Geospatial Resource Access Analysis in Hedaru, Tanzania

surfscreechingSoftware and s/w Development

Dec 11, 2013 (3 years and 4 months ago)

75 views

1




Geospatial Resource Access
Analysis in Hedaru,
Tanzania





Integration

of

GIS

and

Stakeholders

to

Guide

D
evelopment











August

2013


2


Acknowledgements:


I
o
w
a

Sta
t
e

Un
i
v
ersit
y
,

Emp
o
w
er

T
anzania

Inc.
,

E
u
r
opean

Space

Agenc
y
,

The

Uni
t
ed

R
epublic

of

T
anzania,

Same

Dist
r
ict

Office,

T
anzania

Abstract:

In 2012
,

w
o
r
ld

leaders

met

f
or

the

Uni
t
ed

Nations

Con
f
e
r
ence

on

Sustainable

D
e
v
elopment (Rio+20)

and

decla
r
ed

the

need

f
or

an

inc
r
eased

collabo
r
at
i
v
e

ef
f
ort

t
o

p
ro
vide

ru
r
al

agricultu
r
alists

with access

t
o

health

ca
r
e,

social

services,

education

and

t
r
aining,

and

w
a
t
er

in

o
r
der

t
o

c
r
ea
t
e

g
r
ea
t
er

g
lobal
f
ood

securit
y

(United Nations
2012)
.

This

r
esea
r
ch

stu
d
y

e
x
amines

the

geospatial

t
r
ends

of

f
ood

securit
y
,

w
a
t
er

access,

health ca
r
e

access,

and

agricultu
r
al

p
r
actices

in

the
ru
r
al

Hedaru Valley of
northeastern
T
anzania

f
or

the

purpose

of

c
r
eating

a

comp
r
ehens
iv
e

d
e
v
elopment

f
r
ame
w
o
r
k

in that region
.

This

i
nv
esti
g
ation

w
as

conduc
t
ed

in

th
r
ee

villages

with

dif
f
e
r
ent

natu
r
al

r
esou
r
ces

and

w
as

comple
t
ed

th
r
ou
g
h

t
w
o

primary

methods:

1)

the

incorpo
r
ation

of

local

kn
o
w
ledge

and

mapping

of

r
egional

inf
r
astructu
r
es
,

such

as

schools,

clinics,

r
oads,

and

w
a
t
er

access

points,

and

2)

the

administ
r
ation

of

80

quantitat
i
v
e

sur
v
ey
s

t
o

r
andom
l
y

selec
t
ed

households
,

along

with

14

qualitat
i
v
e

f
oll
o
w

up

f
ocus

g
r
oups

f
or

the

household

participants.

B
y

r
eco
r
d
ing

a

p
h
y
sical

location

t
o

all

sur
v
e
y
ed

households

and

associa
t
ed

r
emo
t
e
l
y

sensed

data,

r
esea
r
chers

w
e
r
e

able

t
o

e
x
amine

l
i
v
elihood

spatial

distinctions

based

off

p
ro
ximities

t
o

w
a
t
e
r
,

clinics,

and

schools.

This

r
esea
r
ch

documents

the

cor
r
elation

among

health

ca
r
e

access,

w
a
t
er

a
v
ailabilit
y
,

and

education

access

t
o

achi
e
ve
f
ood

securit
y
.

This

in
f
ormation

has been provided to

local

d
e
v
elopment

p
r
actitioners,

enabling

them

t
o

d
e
v
elop

comp
r
ehens
iv
e

st
r
a
t
egies

and

ta
r
get

the

most

inhibiting

barriers.

Mo
r
e
o
v
e
r
,

the

r
esea
r
ch

p
ro
vides

a

t
empla
t
e

f
or

the

implementation

of

participa
t
ory

GIS

in

ru
r
al

ag
r
arian

r
egions

as

a

t
ool
f
or

d
e
v
elopmen
t
.




Primary

I
n
v
esti
g
a
t
ors:

D
y
lan

Cla
r
k
1
,

Deepak

P
r
em
k
umar
1
,

D
r
.

R
obert

Mazur
2
,

Eli Kisimbo
3

1
I
o
w
a

St
a
t
e

Uni
v
ersi
t
y

S
tuden
t
,

2
I
o
w
a

St
a
t
e

Uni
v
ersi
t
y

F
acul
t
y
,
3
Empower Tanzania Inc.

