[Next Section] [Previous Section] [Contents]

METHODS

Three objectives undertaken include the (1) testing of criteria to estimate paleosubsidence, (2) compilation of a regional subsidence database, and (3) prediction coastal flooding effects from coseismic subsidence in several bays. The methods utilized for objective (1) are discussed below. Methods to achieve objectives (2) and (3) are discussed in the Results section.

Coastal wetlands in Grays Harbor estuary, Washington, Necanicum estuary, Oregon, Tillamook Bay, Oregon, and Siletz Bay, Oregon are investigated for comparisons of the most recent subsidence record and adjacent modern settings of tidal wetlands. Buried wetland deposits are identified in cores and cutbanks from these bays. Modern tidal wetlands are similarly cored and surveyed for elevation relative to modern sea level. The most recently buried wetland deposit is used as fragile plant parts and microfossils can be destroyed through taphonomic processes in older buried wetland deposits. The criteria used for evaluating subsidence in this study are (1) plant macrofossils, (2) abundance of organic content (3) and diatom assemblages across the burial contact. Plant macrofossil changes include (1) presence or absence of diagnostic plant roots, and (2) abrupt changes in organic content across a burial contact. To compliment plant macrofossil evidence of supratidal (freshwater) settings, diatoms are used that allow for discrimination between fresh, brackish, and marine water conditions.

A sampling strategy was designed to help test the accuracy of tidal level indicators outlined above. This strategy included the collection of samples from modern wetland environments and the latest buried wetland deposit.

Land surveying was completed using a Sokkia Set-4 Total Station. Accuracy between sample sites within a marsh traverse is ± 1 cm (0.4 in.) which is presumed for short traverses less than 30 m. For traverses where the total station distance to the site is greater than 30 m, accuracy is presumed to be ± 1-5cm (0.4-2.0 in.). Modern marsh elevations are surveyed from known benchmarks. Accuracy of sites tied into local benchmarks is ± 10 cm. All elevations are surveyed to the National Geodetic Vertical Datum of 1929 (NGVD 29), except for the site at Elliot Slough in Grays Harbor. Vertical control at Elliot Slough is determined from local surveying completed by the city of Aberdeen to a city datum and converted to NGVD 29. Horizontal coordinates (UTM) are estimated from U.S.G.S. 7.5 minute topographic maps to the nearest 10 m.

Four wetland community divisions are identified by plant assemblages. The divisions are colonizing, low marsh, high marsh, and forest edge. Colonizing marsh (CM) sites are identified as areas of a transition zone where low marsh plant species are colonizing a tidal flat. Low marsh (LM) communities are classified by assemblages of Distichlis spicata, Salicornia Virginica, and Triglochen maritima (Jefferson, 1975). High Marsh (HM) communities are classified by assemblages of Deschampsia caespitosa, Potentilla pacifica, and Grindelia integrifolia (Jefferson, 1975). Generally a change from high to low marsh is identified by a change in plant assemblage. The forest edge (FE) community is generally classified by the presence of Sitka spruce (Picea sitchensis), although communities of Willow and Alder are also used. Five sites were sampled for laboratory analysis of organic content at each community. Three of the five sites were used for analysis of diagnostic diatoms.

Buried wetland samples were collected from cutbanks or from a 2.54 cm wide, 1 m long, gouge corer. Representative samples from cores and cutbank exposures are taken directly above and below the 300-year burial contact. Cutbank horizons and cores are logged to the nearest centimeter for lithology, contacts, and identifiable plant macrofossils. Wherever possible, buried marsh deposits were sampled from cutbanks to reduce the possibility of microfossil contamination across the 300-year burial contact from coring. At sites where cutbanks are not exposed, multiple cores were taken to verify representative stratigraphy. Samples from above and below burial contacts are analyzed for diatoms and measurement of organic content.

Most of the buried samples are taken from localities where radiocarbon age determination has been completed (Atwater, 1988b; Darienzo, 1991; Peterson and others, 1996). At Tillamook Bay, where age determinations have not previously been completed on buried deposits, samples are submitted for radiocarbon (C-14) and atomic mass spectrometry (AMS) dating. The C-14 and AMS samples were frozen for storage prior to processing for shipping. Pre-shipment processing included: (1) thawing, (2) lightly rinsing with deionized water, (3) removing contaminating root hairs, (4) drying, and (5) wrapping in aluminum foil.

The laboratory techniques that were used to test the tide level indicators are described below.

Total organic contents of buried and modern marsh samples are determined by the loss on ignition method (Darienzo, 1991) that is modified from the technique used by Franklin and others (1973). Loss on ignition of modern marsh samples are completed to test the accuracy and precision of peat content as an indicator of marsh elevation. The combusted organic carbon is reported as a percentage of the air-dried (W₀) weight. Preliminary tests on samples showed no difference between the method below and that used by Darienzo (1991) where samples were combusted at 450⁰ C for 5 hours.

Loss on Ignition methods involved:

(1)

Analysis of diatoms provides a method of independently establishing changing salinity conditions in tidal to supratidal regimes. Descriptions of the sampling, microscope slide preparation, taxa identification, and the procedure used to identify diatom species, are outlined below.

