Land Cover, Land Use of two bioluminescent bays in

Land Cover, Land Use of two
bioluminescent bays in Puerto Rico
Undergraduate Research
By: Yadira Soto Viruet
Advisor: Fernando Gilbes Santaella, Ph.D
Outline
Introduction
z Objectives
z Study
St d Area
A
z Methodology
z Result
z Discussion
z Conclusion
z
Introduction
z
Land around Puerto Mosquito Bay in Vieques and
La Parguera Bioluminescence Bay in Lajas, Puerto
Rico are used to produce Land Use/ Land Cover
maps..
maps
z
Land use is defined as the use of land by humans
humans..
z
Land cover is designates only the vegetation
(natural or man made).
made).
z
Examples: urban structures,
structures forest,
forest agriculture
agriculture,
water resources, vegetations index and others.
z
The Land Use/Land Cover maps have the
capacity to illustrate the interaction between
human and the surroundings.
z
In Puerto Rico, general maps of Land Uses/Land
Cover were p
produced in 1999 for La Parguera
g
and Vieques Island.
Land Use/ Land Cover Maps, 1999
Linda Velez,1999
z
The Bioluminescence is the emission of light by
living organism (Pyrodinium bahamense).
bahamense).
z
Conditions aiding the accumulation of these
organisms
o ga s s are:
are
a
e: sshallow
a o bas
basin with low
o tidal
da and
a d
narrow mouth, surrounded by mangroves.
mangroves.
z
La Parguera Bioluminescent Bay has shown a
decrease in its bioluminescence by approximately
80%
80
% during recent years compared with Puerto
Mosquito Bay
Bay..
Objectives
z
The main objective of this research was produce
and compare
p
Land Use and Land Cover, specific
p
maps for the two bioluminescence bays
bays;; using
Remote Sensing and GIS tools.
tools.
Study Area
Adapted from Walker, 1997
IKONOS 1m
La Parguera Bioluminescence Bay
•The
Th geomorphology
h l
off
the area is the result of
deformation of
Limestone in early
Cretaceous (Glynn,
1973).
•The bottom sediments
are composed by fine
mud, marine plants and
mangrove leaves
leaves.
Scale 1,2000 (Volckmann, 1984)
Puerto Mosquito
q
Bay
y
Scale 1,2000 (USGS Maps,2001)
• The geology of Vieques consists in intrusive rocks with carbonate
sediments, the bottom sediments are composed by clay.
• QTu – Sedimentary deposits undivided Marine limestone.
Methodology
z
Environment for Visualizing Images (ENVI)
z
Geographic Information System (GIS)
z
IKONOS sensor provides spatial resolution at 1
meter and in multispectral mode, it provides
imagery at 4 m spatial in four spectral bands.
z
The images was displayed using all available
bands; red, green, blue (RGB) and infrared band
(IR).
(IR)
IKONOS, 1m
IKONOS,
1m
(Campbell, 2002)
(Campbell, 2002)
z
Supervised classification
Parallelepiped
z Minimum Distance
z Mahalanobis Distance
z Maximum Likelihood
z
z
C f i
Confusion
M t i method
Matrix
th d
z
z
using ground truth image and ROI’s.
H d l
Hydrology
information
information,
i f
ti , USGS publication
bli ti files
fil
Regions of Interest (ROI’s)
• The process of using known pixels identity to classify
unknown pixel identity.
identity
Index of Vegetation (NDVI)
NDVI= Infrared (IR)-Red (R) / Infrared (IR)+ Red (R)
Results
Maximum Likehood
Assumes that the statistics
for each class in each
band are normally
distributed and calculates
the probability that a given
pixel belongs to a specific
class.
(Campbell, 2002)
Minimum Distance
Uses the mean vectors
of each ROI’s and
calculates the
Euclidean distance
from each unknown
pixel to the mean
vector for each class.
(Campbell, 2002)
Parallelepiped classification
Ranges of values within
training data to define
regions within a
multidimensional data
space.
(Campbell, 2002)
Mahalanobis Distance
Mahalanobis Distance
classification is a direction
sensitive distance classifier
that uses statistics for each
class.
Minimum Distance
Mahalanobis Distance
Maximum Likehood
Distribution of Classes
Distribution of Classes %
L Parguera
La
P
Bioluminescence
Bi l
i
Bay
B
Distribution of Classes %
Puerto Mosquito Bay
1%
12%,
5%
11%
9%
55%
57%
27%
23%
Index of Vegetation
Darker green color represent more density of vegetation
Validation process
z
Accuracy is calculated by the sum of the number of pixels
classified correctly divided by the sum of all the pixels in the
entire ground truth classes.
z
Confusion matrix for La Parguera Bioluminescence Bay
z Ground Truth ROI’s, was 99%
z Kappa coefficient was 0.80.
z The Kappa coefficient is a measure of the proportional
improvement by the 36 classifier over a purely random
assignment
i
t off classes.
l
z
Puerto Mosquito Bay
z Ground
G
d Truth
T th ROI’s
ROI’ 90%
z Kappa Coefficient was 0.5
z
The confusion
f
matrix, using Ground
G
Truth Image, was 100%
%
and the kappa coefficient was 1 for both bays.
