Tải bản đầy đủ - 0 (trang)
Measuring Habitat Changes in Barrier Island Marshes: An Example from Southeastern North Carolina, USA

Measuring Habitat Changes in Barrier Island Marshes: An Example from Southeastern North Carolina, USA

Tải bản đầy đủ - 0trang

392



J.N. Halls



coastal ecosystems (Benfield et al. 2005, Conway 2005, Phinn and Stanford 2001,

Ramessur 2002). The purpose for this chapter is to describe methods for mapping

and analyzing barrier island salt marshes.



17.1.1 Population Growth in Coastal North Carolina

Population growth and decline varies throughout the United States. One popular

method for analyzing the changing population is the national Census of Population

which occurs every 10 years. From 1980 to 1990, the urban areas of the South,

West, and coastal Northeast gained population while large Midwestern cities and

rural areas substantially declined. From 1990 to 2000, population change was still

largest along the coasts, but the non-coastal cities of Las Vegas, Phoenix, Dallas,

and Chicago also grew substantially.

In North Carolina, population growth has been steady, as reflected throughout the

southern United States. However, across the state there are spatial patterns of population decline in the rural areas, large population growth in the largest urban centers of Charlotte and Raleigh, and population growth along the southeastern coast

(Fig. 17.1). Along the southeastern coast, Wilmington has experienced rapid growth

in the city, bedroom communities, and surrounding beaches (Fig. 17.2). The urban

area has spread from the City of Wilmington to include a majority of the surrounding counties spreading along major transportation routes and the coastal retirement

destinations. The study area chosen for investigation is a typical developed barrier



Fig. 17.1 North Carolina change in population, by Census Block Group, from 2000 to 2005



17



Measuring Habitat Changes in Barrier Island Marshes



393



Fig. 17.2 Southeastern North Carolina change in population, by Census Block Group, from 2000

to 2005



island located in southeastern North Carolina. Topsail Island was first developed primarily with vacation homes with few year-round residents. However, the population

has steadily increased over time and currently has a much larger resident population

comprised of retirees and working professionals (Fig. 17.3).



Fig. 17.3 Location of Topsail Island, North Carolina. The island straddles two counties, Onslow

and Pender, which can lead to difficulties in locating comprehensive aerial photography. The portion of the island that was studied is highlighted in a series of 0.5 km width boxes from the Surf

City bridge south to New Topsail Inlet



394



J.N. Halls



17.1.2 Coastal GIS and Remote Sensing

Each local planning agency is GIS savvy and has developed several data layers

for monitoring growth and mapping infrastructure. For example, the City of Wilmington and New Hanover County have an Internet mapping website for people to

visualize a variety of data and also a link to download data for use by other GIS

users. Both Brunswick and Pender counties also have a planning and GIS department where they report to the county commissioners on growth, zoning, infrastructure and other mandated activities.

Other than local government GIS data, remote sensing imagery is another possible data source for investigating land use/land cover change. For example, one

popular satellite system is Landsat (the older Multispectral Scanner and the more

recent Thematic Mapper and enhanced Thematic Mapper) which is useful for mapping both urban and natural areas. There are many types of commercial satellite

and airborne imaging sensors, but the most popular in the United States are Landsat

and SPOT. SPOT, a French satellite, is most appropriate for mapping urban areas

because it has a relatively high spatial resolution (color is 20 m by 20 m cell size

and black and white is 10 m by 10 m cell size) which is needed when discerning

urban objects such as small buildings, roads, etc. Landsat (such as Landsat 5 and 7)

is better for identifying natural habitats because of its ability to discriminate various types of vegetation, although the spatial resolution is coarser (30 m by 30 m cell

size) (Alphan and Yilmaz 2005, Donoghue and Mironnet 2002, Phinn and Stanford

2001, Shi et al. 2002, Ucuncuoglu et al. 2006, Vanderstraete et al. 2006). There

is also aerial photography which is an excellent source of data that enables more

detailed mapping and can provide a longer historical record than satellite imagery

(Al-Bakri et al. 2001, Higginbotham et al. 2004, Jones 2006, Zharikov et al. 2005).

However, with the large scale of aerial photography this translates to small cell size

and consequently much more data. So, aerial photography is usually analyzed when

the study area is relatively small.

