted SRTM elevation errors; (f) SARR histogram of predicted
SRTM errors.

5.3.2 Realizations

To better understand the effect of DEM data uncertainty on inundation modeling, DEM

realizations were constructed from the SRTM data and the error model as shown below:

DEMreal = SRTM − Regression − predictederror − Residualrealization

(5.5)

The regression kriging predicted SRTM error for SARR was subtracted from the orig-

80

Parameter
Intercept
SRTM Eleva-
tion
VCF
SRTM Slope

Estimate
-0.1381996
0.8008351

0.0054241
-0.0803025

Std. Error
0.0188362
0.0050060

0.0006788
0.0103973

t value
-7.337
159.976

7.991
-7.723

Pr(> |t|)
2.36e-13 ***
< 2e-16 ***

1.49e-15 ***
1.24e-14 **

Table 5.1: Regression coefficients for regression model variables

inal SRTM data. Unconditional error model residual realizations were generated and

subtracted as well. QQ plots were used to check the normality assumptions of the error

model. The QQ plot for an initial Gaussian random noise realization as shown in Figure

5.5(a) and (b) is straight and, therefore, matches expectations, while that for the filtered

realization is not as normal, but its shorter tails are acceptable. I generated 50 realizations

of error model residual maps and used the equation above to produce DEM realizations

for SARR; one of these is depicted in Figure 5.5(c).

5.3.3 Inundation Modeling

The spatial extent of inundation for a 1m sea level rise in SARR region was calculated

using the original SRTM DEM as mapped in Figure 5.6, (c),(e). In this comparison, each

map shows three classes: water, where areas were at or below 0m of elevation; land,

where areas are above 1m; flooded, where areas are between 0 and 1m due to being

inundated by this amount of flooding. While only 28% of the land surface is indicated

as flooded using the raw SRTM, DEM realizations show a substantial difference with

nearly 50% of the land area being flooded for all realizations. Figure 5.6,(d),(f) depicts a

histogram of the proportion of land area that remains dry for these realizations as well as

that for the raw SRTM.

In this study, I used the inaccurate but globally available SRTM as the base elevation

81

Figure 5.5: SRTM realizations: (a) random Gaussian noise for the SARR and spatially autocor-
related noise following convolution filtering; (b) QQ plots of random and convolution-filtered
Gaussian noise; (c) DEM realization number 10; (d) bathtub modeling SRTM realization number
10.

data for SARR and transfer a SRTM regression kriging error model from the MRDR. This

model could account for SRTM error due to vegetation and topographic factors while also

characterizing the spatial autocorrelation of the regression residuals. I applied a stan-

dard geospatial bathtub model to multiple realizations of this elevation model to assess

and visualize the impact of one meter of SLR in SARR. The key principle was to em-

ploy regression kriging modeling with unconditional simulation of the residual process

via convolution filtering to create propagate the impact of SRTM error to the geospatial

82

Figure 5.6: (a) SARR map of regression-adjusted SRTM elevation; (b) histograms of adjusted
SRTM ; (c) SARR probability map for land remaining dry following bathtub modeling of 1m in-
undation (d) histogram of flooding probability; (e) SARR bathtub model using raw SRTM data; (f)
Proportion of SARR predicted to remain dry following 1m inundation: raw DEM is dashed line,
histogram are areas across 50 realizations.

inundation model in SARR. To ensure robust results, 50 realizations of the elevation sur-

face were generated. This study is the first one to transfer an error propagation model

for bathtub modeling between different places in the world and to evaluate the effects of

SRTM error on SLR inundation of near-coastal riverine regions. I think that the use of the

MRDR error model in SARR is justified for several reasons:

1. SRTM 30 is a global DEM product that covers both regions. I assume that, because

the sensors that produced it are the same, its error properties in both areas are simi-

83

lar.

2. The topography of both regions is similar. Both regions are low elevation between

0 to 35m. Elevation is relatively high around the rivers and lower away from the

rivers. The northwestern parts of both regions are higher than the southeastern,

which both end in gulfs.

3. I can get the parameters to explain SRTM error in both regions like vegetation cover

and the slope from SRTM itself and from Landsat imagery.

4. I have surface reference elevation data for MRDR, but I don’t have any reference

elevation data for SARR.

Geostatistical simulation is computationally expensive, especially over large regions.

In this study, a Gaussian convolution filter was applied to random noise raster, which

rapidly generated spatially autocorrelated DEM error realizations. The results showed

that the methodology improved simulation efficiency while guaranteeing a relative nor-

mality of residuals in each simulation. The comparison of geospatial inundation model-

ing for raw SRTM 30 m and adjusted SRTM 30 m, as shown in Figure 5.6,(f) for 1m SLR,

and the impact of SRTM error on flooding estimation is stark. The raw SRTM was flooded

around 28%, but adjusted SRTM will be more than 45% flooded. This finding has poten-

tial implications for other studies that use SRTM to estimate SLR. were just one study

(Kulp and Strauss (2016)). This study used SRTM to assess the performance of NOAA’s

Global Land 1-km Base Elevation Project (GLOBE) in the US, but this study didn’t reach

any definitive conclusion because of the spatial heterogeneity in Shuttle Radar Topogra-

phy Mission (SRTM) 30 m modeling quality which used in this study4(Kulp and Strauss

(2016)).

84

5.4 Limitations

A major limitation for conducting SLR modeling in SARR is the data. For comparison,

MRDR has high-resolution, high accuracy 10 m DEMs derived from lidar, while tide sta-

tion gauges in the MRDR can give regular sea level elevations in this region at high tem-

poral frequency over many years (NOAA (2018)). In contrast, in SARR there are several

studies concerned with the intrusion of seawater in the region and how the river became

salty because of a lack of fresh water from upstream water, local demand, and seawater

intrusion, but there is no reliable, long term gauge or station to provide the data I can

use to measure SLR, or high-quality, freely available DEM. This study was the first about

SLR impact in SARR. In SARR I have just SRTM 30m, a global data product with sub-

stantial error. After developing the regression kriging error propagation for the MRDR,

I duplicate the modeling and fitting the model to the mean and handles spatially auto-

correlated residual. The residual error surface is unconditional and residual realizations

match properties of the model but do not have local reference data for validation. I must

assume that error properties truly are similar between the two regions; if this is not the

case, then the effects of error on inundation model are misspecified. Second, the result-

ing surface in the regression kriging, like any interpolation procedure, does not represent

slope and aspect. In both regions, the surface is smoother than depicted, as slopes and

aspect are poorly represented.

Another limitation of the study is the methods described in this study can only be

applied in near-coastal riverine regions. Therefore, any differences between enforcement

methods at the propagation regression kriging modeling can only be said to be transfer-

able to other areas with similar geographical characteristics, especially topography and

85

vegetation cover. There are multiple SLR models available for the data-rich MRDR, in-

cluding more formal hydrological modeling, but this study is limited to bathtub modeling

in order provide a fair comparison between MRDR and SARR.

5.5 Conclusion

The study results presented in this paper furnished valuable information on the error

propagation modeling of SRTM 1 to characterize sea level rise in SARR. It is the first to

apply error propagation modeling in this context. Model output included a probability of

inundation for each location in the SARR, and a comparison of total inundated area. The

approach was effective for determining and mapping the land which would be inundated

in the region from a 1m rise in sea level based on bathtub modeling. Most critically,

in SARR, error propagation modeling found much larger areas were at risk of flooding

than would be estimated with raw SRTM. SRTM, with its global coverage, can be used

to model SLR if error propagation modeling is applied properly. The approach will give

reasonable flooding risk for regions facing SLR. This research provides an appropriate

example of the successful application of transferring a geostatistical error model for error

propagation modeling to assess SLR inundation on near-coastal riverine regions.

