Eelgrass and Seaweed Assessments at Alaska Peninsula-Becharof NWR

Progress Summary- March 2011

Kyle Hogrefe, U.S. Geological Survey-Alaska Science Center, 4210 University Drive, Anchorage, AK 99508, 907-786-7416, Fax: 907-786-7401, khogrefe@usgs.gov David H. Ward, U. S. Geological Survey-Alaska Science Center, 4210 University Drive, Anchorage, AK 99508, 907-786-7097, Fax: 907-786-7070, dward@usgs.gov Tyrone Donnelly, U. S. Geological Survey-Alaska Science Center, 4210 University Drive, Anchorage, AK 99508, 907-786-7176, tfdonnelly@usgs.gov Lucretia Fairchild, U. S. Geological Survey-Alaska Science Center, 4210 University Drive, Anchorage, AK 99508, 907-242-3432, lucretiafairchild@yahoo.com

Ron Britton, U. S. Fish and Wildlife Service, PO Box 277, King Salmon, Alaska 99613, 907-246-3339, ron_britton@fws.gov

2

Introduction Seagrasses play an essential role in the health of estuarine and coastal ecosystems

through their high productivity, stabilization and enrichment of sediments, and support of a

complex trophic web (Hemminga and Duarte 2000). These tidally inundated meadows provide

high-value ecosystem services compared with other marine and terrestrial habitats (Costanza et

al. 1997). They are often the dominant primary producers of coastal areas and any change to

their distribution may have implications for ecosystem functions (e.g. biochemical fluxes, trophic

tranfers, nutrient cycling) and diversity of species (Duffy 2006). Seagrasses are also among the

most threatened marine habitats (Short and Wyllie-Echevarria 1996, Orth et al. 2006) because of

their sensitivity to shifts in water clarity, temperature, and depth, hydrology and wave action

(Short and Neckles 1999). Therefore, it is important to monitor the health of these critically

important habitats.

Eelgrass (Zostera marina) is the dominant seagrass in Alaska and likely the predominant

marine macrophyte of coastal areas bordering the Alaska Peninsula and Becharof National

Wildlife Refuge (APBNWR) along the south side of the Alaska Peninsula. Anecdotal

observations over more than 600 km of coastline suggest that eelgrass is abundant in protected

shallow water embayments and inlets, and is likely an important reason for the rich diversity of

plant and animal species that characterizes the region. Nevertheless, we know little about the

status and trends in this population. Estimates of annual and seasonal variation in the abundance

and distribution of eelgrass are lacking, but are fundamental for protecting and understanding

threats (e.g., dredging, fishing activities, climate change) and the value (biological and

economical) of the resource to the region. As part of a broad scale program to inventory eelgrass

and test methodology to monitor long-term trends in the health of this seagrass along the coast of

the APBNWR, we assess distribution and abundance of eelgrass at Chignik Lagoon, a key

stopover site for migratory birds and an important commercial salmon fishery. Here, we report

on progress to map spatial extent of eelgrass and assess its abundance relative to the presence of

associated seaweeds and macro-invertebrates in June 2010.

3

Methods Eelgrass Mapping:

Landsat satellite imagery was downloaded from the USGS, Earth Resource and

Observation Science Center (EROS) to develop baseline maps for assessment of areal extent of

eelgrass in coastal waters adjacent to APBNWR. A Landsat ETM+ (Enhanced Thematic Mapper

Plus) image from August 28, 1999 was used to assess eelgrass in Chignik Lagoon and Mud Bay.

The image was acquired at a relatively low tide of + 0.15 m mean lower low water (mllw) as

determined from tidal predictions at Anchorage Bay, 8 km to the west of Chignik Bay. The

image was projected in UTM (Universal Transverse Mercator) Zone 4 North using the WGS 84

(1984 World Geodetic System) datum and had a spatial resolution of 30 m. After preprocessing,

we classified embayment areas into 3 major land cover types: eelgrass, sand/mud, and deep

water (Figure 1).

The image was preprocessed to calibrate for at-sensor radiance, correct for atmospheric

path interference and check for georeferencing accuracy. Radiance calibration was performed

using the ENVI 4.7 Landsat calibration tool following calibration factors and formulas

established in Chander et al. (2009). The images were corrected for atmospheric interference

using the “dark pixel subtraction” method (Chavez 1988). We verified the EROS georeferencing

by comparing the position of prominent landmark features in the image with the position of those

same features found on a NOAA nautical chart. We detected only a small (<1 pixel) offset

between the image and the chart, and no additional georectification was made. However, careful

field collection of ground control points using global positioning system (GPS) units may

improve the spatial accuracy of the imagery and any products.

