&EPA
          United States
          Environmental Protection
          Agency
            Office of Research and
            Development
            Washington DC 20460
EPA/600/R-01/009
December 1999
Innovative Technology
Verification Report
          Sediment Sampling
          Technology

          Art's Manufacturing &
          Supply, Inc.
          Split Core Sampler for
          Submerged Sediments

-------
                                  EPA/600/R-01/009
                                   December 1999
  Innovative Technology
     Verification Report
Art's Manufacturing & Supply, Inc.,
       Split Core Sampler for
       Submerged Sediments
                Prepared by

               Tetra Tech EM Inc.
               Chicago, Illinois

             Contract No. 68-C5-0037
               Dr. Stephen Billets
            Environmental Sciences Division
          National Exposure Research Laboratory
           Office of Research and Development
          U.S. Environmental Protection Agency
            Las Vegas, Nevada 89193-3478

-------
                                       Notice
This document was prepared for the U.S. Environmental Protection Agency (EPA) Superfund
Innovative Technology Evaluation Program under Contract No. 68-C5-0037. The document has
been subjected to the EPA's peer and administrative reviews and has been approved for publication.
Mention of corporation names, trade names, or commercial products does not constitute endorsement
or recommendation of specific products for use.

-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                Office of Research and Development
                     Washington, DC 20460
               ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM
                                  VERIFICATION STATEMENT
                                                                                                     ^A
                                                                                            ii   i^.^T
   TECHNOLOGY TYPE:    SEDIMENT SAMPLER

   APPLICATION:           CORE SAMPLING OF SEDIMENT

   TECHNOLOGY NAME:   ART'S MANUFACTURING & SUPPLY, INC.,
                              SPLIT CORE SAMPLER FOR SUBMERGED SEDIMENTS

   COMPANY:               ART'S MANUFACTURING & SUPPLY, INC.
   ADDRESS:                105 HARRISON
                              AMERICAN FALLS, IDAHO 83211

   WEB SITE:                http://www.ams-samplers.com

   TELEPHONE:             (208) 226-2017
  VERIFICATION PROGRAM DESCRIPTION

  The U.S. Environmental Protection Agency (EPA) created the Superfund Innovative Technology Evaluation (SITE) and
  Environmental Technology Verification (ETV) Programs to facilitate deployment of innovative technologies through
  performance verification and information dissemination. The goal of these programs is to further environmental protection
  by substantially accelerating the acceptance and use of improved and cost-effective technologies. These programs assist and
  inform those involved in design, distribution, permitting, and purchase of environmental technologies. This document
  summarizes results of a demonstration of the Split Core Sampler for Submerged Sediments (Split Core Sampler) designed
  and fabricated by Art's Manufacturing & Supply,  Inc.

  PROGRAM OPERATION

  Under die  SITE and ETV Programs, with the full participation of the technology developers, the EPA  evaluates and
  documents the performance of innovative technologies by developing demonstrationplans, conducting field tests, collecting
  and analyzing demonstration data, and preparing reports. The technologies  are evaluated under rigorous quality assurance
  (QA) protocols to produce well-documented data of known quality.  The EPA National Exposure Research Laboratory,
  which demonstrates field sampling, monitoring,  and  measurement technologies, selected Tetra Tech EM Inc. as the
  verification organization to assist in field testing two sediment sampling technologies. This demonstration was funded by
  the SITE Program.

  DEMONSTRATION DESCRIPTION

  In April and May 1999, the EPA conducted a field demonstration of the Split Core Sampler along with one other sediment
  sampler. This verification statement focuses on the Split Core Sampler; a similar statement has been prepared for the other
  sampler. The performance and cost of die  Split Core Sampler were compared to those of two conventional samplers (the
  Hand Corer and Vibrocorer), which were used as reference samplers. To verify a wide range of performance attributes, the
  Split Core Sampler demonstration had both primary and secondary objectives. Primary objectives for this  demonstration
  included evaluating the sampler's ability to (1) consistently collect a given volume of sediment, (2) consistently collect
  sediment in a given depth interval, (3) collect samples with consistent characteristics from a homogenous layer of sediment,
  (4) collect a representative sample from a clean sediment layer below a contaminated sediment layer, and (5) be adequately
  decontaminated. Additional primary objectives were to measure  sampling time and estimate sampling costs.  Secondary
  objectives included (1) documenting the skills and training required for sampler operation, (2) evaluating die sampler's
  ability to collect samples under a variety of site conditions,  (3) assessing the sampler's ability to collect an undisturbed
  sample, (4) evaluating sampler durability, and (5) documenting the availability  of the sampler and its spare parts. To ensure
  data usability, data quality indicators for precision, accuracy, representativeness, completeness, and comparability were also
  assessed based on project-specific QA objectives.
EPA-VS-SCM-37                The accompanying notice is an integral part of this verification statement.                 December 1999
                                                    iii

-------
  The Split Core Sampler was demonstrated at sites in EPA Regions  1 and 5.  At the Region 1 site, the sampler was
  demonstrated in a lake and wetland. At the Region 5 site, the sampler was demonstrated in a river mouth and freshwater
  bay.  Collectively, the two sites provided multiple sampling areas with the different water depths, sediment types, sediment
  contaminant characteristics, and sediment thicknesses  necessary to properly evaluate  the sampler.  Based on the
  predemonstration investigation results, demonstration objectives, and site  support facilities available, (1) the Hand Corer
  was used as the reference sampler in the lake, wetland, and freshwater bay and (2) the Vibrocorer was used as the reference
  sampler in the river mouth. A complete description of the demonstration and a summary of its results are available in the
  "Innovative Technology Verification Report: Sediment Sampling Technology—Art's Manufacturing & Supply, Inc., Split
  Core Sampler for Submerged Sediments" (EPA/600/R-01/009).

  TECHNOLOGY DESCRIPTION

  The Split Core Sampler is an end-filling sampler designed to collect undisturbed core samples of sediment up to a maximum
  depth of 4 feet below sediment surface  (bss).  The sampler collects samples from the sediment surface downward, not at
  discrete depth intervals. Sampler components  include one or more split core tubes, couplings for attachment to additional
  split core tubes, a ball check valve-vented top cap, a coring tip, one  or more extension rods, and a cross handle. All these
  components are made of stainless steel; carbon-steel extension rods are also available from the developer. The sampler may
  be used with a core tube liner to facilitate removal of an intact sample from the split core tube. To collect a sediment sample,
  the sampler can be either manually pushed into the sediment using the cross handle or hammered into the sediment using
  a slide-hammer or an electric hammer.  The check valve in the sampler's top cap allows water to exit the sampler during
  deployment and creates a vacuum to help retain a sediment core during sampler retrieval.  The sampler can be retrieved by
  hand, by reverse hammering using the slide-hammer, or by using a tripod-mounted winch.

  VERIFICATION OF PERFORMANCE

  Key demonstration findings are summarized below for the primary objectives.

  Consistently Collecting a Given Volume of Sediment. In the shallow depth interval (0 to 4 inches bss), to collect a specified
  number of samples, the Split Core Sampler required 7 percent more attempts than expected (46 actual versus 43 expected),
  whereas the reference  samplers required 14 percent more attempts than expected (49 actual versus 43 expected).  In the
  moderate depth interval (4 to 32 inches bss), the Split Core Sampler required 38 percent more attempts than expected (40
  actual versus 29 expected), but the reference samplers required 156 percent more attempts than expected (64 actual versus
  25 expected).

  For the shallow depth interval, mean sample recoveries ranging from 89 to 100 percent were achieved by the Split Core
  Sampler, whereas mean sample recoveries for the reference samplers ranged from 85 to 100 percent. The variation in sample
  recoveries as measured by their relative standard deviations (RSD) ranged from 0 to 26 percent for the Split Core Sampler,
  whereas the reference samplers' RSDs ranged from 0 to 33 percent. For the moderate depth interval, mean sample recoveries
  ranging from 37 to 100 percent were achieved by  the Split Core Sampler, whereas the reference samplers'  mean sample
  recoveries ranged from 21 to 82 percent. The  RSDs for the Split Core Sampler ranged from 0 to 51 percent, whereas the
  reference samplers' RSDs ranged from 3 to 161  percent.

  Consistently Collecting Sediment in a Given Depth Interval: Both the Split Core Sampler and reference samplers collected
  samples in shallow and moderate depth intervals in all demonstration areas, which contained various sediment types. No
  sampler was able to collect samples in the deep depth interval (4 to 11 feet bss).  For the shallow depth interval, the Split
  Core Sampler's actual core lengths equaled the  target core length in 96 percent of the total sampling attempts. The reference
  samplers' actual core lengths equaled the target core length in 94 percent of the total sampling attempts. For the moderate
  depth interval, the Split Core Sampler's actual core lengths equaled the target core length in 39 percent of the total sampling
  attempts. The reference samplers' actual core lengths equaled the  target core length in 13 percent of the total sampling
  attempts.

  Collecting Samples with  Consistent Characteristics from a Homogenous Layer of Sediment: Based on particle size
  distribution results,  both  the Split Core Sampler and reference samplers  collected samples with consistent physical
  characteristics from two homogenous layers of sediment (a sandy silt layer and a clayey silt layer).

  Collecting a Representative Sample from a Clean Sediment Layer Below a Contaminated Sediment Layer: In sampling
  a clean sediment layer below a contaminated sediment layer, the Split Core Sampler and reference sampler (the Hand Corer)
  collected samples whose contaminant concentrations were statistically different at a significance level of 0.05.  Arsenic
  concentrations in the samples collected by the Split Core Sampler were less than those in the samples collected by the Hand
  Corer. However, because of the greater opportunity for sample compaction in the Split Core Sampler, no conclusion could
  be drawn regarding this sampler's ability to collect representative samples  from a clean layer below a contaminated layer.
EPA-VS-SCM-37                 The accompanying notice is an integral part of this verification statement.                   December 1999
                                                       iv

-------
  Sampler Decontamination. Both the Split Core Sampler and reference samplers demonstrated the ability to be adequately
  decontaminated after sampling in areas contaminated with either polychlorinated biphenyls or arsenic.

  Sampling Time. Compared to the reference samplers, the Split Core Sampler reduced sampling time by 15 to 52 percent
  in three of the four areas sampled but increased the sampling time by 8 percent in the remaining area.

  Sampling Costs. Of the sampling costs estimated for two of the four areas sampled, in one area the sampling costs for the
  Split Core Sampler were 95 percent less than those for the reference sampler (the Vibrocorer), and in the other area the
  sampling costs for the Split Core Sampler were 8 percent more than those for the reference sampler (the Hand Corer).

  Key demonstration findings are summarized below for the secondary objectives.

  Skill and Training Requirements. The Split Core Sampler, like the Hand Corer, is easy to operate and requires minimal
  skills and training. However, operation of the Vibrocorer is relatively complicated and requires moderate skills and training.
  The Split Core Sampler was operated by one person, whereas the Hand Corer was operated by one or two persons and the
  Vibrocorer was operated by two persons.  When more than two extension rods were required, the Split Core Sampler and
  Hand Corer were operated using a tripod-mounted winch.  The Vibrocorer operation required a motor-operated winch
  because of the weight of the sampler.

  Sampling Under a Variety of Site Conditions'. Both the Split Core Sampler and reference samplers collected samples in
  shallow and moderate depth intervals in all demonstration areas, which contained various sediment types. No sampler was
  able to collect samples in the deep  depth interval (4 to 11 feet bss).  For more efficient recovery of samples, an electric
  hammer should be used to induce vibrations in the Split Core Sampler; a 110-volt power supply is required to operate the
  electric hammer. The Vibrocorer requires a three-phase, 230- or 440-volt, 50- to 60-hertz power supply, which is a sampler
  limitation if the power supply fails.  The Hand Corer does not require a power supply.

  Collecting an Undisturbed Sample: Based on visual observations, both the Split Core Sampler and reference samplers
  collected partially compressed core samples of consolidated and unconsolidated sediments from the  sediment surface
  downward. Samples collected by both the Split Core Sampler and reference samplers in moderate and deep depth intervals
  may be of questionable representativeness because of core shortening and core compression. Sediment stratification was
  preserved for both consolidated and unconsolidated sediments in the samples collected by  the Split Core  Sampler and
  reference samplers.

  Sampler Durability and Availability. Based on their materials of construction and engineering designs, both the Split Core
  Sampler and reference samplers are considered to be sturdy. The Split Core Sampler and its support equipment  are not
  expected to be available in local retail stores.  Similarly, the primary components of the Hand Corer and Vibrocorer are not
  expected to be available in local retail stores; extension rods for the Hand Corer may be locally available.

  Based on the demonstration results, the Split Core Sampler can be operated by one person with minimal skills and training.
  For more efficient recovery of samples, an electric hammer should be used to induce vibrations in the sampler.  When more
  than two extension rods are used, a winch is recommended for sampler operation. The sampler is designed to  collect
  sediment samples up to a maximum depth of 4 feet bss and, based on visual observations, collects partially compressed
  samples of both consolidated and unconsolidated sediments from the sediment surface downward; sample representativeness
  may be questionable because of core  shortening and core compression. The sampler preserves sediment stratification inboth
  consolidated and unconsolidated sediment samples. The Split Core Sampler is a good alternative to conventional sediment
  samplers.  As with any sampler selection, the user must determine the appropriate sampler for a given application based on
  project-specific data quality objectives.
  Gary J. Foley, Ph.D.
  Director
  National Exposure Research Laboratory
  Office of Research and Development
   NOTICE: EPA verifications are based on an evaluation of technology performance under specific, predetermined criteria and
   appropriate quality assurance procedures. The EPA makes no expressed or implied warranties as to the performance of the technology
   and does not certify that a technology will always operate as verified.  The end user is solely responsible for complying with any and
   all applicable federal, state, and local requirements.
EPA-VS-SCM-37                 The accompanying notice is an integral part of this verification statement.                  December 1999

                                                        V

-------
                                      Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
nation's natural resources.  Under the mandate of national environmental laws, the agency strives
to formulate and implement actions leading to a compatible balance between human activities and
the ability of natural systems to support and nurture life. To meet this mandate, the EPA Office of
Research and Development  provides  data  and scientific support that  can be used to solve
environmental problems, build the scientific knowledge base needed to manage ecological resources
wisely, understand how pollutants affect public health, and prevent or reduce environmental risks.

The National Exposure Research Laboratory (NERL) is the agency's center for investigation of
technical and management approaches for identifying and quantifying risks to human health and the
environment. Goals of the laboratory's research program are to (1) develop and evaluate methods
and technologies for characterizing and monitoring air, soil, and water; (2) support regulatory and
policy decisions; and (3) provide the scientific support needed to ensure effective implementation
of environmental regulations and strategies.

The EPA Superfund Innovative Technology Evaluation (SITE) Program evaluates  technologies
designed for  characterization and  remediation  of contaminated Superfund  and Resource
Conservation and Recovery Act sites. The SITE Program was created to provide reliable cost and
performance data in order to speed acceptance and use of innovative remediation, characterization,
and monitoring technologies by the regulatory and user community.

Effective measurement and  monitoring technologies  are  needed to  assess  the  degree of
contamination at a site, provide data that can be used to determine the risk to public health or the
environment, supply the necessary cost  and performance data  to select the  most appropriate
technology, and monitor the success or failure of a remediation process. One component of the EPA
SITE Program, the Monitoring and Measurement Technology (MMT) Program, demonstrates and
evaluates innovative technologies to meet these needs.

Candidate technologies can originate within the federal government or the private sector. Through
the SITE Program, developers are given the opportunity to conduct a rigorous demonstration of their
technologies under actual field conditions. By completing the  demonstration and distributing the
results, the agency establishes a baseline for acceptance and use of these technologies.  The MMT
Program is administered by the Environmental Sciences Division of NERL in Las Vegas, Nevada.

                                            Gary J. Foley, Ph.D.
                                            Director
                                            National Exposure Research Laboratory
                                            Office of Research and Development
                                           VI

-------
                                      Abstract
The Split Core Sampler for Submerged Sediments (Split Core Sampler) designed and fabricated by
Art's Manufacturing & Supply, Inc., was demonstrated under the U.S. Environmental Protection
Agency (EPA) Superfund Innovative Technology Evaluation Program in April and May 1999 at sites
in EPA Regions 1 and 5, respectively.  In addition to assessing ease of sampler operation, key
objectives of the demonstration included evaluating the sampler's ability to (1) consistently collect
a given volume of sediment, (2) consistently collect sediment in a given depth interval, (3) collect
samples with consistent characteristics from a homogenous layer of sediment, and (4) collect
samples under a variety of site conditions. This report describes the demonstration results for the
Split Core Sampler and two conventional samplers  (the Hand Corer and Vibrocorer) used as
reference samplers.  During the demonstration, the  Split Core Sampler performed as well as or better
than the  reference  samplers.  Based on  visual observations, both the Split Core Sampler and
reference  samplers collected partially compressed  samples  of consolidated and unconsolidated
sediments from the sediment surface downward;  sample representativeness may be questionable
because of core shortening and core compression. Sediment stratification was preserved for both
consolidated and unconsolidated sediment samples collected by the Split Core Sampler and reference
samplers.  No sampler was able to collect samples in the deep depth  interval (4 to 11 feet below
sediment surface).  The average sampling time was less for the Split Core Sampler than for the
reference samplers. Sampling costs for the Split Core Sampler were 8 percent greater than those for
the Hand Corer and 95 percent less than those for the Vibrocorer.
                                          vn

-------
                                      Contents



Chapter                                                                         Page

Notice	ii

Verification Statement	iii

Foreword 	 vi

Abstract 	vii

Figures	xiii

Tables	xv

Abbreviations, Acronyms, and Symbols	xvii

Acknowledgments  	  xix

1      Introduction	1
        1.1    Description of the SITE Program  	1
        1.2    Scope of the Demonstration	4

2      Description of the Innovative Sediment Sampler	6
       2.1    Sampler Description  	6
       2.2    General Operating Procedures	8
       2.3    Advantages and Limitations  	8
       2.4    Developer Contact Information	9

3      Demonstration Site Descriptions  	10
       3.1    EPA Region 5 Site (Site 1)  	10
              3.1.1    Site l,Areal 	10
              3.1.2    Site 1, Area 2 	11
       3.2    EPA Region 1 Site (Site 2)  	12
              3.2.1    Site 2, Area 1 	12
              3.2.2    Site 2, Area 2 	12

4      Demonstration Approach  	13
       4.1    Demonstration Objectives	13
                                         Vlll

-------
                              Contents (Continued)
Chapter                                                                         Page

       4.2    Demonstration Design	14
       4.3    Field Sampling and Measurement Procedures	17
       4.4    Laboratory Sample Preparation and Analysis Methods  	22

5      Description of the Reference Sediment Samplers 	26
       5.1    Hand Corer	26
              5.1.1   Technology Description	26
              5.1.2   General Operating Procedures 	27
              5.1.3   Advantages and Limitations	28
       5.2    Vibrocorer	28
              5.2.1   Technology Description	28
              5.2.2   General Operating Procedures 	29
              5.2.3   Advantages and Limitations	30

6      Performance of the Split Core Sampler 	31
       6.1    Primary Objectives  	31
              6.1.1   Ability to Consistently Collect a Specified Volume of Sediment  ....  32
                     6.1.1.1 Number of Sampling Attempts Required  	32
                     6.1.1.2 Volume of Sediment Collected	34
              6.1.2   Ability to Consistently Collect Sediment in a Specified Depth
                     Interval  	37
              6.1.3   Ability to Collect Multiple Samples with Consistent Physical or
                     Chemical Characteristics, or Both, from a Homogenous Layer of
                     Sediment	39
              6.1.4   Ability to Collect a Representative  Sample from a Clean Sediment
                     Layer Below a Contaminated Sediment Layer	42
              6.1.5   Ability to be Adequately Decontaminated	44
              6.1.6   Time Requirements for Sample Collection Activities 	44
       6.2    Secondary Objectives  	46
              6.2.1   Skill and Training Requirements for Proper Sampler Operation	46
              6.2.2   Ability to Collect Samples Under a Variety of Site Conditions	47
              6.2.3   Ability to Collect an Undisturbed Sample 	48
              6.2.4   Durability Based on Materials of Construction and Engineering
                     Design	48
              6.2.5   Availability of Sampler and Spare Parts	48
       6.3    Data Quality	48
              6.3.1   Field Measurement Activities	49
              6.3.2   Laboratory Analysis Activities	49

7      Performance of the Reference Samplers	53
       7.1    Primary Objectives  	53
              7.1.1   Ability to Consistently Collect a Specified Volume of Sediment  ....  54
                     7.1.1.1 Number of Sampling Attempts Required  	54

-------
                              Contents (Continued)
Chapter                                                                         Page

                     7.1.1.2 Volume of Sediment Collected	56
              7.1.2  Ability to Consistently Collect Sediment in a Specified Depth
                     Interval  	58
              7.1.3  Ability to Collect Multiple Samples with Consistent Physical or
                     Chemical Characteristics, or Both, from a Homogenous Layer of
                     Sediment	60
              7.1.4  Ability to be Adequately Decontaminated	63
              7.1.5  Time Requirements for Sample Collection Activities 	64
       7.2    Secondary Objectives  	65
              7.2.1  Skill and Training Requirements for Proper Sampler Operation	65
              7.2.2  Ability to Collect Samples Under a Variety of Site Conditions	66
              7.2.3  Ability to Collect an Undisturbed Sample 	67
              7.2.4  Durability Based on Materials of Construction and Engineering
                     Design	67
              7.2.5  Availability of Sampler and Spare Parts	67
       7.3    Data Quality	68
              7.3.1  Field Measurement Activities	68
              7.3.2  Laboratory Analysis Activities	69

8      Economic Analysis	71
       8.1    Issues and Assumptions	71
              8.1.1  Sampler Costs	71
              8.1.2  Labor Costs	71
              8.1.3  IDW Disposal Costs	72
              8.1.4  Support Equipment Costs	72
              8.1.5  Costs Not Included	73
       8.2    Split Core Sampler Costs	74
              8.2.1  Sampler Cost	74
              8.2.2  Labor Cost  	74
              8.2.3  IDW Disposal Cost	74
              8.2.4  Support Equipment Cost	74
              8.2.5  Summary of Split Core Sampler Costs	76
       8.3    Hand Corer Costs	76
              8.3.1  Sampler Cost	76
              8.3.2  Labor Cost  	76
              8.3.3  IDW Disposal Cost	76
              8.3.4  Support Equipment Cost	77
              8.3.5  Summary of Hand Corer Costs	77
       8.4    Vibrocorer Costs 	77
              8.4.1  Sampler Cost	77
              8.4.2  Labor Cost  	77
              8.4.3  IDW Disposal Cost	77
              8.4.4  Support Equipment Cost	78

-------
                              Contents (Continued)
Chapter
Pas
              8.4.5   Summary of Vibrocorer Costs  	78
       8.5    Comparison of Economic Analysis Results	78

9      Summary of Demonstration Results	79
       9.1    Primary Objectives 	79
       9.2    Secondary Objectives  	84

10     References	86

Appendix A   Developer's Claims for the AMS Split Core Sampler for Submerged
              Sediments	88
       A.I    Updates or Improvements to the Split Core Sampler  	88
       A.2    Prior Deployment of the Split Core Sampler	89
       A.3    Developer Comments on the SITE Demonstration	89

Appendix B    Performance and Cost of the Ekman Grab  	90
       B.I    Description of the Ekman Grab	90
              B.I.I   Sampler Description	90
              B.1.2   General Operating Procedures  	91
              B.I.3   Advantages and Limitations	91
       B.2    Description of the Demonstration Sites	92
       B.3    Demonstration Approach	92
              B.3.1   Demonstration Objectives 	92
              B.3.2   Demonstration Design   	93
              B.3.3   Field Sampling and Measurement Procedures	94
       B.4    Performance of the Ekman Grab	96
              B.4.1   Primary Objectives	96
                     B.4.1.1 Ability to Consistently Collect a Specified Volume of
                            Sediment  	97
                     B.4.1.2 Ability to Consistently Collect Sediment in a Specified
                            Depth Interval   	99
                     B.4.1.3 Ability to Collect Multiple Samples with Consistent
                            Physical or Chemical Characteristics, or Both, from a
                            Homogenous Layer of Sediment	100
                     B.4.1.4 Ability to be Adequately Decontaminated  	101
                     B.4.1.5 Time  Requirements for Sample Collection Activities	101
                     B.4.1.6 Costs Associated with Sample Collection Activities  	103
              B.4.2   Secondary Objectives	105
                     B.4.2.1 Skill and Training Requirements for Proper Sampler
                            Operation	105
                     B.4.2.2 Ability to Collect Samples Under a Variety of Site
                            Conditions 	105
                     B.4.2.3 Ability to Collect an Undisturbed Sample	106
                                         XI

-------
                              Contents (Continued)
Chapter                                                                        Page

                     B.4.2.4 Durability Based on Materials of Construction and
                            Engineering Design	107
                     B.4.2.5 Availability of Sampler and Spare Parts  	107
              B.4.3  Data Quality  	107
                     B.4.3.1 Field Measurement Activities 	107
                     B.4.3.2 Laboratory Analysis Activities 	108
       B.5    References 	108

Appendix C    Statistical Methods 	109
       C.I    Wilk-Shapiro Test	109
       C.2    Wilcoxon Signed Rank Test 	110
       C.3    References 	113
                                         xn

-------
                                       Figures
Figure                                                                           Page
2-1.    Split Core Sampler 	7
4-1.    Site 1 sampling locations  	18
4-2.    Site 2 sampling locations  	19
5-1.    Hand Corer	27
5-2.    Vibrocorer	29
6-1.    Percent sample recoveries for Split Core Sampler at Site 1	35
6-2.    Percent sample recoveries for Split Core Sampler at Site 2	36
6-3.    Split Core Sampler sample particle size distribution results for
       S1A2 (freshwater bay)  	40
6-4.    Split Core Sampler sample arsenic and particle size distribution results for
       S2A1 (lake)	41
6-5.    Comparison  of Split Core Sampler and reference sampler arsenic concentration
       results for S2A1 (lake)  	43
7-1.    Percent sample recoveries for Vibrocorer and Hand Corer at Site 1 	56
7-2.    Percent sample recoveries for Hand Corer at Site 2	57
7-3.    Hand Corer sample particle size distribution results for S1A2 (freshwater bay)	61
7-4.    Hand Corer sample arsenic and particle size distribution results for S2A1 (lake)	62
B-l.    Ekman Grab  	91
B-2.    Sampling locations for Ekman Grab demonstration	95
B-3.    Percent sample recoveries for Ekman Grab in S1A1 (river mouth),
       S1A2 (freshwater bay), and S2A1 (lake) 	100
B-4.    Ekman Grab sample analytical results for S1A1 (river mouth) and S2A1  (lake)  ....  102
C-l.    Wilk-Shapiro test plot for core length measurements in S1A2 (freshwater bay)	110
C-2.    Wilk-Shapiro test plot for core length measurements in S2A2 (wetland)  	Ill
C-3.    Statistix® output for Hand Corer sample data for S2A2 (wetland)	112
                                         Xlll

-------
                            Figures (Continued)
Figure                                                                     Page

C-4.    Statistix® output for Hand Corer and Split Core Sampler sample data for
       S2A1 (lake)	112
                                      xiv

-------
                                      Tables
Table                                                                           Page
3-1.    Demonstration Area Characteristics	11
4-1.    Innovative Sediment Sampler Demonstration Design  	15
4-2.    Rationale for Sampling Approach  	20
4-3.    Sample Matrix	23
4-4.    Laboratory Sample Preparation and Analysis Methods	24
4-5.    Laboratory Quality Control Checks 	25
6-1.    Comparison of Expected and Actual Number of Sampling Attempts for
       Split Core Sampler at Site 1 	33
6-2.    Comparison of Expected and Actual Number of Sampling Attempts for
       Split Core Sampler at Site 2 	34
6-3.    Percent Sample Recovery Summary Statistics for Split Core Sampler  	37
6-4.    Comparison of Target and Actual Core Length Data for Split Core Sampler	38
6-5.    Particle Size Distribution Summary Statistics for Split Core Sampler	42
6-6.    Time Required to Complete Sampling Activities for Split Core Sampler	45
6-7.    Summary of Quality Control Checks and Acceptance  Criteria for Field and
       Laboratory Parameters  	50
7-1.    Comparison of Expected and Actual Number of Sampling Attempts for
       Reference Samplers at Site 1 	54
7-2.    Comparison of Expected and Actual Number of Sampling Attempts for
       Reference Sampler at Site 2 	55
7-3.    Percent Sample Recovery Summary Statistics for Reference Samplers  	58
7-4.    Comparison of Target and Actual Core Length Data for Reference Samplers 	59
7-5.    Particle Size Distribution Summary Statistics for Hand Corer   	63
7-6.    Time Required to Complete Sampling Activities for Reference Samplers  	64
8-1.    Comparison of Investigation-Derived Waste Quantities Generated by
       Split Core Sampler and Reference Samplers 	72
                                         xv

-------
                              Tables (Continued)
Table                                                                        Page
8-2.    Split Core Sampler Cost Summary	75
8-3.    Hand Corer Cost Summary for S2A1 (Lake)	76
8-4.    Vibrocorer Cost Summary for S1A1 (River Mouth) 	77
8-5.    Comparison of Costs for Split Core Sampler and Reference Samplers	78
9-1.    Summary of Results for Primary Objectives 	80
9-2.    Summary of Results for Secondary Objectives  	82
B-l.    Ekman Grab Demonstration Design	93
B-2.    Rationale for Sampling Approach  	96
B-3.    Ekman Grab Sample Matrix	97
B-4.    Comparison of Expected and Actual Number of Sampling Attempts for
       Ekman Grab at Site 1  	98
B-5.    Comparison of Expected and Actual Number of Sampling Attempts for
       Ekman Grab in S2A1 (Lake) 	98
B-6.    Percent Sample Recovery Summary Statistics for Ekman Grab  	99
B-7.    Comparison of Target and Actual Sediment Thickness Data for Ekman Grab 	101
B-8.    Particle Size Distribution Summary Statistics for Ekman Grab	103
B-9.    Time Required to Complete Sampling Activities for Ekman Grab  	104
B-10.   Ekman Grab Cost Summary	105
C-l.    Data Sets for Example Wilk-Shapiro Test Calculations	110
C-2.    Hand Corer Sample Data for 4- to 12-Inch Below Sediment Surface
       Depth Interval in S2A2 (Wetland)  	Ill
C-3.    Hand Corer and Split Core Sampler Sample Data for 10- to 30-Inch
       Below Sediment Surface Depth Interval in S2A1 (Lake)	112
                                       xvi

-------
                   Abbreviations, Acronyms, and Symbols
>           Greater than
<,           Less than or equal to
±           Plus or minus
<           Less than
AMS       Art's Manufacturing & Supply, Inc.
ASTM      American Society for Testing and Materials
BS/BSD     Blank spike/blank spike duplicate
bss         Below sediment surface
CFR        Code of Federal Regulations
DER       Data evaluation report
EPA        U.S. Environmental Protection Agency
ETV        Environmental Technology Verification
FLAA      Flame atomic absorption
ft           Foot
ft/s         Foot per second
GLNPO     Great Lakes National Program Office
ICP         Inductively coupled argon plasma
IDW        Investigation-derived waste
ITVR       Innovative technology verification report
L           Liter
Ib           Pound
mg/kg       Milligram per kilogram
mg/L       Milligram per liter
mL         Milliliter
MMT       Monitoring and Measurement Technology
MS/MSD    Matrix spike/matrix spike duplicate
NA         Not applicable
NERL      National Exposure Research Laboratory
ORD       Office of Research and Development
OSWER     Office of Solid Waste and Emergency Response
PCB        Polychlorinated biphenyl
PE         Performance evaluation
PSD        Particle size distribution
PSR        Percent sample recovery
QA         Quality assurance
QA/QC     Quality assurance/quality control
QC         Quality control
RPD        Relative percent difference
                                        XVII

-------
           Abbreviations, Acronyms, and Symbols (Continued)

RSD        Relative standard deviation
S1A1        Site 1, Area 1
S1A2        Site 1, Area 2
S2A1        Site 2, Area 1
S2A2        Site 2, Area 2
SITE        Superfund Innovative Technology Evaluation
SOP        Standard operating procedure
Statistix®    Statistix® for Windows, Version 2.0
TCLP       Toxicity characteristic leaching procedure
Tetra Tech   Tetra Tech EM Inc.
TSA        Technical  system audit
                                      XVlll

-------
                               Acknowledgments
This report was prepared for the U.S. Environmental Protection Agency (EPA) Superfund Innovative
Technology Evaluation (SITE) Program under the direction and coordination of Dr. Stephen Billets
and Dr. Brian Schumacher of the EPA National Exposure Research Laboratory—Environmental
Sciences Division in Las Vegas, Nevada.  The SITE Program thanks Mr. Joseph LeMay and
Mr. Andy Beliveau of EPA Region 1, Mr. Robert Paulson of the Wisconsin Department of Natural
Resources, and Mr. Marc Tuchman and Mr. Scott Cieniawski of the EPA Great Lakes National
Program Office for their support in conducting field activities for this project. Mr. Jonathan Kuhns
of Hawk Consulting and Dr. Larry Jackson of Environmental Quality Management served as the peer
reviewers of this report.