3





Abbreviations


AfDB: African Development Bank

DFI
D: Department of International Development

ESA: European Space Agency

ETI: Empower Tanzania Inc.

FAO
: Food and Agriculture Organization

FGD: Focused Group Discussion

GDEM: Global Digital Elevation Map

GIS: Global
Information Technology

GWR: Geographically Weighted Regression

IRB: Institutional Review Board

ISU: Iowa State University

IUCN: International Union for Conservation of Nature

NDVI: Normalized Difference Vegetation Index

NGO: Nongovernmental Organization

NO
AA
: National oceanic and Atmospheric Administration

PBWB: Pangani Basin Water board

PGIS
: Participatory Geographic Information Systems

RSD: Remotely Sensed Data

UN
EP: United Nations Environment

Programme

USGS: United States Geographic Survey






4


Introduction


Populations around the world are feeling the inc
reased effects of anthropogenic
-
forced
environmental changes

and rapid population movements

(Satterthwaite 2009)

(Christensen J.
H. 2007)

(United Nations Economic and Social Affairs 2004)
. These environmental and social

shifts

are
having an elevated impact on

the livelihoods of agriculturalists

and pastoralists in
developing count
ries

(Burke M B 2009)

(Food and Agriculture Organization 2008)

(Washingron R. 2006)

(Battisti D. 2009)
. One of these critically a
ffected regions

is

the
lowlands of the Pan
gani River Basin

in

Same,
Tanzania
. The compounding effects of r
educed
glacial and snow runoff from

Mount Kilim
a
njaro and Mount Meru
, increased drought
prevalence, and elevated population growth have culminated
in se
ver
e

resource constraints

(Pangani Basin Water Board, 2010)

(Mbonile J n.d.)

(World Health Organization 2011)

(Turpie
J. K. 2005)
.

As the fluvial systems swiftly change and
migration

from the
highlands
burgeons
,
populations will increasingly
experience

water and food shortages.


This
research was
esta
blished as it became clear that

Empower Tanzania Inc., the

in
-
country

partner and
Tanzanian n
on
-
governmental
o
rganization (NGO), n
eeded further quantitative
information about the livelihoods of the people living in

the Hedaru Valley. After
initial
examination of the stakeholders in the area, it became obvious that despite a broad range

of
resources and
development
organizations working in the region, there was a

general lack of
empirical data

and
cross
-
disciplinary information that the organizations
could utilize
to guide
their policies and programs

(National Res
earch Council Staff 2002)
.


This
research

integrates

various tools

u
sually used in solidarity

t
o gain a broad,
comprehensive understanding of the
regional livelihood constraints

and
, in a broader context,
trends between natural resource proximities and social indicators.
Through the combined use
of quantitative and qualitative surveys implemented through
Participatory
Geographic
I
nformation
S
ystems (PGIS)
, quantitative
time series analysis of Landsat
-
based

Normalized
Difference
Vegetation

Index (NDVI)
images
, and statistical and geospatial analysis
,

researchers
added to the acade
mic understanding of the region

and demonstrated a

new method for future
research
-
based development.


PGIS
has been shown to
help
distill indigenous knowledge, examine spatial trends, and enable
issue and context driven research

(Dunn E. 2007)

(Chirowodza 2009)
.

PGIS was used as the
backbone of this research to gain a deep un
derstanding of critical regional resources, record
indigenous knowledge, and implement qualitative
f
ocus
g
roup
d
iscussions (FGD
s
).
In addition
,

through utilizing

the

NVDI time series, researchers were able to examine changes in the
regional vegetation density from 1987 to 2011.
Use of NDVI images has

been shown

to be a
cost effective method

of analyzing vegetation change related to societal or environmental
factors

(I. M. Geerken R. 2004)

(Cohen W. 2004)

(Michalak W. Z. 1993)
.


This research was done with the collaboration of
the
European Space Agency (ESA)
,
Iowa State
University

(ISU)
, and
Empower Tanzania Inc
.

(ETI)
.
The data
provide
s

further insights into the
health, education, water,
and nutrition

pressures
faced by people

living in the
region
.



