Mapping Chaparral
In Perth, Santa Monica
and Cape town
From Space
Jawhara Tariq
Ryan Haynes
Nilu Taeed
Prof. Gillespie
UCLA Department of Geography
Introduction
Chaparral is shrubland found within California, South Africa, Mexico, and Australia. The plants have small and hard leaves that hold moisture, and some of the plants include shrub oak, Yucca Wipple, and poison oak (blueplanetbiomes). The name “chaparral” is a Spanish word referring to scrub oak. This vegetation primarily inhabits Mediterranean regions ranging from flat terrain to mountain slopes (Doug). Most chaparral vegetation occurs regionally along the Pacific Coast to the mountain foothills from Southern Oregon to Baja California (RCIP, 18). The chaparral is dominant within Southern California (Biosbcc). The chaparral is composed of evergreen shrubs and occurs on dry and gravelly slopes with little soil or fairly heavy soils. Chaparral occurs in large continuous stands or in patchwork of other coastal vegetation habitats (RICP, 20). However, it must be noted that chaparral that occurs in mountains and high elevation are larger and more continuous while chaparral near the coast tends to be in disjoint patches (RICP, 20).
Slope,
elevation, soil type, and fire are the main geographical factors that influence
chaparral development (Biosbcc). The dominant force within chaparral is fire,
and most species are able to persist or have adapted as auto successional
species. Chaparral is a fire dependent community, and they are either
resprouters, obligate seeders or both within their respective communities
(RICP, 20). The position on a slope is a
selective factor in the chaparral environment, as it determines the exposure
the vegetative community will receive (Biosbcc). South facing slopes leave rise to less
diversity whilst north facing slopes diversity is high, because the mortality
rate of species is more frequent and inevitable (Biosbcc). Soil texture and structure
determines what type of plant life can be supported; chaparral soils are
typically on top of bedrock (Biosbcc).
There
is no distinct threat to chaparrals because they are widely distributed and
support various species. It has been
argued that fire suppression has been a threat to chaparral, as the chaparral
needs fire in order to be autosuccessional (RICP, 27). Conversely, despite the
fact that chaparrals are fire dependent, over-burning of these areas can lead
to its decline (California Chaparral). These ecosystems are highly sensitive to
fire, and because of the diversity and variety between chaparrals, there is no
set prescription to ensure that a specific community thrives. When chaparrals
cannot replenish quickly enough, their lack of existence leads way for
non-native grasses to invade the burned area, transforming the locations to
hubs of invasive flora species (California Chaparral). Furthermore, areas of dead chaparral can lead
to large fires that can cause a problem both ecologically and economically.
Using
satellite imagery, we compared chaparral sizes between three different areas
for 2000 and 2014. Within this fourteen
year window, we were able to conclude that chaparral has not only decreased in
some areas (most likely due to fire suppression), but areas have most likely
transformed into other vegetation. This is easy to be concluded because certain
vegetation that should have been chaparral was no longer picked up on the same
range in NDVI’s.
The Study
The
study included three chaparrals from the western coastal areas of Australia,
the Cape Town area of South Africa, and the coast of California, specifically
within the Santa Monica Mountains. The study will use Satellite Remote Imagery
to detect change in chaparral in the described areas between 2000 and
2014. The result will show the NDVI’s
and thermal map to show the change in vegetation for the areas during the
prescribed time. A change detection map
will also be computed in order to better depict the amount of change between
the chaparral ecosystems for our respective areas.
Methods:
The Study area was chosen because
of its large population near an area that has abundant Chaparral. It was
determined that two images would be used; both images would be about one decade
apart. So the extent of the Chaparral could be compared between two time
periods. Therefore images were taken from 2002/2003 and 2013/2014. It
was decided that both images would
be taken from the months during the dry season of the areas, in
hopes the images would not contain any cloud obstructions. For
Australia and South Africa, this meant February. For Santa Monica, images were acquired
during July and September respectively to avoid coastal cloud cover.