86

Chapter 6

Conclusion

The threat of global warming and changing sea level rise has led to many studies on its

impact in some areas, like the MRDR in southern Louisiana, USA. However, it is also

necessary to assess the impact in other coastal areas like the SARR in southern Iraq. Con-

sequences of sea level rise in the coastal regions investigated in this dissertation are com-

plex, with a range of indicators and impacts. The most important indicators of SLR in both

regions are loss of land within the MRDR, declining vegetation cover within both regions,

reduction in land-based human activity such as agriculture and settlement within both

regions, ecosystem change around both regions, increase in groundwater salinity around

both regions, and increase in water salinity in both rivers. In this research, I sought to

explore the potential impacts of SLR on the SARR in southern Iraq and conduct a com-

parison with the MRDR in southern Louisiana, USA. To accurately model and predict

SLR impact in the near-coastal riverine regions of the world, high quality, high-resolution

elevation data is needed, especially for geospatial inundation modeling (bathtub model-

ing). Unfortunately, there is no freely available high-resolution data for all of the world

and therefore enhancing SRTM 1 Arc Second by developing error propagation models is

important for those regions.

Previous studies of geospatial inundation modeling (bathtub modeling) for SLR using

high-resolution data such as LiDAR were applied for limited geographical areas around

87

coastal cities which are vulnerable to increasing flood risks, including the Greater Lon-

don region (United Kingdom), the Lagos State area (Nigeria) and Brighton, the northern-

most suburb of Brisbane City (Australia) (Yunus et al. (2016); Leon et al. (2014); Sande

et al. (2012)). In contrast there is a study which used SRTM to assess the performance of

NOAAs Global Land 1-km Base Elevation Project (GLOBE) in the US, but this study does

not reach any definitive conclusion because of the spatial heterogeneity in SRTM quality

for data used in this study (Kulp and Strauss (2016)).

In this research, I clarify the connections between the environmental degradation in

the SARR and SLR. These connections are complex, and the range of potential drivers

present challenges for understanding the causes, consequences, and trajectories of this

marginal region. A coupled human and natural systems framework was established in

an effort to better comprehend the factors influencing the previous five decades of envi-

ronmental degradation in the SARR. I used decadal Landsat imagery across that region

to assess the geographic distribution of vegetation, how it is associated with the regional

environment, and how it has been changing. Additionally, an array of factors was iden-

tified as contributing to the degradation of vegetation in this dynamic, peripheral area.

Global climate change plays a critical part in directly and indirectly driving the most im-

portant factors and presents the most serious problem now facing the SARR. The scope

of this problem is significant due to the fact that climate change is a global problem as

opposed to a regional problem, and is not directly addressable by regional policy. In the

second case, I created error propagation modeling to assess SRTM 1 Arc Second error in

the MRD region. Following this assessment, I applied inundation modeling to determine

the flooding area due to SLR. The modeling developed for the MRD region appears to

have a viable application in similar near coastal riverine regions around the world.

88

In addition, I transferred error propagation modeling from the MRDR to the SARR, in

an effort to understand the impact of SLR in this region by applying inundation modeling.

The error propagation model is comprised of three phases. First, several parameters from

the MRDR, which is used to produce an expected DEM error surface for the SARR. The

result for the SARR is obtained by subtracting the surface from the SRTM DEM. Second,

spatially autocorrelated, mean-zero rasters with a Gaussian covariance model capture

residuals from the error model, which are replicated for the SARR. My goal here was to

better characterize the extent of flooding in the SARR associated with 1 m of SLR using

an inundation modeling (bathtub model) by enhancing the SRTM 1 Arc Second DEMs by

duplicating with an SRTM error model.

6.1 Methods

6.1.1 Coupled human and natural systems (CHANS)

CHANS is a popular, recent approach for investigators to understand responses, nonlin-

earities, thresholds, and legacy effects across many spatial, temporal, and organizations

scales. The CHANS framework has been enormously valuable as it permits a nontradi-

tional interdisciplinary approach. I used a CHANS framework to measure SLR indica-

tors in SARR and the associations of SLR with the human system (war, government land

policy, upstream hydroengineering, pollution, and degradation) and natural system (me-

teorological factors associated with drought and groundwater salinity). In this study, I

utilized numerous measures of ecological, economic and social systems following a re-

view of the CHANS concept.

89

6.1.2 The inundation model (bathtub modeling) and error propagation

DEM error was in part removed and in part accounted for by utilizing the error propa-

gation model, which fed into an inundation model that allowed for an assessment of the

SLR influences in the SARR. To understand the impact of the SLR, I have to apply the in-

undation model in this complex region, I conducted an error assessment and developed

statistical error modeling for SRTM for the MRDR and implemented this model to the

SARR in an effort to improve the quality of elevation data. In this study, the vegetation

index derived from Landsat combined with elevation and slope are used to account for

SRTM error in the MRDR. Based on a set of 10 thousand points in the MRDR, an error

model was developed, which then was authenticated and used to adjust the SRTM, and

then was used to generate an enhanced DEM for an inundation model. The regression

kriging-based SRTM error model effectively captures error through correlated environ-

mental variables and residual spatial autocorrelation collectively.

Next, I developed error propagation models to enhance SRTM 1 Arc Second for mod-

eling SLR in the SARR by applying the model from chapter four, developed for the MRD

region, to the SARR. I employed process convolution to develop the residual simulations

in an efficient manner. The convolution filter reproduces a spatial Gaussian covariance

function matching that observed in the MRDR. The MRDR predictions of the SRTM ele-

vation shows the error between 20 m and 25m, and the histogram of model predictions

was approximately one meter. In the second prediction, SRTM elevation SARR depicts

the error between 51 m and -50 m, and the histogram of model predictions was again

approximately one meter. The adjusted SRTM elevation errors were on two meters, and

the modified errors distribution ranged between 15 m and -11 meters. Fifty simulations

90

of 2-D autocorrelated Gaussian noise process were produced, which was used as input

to the inundation modeling experiment. In SARR I dont have reference digital elevation

data, so this study is transferred the error propagation modeling from the MRDR, which

has reference NED data digital elevation, to the SARR.

6.2 Dissertation Objectives

This dissertation has four objectives each of which uses many geospatial methods. The

dissertation objectives are:

Objective one.

Compare and contrast MRDR, SARR, and identify whether these regions have envi-

ronmental degradation because of SLR, based on the theoretical framework and findings

from previous studies. To achieve this objective, the following steps are used:

1. Characterize the geomorphology and climate in both regions and show how that is

correlated with SLR in both regions.

2. Identify human activity and economic importance in both regions and also identify

how the human system accelerates SLR risk in both regions.

3. Assess the different ways that SLR may affect both regions through rising salinity

in groundwater, pressure from subsidence erosion and regional flooding.

With respect to the two chapter, I found that both regions are similar in important geo-

morphological, environmental, economic and climatic characteristics. For example, both

are geologically recent creations from the deposits of several of the worlds greatest rivers

(Tigris, Euphrates and the Mississippi Rivers) flowing into gulfs (the Persian Gulf and

91

the Gulf of Mexico). Both regions are important regional economic centers for agricul-

tural production, oil production, fishing, and marine transport. Both rivers have natural

deltas that have formed where the mouth of each river meets its gulf. Also, land areas

on either side of these rivers are highest along the edges of the channels and decrease

in elevation farther from the river. Delta development for both rivers is an integrated

and long-term process that has given both regions a unique mix of ecosystems, habitats

and landforms. In this chapter I show that both regions face increasing pressure from

subsidence, erosion, sea-level rise and rising salinity in groundwater, leading to coastal

retreat in these regions because of the climate change and human activities there. SLR

affects these regions in somewhat different through flooding, pressure from subsidence

and erosion, and rising salinity in groundwater.