This report was prepared for the EPA by Dr. Kirankumar Topudurti, Mr. Eric Monschein, and
Mr. Andrew  Bajorat of Tetra Tech EM Inc. Special acknowledgment is given to Ms. Jeanne
Kowalski, Mr. Jon Mann, Mr. Stanley Labunski, Ms. Sandy Anagnostopoulos, Ms. Amy Stephen,
Mr. Gary Sampson, and Mr. Bob Overman for their assistance during the preparation of this report.

-------
                                              Chapter 1
                                            Introduction
The U.S. Environmental Protection Agency (EPA) Office
of  Research  and  Development's  (ORD)  National
Exposure Research Laboratory (NERL) has conducted a
demonstration of an innovative sediment sampler known
as the Split Core Sampler for Submerged Sediments,
a  core  sampler  designed and  fabricated  by Art's
Manufacturing & Supply, Inc. (AMS), of American Falls,
Idaho. In this innovative technology verification report
(ITVR), the AMS Split Core Sampler  for Submerged
Sediments is referred to as  the Split Core Sampler.  The
demonstration was conducted under the EPA Superfund
Innovative Technology Evaluation (SITE) Program at two
sites during the last week of April and first week of May
1999. The purpose of this demonstration was to obtain
reliable  performance and cost data  on the Split Core
Sampler in order to (1) achieve a better understanding of
the  sampler's  capabilities  relative  to conventional
sediment samplers and (2) provide an  opportunity for the
sampler to enter the  marketplace  and  compete  with
conventional samplers without long delays.

This  ITVR presents the  performance  results  of the
demonstration and associated costs for the Split Core
Sampler.  Specifically, this report describes  the SITE
Program and the scope of the demonstration (Chapter 1),
innovative  sediment sampler that was demonstrated
(Chapter 2), two demonstration sites (Chapter 3),
demonstration  approach   (Chapter  4), conventional
sediment samplers used as reference samplers during the
demonstration (Chapter 5), performance of the innovative
sampler  (Chapter  6),  performance  of  the  reference
samplers (Chapter 7), economic analysis for the innovative
and reference samplers (Chapter 8), demonstration results
in summary form (Chapter 9), and references used to
prepare the ITVR (Chapter 10). AMS  claims for, updates
on, and information on previous deployments of the
innovative  sampler  are   provided  in   Appendix  A.
Appendix B presents performance results for the Ekman
Grab, a conventional grab sampler that was included in the
demonstration because grab samplers are commonly used
to collect  surficial sediment in order to assess the
horizontal  distribution  of  sediment  characteristics.
Appendix C describes the statistical methods  used, as
appropriate, to address the primary objectives for the
demonstration.

1.1    Description of the SITE Program

Performance verification  of innovative environmental
technologies is an integral part of the regulatory and
research mission of the EPA. The SITE Program was
established by the EPA  Office of Solid  Waste  and
Emergency Response  (OSWER) and  ORD under the
Superfund Amendments and Reauthorization Act of 1986.
The primary purpose of the SITE Program is to promote
acceptance and use of innovative sampling, monitoring,
measurement, and treatment technologies.

The  overall goal  of  the  SITE Program is to conduct
research  and  performance  verification   studies  of
innovative technologies that may be used to achieve long-
term protection of human health and  the environment.
The  various  components  of the  SITE Program are
designed to  encourage  development,  demonstration,
acceptance, and use of innovative sampling, monitoring,
measurement, and treatment technologies. The program is
designed to meet four primary objectives:  (1) identify
and remove obstacles to development and commercial use
of innovative technologies, (2) support a development
program  that   identifies   and  nurtures  emerging
technologies,   (3)  demonstrate  promising  innovative
technologies to establish reliable performance  and cost
information for site characterization and cleanup decision-

-------
making, and (4) develop procedures and  policies  that
encourage use of innovative technologies at Superfund
sites as well as at other  waste sites and  commercial
facilities.

The  intent  of a  SITE demonstration  is to obtain
representative, high-quality performance and cost data on
one or more innovative technologies  so that potential
users can  assess a given technology's  suitability for a
specific application.   The SITE Program  includes the
following elements:

•   Monitoring and Measurement Technology (MMT)
    Program—Evaluates  technologies that  sample,
    detect, monitor, and measure hazardous and toxic
    substances.  These technologies  are  expected to
    provide better, faster, and more cost-effective methods
    for   producing   real-time  data  during   site
    characterization and remediation  studies  than do
    conventional technologies.

•   Remediation   Technology   Program—Conducts
    demonstrations of innovative treatment technologies
    to provide reliable performance, cost, and applicability
    data for site cleanups.

•   Technology  Transfer  Program—Provides   and
    disseminates technical information in  the  form of
    updates,  brochures,  and  other  publications   that
    promote the SITE Program and technologies. It  also
    offers  technical assistance, training, and workshops
    to support the technologies.

The  innovative sediment  sampler demonstration  was
conducted as part of the MMT Program, which  provides
developers of  innovative  hazardous  waste sampling,
monitoring,  and  measurement  technologies  with  an
opportunity  to   demonstrate  their  technologies'
performance  under  actual  field conditions.   These
technologies may be used to sample, detect, monitor, or
measure hazardous and toxic substances in soil, sediment,
waste material, or groundwater. The technologies include
chemical sensors  for  in situ (in place) measurements,
groundwater samplers, soil and sediment samplers, soil gas
samplers,   laboratory   and   field-portable  analytical
equipment, and other systems that support field sampling
or data acquisition and analysis.
The MMT Program promotes acceptance of technologies
that  can be used  to  accurately assess the degree of
contamination at a site, provide data to evaluate potential
effects on human health and the environment, apply data
to assist in selecting the most appropriate cleanup action,
and monitor the  effectiveness of a remediation process.
The  program places a  high priority  on  innovative
technologies that provide more cost-effective, faster, and
safer methods for producing real-time or near-real-time
data than do conventional technologies. These innovative
technologies are demonstrated under field conditions, and
the   results  are  compiled, evaluated,  published,  and
disseminated by ORD.  The  primary objectives of the
MMT Program are as follows:

•   Test field sampling and analytical technologies that
    enhance  sampling,   monitoring,    and   site
    characterization capabilities

•   Identify  performance   attributes  of   innovative
    technologies to address field sampling,  monitoring,
    and characterization problems in a more cost-effective
    and efficient manner

•   Prepare  protocols, guidelines,  methods, and other
    technical publications that enhance acceptance of
    these technologies for routine use

The MMT Program is administered by the Environmental
Sciences Division of NERL in Las Vegas, Nevada.  The
NERL is the EPA's center for investigation  of technical
and  management  approaches  for  identifying   and
quantifying risks to human health  and the environment.
The NERL's mission components include (1) developing
and evaluating methods and technologies for sampling,
monitoring,  and characterizing water,  air, soil,  and
sediment; (2) supporting regulatory and policy decisions;
and (3) providing the technical support needed to ensure
effective implementation of environmental regulations and
strategies. By demonstrating selected innovative sediment
samplers, the MMT Program is supporting development
and evaluation of methods and technologies for sampling
and characterizing sediment.

The MMT Program's technology performance verification
process is designed to conduct demonstrations that will
generate high-quality data that potential users can employ

-------
to verify technology performance and cost. Four key steps
are inherent in the process: (1) needs identification and
technology selection, (2) demonstration planning and
implementation,   (3)  report   preparation,  and
(4) information distribution.

The first step of the technology performance verification
process begins with identifying technology needs of the
EPA and regulated community. The EPA regional offices,
the U.S. Department of Energy, the U.S. Department of
Defense, industry, and state environmental regulatory
agencies are  asked  to  identify technology needs  for
sampling, monitoring, and measurement of environmental
media. Once a need is identified, a search is conducted to
identify suitable technologies that will address the need.
The technology search and identification process consists
of examining industry and trade publications, attending
related conferences,  exploring leads from technology
developers and industry experts, and reviewing responses
to Commerce Business Daily announcements. Selection of
technologies  for field testing includes evaluation of the
candidate  technologies  based on  several criteria.   A
suitable technology for field testing

•   Is designed for use in the field

•   Is  applicable  to  a  variety  of  environmentally
    contaminated sites

•   Has potential  for solving problems  that  current
    methods cannot satisfactorily address

•   Has estimated costs that are competitive with those of
    current methods

•   Is likely to achieve better results than current methods
    in areas such as data quality and turnaround time

•   Uses techniques that are easier and safer than  current
    methods

•   Is commercially available

Once  candidate   technologies  are  identified, their
developers are  asked to  participate  in  a developer
conference.   This  conference gives  the developers an
opportunity to describe their technologies' performance
and to learn about the MMT Program.
The   second  step   of the  technology  performance
verification process  is  to  plan  and  implement  a
demonstration that will generate high-quality  data that
potential  users  can  employ  to  verify  technology
performance and cost. Demonstration planning activities
include  a predemonstration  sampling  and  analysis
investigation that assesses  existing conditions  at  the
proposed demonstration site or sites. The objectives of the
predemonstration investigation are to (1) confirm available
information on  applicable  physical,  chemical,  and
biological characteristics of contaminated media at the
sites to justify selection of site areas for the technology
demonstration; (2) provide the technology developers with
an opportunity to evaluate the areas and identify logistical
requirements;   (3)  determine  the  overall   logistical
requirements for conducting  the  demonstration; and
(4) provide the analytical laboratories with an opportunity
to identify any  matrix-specific  analytical   problems
associated  with  contaminated  media  and  propose
appropriate solutions. Information generated through the
predemonstration investigation  is used to  develop  the
demonstration  design  and  sampling  and  analysis
procedures.

Demonstration planning activities also include preparation
of a demonstration plan that describes the procedures to be
used to verify the performance and cost of each innovative
technology.    The  demonstration plan   incorporates
information generated  during  the  predemonstration
investigation as well as input from technology developers
and demonstration site representatives. The demonstration
plan  also incorporates the quality assurance and quality
control (QA/QC) elements needed to  produce data of
sufficient quality to document the performance and cost of
each technology.

During the technology performance verification process,
each innovative technology is evaluated independently
and,  when possible, against a reference technology. The
performance of a developer or innovative technology is
not compared to that of another developer or innovative
technology.  Rather, demonstration data  are used  to
evaluate the performance, cost, advantages, limitations,
and field applicability of each technology.

As part of the third step of the technology performance
verification process, the EPA publishes a verification
statement and a detailed evaluation of each technology in

-------
an ITVR.  To ensure its quality, the ITVR is published
only after comments from the technology developer and
external peer reviewers are satisfactorily addressed.  All
demonstration  data used to evaluate each innovative
technology are summarized in  a data evaluation report
(DER) that constitutes a record of the demonstration. The
DER is not published by the EPA, but an unpublished
copy may be obtained by contacting the EPA project
manager, Dr. Stephen Billets.

The fourth step of the technology performance verification
process is to distribute  demonstration information.  The
EPA distributes  ITVRs free of charge through direct
mailings, at conferences, and on the Internet to benefit
technology developers  and potential  technology users.
ITVRs are available on the Internet through the Hazardous
Waste  Clean-Up Information web  site supported by the
EPA  OSWER   Technology   Innovation  Office
(http://www.clu-in.org). Additional information on the
SITE  Program  is  provided  at   the  ORD  web  site
(http://www.epa.gov/ORD/SITE).

1.2     Scope of the Demonstration

Environmental  sediment  sampling  is  conducted  to
characterize sediment at a particular location. Sediment
characterization  may involve  biological analyses (for
biological  availability  and  benthic  biota), chemical
analyses (for organic and inorganic contaminants), and
physical analyses (for  color, texture, and  particle size
distribution  [PSD]).  Sediment samplers are typically
designed to collect discrete samples of sufficient quantity
and quality at a predetermined depth relatively easily and
in a  reasonable amount of time. Although  the samplers
now  being  used  meet   most  sediment  sampling
requirements, innovative samplers maybe faster and easier
to operate, less expensive, and more accurate and precise.

The  MMT Program members involved in the Split Core
Sampler demonstration included the EPA NERL, the EPA
National Risk Management Research Laboratory, EPA
Region   1,  the  Wisconsin Department  of  Natural
Resources, the EPA Great Lakes National Program Office
(GLNPO),andAMS.

The  performance  of  the  Split  Core  Sampler  was
demonstrated and  compared to  that of conventional
sediment samplers in order to provide evidence that the
Split Core Sampler worked as intended and to facilitate its
use.   The conventional sediment samplers, which are
referred to as reference samplers herein, are described in
Chapter 5. For the demonstration, either a Hand Corer or
a Vibrocorer was used as a reference sampler, depending
on site conditions and sampler availability.

In addition to the Split Core Sampler, AMS was given the
opportunity to substitute one alternate innovative sampler
if AMS believed that the alternate sampler was better
suited forthe conditions and objectives being addressed in
a particular sampling  area.  Because  the  Split  Core
Sampler was not designed to collect core samples more
than 48 inches  below  sediment surface  (bss), AMS
attempted to demonstrate the  AMS Dual Tube Liner
Sampler in one demonstration area to collect samples in
the 4- to 6-foot  bss depth interval.   However, while
attempting to deploy the Dual Tube Liner Sampler during
a practice run,  AMS could not control the sampler's
deployment into the sediment because of its heavy weight.
As a result,  AMS elected not to demonstrate  the  Dual
Tube Liner Sampler in this area.

AMS also attempted to demonstrate the Dual Tube Liner
Sampler in one demonstration area in  order to collect
samples in the 9- to 11-foot bss depth interval. However,
because AMS did not have all the sampler components on
hand at the time,  AMS could not demonstrate the Dual
Tube Liner Sampler in this area.

A  conventional grab sampler was also  included in the
demonstration because grab samplers are commonly used
to  collect surficial  sediment  in  order to  assess the
horizontal distribution of sediment characteristics. The
Ekman Grab, a commonly used grab sampler, was chosen
for the demonstration.   Performance   and  cost data
collected forthe Ekman Grab are not be compared to those
for the  Split Core Sampler but rather  are presented in
Appendix B as supplemental information.

The  demonstration  had  both  primary  and secondary
objectives.  The primary objectives were critical to the
technology  evaluation and required use of quantitative
results  to  draw  conclusions   regarding  technology
performance.  The  secondary objectives pertained to
information that was useful but did not necessarily require
use of quantitative results to draw conclusions regarding
technology performance.  Based on available historical

-------
data for the demonstration sites, the primary objectives
required use of chemical and physical characterization of
sediment but not biological characterization. The primary
and secondary objectives are presented in Chapter 4.

To meet the demonstration objectives, individual areas at
two sites were selected for conducting the demonstration.
The first site is referred to as Site 1; it included two areas
and lies in EPA Region 5. The second site is referred to as
Site 2; it included two areas and lies in EPA Region 1.
These sites and areas are described in Chapter 3.

In preparation for the demonstration, a predemonstration
sampling and analysis  investigation was completed at
the two sites in February 1999.  The  purpose of this
investigation was  to assess whether  the  sites  were
appropriate for evaluating the Split Core Sampler based on
the demonstration  objectives.  The  demonstration was
conducted during the last week of April and first week of
May 1999. The procedures used to verify the performance
and cost of the Split Core Sampler are summarized in a
demonstration plan completed in April 1999 (EPA 1999).
The  demonstration plan also incorporates the QA/QC
elements needed to generate data of sufficient quality to
document innovative and reference sampler performance
and cost. The plan is available on the Internet through the
ORD web  site (http://www.epa.gov/ORD/SITE).

-------
                                               Chapter 2
                        Description of the Innovative Sediment Sampler
Core samplers are commonly used to collect sediment
profiles in order to assess the  vertical  distribution of
sediment characteristics. Based on the method of sample
collection, core samplers may be broadly classified in two
categories:   (1)   side-filling   core   samplers  and
(2) end-filling core samplers (Faegri and Iversen 1989). A
side-filling core sampler is operated by first driving the
sampler to a particular depth. The core tube is then rotated
clockwise to fill the tube by cutting out a segment of
sediment. A large cover plate attached to the core tube
holds the  sampler stationary while the  tube  rotates
clockwise to collect the sediment.  Resistance offered by
the sediment keeps the cover plate stationary, allowing the
core tube to  rotate.  Examples  of side-filling samplers
include the Russian sampler and the Hiller sampler (Faegri
and Iversen  1989).  Additional details  on side-filling
samplers are provided by  Environment Canada  (1994),
Faegri and Iversen (1989), Aaby and Digerfeldt  (1986),
Jowsey (1966), and Belokopytov and Beresnevich (1955).

An end-filling core sampler typically consists of one or
more core tubes or a box that collects sediment from the
bottom end of the sampler as it is pushed through the
sediment.  An end-filling sampler generally  collects
sediment from the sediment surface down to a particular
depth.  Once the core sample is extruded through the end
of the sampler, a discrete depth interval of the core sample
may be  subsampled. Examples of end-filling samplers
include the Hand Corer, Split Core Sampler, Dual Tube
Liner Sampler, and Vibrocorer. Additional details on end-
filling samplers  are provided by Environment Canada
(1994),  Blomqvist (1991), Faegri and Iversen  (1989),
Aaby and Digerfeldt (1986), and Downing (1984).

This chapter describes the Split Core Sampler designed
and fabricated by AMS.   This  end-filling sampler is
designed to collect undisturbed, cylindrical core samples
of various types of sediment, including saturated sands and
silts, to a maximum depth of 48 inches bss. The sampler
is designed to collect sediment with a particulate diameter
not exceeding 2/3 inch. Sections 2.1 through 2.4 describe
the Split Core Sampler, discuss its general operating
procedures, outline its advantages  and limitations, and
provide  developer  contact  information.     Similar
information for the reference samplers used during the
demonstration is provided in Chapter 5.

2.1    Sampler Description

Components of the  Split Core Sampler include (1) 6- and
12-inch-long pairs of 300-series, stainless-steel split core
tubes with interlocking, recessed  channels and male,
square-threaded  ends;  (2) a  400-series,  stainless-steel
coring tip; (3) a plastic basket retainer with flexible leaves;
(4) a ball check valve-vented top cap; (5) a male, threaded
top cap coupling; (6) a female, square-threaded coupling
for attachmentto additional stainless-steel split core tubes;
and (7) stainless-steel  or carbon-steel (4130 alloy) AMS
extension rods available in 3-, 4-, and 5-foot lengths (see
Figure 2-1). The sampler can be operated with a stainless-
steel or rubber-coated AMS cross handle, an AMS slide-
hammer (6, 10, or 19 pounds [lb]), or an electric hammer.
The  sampler can also  be equipped with core tube liners
that fit inside the split core tubes and facilitate removal of
an intact sample. Core tube liners are available in plastic,
stainless  steel, brass, aluminum, and Teflon®; end caps
made of plastic are also available.  Additional  support
equipment for sampler deployment may include an SDS
Max self-locking adapter for attaching an electric hammer
to the extension rod, an AMS Sample Preparation Station
for splitting core tube liners and examining samples, and
an extrusion  rod.  The extrusion rod used during the
demonstration consisted of a  plunger attached  to  a

-------
                                                          Cross handle
                                                          Extension rod
                                                        Top cap coupling
                                                           Ball check
                                                          valve- vented
                                                             top cap
                                                        'Split core tubes
                                                         Basket retainer
                                                            Coring tip
                                                                                              Threaded end
Figure 2-1. Split Core Sampler.

-------
graduated rod. An AMS tripod-mounted winch may also
be used to assist the sampling technician in dislodging and
retrieving the sampler from the sediment.

The assembled Split Core Sampler has an inside diameter
of 2 inches and is designed to collect about 50 milliliters
(mL) of sediment per inch of core tube length. The fully
equipped sampler (including one pair of 2-inch-diameter,
12-inch-long split core tubes; the ball check valve-vented
top cap; the coring tip; the coupling; one 12-inch-long,
disposable, plastic core tube liner with end caps; one
4-foot-long, carbon-steel extension rod; and one rubber-
coated cross handle) weighs about 10 Ib.

The Split Core Sampler can be  either manually pushed
into sediment using the AMS cross handle or hammered
into sediment using the AMS slide-hammer or an electric
hammer.  The sampler can be removed from  sediment
either manually,  by reverse hammering using the AMS
slide-hammer, or with the AMS tripod-mounted winch.

The  Split  Core Sampler is  innovative  because  it
incorporates a ball  check valve-vented  top  cap that
(1) allows air and water to  exit  the sampler  during
deployment, (2) prevents water from entering the sampler
during retrieval, and (3) creates a vacuum to help retain
a sediment core during sampler retrieval. Also, the coring
tip of the sampler has been modified from earlier versions
of the sampler to accommodate the plastic basket retainer,
which  is designed to help prevent sample  loss as the
sampler is retrieved.

2.2    General Operating Procedures

The Split Core Sampler can be operated by one person
from a platform, from a boat, or while wading in shallow
water.  Depending on sampler decontamination require-
ments and sampling conditions such as water depth and
sediment type, the AMS stainless-steel extension rods or
the stronger,  more  widely used AMS  carbon-steel
extension rods are attached to the  sampler before  its
deployment. During sampler assembly, a core tube liner
may be inserted into  the split core tube.  The core tube
liner holds and stores the sample for later examination.

The fully assembled sampler is manually lowered into the
water in such a way that the coring tip is placed on the
sediment surface.  The speed of sampler deployment to the
sediment surface should be controlled to (1) allow air and
water to escape from the sampler through the ball check
valve in order to avoid bow wave formation, which could
disturb flocculent or unconsolidated sediment that might
be near the sediment surface, and (2) ensure  that the
pressures inside and outside the  sampler  are equalized
when the sampler touches the sediment surface.

The sampler can then be either manually pushed with the
AMS cross handle or driven with the AMS slide-hammer
or an electric hammer to the desired sediment depth. An
electric hammer, which induces vibrations in the sampler,
should be used when possible for more efficient recovery
of sediment samples.  The  sampling technician should
practice sampler deployment to determine whether the
AMS slide-hammer or an  electric hammer is  needed.
Irrespective of the deployment  mechanism used, the
sampler should be  driven into the sediment in a steady
manner.

The sampler is removed from the sediment either manually
by reverse  hammering using the AMS slide-hammer or
with the AMS tripod-mounted winch. The sampler should
be raised out of the water either manually or using the
AMS tripod-mounted winch  when the weight of the
sampler and extension rods requires it. When the sampler
is being retrieved, the sampler should be kept vertical, and
the rate of retrieval  should be kept as steady as possible to
minimize resuspension and disruption of the sediment.
The tapered coring tip, the plastic basket retainer at the
bottom of the sampler, and the partial vacuum created by
the ball check valve-vented top cap retain the  sediment
core within the split core tube.

Once the  sampler  has  been  retrieved,  either  the
interlocking split core tubes are disassembled or the coring
tip or top cap is removed to allow removal of the core tube
liner. The  sediment core enclosed in the core tube liner
can be sealed in the core tube liner using  two core tube
liner end caps or can be extruded for further examination
and processing. The sediment core may be extruded by
pushing the sample  out one end of the core tube liner with
an extrusion rod or by cutting the core tube liner open
longitudinally, if the liner is made of plastic, using the
AMS Sample Preparation Station.

2.3    Advantages and Limitations

The  Split Core  Sampler is easy  to operate, requiring
minimal skills and training. Sampler assembly and sample

-------
collection procedures can be learned in the field with a
few practice attempts.  In addition, a written standard
operating procedure (SOP) accompanies the sampler when
it is procured. The sampler can be operated by one person
in both shallow (wading) and deep water depths because
of its  lightness (10 Ib).   In addition,  the  sampling
technician can use one or a combination  of the 6- and
12-inch-long pairs of stainless-steel split  core tubes to
collect 6- to  48-inch-long sediment  cores.  Sampler
operation is especially simple when a plastic core tube
liner is used  because the  sampler  does not require
complete  disassembly  to  extrude  the  sample  and
reassembly after each sampling attempt. Only the coring
tip or top cap has to be detached in order to remove the
core tube liner containing the sediment core. Use of the
disposable  liner  also  minimizes  the risk  of  cross-
contamination between sampling locations.

Another advantage of the Split Core Sampler is the ball
check valve-vented top cap.  This top cap  is  designed to
(1)  allow air and water  to  exit  the  sampler  during
deployment, (2) prevent water from entering the sampler
during retrieval, and (3) create a vacuum to help retain a
sediment core during sampler retrieval. Collectively, these
design features increase the likelihood of collecting an
undisturbed sample.

A limitation of the Split Core Sampler is  that during
sampler deployment, the core tube liner  is  exposed to
different layers of sediment contamination.  Contaminants
may adhere to the exposed surface of the liner while the
sampler passes through different layers of sediment. Also,
the ball check valve-vented top cap may become clogged
if the sampler is deployed in such a way that the top cap is
below the sediment surface.  Therefore, the ball  check
valve must be inspected after every sampling attempt; if
the valve is clogged, the top cap should be  removed to
allow adequate cleaning of the valve.

The Split Core Sampler cannot collect discrete samples
from various  sediment depths.  Core  samples must be
collected from the sediment surface downward. Because
end-filling samplers such as the Split Core Sampler must
collect samples from the sediment surface downward, the
Split Core Sampler is subject to core shortening. Core
shortening  occurs  when  the  length  of sediment core
collected is less than the depth of sampler penetration into
the sediment.   Core  shortening may occur when the
friction of the sediment against the inside wall of the core
tube  increases  with  increasing depth  of  sediment
penetration, causing lateral displacement of sediment and
resulting in  gradually thinner increments of sediment
entering the sampler. Because not all layers are uniformly
sampled, core shortening can introduce sampling bias.

For efficient recovery of sediment samples, an electric
hammer  should  be used  to  induce  vibrations  in  the
sampler.  Because an external power source is required to
operate the electric hammer, the sampling platform must
be able to accommodate the weight and size of a portable
generator. Furthermore, if the sampling platform is not
already equipped with a winch system and an access hole,
use of the AMS tripod-mounted winch or similar device
limits the sampling platform  locations from which the
sampler can be deployed.  Specifically, use of the tripod-
mounted winch requires that  the sampling platform be
equipped with a hole  over which the tripod-mounted
winch can be placed and through which the sampler can be
deployed.

2.4    Developer Contact Information

Additional information about the Split Core Sampler can
be obtained from the following source:

Mr. Brian Anderson
Art's Manufacturing & Supply, Inc.
105 Harrison
American Falls, ID 83211
Telephone: (208) 226-2017
Fax: (208) 226-7280
E-mail: briana@bankpds.com
Internet:  www.ams-samplers.com

-------
                                              Chapter 3
                                 Demonstration Site Descriptions
This  chapter  discusses  the two  sites  selected  for
conducting the Split Core Sampler demonstration.  The
first site is referred to as Site 1  and includes two areas
along a river in EPA Region 5. The second site is referred
to as Site 2 and includes two areas along a river in EPA
Region 1.  After a review of the information available on
these and other candidate sites, Sites 1 and 2 were selected
based on the following criteria:

•   Site Diversity—Each site  consisted  of multiple
    sampling areas with the different water depths, flow
    regimes,  sediment  types,  sediment contaminant
    characteristics, and sediment thicknesses necessary to
    evaluate the Split Core Sampler.

    Access and Cooperation—Site representatives were
    interested  in  supporting  the  demonstration  by
    providing historical data and site access.

In February  1999, a  predemonstration sampling and
analysis investigation was conducted to assess existing site
conditions and to confirm information provided by EPA
Regions 1  and 5.  The predemonstration investigation
results summarized in Table 3-1 were used to develop the
demonstration design for the innovative and reference
samplers.    The  following sections provide  brief
descriptions of the two demonstration sites.