Classification of the imagery was conducted using an unsupervised isodata clustering

algorithm to identify statistically separable spectral classes that were then used in a supervised

maximum likelihood analysis (Ward et al. 1997, Ozesmi and Bauer 2002). Visual interpretation

of the imagery and information from the 2010 field survey were used to choose isodata clusters

covering regions that provided the cleanest examples of the three major land cover types. These

clusters were then used to extract training data for a maximum likelihood classification that

assigned every pixel within the embayments to one of the three cover types. When considering

the full spectral range of the image, we found that eelgrass covered by optically deep water (> 1

m of depth) was misidentified as deep water over unvegetated substrate because of similarity in

4

spectral signals. Therefore we repeated the classification process, focusing on just the region

classified as deep water, to narrow the spectral range being analyzed and enhance differentiation

between submerged eelgrass and submerged unvegetated. This technique increased the eelgrass

coverage estimates by 409 ha in Chignik Lagoon and 44 ha in Mud Bay and the accuracy of

these estimates by 2% and 10 %, respectively.

Mapping accuracy of eelgrass extent was evaluated using the percent cover

determinations made during the 2010 field survey. For this exercise, we simplified estimates of

eelgrass percent cover to presence (> 5% eelgrass cover) or absence (< 5% eelgrass cover)

categories to approximate the cover needed to produce a spectral signal for eelgrass (Valta-

Hulkkonen et al. 2003). Finally, we created maps of eelgrass density (percent cover) and

abundance (standing crop) from the field survey data using the inverse distance weighted (IDW)

interpolation method. The IDW interpolation method applies the assumption that locations in

close proximity are more likely to be similar than those farther apart to create a raster surface for

an entire area from localized point data (Valley et al. 2005).

5

Figure 1: Landsat imagery used to estimate eelgrass spatial extent in Chignik Lagoon and Mud Bay (A) shown with land cover classifications and field survey points (B).

6

Field Surveys:

Abundance of eelgrass and macro-seaweeds, and the presence of selected macro-

invertebrate species were assessed in Chignik Lagoon and Mud Bay to a depth of about -2.5 m

mllw in June 2010. Although eelgrass grows deeper in the subtidal, this restricted sampling area

encompassed the majority of the distribution of eelgrass in these embayments. We used a point

sampling approach with a systematic random design, where points were distributed across

subtidal and intertidal areas (Figure 1). This design allowed for a proportional assessment of

cover types and related physical attributes (see below) within each of the embayments. Points

were located in the field by boat using a GPS unit and sampled by snorkeling in dry suits during

high tide. At each point, we estimated water temperature (surface) and depth, salinity (surface),

substrate type and depth, water clarity as determined with a 20 cm-diameter secchi disk. We

assessed percent cover of eelgrass and seaweeds within four, 0.25 m2 quadrats that were spaced

approximately 5 m apart in cardinal directions around the point. We also determined the

presence/absence of selected sessile macro-invertebrate species (i.e., bivalves, gastropods,

sponges, starfish, and crabs) within the quadrats. Cover was defined as the portion of the quadrat

area obscured by a particular macrophyte while viewed in water from above. If eelgrass was

present, 5 representative shoots were collected from each of the quadrats and measured for width

and length. If seaweeds were present, we estimated cover for all species combined and for the

dominant seaweed species within each of the four quadrats. Seaweeds were identified in the

field, but, when this was not possible; specimens were collected and later identified by Dr.

Sandra Lindstrom of the University of British Columbia.

To minimize among-observer differences in estimates of macrophyte cover we assigned a

cover score between 0 and 5 based on the Braun-Blanquet (BB) visual estimation technique

(Braun-Blanquet 1972). From these cover estimates we computed three statistics for eelgrass

and total seaweeds: density, abundance and frequency of occurrence according to Fourqurean et

al. (2001). We estimated aboveground biomass of eelgrass in each year to establish a

relationship between percent cover (BB score: 1-5) x average shoot length (cm) and above-

ground biomass of eelgrass (g dry wt/m2) within each embayment. Percent cover was estimated

for eelgrass in 25, 0.25m2 quadrats and then shoots were clipped and frozen. Later in the

laboratory the samples were cleaned, measured, and dried to constant mass for weighing. We

report means and standard errors.