The images were
selected from the summer months also so there would be no influx of the
vegetation because of the possibility of rain. Therefore using the Google Earth Application the study site
was located and determined. Then using the USGS database Glovis, several
satellite images were acquired and downloaded, one from 2013/2014 and one from 2002/2003.
The 2013/2014 image was acquired from the Landsat 8 Satellite with the Operational
Land Imager and Thermal Infrared Sensor. The Spectral Resolution of Landsat 8
is 11 bands with: band 1 being Ultra Blue, Band 2 though 4 being visible, Band
5 being Near Infrared, band 6 and 7 are SWIR, Band 8 being Panchromatic, Band 9
is Cirrus, and Band 10 and 11 are thermal using the Thermal Infrared Sensor.
The Spatial Resolution for Landsat 8 is: band 1 though 7 and 9 have a
resolution of 30 X 30 meters, band 8 has a resolution of 15 X 15 meters, and
bands 10 and 11 have a resolution of 100 X 100 meters. For the 2002/2003 image the
Landsat 4, 5 Satellite was used; the Landsat 8 Satellite could not be used for
the 2002/2003 image because the Landsat 8 Satellite was not put into orbit until
2013/2014. Landsat 4,5 use the Multispectral Scanner. The Spectral Resolution of
Landsat 4, 5 is 7 bands. Band 1 though 3 are visible, bands 4 and 5 and
infrared, and band 6 is thermal. The Spatial resolution of Landsat 4, 5 is 60 x
60 meters for each band.
After
the images were acquired from the USGS database Glovis they were imported into
the computer program ENVI Classic. To see the extent of the chaparral in the city
of Perth an NDVI was created. A NDVI stands for Normalized Difference
Vegetation Index, an NDVI separates “greenness” from the image and displays
healthy vegetation (Jensen). Therefore using the red and near infrared band an NDVI was
created for both the 2002/2003 and 2013/2014 images. Since an NDVI creates a black and
white image, a density slice was overlaid on the NDVI transformation. Then it
was determined which color on the density slice was displaying the chaparral.
Therefore all of the other slices were deleted, and the map only displayed
chaparral and vegetation similar to it.
A
thermal image was then created for both the 2002/2003 and 2013/2014 image. For the 2014
image, which comes from the Landsat 8 Satellite. The thermal image uses the
band 10. The image was then classified using the ENVI color-mapping tool, using
the tool, the red thermal color spectrum was used to create the map. To create
a thermal map for the 2002/2003 image, band 6 was used. The same steps were taken to
create the thermal image as the 2013/2014 map.
A
change detection map was created to show the change in chaparral vegetation.
Taking the two NDVI’s images a change detection map was created, the map shows
the percent change of vegetation between the 2002/2003 and 2013/2014. Using the classification
tool the positive change in vegetation was pulled out of the image and
displayed. The change detection map shows the percentage of change within each
pixel of the map. Therefore a positive increase would show that a pixel’s
brightness increased while a negative change would show that the pixel
decreased in brightness.
Perth Australia
Perth Australia is located at 31 57’ .8 S and 115 51’ .32 E
in the Western State of Australia. Perth is considered Mediterranean climate
with dry hot summers during the months of December tell March, and wet cold
winters from June tell September. The vegetation that surrounds the city is
called Kwongan Scrub, which is similar to the Chaparral that can be found in
California (WWF.org).
 |
| Study Area, Google Earth 2014 |
Results:
 |
2003 NDVI- Green in Chaparral
|
 |
| 2014 NDVI Green Displaying Chaparral |
 |
| 2014 Thermal |
 |
| 2003 thermal- White hot, dark red cool |
 |
| Change detection map of 2003 and 2014 NDVI |
 |
| 2014 Unsupervised Classification- Green Chaparral |
 |
| 2003- Unsupervised Classification- Green Chaparral |
 |
| Change Detection showing areas with high increase in vegetation |
To determine the actual percent of change between the 2003 and 2014 map two unsupervised classification were created. The Classification was created using the two NDVI maps from above. After the classification was created they were edited so that chaparral would appear green on the classification. The two classifications where then entered in the basic tool called- Change Detection Statistics. To get the statistics the two bands that displayed Chaparral had to be matched and all other bands had to be removed. This then gave us a table showing the percent change in the pixels.