3.1    EPA Region 5 Site (Site 1)

Site 1 consists of sections of a river in EPA Region 5.
Two areas along the river were selected as demonstration
areas. These areas and the sampling platforms used are
briefly described below and are shown in Figure 4-1.
3.1.1  Site 1, Area 1

Site 1, Area 1 (S1A1) lies at the river mouth, which is
about 0.5 mile wide.  The area generally represents an
open-water condition.  During the  demonstration, the
average water velocity in this area was equal to or less
than 0.07 foot per second (ft/s). The water depth in the
vicinity of S1A1 ranged from about 5 to 6 feet. Sampling
in  S1A1  was  conducted using  the  EPA  GLNPO's
Mudpuppy, a 32-foot-long, 8-foot-wide, twin-motor, flat-
bottom boat specifically designed for sediment sampling
in rivers and harbors.  The boat is equipped  with  a
vibrocoring unit supported by an A-frame and winch that
allows collection of sediment cores  up  to 15 feet long.
Additional features  that make the Mudpuppy a suitable
platform for conducting vibrocoring or other sediment
sampling include the following:

    A sampling platform at the bow of the boat with a hole
    in  the  middle  wide  enough  to accommodate the
    vibrocoring unit

    Adequate deck space for subsampling and processing
    15-foot-long core samples

    A differentially corrected global positioning system
    with submeter  accuracy that allows  precise  and
    accurate determination of sampling locations

    Four anchor lines for maintaining the boat's position
    over sampling locations

    An electrical power source for support equipment
                                                    10

-------
Table 3-1.  Demonstration Area Characteristics
Demonstration Area
S1A1 (river mouth)
S1A2 (freshwater bay)
S2A1 (lake)
S2A2 (wetland)
Average Water
Velocity3 (ft/s)
<0.07
<0.05
<0.05
< 0.05 to 0.7
Water Depth3
(ft)
5 to 6
2
18
0.5 to 1.5
Predemonstration Investigation Results
Target Sampling
Depth Interval
(inches bss)
Oto4
4 to 12
Oto6
12 to 36
Oto4
10 to 30
4 to 12
Contaminant
PCBs
PCBs
PCBs
None"
Arsenic
Arsenic
Arsenic
Physical Characteristics
Unconsolidated sediment containing primarily
sand with some silt and little clay
Consolidated sediment containing primarily sand
and silt with some clay
Unconsolidated sediment containing primarily
sand and silt with some clay
Consolidated sediment containing primarily silt
with some sand and clay
Unconsolidated sediment containing primarily silt
with some sand and clay
Consolidated sediment containing primarily sand
with some silt and little clay
Consolidated sediment containing primarily sand
with some silt and little clay
Notes:

<     =  Less than or equal to
<     =  Less than
bss   =  Below sediment surface
ft     =  Foot
ft/s    =  Foot per second
PCB   =  Polychlorinated biphenyl
    Average water velocity and water depth represent data collected during the actual demonstration.

    No measurable PCB contamination was present in this depth interval.
Predemonstration investigation sample analytical results
for S1A1 indicated that polychlorinated biphenyl (PCB)
contamination in the 0- to 4-inch bss depth interval was
minimal. However, the 4- to 12-inch bss depth interval in
this area had the highest levels of PCB contamination of
any depth interval sampled during the predemonstration
investigation.  Based on the PSD data, sediment in the 0-
to 4-inch bss depth interval was predominantly sand with
some silt and  little clay.  PSD in the 4- to 12-inch bss
depth interval was predominantly sand and silt with some
clay.  Sediment in the 0- to 4-inch bss depth interval was
Unconsolidated and became increasingly consolidated with
depth.  During the demonstration, a clay hardpan was
encountered at about 5 feet bss in  the sampling area.
Based  on  the   PCB  and   PSD  data   from  the
predemonstration investigation, the sediment in the 0- to
4-inch bss  depth interval  in  S1A1 appeared  to be
chemically and physically homogenous.  However, the
sediment in the 4- to 12-inch bss depth interval in this area
did  not  appear  to  be  as  chemically  or  physically
homogenous as was the case in Site 1, Area 2 (S1A2).
                     3.1.2   Site 1, Area 2

                     S1A2 is about 11 miles upstream of S1A1.  The river is
                     about 2,000 feet wide in S1A2.  A small, protected bay is
                     present along the  river channel's bank at this location.
                     This bay has a very slow-moving current and, because of
                     its  configuration,  backflow conditions.    During  the
                     demonstration, the average water velocity in the area was
                     less than 0.05  ft/s. The water depth in the bay was about
                     2 feet.  Sampling in S1A2 was conducted within the bay
                     using an 18-foot-long,  4-foot-wide, flat-bottom Jon boat.
                     The boat was equipped with a single engine, a set of oars,
                     and a  single anchor line for positioning the boat over
                     sampling locations. The Mudpuppy could not be used to
                     conduct sampling in S1A2 because the water in this area
                     was too shallow (the Mudpuppy requires a minimum water
                     depth of about 3 feet).

                     Predemonstration investigation sample analytical results
                     for S1A2 indicated that PCB contamination in the 0- to
                     6-inch bss depth interval was minimal but greater than that
                                                     11

-------
in the  0- to  4-inch  bss  depth  interval  in  S1A1.
Furthermore, the 12- to 36-inch bss depth interval in S1A2
had no measurable PCB contamination.  Sediment in the
0- to 6-inch bss depth interval was predominantly sand and
silt with some clay.  Sediment in the 12- to 36-inch bss
depth interval was predominantly silt with some sand and
clay.  Sediment in the top few inches was unconsolidated
and became consolidated with increasing depth. Based on
the PSD  data from the predemonstration investigation,
sediment in the 12- to 36-inch bss depth interval in S1A2
appeared to be the most physically homogenous at Site 1.

3.2    EPA Region 1 Site (Site 2)

Site 2 consists of sections of a river in EPA Region 1. The
river, which has a moderate flow, runs through a low-lying
wetland area and empties into a lake. Two areas along the
river were selected as demonstration areas. These areas
and the sampling platforms  used are briefly described
below and are shown in Figure 4-2.

3.2.1   Site 2, Area 1

Site 2, Area 1 (S2A1) is a lake  located about 5 miles
downstream of Site  2,  Area 2  (S2A2).  During the
demonstration, the average water velocity in the area was
less than 0.05 ft/s, and the water depth was about 18 feet.
Sampling in S2A1 was conducted using a 30-foot-long, 8-
foot-wide pontoon boat. The pontoon boat was equipped
with a single engine and eight anchor lines for positioning
the boat over sampling locations. In addition, a 6-inch-
diameter  hole was provided in the middle of the boat to
allow use of a core sampler with a tripod-mounted winch.
The front and sides of the boat would not accommodate a
tripod-mounted winch.

Predemonstration  investigation sample analytical results
for S2A1 indicated that the 0- to 4-inch bss depth interval
in this area had more  consistent and  higher levels of
arsenic contamination and more consistent PSD than was
the case in S2A2. Arsenic  contamination in the 0- to
4-inch bss depth interval in S2A1 was  an order  of
magnitude greater than that in the 10- to 30-inch bss depth
interval.  Sediment in the 0- to 4-inch bss depth interval
was predominantly silt with some sand and clay. Sediment
in this depth interval was unconsolidated. Sediment in the
10- to 30-inch bss depth interval was predominantly sand
with some silt and little clay.  Based on the arsenic and
PSD data from the predemonstration investigation, the
sediment in the 10- to 30-inch bss depth interval in S2A1
appeared to be  the  most  chemically  and physically
homogenous sediment at Site 2.

3.2.2  Site 2, Area 2

S2A2 is near a low-lying wetland along the river.  This
area is about 5 miles upstream of S2A1. The river channel
is about 10 feet wide in S2A2. Water flow in this area is
low to moderate, reflecting seasonal variations.  During
the demonstration, the average water velocity in the area
ranged from less than 0.05 to 0.7 ft/s, and water depths in
the area ranged from about 0.5 to 1.5  feet.  Sampling in
S2A2 was conducted from wood planks fastened to two
aluminum ladders extended across  the  river channel.
Depending on the individual needs  of each sampling
technician, (1) samples were collected off the side of one
ladder or (2) the sampling technician stood with one foot
on each ladder to collect samples between the ladders.

At the time of the predemonstration investigation, the top
4 to 8 inches of sediment in S2A2 contained  organic
matter, primarily  decomposed leaves and wood chips.
Predemonstration investigation sample analytical results
for S2A2 indicated that levels of arsenic contamination
from the bottom of the organic layer down to 12 inches bss
were nonuniform and lower than the levels in S2A1.  In
S2A2, sediment in the 4- to 12-inch  bss depth interval
(below the organic layer) was predominantly sand with
some silt and little clay.  Sediment in this depth interval
was highly consolidated. Based on the arsenic and PSD
data from the predemonstration investigation, S2A2 did
not appear to be as chemically or physically homogenous
as S2A1.  In addition, historical data provided by EPA
Region  1  indicated that a  30-foot-thick layer  of peat
existed below the  sediment layer in S2A2.
                                                    12

-------
                                               Chapter 4
                                     Demonstration Approach
This  chapter  presents  the  demonstration  objectives
(Section 4.1), design (Section  4.2), field  sampling  and
measurement procedures (Section 4.3), and  laboratory
sample preparation and analysis methods (Section 4.4).

4.1    Demonstration Objectives

The main intent of the SITE MMT Program is to develop
reliable performance and  cost  data on   innovative
technologies.   A  SITE demonstration  must  provide
detailed and reliable performance and cost data so  that
potential technology users  have adequate information to
make  sound  judgments  regarding  a  technology's
applicability to  a  specific  site  and  to  compare  the
technology to alternatives.

The Split Core Sampler demonstration had both primary
and secondary objectives. Primary objectives were critical
to the  technology evaluation and  required  use  of
quantitative  results  to  draw  conclusions  regarding
technology performance. Secondary objectives pertained
to information that was useful but did not necessarily
require use of quantitative results to  draw conclusions
regarding technology performance.

The  primary objectives for the  innovative sediment
sampler demonstration were as  follows:

PI.   Evaluate  whether  the  sampler  can consistently
      collect a specified volume of sediment

P2.   Determine whether the sampler can consistently
      collect samples in a specified depth interval

P3.   Assess the  sampler's  ability to collect multiple
      samples  with  consistent physical  or chemical
     characteristics, or both, from a homogenous layer of
     sediment

P4.  Evaluate  whether  the  sampler  can  collect  a
     representative sample from a "clean" sediment layer
     that is below a contaminated sediment layer

P5.  Assess  the  sampler's ability to be  adequately
     decontaminated between sampling areas

P6.  Measure  the  time  required  for each  activity
     associated with sample collection (sampler setup,
     sample  collection,  sampler  disassembly,  and
     sampler decontamination)

P7.  Estimate costs associated  with sample collection
     activities  (sampler, labor,  supply, investigation-
     derived  waste  [IDW]   disposal,  and  support
     equipment costs)

The  secondary  objectives for the innovative  sediment
sampler demonstration were as follows:

SI.  Document the skills and  training  required  to
     properly operate the sampler

S2.  Evaluate the sampler's ability to  collect samples
     under a variety of site conditions

S3.  Assess  the   sampler's   ability  to   collect  an
     undisturbed sample

S4.  Evaluate  the  sampler's durability based  on its
     materials of construction and engineering design

S5.  Document the availability of the sampler and spare
     parts
                                                    13

-------
The  objectives  for the demonstration were developed
based on input from MMT Program members, general user
expectations  of  sediment  sampler  capabilities,
characteristics of the  demonstration  areas, the time
available to complete the demonstration, and  sampler
capabilities that AMS intended to highlight.

4.2    Demonstration Design

In February  1999,  a predemonstration sampling and
analysis investigation was conducted to assess  existing
conditions and confirm available information on physical
and chemical characteristics in each demonstration area.
Based   on  information  from   the  predemonstration
investigation as well  as  available historical  data, a
demonstration design was developed  to  address  the
demonstration  objectives.     Input   regarding  the
demonstration design was obtained from demonstration
site representatives and AMS.  Table 4-1 summarizes the
demonstration design.

AMS  operated  the  Split  Core   Sampler in  each
demonstration area. The EPA made observations and took
measurements to  evaluate the Split Core Sampler in
accordance  with  the   demonstration  objectives.
In addition, a reference sampler  was selected for each
demonstration area either because the sampler had been
successfully used to collect  sediment  samples in  the
particular demonstration area or  because it is typically
used to collect  sediment  samples under the conditions
encountered in the particular area.  The Vibrocorer was
used as the reference sampler in S1A1. The Hand Corer
was used as the reference sampler in S1A2, S2A1, and
S2A2.   Similarly, the  sampling platforms  used were
selected based on their availability but not necessarily
based on sampler requirements.  For example, in S1A1,
the EPA GLNPO's Mudpuppy was  used because it was
available free of charge  from EPA Region 5.  During the
demonstration,  each  reference sampler was evaluated
under the same conditions and objectives as the Split Core
Sampler. All the sampling activities conducted by AMS
for the Split Core Sampler were  also conducted for the
reference  samplers  by the sampling  technicians (for
example, the EPA GLNPO operated the Vibrocorer).
During the use of each reference  sampler, the EPA also
took the  same  measurements  and made the  same
observations  as  were  performed for the  Split Core
Sampler.

The Split Core Sampler and reference sampler for Site 2
were not designed to collect core samples from the 9- to
11-foot bss sampling depth interval. Therefore, in this
sampling depth  interval, the Split  Core  Sampler and
reference sampler were not used. Because the Split Core
Sampler was not designed to collect sediment cores in the
4- to 6-foot bss sampling depth interval at Site 1, the Split
Core Sampler was also not demonstrated in this sampling
depth interval. In addition, the Dual Tube Liner Sampler
was not demonstrated in these sampling depth intervals for
the reasons stated in Section 1.2.

The approach used to address each primary objective for
the innovative and reference core samplers is discussed
below.   Because  of  varying  sampler  features,  the
characteristics of the demonstration areas, and the limited
time available for the field demonstration, not all primary
objectives were addressed in each  demonstration area.
However, the Split Core Sampler and a reference sampler
were evaluated under three  or more primary objectives in
each demonstration area.

•   To  address  primary objective  PI,  a  volume  of
   sediment to  be collected was   specified for each
   sampling depth interval.  The volume  specified was
   based on analytical requirements for characterizing the
   sample or on the design volume of the sampler for the
   particular sampling depth interval. If after one attempt
   the sampler had not retrieved the specified volume of
   sediment, additional attempts were made to retrieve
   the  specified volume.   The number of  attempts
   required and the volume of sediment collected in each
   attempt at a given location within an area were noted.

•   Primary objective P2 was addressed by verifying that
   each sediment sampler was able to consistently sample
   a specified depth interval. For each sampler, the depth
   of sampler deployment, total sample length, and
   sample length within the specified depth interval were
   noted. Various  site conditions,  including sediment
   depth, water depth, and sediment composition, were
   considered in addressing P2 in each demonstration
   area.
                                                   14

-------
Table 4-1. Innovative Sediment Sampler Demonstration Design
Demonstration
Area
S1A1
(river mouth)
S1A2
(freshwater
bay)
S2A1
(lake)
S2A2
(wetland)
Target Sampling
Depth Interval (bss)
0 to 4 inches
6 to 12 inches
4 to 6 feet
0 to 4 inches
12 to 32 inches
0 to 4 inches
10 to 30 inches
4 to 12 inches
9 to 1 1 feet
Primary Objective
P1 Volume
P2 Depth interval
P6 Sample collection time
P1 Volume
P2 Depth interval
P5 Decontamination
P6 Sample collection time
P7 Cost
P1 Volume
P2 Depth interval
P6 Sample collection time
P1 Volume
P2 Depth interval
P6 Sample collection time
P1 Volume
P2 Depth interval
P3 Consistent samples
from a homogenous
layer
P6 Sample collection time
P1 Volume
P2 Depth interval
P3 Consistent samples
from a homogenous
layer
P4 Clean layer below
contaminated layer
P5 Decontamination
P6 Sample collection time
P7 Cost
P1 Volume
P2 Depth interval
P3 Consistent samples
from a homogenous
layer
P4 Clean layer below
contaminated layer
P6 Sample collection time
P1 Volume
P2 Depth interval
P6 Sample collection time
P1 Volume
P2 Depth interval
P6 Sample collection time
Sampling Parameter
(Matrix)
Core length and
volume (sediment)
PCBs, volume, and
core length (sediment)
PCBs (final rinsate)
Core length and
volume (sediment)
Core length and
volume (sediment)
PSD, volume, and core
length (sediment)
Arsenic, PSD, volume,
and core length
(sediment)
Arsenic (final rinsate)
Arsenic, PSD, volume,
and core length
(sediment)
Core length and
volume (sediment)
Core length and
volume (sediment)
Volume Required
per Sample
Design volume3
250 ml
1 L
Design volume
Design volume
250 ml
250 ml
500ml
250 ml
Design volume
Design volume
Sampler
Split Core Sampler
Vibrocorer
Split Core Sampler
Vibrocorer
Dual Tube Liner Sampler
Vibrocorer
Split Core Sampler
Hand Corer
Split Core Sampler
Hand Corer
Split Core Sampler
Hand Corer
Split Core Sampler
Hand Corer
Split Core Sampler
Hand Corer
Dual Tube Liner Sampler
Notes:

bss    =
L
mL
PCB   =
PSD   =
Below sediment surface
Liter
Milliliter
Polychlorinated biphenyl
Particle size distribution
     For a given depth interval, the design volume corresponds to 100 percent sample recovery.
                                                               15

-------
Primary objective P3 was addressed by  analyzing
samples collected in a homogenous sediment layer for
arsenic  or PSD.  P3 was addressed  in the deeper
sampling depth interval in S1A2 and in both sampling
depth intervals in S2A1.  These areas and intervals
were chosen for this purpose because, according to the
analytical results for predemonstration investigation
samples, these intervals exhibited relatively consistent
chemical or physical characteristics  or both.

Primary objective P4 was addressed by evaluating
whether a sample could be collected from a layer of
sediment   with  relatively  low  contaminant
concentrations   (a   "clean"  layer)   beneath  a
"contaminated"  layer   of   sediment  that  had
significantly  higher  contaminant  concentrations
without cross-contaminating  the clean layer sample.
P4 was addressed in S2A1 because, according to the
results of the predemonstration investigation, a clean
layer of sediment was present beneath  a relatively
contaminated  layer  of  sediment.   During  the
demonstration, sediment samples were collected from
each layer and analyzed  for arsenic.  The analytical
data for these samples were used to determine whether
sediment from the  contaminated  layer  had  been
carried into the clean layer during sampler deployment
and retrieval.

Primary objective P5 was addressed by  collecting
samples of equipment rinsate (water) during the final
stage of core sampler decontamination.   P5  was
addressed  in the deeper sampling depth interval in
S1A1 and in the shallower sampling depth interval in
S2A1 because sediment  in these areas and intervals
contained  the highest observed  concentrations of
PCBs   and   arsenic,  respectively,  among  the
demonstration areas.    Decontamination of each
sampler demonstrated in  a given area was performed
after all samples had been collected  in that area.

Primary objective P6 was addressed by measuring the
time required for each activity associated with sample
collection, including sampler setup, sample collection,
sampler disassembly, and sampler decontamination.
P6 was addressed in all demonstration areas to satisfy
this objective under a variety of site  conditions.
•   Primary objective P7 was addressed in  S1A1  and
    S1A2 by estimating the costs associated with sample
    collection activities, including sampler, labor, IDW
    disposal, and support equipment costs. The following
    costs associated with collection of all the investigative
    samples in each area where P7 was addressed were
    accounted for:

    1.  The sampler cost was estimated based on price
       lists for purchasing  each  sediment sampler;
       disposable, plastic core liners (if applicable); and
       support  equipment.    Leasing  costs  for  the
       samplers  were  not  considered  because  the
       samplers are unavailable for leasing.

    2.  The labor cost was estimated based on the number
       of  people  required to  operate each sediment
       sampler and the time required to conduct sampling
       activities  (sampler setup,  sample   collection,
       sampler   disassembly,    and   sampler
       decontamination).

    3.  The IDW  disposal   cost was  estimated  for
       specified areas.  A volume  of sediment to be
       collected was specified for each demonstration
       area where P7 was addressed. For each such area,
       any sediment collected by a sampler that was not
       required for analytical purposes was considered to
       be  IDW.  For example, the  sediment collected
       above and below the specified depth interval and
       the portion of a sample exceeding the specified
       volume within  a  given  depth  interval were
       considered to be IDW.

    4.  The support equipment cost was estimated based
       on  the rental or purchase cost of any additional
       equipment required for sample collection, such as
       generators or winches needed at the time of the
       demonstration.

Secondary objectives SI, S2, and S3 were addressed in all
the demonstration areas where  a given sampler was
evaluated because no additional sampling was required to
address them.  Secondary objectives  S4 and S5 were not
area-dependent; they were addressed  based on developer
information as  well   as  observations  of  sampler
performance during the demonstration. The approach used
to address each secondary objective is discussed below.
                                                16

-------
•   Secondary objective SI was addressed by observing
    and noting the skills required to operate each sampler
    during the demonstration, how easy the sampler was
    to operate, and the sampler's approximate weight and
    by discussing  any necessary sampling technician
    training with the developer.

•   Secondary objective S2 was addressed by determining
    each  sampler's ability to collect sediment  samples
    given the variety of sampling platforms, water depths,
    sediment  depths, sediment compositions, and  flow
    conditions encountered in the demonstration areas.

•   Secondary objective S3 was addressed based on visual
    observations made during sampling or after a sediment
    sample had been extruded from a sampler.

•   Secondary objective S4 was addressed by noting each
    sampler's  materials  of construction.   Sediment
    sampler failures or repairs that were necessary during
    use of the sampler were also noted.

•   Secondary objective S5 was addressed by discussing
    the availability of replacement  samplers  with the
    developer and determining whether spare parts were
    available  in a  retail  store  or  only  through  the
    developer. In addition, when replacement samplers or
    spare parts were required during the demonstration,
    their  availability was noted.

4.3    Field Sampling and Measurement
       Procedures

This section presents  field sampling and measurement
procedures  used  during  the   Split  Core   Sampler
demonstration.   Specifically, this section summarizes
demonstration sampling locations;   sample collection,
sample preparation, and measurement procedures; and
field QC  procedures. Additional details about the sample
collection,   sample   preparation,   and  measurement
procedures are  presented in the demonstration  plan
(EPA 1999).   The demonstration plan is  available on
the  Internet  through   the  ORD  web   site  (http://
www.epa.gov/ORD/SITE).

Sediment samples were  collected at Site 1  for  PCB
analysis,  at Site 2 for arsenic analysis, and at both sites for
PSD  analysis.     The  sampling  locations  in   each
demonstration area are presented in Figures 4-1  and 4-2.
Table  4-2  lists  the  target  sampling  depth intervals,
numbers  of  investigative  samples,   and  analytical
parameters for each demonstration area and provides the
rationale for their selection. In general, the rationale for
choosing the number of samples to be collected in each
area was based on the objectives to be addressed, the
analyses to  be  conducted  to  address  one  or more
objectives, the time required to collect samples, and the
cost of each analysis.  When five samples were to be
collected in a sampling area, samples were collected in the
four corners and center of the area; when ten samples were
collected in a sampling area, the additional five samples
were   collected  at  locations   randomly  distributed
throughout the area.

Many of the field measurements made to support the
primary objectives (see Section 4.2) were simple, standard
measurements  and do not require additional explanation.
These  measurements  include  the  volume  of IDW
generated, number of sampling technicians, number of
sampling  attempts per location,  volume  of  sediment
collected, time required for sample collection activities,
volume of fuel consumed to operate motorized sampling
or support equipment, core length, sampling area grid size,
and water velocity.  However, several field measurements
were  made  to  address  demonstration-specific
requirements,  and  additional  explanation  of these
measurements  is  warranted to enhance understanding of
the  sampler performance results presented in Chapters 6
and 7.  These field measurements are summarized below
by objective.

•   To  address primary objective PI,  the volume  of
    sediment sample  from a given depth interval was
    measured, and then any unrepresentative material was
    removed from the sediment sample and collected as
    IDW.   Unrepresentative  material  included  sticks,
    shells, and stones. After removal of unrepresentative
    material, if not enough sediment was  left to meet
    analytical sample  volume requirements, the sampling
    technician  collected   additional  cores from  the
    sampling location.

•   To address primary obj ective P2, the depth of sampler
    deployment was measured by allowing the sampling
    technician  to lower the sampler to the surface of the
    sediment.  Once the sampling technician felt that he
    had identified the  sediment surface, a mark was made
    on the sampler cable or extension rod using  a fixed
                                                    17

-------
                     S1A1 (river mouth)
               Target sampling depth intervals:
                0 to 4 inches, 6 to 12 inches,
                       and 4 to 6 feet
                   below sediment surface

           Approximate scale:  1 inch = 1,200 feet
ABODE
1
2
3
4
5
*v.
0



0









0


10 feet





0



0


I
I

    Based on the demonstration design,
   no samples from the 0- to 4-inch and
    4- to 6-foot below sediment surface
depth intervals required laboratory analysis.
                   S1A2 (freshwater bay)
               Target sampling depth intervals:
               0 to 4 inches and  12 to 32 inches
                   below sediment surface

            Approximate scale: 1 inch = 1,200 feet
                                                                              • 10 feet •
   Based on the demonstration design,
     no samples from the 0- to 4-inch
  below sediment surface depth interval
       required laboratory analysis.
                                                                   Legend
                                                                     O   Polychlorinated biphenyls
                                                                      •   Particle size distribution
                                                                       ="" Flow direction
Figure 4-1.  Site 1 sampling locations.
                                                            18

-------
                       S2A1 (lake)
             Target sampling depth intervals:
            0 to 4 inches and 10 to 30 inches
                 below sediment surface

         Approximate scale:  1 inch = 1,200 feet
                                                                  ABODE
1
2
\
«
O
«
O

O

«


O

e
O
•

t
I

                    S2A2 (wetland)
                              Wetland area
             Target sampling depth intervals:
              4 to 12 inches and 9 to 11 feet
                 below sediment surface

            Approximate scale:  1 inch = 20 feet
                                                                                             Sampling locations
                                                                                             are represented by
                                                                                               shaded areas.
                                                                                               Based on the
                                                                                           demonstration design,
                                                                                            no samples required
                                                                                             laboratory analysis.
                                                                    Legend
                                                                      O   Arsenic only
                                                                      ®   Both arsenic and
                                                                          particle size distribution
                                                                    •^— Flow direction
Figure 4-2. Site 2 sampling locations.
                                                            19

-------
Table 4-2.  Rationale for Sampling Approach
Demonstration
Area
S1A1 (river
mouth)
S1A2
(freshwater bay)
S2A1 (lake)
S2A2 (wetland)
Target Sampling
Depth Interval
(bss)
0 to 4 inches
6 to 12 inches
4 to 6 feet
0 to 4 inches
12 to 32 inches
0 to 4 inches
10 to 30 inches
4 to 12 inches
9 to 1 1 feet
Number of Investigative
Samples per Sampler3
(Analytical Parameter)
5 (NA)
5 (PCBs)
1 (PCBs)
5 (NA)
5 (NA)
5 (PSD)
10 (Arsenic)
5 (PSD)
1 (Arsenic)
10 (Arsenic)
5 (PSD)
5 (NA)
5 (NA)
Matrix
Sediment
Sediment
Equipment
rinsate
Sediment
Sediment
Sediment
Sediment
Equipment
rinsate
Sediment
Sediment
Sediment
Rationale
Analytical samples not collected because only primary
objectives P1 (volume), P2 (depth interval), and P6 (sample
collection time) were addressed
Verify that contamination was present
Determine whether a sampler could be adequately
decontaminated (primary objective P5)
Analytical samples not collected because only primary
objectives P1 (volume), P2 (depth interval), and P6 (sample
collection time) were addressed
Analytical samples not collected because only primary
objectives P1 (volume), P2 (depth interval), and P6 (sample
collection time) were addressed
Determine whether a sampler could collect consistent samples
from a homogenous layer of sediment (primary objective P3)
with consistent physical characteristics
Determine whether a sampler could collect consistent samples
from a homogenous layer of sediment (primary objective P3)
with consistent physical and chemical characteristics and
determine whether a sampler could collect sediment samples
from a clean layer of sediment located below a layer of
contaminated sediment (primary objective P4)
Determine whether a sampler could be adequately
decontaminated (primary objective P5)
Determine whether a sampler could collect consistent samples
from a homogenous layer of sediment (primary objective P3)
with consistent physical and chemical characteristics and
determine whether a sampler could collect sediment samples
from a clean layer of sediment located below a layer of
contaminated sediment (primary objective P4)
Analytical samples not collected because only primary
objectives P1 (volume), P2 (depth interval), and P6 (sample
collection time) were addressed
Notes:

bss
NA
PCB
PSD
Below sediment surface
Not applicable
Polychlorinated biphenyl
Particle size distribution
    The number of investigative samples varied depending on the analytical parameters and the objectives addressed in each demonstration area.
    Ten investigative samples were collected and analyzed for arsenic to address primary objectives P3 and P4. However, only five investigative
    samples were collected and analyzed for PSD to address primary objective P3 because the variability associated with PSD is typically less than
    that associated with arsenic concentrations.
    reference point (the water surface, boat side, or boat
    floor). Another mark was made higher on the cable or
    extension rod indicating the depth corresponding to
    the sampling  technician's  estimate of the  depth to
    which the  sampler  should be  driven to collect a
    sediment sample from the  specified sampling depth
                                                       interval. The sampler was then lowered to this depth,
                                                          20

-------
and a sample was collected. For measurement of the
total core  length  retrieved and  the  core length
retrieved in the sampling depth interval, no correction
was made for sample compression or expansion that
might have  taken place during sample collection.

To address primary obj ectives P3 and P4, excess water
overlying   the  sediment  samples  was  carefully
decanted before the samples  were transferred to
stainless-steel bowls and homogenized. The decanting
step  ensured that the sediment  samples would have
adequate percent solids for analysis. Homogenization
involved stirring the material with a stainless-steel
spoon for 4 minutes or longer until the sediment
attained uniform color, texture, and residual water
distribution. Sample containers were then filled using
a quartering technique in  which the homogenized
sample present in the stainless-steel bowl was divided
into quadrants.  Each sample container was  filled by
using a  spoon to alternately transfer sediment from
one quadrant and then from the opposite quadrant
until the sample container was filled.  Any unused
sediment was collected as IDW.