7

Results and Discussion: The Landsat imagery was a good data source for the assessment of eelgrass spatial extent

in coastal waters adjacent to Alaska Peninsula and Becharof NWR. We searched the EROS

Landsat archive from 1990 through 2010 for imagery with minimal cloud cover acquired during

peak eelgrass biomass (June – August) and low tide to increase our accuracy by mapping

eelgrass while it was most evident. The 1999 imagery was the best fit for these criteria and

permitted an initial estimate of eelgrass spatial extent in

Chignik Lagoon and Mud Bay (Table 1). We were able to

differentiate eelgrass from other cover types (i.e. water and

unvegetated tidal flats) with a high degree of accuracy based

on field survey point data not designed for ground truthing

the imagery (see below). With additional ground-truthing of

the classifications a useful baseline map can be created to

evaluate future habitat change at these embayments. Images

from 2006 and 2009, with portions of the study areas

obscured by clouds or containing data gaps caused by sensor

malfunction, were obtained and could be used to create a

combined product for a more recent estimate of eelgrass

extent.

Chignik Lagoon:

Mapping.

Eelgrass represented the largest land cover type in Chignik Lagoon, comprising 54% of

its spatial extent (Figure 1; Table 1). Although distribution was fairly consistent across most of

the lagoon, eelgrass was absent near the mouth, in deeper channels and close to the inlet of the

Chignik River. Distribution was more irregular in shallow areas near the head of the lagoon

where mud flats became more predominant.

We found only a small degree of error in the classification of eelgrass extent based on an

accuracy assessment using 104 field survey points (Table 2). Of these, 66 of the 72 “Eelgrass”

points (91.7% producer’s accuracy) and 30 of the 32 “No Eelgrass” (93.8% producers accuracy)

were classified correctly for an overall accuracy of 92.3%. Of the 6 misidentified “Eelgrass”

Table 1. Land cover class summary for Chignik Lagoon and Mud Bay

(ha) Chignik

Bay Mud Bay

Eelgrass 2414 188

Sand / Mud 905 58

Deep Water 1150 30

Total 4469 276

(% of total)

Eelgrass 54 68

Sand / Mud 20 21

Deep Water 26 11

Total 100 100

8

points, 1 was classified as

sand/mud while the other

5 were classified as deep

water. All of the 8

misidentified points were

located in transition zones

between cover types

where differing resolution

between the field data (1

m) and the satellite image

(30 m) was the likely source of error.

IDW interpolations were performed using survey point data to map the variability in

density (percent cover) and abundance (standing crop) of eelgrass across the lagoon (Figure 2).

The main area of high eelgrass density (75 – 100 percent cover) was spread across the northwest

two-thirds of the lagoon, though eelgrass was notably absent in the deeper channels and close to

the lagoon mouth, with secondary concentrations to the southeast and southwest of Chignik

Island. Eelgrass abundance followed the same general pattern, but the areas of greatest

abundance (137 – 229 g dry wt/m2) were more concentrated. Eelgrass became increasingly

sparse and less abundant while mudflats became more persistent in the shallow southwest third

of the lagoon.

Distribution of eelgrass is most influenced by light availability (i.e., water depth and

clarity), wave and tidal action, sediment transport, and nutrient availability (Hemminga and

Duarte 2000). Therefore, the absence of eelgrass near the lagoon entrance and in deep channels

may be due to water depths too deep for light penetration and photosynthesis in combination

with strong wave and tidal action that prevent eelgrass from taking root. Toward the head of the

lagoon, fine grain sediments from fresh water inputs, particularly the Chignik River, may

increase turbidity thereby reducing light penetration and lowering eelgrass productivity.

Increased desiccation from exposure to air and direct sunlight, during low tide may also limit

productivity in the lagoon’s shallow southeastern region. Eelgrass was most abundant at

moderate water depths, particularly where the spit across the lagoon mouth, and sand bars protect

against extensive wave action while the channels deliver essential nutrients.

Table 2: Eelgrass classification error for Chignik Lagoon, 1999 Landsat imagery

08/08/1999 ETM+ Classification

Field Survey Reference Data

Eelgrass No

Eelgrass

Total % (User's Accuracy)

Eelgrass 66 2 68 97.1

No Eelgrass 6 30 36 83.3 Total Correct 96 Total Survey Points 72 32 104

% Correct (Producer's Accuracy)

91.7 93.8 92.3

9

Figure 2. Inverse distance weighted interpolations of eelgrass percent cover (A) and standing crop (B) using 2010 field survey data.