 |
| Chart showing Percentage Change in Change Detection Map |
Results:
When comparing the two NDVI images they reveal that there was an
actual increase in vegetation between the 2003 and 2014 images. The thermal images
from 2003 and 2014 did not reveal why there has been an increase, but gives
evidence of the changes of the city. The thermal map shows no real increase in urbanization. The thermal map shows a slight increase in
the heat at the downtown sector of the city. That in the 2014 map we can see an
increase in the heat coming from downtown, and a decrease in the suburb sector
near the south part of the image. The change detection map reveals that there
has been a positive increase in vegetation between 2003 and 2014, demonstrated
in red we can see the increase of vegetation. Separating the positive increase
from the change detection we can clearly see the areas that have experienced a
growth in vegetation. Using both change detection maps, the unsupervised
classification, and the change detection statics it can be determined that
there was actually a 32.75% difference between the 2003 NDVI and the 2014 NDVI.
By using the change detection maps we can determine that this was a positive
32.75% change in Chaparral. Meaning that that is more vegetation in 2014 then
there was in 2003.
Santa Monica
Study Area:
Chaparral is the dominant vegetation in the
Santa Monica Mountains. It is a highly important vegetation community to
California, and is most extensive in coastal areas. These communities are
important because it represents a fire hazard (Roberts. 2). My study area presents the Santa Monica
Mountains. They extend 73km from Oxnard alluvial plain to the west Los Angeles
River (Robert, 3) The coordinates are roughly 34◦07’13”N and 118◦55’54”W(NPS).
Mixed chaparral is found throught the Santa Monica Mountains on moist, north
facing slopes(NPS).
Results:
 |
| False Color Image 2003 |
 |
| False Color Image 2013 |
The above images are false color images. Urban and places emitting high heat come in whites to purples. Vegetation is green, and dry vegetation is orange. Chaparral in the area would come up on the above images in oranges-greens.
 |
| NDVI 2003 |
 |
| NDVI 2013 |
 |
| Change Detection between 2003 and 2013 |
 |
| Thermal Map 2003 |
 |
| Thermal Map 2013 |
When comparing the NDVI images, it appears that chaparral
has increased from 2003 and 2014.
However, research on chaparral indicates that reliance on satellite
imagery may be completely indicative of chaparral. It must be noted that by
nature of where the chaparral is located and its make-up, abundance changes
based upon climate and the time of year.
Therefore, it could be that these images I chose in particular were not the
most indicative of the chaparral during their respective years. When looking at
the two NDVI’s there are some similarities between the images. Chaparral
continues to exist within many of the same areas. The change detection map
gives indication of that. It shows some slight change in the east, in which
there is a lessening of chaparral in 2014. However, for the most part it gives
a positive change, or increase in chaparral for 2014. The thermal maps give insight into this as
well. It must be noted that darker areas of the thermal map represent lusher
vegetation and lighter areas represent areas with more heat, such as urban
spaces. It must be noted however, that there may have been something altered
from my original image. When one compares my RGB color images with my thermal
images, one would assume that my thermal images would have more dark areas
indicating the vegetation. However, this was not the case. This skewed my
results significantly.
Capetown
Study Area:
The city of Capetown, is a bio diversity hot spot as well as a unique Cape floristic region, the city has one of the highest bio diversity of any area in the world and an estimated 2,2200 species of plants confined to Table Mountain. It’s home to 19 different vegetation types, which are completely endemic to the city due to the fact the city is uniquely located at the convergence of various micro-climates.Urban growth and sprawl has covered much of these ecosystems over time.
 | ||
| The Great Escarpment and the area along the Southern coast that makes the Cape Floral Kingdom |
Figures:
 |
| Thermal Map, 2013 |
 |
| Thermal Map, 2002 |
 |
| Change Percentage Map from NDVI 2002-2013 |
 |
Color Map of Capetown, 2013 |
 |
NDVI 2002 |
Results:
Urban growth and sprawl has covered much of the ecosystems over time. When comparing the two NDVI images, it revealed that there was an actual decrease in vegetation between the year 2002 and 2013 map. The thermal heat maps from the year 2002 and 2013, reveal that there has been an increase in urban growth in the city. but also gives evidence of land coverage. The thermal maps shows an increase in the heated areas over time and growth near the city bowl( the natural amphitheatre-shaped area bordered by Table mountain, Lions Head, Signal Hill, and Devils Peak). When comparing the two red/blue/green color maps of the years 2002 and 2013, it very clear that the Urban Areas which are represented in the brighter yellow and red colors have drastically increased.