Objective two. Use a CHANS perspective to characterize and delineate a clear conceptual

model for factors impacting SARR and identify how human and natural systems are asso-

ciated with SLR risk in this region. For achieving this objective, these steps are followed:

1. Identify primary human and environmental drivers in SARR system.

2. Characterize NDVI as a proxy for vegetation quality and quantity and measure it

over a five decade period.

3. Identify spatial and temporal changes in NDVI throughout the SARR.

4. Link observed changes to dominant drivers of change in the SARR.

Chapter threes findings in this objective are: this chapter supports the importance of

the CHANS viewpoint because it recognizes not only the difficulties of this region but

also the solutions needed. First, the study outlines a distinct view of the SAR region in an

92

effort to identify the facts needed so that CHANS can be applied for a deeper understand-

ing of the problems and solutions needed. Second, my research identified the significance

of the SLR, which a primary natural systems driver for this region, and how the impacts

of SLR have accelerated over time. This study supported the importance of CHANS by

allowing for confirmation and understanding that the Shatt al-Arab River region is suffer-

ing. Environmental degradation occurred in the SARR from 1975 to 2017, and this study

concentrates on both the human and natural complications while determining which one

is accountable for the environmental degradation.

Objective three. Enhance SRTM Global DEM to better model SLR and characterize land

change in MRDR. For this objective to be accomplished, these steps are followed:

1. Determine how SRTM error correlates with canopy cover (derived from Landsat 7).

2. Identify the extent to which SRTM error is reduced using canopy cover and other

globally available covariates.

3. Assess the effectiveness of geostatistical models in reproducing inundation model

results using high-accuracy USGS NED in this region.

Chapter fours followed an extensive process to address this objective: The first phase

was to extract the 10,000 random point sample from reference NED elevations, SRTM, and

NDVI, and then use the outcome to develop a regression model for SRTM error across the

region. The residuals of spatial structure from the regression model were spatially mod-

eled to fit both short-range and long-range structure of two Gaussian variogram models.

The first variogram model parameters were: partial sill = 0.1, nugget = 0.04 and range

= 800, whereas partial sill = 0.12, nugget = 0 and range = 4000, were parameters for the

93

second variogram model. The regression model was validated using 10,000 other ran-

dom locations. I applied the inundation model of 1m SLR for the adjusted SRTM DEM

and the original SRTM DEM before error modeling. The output of the modeling maps

shows three classes: land areas which are higher than 1m; water areas which were at or

under 0m of elevation; and flooded areas between 1m and 0 which are flooded by this

quantity of inundation. The difference between models is unambiguous: around 52% in

the adjusted SRTM will be flooded, while in the original SRTM 8.8% will be flooded.

Objective four. Develop error propagation models to enhance SRTM 30 for modeling sea

level rise and characterizing the land change in SARR. This is achieved via the following

steps:

1. Apply the error model from objective three, developed for MRDR, to SARR.

2. Determine and map land flooding due to SLR in SARR and identify the land which

will become inundated based on the bathtub model.

3. Apply a valid geostatistical model to account for SRTM error and asses uncertainty

in SLR inundation model.

Chapter fives findings on this objective are as follows. After unconditional simulation

with convolution filtering for SRTM, I applied the inundation model with SLR set to 1 m

to each realization. This produced fifty output flooding maps, enabling an understanding

of the likelihood of flooding for each cell in the region. The output maps show the areas

below 1m and therefore flooded by this volume of water. Around 45% of the regions land

will be flooded for all realizations. For the raw SRTM, 27% of the land surface is shown

as flooded. SRTM error seems to give conservatively biased estimates of the flooded area.

94

6.3 Future Research

Modeling SLR impact in near-coastal riverine regions is essential for these sensitive ar-

eas, so focus should be on this problem. By assessing the utility of geospatial methods

for monitoring and modeling these impacts through the integration of land changes that

appear through DEM and satellite imagery analysis for different periods of time with the

values of water salinity, SLR rate, and land subsidence rate in the river regions, analysts

can understand what the effects of these changes may be. For my future research, the

error models should be tested in areas with similar geographical characteristics and data

like that in the Mississippi River Delta region. That will help determine if the error mod-

eling can be used other sensitive coastal regions in the world Another important direction

is to use relationships between SRTM DEMs 1 Arc Second and classified satellite imagery

of land change on the SARR over recent decades, e.g., 1975, 1985, 1995, 2005, and 2018, to

predict the impact of SLR on land use over the next several decades. The coming decades

will see important changes in many understudied near coastal areas like the SARR, and

the use of spatial data, while accounting for error, will play an important role in helping

to understand and manage those changes.

95

BIBLIOGRAPHY

96

Bibliography

A. D. Abdullah, I. Masih, P. van der Zaag, U. F. Karim, I. Popescu, and Q. A. Suhail.
Shatt al arab river system under escalating pressure: a preliminary exploration of the
issues and options for mitigation. International Journal of River Basin Management, 13(2):
215–227, 2015. doi: 10.1080/15715124.2015.1007870. URL https://doi.org/10.1080/
15715124.2015.1007870.

S. Al-Asadi. The future of freshwater in shatt al- arab river (southern iraq).

Journal of
Geography and Geology, 9(2):24, 2017. ISSN 1916-9787. doi: 10.5539/jgg.v9n2p24. URL
http://ccsenet.org/journal/index.php/jgg/article/view/68135.

N. Al-Bahili, J. Al-Maliki, and A. Al-Rubaye. The effect of vegetation on the stream bank
erotion of shatt al-arab river, south iraq. Marsh Bulletin, 5(1):1–13, 2010. URL http:
//pubs.er.usgs.gov/publication/pp1815.

I. Al Isawi. The geographic potentialities in shatt al-arab district and their suitability for
planting the suggested field produce. Dissertation, University of Basrah, College of Arts,
2011.

H. Al-Mahmood. The monthly variations of discharge and effect that on a total dissolved
suspended and salinity in shatt al-arab river (south of iraq). Iraqi Journal of Science, 50:
355–368, 2009. URL http://repository.uobaghdad.edu.iq/ArticlePrint.aspx?ID=
2170.

Z. Al-Saaidy. The potentials and the agricultural production in basrah during (1990-2006)
and it’s developing procedures in future. Graduate Thesis and Dissertations, The Uni-
versity of Basrah, College of Administration and Economics., 2008.

A. Alothman, M. Bos, R. Fernandes, and M. Ayhan. Sea level rise in the north-western
part of the arabian gulf.
ISSN 0264-3707.
doi: https://doi.org/10.1016/j.jog.2014.09.002. URL http://www.sciencedirect.com/
science/article/pii/S0264370714001264.

Journal of Geodynamics, 81:105 – 110, 2014.

M. Antonellini, P. Mollema, B. Giambastiani, K. Bishop, L. Caruso, A. Minchio, L. Pel-
legrini, M. Sabia, E. Ulazzi, and G. Gabbianelli. Salt water intrusion in the coastal
aquifer of the southern po plain, italy. Hydrogeology Journal, 16(8):1541, Jun 2008.
ISSN 1435-0157. doi: 10.1007/s10040-008-0319-9. URL https://doi.org/10.1007/
s10040-008-0319-9.

97

F. Baret, M. Weiss, R. Lacaze, F. Camacho, H. Makhmara, P. Pacholcyzk, and B. Smets.
Geov1: Lai and fapar essential climate variables and fcover global time series cap-
italizing over existing products. part1: Principles of development and production.
Remote Sensing of Environment, 137:299 – 309, 2013.
ISSN 0034-4257. doi: https:
//doi.org/10.1016/j.rse.2012.12.027. URL http://www.sciencedirect.com/science/
article/pii/S0034425713000813.