10

Eelgrass and Seaweed Abundance.

Chignik Lagoon was characterized by cold water temperature, very low salinity, shallow

substrate depths, and deep water depths (Table 3). Average surface water temperature was 10.0

+ 0.2 o C (range = 8 – 14o C) in June with colder temperatures near river outflows and warmer

temperatures in the center of the lagoon. Outflow from the Chignik River and Mallard Creek

were strong in June resulting in a relatively low average surface salinity of 12.3 + 0.8 ppt (range

= 1 – 25 ppt). Salinity measurements were lowest near the river outflows and highest at the

mouth of the lagoon (Table 3; Figure 3). The overall salinity was lower in Chignik Lagoon than

in other embayments measured to date along the Alaska Peninsula (n = 4; Table 3; Hogrefe et al.

2011). Although not reflected in the average water depths (Table 3) Chignik Lagoon contained

relatively deep water channels averaging around 2 m in the central portion and 5 m near the

mouth of the lagoon. Average water depths at our sample points were not representative of the

lagoon because measurements were restricted to intertidal flats with depths above 2.5 m mllw.

The substrate was composed primarily of 86% fine sediments (53% mud and 33% sand) and

14% cobble rock.

Eelgrass was present on 72% of points (n = 104) scattered across the entire bay, and

when present, eelgrass was abundant (average abundance score = 4.4, ca.85% cover; (Table 3;

Figure 2). Average tide height of points containing eelgrass was -0.2 + 0.01 m with points

spread across intertidal flats from +1.0 to -1.3 m mllw. Above-ground biomass averaged 99.3 +

6.8; range = 2 to 229 gdrywt/m2, a level that was 45% greater in Chignik Lagoon than in Mud

Bay but nearly half the above-ground biomass of eelgrass in Kinzarof Lagoon (Table 3).

Variation in above-ground biomass was positively correlated with shoot lengths (R2 = 0.48; P <

0.001) and salinity (R2 = 0.87; P < 0.001) and negatively correlated with intertidal depth (R2 =

0.24; P < 0.001; greater abundance with increasing intertidal depth to about -1.0 m mllw).

Average length of eelgrass shoots was 62.9 + 3.9 cm (range = 14.2 – 181 cm) in early summer.

On average, shoots were longer in this lagoon than in all other embayments surveyed to date in

southwest Alaska, including Izembek, Togiak, and Yukon Delta NWRs (Table 3; Hogrefe et al.

2011). The longest shoots, averaging >1 m in length, were found in the first third of the lagoon

Seaweeds, representing 37 different species, were sparsely distributed in Chignik

Lagoon, occurring on 62% of points (n = 95) and nearly always in association with eelgrass

(99% of occurrences). When present, seaweed abundance was also relatively low (mean

11

abundance score = 2.21 + 0.12; ca. 30% cover) compared to eelgrass abundance and to seaweed

abundance in other embayments (Table 3). The most common seaweeds found in the lagoon

were green blades, comprised mostly of Kornmannia zostericola and to a lesser extent Ulvaria

obscura var. blyttii, and Monostroma grevillei. Kormannia sp. was found unattached in clumps

at the base of eelgrass shoots or attached to eelgrass blades. The next most common seaweeds

were epiphytic diatoms, Neorhodomela oregona, and Eudesme virescens. Epiphytic diatoms

were most abundant in the low intertidal beds of eelgrass near the island and in central portions

of the bay. Macro-invertebrates (presence of >1 species) were found on an average of 41% of

points (n = 95) and always in association with eelgrass. Occurrence of macro-invertebrates in

Chignik Lagoon was similar to Nanvak Bay (40% of points) in Togiak NWR and lower than at

Izembek and Kinzarof lagoons in Izembek NWR (ca 65% of points). The most common macro-

invertebrates in Chignik Lagoon were gastropods (39% of points), mussels (12% of points), and

marine isopod crustaceans (12% of points). We did not detect any crabs, starfish, or sponges

within the survey points.