To address  primary objective P5, the nondisposable
components of each sampler were decontaminated by
scrubbing them with an Alconox solution,  washing
them with potable water, and then rinsing them with
deionized water. At Site 1, 3 L of rinsate per sampler
was  generated to meet analytical  and QC volume
requirements  for PCB analysis. At Site 2, 2  L of
rinsate per sampler was generated to meet analytical
and QC volume requirements for arsenic analysis. All
deionized water used to generate rinsate samples was
from one lot of water identified by a lot number.  To
verify  that any  contamination  detected  by  the
laboratory in the rinsate samples was not present in
the  deionized  water or the  result of field sample
collection procedures at Sites 1 and 2, samples of this
water were  sent to the laboratory for PCB and arsenic
analyses, respectively, along with the rinsate  samples.
Deionized water samples were collected once at each
demonstration site  during  collection of sediment
samples.

To address  primary objective P6, timing of sampler
setup began  when a  sampling  technician began
assembling  a given  sampler and  ended when the
sampler was completely assembled and any additional
equipment necessary for sampling using the sampler
    had been collected and was ready to be transported to
    the sampling location. If additional time was required
    to set up the sampler at the sampling location, this
    time  was measured and included  in the total setup
    time.

    Timing of sample collection began when the sampler
    was ready to be deployed and ended when the sample
    had been retrieved; extruded from the sampler; and
    submitted   for  measurement,  preparation,  and
    distribution  into  the appropriate  containers  for
    analysis.    If additional sampling attempts  were
    required to  collect the specified sample volume, the
    time required to complete these attempts was added to
    the sample  collection time.  If any portion  of the
    sampler was disassembled to extrude a sample and
    reassembled before the next sample was collected, the
    time  required for disassembly and reassembly was
    included in the total sample collection time. Between
    sampling attempts and locations, if a sampler had any
    sediment adhering to it, the sampler was rinsed at the
    sampling location using  surface water.  The time
    required for rinsing was also added to the total sample
    collection time.   Sample collection time did  not
    include the time needed to position the  sampling
    platforms at specific sampling locations.

    Timing of  sampler disassembly  began when all
    samples had  been  collected or extruded and  the
    sampling technician began disassembly of the sampler.
    The  timing ended  when  the  sampler  had been
    completely  disassembled and was ready  to be
    decontaminated.

    Timing of sampler decontamination began when the
    nondisposable  components of each sampler were
    decontaminated by scrubbing them with an Alconox
    solution.  The timing continued until the sampling
    technician   considered   the  sampler  to  be
    decontaminated to the degree that a sample of the final
    rinsate could be collected to address primary obj ective
    P5.  Sampler decontamination occurred once in each
    demonstration area after all samples were collected
    and the sampler was disassembled.

QC checks for field measurements were used to evaluate
the quality of field activities.  In general, the QC checks
were used to assess the representativeness of the samples
and to ensure that the degree to which the analytical data
were representative of actual  site conditions was known
                                                21

-------
and  documented.   QC  checks  for  field parameters
consisted  of the  time required  for sample collection
activities and the water velocity.  Field QC checks for
laboratory parameters consisted of temperature blanks (in
shipments that contained samples for PCB analysis) and
field replicates. Field replicates were collected to evaluate
whether a sample was adequately homogenized in the field
prior to filling of sample containers.   Field  replicate
samples included field duplicates (rinsate)  for PCB and
arsenic analyses and field triplicates (sediment) for PCB,
arsenic, and  PSD analyses.  Table 4-3 identifies the
planned numbers  of investigative samples and field
replicate samples. Field replicate samples were submitted
for laboratory analysis as  blind  samples  (that is, the
laboratories did not know which samples were replicates).
Acceptance criteria and associated corrective actions for
field QC checks are presented in the demonstration plan
(EPA 1999).

During the demonstration, the EPA conducted an internal
technical system audit (TSA)  of field sampling  and
measurement systems.  The  following activities were
audited during the field TSA:  sample collection; sample
preparation; field measurements; field documentation;
decontamination;  and sample labeling, packaging,  and
shipping.

A summary discussion of whether the field QC procedures
generated data that met the demonstration objectives  is
presented in Sections 6.3 and  7.3 for the innovative and
reference samplers, respectively.  More detailed infor-
mation is provided in the DER (Tetra Tech EM Inc. [Tetra
Tech] 1999b).

4.4    Laboratory Sample Preparation and
       Analysis Methods

In selecting  appropriate  methods  for preparing  and
analyzing the demonstration samples from Sites 1 and  2,
the  specific analytes of  interest,  the  laboratories'
experience in analyzing the predemonstration samples, and
the  target  reporting  limits  required  to  address  the
demonstration objectives  were   taken into   account.
Table 4-4 summarizes the laboratory sample preparation
and analysis methods used for the demonstration.
Laboratory QC checks  were used to demonstrate the
absence of interferants and contamination from laboratory
glassware and reagents,  to verify that the measurement
systems were in control, to  evaluate the precision and
accuracy  of  laboratory  analyses,  and to  ensure the
comparability of data. Laboratory-based QC checks other
than those associated with instrument calibration consisted
of method blanks, surrogates, MS/MSDs,  extract and
digestate duplicates, blank spike/blank spike duplicates
(BS/BSD), interference check analyses, serial dilutions,
postdigestion spikes, repeat  analyses, and  performance
evaluation (PE) samples.  Table 4-5 summarizes the
laboratory QC checks used for the demonstration and their
purpose.   The frequencies, acceptance  criteria,  and
corrective actions for QC checks are presented  in the
demonstration plan (EPA 1999).

Predemonstration and in-process TSAs of the laboratories
used  for  the  demonstration  were  conducted.    The
following activities were audited:  sample  receipt and
sample  storage;  internal chain  of custody;  sample
extraction, digestion,  and  cleanup; sample  analysis;
standards  preparation  and  storage;  calibration;  QC
procedures; and data reduction, validation, and reporting.

Predemonstration  and in-process performance audits of
laboratory activities were also conducted for PCB and
arsenic analyses. During each audit, (1) two PE samples
(one low-level and one  high-level) each for PCBs and
arsenic were obtained for the  sediment matrix and (2) one
low-level  PE sample  each  for PCBs and arsenic was
obtained for the aqueous matrix.  The PE samples were
submitted to the laboratory as  double-blind samples for
analysis.

A summary discussion  of whether the laboratory QC
procedures generated  data that met  the demonstration
objectives is presented in Sections 6.3 and 7.3 for the
innovative and reference samplers, respectively.  More
detailed information is provided in the DER (Tetra Tech
1999b).
                                                    22

-------
        Table 4-3.  Sample Matrix
Demonstration
Area
S1A1
(river mouth)
S1A2
(freshwater
bay)
S2A1
(lake)
S2A2
(wetland)
Target Sampling
Depth Interval
(bss)
0 to 4 inches
6 to 12 inches
4 to 6 feet
0 to 4 inches
12 to 32 inches
0 to 4 inches
10 to 30 inches
4 to 12 inches
9 to 1 1 feet
Sampler
Split Core Sampler
Vibrocorer
Split Core Sampler
Vibrocorer
Dual Tube Liner Sampler
Vibrocorer
Split Core Sampler
Hand Corer
Split Core Sampler
Hand Corer
Split Core Sampler
Hand Corer
Split Core Sampler
Hand Corer
Split Core Sampler
Hand Corer
Dual Tube Liner Sampler
Analytical
Parameter
NA
PCBs
NA
NA
PSD
Arsenic
PSD
Arsenic
PSD
NA
NA
Sediment Samples
Number Per Sampler
Investi-
gative
Samples
5
5
5
5
5
10
5
10
5
5
5
MS/MSD
Samples3
Field
Triplicate
Samples"
Laboratory
Analyses
Total
Number of
Analyses
Equipment Rinsate Samples
Number Per Sampler
Equipment
Rinsate
Samples
Field
Duplicate
Samples0
Laboratory
Analyses
Total
Number of
Analyses
Samples were not analyzed for PCBs, arsenic, or PSD. The rationale for the number of samples is
provided in Table 4-2.
1
2
11
22
1
1
2
4
Samples were not analyzed for PCBs, arsenic, or PSD. The rationale for the number of samples is
provided in Table 4-2.
Samples were not analyzed for PCBs, arsenic, or PSD. The rationale for the number of samples is
provided in Table 4-2.
NA
2
NA
2
NA
1
3
1
3
1
7
20
7
20
7
14
40
14
40
14
NA
1
NA
0
NA
NA
1
NA
0
NA
0
2
0
0
0
0
4
0
0
0
Samples were not analyzed for PCBs, arsenic, or PSD. The rationale for the number of samples is
provided in Table 4-2.
Samples were not analyzed for PCBs, arsenic, or PSD. The rationale for the number of samples is
provided in Table 4-2.
to
OJ
        Notes:

        bss
        MS/MSD
Below sediment surface
Matrix spike/matrix spike duplicate
NA    =   Not applicable
PCB   =   Polychlorinated biphenyl
PSD   =   Particle size distribution
             MS/MSD samples were collected for PCB and arsenic analyses and were designated in the field.  MS/MSD samples were not collected for equipment rinsate samples because the additional
             volume required for the analysis may have diluted any contamination present to concentrations below laboratory detection limits. Sediment MS/MSD samples did not require additional sample
             volume.

             Field triplicate sediment samples were collected by filling three sample containers with homogenized sediment. A sufficient volume of sediment for field triplicate samples was collected as
             described in the approach for addressing primary objective P1 in Section 4.2. Field triplicate samples were submitted for analysis as blind samples.
             Field duplicate equipment rinsate samples were collected by filling one additional container for PCB or arsenic analysis.  Field duplicate samples were submitted for analysis as blind samples.

-------
Table 4-4. Laboratory Sample Preparation and Analysis Methods
Parameter (Matrix)
                                       Method Reference3
                                                                               Method Title
PCBs (sediment)
PCBs (equipment rinsate)
Arsenic (sediment)
Arsenic (equipment rinsate)
PSD (sediment)
                  SW-846 Method 3550B (extraction)

                  SW-846 Method 3665A" (cleanup)

                  SW-846 Method 36606° (cleanup)

                  SW-846 Method 8082 (analysis)

                  SW-846 Method 351OC (extraction)

                  SW-846 Method 3665Ab (cleanup)

                  SW-846 Method 8082 (analysis)

                  SW-846 Method 3050B (digestion)

                  SW-846 Method 601 OB (analysis)

                  SW-846 Method 301OA (extraction)


                  SW-846 Method 601 OB (analysis)

                  ASTM Method D 422-63
                  (Reapproved in 1990)
                                                                   Ultrasonic Extraction

                                                                   Sulfuric Acid/Permanganate Cleanup

                                                                   Sulfur Cleanup

                                                                   PCBs by Gas Chromatography

                                                                   Separatory Funnel Liquid Extraction

                                                                   Sulfuric Acid/Permanganate Cleanup

                                                                   PCBs by Gas Chromatography

                                                                   Acid Digestion of Sediment, Sludges, and Soils

                                                                   Inductively Coupled Plasma-Atomic Emission Spectrometry

                                                                   Acid Digestion of Aqueous Samples and Extracts for Total Metals
                                                                   for Analysis by FLAA or ICP Spectroscopy

                                                                   Inductively Coupled Plasma-Atomic Emission Spectrometry

                                                                   Standard Method for Particle-Size Analysis of Soils
                                                                   (with hydrometer option)
Notes:

ASTM
EPA
FLAA
ICP
PCB
PSD
American Society for Testing and Materials
U.S. Environmental Protection Agency
Flame atomic absorption
Inductively coupled argon plasma
Polychlorinated biphenyl
Particle size distribution
    SW-846 reference: EPA 1996; ASTM reference: ASTM 1998

    SW-846 Method 3665A is used whenever elevated baselines or overly complex chromatograms prevent accurate quantitation of Aroclors.  The
    laboratory routinely performed sulfuric acid cleanup on PCB sample extracts using SW-846 Method 3665A.

    The laboratory detected elevated  levels of sulfur in  predemonstration investigation samples analyzed for PCBs.  Therefore, the laboratory
    monitored PCB chromatograms for the presence of sulfur and cleaned up the extracts using SW-846 Method 3660B when sulfur was detected.
                                                              24

-------
Table 4-5.  Laboratory Quality Control Checks
Quality Control Check
             Parameter
        Matrix
                          Purpose
Method blanks


Surrogates



MS/MSDsa


Extract duplicates


Digestate duplicates


BS/BSDs
Interference check
analyses

Serial dilutions
Postdigestion spikes

Repeat analyses

PE samples
         PCBs and arsenic   Sediment and rinsate
         PCBs
Sediment and rinsate
         PCBs and arsenic   Sediment
         PCBs


         Arsenic
Sediment and rinsate


Sediment and rinsate
         PCBs and arsenic   Sediment and rinsate
         Arsenic


         Arsenic
Sediment and rinsate


Sediment and rinsate
         Arsenic            Sediment and rinsate

         PSD               Sediment

         PCBs and arsenic   Sediment and water
Verify that steps in the analytical procedures did not introduce
contaminants that affected analytical results

Determine whether significant matrix effects existed within the
samples and measure the efficiency of recovery of analytes in
sample preparation and analysis

Determine the accuracy and precision of the analytical results with
respect to the effects of the sample matrix

Determine the precision associated with laboratory analytical
procedures following sample extraction

Determine the precision associated with laboratory analytical
procedures following sample digestion

Determine whether observed deviations for MS/MSDs and for
extract and digestate duplicate samples were caused by a matrix
effect

Evaluate the validity of the interelement correction factors
Determine whether significant physical or chemical interferences
existed as a result of the sample matrix

Determine whether a matrix effect should be expected

Evaluate the precision of hydrometer readings

Determine the accuracy associated with the laboratory analytical
procedures for low-level and high-level concentrations
Notes:

BS/BSD   =
MS/MSD  =
PCB
PE
PSD
Blank spike/blank spike duplicate
Matrix spike/matrix spike duplicate
Polychlorinated biphenyl
Performance evaluation
Particle size distribution
     MS/MSD samples were not collected for equipment rinsate samples because the additional volume required for the analysis may have diluted
     any contamination present to concentrations below laboratory detection limits.  In addition, MS/MSDs are not typically collected for rinsate
     samples.
                                                                25

-------
                                              Chapter 5
                       Description of the Reference Sediment Samplers
This  chapter  describes  two  conventional  sediment
samplers that were used as reference samplers during the
demonstration. Each reference sampler was chosen based
on its proven ability to meet the various demonstration
objectives presented in Section 4.1.  Specifically, two core
samplers were selected  as reference samplers: the Hand
Corer and the Vibrocorer.

The  Hand  Corer is  a commonly used core  sampler
designed to obtain sediment samples in a variety of lake
and  river environments.   The   sampler  can collect
continuous sediment cores to a depth of about 36 inches
bss.  Based on the predemonstration investigation results,
demonstration objectives, and  site  support  facilities
available, the Hand Corer was selected as the reference
sampler for S1A2, S2A1, and S2A2.

The  Vibrocorer  is a core sampler designed to obtain
sediment samples in a variety of shallow and deep river,
lake, and ocean  environments.   The  sampler  has been
successfully used by the EPA at several contaminated sites
in Region 5.  Based on the predemonstration investigation
results,  demonstration   objectives,  and  site  support
facilities available, the  Vibrocorer was selected as the
reference sampler for S1A1.

Sections 5.1 and 5.2 provide descriptions, discuss general
operating  procedures,  and  outline   advantages  and
limitations of the Hand Corer and Vibrocorer used in the
demonstration.

5.1     Hand Corer

The Hand Corer selected as a reference sampler for the
demonstration  is  designed to   collect undisturbed,
cylindrical core samples from various types of sediment,
including saturated sands and silts, to a depth of about
36 inches bss in stagnant or swiftly moving water.

5.1.1  Technology Description

Components of the Hand Corer include (1) a Lexan™ nose
piece; (2) a 36-inch-long, stainless-steel core tube; (3) a
stainless-steel  head piece with a flutter valve; (4) two
detachable, stainless-steel handles; and (5) a clevis (see
Figure 5-1).  For deployment in deep water, the Hand
Corer can be equipped with a guide rope or extension rods
and  a turning handle.   The Hand  Corer  can also  be
equipped with disposable, clear plastic core tube liners
that fit inside the core tube (these liners are not shown in
Figure 5-1).

Support equipment for sampler deployment may include
a tripod-mounted winch for (1) controlling the rate  of
sampler deployment and retrieval; (2) minimizing the
physical stress on the  sampling technician, particularly
during sampler retrieval and during intense or extended
sampling events; and (3) preventing the sampler from
sinking  too deeply  into  the  sediment  to  obtain  a
representative sample.

The stainless-steel core tube has a 2-inch outside diameter
and is designed to collect about 50 mL of sediment per
inch of core tube length; the maximum design volume of
the core tube is about 1,800 mL. The fully equipped Hand
Corer, including the nose piece, core tube, head piece with
flutter valve, handles, and clevis, weighs about  12 Ib.
Each 5-foot-long extension rod and a turning handle weigh
about 5 and 2 Ib, respectively.

In water less than 20 feet deep, the Hand Corer may  be
manually deployed and driven into the sediment using the
                                                   26

-------
   Flutter valve
                                   Clevis
  Detachable
   handles
                                       • Head piece
                                    , Core tube
       Nose piece
Figure 5-1. Hand Corer.
handles and the necessary length of extension rods.  In
water more than 20  feet deep, the  sampler  may  be
deployed using a guide rope attached to the clevis and a
weight attached to the core tube. During sampler retrieval,
a sediment core is retained within the core tube by a partial
vacuum created by the closed flutter valve.

5.1.2   General Operating Procedures

The Hand Corer can be operated in shallow water by one
person from a platform, from a boat, or while wading. For
sampling in deep water, two sampling technicians are
recommended to control the weight of the sampler and
extension rods and to conduct efficient sampling. During
sampler assembly, a plastic core tube liner may be inserted
into the core tube.  Core tube liners hold and  store the
sample for later examination.  Depending on the water
depth and flow  conditions,  either the handles and the
necessary number of extension rods or the guide rope can
be used to deploy the Hand Corer to the sediment surface.
The speed of sampler deployment to the sediment surface
should be controlled to avoid bow wave formation, which
could disturb flocculent or unconsolidated sediment that
might be near the sediment surface (Blomqvist 1991).

The sampler may be driven into the sediment by manual
force on the handles or by gravity penetration. In general,
the sampler should be driven  into the sediment in a steady
and uninterrupted manner.   The sampler is manually
retrieved by pulling upward on the handles, extension
rods, or guide rope, as appropriate.  When samples are
being collected in shallow water depths, the flutter valve
should be manually closed once the Hand Corer reaches
the desired sediment depth.  When the sampler is being
retrieved from deep water depths, the upward motion of
the submerged  sampler causes  the flutter valve  to
automatically close.  The tapered nose piece and partial
vacuum created by the flutter valve  retain the sediment
core within the plastic core tube liner. When the weight of
the sampler and extension  rods  requires  it, a tripod-
mounted winch  should be used to  control the rate of
sampler retrieval. The sampler should be kept vertical and
the rate of retrieval should be kept as steady as possible to
minimize resuspension  and disruption of the sediment.

After sampler retrieval, the nose piece or head piece is
removed to allow removal of the plastic core tube liner.
The sediment core enclosed in the core tube liner may be
either sealed in  the core tube  using two core caps or
extruded for further examination and processing. The
sediment core may be removed by pushing the sample out
one end of the core tube liner  with an extrusion rod. Prior
to sampling, some sampling technicians cut the core tube
liner twice longitudinally and tape the liner together with
vinyl electrical tape before inserting the liner into the core
tube.  In this case, after a sample is collected, the tape
holding the two halves of the  core tube  liner is  cut,
splitting the liner in half and exposing the sediment core.
                                                    27

-------
5.1.3  Advantages and Limitations

An advantage of the Hand Corer is that it is easy to
operate, requiring minimal skills and training. Sampler
assembly and sample collection procedures can be learned
in the field with a few practice attempts. In addition, a
written SOP typically accompanies the sampler when it is
procured.  The sampler can be operated by one person in
shallow (wading) water depths because of its  light weight
(12 Ib). Sampler operation is especially simple when a
core tube liner is used  because the sampler does not
require complete disassembly to extrude the  sample and
reassembly after each sampling  attempt.  Only the nose
piece or head piece requires detachment to  remove the
plastic core tube liner containing the  sediment core. Use
of the disposable liner also minimizes the risk of cross-
contamination between sampling locations.

Another advantage of the Hand Corer is the flutter valve
in the head piece. The flutter valve is designed to allow
water to  exit the top of the core tube during sampler
deployment,  thus minimizing   potential   bow  wave
formation near the sediment surface.  During sampler
retrieval, the sediment core is retained within the core tube
by a partial vacuum created by the closed flutter valve.
Collectively, these design features increase the likelihood
of collecting an undisturbed sample.

A limitation of the Hand Corer is that during sampler
deployment, the plastic  core  tube  liner is  exposed to
different layers of sediment contamination. Contaminants
may adhere to the exposed surface of the liner while the
sampler passes through different layers of sediment. Also,
the flutter valve may become clogged if the sampler is
deployed in such a way that the flutter valve is driven
into the sediment. Specifically, sediment and nonsed-
imentaceous materials (leaves, plant roots, or small stones)
may become trapped between the flutter valve and core
tube, resulting in partial or complete  loss of vacuum and
eventually partial or complete loss of the sediment sample.

Another limitation of the Hand Corer is that it  cannot
collect discrete samples  from various  sediment  depths.
Core samples must be collected from the sediment  surface
downward. Because end-filling samplers such as the Hand
Corer must collect samples from the sediment  surface
downward, the Hand  Corer is subject to core shortening.
Core shortening occurs when the length of sediment core
collected is less than the depth of sampler penetration into
the sediment.   Core  shortening may occur when the
friction of the sediment against the inside wall of the core
tube  increases  with  increasing depth  of  sediment
penetration, causing lateral displacement of sediment and
resulting in gradually thinner  increments of sediment
entering the sampler. Because not all layers are uniformly
sampled, core shortening can introduce sampling bias.

Furthermore, use of a tripod-mounted winch limits the
sampling platform locations from which the sampler can
be deployed. Specifically, the sampling platform must be
equipped with a hole over which the tripod-mounted
winch can be placed and through which the sampler can be
deployed.

5.2    Vibrocorer

The Vibrocorer is designed to collect sediment cores in
deep river, lake, and ocean environments.  The sampler is
designed to operate in shallow and deep water conditions
and to provide complete and continuous sediment profile
collection to a maximum depth  of 4,000 feet beneath the
water surface. According to the EPA GLNPO, the sampler
is designed to collect sediment cores to a depth of 15 feet
bss in packed sand and to a depth of 20 feet bss in silt and
clay;  however,  sediment  cores have  been successfully
collected to a depth of 35 feet bss using the Vibrocorer.

5.2.1  Technology Description

Components of the Vibrocorer  include (1) an anodized-
aluminum, pressure-housed vibrohead with a terminal for
an electric cable; (2) a disposable, 10-foot-long, 4-inch-
diameter, clear plastic core tube; (3) a core tube clamp;
and (4) a guide rope (see Figure 5-2).  The sampler is also
equipped with a check valve in  the vibrohead and a core
nose at the  bottom end of the core tube (the check valve
and core nose are not shown in Figure 5-2).  Core tube
sectioning and extraction are performed using a hand-held
or battery-powered electric saw. The Vibrocorer requires
a three-phase,  230- or 440-volt, 50- to 60-hertz electric
current. The sampler must be supplied with power from a
power source through an electric cable and a control box.
The Vibrocorer must be operated from a boat, dock, or
platform with enough working space to accommodate an
A-frame of adequate size.
                                                   28

-------
                                      Guide rope
                                           Vibrohead
 Electric cable
    Core tube clamp
      Core tube
Figure 5-2. Vibrocorer.
The  typical weight  of a fully equipped  Vibrocorer,
including the vibrohead and  core tube, is about 150 Ib.
Core tubes are available in lengths up to  15 feet with a
4-inch diameter and up to 20  feet with a 3-inch diameter.
If a 15-foot-long core sample is required, the core tube
must be 16 feet long because 6 inches is lost when the core
tube is inserted into the vibrohead and 6 inches is lost
when the core nose is attached.

The Vibrocorer is deployed to the sediment surface using
the A-frame and  winch.  Once the sampler is deployed to
the  sediment  surface and  supplied with  power, the
vibrohead vibrates at a frequency of up to 3,450 vibrations
per minute, depending on the power supply. The vibrating
motion of the vibrohead drives the core tube vertically
downward  into the sediment.  The sampler is retrieved
mechanically using the  A-frame  and winch.  During
sampler retrieval, the check valve in the vibrohead creates
a vacuum that, along with the core nose, retains sediment
within the core tube.

5.2.2   General Operating Procedures

The Vibrocorer must be operated by at least two persons
from a boat, dock, or platform. To prepare for sampler
deployment, the vibrohead is raised using the A-frame and
winch, and the core tube is secured to the vibrohead at the
core tube clamp.  Again using the A-frame and winch, the
sampler is deployed to the desired sampling position; the
vibrohead should then be supplied with power and allowed
to vibrate.   The speed  of sampler deployment to the
sediment surface should be controlled to avoid bow wave
formation,  which  could  disturb  flocculent   or
unconsolidated sediment that might be near the sediment
surface (Blomqvist 1991).

As the vibrohead vibrates, the core tube is gradually forced
downward  into the sediment.  Once the  core tube is
deployed to the desired sediment depth, the power can be
turned  off and the vibrohead can be allowed to stop
vibrating. Now the sampler can be mechanically removed
from the sediment using the A-frame and winch. During
sampler retrieval, the check valve in the vibrohead creates
a vacuum that, along with the core nose, retains sediment
within the core tube. Once  the core tube is retrieved from
the  water,  water remaining in the top of the core tube
should be drained by drilling holes in the core tube at the
sediment-water interface  with an electric or  battery-
powered drill. To remove the core tube from the core tube
clamp, four nuts that secure the core tube in place must be
removed.   Afterward, the core tube is placed on the
sampling platform to extract the sediment.  To extract the
core sample, horizontal sections of the core tube should be
cut using an electric or battery-powered saw.
                                                    29

-------
5.2.3  Advantages and Limitations

Advantages of the Vibrocorer include its ability to collect
sediment samples up  to 4,000  feet beneath the water
surface.  In addition,  the vibrohead component of the
sampler allows  core tube penetration into the sediment
without manual  labor.  Sampler deployment and retrieval
are controlled with an A-frame and winch. Furthermore,
use  of new core tubes  for  each sampling  attempt
minimizes  the  risk  of  cross-contamination between
sampling locations.

A limitation of the Vibrocorer is that during  sampler
deployment, the disposable core tube  is exposed to
different layers of sediment contamination. Contaminants
may  adhere to the exposed surface of the core tube while
the sampler passes through different layers of sediment.
In addition, the sampler cannot collect discrete samples
from  various sediment  depths;  core samples must be
collected from the sediment surface downward.  As a
result, samples collected with the Vibrocorer are subject
to core shortening as described in Section 5.1.3.

Another limitation of the Vibrocorer is that  it must be
operated by at least two persons from a boat, dock, or
platform. If the sampler is being operated from a boat and
the boat drifts away from the deployed Vibrocorer, the
tension on the winch cable could pull the Vibrocorer over
and damage it, or the electric cable could snap and cause
an electrical short circuit. Also, if the boat drifts while the
Vibrocorer is deployed, extracting the core tube from the
sediment would be difficult because the winch cable from
the sampler to the boat would not be vertical;  as a result,
the core tube could be bent and the sediment sample could
be lost.
                                                    30

-------
                                               Chapter 6
                             Performance of the Split Core Sampler
To  verify a wide range of performance attributes, the
innovative  sediment sampler demonstration had  both
primary and secondary objectives.  Primary objectives
were  critical to the technology  evaluation and were
intended to  produce  quantitative  results regarding
technology performance. Secondary objectives provided
information that was useful but  did  not  necessarily
produce  quantitative  results  regarding   technology
performance. The approach used to address each primary
and secondary objective for the  Split Core Sampler and
reference samplers is discussed in Chapter 4. This chapter
describes the performance of the Split Core Sampler based
on the primary obj ectives (excluding costs associated with
sample collection activities) and secondary objectives.
This  chapter  also  discusses  the  data  quality  of
demonstration results for the Split Core Sampler.

The performance of the reference samplers is discussed in
Chapter  7,  costs  associated with sample collection
activities (primary objective P7) are presented in Chapter
8, and the  performance of the Split Core Sampler and
reference samplers  is compared in summary  form in
Chapter 9.

6.1    Primary Objectives

This section discusses the performance results for the Split
Core Sampler based on the primary objectives stated in
Section 4.1 except for primary  objective P7 (sampling
costs), which is addressed in Chapter  8.   Primary
objectives PI through P6 required evaluation of the  Split
Core Sampler's

P1.   Ability to consistently collect a specified volume of
     sediment
P2.  Ability to  consistently  collect  sediment in  a
     specified depth interval

P3.  Ability to collect multiple samples with consistent
     physical or chemical characteristics, or both, from a
     homogenous layer of sediment

P4.  Ability to collect a representative sample from a
     clean sediment layer below a contaminated sediment
     layer

P5.  Ability to be adequately decontaminated

P6.  Time requirements for sample collection activities

To address primary objectives PI through P6,  samples
were collected from four different areas: (1) S1A1, a river
mouth; (2) S1A2, a small, freshwater bay; (3)  S2A1, a
lake; and (4) S2A2, a wetland. A sampling technician
designated by AMS used the Split Core Sampler to collect
samples from the  following target depth intervals: 0 to
4 and 6 to  12 inches bss in S1A1, 0  to 4 and 12 to
32 inches bss in S1A2, 0 to 4 and 10 to  30 inches bss in
S2A1, and 4 to 12 inches bss in S2A2.  Multiple depth
intervals were simultaneously sampled in a given attempt
if the sampler was long enough to reach these intervals.
For  example,  in   S2A1,  sediment  samples  were
simultaneously collected in the 0- to 4- and 10- to 30-inch
bss depth intervals until sample volume requirements were
met for the 10-to 30-inch bss depth interval. If additional
sample volume was still needed for the  0- to 4-inch bss
depth interval, additional sampling attempts were made in
only that depth interval. Because the Split Core  Sampler
is not designed to collect samples at a depth below 4 feet
bss, the sampler was not used in the 4- to 6-foot bss depth
interval in S1A1 or in the 9- to  11-foot bss depth interval
in S2A2.   The demonstration areas and target depth
                                                   31

-------
intervals are described in greater detail in Chapters 3 and
4. The numbers of investigative and QC samples collected
in each demonstration area, sediment sample volumes
required, and sample analytical parameters are discussed
in Chapter 4.