To investigate how satellite imagery and aerial photography can be utilized for

mapping land cover change, two studies were conducted where Landsat 5 imagery

was used to investigate change in New Hanover County and aerial photography was

used to map changes in marsh habitats at Topsail Island.



17.2 Physical Geography of Barrier Islands

Barrier islands and coastal salt marshes are complex ecosystems that move and

change through time in response to many factors. For example, hurricanes bring

strong winds, rain, and storm surge which can greatly change the distribution of

surficial deposits (Nordstrom et al. 2006). Through time the islands can migrate and

inlets change their positions.

There are many reasons for investigating how back-barrier marsh systems change

through time. For example, they provide protection for the mainland during storms



17



Measuring Habitat Changes in Barrier Island Marshes



395



by absorbing the tidal surge and providing a stabilizing environment for storm overwash. These environments are also economically and environmentally important

ecosystems because they provide fish nursery habitats, bird nesting and foraging

sites, and act as a filter for chemicals entering the ecosystem. In the southeastern

United States salt marshes are typically found in tandem with barrier islands. This

ecosystem includes the beach, dunes, vegetated zones, maritime forest, swampy terrains, tidal flats, and low-lying salt marshes (Bates and Jackson 1984). Researchers

have identified several factors related to marsh stability: geomorphology, elevation,

vegetation, hydrologic conditions, frequency of tropical storms, tidal range, and sediment supply (Goodbred and Hine 1995, Davidson-Arnott et al. 2002). If estimates

are correct and sea level rise is increasing at 1.9 cm/year (Davis 1994), then the salt

marshes in this region require a substantial amount of sediment, either from overwash or other transport mechanisms, to sustain their existing size.

In addition to the geologic and geomorphic processes of marsh formation, there

has been a steady increase in coastal development along all coasts of the United

States (Titus 1990) and it is yet to be determined what impact this urbanization

has on back barrier marshes (Bertness et al. 2004). Therefore, a study was undertaken to map back barrier marshes in order to quantify change as well was

compute various spatial measures to identify patterns in how these marshes have

changed through time. To understand how barrier island marsh habitats change

through time Topsail Island, located in southeastern North Carolina, was investigated. The island is part of a chain of barrier islands in the geologic system

known as the Georgia Bight. Topsail Island is a 30km barrier island that was

initially used as a military rocket testing site in the 1940s and is now primarily single-family vacation homes but has an increasing population of year-round

residents.



17.3 GIS Database Development

It has become quite common to utilize the tools available in remote sensing and

GIS software for mapping coastal habitats such as salt marshes (Zharikov et al.

2005, Dech et al. 2005, Jupiter et al. 2007). For Topsail Island, the GIS development began with a detailed survey of all local, regional, state, and federal agencies

that commonly acquire aerial photography. Many dates of photography were identified, however only those dates where photography covered all of the back barrier marshes, were of similar scale (1:12,000 and 1:20,000), were at similar tidal

stages, and same time of year were used in the study. The most recent photography

was 1998, it was already rectified into orthophotography, and was near-infrared.

All other years (1938, 1949, 1956, 1971, and 1986) were in analog format which

required scanning and rectifying. After several tests at varying resolutions, it was

determined that scanning the aerial photographs at 400 dpi was sufficient for the

scale, interpretation and digitization of the marshes.



396



J.N. Halls



17.3.1 Interpretation and Digitizing

Once in a GIS, each of the photographs were interpreted, digitized and then compared. Although the scanning produced pixels with 1 m spatial resolution, the scale

of the photography dictated the minimum mapping unit based on standards developed by the US Geological Survey. Therefore, the aerial photography was precisely

mapped where the smallest marsh polygon interpreted and digitized was less than

0.1 hectare. The land cover classification scheme was: marsh, upland, water, and

barrier island (Fig. 17.4). To aid in the interpretation process, field work was conducted where hundreds of sites were visited and a comparison was made between

the real land cover types (predominantly wetlands) and the photography. The field

work was imperative for becoming comfortable interpreting the imagery. Although

photointerpretation and digitizing are labor intensive and time consuming, several

image processing classification techniques were tested but did not produce acceptable results. For example, unsupervised and supervised analyses were tested using

a variety of cluster algorithms and training site selection trials. Unfortunately, the

image processing algorithms were not able to distinguish marshes from water with

any consistency because the photography had little spectral variety. Perhaps future

tests using object-oriented classification rather than pixel spectral analysis will yield

improved habitat mapping (Laliberte et al. 2004, Lathrop et al. 2006).