12

Table 3. Abundance of eelgrass and seaweed at Chignik Lagoon and Mud Bay in June and Kinzarof Lagoon in August, 2010. Chignik Lagoon Mud Bay Kinzarof LagoonA n Mean SE n Mean SE n Mean SE Physical parameters Water Temperature (C) 77 9.95 0.17 34 10.00 0.11 90 12.13 0.09 Salinity (ppt) 74 12.28 0.84 34 24.25 1.05 91 26.50 0.18 Water depth (cm) 72 81.03 4.18 34 83.79 6.95 102 99.44 4.88 Tidal depth, mean lower low water (m) 74 -0.19 0.07 37 0.10 0.11 101 0.40 0.08 Substrate depth (cm) 84 3.92 0.32 39 7.96 0.98 102 3.25 0.41 Seagrass Zostera marina Standing crop (g dry wt/m2) when present 74 99.27 6.76 30 67.97 7.29 70 207.23 23.15 Vegetative Shoots Density (BB score: between 0 and 5) 94 3.44 0.21 42 2.56 0.33 103 2.36 0.20 Abundance (BB score: between 1 and 5) 74 4.43 0.12 30 3.79 0.24 75 3.57 0.16 Frequency (between 0 and 1) 94 0.74 0.04 42 0.64 0.45 103 0.61 0.04 Shoot length (cm) 75 62.92 3.94 31 59.00 3.25 70 54.59 4.57 Shoot sheath (cm) 74 15.47 1.05 31 15.66 0.97 70 11.66 0.99 Seaweeds All species combined Vegetative Shoots Density (BB score: between 0 and 5) 95 0.96 0.12 42 2.52 0.29 103 2.54 0.15 Abundance (BB score: between 1 and 5) 58 2.21 0.12 37 3.37 0.23 96 2.87 0.13 Frequency (between 0 and 1) 95 0.41 0.04 42 0.70 0.06 103 0.85 0.03 A Located 200km west of Chignik along the south side of Alaska Peninsula in Izembek NWR

13

Figure 3. Inverse distance weighted interpolations salinity using 2010 field survey data.

14

Mud Bay:

Mapping.

Classified Landsat imagery indicated that eelgrass meadows were the predominant cover

type in Mud Bay, comprising 68% of its spatial extent with a fairly even distribution across the

bay (Figure 1; Table 1). As determined from the field survey data, our map was an accurate

assessment of eelgrass distribution in this embayment (Table 4). Of the 40 survey points, 27 of

the 28 “Eelgrass” points (96.4% producer’s accuracy) and 9 of the 12 “No Eelgrass” points (75%

producer’s accuracy)

were classified correctly

for an overall accuracy of

90.0%. Again, each of

the 4 misidentified points

were located in transition

zones between cover

types so that differing

resolution between data

sources was the likely

cause of the error.

The IDW interpolations of survey data indicated that the main area of high eelgrass

density (75 – 100 percent cover) and abundance (94 - 149) was in the central portion of the bay

(Figure 2). The general pattern of eelgrass distribution was similar to that of Chignik Lagoon

with the greatest abundance occurring in protected areas with moderate water depth and at a

distance from fresh water inlets.

Eelgrass and Seaweed Abundance

In contrast to Chignik Lagoon, Mud Bay was characterized by relatively high salinity

(range = 2.2 – 30 ppt), deep substrate depths (range = 1 – 20 cm), and shallow water depths (range =

4 – 168 cm; Table 3; Figure 3). Freshwater influx was minimal in June with inputs primarily

confined to small streams at the head of the bay. Water temperature averaged 10.0 + 0.1 o C

(range = 9 – 12o C) and was similar to levels found in Chignik Lagoon or just outside of these

embayments in Chignik Bay. As the name implies, Mud Bay was comprised primarily of fine

Table 4: Eelgrass classification error for Mud Bay, 1999 Landsat imagery

08/08/1999 ETM+ Classification

Field Survey Reference Data

Eelgrass No

Eelgrass Total

% (User's

Accuracy)

Eelgrass 27 3 30 90.0

No Eelgrass 1 9 10 90.0 Total Correct 36 Total Survey Points 28 12 40

% Correct (Producer's Accuracy)

96.4 75.0 90.0

15

sediments (63% mud and 22% sand) with depths averaging twice those of Chignik Lagoon

(Table 3).

Mud Bay also contrasted from Chignik Lagoon in that seaweeds were the dominant

marine macrophyte, comprising 88% of points (n = 42). We detected >29 different species of

seaweeds with green seaweeds being the most dominant (71% of points) and most abundant. Of

these, K. zostericola (57% of points) occurred more often than Cladophora spp. (48% of points),

but Cladophora spp. were more abundant when present (average abundance score: Cladophora

spp. = 3.3, ca.60% cover vs Kornmannia spp. = 1.6, ca. 20% cover). Other common seaweed

species were Alaria spp. (33% of points), Saccharina spp. (29% of points), and Dictyosiphon

foeniculaceus (17% of points). Epiphytic diatoms, which were abundant in Chignik were rarely

(1% of points) detected in Mud Bay. Seaweeds were frequently associated with eelgrass (81% of

points), but it was not uncommon for them to occur in areas void of eelgrass (19% of points) and

containing rockier substrates.