M. Barlage and X. Zeng. The impact of fractional vegetation cover and leaf area index on
warm season precipitation variability in global ensemble simulations. Arizona Journal
of Hydrometeorology, 01 2003.

P. M. Barlow. Ground water in freshwater-saltwater environments of the atlantic coast.
Tech. Report, U.S. Department of the Interior, 2003. URL http://pubs.er.usgs.gov/
publication/cir1262. Report.

J. A. Barras. Land area changes in coastal louisiana after the 2005 hurricanes: A series of
three maps. U.S. Department of the Interior — U.S. Geological Survey, 2006. ISSN 2006-
1274. URL http://pubs.er.usgs.gov/publication/ofr20061274. Report.

J. A. Barras, J. C. Bernier, and R. A. Morton. Land area change in coastal louisiana: A mul-
tidecadal perspective (from 1956 to 2006). Tech. Report, U.S. Department of the Interior,
2008. URL http://pubs.er.usgs.gov/publication/sim3019. Report.

P. Beaumont. Agricultural and environmental changes in the upper euphrates catch-
ment of turkey and syria and their political and economic implications. Applied
Geography, 16(2):137 – 157, 1996.
ISSN 0143-6228. doi: https://doi.org/10.1016/
0143-6228(95)00033-X. URL http://www.sciencedirect.com/science/article/pii/
014362289500033X.

K. J. Bhang, F. W. Schwartz, and A. Braun. Verification of the Vertical Error in C-Band
SRTM DEM Using ICESat and Landsat-7, Otter Tail County, MN. IEEE Transactions on
Geoscience and Remote Sensing, 45:36–44, Jan. 2007. doi: 10.1109/TGRS.2006.885401.

R. Bivand. spgwr: Geographically weighted regression. CRAN Task View: Analysis of Spa-
tial Data, 01 2017. URL https://cran.r-project.org/web/packages/spgwr/index.
html.

L. Bizikova, D. Roy, D. Swanson, H. D. Venema, and M. McCandless. The water energy
food security nexus : Towards a practical planning and decision-support framework for
landscape investment and risk management. In The International Institute for Sustainable
Development (IISD), 2013.

M. D. Blum and H. H. Roberts. Drowning of the mississippi delta due to insufficient
sediment supply and global sea-level?rise. Nature Geoscience, 2:488 EP –, Jun 2009. URL
http://dx.doi.org/10.1038/ngeo553.

M. D. Blum and H. H. Roberts.

The mississippi delta region: Past, present,
Annual Review of Earth and Planetary Sciences, 40(1):655–683,

and future.

98

2012. doi: 10.1146/annurev-earth-042711-105248. URL https://doi.org/10.1146/
annurev-earth-042711-105248.

A. G. Bobba. Numerical modelling of salt-water intrusion due to human activities and
sea-level change in the godavari delta, india. Hydrological Sciences Journal, 47(sup1):
S67–S80, 2002. doi: 10.1080/02626660209493023. URL https://doi.org/10.1080/
02626660209493023.

L. Brandimarte, I. Popescu, and N. K. Neamah. Analysis of fresh-saline water interface
at the shatt al-arab estuary. International Journal of River Basin Management, 13(1):17–25,
2015. doi: 10.1080/15715124.2014.945092. URL https://doi.org/10.1080/15715124.
2014.945092.

T. Buday, I. Kassab, and S. Jassim. The Regional Geology of Iraq: Tectonism, magmatism
and metamorphism. The Regional Geology of Iraq. State Organization for Minerals,
Directorate General for Geological Survey and Mineral Investigations, 1987. URL
https://books.google.com/books?id=2VxPAQAAIAAJ.

V. R. Burkett, D. B. Zilkoski, and D. Hart. Sea-level rise and subsidence: Implications
for flooding in new orleans, louisiana. U.S. Geological Survey Subsidence Interest Group
Conference: proceedings of the Technical Meeting, Galveston, Texas, November 27-29, 2001, 06
2003. URL http://pubs.er.usgs.gov/publication/2000794.

C. Cabanes, A. Cazenave, and C. Le Provost. Sea level rise during past 40 years de-
termined from satellite and in situ observations. Science, 294(5543):840–842, 2001.
ISSN 0036-8075. doi: 10.1126/science.1063556. URL http://science.sciencemag.org/
content/294/5543/840.

C. H. Chabreck. Vegetation, water and soil characteristics of the louisiana coastal region.
LSU Agricultural Experiment Station Reports, 664, 1972. ISSN 0036-8075. URL https:
//digitalcommons.lsu.edu/agexp/147/.

R. Christensen. Advanced linear modeling. multivariate, time series, and spatial data;
nonparametric regression and response surface maximization. 2nd ed. Springer Science
and Business Media, 01 2001.

P. U. Clark, J. D. Shakun, S. A. Marcott, A. C. Mix, M. Eby, S. Kulp, A. Levermann, G. A.
Milne, P. L. Pfister, B. D. Santer, D. P. Schrag, S. Solomon, T. F. Stocker, B. H. Strauss,
A. J. Weaver, R. Winkelmann, D. Archer, E. Bard, A. Goldner, K. Lambeck, R. T. Pier-
rehumbert, and G.-K. Plattner. Consequences of twenty-first-century policy for multi-
millennial climate and sea-level change. Nature Climate Change, 6:360 EP –, Feb 2016.
URL http://dx.doi.org/10.1038/nclimate2923. Perspective.

B. R. Couvillion, J. A. Barras, G. D. Steyer, W. Sleavin, M. Fischer, H. Beck, N. Trahan,
B. Griffin, and D. Heckman. ”land area change in coastal louisiana from 1932 to 2010.
Tech. Report, U.S. Department of the Interior, 2011. URL https://pubs.usgs.gov/sim/
3164. Report.

99

G. B. Cressey. The shatt al-arab basin. Middle East Journal, 12(4):448–460, 1958.

00263141, 19403461. URL http://www.jstor.org/stable/4323054.

ISSN

S. Dasgupta, B. Laplante, C. Meisner, D. Wheeler, and J. Yan. The impact of sea level rise
on developing countries: a comparative analysis. Climatic Change, 93(3):379–388, Apr
2009.
ISSN 1573-1480. doi: 10.1007/s10584-008-9499-5. URL https://doi.org/10.
1007/s10584-008-9499-5.

M. El Raey. Impact of sea level rise on the arab region. University of Alexandria and Regional
Center for Disaster Risk Reduction, Arab Academy of Science, Technology and Maritime Trans-
port, 2010. URL http://www.arabclimateinitiative.org/Countries/egypt/ElRaey_
Impact_of_Sea_Level_Rise_on_the_Arab_Region.pdf.

J. Emel, J. Plisinski, and J. Rogan. Monitoring geomorphic and hydrologic change at mine
sites using satellite imagery: The geita gold mine in tanzania. Applied Geography, 54:243
– 249, 2014. ISSN 0143-6228. doi: https://doi.org/10.1016/j.apgeog.2014.07.009. URL
http://www.sciencedirect.com/science/article/pii/S0143622814001659.

A. M. Essa. The use of diatom indices for the assessment of shatt al-arab river water
quality. Journal of Basrah Researches ((Sciences)), 38:114–124, 2012. URL https://bit.
ly/2NIdnD1.

T. G. Farr, P. A. Rosen, C. Edward, C. Robert, D. Riley, H. Scott, K. Michael, P. Mimi,
R. Ernesto, R. Ladislav, S. David, S. Scott, S. Joanne, U. Jeffrey, W. Marian, O. Michael,
B. Douglas, and A. Douglas. The shuttle radar topography mission. Reviews of
Geophysics, 45(2), May 2007. doi: 10.1029/2005RG000183. URL https://agupubs.
onlinelibrary.wiley.com/doi/abs/10.1029/2005RG000183.