17.3.2 Accuracy Assessment of Photointerpretation and Digitizing

Performing an accuracy assessment is an important part of any change detection

analysis or other type of temporal spatial analysis (Couto 2003; Hughes et al.

2006). After all of the photographs for the 6 years were digitized and checked for

logical consistency and topological correctness an accuracy assessment was conducted where 140 points were randomly located in the study area and the digitized

land cover classes were compared to the aerial photography. Using an error matrix, an overall accuracy greater than 80% was computed for each year which was

acceptable.



17.4 Change Detection

The back barrier marsh habitats of Topsail Island changed from 1938 to 1998, but

the changes were not systematic across the study area (Fig. 17.4). When summarizing the total area of marsh, it steadily decreased from 1938 to 1998 (Fig. 17.5).

In fact, by simply calculating regression statistics between marsh area and time,

there was a strong negative linear relationship between the area of marsh and time

(y = −2.112x + 4739, R2 = 0.961). However, these summary results do not fully

describe how the marsh has changed through time. There are several analytical



17



Measuring Habitat Changes in Barrier Island Marshes



397



Fig. 17.4 Topsail Island interpreted and digitized habitat maps in (a) 1938, (b) 1949, (c) 1956, (d)

1971, (e) 1986, and (f) 1998



398



J.N. Halls



Fig. 17.5 Area of marsh from 1938 to 1998



techniques that were employed to further investigate how the Topsail marshes have

changed including change detection analyses and landscape fragmentation.

To compare how the land cover changed from one time period to the next transition matrices, also referred to as cross-tabulation matrices, were created. The technique used in this study is known as the post-classification comparison where the

input data layers have been independently classified/interpreted and then the results

are compared, or overlaid (Jensen 1996). Using this approach, classification matrices document how each land cover class changed from one time period to the next

(Table 17.1). The diagonal cells (shown in grey) contain the area (in hectares) that

did not change from time 1 to time 2 and conversely, the off-diagonal cells document the area that changed from time one to time two and how the area changed.

So, the diagonal values can be considered persistent whereas the off-diagonal values document the areas that have transitioned from one class to another. Although

documenting the area of change is useful, it is best to represent this change in percentages so that further statistical analyses can be computed (Table 17.2). The total

percentage column is the summation of each habitat row in time 1 and likewise

the total percentage row is the summation of each habitat column in time 2. The

percentage of habitat lost is the summation of the off-diagonal row values and the

percentage of habitat gained is the summation of off-diagonal column values. The

total net change is the difference between the total in time 1 and total in time 2.

However, the net change does not describe how the habitats have persisted versus

changed, or transitioned, to another habitat type.

The change detection cross-tabulation matrices revealed that 71% (22.47/31.75)

of the marsh in 1938 remained marsh in 1949; this dropped to 65% in 1956, 58% in

1971, 65% in 1986 and 73% in 1998. To visualize how the marshes have changed

through time, we can track the persistent marsh (marsh that remained from time 1

to time 2), the marsh that didn’t remain (lost marsh) and new marsh (marsh that was

gained) (Fig. 17.6). Although it is interesting to map the persistent, gained, and lost

marsh there isn’t a clear spatial pattern. From a vulnerability standpoint, it would

be best if we could clearly decipher a pattern of marsh loss versus gain (Fig. 17.7).



17



Measuring Habitat Changes in Barrier Island Marshes



399



Table 17.1 Cross-tabulation matrices for each date of aerial photography (in hectares)



1938



marsh

marsh

upland

water



449.2

45

118.1



1949



marsh

marsh

upland

water



402.6

42.6

143.9



1956



marsh

marsh

upland

water



338.8

47.6

192.9



1971



marsh

marsh

upland

water



378.3

38.3

134.8



1986



marsh

marsh

upland

water



405.6

17.4

80.6



1949

upland

28.6

215.7

17.6

1956

upland

36.5

202.9

24.3

1971

upland

53.9

180.2

28.7

1986

upland



water

156.8

20.7

947

water

176.7

13.5

961.6

water

192.2

33.5

922.9

water



49.7

151.8

205.1

17.3

17.1 1,003.70

1998

upland

water

34.4

114.1

239.9

14.6

10 1,087.20



However, there clearly isn’t a consistent spatial pattern to marsh loss other than that

there is more loss than gain. So, further analysis into how the habitats are changing

is necessary in order to more clearly understand the changing landscape.