Eelgrass occurred on slightly over 71% of points in Mud Bay, and when present, eelgrass

was abundant (mean abundance score = 3.8 + 0.24, ca. 70% cover). On average shoot lengths

were shorter and above-ground biomass was lower in Mud Bay than in Chignik or Kinzarof

lagoons (Table 3). Macro-invertebrates (presence of >1 species) were more common in Mud

Bay than in Chignik Lagoon, occurring on 57% of points (n = 42) and almost always (90% of

points) in association with eelgrass. The most common macro-invertebrates in Mud Bay were

gastropods (57% of points), amphipod crustaceans (10% of points), and mussels (2% of points).

We did not detect any crabs, starfish, or sponges.

Recommendations: Mapping:

• Ground truth and finalize the maps of eelgrass extent in Chignik Lagoon and Mud Bay. • Acquire photography and/or satellite imagery to map eelgrass extent in other embayments that are important to APBNWR.

Field Surveys:

• Continue environmental monitoring (e.g., dissolved O2, water temperature, salinity, water clarity) in Chignik Lagoon.

16

Acknowledgements: Funding and in-kind support for this project was provided by the Alaska Peninsula and

Becharof National Wildlife Refuge (APBNWR), and U. S. Geological Survey-Alaska Science

Center. We are extremely grateful to staff at APBNWR for their generosity and kind assistance

throughout project. We thank Sandra Lindstrom (University of British Columbia) for her

assistance with identification of seaweeds. Field collections would not be possible without the

generous help of the crew on the USFWS vessel, R/V Arlluk including Alvin Pederson, Orville

Lind and Dominique Watts.

Literature Cited: Braun-Blanquet, J. 1972. Plant sociology: the study of plant communities. Hafner Press, New York. Chander, G, BL Markham, and DL Helder. 2009. Summary of current radiometric calibration coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI sensors. Remote Sensing of Environment 113: 893-903. Chavez, PS. 1988. An Improved Dark-Object Subtraction Technique for Atmospheric Scattering Correction of Multispectral Data. Remote Sensing of Environment. 24:459-479. Costanza, R et al. 1997. The value of the world’s ecosystem services and natural capital. Nature 387: 253-260. Costanza, R et al. 1997. The value of the world’s ecosystem services and natural capital. Nature 387: 253-260. Duffy, JE. 2006. Biodiversity and the functioning of seagrass ecosystems. Marine Ecology Progress Series 311:233-250. Fourqurean, JW, A Willsie, CD Rose, and LM Rutten. 2001. Spatial and temporal pattern in seagrass community composition and productivity in south Florida. Marine Biology 138:341-354. Hemminga, MA, and CM Duarte. 2000. Seagrass ecology, Cambridge (United Kingdom): Cambridge University Press. Hogrefe, KH, DH Ward, T Donnelly and L Fairchild. 2011. Eelgrass and Seaweed Assessments at Izembek NWR, 2007 – 2010. Progress Summary – March 2011 Orth, RJ., et al. 2006. A global crisis for seagrass ecosystems. BioScience, 56, 987–996.

17

Ozesmi, SL and ME Bauer. 2002. Satellite remote sensing of wetlands. Wetlands Ecology and Management 10:381-402 Short, FT, and S. Wyllie-Echeverria. 1996. A review of natural and human-induced disturbance of seagrasses worldwide. Environmental Conservation 23:17–27. Short, FT, and HA Neckles. 1999. The effects of global climate change on seagrasses. Aquatic Botany 63:169-196. Valley, RD, MT Drake, and CS Anderson. 2005. Evaluation of alternative interpolation techniques for the mapping of remotely-sensed submersed vegetation abundance. Aquatic Botany. 81:1:13-25 Valta-Hulkkonen, K, P Pellikka, H Tanskanen, A Ustinov, and O Sandman. 2003. Digital false colour aerial photographs for discrimination of aquatic macrophyte species. Aquatic Botany 75:71-88 Ward, DH., CJ Markon, and DC Douglas. 1997. Distribution and stability of eelgrass beds at Izembek Lagoon, Alaska. Aquatic Botany 58:229-240.33