5


Figure 1

The Headru Valley is located in southeaster
n

Tanzania
,

in the Same District

Study Site


The three population centers that were chosen
for the research
are located in the Hedaru
W
ard
in

southern

Same District, Tanzania. T
hese villages

Gunge, Katahe, and Mabilion

h
ad
varying proximities
to water, centralized markets,
schools,
health facilities, main roads, and
irrigated fields. A
dditionally,
two of the villages were primarily
inhabited by the
agrarian
-
based Pare tribe, while Katahe was

inhabited by the

pastorally
-
based

Masai
. All of these
differences played a role as variables to assess trends that could be extrapolated further in the
region.


The Hedaru W
ard is approximately a 41,000 hectare
s

of mostly arid and sparse lands with
approximately 50,000
residences
. Very little
detailed data were

available
for

this
specific
region prior to this study,
although some general

information was availab
le
at the larger
regional district scale.



The Same District, approximately 150km
south of

Mount Kilim
a
njaro
, has

a diverse
topography

and landscape
.
In the southeast, the 1500 meter
high
South
Pare Mountain range
towers

over the western 500 meter

high

arid planes. Running through lowlands, the Pangani
river provides for most of the life in the district. During the bimodal rainy
seasons,

the
mountainous side of the region receives an annual average of 1500mm of rainfall, while the
lowlands receive aro
und 500mm

(Kinoti 2010)

(Mutabazi K. D. 2006)
. Gunge, Katah
e, and
Mabilion are all within 8
km of the Pangani river in the lowlands of the district.

Although a
national highway runs through the region about 10km from the villages
,

infrastructure,
utilities,
and

social services in the region

are sparse.

(
figure

1)
.


For all GIS aspects of the research, data was collected and processed using Geographic
Coordination System WGS 1984 or Projected C
oordinate System WGS 1984. Al
though more
accurate
corrections could have been made with a localized coo
rdinate system, there is no
standard system for the Tanzania region. The research area is in UTM zone 37M, central point
Lat:

-
4.56,
Long:
37.89
.


6


Methodology


This research consisted of three central phases; (
I
) literature and
geographic information
system (
GIS
)

data review
,

(II
) in
-
field PGIS data collection, survey administration,
and
focus
group facilitation,
and
(
III
)
field
-
collected

d
ata ana
lysis,
N
ormalized
D
ifference Vegetation
I
ndex (NDVI)

series comparison,
recommendations, an
d distribution
of information.


Phase I:

Researchers began this study with a review of
literature

and geospatial data

on the
region
.

Initially, demographic data was collected from previous censuses through the
Tanzanian Statistics and Census Bureau

(National Bureau of Statistics 2005)

(National Bureau
of Statistics, Tanzania Ministry of Planning, Economics and Empowerment 2006)
. Agricultural,
hydrologic, and climatic data was collected through the

Food and Agri
culture Organization
(F
AO)

(Food and Agriculture Organization 2011)
, United Nations Environment

Program
m
e
(UNEP), International Union for Conservation of Nature (IUCN), and the Pangani Basin Water
B
oard (PBWB). In addition, GI
S data was acquired from
the National Oceanic and At
mospheric
Administration (NOAA),
the
African Development Bank (
Af
DB
)
, the United States Geographic
Survey (USGS)
,
and the E
uropean Space Agency

(ESA)
.
Prior to formulating quantitative and
qualitative surveys, researchers created regional maps using
Global Digital Elevation Map
(
GDEM
)

data
,

L
andsat

4,

5, and 7 images, SPOT data (

Data provided by the European Space
Agency
), and shapefiles

made available from the Wor
l
d Bank. These maps and geospatial
information
were

used to develop an understanding of resource dispersion, infrastructure
placement, and watershed data

(Jensen J. R. 2004)
.
Additionally,
Landsat

data was used to gain
an understanding of the region

s soil type

(bands 1
-
2
-
3 and 1
-
2
-
4)
, vegetation density

(bands
7
-
4
-
2

and 1
-
4
-
7)
, and
large scale regional infrastructure (bands 7
-
4
-
2)

(Michalak W. Z 1993)

(Cohen W. 2004)
.
It was observed the region had

iron

clan

soils,
a
dearth

o
f

paved roads, and
minimal vegetation

(Scepan J. 1999)
.
Finally
, in order to understand where the villages were
geospatially
located
prior to field research,
a
n

ESA
image content release was approved for the
research, enabling the use of SPOT data. This provided a 5m

panchromatic

view of the region
and detecti
on of h
uts,
something

Landsat

3
0
m images
could
not
allow
.