Discussion
z
Supervised classification, the analyst has control of
the selected classes depending of the purpose of
the research.
z
All supervised classifications, were tested to
d t
determine
i th
the b
bestt spectral
t l response ffor each
h
image and classification method.
z
For La Parguera Bioluminescence Bay the
percentage distribution were similar in the two
bays, except for urban or built up.
z
In La Parguera Bioluminescence Bay has more
interaction with anthropogenic activities than
Puerto Mosquito Bay
Bay..
z
It iis visible
i ibl to
t recognize
i
that
th t La
L Parguera
P
Bioluminescence Bay has less density of
vegetation than Puerto Mosquito Bay
Bay..
z
The geological characteristic take an important
role for the distribution and impute of sediments
that effect into the bay
bay..
Conclusion
z
The capacity to recognize and updates these land use
and land cover areas are important tools development of
priorities
i iti ffor th
the managementt and
d ffor ffuture
t
research.
h
z
The fact that the presence of the urban or anthropogenic
activities affects the bioluminescent bays, also contribute
to reduce the bioluminescent.
bioluminescent.
z
Human activities have a substantial impact on the
hydrology characteristic of the bay
bay, also the sediment
impute to the bay has been variable due to human
activities.
z
The destruction of vegetation and the transit of
boats can also affect the bay
bay.
z
For future research,
research is suggested used older
images and make supervised classification to
determine changes in time and vegetation.
z
The IKONOS sensor is a great tool to obtain
i
images
with
ith hi
high
h quality,
lit classification
l
ifi ti and
d iindex
d
of vegetation was possible using the spectral
information available in the images.
z
Webpage:
p g http://gersview.uprm.edu/website/
p g
p
Acknowledgments
z
z
z
I want to thank Dr. Fernando Gilbes Santaella for
support and help in working on this research.
Thanks William Hernández to p
provide the USGS
information and support me.
This p
project
j
was funded by
y University
y of Puerto
Rico, Sea Grant College Program.
References
z
z
z
z
z
z
z
z
z
z
z
Bawiec, J. W., 2001, Geology, geochemistry, geophysics, Mineral Occurrences and
Mineral Resource Assessment for Commonwealth of Puerto Rico. U.S. Geological
Survey OpenOpen- File Report p. 9898-38.
Campbell, J.A, 2000, Introduction to Remote Sensing, The Guilford Press. p. 319319-586
Connelly X.
Connelly,
X M.
M 1993,
1993 Batimetría de la Bahía Puerto Mosquito.
Mosquito Vieques,
Vieques Puerto Rico.
Rico
In: Walker, L.A., 1997, Populations Dynamics of Dinoflagellates in Two
Bioluminescent bays: Bahía Fosforecente and Puerto Mosquito, Vieques. M.S.
thesis, University of Puerto Rico, Mayagüez. p.51.
Foddy, G. and Curran P. 1994, Environmental Remote Sensing from Global to
Regional Scales. p. 238
Glynn, P.W. 1973. Ecology of a Caribbean Coral Reef. The Porites Reef
Reef--Flat
Biotope:Part 1. Meterology and Hydrography. Marine Biology.20: p. 297297-318
Green E., Alasdair E., Clark C. and J. Mumby,2000. Remote Sensing Handbook for
Tropical Coastal Management
Management. Unesco Publishing
Publishing. Coastal management source
books 3
Margalef, R., 1961. Hidrografía y fitoplancton de un área marina de la costa
meridional de Puerto Rico. Investigación Pesquera. 18:p.3318:p.33-96.
Seixas Guerrero,, C.E.,, 1988,, Patrones de distribución especial
p
y sucesión temporal
p
en poblaciones de dinoflagelados de la Bahía Fosforescente, Puerto Rico. Ph.D
tesis, Universidad de Puerto Rico, Mayagüez.88.
Volckmann, R. P., 1984, Geologic Map of the Cabo Rojo and La Parguera
Bioluminesce Bay Quadrangles, Southwest of Puerto Rico: U.S.G.S. Miscellanous
Geologic Invest.
Invest Map II-1557,
1557 scale 1:20
1:20,000.
000
Walker, L.A., 1997, Population dynamics of Dinoflagellates in two bioluminescent
bays: Bahia Fosforescente, La Parguera Bioluminescence Bay and Puerto Mosquito
Bay Vieques. V1. p. 51