During the demonstration, AMS used two different Split
Core  Sampler core tube  lengths to collect sediment
samples: 6 and 12 inches. The Split Core Samplers were
configured by screwing togetherthe appropriate number of
6- and 12-inch-long core tubes to achieve the desired core
tube length for a given sampling scenario.  Both the 6- and
12-inch-long core tubes had a  2-inch inside diameter.
AMS chose which core tubes to use based on site and area
conditions and sampling requirements identified in the
demonstration plan. The sampling technician was also
provided an opportunity to practice sample collection at
each  demonstration until  he  felt confident  enough to
initiate demonstration sampling.  The Split Core Sampler
is described in Chapter 2.

The  demonstration  results for the  Split Core Sampler
under primary objectives PI, P2, and P4  were evaluated
using the Wilk-Shapiro test to  determine whether the
results were normally  distributed. Because most of the
results were not normally distributed, the Wilk-Shapiro
test was used in an attempt to evaluate whether the results
followed a lognormal distribution. The test revealed that
the results either were not lognormally distributed or could
not be tested for lognormality because  they contained
values that were equal to zero. For these  reasons, a
parametric test such as the paired Student's t-test was not
used to perform hypothesis testing. The Wilcoxon signed
rank test, a nonparametric test for paired samples that
makes no assumptions regarding distribution, was used as
an  alternative to the  Student's t-test.   Although the
Wilcoxon signed rank test has been historically accepted
as a nonparametric test, it is not as powerful as the
Student's t-test because the Wilcoxon signed rank test
does not account for the magnitude of difference between
sample pair results. Despite this limitation, the Wilcoxon
signed rank test was more appropriate than the Student's
t-test for evaluating the demonstration results. A computer
program known  as Statistix® for Windows, Version 2.0
(Statistix®)  developed  by  Analytical  Software  of
Tallahassee,  Florida,  was used to  perform statistical
evaluations  of the demonstration  results  (Analytical
Software 1996).  Appendix C provides details  on the
statistical methods used for data evaluation.
6.1.1  Ability to Consistently Collect a Specified
        Volume of Sediment

Primary objective PI involved evaluating the Split Core
Sampler's ability to consistently collect a specified volume
of sediment. This objective was addressed by comparing
(1) the actual number of sampling attempts required to
collect a specified volume of sediment to the expected
number of attempts (rounded to the nearest higher integer)
at each sampling location in each target depth interval and
(2)  the actual  volume  of sediment  collected in  the
specified target depth interval in  each attempt to  the
calculated sampler volume (design volume) for the depth
interval.    The expected  number of attempts  was
determined by dividing the specified sample volume by the
design volume for the depth interval. The results of these
comparisons are summarized below.

6.1.1.1 Number of Sampling Attempts Required

Tables 6-1 and 6-2 present the expected and actual number
of sampling attempts for each depth interval at Sites 1 and
2, respectively.  Initially, the Wilcoxon signed rank test
was used to determine whether the difference  between the
expected and actual number of attempts was statistically
significant.  However, the conclusions drawn from  the
Wilcoxon signed  rank test were  inconsistent with  the
conclusions reached in comparing the expected and actual
number of attempts.  This discrepancy was primarily due
to the test's inability to account for the magnitude of the
difference between data pairs (see Appendix C for an
example).

Based on the number of sampling attempts required in
S1A1, the Split  Core Sampler performed well in the 0- to
4- and  6- to  12-inch bss  depth intervals,  where  the
expected number of attempts equaled the actual number of
attempts.

The  Split Core Sampler's performance in  S1A2 was
similar to that  in S1A1.  One additional attempt was
required in the 12- to 32-inch bss depth interval at one of
the five sampling locations in S1A2.

In S2A1, which was the first area sampled during  the
demonstration, the Split Core Sampler performed well in
the 0-to 4-inch bss depth interval but did not perform well
in the 10- to 30-inch bss depth interval.  In the 0- to 4-inch
bss  depth interval,  the  Split Core  Sampler  required
                                                    32

-------
Table 6-1.  Comparison of Expected and Actual Number of Sampling Attempts for Split Core Sampler at Site 1

                                             Number of Attempts in S1A1 (River Mouth)

Location
1A
1E
3C
5A
5E
Total
0- to 4-Inch bss
Expected
1
1
1
1
1
5
Depth Interval
Actual
1
1
1
1
1
5
6- to 12-Inch bss
Expected
1
3
1
3
1
9
Depth Interval
Actual
1
3
1
3
1
9
                                            Number of Attempts in S1A2 (Freshwater Bay)
                           0- to 4-Inch bss Depth Interval
Location
1A
1E
3C
5A
5E
Total
Expected
1
1
1
1
1
5
Actual
1
1
1
1
1
5
Expected
1
1
1
1
1
5
Actual
1
1
1
2
1
6
                 12- to 32-Inch bss Depth Interval
Note:

bss  =  Below sediment surface
36  attempts,  whereas  33  attempts  were  expected.
Specifically, one additional attempt was required in the 0-
to 4-inch bss depth interval at three of the ten sampling
locations. In the  10- to 30-inch bss depth interval, the
Split  Core  Sampler  required  20  attempts, whereas
10 attempts were expected.  However, the actual number
of attempts exceeded the expected number of attempts by
more than one attempt at only three of the ten sampling
locations. Forthe 10-to 3 0-inch bss depth interval at four
of the ten locations, one attempt was added to the actual
number of recorded attempts because the volume of
sediment retrieved was  less than  the volume required.
Because  the samples were  extruded in the sample
management area and  not on the sampling platform and
because the demobilization activities for the day  were
completed before sample extrusion began, a field decision
was  made not to  collect  the  additional volumes of
sediment required.  The addition of one  attempt was
based on the  average volume  of sediment  (375  mL)
collected in each attempt in the 10- to 30-inch bss depth
interval in S2A1, which was greater than the deficit
(100 mL) at any of the four locations. The additional
attempts in this depth interval may  be attributable to
(1) error in assessing the location ofthe sediment surface,
which  might  have resulted  in the  actual depth  of
penetration  being less  than the measured depth  of
penetration; (2) deficient entry of sediment into the core
tube (core shortening); (3) the sediment consisting of high
levels of silt (50 to 67 percent), which might have caused
plug formation in the coring tip that inhibited  further
sediment retrieval; or (4) sediment loss during sampler
retrieval.

In S2A2, the Split Core Sampler performed well in the 4-
to 12-inch bss depth interval, where the expected number
of attempts equaled the actual number of attempts.

Based on the number of sampling attempts required in all
four demonstration areas and multiple sampling depth
intervals, the Split Core Sampler demonstrated the ability
to consistently collect a specified volume of sediment
except in the 10- to 30-inch bss depth interval in S2A1.
Overall, the  sampler required  only  19  percent more
attempts than  expected  (86  actual  attempts  versus
                                                    33

-------
Table 6-2. Comparison of Expected and Actual Number of Sampling Attempts for Split Core Sampler at Site 2
                                                Number of Attempts in S2A1 (Lake)

Location
1A
1B
1E
2A
2C
2D
2E
3A
3B
3E
Total
0-to
Expected
3
2
3
4
3
4
4
3
2
5
33
4-Inch bss Depth Interval
Actual
3
2
3
5
3
5
4
3
2
6
36
10- to 30-Inch bss
Expected
1
1
1
1
1
1
1
1
1
1
10
Depth Interval
Actual3
1
1
3
2
2
2
4
1
1
3
20
Location
Total
                        Number of Attempts in S2A2 (Wetland)
                           4-to 12-Inch bss Depth Interval
Expected
Actual
1A
1E
3C
5A
5E
1 1
1 1
1 1
1 1
1 1
Notes:

bss  =   Below sediment surface

a    At sampling locations 1E, 2C, 2E, and 3E, one attempt was added to the actual number of recorded attempts in order to account for the sample
    deficit compared to the sample volume required.  Refer to Section 6.1.1.1 for additional explanation.
72  expected  attempts).    The  Split  Core  Sampler
performance results are comparable to those presented in
Section 7.1.1.1 for the reference samplers at Site 1 and
superior to those for the reference sampler at Site 2.

6.1.1.2  Volume of Sediment Collected

The  volume  of sediment collected  by the Split Core
Sampler in  each  sampling  attempt in a given  depth
interval was divided by the corresponding design volume,
and the  resulting ratio was multiplied by 100 to estimate
the percent sample recovery (PSR). The relative standard
deviation (RSD) of the PSRs was calculated to evaluate
                                the  ability of the  sampler to consistently  collect a
                                specified volume of sediment; if the sampler were to
                                recover an identical volume of sediment in every attempt,
                                the  RSD would  equal zero. To properly evaluate the
                                sampler's performance, both PSRand RSD results should
                                be considered because a low RSD, which indicates the
                                sampler's performance was consistent, may be based on
                                consistently low PSRs. Figures 6-1 and 6-2 present PSRs
                                for the Split Core Sampler at Sites 1 and 2, respectively.
                                Table 6-3 presents PSR summary statistics (range, mean,
                                and RSD) for the  Split Core  Sampler at both  Sites 1
                                and 2.

                                Based on the PSR information  for S1A1, the Split Core
                                Sampler performed well in the  0- to 4- and 6- to  12-inch
                                                    34

-------
         S1A1,
                    S1A1 (river mouth)
       Total number of attempts: 5
         Oto20     >20to40     >40to60     >60to
                  Percent sam|Jtellfecovery
        S1A2,
                                                                Total number of attempts: 5
        Oto20      >20to40    >40to60    >60to

                 Percent sampflsifecovery
                                                                                                    >80to100
         Oto20     >20to40     >40to60     >60to80    >80to100
                  Percent samfffSRecovery

Note:

bss  =   Below sediment surface

Figure 6-1. Percent sample recoveries for Split Core Sampler at Site 1.
                                                                  Oto20
                 >20 to 40     >40 to 60     >60 to 8
                 Percent samfftSRiecovery
                                                                                                    >80to100
bss depth intervals.  Specifically, as shown in Table 6-3,
each attempt in the 0- to 4-inch bss depth interval had a
PSR of 100, and the PSRs for the 6- to 12-inch bss depth
interval ranged from 83 to 100 with a mean PSR of 93.
For the 0- to 4-inch bss depth interval, the RSD was zero.
For the 6- to  12-inch bss depth interval, the RSD was low
(10 percent) because the recoveries fell in a narrow range
(83 to 100 percent).  Although no RSD criterion has been
set for determining the ability to consistently sample a
specified volume of sediment, an RSD of 30 percent or
less is considered to be acceptable.  Based on the RSDs,
the Split Core Sampler was able to consistently sample the
0- to 4- and 6- to  12-inch bss depth intervals in S1A1.

In S1A2, the  Split Core Sampler performed well in the 0-
to 4-inch bss depth interval but did not perform well in the
12- to 32-inch bss depth interval.  In the 0- to 4-inch bss
depth interval, the Split Core Sampler achieved a PSR of
100 in every attempt. However, in the 12- to 32-inch bss
depth interval, PSRs ranged from 20 to 60 and had a mean
of  only  37, as  shown in Table 6-3.   As shown  in
Figure 6-1, four of the six attempts in this interval fell in
the greater than 20 to 40 percent range,  and  one attempt
fell in each of the 0 to 20 and greater than 40 to 60 percent
ranges.  In the 12- to 32-inch bss depth interval, the RSD
was 37 percent, which exceeded the 30 percent RSD
guideline.  The low recoveries in  the 12- to  32-inch bss
depth interval may be attributable to (1) error in assessing
the location of the sediment surface, which might have
resulted in the actual depth of penetration being less than
the measured depth of penetration; (2) core shortening;
(3) the sediment consisting of high levels of silt and clay,
                                                      35

-------
         S2A1, Qi
         Oto20   >20to40   >40to60   >60to80   >80to100    >100

                 Percent sample recovery
        S2A1,
         Oto20    >20to40   >40to60   >60to80   >80to100   >100
                 Percent samptSBecovery
        S2A2,
                    S2A2 (wetland)
        Total number of attempts: 5
        Oto20    >20to40   >40to60   >60to80   >80to100   >100
                 Percent samptsrecovery
Notes:

bss  =  Below sediment surface

Percent sample recoveries exceeding 100 resulted from either the
volumetric measurement error associated with the presence of void
spaces when the sediment was transferred to a graduated container or
sediment compaction in the core tube.

Figure 6-2.  Percent sample recoveries for Split Core Sampler at
          Site 2.
which might have caused plug formation in the coring tip
that inhibited further sediment retrieval; or (4) sediment
loss  during sampler retrieval.  Based on the RSDs, the
Split Core Sampler was able to consistently sample the 0-
to 4-inch bss depth interval.

In S2A1, the Split Core Sampler performed well in the 0-
to 4-inch bss depth interval but did not perform well in the
10- to 30-inch bss depth interval. As shown in Table 6-3,
PSRs for the 0- to 4-inch bss depth interval ranged from 25
to 125 with a mean of 89.  As shown in Figure 6-2, 23 of
the 36 attempts in this interval had PSRs greater than 80,
and  31  of the  36 attempts had PSRs greater than 60.
Because most of the PSRs fell in a narrow range, the  RSD
of 26 percent was less than the 30 percent RSD guideline.
In the 10- to 30-inch bss depth interval, the PSRs  ranged
from 0 to 75 with a mean of 38. As shown in Figure 6-2,
12 of the 16 attempts in  this interval had PSRs in the
greater than 20 to 40 and greater than 40 to 60 ranges. An
RSD of 51 percent was calculated for the 10- to 30-inch
bss depth  interval, which  indicates a  high degree of
inconsistency.   The  failures in  this interval may be
attributable to the reasons cited above for S1A2  except
that in S2A1, the sediment did not consist of as much clay
as did  the sediment  in  S1A2 and thus  provided less
opportunity for plug formation.

In S2A2, the Split Core Sampler performed well in the 4-
to 12-inch bss depth interval. A PSR of 100 was achieved
in every attempt in this interval, resulting in an RSD of
0 percent.

Based on the volumes of sediment collected in all four
demonstration areas and multiple sampling depth intervals,
the  Split Core Sampler  demonstrated  the ability to
consistently collect a specified volume of sediment. An
RSD below 30 percent was observed for five of the seven
sampling depth intervals.  Of the two remaining depth
intervals, the RSDs ranged from 30 to 50 percent in one
depth interval and were  greater than 50 percent in the
other.   The sampler performed well in sampling depth
intervals that did  not  exceed  12  inches  bss: RSDs of
0 percent were observed for three such depth intervals, and
RSDs of 10 and 26 were observed for the two remaining
depth intervals.  In these  depth intervals, the Split  Core
Sampler collected more  than 80  percent of its  design
volume  in 47 of 60 attempts (78 percent).  The sampler
had  mixed results in sampling  depth  intervals  that
exceeded 12 inches bss: RSDs of 37 and 51 percent  were
                                                     36

-------
Table 6-3. Percent Sample Recovery Summary Statistics for Split Core Sampler
Demonstration Area
S1A1 (river mouth)
S1A2 (freshwater bay)
S2A1 (lake)
S2A2 (wetland)
Target Depth Interval (inches bss)
Oto4
6 to 12
Oto4
12 to 32
Oto4
10 to 30
4 to 12
Actual Number of Attempts
5
9
5
6
36
16
5
PSR Range3
100
83 to 100
100
20 to 60
25 to 125
Oto75
100
Mean PSR
100
93
100
37
89
38
100
RSD (%)
0
10
0
37
26
51
0
Notes:

bss   =
PSR   =
RSD   =
Below sediment surface
Percent sample recovery
Relative standard deviation
    PSRs exceeding 100 resulted from either the volumetric measurement error associated with the presence of void spaces when the sediment was
    transferred to a graduated container or sediment compaction in the core tube.
observed in the  12- to 32- and 10- to 30-inch bss depth
intervals in S1A2 and S2A1, respectively. For these depth
intervals, the S1A2 RSD exceeded the 30 percent RSD
guideline by only 7 percentage points, but the S2A1 RSD
exceeded the guideline by 21 percentage points.  The Split
Core Sampler again performed as well as or better than the
reference samplers (see Section 7.1.1.2).

6.1.2  Ability to Consistently Collect Sediment in
       a Specified Depth Interval

Primary objective P2  involved evaluating the Split Core
Sampler's  ability to consistently collect sediment  in a
specified depth interval. This objective was addressed by
comparing actual and target core lengths for each depth
interval. The target core length for a sample was equal to
the distance between the upper and lower boundaries of a
depth interval.  Because the  core length measurements
presented in this section do not account for void space in
the core or rounding error, an  attempt may have achieved
an actual core length that equaled the target core length
but may not have resulted in a PSR of 100.

Because of difficulties in assessing  the location of the
sediment surface, the  sampling technician chose to push
the Split  Core  Sampler  beyond  the specified depth
intervals. Consequently, accuracy in targeting a specified
depth interval may have been compromised.  To assess
overall accuracy in targeting specified depth  intervals,
                                              core  lengths  were  compared to  depths of sampler
                                              deployment;  if a  core length  equals  the   depth  of
                                              deployment,  one may  conclude that the core length
                                              accurately reflects the specified depth interval. However,
                                              in most cases for the Split Core Sampler and for the
                                              reference samplers, the core lengths were shorter than the
                                              depths of deployment, indicating the occurrence of core
                                              shortening  or loss of sample  during sampler retrieval.
                                              Because  core  shortening plays  a  significant  role  in
                                              sediment sampling using end-filling samplers and because
                                              both the Split Core Sampler and reference samplers are
                                              end-filling samplers, core shortening is briefly described
                                              below.

                                              Core shortening, which primarily involves deficient entry
                                              of sediment into the core tube during sampler penetration,
                                              occurs because friction between sediment  and the inside
                                              wall of the sampler gradually increases as the core tube
                                              penetrates the sediment, resulting in gradual thinning of
                                              the core by lateral extrusion in front of the  core tube. As
                                              the friction changes with the  depth of penetration, the
                                              extent of core  shortening also changes.   Thus, not  all
                                              sediment layers may be uniformly represented within a
                                              given sample, and the actual core length will be less than
                                              the depth of sampler deployment (Blomqvist 1991).  Core
                                              shortening is more likely to affect sampling attempts in
                                              deeper intervals than in shallower intervals. The degree of
                                              core shortening was probably somewhat reduced for the
                                              Split Core Sampler because an electric hammer was used
                                              to induce vibrations in the sampler in order to reduce the
                                              friction generated upon sediment entry into the core tube.
                                                     37

-------
Table 6-4 presents the number of attempts in which the
actual core length equaled the target core length, target
core lengths, and mean actual core lengths. Initially, the
Wilcoxon signed rank test was to  be used to determine
whether differences between the actual and target core
lengths were statistically significant. However, review of
the Wilcoxon signed rank test revealed that the test results
for many of the data sets  were inconsistent with the
conclusions reached in  comparing the  actual and target
core lengths  for the  reason  stated  in Section  6.1.
Therefore,  primary  objective  P2 was addressed  by
evaluating (1) the number of attempts in which the actual
core length equaled the target core length and (2) the
difference between the  target core length and  the mean
actual core length.

In S1A1, the actual core lengths equaled the target core
lengths for all attempts in the 0- to 4- and 6- to 12-inch bss
depth intervals.  The average core length retrieved in this
area was  about 21 percent shorter than the  depth of
sampler deployment, which is not significant.

In S1A2, actual core lengths equaled the target core length
in all attempts in the 0- to 4-inch bss depth interval but
failed to do so for any of the attempts in the 12- to 32-inch
bss depth interval.  Samples collected in the latter interval
ranged in core length from 4.5 to 13.5 inches with a mean
core length of 9 inches. The failures to obtain  the target
core length in this interval may be attributable to (1) error
in assessing the  location of the sediment surface, which
might have resulted in the actual depth of penetration
being less than the measured depth of penetration; (2) core
shortening; (3) the sediment consisting of high levels of
silt and clay, resulting in formation of a plug in the coring
tip that inhibited further sediment retrieval; (4) sediment
compaction in the core tube; or (5) sediment loss during
sampler retrieval. The average core length retrieved in this
area was about 35 percent shorter than the  depth of
sampler deployment.

The  results observed  in  S2A1 were similar to those
observed in S1A2.  In the 0- to 4-inch bss depth interval in
S2A1,  samples  collected by  the Split Core  Sampler
equaled the target core length in  34  of 36  attempts;
consequently, the mean actual core length calculated for
this interval (3.9 inches rounded to 4  inches) compared
favorably to the target core length of 4 inches. However,
none of the samples collected during the 16 attempts in the
10- to 30-inch bss depth interval equaled the target core
length. The actual core lengths  retrieved from this depth
interval ranged from 0 to 16.5 inches, resulting in a mean
core length of 10 inches that compared unfavorably to the
target core length of 20 inches. The sampler failures in the
deeper interval in S2A1 may be attributable to the reasons
cited above for S1A2 except that in S2A1, the sediment
did not consist of as much clay as did the sediment in
S1A2  and  thus  provided less  opportunity  for plug
formation.  For both core tube lengths used, the average
core length retrieved in this area was about 69 percent
shorter than the depth of sampler deployment.
Table 6-4. Comparison of Target and Actual Core Length Data for Split Core Sampler
Number of Attempts in Which Actual
Target Depth Interval Core Length Equaled Target Core Target Core Length
Demonstration Area (inches bss) Length/Total Attempts (inches)
S1A1 (river mouth)
S1A2 (freshwater bay)
S2A1 (lake)
S2A2 (wetland)
Oto4
6 to 12
Oto4
12 to 32
Oto4
10 to 30
4 to 12
5/5
9/9
5/5
0/6
34/36
0/16
5/5
4
6
4
20
4
20
8
Mean Actual Core Length
(inches)
4
6
4
9
10
8
Notes:

bss  =  Below sediment surface

a    The calculated mean actual core length (3.9 inches) was rounded to the nearest integer.
                                                     38

-------
In S2A2, the actual core lengths equaled the target core
length for all attempts in the  4- to 12-inch bss depth
interval.  The average core length retrieved in this area
was about 68 percent shorter than the depth of sampler
deployment.

In summary, the demonstration results indicate that the
Split  Core Sampler was  able to  consistently  collect
sediment in sampling depth intervals that did not exceed
12 inches but did not perform well in the depth intervals
below 12 inches bss.  For sampling depth intervals not
exceeding 12 inches bss, 58 of the 60 actual core lengths
matched  the target core lengths. For the depth intervals
below 12 inches bss, none of the 22 actual core lengths
matched the target core length. The Split Core Sampler
performed as well as or better than the reference samplers
(see Section 7.1.2).

6.1.3  Ability to Collect Multiple Samples with
       Consistent Physical or Chemical
       Characteristics, or Both, from a
       Homogenous Layer of Sediment

Primary objective P3 involved evaluating the Split Core
Sampler's  ability to collect  multiple  samples  with
consistent physical or chemical characteristics, or both,
from a homogenous layer of sediment.  This objective
was addressed  by calculating  the RSD values  for the
sample analytical results for the 12- to 32-inch bss depth
interval in S1A2 and the 0- to 4- and 10- to 30-inch bss
depth intervals in S2A1.  Based on the predemonstration
investigation results, these three  depth intervals were
determined to be homogenous in terms of their physical
characteristics,  and the two S2A1 depth intervals were
determined to be homogenous in terms of their chemical
characteristics.

Figure 6-3 presents the demonstration analytical results
for PSD in the 12- to 32-inch bss depth interval in S1A2.
Figure 6-4 presents the demonstration analytical results
for arsenic and PSD in the 0- to 4- and 10- to 30-inch bss
depth intervals  in S2A1.  The demonstration analytical
results for arsenic appeared to contain statistical outliers
that indicated that the two S2A1 depth intervals might not
be chemically homogenous. For this evaluation, outliers
are defined as sample analytical results that are not within
two standard deviations of the mean.  The only outlier for
samples  collected by the Split Core Sampler was the
120 milligrams per kilogram (mg/kg) of arsenic in the 0- to
4-inch bss  depth  interval.  However,  outliers were also
found in the arsenic analytical results for samples collected
by the reference  sampler (see Section 7.1.3), providing
further evidence that the two S2A1 depth intervals may not
be chemically homogenous. A similar analysis performed
for the PSD data revealed no statistical outliers. Therefore,
the Split Core Sampler was evaluated based only on its
ability to collect multiple samples with consistent physical
characteristics.   RSDs were  calculated for each depth
interval based on the PSD analytical results for all locations
sampled.

RSDs calculated for the PSD  data were compared to the
laboratory  acceptance  criterion of  15 percent for field
triplicates (which was based on  historical information)
because  RSDs less than or equal to  15 percent for all
samples collected in a given depth interval and area may be
more attributable to the laboratory's precision than to the
sampler' s ability to collect multiple samples with consistent
physical characteristics. When the RSD for all samples in
a depth  interval  was  greater than 15  percent, it was
compared to the measured RSD for the field triplicates,
which  were prepared  by first  homogenizing  and then
subsampling  the  sediment collected  in a given depth
interval, location, and area. An RSD for all samples that is
less  than the RSD for  field triplicates may be  more
attributable to the laboratory's analytical procedure, the
sample  homogenization  procedure  implemented in the
field, or both rather than the  sampler's ability to collect
physically consistent samples. However, PSD parameters
with means less than 10 percent were not evaluated in this
manner because at low levels,  the analytical method is not
as precise; as a result, it will generate high RSD values and
may not actually reveal  whether multiple  samples with
consistent  physical   characteristics   were  collected.
Table 6-5 presents PSD summary statistics (range, mean,
and RSD) calculated for the samples and field triplicates
collected in  each depth interval  relevant to  primary
objective P3.

For the 12-to 32-inch bss depth interval in S1A2, the RSDs
for the silt and clay were  below the 15 percent laboratory
acceptance criterion.  The mean for the sand results for
samples collected from the depth interval  was less than
10 percent and  was not evaluated using  the criterion.
However, the sand results exhibited a tight range  (2 to
5 percent).
                                                    39

-------
              12- to 32-inch bss depth interval
Location 1A
Sand: 2%
Silt: 74%
Clay: 24%




Location 5A
Sand: 5%
Silt: 72%
Clay: 23%











Location 3C
Sand: 3%
Silt: 66%
Clay: 31 %











Location 1 E
Sand: 5%
Silt: 70%
Clay: 25%




Location 5E
Sand: 2%
Silt: 71%
Clay: 27%
Note:

bss  =   Below sediment surface

Figure 6-3. Split Core Sampler sample particle size distribution results for S1A2 (freshwater bay).
For the 0- to 4-inch bss depth interval in S2A1, the RSDs
for the sand and silt results were below the 15 percent
laboratory acceptance criterion.  The RSDs for the clay
results were  also below the  15  percent laboratory
acceptance criterion, despite having a  mean (9 percent)
that was less than 10 percent.

For the 10-to 30-inch bss depth interval in S2A1, the RSD
for the silt results was below the 15 percent laboratory
acceptance criterion.  However, the  23 percent RSD for
the sand results for this  depth  interval was above the
laboratory acceptance criterion and significantly above the
measured RSD for field triplicates (6 percent). Therefore,
some of the variation  in the  sand  results  may  be
attributable to the Split Core Sampler's ability to collect
samples  with consistent physical characteristics.  The
variation, however, was not considered significant because
it was only 8 percentage points greater than the laboratory
acceptance criterion.  The mean for the clay results for
samples collected from this depth interval was less than
10 percent and was not  evaluated using the criterion.
However, the clay results exhibited a tight range (6 to
10 percent).
                                                      40

-------
                0- to 4-inch bss depth interval
Location 1A
Arsenic: 24 mg/kg
Sand: 34%
Silt: 57%
Clay: 8%
Location 2A
Arsenic: 48 mg/kg

Location 3A
Arsenic: 30 mg/kg
Sand: 42%
Silt: 49%
Clay: 9%
Location 1 B
Arsenic: 40 mg/kg



Location 3B
Arsenic: 1 6 mg/kg



Location 2C
Arsenic: 36 mg/kg
Sand: 39%
Silt: 52%
Clay: 8%





Location 2D
Arsenic: 30 mg/kg




Location 1 E
Arsenic: 120
mg/kg
Sand: 29%
Silt: 59%
Location 2E
Arsenic: 47 mg/kg

Location 3E
Arsenic: 47 mg/kg
Sand: 36%
Silt: 54%
Clay: 10%
                10-to 30-inch bss depth interval
Location 1A
Arsenic: 5.0 mg/kg
Sand: 26%
Silt: 64%
Clay: 10%
Location 2A
Arsenic: 4.6 mg/kg
Location 3A
Arsenic: 5.3 mg/kg
Sand: 25%
Silt: 67%
Clay: 8%
Location 1 B
Arsenic: 5.3 mg/kg



Location 3B
Arsenic: 5.0 mg/kg




Location 2C
Arsenic: 5.4 mg/kg
Sand: 27%
Silt: 67%
Clay: 6%






Location 2D
Sample not
analyzed



Location 1 E
Arsenic: 4.7 mg/kg
Sand: 31%
Silt: 60%
Clay: 9%
Location 2E
Arsenic: 4.7 mg/kg
Location 3E
Arsenic: 5.2 mg/kg
Sand: 42%
Silt: 50%
Clay: 8%
Notes:

bss     =    Below sediment surface
mg/kg   =    Milligram per kilogram

The particle size distribution results for a given sample may not total 100 percent because of rounding or because some sediment did not pass through
a U.S. Standard No. 4 sieve and was classified as gravel rather than sand, silt, or clay.
Figure 6-4. Split Core Sampler sample arsenic and particle size distribution results for S2A1 (lake).
                                                                41

-------
Table 6-5.  Particle Size Distribution Summary Statistics for Split Core Sampler
Demonstration Area
S1A2 (freshwater bay)


S2A1 (lake)





Target Depth Interval
(inches bss)
12 to 32


Oto4


10 to 30


Parameter
Sand
Silt
Clay
Sand
Silt
Clay
Sand
Silt
Clay
Number of
Samples
5
5
5
5
5
5
5
5
5
Range (%)
2 to 5
66 to 74
23 to 31
29 to 42
49 to 59
8 to 10
25 to 42
50 to 67
6 to 10
Mean (%)
3
71
26
36
54
9
30
62
8
RSD (%)
(All Samples)
46
4
13
14
7
10
23
11
18
RSD (%)
(Field Triplicates)
54
6
9
4
2
6
6
3
13
Notes:

bss   =  Below sediment surface
RSD   =  Relative standard deviation
In summary, the Split Core Sampler met primary obj ective
P3 criteria except for an 8 percentage point exceedance in
the RSD for sand results for the 10- to 30-inch bss depth
interval in S2A1.  Therefore, it was concluded that the
Split Core Sampler was able to collect multiple samples
with consistent physical characteristics.