Urban growth and sprawl has covered much of the ecosystems over time. When comparing the two NDVI images, it revealed that there was an actual decrease in vegetation between the year 2002 and 2013 map. The thermal heat maps from the year 2002 and 2013, reveal that there has been an increase in urban growth in the city. but also gives evidence of land coverage. The thermal maps shows an increase in the heated areas over time and growth near the city bowl( the natural amphitheatre-shaped area bordered by Table mountain, Lions Head, Signal Hill, and Devils Peak). When comparing the two red/blue/green color maps of the years 2002 and 2013, it very clear that the Urban Areas which are represented in the brighter yellow and red colors have drastically increased.
Discussion
Cape town:
When mapping my vegetation I found it interesting to note it was difficult to distinguish my chaparral from other green vegetation. The main vegetation type of Western Cape is fynbos, meaning “fine bush”, which grows on nutrient-depleted, usually sandstone derived acidic soils and some evergreen , flowering shrubs. Fynbos makes up almost 80% of all the vegetation of the Cape Floral Kingdom, and it shares a striking similarity to Southern California's chaparral, however due to relatively poor quality of the soil, the shrubs are generally smaller and spaced farther apart, making it easier to move through than is the case with chaparral. “ Like our chaparral, this vegetation type is well adapted to fires, and even requires a cycle of fires of from six to forty-five years to prevent the plants from aging and degenerating, and allowing the invasion of thicket and forest. It is not necessary here nor will I attempt to delineate any differences between the various fynbos types or plant communities, but let it suffice to say that of the approximately 8,600 species in the Cape region, some 7,000 of them are present in the fynbos biome. Thus the relationship between the fynbos biome and the Cape Floral Kingdom is sufficiently great that they can be forgiven who associate the two as one ( Michael L Charters).”
When mapping my vegetation I found it interesting to note it was difficult to distinguish my chaparral from other green vegetation. The main vegetation type of Western Cape is fynbos, meaning “fine bush”, which grows on nutrient-depleted, usually sandstone derived acidic soils and some evergreen , flowering shrubs. Fynbos makes up almost 80% of all the vegetation of the Cape Floral Kingdom, and it shares a striking similarity to Southern California's chaparral, however due to relatively poor quality of the soil, the shrubs are generally smaller and spaced farther apart, making it easier to move through than is the case with chaparral. “ Like our chaparral, this vegetation type is well adapted to fires, and even requires a cycle of fires of from six to forty-five years to prevent the plants from aging and degenerating, and allowing the invasion of thicket and forest. It is not necessary here nor will I attempt to delineate any differences between the various fynbos types or plant communities, but let it suffice to say that of the approximately 8,600 species in the Cape region, some 7,000 of them are present in the fynbos biome. Thus the relationship between the fynbos biome and the Cape Floral Kingdom is sufficiently great that they can be forgiven who associate the two as one ( Michael L Charters).”
Perth:
There could be several reasons for the increase of
vegetation between 2003 and 2014. Several of these reasons are: drought, error,
or climate change. Error is a probable reason for the increase in vegetation
because the two images were taken from two different satellites and therefore
the quality of the images differs. Since the 2003 image comes from Landsat
4,5 and therefore is a lower quality then the 2004 image there is a greater chance for error. Another reason for
the actual increase in vegetation over the 11-year period could be because of
climate change. Since 1970 there has been a shift in the climate of Perth,
where there has been a gradual decrease in winter rain by 20% some years. While
winter rains decrease though, there has been an increase in summer rains, where
storms come in and provide large amounts of water during the dry season (IOCI
2002). This could explain an increase in vegetation, that if the climate is
shifting so there is a gradual increase in rain every summer, therefore we can expect more vegetation. Another explanation for the increase in vegetation
between 2014 and 2003 is drought. During 2001 tell 2012 Australia was considered
in a serious drought. The drought was officially announced over in 2012.