C. L. Foote, K. Crane, W. Block, and S. Gray. Economic policy and prospects in iraq (may
4, 2004). FRB of Boston Public Policy Discussion Paper No. 04-1, pages 47–70, 2004. URL
http://dx.doi.org/10.2139/ssrn.887920.

J. Fox, R. R. Rindfuss, S. J. Walsh, and V. Mishra. People and the environment: Ap-
proaches for linking household and community surveys to remote sensing and gis.
Springer Science and Business Media, 2003. URL https://www.springer.com/us/book/
9781402073229.

D. O. Fuller. Trends in ndvi time series and their relation to rangeland and crop
International Journal of Remote Sensing, 19(10):
production in senegal, 1987-1993.
2013–2018, 1998. doi: 10.1080/014311698215135. URL https://doi.org/10.1080/
014311698215135.

W. E. Galloway and D. K. Hobday. Coal. in: Terrigenous clastic depositional systems.

Terrigenous Clastic Depositional Systems, pages 253–297, 1983.

W. E. Galloway, T. L. Whiteaker, and P. Ganey-Curry. History of cenozoic north american
drainage basin evolution, sediment yield, and accumulation in the gulf of mexico basin.
Geosphere, 7(4):938, 2011. doi: 10.1130/GES00647.1. URL http://dx.doi.org/10.1130/
GES00647.1.

100

P. Gamba, F. Dell’Acqua, and B. Houshmand. Srtm data characterization in urban areas.
International Archives of Photogrammetry Remote Sensing and Spatial Information Sciences.,
34, 05 2002.

F. Geri, V. Amici, and D. Rocchini. Human activity impact on the heterogeneity of a
mediterranean landscape. Applied Geography, 30(3):370 – 379, 2010.
ISSN 0143-6228.
doi: https://doi.org/10.1016/j.apgeog.2009.10.006. URL http://www.sciencedirect.
com/science/article/pii/S0143622809000721.

D. B. Gesch. Analysis of lidar elevation data for improved identification and delineation
of lands vulnerable to sea-level rise. Journal of Coastal Research, pages 49–58, 2009. doi:
10.2112/SI53-006.1. URL https://doi.org/10.2112/SI53-006.1.

D. B. Gesch, M. J. Oimoen, S. K. Greenlee, C. A. Nelson, M. J. Steuck, and D. J. Tyler. The
national elevation data set. Photogrammetric Engineering and Remote Sensing, 68(1):5–11,
2002. URL http://pubs.er.usgs.gov/publication/70156331.

G. R. Gibson, J. B. Campbell, and R. H. Wynne. Three decades of war and food insecu-
rity in iraq. Photogrammetric Engineering and Remote Sensing, 78(8):885–895, 2012. ISSN
0099-1112. doi: doi:10.14358/PERS.78.8.895. URL https://www.ingentaconnect.com/
content/asprs/pers/2012/00000078/00000008/art00008.

G. R. Gibson, J. B. Campbell, and C. E. Zipper. Sociopolitical influences on cropland area
change in iraq, 20012012. Applied Geography, 62:339 – 346, 2015. ISSN 0143-6228. doi:
https://doi.org/10.1016/j.apgeog.2015.05.007. URL http://www.sciencedirect.com/
science/article/pii/S0143622815001277.

G. R. Gibson, N. L. Taylor, N. C. Lamo, and J. K. Lackey. Effects of recent instability
on cultivated area along the euphrates river in iraq. The Professional Geographer, 69(2):
163–176, 2017. doi: 10.1080/00330124.2016.1194216. URL https://doi.org/10.1080/
00330124.2016.1194216.

L. Giosan and A. M. Freeman. How Deltas Work: A Brief Look at the Mississippi River Delta
in a Global Context, pages 29–32. Springer Netherlands, Dordrecht, 2014.
ISBN 978-
94-017-8733-8. doi: 10.1007/978-94-017-8733-8 3. URL https://doi.org/10.1007/
978-94-017-8733-8_3.

J. B. D. M. J. R. K. R. Good, B. and R. Wilson. Louisiana’s major coastal navigation chan-
nels. Louisiana Department of Natural Resources, Baton Rouge, LA 70802, 1995. URL
http://coastal.la.gov/a-common-vision/2012-coastal-master-plan/.

S. N. Goward and S. D. Prince. Transient effects of climate on vegetation dynamics:
Satellite observations. Journal of Biogeography, 22(2/3):549–564, 1995. ISSN 03050270,
13652699. URL http://www.jstor.org/stable/2845953.

S. Hajian and Z. Hamidi-Esfahani. Date Palm Status and Perspective in Iran, pages 19–
ISBN 978-94-017-9707-8. doi: 10.1007/

47. Springer Netherlands, Dordrecht, 2015.
978-94-017-9707-8 2. URL https://doi.org/10.1007/978-94-017-9707-8_2.

101

R. A. M. Hamid. The concentration of salt in shatt al-arab. Journal of Basrah researchers

(The Humanities), 39, 2014. ISSN 18172695.

J. E. Hansen. Scientific reticence and sea level rise. Environmental Research Letters, 2(2):

024002, 2007. URL http://stacks.iop.org/1748-9326/2/i=2/a=024002.

S. Hassanzadeh, A. Kiasatpour, and F. Hosseinibalam. Sea-level response to atmospheric
forcing along the north coast of persian gulf. Meteorology and Atmospheric Physics, 95
(3):223–237, Feb 2007. ISSN 1436-5065. doi: 10.1007/s00703-006-0213-8. URL https:
//doi.org/10.1007/s00703-006-0213-8.

K. M. Hazzouri, J. M. Flowers, H. J. Visser, H. S. M. Khierallah, U. Rosas, G. M. Pham, R. S.
Meyer, C. K. Johansen, Z. A. Fresquez, K. Masmoudi, N. Haider, N. El Kadri, Y. Idagh-
dour, J. A. Malek, D. Thirkhill, G. S. Markhand, R. R. Krueger, A. Zaid, and M. D.
Purugganan. Whole genome re-sequencing of date palms yields insights into diversi-
fication of a fruit tree crop. Nat Commun, 6:8824, Nov 2015. ISSN 2041-1723. doi: 10.
1038/ncomms9824. URL http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4667612/.
26549859[pmid].

T. Hengl, G. Heuvelink, and A. Stein. Comparison of kriging with external drift and

regression-kriging. Technical Note, 08 2003.

T. Hengl, B. Bajat, D. Blagojevi, and H. I. Reuter. Geostatistical modeling of topog-
raphy using auxiliary maps. Computers and Geosciences, 34(12):1886 – 1899, 2008.
ISSN 0098-3004. doi: https://doi.org/10.1016/j.cageo.2008.01.005. URL http://www.
sciencedirect.com/science/article/pii/S0098300408001106.

M. Hofton, R. Dubayah, J. B. Blair, and D. Rabine. Validation of srtm elevations
over vegetated and non-vegetated terrain using medium footprint lidar. Photogram-
metric Engineering and Remote Sensing, 72(3):279–285, 2006.
ISSN 0099-1112. doi:
doi:10.14358/PERS.72.3.279. URL https://www.ingentaconnect.com/content/asprs/
pers/2006/00000072/00000003/art00005.

C. E. Hughes, P. Binning, and G. R. Willgoose. Characterisation of the hydrology of an
estuarine wetland.
ISSN 0022-1694. doi:
https://doi.org/10.1016/S0022-1694(98)00194-2. URL http://www.sciencedirect.
com/science/article/pii/S0022169498001942.