6.1.4  Ability to Collect a Representative Sample
       from a Clean Sediment Layer Below a
        Contaminated Sediment Layer

To evaluate whether the Split Core Sampler could collect
representative  samples from a clean sediment layer that
was  below  a contaminated sediment  layer (primary
objective P4), samples were collected from both clean and
contaminated layers using the Split Core Sampler and the
Hand  Corer  (a reference  sampler).    Because  the
predemonstration investigation results indicated that the
10- to 30-inch bss depth interval  in S2A1  contained
arsenic concentrations that were an  order of magnitude
less than those in the  0- to 4-inch bss depth interval  in
S2A1, the 10- to 30- and 0- to 4-inch bss depth intervals
were  considered to be clean and contaminated layers,
respectively.  Difficulties were encountered in assessing
the location of the sediment surface in this demonstration
area because a black, gelatinous material was present near
the sediment surface.   In addition,  the location  of the
sediment surface varied significantly  at several of the grid
locations.   This variation may have been caused by
previous   sampling  attempts  made   during  the
demonstration.
Samples collected from both depth intervals were analyzed
for arsenic. The contaminated layer concentrations were
used only to document that a contaminated layer existed
above the clean layer. The clean layer concentrations were
used to compare the Split Core Sampler's performance
with that of the Hand Corer. To make this comparison, the
null hypothesis was that the mean difference between the
Split  Core Sampler and Hand  Corer sample  arsenic
concentrations for the clean layer equaled zero.   The
alternative  hypothesis was  that  the  mean  difference
between the Split Core Sampler and Hand Corer sample
arsenic concentrations for the clean layer was not equal to
zero. A two-tailed Wilcoxon signed rank test was used to
compare the Split Core Sampler and Hand Corer sample
concentrations.

Figure 6-5 presents the  arsenic  concentrations in the
samples collected by the Split Core Sampler and the Hand
Corer in both depth intervals in S2A1.  Figure 6-5 also
presents the difference between the arsenic concentrations
in the samples collected by the two samplers in the 10- to
30-inch bss depth interval at each sampling location by
subtracting the arsenic concentration in the Hand Corer
sample from that in the Split Core Sampler sample.  Each
negative difference indicates that the sample collected by
the Split  Core  Sampler was  less  impacted  by the
contaminated layer than the sample collected by the Hand
Corer; each positive difference  indicates that the reverse
was true.

The sample analytical results showed that the 0-to 4-inch
bss depth  interval  contained  arsenic concentrations
                                                    42

-------
               0- to 4-inch bss depth interval
Location 1A
SCS: 24 mg/kg
HOC: 250 mg/kg
Location 2A
SCS: 48 mg/kg
HOC: 190 mg/kg
Location 3A
SCS: 30 mg/kg
HOC: 140 mg/kg
Location 1 B
SCS: 40 mg/kg
HOC: 130 mg/kg


Location 3B
SCS: 16 mg/kg
HOC: 140 mg/kg


Location 2C
SCS: 36 mg/kg
HOC: 120 mg/kg




Location 2D
SCS: 30 mg/kg
HOC: 130 mg/kg


Location 1 E
SCS: 120 mg/kg
HOC: 190 mg/kg
Location 2E
SCS: 47 mg/kg
HOC: 150 mg/kg
Location 3E
SCS: 47 mg/kg
HOC: 130 mg/kg
               10- to 30-inch bss depth interval
Notes:

bss
Diff
HOC
mg/kg
SCS
Location 1A
SCS: 5.0 mg/kg
HOC: 24 mg/kg
Diff: -19 mg/kg
Location 2A
SCS: 4.6 mg/kg
HOC: 8.3 mg/kg
Diff: -3.7 mg/kg
Location 3A
SCS: 5.3 mg/kg
HOC: 7.2 mg/kg
Diff: -1.9 mg/kg
Location 1 B
SCS: 5.3 mg/kg
HOC: 8.5 mg/kg
Diff: -3.2 mg/kg


Location 3B
SCS: 5.0 mg/kg
HOC: 8.2 mg/kg
Diff: -3.2 mg/kg



Location 2C
SCS: 5.4 mg/kg
HOC: 9.7 mg/kg
Diff: -4.3 mg/kg






Location 2D
SCS: Sample not
analyzed
HOC: 13 mg/kg



Location 1 E
SCS: 4.7 mg/kg
HOC: 16 mg/kg
Diff: -11. 3 mg/kg
Location 2E
SCS: 4.7 mg/kg
HOC: 7.2 mg/kg
Diff: -2.5 mg/kg
Location 3E
SCS: 5.2 mg/kg
HOC: 52 mg/kg
Diff: -46.8 mg/kg
Below sediment surface
Difference between arsenic concentrations in Split Core Sampler and Hand Corer samples
Hand Corer
Milligram per kilogram
Split Core Sampler
Figure 6-5.  Comparison of Split Core Sampler and reference sampler arsenic concentration results for S2A1 (lake).
                                                             43

-------
significantly greater than those in the 10- to 30-inch bss
depth interval.  For the  Split Core  Sampler, the arsenic
concentrations  ranged  from  16  to  48  mg/kg (not
considering  the anomalous result  of 120 mg/kg for
Location IE) and from 4.6 to 5.4 mg/kg in the 0- to 4- and
10- to 30-inch bss depth intervals, respectively.  For the
Hand Corer, the arsenic concentrations ranged from 120 to
250 mg/kg and from 7.2 to 24 mg/kg (not considering the
anomalous result of 52 mg/kg for Location 3E) in the 0- to
4- and 10- to 30-inch bss  depth intervals, respectively.
Explanation of these anomalies was beyond the scope of
the demonstration.

Comparison of the arsenic concentration ranges  showed
that the Split Core Sampler sample concentration range
was  less than that for the  Hand Corer for each depth
interval.  Based on the limited data available,  arsenic
concentrations  in  S2A1  appeared to  decrease  with
increasing sediment depth. This observation suggests that
significant compaction occurred in the  sediment samples
collected by the Split Core Sampler.  Sediment sample
compaction in the core tube of the Split Core Sampler is
likely because (1) the depth of sampler deployment was
greater than the core tube length; (2) the top of the  sampler
is closed by the top cap, and escape of sediment is limited
by the ball check valve; and (3) vibrations induced by the
electric hammer promote sample compaction.  Sediment
compaction in the Hand Corer core tube is not expected to
be as significant because the latter two factors do not
apply to the Hand Corer.

For the 10- to 30-inch bss depth  interval, the  arsenic
concentrations in samples collected using the Split Core
Sampler were less than the concentrations in samples
collected  using the  Hand  Corer  for  each  paired
observation. A statistical comparison  of the Split Core
Sampler and Hand Corer sample arsenic concentrations for
the clean layer using the  Wilcoxon  signed  rank test
showed that the arsenic concentrations were different at a
significance level  of 0.05  and that there was  only a
0.9 percent probability that the concentrations were not
different. This conclusion seems reasonable based on the
average difference between the Split Core Sampler and
Hand Corer sample concentrations, which was about
-11 mg/kg. This average difference was skewed by the
anomalous paired observation for Location 3E (5.2 and
52 mg/kg of arsenic in the Split Core Sampler and Hand
Corer samples, respectively). If the paired observation for
Location 3E is not considered, the average difference in
concentrations  is  about  -6.1  mg/kg, which  is  still
significant because  the reporting  limit for arsenic was
1.0 mg/kg.

In summary, although the Split Core  Sampler sample
concentrations for the clean layer appeared to be less than
the Hand Corer sample concentrations, because of sample
compaction  in the  Split Core  Sampler core tube, no
conclusion could be  drawn  regarding the Split Core
Sampler's ability to collect representative samples from a
clean layer that is below a contaminated layer.

6.1.5  Ability to be Adequately Decontaminated

Primary objective P5 involved evaluating the Split Core
Sampler's ability to be adequately decontaminated (see
Section 4.3).  This objective was addressed by  collecting
equipment rinsate samples after sampler decontamination
activities  in S1A1 and S2A1.  Specifically,  the 6-  to
12-inch bss depth interval in S1A1 and the 0- to 4-inch bss
depth interval in S2A1 were chosen to address P5 because
they contained high concentrations of PCBs and arsenic,
respectively.  Although it was intended that the evaluation
be limited to these depth intervals, this was not possible
because  AMS  simultaneously collected  samples  in
multiple depth intervals. However, this deviation did not
impact the  primary objective.   If the sampler were
adequately decontaminated, the  analytical results for the
equipment rinsate samples  would be below the analytical
laboratory's  reporting limits.  To  ensure that  the water
used to decontaminate the sampler was not contaminated,
decontamination  water blanks were  also  analyzed.
Contaminant concentrations in both the equipment rinsate
samples and decontamination water blanks were below the
laboratory reporting limits for PCBs (1 part per billion)
and arsenic (10 parts per billion).   Thus, the Split Core
Sampler  demonstrated  the  ability to be  adequately
decontaminated.

6.1.6  Time Requirements for Sample Collection
       Activities

Primary objective P6 involved evaluating the Split Core
Sampler's time requirements  for sample   collection
activities. These requirements were evaluated  in all four
demonstration areas but were not specifically evaluated by
depth interval  because samples  were simultaneously
collected in  multiple depth intervals to reduce the total
sample collection time. One technician was required for
                                                    44

-------
sampler  setup, sample collection,  sample  extrusion,
sampler disassembly, and sampler decontamination in each
of the four demonstration  areas. The amounts of time
required  to complete  these  activities  are  shown in
Table 6-6.  The time  measured for sample collection
activities did not include the time taken for mobilization,
demobilization, and maneuvering the sampling platforms
to access sampling locations because these activities were
not specific to the sampler; they were either site- or
weather-related.

The sampler setup time ranged from  2 to 18 minutes. In
S1A1 and S1A2, only 2 minutes was required for sampler
setup because only a few extension rods (three in S1A1
and  two  in  S1A2) were required.   Because  seven
extension rods and an AMS tripod-mounted winch were
required in S2A1, a greater amount of time (11 minutes)
was required.  Although only three extension rods were
used in S2A2, 18 minutes was required for sampler setup.
The main time requirement in this area was for the setup
of the AMS tripod-mounted winch on a sampling platform
that did not have enough space for easy winch setup.

The amount of time needed for sample collection ranged
from 35  to 444 minutes and was  mostly a  function of
(1) how  many attempts were required in  each depth
interval and (2) demonstration area characteristics such as
water depth and target sampling depth intervals. In S1A1,
where the water depth was about 5 to 6 feet and the target
sampling depth interval was 0 to 12 inches bss, the sample
collection time per attempt ranged from 2 to 7 minutes. In
S1A2, where the water depth interval was about 2 feet and
the target sampling depth interval was 0 to 32 inches bss,
the sample collection time  per attempt ranged from  4 to
9 minutes.  Two different core lengths were collected in
                          S2A1, where the  water depth  was  about  18  feet.
                          Collecting samples from 0 to 4 inches bss required 5 to
                          10 minutes per attempt, and collecting samples from 0 to
                          30 inches bss required 10 to 15 minutes per attempt.  In
                          S2A2, where the water depth ranged from 0.5 to 1.5 feet
                          and the target sampling depth interval was 4 to 12 inches
                          bss, the sample collection time per attempt ranged from
                          9 to 18 minutes.  The additional time needed for each
                          attempt in S2A2 was associated with (1) the  sampler's
                          depth of deployment and (2) the type of sampling platform
                          used. Because of the heterogeneity of the sample matrix,
                          the sampler was driven deeper than 4 feet bss to efficiently
                          recover the sediment. In addition, the sampling platform
                          in S2A2 did not have adequate work space to perform
                          sample collection activities quickly.

                          The amount of time needed for sample extrusion ranged
                          from 5 to 74 minutes and was strictly a function of how
                          many  sample  cores  were  collected  in  each  area.
                          Approximately 1 to 2 minutes was needed for extrusion of
                          each sample.

                          The amount of time needed for sampler disassembly was
                          not recorded in any of the areas; sampler setup time was
                          used  as  a  substitute for sampler disassembly  time.
                          Therefore, the  amount  of time required  for sampler
                          disassembly was assumed to range from 2 to  18 minutes.

                          The  amount of time  needed  for  Split Core Sampler
                          decontamination was evaluated only in S1A1 and S2A1.
                          In  S1A1,   17  minutes  was  needed  for  sampler
                          decontamination, while 99 minutes was needed in S2A1.
                          The difference between the amounts of time needed for
                          sampler  decontamination in  these  two areas can  be
                          accounted for by the following factors:
Table 6-6. Time Required to Complete Sampling Activities for Split Core Sampler

                                                        Time Required (minutes)
Activity
S1A1 (River Mouth)    S1A2 (Freshwater Bay)
                    S2A1 (Lake)
                S2A2 (Wetland)
Sampler setup
Sample collection
Sample extrusion
Sampler disassembly
Sampler decontamination
Total
        2
       36
       11
        2
       17
       68
     2
     35
     6
     2
Not evaluated
     45
 11
444
 74
 11
 99
639
     18
     64
      5
     18
Not evaluated
    105
                                                    45

-------
•   Three extension  rods were required in S1A1,  but
    seven extension rods were required in S2A1.

•   One 6-inch-long core tube and one 12-inch-long core
    tube  were used in  S1A1, whereas two  additional
    12-inch-long core tubes were used in S2A1.

•   S2A1 was  the  first area  sampled  during  the
    demonstration, and thus the sampling technician was
    implementing the decontamination procedure for the
    first time in this area.

Based on the demonstration results, a technician familiar
with the Split Core Sampler would need 2 to 18 minutes
for sampler setup, depending on  (1)  the number  of
extension rods required, (2) the number of core tubes used,
and (3) whether  or not the AMS tripod-mounted winch
was required. Sample collection time increases with water
depth; depth of sampler deployment; and to some extent,
the type of sampling platform used.  Approximately 5 to
10 minutes per sampling attempt could be expected for
sample collection along with an additional 1 to 2 minutes
for sample extrusion of each sample.  It is estimated that
sampler disassembly would take about the same amount of
time as sampler setup.  Sampler decontamination times in
S1A1 (17 minutes)  and S2A1 (99  minutes)  differed
greatly;   the  amount of time  needed  for  sampler
decontamination  is a function of the number of core tubes
and extension rods required. When sediment sampling is
planned,  the time required for setting up the sampling
platform and for maneuvering the platform to position the
sampler  at the  sampling  location  would  have to be
considered in addition to the times presented above.

6.2    Secondary Objectives

This section discusses the performance results for the Split
Core Sampler based on the secondary objectives stated in
Section 4.1.  Secondary objectives S1 through S5 required
evaluation of the  Split Core Sampler's

SI.  Skill and training requirements for proper sampler
     operation

S2.  Ability to  collect  samples under a variety of site
     conditions

S3.  Ability to collect an undisturbed sample
S4.  Durability based on materials of construction and
     engineering design

S5.  Availability, including spare part availability

Secondary  objectives  were  addressed  based  on
(1) observations of the Split Core Sampler's performance
during the demonstration and (2) information provided by
AMS.

6.2.1  Skill and Training Requirements for
       Proper Sampler Operation

The  Split Core  Sampler is easy to operate, requiring
minimal skills and training. Sampler assembly and sample
collection procedures can be learned in the field with a
few  practice attempts.   In  addition, a  written  SOP
accompanies the sampler when it is procured.   Sampler
operation is simple because the sampler does not require
complete disassembly and reassembly after each sampling
attempt. Only the coring tip requires removal to extrude
the core tube liner containing the sediment core.  The
sampler can be  operated by one person in both shallow
(wading) and deep water depths because of its lightness;
the fully equipped sampler, including one pair of 2-inch-
diameter,  12-inch-long split core tubes; the top cap with
the ball check valve; the  coring  tip; the coupling; one
12-inch-long, disposable, plastic core tube liner; one
4-foot-long,  carbon-steel AMS extension  rod; and one
rubber-coated AMS cross handle, weighs about 10 Ib. For
water depths requiring use of additional extension rods,
each 3-foot- and 4-foot-long extension rod weighs about
1.5 and 2 Ib, respectively. Support equipment used during
the demonstration included an AMS  slide-hammer, an
electric hammer (Bosch Model 11223 EVS) and power
source, an SDS Max self-locking adaptor for attaching the
electric hammer to the extension rod, an  AMS tripod-
mounted winch, and  an extrusion rod for extruding
sediment from the disposable, plastic core tube liner. For
applications  requiring use of an electric  hammer, the
electric hammer (37 Ib) and power source (if portable,
such as a generator) will likely be the heaviest pieces of
equipment required to operate the sampler.

During the demonstration, minimal strength and stamina
were required to collect  samples with the Split Core
Sampler  in  shallow  and  moderate  depth  intervals
containing  both  unconsolidated  and  consolidated
sediments. In all sampling areas, except S1A2, minimal
                                                   46

-------
strength and stamina were required to drive the sampler to
the desired depth intervals with the electric hammer.  In
S1A2, the sampler was manually driven by hand or with
an AMS slide-hammer to the target depth  interval; the
sampling platform used was too small to accommodate the
weight of  a  portable generator, and  thus  an electric
hammer could not be used. Minimal strength and stamina
were also required to retrieve the sampler at all  sampling
locations. An AMS tripod-mounted winch was used to
dislodge the  sampler from the sediment at four of ten
locations in S2A1 and five of five locations in S2A2. The
sampling technician  chose  to use the  tripod-mounted
winch when the strength and stamina required to  manually
dislodge the sampler were too great.

Previous sediment sampling  experience is beneficial in
selecting the  most appropriate support equipment for a
given Split Core Sampler application. For example, AMS
decided  to  use  an  electric hammer after several
unsuccessful  practice attempts to collect samples using
hand-  or slide-hammer-assisted  driving  at  the  first
sampling location in S2A1. The electric hammer was used
to induce vibrations to the sampler in all demonstration
areas where 110-volt power was available (S1A1, S2A1,
and  S2A2),  resulting in more  efficient  recovery  of
samples. As a result of the same practice attempts, AMS
also decided not to use the plastic basket retainers because
the retainers were too stiff to be effective for the sediment
to be  sampled during  the  demonstration.   Previous
sediment  sampling  experience  is also beneficial  in
accurately assessing the location of the sediment surface
using the sampler, as is the case with other samplers.

6.2.2  Ability to Collect Samples Under a  Variety
       of Site Conditions

The Split Core Sampler demonstrated its ability  to collect
sediment samples under all conditions encountered during
the demonstration, which included a variety of sampling
platforms, water depths,  sediment depths, and  sediment
compositions. During the demonstration, the  range of
sampling platforms used included wooden planks fastened
to ladders in S2A2; an 18-foot-long, 4-foot-wide Jon boat
in S1A2; a sturdier, 30-foot-long, 8-foot-wide pontoon
boat in S2A1; and the EPA GLNPO Mudpuppy in S1A1.
Because the sampler requires an external power source to
operate the electric hammer, the electric hammer was not
used in S1A2; specifically, the Jon boat was too small to
accommodate the weight of a portable generator.  Sampler
operation was feasible from any location on the sampling
platforms used in S1A1, S1A2, and S2A2. However, a
tripod-mounted winch was used to dislodge the sampler
from the sediment at four of ten sampling locations in
S2A1 and five of five sampling locations in S2A2. Use of
the tripod-mounted winch in  S2A1 dictated that  the
sampler be deployed through a 6-inch-diameter hole that
had to be cut in the center of the pontoon boat.  In S2A2,
the tripod-mounted winch had to be positioned over the
sampling locations by straddling the wooden planks.

Because of the lightness of the sampler and extension rods,
water depth had no significant impact on the  sampling
technician's ability to deploy and retrieve the sampler.
Water  depths  encountered  during  the  demonstration
ranged from about 0.5 foot in S2A2  to about 18 feet in
S2A1.  However, as with other samplers, the  sampling
technician's ability to assess the location of the sediment
surface  using the  Split Core Sampler decreased with
increasing water depth and turbidity.  Because of the
significant water depth and turbidity in S1A1, S1A2, and
S2A1, the sampling technician could not see the sediment
surface  from the sampling platforms.   An underwater
video camera may have enabled the sampling technician to
accurately assess the location of the sediment surface in
these areas (Blomqvist 1991).

The  Split Core Sampler  was able to collect  sediment
samples in all shallow and moderate depth intervals (up to
36 inches bss) in each demonstration area. However, in all
sampling attempts, the sampling technician chose to drive
the sampler beyond the specified sampling depth intervals
in order to retrieve sediment within the specified intervals.
Furthermore, as stated in  Section  6.1.1.1, the Split Core
Sampler required 20 attempts in the  10- to 30-inch bss
depth interval in S2A1, whereas 10 were expected to be
required. These limitations may be attributable to (1) error
in assessing the location of the sediment surface, which
might have resulted in  the actual depth  of penetration
being less  than the measured depth of penetration;
(2) deficient entry of sediment into the core tube (core
shortening); (3) sediment plug formation in the coring tip
that inhibited further sediment retrieval; or (4) sediment
loss during sampler retrieval.
                                                   47

-------
6.2.3  Ability to Collect an Undisturbed Sample

During  the  demonstration,  the  Split Core  Sampler
consistently  collected sediment samples in which the
sediment stratification was preserved; however, based on
visual observations, the samples appeared to have been
compacted.  Bow wave disturbance near the  sediment
surface did not occur in S2A2; the water depth  (0.5 to
1.5 feet) and low turbidity in this area allowed visual
confirmation of the location of the sediment surface. Bow
wave disturbance  near the sediment  surface  in  the
remaining demonstration areas was unlikely because the
speed of sampler deployment was controlled.  Also, as
mentioned above, sediment stratification was preserved in
samples collected in these areas.

The total core length retrieved in each attempt  using the
Split Core Sampler was less than  the depth of sampler
deployment. The difference between the total core length
retrieved and the depth of sampler deployment  ranged
from 2 to 7 inches in S1A1, 8.5 to 15.5 inches in  S1A2, 12
to 48 inches in S2A1, and 29 to 46.5 inches  in S2A2.
These differences indicate that sampling bias might have
occurred during sample collection in a given target depth
interval.

6.2.4  Durability Eased on Materials of
       Construction and Engineering Design

The primary components of the Split Core Sampler used
during the demonstration included (1) 6- and 12-inch-long
pairs of 300-series, stainless-steel  split core tubes with
interlocking, recessed channels and male, square-threaded
ends; (2) a 400-series, stainless-steel coring tip;  (3) a ball
check valve-vented top cap; (4) a female, square-threaded
coupling for attachment to additional stainless-steel split
core  tubes; and (5)  3- and 4-foot-long AMS extension
rods  made  of  stainless  steel  or  carbon  steel  (see
Figure 2-1).  The sampler was operated with the AMS
slide-hammer, the rubber-coated AMS cross handle, and
the electric hammer. In addition, the sampler was used
with disposable, plastic core tube liners and end caps to
facilitate removal and transport of an intact sample from
the split core tubes.  Based on observations made during
the demonstration, the  Split Core Sampler is  a sturdy
sampler;  none  of  the stainless-steel or carbon-steel
components  of the  sampler  was damaged or required
repair or replacement.  However,  on several occasions
sediment was caught in the ball check valve. As a result,
cleaning of the top cap using surface water or disassembly
of the top cap was required to clear the obstruction.

During the demonstration, the Split Core Sampler was
equipped  with  varying  lengths  of stainless-steel and
carbon-steel AMS extension rods. A total length of up to
11 feet of extension rods was used to collect samples in
S1A1, S1A2, and S2A2.  In these three areas, no bending
or bowing of the extension rods was observed.  In S2A1,
seven extension rods were coupled together to a combined
length of about 27 feet. Throughout most of the sampling
in S2A1, significant bowing of the coupled extension rods
was observed;  however, the rods were not damaged.

The  only  sampler component  damaged during  the
demonstration was the disposable, plastic core tube liner.
In S2A1 and S2A2, the sampling technician had to pound
one end of the core tube  liner against the side of the
stainless-steel bowl in order to extrude the sample in the
bowl. In a few cases, the degree of pounding required
resulted in the core tube liner cracking or breaking. To
rectify this problem, the sampling technician employed an
extrusion rod  in S1A1 and S1A2 to push the sample out
the bottom of the core tube liner.

6.2.5  Availability of Sampler and Spare Parts

No primary component of the Split Core Sampler required
replacement or servicing during the demonstration.  Had
a primary sampler component required replacement, it
would not have been available  in  local retail stores.
Sampler components may be  obtained from  AMS by
overnight  courier in 2  days or less, depending on the
location of the sampling site.

6.3    Data Quality

The overall QA objective for the demonstration was to
produce well-documented  data of known quality.  The
TSAs conducted to evaluate data quality did not reveal any
problems  that  would  make  the  demonstration  data
unusable.   The scope  of these TSAs  is described in
Sections 4.3 and 4.4 of this ITVR.

This  section   briefly  discusses  the data quality  of
demonstration results for the Split Core Sampler; more
detailed information is provided in the DER (Tetra Tech
1999b).  Specifically, the data quality associated with the
field measurement activities is discussed first, followed by
                                                   48

-------
the data quality associated with the  laboratory analysis
activities.

6.3.1  Field Measurement A ctivities

Field measurement  activities  conducted during  the
demonstration  included  measurement  of  the  time
associated  with  sample  collection activities,  water
velocity, water depth, core length, volume of IDW, volume
of sediment collected in a given sampling attempt, and
depth of sampler deployment.  Of these measurement
parameters, specific acceptance criteria were set for the
precision associated with the  time and  water velocity
measurements only (EPA  1999).  All time and water
velocity measurements made during the demonstration met
their respective criteria (see Table 6-7). Of the remaining
parameters,  some  difficulties were  encountered  in
measuring the volume of sediment collected in a given
sampling attempt and the depth of sampler deployment,
which are discussed below.

To measure the volume of sediment collected in a given
sampling attempt, the  sediment sample was transferred
into a 2-L container graduated in increments of 20 mL.
The container was tapped on a hard surface to  minimize
the presence of void spaces in the sample, the sample
surface was made even using a spoon, and the volume of
the sample was measured.   However, because the void
spaces could not be completely eliminated, the volumetric
measurements are believed  to  have a positive  bias that
resulted in overestimation  of PSRs.   Because the total
volume of the void spaces  could not be measured, its
impact on the  PSR results could  not  be  quantified.
However,  because the same volumetric measurement
procedure was used for both the innovative and  reference
samplers, the PSR results could still be compared.

The depth of sampler deployment was measured with
reference to the sediment surface. To identify the location
of the sediment surface, the sampling technician lowered
the sampler into the water and used the bottom end of the
sampler to feel the sediment surface.  Subsequently, the
technician drove the sampler into the  sediment to a depth
that he estimated to be appropriate to collect a  sediment
sample in the specified depth interval.  Overall during the
demonstration, this approach resulted in an average core
length that was about 21 to 69 percent shorter than the
estimated depth of sampler deployment, indicating that the
sampling technician may have had difficulty assessing the
location of the sediment surface.  Although both the
innovative   and  reference   samplers   used  in  the
demonstration are end-filling samplers that do not collect
uncompressed sediment samples, the degree of sediment
sample compaction in the core tube varied depending on
the sampler used.  In addition,  core shortening that could
occur in both the innovative and reference samplers would
impact the ability of the samplers to uniformly sample the
sediment in a given depth interval; the extent of the core
shortening, however, would depend on the sampler used.
Therefore, conclusions drawn from a comparison of the
sediment characteristics of the samples collected by the
reference samplers with those of the samples collected by
the Split Core Sampler should  be carefully interpreted.

6.3.2  Laboratory Analysis Activities

The laboratory analyses conducted for the demonstration
included  the following: (1)  PCB, arsenic,  and  PSD
analyses of sediment samples and (2) PCB and arsenic
analyses of equipment rinsate  samples.  To evaluate the
data quality of the laboratory  analysis results, field-
generated  QC samples, PE samples, and laboratory QC
check samples were analyzed. The field-generated QC
samples included the  field replicates and temperature
blanks described in Section 4.3  of this ITVR.  The PE
samples and laboratory QC check samples are described in
Section 4.4.  The acceptance criteria for the QC samples
are presented in Table 6-7.

All temperature blanks and field replicates subjected to
PCB and arsenic  analyses met the acceptance criteria,
indicating that the sample homogenization procedure (field
replicates)  and   sample   preservation  procedure
(temperature blanks) implemented in the field met the
demonstration requirements.   However, as stated in
Section 6.1.3, in one case the result of the field triplicate
sample analysis for PSD did not meet the acceptance
criterion.  Despite this failure to meet the acceptance
criterion, the PSD results are considered to be valid for the
reasons detailed in Section 6.1.3.

The  PE sample results for  both the  PCB and arsenic
analyses met the acceptance criteria, indicating that the
analytical laboratory accurately measured both PCBs and
arsenic.