This initial review of
the available information on the physical and social geography of the region also highlighted
the knowledge gaps that existed.

After knowledge gaps were
determined
, investigative

tools
were developed to specifically
cover

those
gaps. These investigative

tools were approved by
the Institutional Review Board (IRB)

at ISU.


Phase II:

This phase c
onsisted of collecting the geospatial data for
critical

infrastructure
,
randomly
selecting 80 households for survey participation, facilitating the distribution of 80
house
hold surveys and 16 focus group discussions (FGD)
, and networking with development
practitioners in the region for future cooperation and research
-
findings distribut
ion.
Important infrastructure w
as

determined based on the input from the District Agriculture
Extension Officers of Hedaru and Mabilion, the
former

Mayor of Hedaru Town, the Head of the
Pangani River Water Board, and
the
Lutheran Church members.
With this input,
researchers

were able to locate wells, churches, schools, health facilities, etc. that play crucial roles in the
surrounding communities. Th
is

infrastructure w
as

then visited, assess
ed
, and geo
-
referenced.


Household selection and survey
distribution was slightly inhibited by the absence of
population density data that could have been overlaid to create a truly random sample.
7


Instead,
researchers

created 19 clusters based on the houses proximity to the Pangani River
(
the
only water source
) and
primary
schools. These two variables were chosen because of
their
hypothesized

impact in household
s


livelihoods. After the clusters were created, a
specified
number

of houses were chosen within the cluster (ba
sed on cluster area)

to
participate in surveying
. Houses w
ere chosen by randomly selecting a point on a road, or path

with
in the cluster
, randomly
generating

a vector (1deg
-
360deg), and

then

lastly randomly
generating

a distance from point (1m
-
60m). Households were then not
ified about the study
and asked to participate
by

their village chair, Agriculture Extension Officer, or
Lutheran
Community
Evangelist. Surveys and focus groups
were then administered over a 7
-
week
period as outlined in the IRB submission.


Phase III:

Fol
lowing all data collection, investigators
analyzed

data by running
linear
regressions with Excel
, analyzing response clustering with

ArcGIS
spatial analyst

tool
s,
examining influences of infrastructure proximities, and
via the

NDVIs

image time series

(Brunsdon C. 1998)

(ESRI 2013)
. F
ocus group discussion
s

were used to assist in confirming
observed trends and
found correlations
.



Linear regressions using Excel were performed for all
quantitative survey responses. The
confidence cutoff was 95%, with a p value <0.05.

As the nature of the study asked a number of
binomial questions, there was some decreased accuracy analysis of the responses.


Additionally, because of the low survey num
ber

and homogeneity in the
region
, there was not
enough variation in the data to run regressions on certain variables

such as malari
a
.


Spatial
cluster analyses

were
run

using ArcGIS
tools

to examine spatial trends of survey
responses. Researchers used both

high/
low clustering
analysis
and
a
hot spot analyst

tools
.
Both were based on a
fixed

distance band conceptualization of spatial relationships and
Euclidian distance.
T
he constant dis
tance band of
842.94 meters
determined as an optimal
band distance, and was used

for all spatial analyses.

Similar to the linear regressions, the
confidence cutoff was 95% for all spatial analyses. Researchers began by running the high/low
clustering too
l to determine variables that where statistically significant. The variables that
showed significance were then analyzed using the hot spot
analyst.


Spatial proximity correlations w
ere conducted using the ArcGIS spatial join tool, to create a
distance fi
eld for each survey location from
infrastructures

proximity values for h
ealth clinics,
churches, schools, the Pangani river, and the main road. All of these infrastructures play
critical roles in providing social services, economic market access, and
water. After
proximities were found for each household, the data was exported to excel for linear
regressions using survey answers as dependent variables and distances as independent
variables.


NDVIs were performed using
Landsat 4, 5, and 7 images from 1
987 to 2011. This additional
component was done to improve on the understanding of regional vegetation change

(Coppin
P. 2004)

(Sadek S. H. A. 1993)

(Fig

1
).

The images used for the NDVI calculat
ions were chosen
with the exclusionary criteria of
<
10% cloud cover. With this constraint
,

and the need to
standardize
season of image capture, the dry season

f
rom October to late January was
selected

(Kinoti 2010)
.