Journal of Hydrology, 211(1):34 – 49, 1998.

IPCC. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the
Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Uni-
versity Press, Cambridge, United Kingdom and New York, NY, USA, 2013. ISBN ISBN
978-1-107-66182-0. doi: 10.1017/CBO9781107415324. URL www.climatechange2013.
org.

M. T. Jabbar and X. Zhou. Eco-environmental change detection by using remote sensing
and gis techniques: a case study basrah province, south part of iraq. Environmental Earth
Sciences, 64(5):1397–1407, Nov 2011. ISSN 1866-6299. doi: 10.1007/s12665-011-0964-5.
URL https://doi.org/10.1007/s12665-011-0964-5.

102

J. B. C. Jackson, M. X. Kirby, W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque,
R. H. Bradbury, R. Cooke, J. Erlandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange,
H. S. Lenihan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner, and R. R.
Warner. Historical overfishing and the recent collapse of coastal ecosystems. Science,
293(5530):629–637, 2001.
ISSN 0036-8075. doi: 10.1126/science.1059199. URL http:
//science.sciencemag.org/content/293/5530/629.

G. Jiapaer, X. Chen, and A. Bao. A comparison of methods for estimating fractional veg-
etation cover in arid regions. Agricultural and Forest Meteorology, 151(12):1698 – 1710,
2011.
ISSN 0168-1923. doi: https://doi.org/10.1016/j.agrformet.2011.07.004. URL
http://www.sciencedirect.com/science/article/pii/S0168192311002176.

X. Jing, W.-Q. Yao, J.-H. Wang, and X.-Y. Song. A study on the relationship between
dynamic change of vegetation coverage and precipitation in beijings mountainous
areas during the last 20 years. Mathematical and Computer Modelling, 54(3):1079 –
1085, 2011.
ISSN 0895-7177. doi: https://doi.org/10.1016/j.mcm.2010.11.038. URL
http://www.sciencedirect.com/science/article/pii/S0895717710005297. Mathe-
matical and Computer Modeling in agriculture (CCTA 2010).

C. O. Justice, J. R. G. Townshend, B. N. H. HOLBEN, and C. J. Tucker. Analysis of the
phenology of global vegetation using meteorological satellite data. International Journal
of Remote Sensing, 6(8):1271–1318, 1985. doi: 10.1080/01431168508948281. URL https:
//doi.org/10.1080/01431168508948281.

S. Kulp and B. H. Strauss. Global dem errors underpredict coastal vulnerability to sea
level rise and flooding. Frontiers in Earth Science, 4:36, 2016. ISSN 2296-6463. doi: 10.
3389/feart.2016.00036. URL https://www.frontiersin.org/article/10.3389/feart.
2016.00036.

J. X. Leon, G. B. M. Heuvelink, and S. R. Phinn.

Incorporating dem uncertainty in
coastal inundation mapping. PLoS One, 9(9):e108727, Sep 2014. ISSN 1932-6203. doi:
10.1371/journal.pone.0108727. URL http://www.ncbi.nlm.nih.gov/pmc/articles/
PMC4177217/.

X. Li, R. Rowley, J. Kostelnick, D. Braaten, J. Meisel, and K. Hulbutta. Gis analysis of
global impacts from sea level rise. Photogrammetric Engineering and Remote Sensing, 75:
807–818, 07 2009.

J. B. Lindsay. Sensitivity of channel mapping techniques to uncertainty in digital elevation
data. International Journal of Geographical Information Science, 20(6):669–692, 2006. doi:
10.1080/13658810600661433. URL https://doi.org/10.1080/13658810600661433.

J. Liu, L. An, S. S. Batie, R. E. Groop, Z. Liang, M. A. Linderman, A. G. Mertig, Z. Ouyang,
and J. Qi. Human Impacts on land Cover and Panda Habitat in Wolong Nature Reserve,
pages 241–263. Springer US, Boston, MA, 2003. ISBN 978-0-306-48130-7. doi: 10.1007/
0-306-48130-8 9. URL https://doi.org/10.1007/0-306-48130-8_9.

103

J. Liu, T. Dietz, S. R. Carpenter, M. Alberti, C. Folke, E. Moran, A. N. Pell, P. Deadman,
T. Kratz, J. Lubchenco, E. Ostrom, Z. Ouyang, W. Provencher, C. L. Redman, S. H.
Schneider, and W. W. Taylor. Complexity of coupled human and natural systems. Sci-
ence, 317(5844):1513–1516, 2007. ISSN 0036-8075. doi: 10.1126/science.1144004. URL
http://science.sciencemag.org/content/317/5844/1513.

T. S. o. Louisiana.

Louisianas comprehensive master plan for a sustainable coast.
Coastal Protection and Restoration Authority, 2012. URL http://coastal.la.gov/
a-common-vision/2012-coastal-master-plan/.

A. A. Mahmood, A. M. Eassa, H. Mohammed, Muayad, and Y. Shubbar, Israa. As-
sessment of ground water quality at basrah, iraq by water quality index (wqi). Jour-
nal of Babylon University/Pure and Applied Sciences, 21:2531–2543, 2013. URL https:
//bit.ly/2NIZsMS.

A. Marina, A. Heidi, B. L. A., B. Nicholas, D. L. E., D. S. A., J. C. A., F. Jos, H. D. S., K. T. T.
A., L. J. (Jack), M. W. J., M. H. D. G., M. J. D. A., M. Michael, P. Guillermo, P. R. Gilmore,
R. C. L., R. N. J., S. David, and U. Gerald. Research on coupled human and natural
systems (chans): Approach, challenges, and strategies. The Bulletin of the Ecological
Society of America, 92(2):218–228, 2011. doi: 10.1890/0012-9623-92.2.218. URL https:
//esajournals.onlinelibrary.wiley.com/doi/abs/10.1890/0012-9623-92.2.218.

E. Mcleod, B. Poulter, J. Hinkel, E. Reyes, and R. Salm. Sea-level rise impact mod-
els and environmental conservation: A review of models and their applications.
Ocean and Coastal Management, 53(9):507 – 517, 2010.
ISSN 0964-5691. doi: https:
//doi.org/10.1016/j.ocecoaman.2010.06.009. URL http://www.sciencedirect.com/
science/article/pii/S0964569110000852.

G. A. Meehl, W. M. Washington, W. D. Collins, J. M. Arblaster, A. Hu, L. E. Buja, W. G.
Strand, and H. Teng. How much more global warming and sea level rise? Science, 307
(5716):1769–1772, 2005.

B. Meyssignac and A. Cazenave. Sea level: A review of present-day and recent-past
changes and variability.
ISSN 0264-3707.
doi: https://doi.org/10.1016/j.jog.2012.03.005. URL http://www.sciencedirect.com/
science/article/pii/S0264370712000464.

Journal of Geodynamics, 58:96 – 109, 2012.

A. J. Mohammed. Remnants of war in the province of basra (a study of geography). Basra

studies journal, 5:115–150, 2008. ISSN 1994-4721. URL https://bit.ly/2KYkNnb.

K. D. Muslih.

Identifying the climatic conditions in iraq by tracking down cooling
events in the north atlantic ocean in the period 30000 bc. Miscellanea Geographica, 18
(3):40 – 46, 2014. URL https://content.sciendo.com/view/journals/mgrsd/18/3/
article-p40.xml.

K. Nageswara Rao, P. Subraelu, T. Venkateswara Rao, B. Hema Malini, R. Ratheesh,
S. Bhattacharya, A. S. Rajawat, and Ajai. Sea-level rise and coastal vulnerability: an

104

assessment of andhra pradesh coast, india through remote sensing and gis.
nal of Coastal Conservation, 12(4):195–207, Nov 2008.
s11852-009-0042-2. URL https://doi.org/10.1007/s11852-009-0042-2.