The analytical results for all laboratory QC check samples
except the  following  met  the  acceptance  criteria:
                                                   49

-------
Table 6-7.  Summary of Quality Control Checks and Acceptance Criteria for Field and Laboratory Parameters
Parameter
Quality Control Check
Matrix
Acceptance Criterion
Field
Time required for sample
collection activities
Water velocity
Cooler temperature
Simultaneous measurements
Consecutive measurements
Temperature blank
Not applicable
Water
Water
RPD< 10
RPD < 20
4±2°C
Laboratory
PCBs
Arsenic
Method blank
Surrogate
MS/MSD
Extract duplicates
BS/BSD
Field triplicates
Field duplicates
PE samples
Interference check solution A
Interference check solution AB
Serial dilution
Method blank
MS/MSD
Postdigestion spike
Digestate duplicates
BS/BSD
Field triplicates
Field duplicates
Sediment and equipment rinsate
Sediment and equipment rinsate
Sediment
Sediment
Equipment rinsate
Sediment
Equipment rinsate
Sediment
Equipment rinsate
Soil
Water
Sediment and equipment rinsate
Sediment and equipment rinsate
Sediment and equipment rinsate
Sediment and equipment rinsate
Sediment
Sediment and equipment rinsate
Sediment and equipment rinsate
Sediment
Equipment rinsate
Sediment
Equipment rinsate
 50 times the
instrument detection limit
< Reporting limit
RPD< 10
Percent recovery: 67 to 109
Percent recovery: 75 to 125
RPD< 10
RPD< 10
Percent recovery: 80 to 120
RPDs 10
Percent recovery: 81 to 1 1 3
RSD < 30
RPD < 20
                                                           50

-------
Table 6-7. Summary of Quality Control Checks and Acceptance Criteria for Field and Laboratory Parameters (Continued)
Parameter
Quality Control Check
Matrix
Acceptance Criterion
Laboratory (Continued)
Arsenic (continued)
PSD
PE samples
Repeat analysis
Field triplicates
Soil
Water
Sediment
Sediment
Actual concentration = 239 mg/kg
Expected recovery3 = 199 mg/kg
Actual recovery11 = 183 mg/kg
Actual concentration = 6.02 mg/kg
Expected recovery3 = 5 mg/kg
Actual recovery11 = 4.81 mg/kg
25.0 to 39.4 parts per billion (certified value:
33.4 parts per billion)
± 1 hydrometer unit
RPD < 15 for sand, silt, and clay
Notes:
±
BS/BSD
mg/kg
MS/MSD
Greater than                      PCB   =
Less than or equal to                PE    =
Plus or minus                     PSD   =
Blank spike/blank spike duplicate       RPD   =
Milligram per kilogram               RSD   =
Matrix spike/matrix spike duplicate
Polychlorinated biphenyl
Performance evaluation
Particle size distribution
Relative percent difference
Relative standard deviation
    The expected recovery is based on typical recoveries of arsenic in soil during multiple interlaboratory studies.

    The actual recovery is the mean arsenic concentration in the PE sample based on four replicate analyses by the proficiency testing laboratory.
(1) MS/MSD samples for analysis for PCBs  in  the
sediment matrix and (2) equipment rinsate  samples for
PCB analysis. These issues and their likely impact on data
quality are discussed below.

For the sediment matrix, in all MS/MSD samples analyzed
for PCBs, Aroclor 1016 was recovered at levels higher
than the upper limit of the acceptance criterion, indicating
a positive bias in the PCB results for sediment samples.
However, the  analytical  laboratory had  no problem
meeting the acceptance criteria for control samples such as
BS/BSDs.   For  this  reason, the  failure  to meet  the
acceptance criterion for MS/MSD  sample  analysis was
attributed to matrix interference. Because Aroclor 1016
was recovered at levels higher than the upper limit of the
acceptance criterion in all MS/MSD samples associated
with both the innovative and reference samplers, the PCB
results could still be compared. The MS/MSD  spiking
compounds (Aroclors 1016 and 1260) were selected based
on the Aroclors  detected  during the predemonstration
investigation   and  as  recommended   in SW-846
Method 8082.
                                           Also for the sediment matrix, in one out of three MS/MSD
                                           pairs analyzed for PCBs, Aroclor 1260 was recovered at a
                                           level less than the lower limit of the acceptance criterion
                                           in the MS sample, but the recovery in the associated MSD
                                           sample was acceptable. Because the investigative samples
                                           contained  only Aroclor  1242,  of  the  two  spiking
                                           compounds used to prepare the MS/MSD  samples, only
                                           the Aroclor 1016 recoveries were considered to be relevant
                                           based on the PCB congener distribution; the Aroclor 1260
                                           recoveries were not considered to be relevant. Therefore,
                                           the low recovery associated with Aroclor 1260 had no
                                           impact on data quality.

                                           In all equipment rinsate samples analyzed for  PCBs,
                                           decachlorobiphenyl (the surrogate) was recovered at levels
                                           lower than the  lower limit of the acceptance criterion,
                                           indicating a negative bias in the PCB results for equipment
                                           rinsate samples. However, the analytical laboratory had
                                           no problem  meeting the  acceptance criteria for control
                                           samples such as PE samples and deionized water blanks.
                                           For this reason,  the failure to meet the surrogate recovery
                                           acceptance criterion for  the  equipment rinsate sample
                                                     51

-------
analysis was attributed to matrix interference. Because the   samples associated with both the innovative and reference
surrogate was recovered at levels lower than the lower   samplers, the PCB results could still be compared.
limit of the acceptance criterion in all equipment rinsate
                                                   52

-------
                                               Chapter 7
                             Performance of the Reference Samplers
To  verify a wide range of performance attributes, the
innovative  sediment sampler  demonstration  had both
primary and secondary objectives.  Primary objectives
were  critical to the technology  evaluation and were
intended to  produce  quantitative  results  regarding
technology performance.  Secondary objectives provided
information that  was useful  but  did not necessarily
produce  quantitative  results  regarding  technology
performance. The approach used to address each primary
and secondary objective for the Split Core Sampler and
reference samplers is discussed in Chapter 4. This chapter
describes the performance of the reference samplers based
on the primary obj ectives (excluding costs associated with
sample collection activities) and secondary objectives.
This  chapter  also  discusses the   data  quality  of
demonstration results for the reference samplers.

The performance of the Split Core Sampler is discussed in
Chapter  6, costs  associated  with  sample collection
activities  (primary objective  P7)  are  presented  in
Chapter 8, and the performance of the Split Core Sampler
and reference samplers is compared in summary form in
Chapter 9.

7.1    Primary Objectives

This section discusses the performance results  for the
reference samplers based on the primary objectives stated
in Section 4.1 except for primary objectives P4 and P7,
which  are  addressed in  Section 6.1.4 and Chapter 8,
respectively. Otherwise, the primary objectives discussed
in this  section  are the  same as those discussed in
Section 6.1.  During the demonstration, the sampling
technicians were provided an opportunity to practice
sample collection at each demonstration area until they felt
confident enough to initiate demonstration sampling.
To address primary objectives, samples were collected
using two different reference samplers, the Vibrocorer in
S1A1 and the Hand Corer in the other areas. The areas
and  depth  intervals sampled are the same as those
described in Section 6.1 except that the 4- to 6-foot bss
and 9- to 11-foot bss depth intervals in S1A1 and S2A2,
respectively,  were  not sampled  using the  reference
samplers. The Vibrocorer had difficulty fully penetrating
the 4- to 6-foot bss depth interval because of the presence
of clay hardpan and was thus unable to collect samples
from this interval in S1A1; the sampling technicians made
only a few attempts and decided not to complete sampling
in this depth interval.  In S2A2, the Hand Corer was not
used for the 9- to  11-foot bss depth interval because it is
not designed to collect samples at depths below 3 feet bss.
Consequently, the reference samplers were not evaluated
with respect to these two depth intervals. The numbers of
investigative  and QC  samples  collected in each area,
sediment sample volumes required, and sample analytical
parameters are discussed in Chapter 4.

The demonstration results for the reference samplers under
primary objectives PI and P2 were evaluated using the
Wilk-Shapiro test to determine whether the results were
normally distributed. Because most of the results were not
normally distributed, the Wilk-Shapiro test was used in an
attempt  to  evaluate whether the  results  followed a
lognormal distribution. The test revealed that the results
either were not lognormally distributed or could not be
tested for lognormality because they contained values that
were equal to zero. For these reasons, the Student's t-test,
a parametric  test, was not used to  perform hypothesis
testing; the Wilcoxon signed rank test, a nonparametric
test, was used as an alternative to the Student's t-test. As
described in Section 6.1,  Statistix® was used to perform
statistical evaluations   of the demonstration   results
(Analytical Software 1996). Appendix C provides details
on the statistical methods used for data evaluation.
                                                    53

-------
7.1.1  Ability to Consistently Collect a Specified
        Volume of Sediment

Primary objective PI involved evaluating the reference
samplers' ability to consistently collect a specified volume
of sediment. This objective was addressed by comparing
(1) the actual number of sampling attempts required to
collect a specified volume of sediment to the expected
number of attempts (rounded to the nearest higher integer)
at each sampling location in each target depth interval and
(2)  the actual  volume of sediment  collected in  the
specified target depth  interval in  each attempt to  the
calculated sampler volume (design volume) for the depth
interval.    The expected  number  of attempts  was
determined by dividing the specified sample volume by the
design volume for the depth interval. The results of these
comparisons are summarized below.

7.1.1.1 Number of Sampling Attempts Required

Tables 7-1 and 7-2 present the expected and actual number
of reference sampler sampling attempts for each depth
interval at  Sites  1 and 2, respectively.  Initially,  the
Wilcoxon signed rank test was used to determine whether
the difference between the expected and actual number of
attempts was  statistically significant.   However,  the
conclusions drawn from the Wilcoxon signed rank test
were  inconsistent with  the  conclusions  reached  in
comparing the expected and actual number of attempts
(see Appendix C for an example).
In S1A1, the Vibrocorer performed well in the 0- to 4- and
6- to  12-inch bss  depth  intervals, where the expected
number of attempts equaled the actual number of attempts.
As  stated above,  the  Vibrocorer had  difficulty fully
penetrating the 4- to 6-foot bss depth interval because of
the presence of clay hardpan and was  thus unable to
collect samples from this  interval in S1A1; the sampling
technicians  made  a few  attempts and  decided not to
complete sampling in this depth interval.

In S1A2, the Hand Corer performed well  in the 0- to
4-inch bss depth interval,  where the expected number of
attempts equaled the actual number of attempts, but did
not perform as  well in  the  12- to 32-inch bss depth
interval. In the 12-to 32-inch bss depth interval, the Hand
Corer required three additional attempts.  The additional
attempts  in  this depth interval may be  attributable to
(1) error in assessing the location of the sediment surface,
which might  have resulted in  the  actual  depth  of
penetration  being  less than the  measured  depth  of
penetration;  (2) deficient  entry of sediment into the core
tube (core shortening); (3) the sediment consisting of high
levels of silt  (63 to 72 percent) and clay (22 to 31 percent),
which might have caused plug formation in the coring tip
that inhibited further sediment retrieval;  or (4) sediment
loss during sampler retrieval.

In S2A1, the Hand Corer again performed better in the
shallower of the two depth intervals sampled. In the 0- to
4-inch bss  depth  interval,  the  Hand   Corer  required
39 attempts, whereas 33 attempts were expected.  In the
10- to 30-inch bss depth interval, the Hand Corer required
more than three times the expected number of attempts to
Table 7-1.  Comparison of Expected and Actual Number of Sampling Attempts for Reference Samplers at Site 1


Location
1A
1E
3C
5A
5E
Total
Number of Attempts
0- to 4-inch bss Depth
in S1A1 (River Mouth)
Interval
Expected Actual
1
1
1
1
1
5
1
1
1
1
1
5
6- to 12-inch
Expected
1
1
1
1
1
5
Using Vibrocorer
bss Depth Interval
Actual
1
1
1
1
1
5
Number of Attempts in S1A2 (Freshwater Bay)
0- to 4-inch bss
Expected
1
1
1
1
1
5
Depth Interval
Actual
1
1
1
1
1
5
12- to 32-inch
Expected
1
1
1
1
1
5
Using Hand Corer
bss Depth Interval
Actual
1
2
3
1
1
8
Note:

bss  =  Below sediment surface
                                                    54

-------
Table 7-2.  Comparison of Expected and Actual Number of Sampling Attempts for Reference Sampler at Site 2
                        Number of Attempts in
                     S2A1 (Lake) Using Hand Corer

Location
1A
1B
1E
2A
2C
2D
2E
3A
3B
3E
Total
0- to 4-inch bss
Expected
3
4
3
2
5
2
2
5
4
3
33
Depth Interval
Actual
3
4
4
2
7
2
2
7
5
3
39
10- to 30-inch bss
Expected
1
1
1
1
1
1
1
1
1
1
10
Depth Interval
Actual
4
5
3
2
3
1
2
6
3
2
31
                                                                 Location
                                                                 1A
                                                                 1E
                                                                 3C
                                                                 5A
                                                                 5E
                                                                 Total
                          Number of Attempts in
                      S2A2 (Wetland) Using Hand Corer
                       4-to 12-inch bss Depth Interval
                       Expected          Actual
Notes:

bss  =  Below sediment surface

a    Sampling was discontinued after the 12 attempts made at this location failed to collect the specified sediment volume.
                                          2
                                         12a
                                          3
                                          2
                                          1
                                         20
collect adequate sample volumes, and the actual number of
attempts equaled the expected number of attempts at only
one of the ten sampling locations. The sampler failures in
S2A1 may be attributable to the reasons cited above for
S1A2 except that in S2A1, the sediment does not consist
of as much clay as does the sediment in S1A2 and thus
exhibited less tendency for plug formation in the coring
tip. Also, during sampler retrieval in S2A1, the sampler's
flutter valve did not seat properly in a few attempts.  This
malfunction resulted in partial or complete loss of vacuum
in the core tube and subsequent sample loss.

In the 4- to 12-inch bss depth interval in S2A2, the Hand
Corer had  significant difficulty in collecting sediment;
20 attempts were recorded, whereas 5 were expected.  Of
the 20 attempts, more than half (12) were recorded at
Location  IE.   Eight attempts were recorded at the
remaining four locations, whereas four were expected.
Moreover,  more  than 20  attempts  would  have been
necessary to complete sampling in  this  depth  interval
because sampling was discontinued at Location  IE after
the 12 attempts made at this location failed to collect the
specified sediment volume.  The Hand Corer experienced
the greatest number of problems in S2A2, perhaps because
this  area contained  significant  amounts  of partially
decomposed reeds and leaves and live vegetation. As a
result, the sediment matrix was highly heterogenous  and
was  difficult to cut through, capture, and  retain.  The
sampler failures in  S2A2 may also be attributed to the
reasons cited above  for S1A2.

In summary, the demonstration results indicate that the
Vibrocorer demonstrated the ability to consistently collect
a specified volume of sediment in the 0- to 4- and 6- to
12-inch bss depth intervals because the number of actual
attempts  equaled the number  of expected  attempts.
However, the Vibrocorer did not collect samples in the 4-
to 6-foot bss depth  interval.  The Hand Corer collected
surficial  sediment  well  but had difficulty  collecting
samples at depths greater than 4 inches bss.  In the two 0-
to 4-inch bss depth intervals, the Hand Corer required only
16 percent  more  attempts than expected (44 actual
attempts versus 38 expected attempts). In contrast, in the
deeper  intervals,  the  Hand  Corer  required  nearly
200  percent more  attempts  than expected (59 actual
attempts versus 20 expected attempts), indicating a high
level of inconsistency in collecting specified volumes of
sediment.
                                                     55

-------
7.1.1.2  Volume of Sediment Collected

The volume  of sediment collected by the  reference
samplers  in each  sampling  attempt in a  given depth
interval was divided by the corresponding design volume,
and the resulting ratio was multiplied by 100 to estimate
the PSR. The RSD of the PSRs was calculated to evaluate
the ability of the reference samplers to consistently collect
a specified volume of sediment; if a  sampler were to
recover an identical volume of sediment in every attempt,
the RSD would equal zero.  Both PSR and RSD results
should be considered to properly evaluate the sampler's
performance because a low RSD, which indicates that the
sampler's performance was consistent, may be based on
                                                 consistently low PSRs. Figures 7-1 and 7-2 present PSRs
                                                 for the reference samplers at Sites 1 and 2, respectively.
                                                 Table 7-3 presents PSR summary statistics (range, mean,
                                                 and RSD) for both sites.

                                                 The Vibrocorer performed well in the 0- to 4- and 6- to
                                                 12-inch bss depth intervals in S1A1. Each attempt in the
                                                 0- to 4-inch bss depth interval had a PSR of 100. In the 6-
                                                 to 12-inch bss depth interval, a narrow PSR range of 79 to
                                                 83 resulted in an RSD of 3 percent, which is less than the
                                                 30 percent RSD guideline.   Although the Vibrocorer
                                                 collected a consistent volume of sediment in this depth
                                                 interval, it  did not collect more than 83  percent of its
                                                 design volume.
         S1A1, 0- to ilrM^m^aph interval
               0- to 4-inch bss depth interval
       Total number of attempts: 5
         Oto20     >20to40     >40to60     >60to80    >80to100

              Vibrocorer percent sample recovery
                                                         S1A2, O-tofaebf  interval
                                                                0- to 4-mcn bss depth interval
                                                                 Total number of attempts: 5
                                                         Oto20      >20to40     >40to60     >60to80     >80to100

                                                              Hand Corer percent sample recovery
        S1A1, 6- to 1^M20to40     >40to60     >60to80     >80to100

     Vibrocorer percent sample recovery
                                                        S1A2, 12-tc£3S3tf{^tB!S&:afejbfri interval
                                                               12-to 32-inch bss depth interval
                                                                                            Total number of attempts: 8
                                                                  Oto20     >20to40     >40to60     >60to80    >80to100

                                                                       Hand Corer percent sample recovery
Note:

bss  =   Below sediment surface
Figure 7-1.  Percent sample recoveries for Vibrocorer and Hand Corer at Site 1.
                                                       56

-------
        Oto20    >20to40   >40to60   >60to80   >80to100
                 Percent sample recovery
                                             >100
               - to 3-mcn  ss  ept  inerva
        Oto20    >20to40   >40to60   >60to80   >80to100   >100
                 Percent sample recovery
        S2A2,
                    S2A2 (wetland)
        Oto20    >20to40   >40to60   >60to80   >80to100    >100
                 Percent sample recovery

Notes:

bss  =  Below sediment surface

Percent sample recoveries exceeding 100 resulted from the volumetric
measurement error associated with the presence of void spaces when
the sediment was transferred to a graduated container.


Figure 7-2. Percent sample recoveries for Hand Corer at Site 2.
In S1A2,  the Hand Corer performed well in the 0- to
4-inch bss depth interval but performed poorly in the 12-
to 32-inch bss depth interval. In the 0- to 4-inch bss depth
interval, the Hand Corer achieved a PSR of 100 in every
attempt. However, in the 12-to 3 2-inch bss depth interval,
PSRs ranged from 15 to 55 and had a mean of only 31, as
shown in Table 7-3. As shown in Figure  7-1, five of the
eight attempts in this interval fell in the greater than 20 to
40 percent range, and two  of the eight attempts fell in the
0 to 20 percent range.  Because the recoveries fell in a
narrow range, the RSD  of 35 percent exceeded the RSD
guideline of 30 percent by only 5 percentage points.

In S2A1,  the Hand Corer performed well in the 0- to
4-inch bss depth interval but did not perform well in the
10- to 30-inch bss depth interval.  As shown in Table 7-3,
PSRs for the 0- to 4-inch bss depth interval ranged from
0 to 100 with a mean of 85. As shown in Figure 7-2, 27 of
the 39 attempts in this interval had PSRs of 80 to 100, and
34 of the 39 attempts had PSRs greater than 60.  Because
most of the PSRs fell  in a narrow range, the RSD of
33  percent compared favorably  to the 30 percent RSD
guideline.  In the 10- to 30-inch bss depth interval, the
PSRs ranged from 0 to 50 with a mean of 21. As shown in
Figure 7-2, most of the PSRs fell  in the 0 to 20 range. An
RSD of 62 percent was calculated for the 10- to 30-inch
bss depth interval, which indicates  a high  degree of
inconsistency.

In the 4- to 12-inch bss depth interval in S2A2, the Hand
Corer had difficulty collecting sediment.  As shown in
Table 7-3, PSRs  for S2A2 ranged from 0 to 125  with a
mean of 22.  This  wide  range of PSRs resulted in an
extremely high RSD of 161 percent. Figure 7-2 shows that
70 percent of the attempts fell in the 0 to 20 PSR range,
which indicates consistently low  recoveries.

In summary, the Vibrocorer performed well in the 0- to 4-
and 6-to 12-inch bss depth intervals, and the Hand Corer
performed well in the shallow depth intervals but not in
the deeper intervals. In the 0- to  4- and 6- to 12-inch bss
depth intervals in S1A1, the Vibrocorer had  RSDs that
were less  than the 30 percent RSD guideline.  The Hand
Corer performed well  in the 0- to 4-inch  bss depth
intervals, in S1A2 and S2A1 for which low RSDs (0 and
33 percent, respectively) were observed. In the 10- to 30-
and 4- to  12-inch bss depth intervals in S2A1 and S2A2,
the RSDs  of 62 and 161 percent, respectively, were well
above the  30 percent RSD guideline.
                                                     57

-------
Table 7-3. Percent Sample Recovery Summary Statistics for Reference Samplers
Demonstration Area
S1A1 (river mouth)
S1A2 (freshwater bay)
S2A1 (lake)
S2A2 (wetland)
Reference Sampler
Vibrocorer
Hand Corer
Hand Corer
Hand Corer
Target Depth Interval
(inches bss)
Oto4
6 to 12
Oto4
12 to 32
Oto4
10 to 30
4 to 12
Actual Number
of Attempts
5
5
5
8
39
31
b
20
PSR Range3
100
79 to 83
100
15 to 55
0 to 100
OtoSO
0 to 125
Mean PSR
100
82
100
31
85
21
22
RSD (%)
0
3
0
35
33
62
161
Notes:
bss   =   Below sediment surface
PSR   =   Percent sample recovery
RSD   =   Relative standard deviation

a    PSRs exceeding 100 resulted from the volumetric measurement error associated with the presence of void spaces when the sediment was
    transferred to a graduated container.


    after the 12 attempts made at this location failed to collect the specified sediment volume.
7.1.2  Ability to Consistently Collect Sediment in
       a Specified Depth Interval

Primary  objective P2 involved evaluating the reference
samplers' ability to consistently collect sediment in a
specified depth interval. This objective was addressed by
comparing actual and target core lengths for each depth
interval.  The target core length for a sample was equal to
the distance between the upper and lower boundaries of a
depth interval.   Because the core length measurements
presented in this section do not account for void space, an
attempt may  have  achieved  an  actual core length that
equaled the target core length but may not have resulted in
a PSR of 100.

Because  of difficulties in assessing the location of the
sediment surface, the sampling technicians chose to push
the  samplers beyond the  specified depth  intervals.
Consequently, accuracy in determining a specified depth
interval may have been compromised. To assess overall
accuracy in determining  specified depth intervals, core
lengths were compared to depths of sampler deployment;
if a core  length equals the depth of deployment, one may
conclude that the  core  length  accurately  reflects  the
specified depth interval.  However, in most cases for the
reference samplers, the core lengths were shorter than the
depths of deployment, indicating the occurrence of core
shortening  or loss of sample during sampler retrieval.
Because core  shortening plays a significant  role  in
sediment sampling using end-filling samplers and because
both reference samplers are end-filling samplers, core
shortening is briefly described below.

Core shortening, which primarily involves deficient entry
of sediment into the core tube during sampler penetration,
occurs because friction between sediment and the inside
wall of the sampler gradually increases as the core tube
penetrates the sediment, resulting in gradual thinning of
the  core by lateral extrusion in front of the core tube.  As
the  friction changes with the depth of penetration,  the
extent of core shortening also changes.  Thus, not all
sediment layers may be uniformly represented within a
given sample, and the actual core length will be less than
the  depth of sampler deployment (Blomqvist 1991). Core
shortening  is more likely to affect sampling attempts in
deeper  intervals than  in  shallower intervals.   Core
shortening  is expected to  be  less  prevalent  for  the
Vibrocorer, because  the  vibrations  produced  by this
sampler reduce the friction generated upon sediment entry
into the core tube.
                                                     58

-------
Table 7-4 presents the number of attempts in which the
actual core length equaled the target core length, target
core lengths, and mean actual core lengths. Initially, the
Wilcoxon signed rank test was to  be used to determine
whether differences between the actual and target core
lengths were statistically significant. However, review of
the Wilcoxon  signed  rank test results revealed that the
results for many of the data sets were inconsistent with the
conclusions reached in  comparing the target and actual
core lengths for the  reasons described in Section  6.1.
Therefore,  primary   objective P2 was  addressed  by
evaluating (1) the number of attempts in which the actual
core length equaled the target core  length and (2) the
difference between the  target  core length and the mean
actual core length.

In S1A1, samples collected by the Vibrocorer equaled the
target core length in five out of five attempts in both the 0-
to 4- and 6- to  12-inch bss depth intervals.  However,
these results are not surprising because the depth of
sampler deployment  was at  least 52 inches for these
attempts.  The Vibrocorer had difficulty fully penetrating
the 4- to 6-foot bss depth interval in S1A1 because of the
presence of clay hardpan and was thus unable to collect
samples in this interval; the sampling technicians made a
few attempts and then decided not to complete sampling in
this interval. The average core length retrieved in this area
was about 23 percent shorter than the depth of sampler
deployment.
In S1A2, samples collected by the Hand Corer equaled the
target core length in all attempts in the 0- to 4-inch bss
depth interval but failed to do so in any of the attempts in
the 12-to 32-inch bss depth interval. Samples collected in
the  latter  interval  ranged in  core length from  3 to
11 inches, with a mean core  length of 7 inches.   The
additional attempts in this interval may be attributable to
(1) error in assessing the location of the sediment surface,
which might have  resulted  in the actual  depth of
penetration being  less  than  the   measured  depth of
penetration;  (2)  core  shortening;  (3)  the  sediment
consisting of high  levels of silt and clay,  resulting in
formation of a plug in the coring tip that inhibited further
sediment retrieval; or (4) sediment loss during sampler
retrieval.  The average core length retrieved in this area
was about 52 percent shorter than the depth of sampler
deployment.

The results observed  in  S2A1 were similar to  those
observed in S1A2. In the 0- to 4-inch bss depth interval in
S2A1, samples  collected by the Hand Corer equaled the
target core length in 36 of 39 attempts; consequently, the
mean  actual core  length  calculated for this interval
(3.7 inches rounded to 4 inches) compared favorably to the
target core length of 4 inches.  However, none of the
samples  collected during the 31 attempts in the  10- to
30-inch bss depth interval equaled the target core length.
The actual core lengths in this depth interval ranged from
0 to 12 inches, resulting in a mean core length of 5 inches
Table 7-4. Comparison of Target and Actual Core Length Data for Reference Samplers
Demonstration Area
S1A1 (river mouth)
S1A2 (freshwater bay)
S2A1 (lake)
S2A2 (wetland)
Number of Attempts in Which Actual
Reference Target Depth Interval Core Length Equaled Target Core Target Core
Sampler (inches bss) Length/Total Attempts Length (inches)
Vibrocorer
Hand Corer
Hand Corer
Hand Corer
Oto4
6 to 12
Oto4
12 to 32
Oto4
10 to 30
4 to 12
5/5
5/5
5/5
0/8
36/39
0/31
3/20
4
6
4
20
4
20
8
Mean Actual Core
Length (inches)
4
6
4
7
5
2
Notes:

bss  =  Below sediment surface

a    The calculated mean actual core length (3.7 inches) was rounded to the nearest integer.
                                                     59

-------
that compared unfavorably to the target core length of
20 inches. The sampler failures in the deeper interval in
S2A1 may be attributable to the reasons cited for S1A2
except that in S2A1, the sediment does not consist of as
much clay as does the sediment in S1A2 and thus provides
less opportunity for plug  formation in the coring tip. In
S2A1, during sampler retrieval the sampler's flutter valve
did not seat properly in a few attempts. This malfunction
resulted in partial or complete loss of vacuum within the
core tube and thus sample loss.  The average core length
retrieved in this area was about 41 percent shorter than the
depth of sampler deployment.

In S2A2, only 3  of the 20 core  lengths collected by the
Hand Corer in the 4- to 12-inch bss depth interval equaled
the target core length. The actual core lengths ranged
from 0 to  8 inches, with a mean core length of 2 inches
that compared poorly to the target core length of 8 inches.
As mentioned above,  the Hand Corer  experienced the
greatest number  of problems in S2A2, perhaps because
this  area  contained  significant amounts of  partially
decomposed reeds  and leaves and live vegetation.  As a
result, the sediment matrix was heterogenous  and was
difficult to cut through, capture, and retain. The average
core length retrieved in this area was about 78 percent
shorter than the depth of sampler deployment.

In summary, the demonstration results indicate that the
Vibrocorer was able to consistently collect sediment from
the 0- to 4- and 6- to 12-inch bss depth intervals in S1A1
because the core lengths  for all attempts in both depth
intervals equaled the target core lengths.  The Hand Corer
collected  surficial  sediment well  but had  difficulty
collecting samples from depths greater than 4 inches bss.
Specifically, samples  collected  in the 0- to 4-inch bss
depth intervals equaled the target core length  in 41 of
44 attempts.  However, the  actual core lengths did not
equal the target core  length for any  of the  samples
collected in the 12- to 32- and 10- to 30-inch bss depth
intervals in S1A2 and S2A1, respectively, and equaled the
target core length in only 3  of 20 attempts in the 4- to
12-inch bss depth interval in S2A2.
7.1.3  Ability to Collect Multiple Samples with
       Consistent Physical or Chemical
       Characteristics, or Both, from a
       Homogenous Layer of Sediment

Primary  objective P3 involved  evaluating the Hand
Corer's ability to collect multiple samples with consistent
physical  or chemical characteristics,  or both,  from a
homogenous  layer of sediment.   This objective  was
addressed by calculating the RSD values for the sample
analytical results for the 12-to 32-inch bss depth interval
in S1A2, and the 0- to 4- and 10- to 30-inch bss depth
intervals   in  S2A1.   Based  on  the predemonstration
investigation  results,  these three depth intervals  were
determined to be homogenous in terms  of their physical
characteristics,  and the two S2A1 depth intervals were
determined to be homogenous in terms of their chemical
characteristics.

For the Hand Corer  samples, Figure 7-3 presents the
demonstration analytical  results for PSD in the 12- to
32-inch bss depth interval in  S1A2,  and  Figure 7-4
presents the demonstration analytical results for arsenic
and PSD  in the 0- to 4- and 10- to  30-inch bss depth
intervals  in S2A1.  The demonstration analytical results
for arsenic contain statistical outliers that indicate that the
two  S2A1  depth intervals  may not  be  chemically
homogenous. For this evaluation, the outliers are defined
as sample analytical  results  that are  not within  two
standard  deviations of the mean; the outliers  include the
250 mg/kg of arsenic in the 0- to 4-inch bss depth interval
and the 52 mg/kg of arsenic in the 10-to 3 0-inch bss depth
interval in S2A1.   Outliers  were  also  found in the
analytical results for samples collected by the Split Core
Sampler  (see Section  6.1.3), providing  further evidence
that the two S2A1 depth intervals may not be chemically
homogenous. A similar analysis performed for the PSD
results revealed no statistical outliers. Therefore, the Hand
Corer was evaluated based only on its ability to collect
multiple samples with consistent physical characteristics.
RSDs were calculated for each depth interval based on the
PSD analytical results for all locations sampled.