Core samples from modern and buried deposits are split for diatom and loss on ignition analyses. Along with loss on ignition samples diatom sample splits are labeled, bagged, and then frozen to preserve the valves. The top several centimeters of each modern wetland sample are used to count the selected diatom species. In buried deposits, the thickness of the sample is reduced to a ~1 cm thick sample taken above and below buried wetland contact.

Samples of modern and buried deposits are prepared for diatom taxa analysis using a technique outlined in (Darienzo, 1991). This process involved:

For each sample, 50 fields (0.26 mm diameter) are scanned. In each field, a whole valve (diatom skeleton) nearest to the microscope cross-hairs is identified at 1250x on the basis of: length (m m), width (m m), number of striae, costae, or aerolae per 10 m m, and distinguishing morphological feature. Descriptions of diatoms used for taxa identification are from Patrick and Reimer (1966; 1975), Jensen (1985), and Hemphill-Haley (1993).

The counting of diatoms included a semi-quantitative analysis of 50 fields (0.26 mm or 260 m m diameter) at 500x for each sample. Within a field, the number of whole valve and identifiable fragments greater than 10 m m are recorded to estimate the relative abundance of diatoms.

Diagnostic diatom species utilized for paleoenvironmental analysis are based on previous diatom studies performed in Pacific Northwest estuaries (Riznyk, 1973; Rao and Lewin, 1976; Whiting and McIntire, 1985; Hemphill-Haley, 1993; Shennan and others, 1995). Diagnostic species (Table 1) are gleaned from studies that described ecology related to each species.

Ecological attributes include tolerance to different levels of salinity. A comparison of buried samples with modern marsh samples is conducted using selected species, the majority of which are identified at both buried and modern wetland settings in Willapa Bay (Hemphill-Haley, 1993) Studies performed in other localities by Hustedt (1927-1966), Hendey (1964), and Foged (1978; 1979; 1981) support the terms assigned to the diagnostic species below.

Comparisons of diatom assemblages from replicate samples of modern wetland settings are undertaken in this study to help establish precision and accuracy in the prediction of paleotidal levels.

The relative abundances of diagnostic diatoms in buried and modern marsh samples are based on the averages of specific diatoms per field. To document saline tidal influence total percentages of polyhalobous (marine) species, mesohalobous (brackish) species, and oligohalobous (freshwater) species are calculated for each wetland community.

Statistical methods from raw diatom data used to support apparent interpretations include the chi-squared (c ²) test for homogeneity, z-test of proportions, and regression analysis to determine if a linear trend exists. All statistics described below are taken from Walpole and Myers (1985).

The chi-squared test for homogeneity null hypothesis is:

H₀: For each salinity level the proportions of diatoms in the forest edge, colonizing, low, marsh high tidal settings are the same.

If the null hypothesis is rejected the accepted hypothesis is:

H₁: For at least one salinity level the proportion of diatoms in the forest edge, colonizing, low, marsh high tidal settings are not the same.

The equation for the c ²-test of homogeneity is:

(2-a)

where o_i is the observed number of diatoms counted and e_i is the expected frequency. The expected frequency e_i is calculated from:

(2-b)

Where c_i and r_j are the rows and columns, respectively. Degrees of freedom (v) are calculated from:

(2-c)

The z-test of proportions is used to compare the proportions of above contact (mud) samples with below contact (peat) samples. Buried wetland samples are also compared to modern wetland sample using the z-test. The null hypothesis tested by the z-statistic is:

H₀: p₁= p₂

Where p₁ and p₂ are the proportions tested. If the null hypothesis is rejected then it is accepted that:

H₁: p₁ ¹ p₂

The z value for test p₁= p₂ is determined by the formula:

(3-a)

where is the pooled estimate of the proportion p given by

(3-b)

where x₁ and x₂ are the number of diatoms from a salinity division in each of the two samples, and n₁ and n₂ represent the total number of diatoms counted. The value is equal to 1-. The z-test uses a two-tailed level of significance (a ) of 0.05 for the critical region of |z|>1.96.

A linear regression analysis of apparent linear-trended data is completed using the null hypothesis that states

H₀: B = 0

If the null hypothesis is rejected it is accepted that statistically there is a linear trend such that

H₁: B ¹ 0

where B is the variable for the equation in the model:

y_i = A + Bx_i + e _I (4-a)

where e _I is the model error with a mean of zero such that an estimated regression line defined by

a + bx (4-b)

where a and b are estimates of the regression coefficients A and B.

The test statistic (t) is defined by

(4-c)

with n-2 degrees of freedom and defined by

=(4-d)

where =(4-e)

From the t statistic the P-value (Probability-value) can be calculated from

2(Area under the curve above the t statistic value) (4-f)

and if P > significance level (a = 0.05): accept H₀.

If P < significance level (a = 0.05): reject H₀ and accept H₁, where the confidence level (1-a ) is 95%.

These methods provide a basis on which paleotidal elevations are evaluated for deposits above and below buried wetland contacts.

[Next Section] [Previous Section] [Contents]