Jour-
ISSN 1874-7841. doi: 10.1007/

R. J. Nicholls and A. Cazenave. Sea-level rise and its impact on coastal zones. Science,
328(5985):1517–1520, 2010. ISSN 0036-8075. doi: 10.1126/science.1185782. URL http:
//science.sciencemag.org/content/328/5985/1517.

NOAA. Data Integration Visualization Exploration and Reporting (DIVER). National Oceanic
and Atmospheric Administration, US Department of Commerce, 2018. URL https:
//www.diver.orr.noaa.gov/.

H. Nomas. The concentration of salt in shatt al-arab. Arts Basra Magazine, 2006.

J. Oksanen and T. Sarjakoski. Error propagation of dem-based surface derivatives. Com-
puters and Geosciences, 31(8):1015 – 1027, 2005. ISSN 0098-3004. doi: https://doi.org/
10.1016/j.cageo.2005.02.014. URL http://www.sciencedirect.com/science/article/
pii/S0098300405000701.

M. C. Peel, B. L. Finlayson, and T. A. McMahon. Updated world map of the kppen-
geiger climate classification. Hydrology and Earth System Sciences, 11(5):1633–1644, 2007.
doi: 10.5194/hess-11-1633-2007. URL https://www.hydrol-earth-syst-sci.net/11/
1633/2007/.

S. Penland and K. E. Ramsey. Relative sea-level rise in louisiana and the gulf of mexico:
1908-1988. Journal of Coastal Research, 6(2):323–342, 1990. ISSN 07490208, 15515036. URL
http://www.jstor.org/stable/4297682.

W. T. Pfeffer, J. T. Harper, and S. O’Neel. Kinematic constraints on glacier contributions
to 21st-century sea-level rise. Science, 321(5894):1340–1343, 2008. ISSN 0036-8075. doi:
10.1126/science.1159099. URL http://science.sciencemag.org/content/321/5894/
1340.

J. Phillips. Climate change and surface mining: A review of environment-human interac-
tions and their spatial dynamics. Applied Geography, 74:95 – 108, 2016. ISSN 0143-6228.
doi: https://doi.org/10.1016/j.apgeog.2016.07.001. URL http://www.sciencedirect.
com/science/article/pii/S0143622816301746.

B. Poulter and P. N. Halpin. Raster modelling of coastal flooding from sealevel rise. In-
ternational Journal of Geographical Information Science, 22(2):167–182, 2008. doi: 10.1080/
13658810701371858. URL https://doi.org/10.1080/13658810701371858.

S. D. Prince.

Satellite remote sensing of primary production:

comparison of re-
International Journal of Remote Sensing, 12(6):
sults for sahelian grasslands 1981-1988.
1301–1311, 1991. doi: 10.1080/01431169108929727. URL https://doi.org/10.1080/
01431169108929727.

R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for

Statistical Computing, Vienna, Austria, 2017. URL http://www.R-project.org/.

105

B. Rabus, M. Eineder, A. Roth, and R. Bamler. The shuttle radar topography missiona
ISPRS Journal
new class of digital elevation models acquired by spaceborne radar.
of Photogrammetry and Remote Sensing, 57(4):241 – 262, 2003.
ISSN 0924-2716. doi:
https://doi.org/10.1016/S0924-2716(02)00124-7. URL http://www.sciencedirect.
com/science/article/pii/S0924271602001247.

S. Rahmstorf. A semi-empirical approach to projecting future sea-level rise. Science, 315
ISSN 0036-8075. doi: 10.1126/science.1135456. URL http://

(5810):368–370, 2007.
science.sciencemag.org/content/315/5810/368.

R. T. Reynolds. Heart of the storm, the genesis of the air campaign against iraq. Air
University Press, Maxwell Air Force Base, Alabama, 1995. URL http://www.au.af.mil/
au/awc/awcgate/au/reynolds.pdf.

J. C. M. M. M. S. J. A. G. O. P. L. A. B. E. B. R. A. S. A. G. Robert J. Hijmans, Jacob van Etten.
raster: Geographic Data Analysis and Modeling, 2017. URL https://cran.r-project.
org/web/packages/raster/index.html.

E. Rodriguez, C. S. Morris, and J. E. Belz. A global assessment of the srtm performance.
Photogrammetric Engineering and Remote Sensing, 72(3):249–260, 2006. ISSN 0099-1112.
doi: doi:10.14358/PERS.72.3.249. URL https://www.ingentaconnect.com/content/
asprs/pers/2006/00000072/00000003/art00002.

M. A. Saad.

Seasonal variations of some physicochemical conditions of shatt al-
arab estuary, iraq. Estuarine and Coastal Marine Science, 6(5):503 – 513, 1978.
ISSN
0302-3524. doi: https://doi.org/10.1016/0302-3524(78)90027-0. URL http://www.
sciencedirect.com/science/article/pii/0302352478900270.

B. Sande, J. Lansen, and C. Hoyng. Sensitivity of coastal flood risk assessments to digital
elevation models. Water, 4(3):568–579, 2012. ISSN 2073-4441. doi: 10.3390/w4030568.
URL http://www.mdpi.com/2073-4441/4/3/568.

M. M. Sherif and V. P. Singh.

Hydrological Processes, 13(8):1277–1287, 1999.

sion in coastal aquifers.
10.1002/(SICI)1099-1085(19990615)13:8(cid:104)1277::AID-HYP765(cid:105)3.0.CO;2-W.

Effect of climate change on sea water intru-
doi:
URL

https://onlinelibrary.wiley.com/doi/abs/10.1002/%28SICI%291099-1085%
2819990615%2913%3A8%3C1277%3A%3AAID-HYP765%3E3.0.CO%3B2-W.

A. Shortridge. Shuttle radar topography mission elevation data error and its relationship
to land cover. Cartography and Geographic Information Science, 33(1):65–75, 2006. doi:
10.1559/152304006777323172. URL https://doi.org/10.1559/152304006777323172.

A. Shortridge and J. Messina. Spatial structure and landscape associations of srtm error.
Remote Sensing of Environment, 115(6):1576 – 1587, 2011. ISSN 0034-4257. doi: https:
//doi.org/10.1016/j.rse.2011.02.017. URL http://www.sciencedirect.com/science/
article/pii/S0034425711000678.

106

A. Singh, T. R. Loveland, D. Del Pietro, H. Partow, P. Peduzzi, and W. C. Silverio Torres.
One planet many people: Atlas of our changing environment, 2005. URL https://
archive-ouverte.unige.ch/unige:32650. ID: unige:32650.

B. Smiley, J. Trofymow, and K. Niemann.

Spatially-explicit reconstruction of 100
years of forest land use and disturbance on a coastal british columbia douglas-fir-
dominated landscape:
Implications for future watershed-scale carbon stock recov-
ery. Applied Geography, 74:109 – 122, 2016. ISSN 0143-6228. doi: https://doi.org/10.
1016/j.apgeog.2016.06.011. URL http://www.sciencedirect.com/science/article/
pii/S014362281630162X.

T. L. Sohl, B. M. Sleeter, Z. Zhu, K. L. Sayler, S. Bennett, M. Bouchard, R. Reker, T. Haw-
baker, A. Wein, S. Liu, R. Kanengieter, and W. Acevedo. A land-use and land-cover
modeling strategy to support a national assessment of carbon stocks and fluxes. Ap-
plied Geography, 34:111 – 124, 2012.
ISSN 0143-6228. doi: https://doi.org/10.1016/
j.apgeog.2011.10.019. URL http://www.sciencedirect.com/science/article/pii/
S0143622811002220.