This was done because excessive atmospheric moisture can skew NDVI
8


responses in multi
plicative ways

(Song C. 2001)
.

Researchers made two exceptions to this, as
images were needed to fill series gaps. This may have introduced some
skewed

results, but
with year
-
to
-
year analysis
,

the
results during these periods
were

consistent.
The NDVI was
then calculated by using Ar
cGIS raster math tool,
applying

the conventional formula:
NDVI=(NIR
-
RED
)/(NIR+
RED
)

(University of Amsterdam, IBED 2011)

(B. N. Geerken R. 2005)
.
These NDV
Is
images
were then compared using zonal
classification change calculator. Three
zones were created

for the purpose of identifying changes more accurately
: irrigated region,
lowland region, and mountain region (above 1000m). The zonal classification change
s

were
compared per year and
by
zone.


The NDVI change comparison was used to assist in developing an understanding of the
region

s stressed natural resources, specifically vegetation. Numerous desertification
problems in Tanzania and in neighboring countries, such as Kenya, have increased in

the past
few decades as populations increase their usage of wood for fuel supply and plow
new fields
for agricultural purposes.
Despite limitation

(Bakr N. 2010)
,
the NDVI comparisons

provide
d

researchers with knowledge to be used in con
junction

with F
GD
s

and quantitative survey
questions about fuel sources and where
wood is harvested
.








Due to

the AfDB,
UK’s
Department for I
nternational Development (
DFID
)
, World Bank, ETI,
and
the
Tanzania
n

Government
expressing

interest in research results during
correspondences

while in
-
country,

data

has been s
ummarized to meet the

specific
needs of these agencies

and
shared
.



Results


The results from this study can be separated into four categories:
the examination of
correlations

and regressions

within the survey responses
, geospatial
relationships between
the responses, FGD summaries, and NDVI results. All of these various methods
,

though
independently analyzed
,

support significant trends.

Imagery Date

Spatial Resolution

Satellite
Sensor

Number of
Bands

Scene ID

January 1, 1987

30m
-
120m

Landsat 5 TM

7

L5167063
06319870101

January 7, 1995

30m
-
120m

Landsat 5 TM

7

L516706306319950107

January 12, 1997

30m
-
120m

Landsat 5 TM

7

LT51670631997012JSA00

February 1, 2010

30m
-
120m

Landsat 5 TM

7

LT51670632010032MLK00

September 9, 2000

30m
-
60m

Landsat 7 TM

7

LE71670632001015SGS00

January 15, 2001

30m
-
60m

Landsat 7 TM

7

LE71670632001015SGS00

January 2, 2002

30m
-
60m

Landsat 7 TM

7

LE71670632002002SGS00

October 26, 2011

30m
-
60m

Landsat 7 TM

7

LE71670632011299ASN00






Figure 2

Landsat image data table for remote sensing NDVI analysis

9



S
urvey c
orrelations

and regressions
:

Through multivariate linear

regression analysis, a number
of statistically significant trends emerged.

To analyze
wellbeing
, an annual
household
income
metric was generated

by monetizing crop yields at market prices.

With the dependent
variables (in the table below) as a function
of
many

variables (from age to gender to how much
land they owned),
statistically significant results were produced, some of which are listed
below

(only the statistically significant variables in the regression are posted)
.
The most
significant impact not
ed was whether or not a family was able to irrigate.
By irrigating,
families were able to increase their annual income c
ould increase by a factor of 1,7
8
1,725 Tsh.
or $1,130
, which is about the GDP per capita of Tanzania

a
tremendous increase in income
.
There is a reverse effect with age; as the age of the respondent increases

by one year
, the
annual crop income decreases by 55,407 Tsh. or $35.16.
Additionally, families that irrigated
faced
an increase in food security and diarrhea prevalence.


These rela
tionships show the huge
potential monetary and food security benefits from an expansion in irrigation, but
additionally

the
need for sanitation and public health
campaign
.


Dependent Variable

Independent Variable


t stat

p value

Coefficient

Number of
meals

Irrigation

2.24

0.0285

0.3746

Diarrhea

Irrigation

1.81

0.0748

0.2235

Annual income

Irrigation

2.73

0.00802

1,7
8
1,725.23

Annual income

Age

-
2.20

0.0317

-
55,407.37


In regard to other health
-
related data, malaria prevalence regressions were not possible due to
the
high incidence of the disease in the

sample size; out of the 80 households surveyed, 76
households reported malaria as a “common illness/disease in their ho
usehold.”
Thus, with
such low variation in the data, it became inaccurate to analyze anything from the regression.