As can be seen in the cross-tabulation matrices (Table 17.2), the amount of persistence in each habitat class far exceeds the amount of change from one habitat class

to another (the diagonals are larger than the off-diagonals), but this is to be expected

in change detection studies. So, although the percentage of persistence is greater

than the percentages that have transitioned to other classes, it is important that we

investigate these transitions in order to determine which transitions are creating the

greatest impact to the landscape. In Table 17.2, the conversion of marsh to water was

an average of 7.9% over the entire study area. The next largest transition was water

converting to marsh (averaging 6.71%), but the largest transitions don’t necessarily

mean these are the most important, or indicative, of how the landscape has changed.

To investigate the habitat change further, there is another technique that can measure how the landscape has changed by quantifying the amount of habitat that has

changed to another class, also known as the amount of class swapping (Pontius et al.

2004). The equation to calculate the amount of swapping is:

S j = 2∗ min(gain, loss)



(17.1)



400



J.N. Halls



Table 17.2 Cross-tabulation matrices for each date of photography (in percent)



marsh

upland

water

total 1956

Gain



marsh

20.08

2.13

7.18

29.39

9.30



marsh

upland

water

total 1971

Gain



marsh

17.02

2.39

9.69

29.10

12.08



marsh

upland

water

total 1986

Gain



marsh

18.95

1.92

6.75

27.62

8.67



marsh

upland

water

total 1998

Gain



marsh

20.24

0.87

4.02

25.13

4.89



1938



marsh

upland

water

total 1949

Gain



marsh

22.47

2.25

5.91

30.63

8.16



1949



1956



1971



1986



1949

upland

1.43

10.79

0.88

13.10

2.31

1956

upland

1.82

10.12

1.21

13.15

3.03

1971

upland

2.71

9.05

1.44

13.20

4.15

1986

upland

2.49

10.28

0.86

13.62

3.35

1998

upland

1.72

11.97

0.50

14.19

2.22



water

7.85

1.04

47.38

56.26

8.88

water

8.81

0.67

47.97

57.46

9.49

water

9.65

1.68

46.36

57.70

11.34

water

7.60

0.87

50.28

58.75

8.47

water

5.69

0.73

54.26

60.68

6.42



total

1938

31.75

14.08

54.17

100.00

total

1949

30.72

12.92

56.36

100.00

total

1956

29.38

13.13

57.49

100.00

total

1971

29.05

13.06

57.89

100.00

total

1986

27.65

13.57

58.78

100.00



Loss

9.28

3.29

6.79



Loss

10.64

2.80

8.39



Loss

12.36

4.07

11.13



Loss

10.09

2.79

7.61



Loss

7.41

1.60

4.52



where, S j is the amount of swapping of class j and gain and loss are the percentages

of greatest gain and largest loss from class j to all other classes. Table 17.3 contains

the overall percentage change, the percentage of gain and loss, the total change (gain

plus loss) and the amount of swapping among classes. The following conclusions

can be drawn from this table:

• The net change reveals the overall loss of marsh habitat in all time periods.

• Water and marsh change more than upland.



17



Measuring Habitat Changes in Barrier Island Marshes



401



Fig. 17.6 A subset of the Topsail study area showing marsh that has persisted, gained, and lost for

(a) 1938 to 1949, (b) 1949 to 1956, (c) 1956 to 1971, (d) 1971 to 1986, and (e) 1986 to 1998



• Although the overall percentage of marsh area has decreased through time

(Fig. 17.5) there has been a much greater percentage of marsh swapping (gain

and loss) across the study area.



17.4.1 Observed Versus Expected Change

Although it is customary to describe how much change has taken place between

time periods, a further investigation into how much change is significant can

be accomplished by calculating the difference between expected and observed

change (Pontius et al. 2004). To determine the importance of the off-diagonals (or



402



J.N. Halls



Fig. 17.7 A portion of the study area showing (a) all marsh areas lost over each time period and

(b) marsh areas that were gained in each time period. Although there was more marsh lost than

gained, there is no clear spatial pattern to the gains and losses



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

Measuring Habitat Changes in Barrier Island Marshes: An Example from Southeastern North Carolina, USA

Tải bản đầy đủ ngay(0 tr)

×