Therefore this would explain the increase of vegetation in 2014 because the
almost ten year drought was over (Cranston 2012).
Santa Monica:
Furthermore, because I am studying a more urbanized area, chaparral is more likely to change due to fire suppression. Chaparrals, being drier shrublands, often lead to fire. In an effort to stop this from spreading to urban Santa Monica, fire suppression and forced fires change the chaparral vegetation communities more often (California Chaparral). Increase in urbanization has caused fires to happen more often, allowing non-native grasses to occupy where chaparrals once more (California Chaparral). Because the chaparral already represents a community of several plants, it could be that the chaparral is not necessarily increasing as indicated in the NDVI, but rather changing. Nonnative grasses and weeds could be picked up in the NDVI, giving the appearance that chaparrals are increasing, but in actuality are being invaded by other species of grass. Despite the increase of chaparral, chaparrals in the Santa Monica Mountains could be experiencing a decline in overall health. The California Chaparral Organization refers to this as “conversion”. Furthermore, although satellite imaging can be highly indicative of plant make up in the Santa Monica Mountains, imaging and spectrometry may need to be improved in order to truly investigate chaparral (Roberts, et al, 20). Vegetation patterns are complicated by natural vegetation, land-use, and urbanization; making it difficult to understand the change of chaparral between 2003 and 2014 (Roberts, et, al 5).
Conclusion:
After comparing the images created for the three different cities: Perth, Santa Monica, and Capetown we can see that some Mediterranean regions are experiencing different effects on their native chaparral. We can see that there is a decrease in chaparral in South Africa's Capetown, but there is a positive increase in Santa Monica and Perth Australia. After comparing the thermal maps between the cities we can see that there has been a growth in urban development around Capetown, but Santa Monica and Perth urban development growth has been minmal. Therefore we can conclude that urban development is playing a huge role in the degradation of chaparral in developing cities like Capetown, but established cities like Perth and Santa Monica's degradation is minimal because the cities are current not expanding into the surrounding chaparral, according to the maps.
Works Citied
"CALIFORNIA CHAPARRAL INSTITUTE." Loss of Chaparral. N.p., n.d. Web. 20 Mar. 2014.
"Chaparral Biome." Chaparral Biome. N.p., n.d. Web. 20 Mar. 2014.
"Chaparral." Habitat Accounts 2 (2002): n. pag. Web.
Charters, Micheal. "Flora of Southern Africa." Flora of Southern Africa. N.p., 2003. Web. 20 Mar. 2014.
Cillers, Sarel S. "Urban Ecology In Cape Town: South African Comparisons and Reflections." Ecology and Society. Research Unit of Environmental Sciences Management, 2012.
"Climate Variability and Change in the South-west." Indian Ocean Climate Initiative (2002): 1-5. Print.
Cranston, Belinda. "Minister Declares End of Drought." The Age. N.p., 27 Apr. 2012. Web. 19 March 2014.
"Google Earth- Perth Subset." N.p.: n.p., 2014. N. pag. Print.int.
Jensen, John R. Remote Sensing of the Environment: An Earth Resource Perspective. Upper Saddle River, NJ: Pearson Prentice Hall, 2007. Print.
Roberts, D.a., M. Gardner, R. Church, S. Ustin, G. Scheer, and R.o. Green. "Mapping Chaparral in the Santa Monica Mountains Using Multiple Endmember Spectral Mixture Models." Remote Sensing of Environment 65.3 (1998): 267-79. Print.
Short, Nicholas M. "Remote Sensing Tutorial Page 3-4." Remote Sensing Tutorial Page 3-4. N.p., n.d. Web. 21 Mar. 2014.
"Wild World Ecoregion Profile." World Wildlife Fund. N.p., n.d. Web. 19 Mar. 2014.