G
eospatial analysis:

Through the geospatial analysis, a number of trends emerged related to the
formerly mentioned correlations.

E
xamining

correlations between household

location

and
their responses to survey questions
,

it was noted that household
position
was correlated with
whether or not they

went to bed hungry (z score=2.26)
.

S
trong correlations were
also
noted
with questions asking if
they felt their water was safe (z=6.26), if they irrigated (z=
9.32
), and if
anyone in the household experienced diarrhea (z=2.
43
). These trends ec
h
o correlations seen
through regressions
,
substantiating
them geospatially

(
Fig.

4
,
5
)
. In addition to these
responses, the expected differences between Masai

(Katahe)

and Pare

(Gunge, and Mabilion)

cultures wer
e seen. Th
e Masai
own
substantially more livestock than the Pare
, 24.8 animals
compared to
the
4.5

average animals per household
, respectively
.



Variable

z value


p score


I猠獯楬⁡⁣桡汬敮来⁦sr⁹潵⁴漠 row⁣ 潰o
?”

4.87

0.000001

“Does your
晡f楬i

irrigate?”

9.32

0.000000

“How many times per month does your household
irrigate?”

6.61

0.000000

“D
楤⁹潵r⁦慭楬y⁵獥⁰敳e楣id敳⁩渠n桥⁰慳琠祥慲
?”

5.88

0.000000

10


“How many chickens does your
family

have?”

-
2.07

0.036718

“Does your household use agroforestry as a farming
technique?”

2.07

0.038645


Do you believe the water is safe to drink directly
from borehole or stream
?”

6.26

0.000000

“How much
water does your household use per day?”

-
2.01

0.044121


How often was there no food to eat of any kind in
your household
?”

2.74

0.006051

“Has anyone in the house gone to bed hungry in the
past month?”

2.26

0.002377


Is diarrhea a common disease/symptom
in your
household?”

2.43

0.014975

“Did you or any household member go a whole day
and night without eating anything because there was
not enough food?”

2.53

0.011368

Yearly income (calculated by yields and commodity
local market value) / land under
cultivation

3.35

0.000804



NDVI
:

Through analysis of Landsat
-
derived NDVI

images from 1987 to 2011, researchers were
able to look at land cover change in the region.
Over the 23 year period, t
he areas that
experienced the most change were in
lowland
ditches and washouts, particularly near Gunge

(
Fig

5)
. These areas contain very little vegetation and sandy soils, increasing their
vulnerability. Additionally,
there were still yearly changes in the vegetation density in the
irrigated region, signaling
room for improvement in agricultural practices in the specified zone
(Fig 6).


In a year
-
to
-
year comparison, it was
noted that more change occurred
between

the years of
2002 to
2011 than

between 1987 and
2002. This could be attributed to changing precipit
ation
patter
n
s, or increased human resource use. The changes noted most during the
2002
to

2011
period were near the
South
Pare Mountains, and near the towns of Gunge and Katahe. During
the qua
litative

survey
s

in which

F
GD
s
,

hous
e
holds were asked where
they harvest wood for
cooking
. M
ost
respondents
reported taking from around the villages
previously,

but currently

they are

being forced to harvest from near the mountains due to sparse growth
and biomass
degradation
near their homes. All year NDVI compa
risons yielded little change on the
southern side of the Pangani river.









Image 3

11
















































Figure 3

ArcGIS Hot Spot Analysis: examining spatial distribution of income per hectares of land in production

ArcGIS Hot Spot Analysis:
representing clusters reporting how often they “go to sleep hungry”


Figure 4

ArcGIS Hot

Spot Analysis: examining spatial distribution of beliefs that the river water is safe

without treating

Figure 5

12




























NDVI means per zone from 1987
-
2011. Note inconsistent year spacing

Figure 6

Figure 5

NDVI image for January 2000 : red represents the lowest vegetation density while green represents
higher density.