R. Springborg. Infitah, agrarian transformation, and elite consolidation in contemporary
iraq. Middle East Journal, 40(1):33–52, 1986. ISSN 00263141, 19403461. URL http://www.
jstor.org/stable/4327247.

G. D. Steyer and D. W. Llewellyn. Coastal wetlands planning, protection, and restoration
act: A programmatic application of adaptive management. Ecological Engineering, 15(3):
385 – 395, 2000. ISSN 0925-8574. doi: https://doi.org/10.1016/S0925-8574(00)00088-4.
URL http://www.sciencedirect.com/science/article/pii/S0925857400000884.

S. Sultan, F. Ahmad, N. Elghribi, and A. Al-Subhi. An analysis of arabian gulf
monthly mean sea level. Continental Shelf Research, 15(11):1471 – 1482, 1995a.
ISSN
0278-4343. doi: https://doi.org/10.1016/0278-4343(94)00081-W. URL http://www.
sciencedirect.com/science/article/pii/027843439400081W.

S. Sultan, F. Ahmad, N. Elghribi, and A. Al-Subhi. An analysis of arabian gulf
monthly mean sea level. Continental Shelf Research, 15(11):1471 – 1482, 1995b.
ISSN
0278-4343. doi: https://doi.org/10.1016/0278-4343(94)00081-W. URL http://www.
sciencedirect.com/science/article/pii/027843439400081W.

L. Tara, S. Ashton, and M. Joseph. The influence of land cover on shuttle radar topography
mission (srtm) elevations in lowrelief areas. Transactions in GIS, 14(4):461–479, 2010. doi:
10.1111/j.1467-9671.2010.01217.x. URL https://onlinelibrary.wiley.com/doi/abs/
10.1111/j.1467-9671.2010.01217.x.

J. G. Titus and C. Richman. Maps of lands vulnerable to sea level rise: modeled elevations
along the us atlantic and gulf coasts. Climate Research, 18(3):205–228, 2001. doi: 10.3354/
cr018205. URL https://www.int-res.com/abstracts/cr/v18/n3/p205-228/.

C. Tripp. A History of Iraq. Cambridge University Press, 3 edition, 2007. doi: 10.1017/

CBO9780511804304.

107

J. Tu. Spatially varying relationships between land use and water quality across an
urbanization gradient explored by geographically weighted regression. Applied Ge-
ography, 31(1):376 – 392, 2011.
doi: https://doi.org/10.1016/
j.apgeog.2010.08.001. URL http://www.sciencedirect.com/science/article/pii/
S0143622810000846. Hazards.

ISSN 0143-6228.

U. N. UN. Oil-for-food. Online, 2003. URL http://www.un.org/Depts/oip/background/

fact-sheet.html.

M. Vermeer and S. Rahmstorf. Global sea level linked to global temperature. Proceedings
ISSN 0027-8424. doi:

of the National Academy of Sciences, 106(51):21527–21532, 2009.
10.1073/pnas.0907765106. URL http://www.pnas.org/content/106/51/21527.

A. Verruijt. A note on the ghyben-herzberg formula. International Association of Scientific
Hydrology. Bulletin, 13(4):43–46, 1968. doi: 10.1080/02626666809493624. URL https:
//doi.org/10.1080/02626666809493624.

J. M. Visser, C. E. Sasser, R. H. Chabreck, and R. G. Linscombe. Marsh vegetation types of
the mississippi river deltaic plain. Estuaries, 21(4):818–828, Dec 1998. ISSN 0160-8347.
doi: 10.2307/1353283. URL https://doi.org/10.2307/1353283.

W. van de Lageweg, S. McLelland, and D. Parsons. Experimental investigations of natural
In EGU General
diatomaceous biofilm behaviour and sediment stabilizing capacity.
Assembly Conference Abstracts, volume 19 of EGU General Assembly Conference Abstracts,
page 1068, Apr. 2017.

H. Wang, Y. P. Hsieh, M. A. Harwell, and W. Huang. Modeling soil salinity distribu-
tion along topographic gradients in tidal salt marshes in atlantic and gulf coastal re-
gions. Ecological Modelling, 201(3):429 – 439, 2007. ISSN 0304-3800. doi: https://doi.
org/10.1016/j.ecolmodel.2006.10.013. URL http://www.sciencedirect.com/science/
article/pii/S0304380006004959.

X. Wang, K. Jia, S. Liang, Q. Li, X. Wei, Y. Yao, X. Zhang, and Y. Tu. Estimating fractional
vegetation cover from landsat-7 etm+ reflectance data based on a coupled radiative
transfer and crop growth model. IEEE Transactions on Geoscience and Remote Sensing, 55
(10):5539–5546, Oct 2017. ISSN 0196-2892. doi: 10.1109/TGRS.2017.2709803.

A. D. Werner and C. T. Simmons. Impact of sealevel rise on sea water intrusion in coastal
aquifers. Groundwater, 47(2):197–204, 2009. doi: 10.1111/j.1745-6584.2008.00535.x. URL
https://onlinelibrary.wiley.com/doi/abs/10.1111/j.1745-6584.2008.00535.x.

E. Williams. Aquatecture: Architectural adaptation to rising sea levels. Graduate Thesis

and Dissertations., 2009. URL http://scholarcommons.usf.edu/etd/85.

C. Wohlfart, B. Mack, G. Liu, and C. Kuenzer. Multi-faceted land cover and land use
change analyses in the yellow river basin based on dense landsat time series: Exem-
plary analysis in mining, agriculture, forest, and urban areas. Applied Geography, 85:73
– 88, 2017. ISSN 0143-6228. doi: https://doi.org/10.1016/j.apgeog.2017.06.004. URL
http://www.sciencedirect.com/science/article/pii/S014362281630813X.

108

W. W. Wood, R. M. Bailey, B. A. Hampton, T. F. Kraemer, Z. Lu, D. W. Clark, R. H.
James, and K. A. Ramadan. Rapid late pleistocene/holocene uplift and coastal evo-
lution of the southern arabian (persian) gulf. Quaternary Research, 77(2):215 – 220,
2012. ISSN 0033-5894. doi: https://doi.org/10.1016/j.yqres.2011.10.008. URL http:
//www.sciencedirect.com/science/article/pii/S0033589411001384.

B. Yassein, K. A Al Asaady, A. Alwaeli, and M. Chaichan. Environmental impacts of salt
IOSR Journal of Environmental Science, Toxicology and
tide in shatt al-arab-basra/iraq.
Food Technology, 10:35–43, 01 2016. URL http://un.uobasrah.edu.iq/papers/30.pdf.

A. P. Yunus, R. Avtar, S. Kraines, M. Yamamuro, F. Lindberg, and C. S. B. Grimmond.
Uncertainties in tidally adjusted estimates of sea level rise flooding (bathtub model) for
the greater london. Remote Sensing, 8(5), 2016. ISSN 2072-4292. doi: 10.3390/rs8050366.
URL http://www.mdpi.com/2072-4292/8/5/366.

S. Zellmer and C. Klein. Mississippi river stories: lessons from a century of unnatural
disasters. SMU Law Review, 60:1471–1538, 06 2007. URL https://digitalcommons.
unl.edu/lawfacpub/12/.

X. Zhang, C. Liao, J. Li, and Q. Sun. Fractional vegetation cover estimation in arid and
International Journal
semi-arid environments using hj-1 satellite hyperspectral data.
of Applied Earth Observation and Geoinformation, 21:506 – 512, 2013.
ISSN 0303-2434.
doi: https://doi.org/10.1016/j.jag.2012.07.003. URL http://www.sciencedirect.com/
science/article/pii/S0303243412001419.

109