13



Discussion


With the in
creasing
strains on the region

both environmental and social

the incorporation
of programs and policies t
argeting key weaknesses
,

and

supportable strengths will be needed.
T
hrough the various forms of
data
analysis
,
a number of trends
have
become clear. All of
the
data eval
uation systems
demonstrated

the strong influence

that irrigation can play

(positive
and negative)

if access
becomes
available. Additionally, regional use of wood
as a fuel source is
beginning to create potential for desertification in the reg
ion. Researchers also noted
knowledge gaps regarding health and sanitation issues pertaining to participants’ context
, as
well as possible endemic
nutritional
deficiencies
.


Through the regressions, it is clear that irrigation can increase agrarian househ
old incomes. In
fact, in this region
,

if a family was able to irrigate
,

they were able to make almost
twice

the
national average household incomes (
$1,375
)
.
Moreover
, through the FGDs
,

it became obvious
that many of the households that irrigated were able to sustain their livelihoods
through

duri
ng periods of drought. One woma
n

who was able to irrigate

in the Gunge region

described
her ability to charge more for her grain during times
of drought due to a regional grain
shortage. In correlation with the
increase of
household income

from irrigation
,

the data shows
that families are also able to
eat more meals per day, curbing some of the local
hunger
.

Spatial
regressions also showed this with most of the Gunge households that irrigate also
reporting
that

they did not go to bed hungry.


The data revealed that households who were able to irrigate also faced
higher

diarrhea
prevalence. It is hypothesized

that this is due to the increased
possibility

of those irrigating
to
be exposed to untreated water and
for
the water to come in contact with their food.


Further water management practices and distribution expansion will be needed in the region.
There wi
ll be more hydrological research needed regarding the

longevity
of Pangani

River
sources as well as quality assessments. Water harvesting techniques have been shown to be
economically effective in the region, and could serve
as an effective intervention

i
n the Hedaru
Valley

(Mutabazi K. D. 2006)
.


In parallel with regional water problems, t
he most prevalent diseases/symptoms in the
surveyed region were malaria (95%),

diarrhea (40%), respiratory issues (23%), and chest
pain (14%). Of those, malaria and diarrhea
are strongly related to water quality and practices.
They have also been shown to decrease in prevalence

with public health education and
distribution of basi
c supplies such as
i
nsecticide
-
t
reated n
ets (ITNs)
. Of the NGOs working in
the area, there are very few working specifically in public health education.

Four to five years
prior to this research, an NGO held educational seminars in th
e region particularl
y aimed at
promotin
g the positive practice of making a latrine near each house to hold fecal waste.
During the data collection in 2012, nearly 95% of households had a latrine near their home.
Unfortunately, this practice has not completely eradicated dia
rrhea or other communicable
diseases, but it does demonstrate opport
unities for knowledge transfer.


14


The holistic resource analysis techniques used for this research are cost effective, transferable,
and show an ability to demonstrate regional development
constraints. Through implementing
these participatory stakeholde
r
-
guided methods in other communities and regions in
developing countries, development practitioners will be able to develop more comprehensive
a
nd sustainable development programs. Further
research
is

needed with
an assurance of
accurate

methodology us
ed. H
owever data correlations and correlations between all forms of
analysis are positive.























15


Works Cited


Bakr N., Weindorf D.C., Bahnassy M.H., Marei S.M. El
-
Badawi M.M.,. "Monitoring land cover
changes in a newly reclaimed area of Egypt using multi
-
temporal Landsat data."
Applied
Geography
, 2010: 592
-
605.

Battisti D., Naylor R. L. "Historical warnings of future food insecurity with unprecedented
seasonal heat."
Science
, 2009: 240
-
246.

Brunsdon C., Fotheringham S., Charlton M.,. "Geographically Weighted Regression
-
Modelling
Spatial Non
-
Stationarity."
Journal

of the Royal Statistical Society. Series D
, 1998: 431
-
443.

Burke M B, Lobell DB, and Guarino L. "Shifts in African crop climates by 2050, and the
implications for crop improvement and genetic resources conservation."
Global
Environmental Change
, 2009: 317
-
325.

Chirowodza, et al. "Using participatory methods and Geographic Information Systems (GIS) to
prepare for an HIV community
-
based trial in Vulindlel, South Africa."
Journal of
Community Psychology
, 2009: 41
-
57.

Christensen J. H., et al.
Regional Climate

Projections Climate Change: The Physical Science Basis
(Contribution of Working Group I to the Fourth Assessment Report of Intergovernmental
Panel on Climate Change).

IPCC Report, Cambridge: Cambridge University Press, 2007.

Cohen W., Goward S. "Landsat's

Role in ecological applications for remote sensing."
BioScience
,
2004: 535
-
545.

Coppin P., Jonckheere I., Nackaerts K., Muys B., Lambin E.,. "Didgital change detection methods
in ecosystem monitoring: a review."
International Journal of Remote Sensing
, 20
04:
1565
-
1596.

Dunn E. "Participatory GIS
-

a people's GIS?"
Progress in Human Geograhy
, 2007: 616
-
637.

ESRI. ArcGIS 10.0, spatial analyst. Redlands, CA, USA. 2013.

Food and Agriculture Organization.
Ferralsols.

Rome: FAO, 2011.

Food and Agriculture Organi
zation.
The State of Food Insecurity in the World.

Rome: FAO, 2008.

Geerken R., Batikha N., Calis D., Depauw E.,. "Differentiaion of rangeland vegitation and
assessment of its status: field investigation and MODIS and SPOT VEGITATION data
analyses."
International Journal of Remote Sensing
, 2005.

Geerken R., Ilaiwi M.,. "Assessment of rangeland degradation and development of a strategy for
rehabilitation."
Remote Sensing of the Environment
, 2004: 490
-
504.

Jensen J. R. "Introductory digital image proces
sing
-
A remote sensing perspective (3rd ed.)."
544. Upper Saddle River, NJ: Prentice Hall, 2004.

Kinoti, Jeniffer. "Estimation of spatial
-
temporal rainfall distribution using remote sensing
techniques: a case study of Makanya catchment, Tanzania."
Internati
onal Journal of
Applied Earth Observation and Geoinformation
, 2010: 290
-
299.

Mbonile J. "Population, Migration, and Water conflict in the Pangani River Basin, Tanzania."
Environmental Change and Security Program
, n.d.

Michalak W. Z. "GIS in land use change

analysis: integration of remotely sensed data into GIS."
Applied Geography
, 1993: 28
-
44.

Mutabazi K. D., Senkondo E. E., et al.,.
Economics of Rainwater Harvesting for Crop Enterprises in
Semi
-
Arid Areas: The Case of Makanya Watershed in Pangani River Bas
in, Tanzania.

Nairobi: Soil and Water Managment Research Network, 2006.

16


National Bureau of Statistics.
Tanzania 2004
-
2005 Demographic and Health Survey Key
Findings.

Dar es Salaam: The Rebublic of Tanzania, 2005.

National Bureau of Statistics, Tanzania Min
istry of Planning, Economics and Empowerment.
Tanzania Census 2002 Analytical Report.

Dar es Salaam: The Republic of Tanzania, 2006.

National Research Council Staff.
Down to Earth: Geographic Information for Sustainable
Development in Africa.

Washington, DC: National Academies Press, 2002.

Pangani Basin Water Board,.
Pangani River Flow Assessment.

Nairobi: International Union for
Conservation of Nature, 2010.

Sadek S. H. A. "Use of Landsat imagery for monitoring agricultural expansion of east
and west
Nile Delta, Egypt."
Egyptian Journal of Soil Science
, 1993: 23
-
24.

Satterthwaite, David. "The Implications of Population Growth and Urbanization for Climate
Change."
Environment and Urbanization
, 2009: 545
-
567.

Scepan J., Menze G., Hansen M. C.,.
"The DISCover Validation image interpretation process."
Photogrammetric Engineering and Remote Sensing
, 1999: 65
-
69.

Turpie J. K., Ngaga Y. M., Karanja F. K.,. "Preliminary economic assessment of water resources
of the Pangani River Basin, Tanzania." Dar e
s Salaam, 2005.

United Nations Economic and Social Affairs.
World Population to 2300.

New York: United
Nations Press, 2004.

United Nations.
The future we want.

Outcome Document, Rio de Janiero, Brazil: UNCSD, 2012.

University of Amsterdam, IBED.
Post
-
class
ification vegitation change detection using NDVI.

Amsterdam: University of Amsterdam, 2011.

Washingron R., et al,. "African climate change: taking the shorter route."
Bulletin of the
American Meteorological Society
, 2006: 1355
-
1366.

World Health Organizati
on.
Report on a regional workshop looking at monitoring approached in
drinking water and sanitation.

Dar es Salaam: UNICEF, 2011.