Table 7-5 presents the PSD summary  statistics (range,
mean,  and RSD) calculated for the  samples and field
triplicates collected using the  Hand Corer in each depth
interval relevant to primary objective P3.  As  stated in
                                                    60

-------
12- to 32 -inch bss de
Location 1A
Sand: 6%
Silt: 72%
Clay: 22%



Location 5A
Sand: 4%
Silt: 68%
Clay: 28%
3th interval







Location 3C
Sand: 3%
Silt: 70%
Clay: 27%







Location 1 E
Sand: 6%
Silt: 63%
Clay: 31%



Location 5E
Sand: 3%
Silt: 67%
Clay: 30%
Note:

bss  =  Below sediment surface

Figure 7-3. Hand Corer sample particle size distribution results for S1A2 (freshwater bay).
Section 6.1.3, RSDs calculated for the PSD results were
compared to  the  laboratory acceptance  criterion  of
15 percent for field triplicates.  When the RSD for all
samples from a given depth interval was greater than
15 percent, it was compared to the measured RSD for the
field triplicates.  An RSD for all samples that is less than
the RSD for field triplicates may be more attributable to
the  laboratory's  analytical  procedure,   the   sample
homogenization procedure  implemented  in  the  field,
or both  rather  than  the sampler's ability  to  collect
physically consistent samples. However, PSD parameters
with means less than 10 percent were not evaluated in this
manner because at low levels, the analytical method is not
as precise; as a result, it will generate high RSD values and
may not  actually reveal whether multiple samples with
consistent were have been collected.

For the 12- to 32-inch bss depth interval in S1A2, the
RSDs for silt and clay results were below the 15 percent
laboratory acceptance criterion.  The mean sand level was
less than 10 percent and was  not evaluated using the
                                                     61

-------
              0- to 4-inch bss depth interval
Location 1A
Arsenic: 250 mg/kg
Sand: 32%
Silt: 63%
Clay: 2%
Location 2A
Arsenic: 190 mg/kg

Location 3A
Arsenic: 140 mg/kg
Sand: 32%
Silt: 63%
Clay: 5%
Location 1 B
Arsenic: 130 mg/kg




Location 3B
Arsenic: 140 mg/kg




Location 2C
Arsenic: 120 mg/kg
Sand: 46%
Silt: 48%
Clay: 2%






Location 2D
Arsenic: 130 mg/kg




Location 1 E
Arsenic: 190 mg/kg
Sand: 26%
Silt: 72%
Clay: 2%
Location 2E
Arsenic: 150 mg/kg

Location 3E
Arsenic: 130 mg/kg
Sand: 29%
Si It: 71%
Clay: 0%
10- to 30-inch bss de|
Location 1A
Arsenic: 24 mg/kg
Sand: 38%
Si It: 61%
Clay: 0%
Location 2A
Arsenic: 8.3 mg/kg

Location 3A
Arsenic: 7.2 mg/kg
Sand: 37%
Silt: 58%
Clay: 4%
5th interval
Location 1 B
Arsenic: 8.5 mg/kg




Location 3B
Arsenic: 8.2 mg/kg




Location 2C
Arsenic: 9.7 mg/kg
Sand: 43%
Silt: 53%
Clay: 3%






Location 2D
Arsenic: 13 mg/kg




Location 1 E
Arsenic: 16 mg/kg
Sand: 35%
Silt: 62%
Clay: 3%
Location 2E
Arsenic: 7.2 mg/kg

Location 3E
Arsenic: 52 mg/kg
Sand: 35%
Silt: 62%
Clay: 3%
Notes:

bss      =    Below sediment surface
mg/kg   =    Milligram per kilogram

The particle size distribution results for a given sample may not total 100 percent because of rounding or because some sediment did not pass through
the U.S. Standard No. 4 sieve and was classified as gravel rather than sand, silt, or clay.


Figure 7-4. Hand Corer sample arsenic and particle size distribution results for S2A1  (lake).
                                                                62

-------
Table 7-5.  Particle Size Distribution Summary Statistics for Hand Corer
Demonstration Area
S1A2 (freshwater bay)


S2A1 (lake)





Depth
(inches bss)
12 to 32


Oto4


10 to 30


Parameter
Sand
Silt
Clay
Sand
Silt
Clay
Sand
Silt
Clay
Number of
Samples
5
5
5
5
5
5
5
5
5
Range (%)
3 to 6
63 to 72
22 to 31
26 to 46
48 to 72
Oto5
35 to 43
53 to 62
Oto4
Mean (%)
4
68
28
33
63
2
38
59
3
RSD (%)
(All Samples)
34
5
13
23
15
18
9
6
60
RSD (%)
(Field Triplicates)
0
3
8
3
6
29
14
2
71
Notes:

bss   =
RSD   =
Below sediment surface
Relative standard deviation
criterion. However, the sand levels exhibited a tight range
(3 to 6 percent).

For the 0- to 4-inch bss depth interval in S2A1, the RSD
for silt levels (15 percent) met the laboratory acceptance
criterion, but the RSD for sand levels (23 percent) did not.
Because the RSD for sand levels exceeded the criterion
but the RSD for sand  levels  in the field  triplicates
(3 percent) met the criterion, some of the variation in the
sand results may be attributable to the Hand Corer's ability
to collect  multiple  samples with  consistent physical
characteristics. The mean clay level in samples collected
in the 0- to 4-inch bss depth interval in S2A1 was less than
10 percent and was not evaluated using  the criterion.
However, the clay levels exhibited a tight range (0 to
5 percent).

For the 10- to 30-inch bss depth interval in S2A1, the
RSDs for sand and silt levels were below the 15 percent
laboratory acceptance criterion. The mean clay level in
samples collected in the depth interval  was less than
10 percent and was not evaluated using  the criterion.
However, the clay levels exhibited a tight range (0 to
4 percent).

In summary, the Hand Corer met the primary obj ective P3
criteria except for an exceedance  in the RSD for sand
levels  in the 0- to 4-inch bss  depth interval in S2A1.
Therefore,  it  was concluded that the  Hand  Corer is
generally able to collect multiple samples with consistent
physical characteristics.
                                              7.1.4  Ability to be Adequately Decontaminated

                                              Primary objective P5 involved evaluating the reference
                                              samplers' ability to be adequately decontaminated.  This
                                              objective was addressed by collecting equipment rinsate
                                              samples after sampler decontamination activities in S1A1
                                              and S2A1.  Specifically,  the  6- to  12-inch bss depth
                                              interval in S1A1 and the 0- to 4-inch bss depth interval in
                                              S2A1 were chosen as the depth intervals where P5 was
                                              evaluated because they contained high concentrations of
                                              PCBs and arsenic, respectively. Although it was intended
                                              that the evaluation of P5 be  limited  to these depth
                                              intervals, because samples were simultaneously collected
                                              in multiple depth intervals, the  primary objective was
                                              addressed for a given area, not for a given depth interval.
                                              However, this deviation did not impact the evaluation of
                                              primary objective P5.

                                              If the reference samplers were adequately decontaminated,
                                              the analytical results for the equipment rinsate samples
                                              would be  below the  analytical laboratory's reporting
                                              limits. To ensure that the water used to decontaminate the
                                              samplers was not itself contaminated,  decontamination
                                              water  blanks   were   also  analyzed.     Contaminant
                                              concentrations in both the equipment rinsate samples and
                                              decontamination water blanks were below the laboratory
                                              reporting limits for PCBs (1 part per billion) and arsenic
                                              (10 parts per billion). Thus, both the Vibrocorer and Hand
                                              Corer  demonstrated  the  ability  to   be   adequately
                                              decontaminated.
                                                     63

-------
7.1.5  Time Requirements for Sample Collection
       Activities

Primary objective P6 involved evaluating the reference
samplers'  time  requirements  for  sample  collection
activities.  These requirements were evaluated in all four
demonstration areas but were not specifically evaluated by
depth interval because  samples  were simultaneously
collected in multiple depth intervals to reduce the overall
sample  collection  time.   For the  Hand  Corer,  one
technician  was  required  for sampler  setup, sample
collection,   sampler   disassembly,  and  sampler
decontamination, except in S2A1 where two technicians
were required for sample collection. For the Vibrocorer,
two technicians  were required for sampler setup  and
sample collection,  and one technician was required for
sampler decontamination in S1A1.  Sampler disassembly
was not necessary because the Vibrocorer is a permanent
fixture aboard the EPA GLNPO' s Mudpuppy and does not
contain components that require disassembly.

The amounts of time required to complete the sampling
activities are shown in Table 7-6. The time measured for
sample collection activities did not include the time taken
for mobilization, demobilization, and maneuvering the
sampling platforms to sampling locations because these
latter activities were not sampler-specific; rather, they
were either site- or weather-related.

To complete sampling activities in  S1A1, the Vibrocorer
required 8 minutes for sampler setup, 124 minutes for
sample collection in  the 0- to  4- and 6- to  12-inch bss
depth intervals (15 to  16 minutes per attempt),  and
10 minutes for sampler decontamination.

For the Hand Corer, sampler setup required 4 minutes in
S1A2. Sampler setup times are not available for S2A1 and
S2A2. In S2A1, the setup time was included in the sample
collection time for one particular sample, and in S2A2, the
setup time was not recorded. However, the setup time
recorded at S1A2 is probably representative of the time
needed for a moderately experienced technician to set up
the Hand Corer; S1A2 was the last demonstration area
sampled with the Hand Corer, so the technician had ample
opportunity to practice sampler setup in other areas.

Sample collection times for the Hand Corer ranged from
47 to 550 minutes in S1A2, S2A1, and S2A2.  Sample
collection with the Hand Corer required 4 to 7 minutes per
attempt  in S1A2 and  S2A2 but 10 to 16  minutes per
attempt in S2A1. More extension rods were required in
S2A1 than in the other two areas because  of the water
depth; five rods were required in S2A1, but only one rod
was  required in  S1A2 and S2A2.  The weight  of the
additional extension rods made use of a tripod-mounted
winch necessary to  hold the  sampler  steady  during
sampling; incorporating the tripod-mounted winch into the
sampling process in S2A1 accounted for the extra time
necessary for sample collection.

Hand Corer disassembly required 2 minutes in S1A2 and
S2A2 but 4 minutes  in S2A1.   The additional time
required  in S2A1 can again be attributed to the use of
additional extension rods in this area.
Table 7-6. Time Required to Complete Sampling Activities for Reference Samplers
                                                      Time Required (minutes)
Activity
Sampler setup
Sample collection
Sampler disassembly
Sampler decontamination
Total
S1A1 (River Mouth)
Vibrocorer
8
124
0
10
142
S1A2 (Freshwater Bay)
Hand Corer
4
47
2
Not evaluated
53
S2A1 (Lake)
Hand Corer
Included in sample collection
550
4
40
594
S2A2 (Wetland)
Hand Corer
Not recorded
163a
2
Not evaluated
165a
Note:
    Hand Corer sampling was completed at four of five sampling locations. At the fifth location, sampling was discontinued after 12 attempts failed
    to collect the specified sediment volume.
                                                    64

-------
Decontamination of the Hand Corer was evaluated only in
S2A1 and required 40 minutes. Because of the numerous
extension rods required in this area, the decontamination
time measured in S2A1 may not be representative.  In
addition, S2A1 was the first demonstration area sampled,
and decreased decontamination times were observed for
the other samplers as the technicians became more familiar
with the decontamination procedures  required for the
demonstration.

In summary, a technician  familiar with the Vibrocorer
would be expected to require 8 minutes for sampler setup,
15  to 16  minutes for each  sampling attempt,  and
10 minutes for sampler decontamination.  A technician
familiar with the Hand Corer would be expected to require
4 minutes for sampler setup, 4 to 7 minutes for  each
sampling attempt, and  2 to 4  minutes  for sampler
disassembly.  However, more time may be necessary for
sample collection depending on the water  depth.  It is
uncertain how much time an experienced technician would
need to adequately decontaminate the Hand Corer, but it
is likely that the technician would require less than the
40  minutes  observed   in  S2A1.    The   amount  of
decontamination time would likely have been less in the
other areas because the technician would have had more
practice in  implementing the required decontamination
procedures  as  well   as  fewer  extension  rods  to
decontaminate.  When sediment sampling activities are
planned,  the time required for setting up the  sampling
platform and for maneuvering the platform to position the
sampler at the sampling  location would  have to be
considered in  addition to the times presented above.

7.2    Secondary Objectives

This section discusses the performance results for the
reference samplers based  on secondary objectives SI
through S5  stated in Section 4.1. Secondary objectives
were  addressed based on observations of the reference
samplers' performance during the demonstration and on
information provided by the EPA GLNPO.

7.2.1  Skill and Training Requirements for
       Proper Sampler Operation

The Hand Corer is easy to operate, requiring minimal
skills and  training.   Sampler  assembly  and sample
collection procedures can be learned in the field with a
few  practice attempts.   In addition, a  written SOP
accompanies the  sampler when it is procured.   The
sampler can be operated by  one person in  shallow
(wading) water depths because of its lightness (12 Ib).
Sampler  operation with  plastic  core  liners  is  simple
because  the  sampler  does  not  require   complete
disassembly and reassembly after each sampling attempt.
Only the nose piece requires removal to extrude the plastic
core liner containing the sediment core.  In water depths
requiring  use of extension rods, sampler  operation
becomes  more  cumbersome because  of the  combined
weight of the stainless-steel sampler and the galvanized-
steel extension rods (5 Ib each).  Because of the heaviness
of the  sampler equipped  with five extension rods, two
personnel and a tripod-mounted winch were  needed to
deploy and retrieve the sampler at each sampling location
in S2A1, where the water  depth was about  18 feet.

During the demonstration, minimal strength and stamina
were required to collect samples with the Hand Corer from
shallow and moderate depth intervals containing both
unconsolidated and consolidated sediments. Specifically,
minimal strength and stamina were required to drive the
sampler into and retrieve it from the 0- to 4-inch bss depth
interval in S1A2 and  S2A1 and the moderate depth
intervals ranging from 10 to 30 and 12 to 32 inches bss in
S2A1 and S1A2,  respectively.   However, moderate  to
significant strength and stamina were required to collect
samples  from  a  depth  interval containing partially
decomposed  reeds  and  leaves  and  live vegetation.
Specifically, moderate to significant strength and stamina
were required to drive the sampler into and retrieve it from
the 4- to 12-inch bss depth interval in S2A2. Sediment in
this interval was consolidated and was predominantly sand
with low water content.  The consolidated  interval
increased the amount of force required to drive the Hand
Corer.  However, the difficulty in driving the sampler was
likely attributable to the sampler's inability to cut through
the  sediment that contained  significant amounts  of
partially decomposed reeds and leaves and live vegetation.

Previous  sediment sampling experience is beneficial in
selecting the most appropriate  support equipment for a
given Hand Corer application. For example, the sampling
technicians chose to use a tripod-mounted winch in S2A1
because of the significant strength and stamina that would
have been required to deploy and retrieve the sampler in
that area if a winch was  not used. Previous  sediment
sampling  experience is  also  beneficial  in accurately
                                                   65

-------
assessing the location of the sediment surface using the
sampler, as is the case with other samplers.

Operation of the Vibrocorer requires moderate skills and
training, and the sampler must be operated by at least two
persons using a sampling platform.   Several hours of
hands-on  training  with an  experienced  Vibrocorer
sampling technician is recommended to learn the proper
operation of the sampler and its support equipment.  In
addition, during the demonstration, the power supply for
the Vibrocorer malfunctioned during sample collection.
The source of the malfunction was identified and corrected
by on-site personnel. Therefore, it is recommended that at
least one of the sampling technicians have electrical and
mechanical  experience   to   be  able  to  correct
malfunctioning support equipment for the  Vibrocorer.
Also, previous sediment sampling experience is beneficial
in assessing the location of the sediment surface using the
sampler, as is the case with other samplers.

During the demonstration, minimal strength and stamina
were required to collect samples with the Vibrocorer in
S1A1.  Although the vibrohead and core tube weigh about
150 Ib, sampler deployment and retrieval were controlled
with an A-frame  and winch  on the EPA GLNPO's
Mudpuppy.  The physical effort required to remove the
core tube from the vibrohead and to extract the sample
from the core tube was minimal.

7.2.2   Ability to Collect Samples Under a Variety
        of Site Conditions

The Hand Corer  demonstrated its  ability  to  collect
sediment samples under all conditions encountered during
the demonstration, which included a variety of sampling
platforms, water depths, sediment depths, and sediment
compositions.   The range of sampling platforms  used
included wooden planks fastened to ladders in S2A2; an
18-foot-long,  4-foot-wide  Jon  boat in S1A2; and  a
sturdier, 30-foot-long, 8-foot-wide pontoon boat in S2A1.
Because the  sampler does  not require electricity or a
tripod-mounted winch for deployment in shallow water,
sampler operation was feasible from any location on the
sampling platforms used in  S1A2 and S2A2.  At S2A1,
however, where the water depth was about 18 feet, two
sampling technicians and a tripod-mounted winch  were
needed  to properly operate the sampler because of the
combined weight of the sampler (12 Ib) and the five
extension  rods and turning  handle (27 Ib).  Use of the
tripod-mounted winch required that a 6-inch-diameter hole
be cut in the center of the pontoon boat to deploy and
retrieve the sampler.

As with other samplers, the ability to assess the location of
the sediment surface with the Hand Corer decreases with
increasing water depth and  turbidity.   Because of the
significant water depth in S2A1 and turbidity in S1A2, the
sampling technicians could not see the sediment surface
from  the  sampling platforms.  An underwater  video
camera may have  enabled the sampling technicians to
accurately assess the location of the sediment surface in
these areas (Blomqvist 1991).

The Hand Corer was able to collect sediment samples in
all shallow  and moderate  depth intervals (less  than
36 inches bss) in  each demonstration area where the
sampler was  deployed.   However,  as  discussed in
Section 7.1.1.1, the actual number of attempts required to
collect the  specified volume of sediment exceeded the
expected number  at most  sampling  locations.    The
additional attempts may be  attributable to  (1) error in
assessing the location of the sediment surface, which may
have resulted in the actual depth of penetration being less
than the measured depth of penetration; (2) deficient entry
of sediment into the core tube (core shortening); (3) plug
formation in the coring tip that inhibited further sediment
retrieval; or (4) partial or complete loss of the sediment
core through the bottom end of the sampler as a result of
partial or complete loss of vacuum in the core tube caused
by incomplete closure of the flutter valve.  Incomplete
closure of the flutter valve was observed during a few
attempts in S2A2 when partially decomposed plant matter
in the 0- to 4-inch bss depth interval became lodged
between the flutter valve and core tube. Core shortening
(in which the actual core length retrieved is less than the
depth of sediment penetration) primarily involves deficient
entry  of sediment  into the core tube  during core tube
penetration.   Physically, sediment friction against the
inside wall of the core tube causes thinning of the core by
lateral extrusion in front of the core tube.  As the friction
changes with  depth, not all sediment layers may be
uniformly represented in the sample (Blomqvist 1991).

The Vibrocorer demonstrated its ability to  consistently
collect sediment samples in the 0- to 4- and 6- to 12-inch
bss depth intervals at all locations in S1A1. As discussed
in Section 7.1.1.1, the actual number of attempts required
to collect the specified volume of sediment in these depth
                                                    66

-------
intervals did  not  exceed the  expected  number at any
sampling locations.   However,  the  sampler could not
collect cores  longer  than 4.4 feet.  The Vibrocorer's
difficulty in collecting sediment in the 4- to 6-foot bss
depth interval may be attributed to the sampler not being
able to penetrate clay hardpan observed  in the sampling
area about 5 feet bss.

The Vibrocorer was unable to collect samples in S1A2, as
was originally intended. The sampler was installed on the
EPA GLNPO's Mudpuppy,  which requires a minimum
water depth of 3 feet for maneuvering. Because the water
depth in S1A2   was  only about  2  feet during the
demonstration, the Mudpuppy was unable to enter the area.

7.2.3  Ability to Collect an Undisturbed Sample

During  the demonstration,  both the Hand Corer and
Vibrocorer consistently  collected  sediment samples in
which the sediment stratification was preserved; however,
based on visual observations, the samples appeared to have
been compacted. Bow wave disturbance near the sediment
surface did not occur in S2A2; the water depth (0.5 to 1.5
feet)  and low turbidity in  this  area  allowed  visual
confirmation of the location of the sediment surface. Bow
wave disturbance  near  the sediment  surface  in the
remaining demonstration areas was unlikely because the
speed of sampler deployment was controlled for each
sampler. As mentioned above, sediment stratification was
preserved for samples collected in these areas.

For both samplers, the total core length retrieved in each
attempt was less than the depth of sampler deployment.
The difference between the total core length retrieved and
the depth of sampler deployment for the  Hand Corer
ranged from 15 to 25 inches in S1A2, 1 to  36 inches in
S2A1, and 12 to 67 inches in S2A2. For the Vibrocorer,
the difference ranged from  10.5 to 38.5 inches.   As
discussed above, these differences may have resulted for
the reasons described in Section 7.2.2.  Furthermore, these
differences indicate  that  sampling  bias  might  have
occurred during sample collection in a given target depth
interval.

7.2.4  Durability Eased on Materials of
        Construction and Engineering Design

The primary components of the Hand Corer include (1) a
Lexan™ nose piece; (2) a 3 6-inch-long, stainless-steel
core tube; (3) a stainless-steel head piece with a flutter
valve; (4) two detachable, stainless-steel handles; and (5) a
clevis (see Figure  5-1).  Based on observations made
during the demonstration, the Hand Corer is a sturdy
sampler; none of the sampler components was damaged or
required repair or replacement during the demonstration.

The Hand Corer was also equipped with varying lengths of
galvanized-steel extension rods during the demonstration.
One extension rod was used to collect samples in shallow
water at S1A2 and S2A2.  In both areas, no bending or
bowing of the extension rod was observed. In S2A1, five
extension rods were coupled together to a combined length
of about  25 feet.  Throughout most of the sampling in
S2A1, minimal bowing of the coupled extension rods was
observed  during  sediment penetration.   During  one
sampling attempt in S2A1, the pontoon boat drifted after
the sampler had been deployed  through the  6-inch-
diameter  hole in the  middle of the boat and had been
driven into the sediment.  The  resulting stress on  the
extension rods caused one of the rods to be damaged at the
threads.

The primary components of the Vibrocorer include (1) an
anodized-aluminum, pressure-housed vibrohead with a
terminal for an electric cable; (2) a disposable,  10-foot-
long, 4-inch-diameter, plastic core tube equipped with a
plastic core catcher; (3) a core tube clamp; and (4) a guide
rope (see  Figure 5-2).  Based on observations made during
the demonstration, the  Vibrocorer is a sturdy sampler;
none  of  the primary components of the  sampler  was
damaged or required repair or replacement  during  the
demonstration. The primary component of the Vibrocorer,
the vibrohead, has an  operating expectancy of about
10,000 hours.  However, as discussed above, the power
supply for the Vibrocorer malfunctioned during sample
collection. The source of the malfunction (moisture in the
control box between the power source and vibrohead) was
identified and corrected by on-site personnel.

7.2.5  Availability of Sampler and Spare Parts

No primary component of  the Hand Corer required
replacement or servicing during the demonstration. Had
a primary sampler component required replacement, it
would not have been available in local retail stores.  As
discussed above, an extension rod was damaged at  the
threads   during  sampling  in  S2A1  and   required
replacement. The replacement rod was acquired within a
                                                   67

-------
few hours in a local retail store.  Replacement extension
rods and primary sampler components may be obtained
from the developer by overnight courier in 2 days or less,
depending on the location of the sampling site. During
sampling in S1A2, the sampling technician was able to
acquire  additional plastic core tube  liners from  the
developer by overnight courier. The developer precut the
plastic core tube liners in response to  a special request
from the sampling technician.  During sampling in S2A1
and S2A2, the sampling technician was able  to have
plastic core tube liners precut at a local machine shop.

No primary  component of  the  Vibrocorer  required
replacement or  servicing during  the demonstration.
However, as discussed above, the power supply for the
Vibrocorer malfunctioned and required servicing.  The
source of the malfunction was identified and corrected by
on-site  personnel within a  few hours.   Had on-site
personnel  been  unable  to  correct  the malfunction,
servicing of the power supply by an off-site electrician
would  have  been  necessary.    Had the  vibrohead
malfunctioned, it would have been packaged and shipped
to the developer for servicing. Because the vibrohead is
pressure-sealed,  servicing  of the  vibrohead is  not
recommended  in the field or  by an unskilled  sampling
technician. Plastic core tubes for the Vibrocorer may be
available from a local plastic manufacturer; however, their
availability should be verified prior to a sampling event,
especially one  in a remote location.  Core tube  catchers
used by GLNPO  can be made  from materials readily
available in a hardware store.

7.3     Data Quality

The overall QA objective for the demonstration was to
produce well-documented data of known quality.  The
TS As conducted to evaluate data quality did not reveal any
problems  that  would  make  the  demonstration  data
unusable.   The scope  of these TSAs is described in
Sections 4.3 and 4.4 of this ITVR.

This  section  briefly discusses  the  data  quality  of
demonstration  results for the  reference samplers; more
detailed information is provided in the DER (Tetra Tech
1999b).  Specifically, the data quality associated with the
field measurement activities is discussed first, followed by
the data quality associated with the laboratory analysis
activities.
7.3.1  Field Measurement A ctivities

Field measurement  activities  conducted during  the
demonstration  included  measurement  of  the  time
associated  with  sample  collection  activities,  water
velocity, water depth, core length, volume of IDW, volume
of sediment collected in a given sampling attempt, and
depth of sampler deployment.  Of these measurement
parameters, specific acceptance criteria were set for the
precision associated  with the time and  water  velocity
measurements only (EPA 1999). All time and water
velocity measurements made during the demonstration met
their respective criteria (see Table 6-7). Of the remaining
parameters,  some  difficulties  were  encountered  in
measuring the volume of sediment collected in a given
sampling attempt and the depth of sampler deployment,
which are discussed below.

To measure the volume of sediment collected in a given
sampling attempt, the sediment sample was transferred
into a 2-L container graduated in increments of 20 mL.
The container was tapped on a hard surface to minimize
the presence  of void spaces in the sample, the sample
surface was made even using a spoon, and the volume of
the sample was measured.  However, because the void
spaces could not be completely eliminated, the volumetric
measurements are believed to have a positive  bias that
resulted in overestimation of PSRs.  Because the total
volume of the void spaces could not be measured,  its
impact on the  PSR results could not  be  quantified.
However, because the same volumetric measurement
procedure was used for both the innovative and reference
samplers, the PSR results could still be compared.

The  depth of sampler deployment was measured with
reference  to  the  sediment  surface.    To identify the
location of the sediment surface, the sampling technicians
lowered the sampler into the water and used the bottom
end  of the  sampler to  feel  the  sediment  surface.
Subsequently, the technicians drove the sampler into the
sediment to a depth that they estimated to be appropriate
to collect a  sediment  sample in the specified depth
interval.   For the Vibrocorer  in S1A1, this approach
resulted  in  an average  core  length that was  about
23 percent shorter than the estimated depth of sampler
deployment, indicating that the sampling technicians may
have had  difficulty assessing the location of the sediment
surface. For the Hand Corer in the remaining three areas,
the average  core length retrieved was shorter than the
                                                   68

-------
estimated depth of sampler deployment, again indicating
that the sampling technicians may have had difficulty
assessing  the  location   of the   sediment   surface.
Specifically, for the Hand Corer in S1A2, S2A1, and
S2A2, the  average  core  length was  shorter  than  the
estimated depth of sampler deployment by 52, 41, and
78 percent, respectively.  Although both the innovative
and reference  samplers used in the demonstration  are
end-filling samplers that  do not collect uncompressed
sediment  samples,  the   degree  of  sediment sample
compaction in the core tube varied depending on  the
sampler used.  In addition, core shortening, which would
impact the ability of the samplers to uniformly sample the
sediment in a given depth interval, occurs to a different
degree depending on the sampler used. For these reasons,
conclusions drawn from a comparison of the  sediment
characteristics  of the samples collected by the reference
samplers with those of the samples collected by the Split
Core Sampler should be carefully interpreted.

7.3.2  Laboratory Analysis Activities

The laboratory analyses conducted for the demonstration
included the  following:  (1)  PCB,  arsenic,  and PSD
analyses of sediment samples and  (2)  PCB and arsenic
analyses of equipment rinsate samples.  To evaluate the
data quality of the laboratory  analysis results,  field-
generated QC  samples, PE samples, and laboratory QC
check samples were analyzed.  The field-generated QC
samples included the  field replicates  and  temperature
blanks described in Section 4.3 of this ITVR.  The PE
samples and laboratory QC check samples are described in
Section 4.4.  The acceptance  criteria for the QC samples
are presented in Table 6-7.

All temperature blanks and field replicates subjected to
PCB and arsenic  analyses met the acceptance criteria,
indicating that the sample homogenization procedure (field
replicates)  and   sample  preservation   procedure
(temperature blanks)  implemented in the field met the
demonstration  requirements.   However, as stated  in
Section 7.1.3, in a few cases the results of field triplicate
sample  analyses for PSD did not  meet the acceptance
criterion.  Despite the failures to  meet the acceptance
criterion, the PSD results are considered to be valid for the
reasons detailed in Section 7.1.3.
The PE sample results for both PCB and arsenic analyses
met the acceptance criteria, indicating that the analytical
laboratory accurately measured PCBs and arsenic.

The analytical results for all laboratory QC check samples
except the  following  met the  acceptance criteria:
(1)  MS/MSD  samples  for analysis for  PCBs  in  the
sediment matrix and (2) equipment rinsate samples for
PCB analysis. These issues and their likely impact on data
quality are discussed below.

For the sediment matrix, in all MS/MSD samples analyzed
for PCBs, Aroclor 1016 was recovered at levels higher
than the upper limit of the acceptance criterion, indicating
a positive bias in the PCB results for sediment samples.
However, the  analytical laboratory  had  no problem
meeting the acceptance criteria for control samples such as
BS/BSDs.   For this reason, the failure to meet  the
acceptance criterion for MS/MSD sample analysis was
attributed to  matrix interference.  Because Aroclor 1016
was recovered at levels higher than the upper limit of the
acceptance criterion in all MS/MSD samples associated
with both the innovative and reference samplers, the PCB
results could still be compared.   The MS/MSD spiking
compounds (Aroclors 1016 and 1260) were selected based
on the Aroclors detected during the predemonstration
investigation   and  as  recommended   in  SW-846
Method 8082.

Also for the sediment matrix, in one out of three MS/MSD
pairs analyzed for PCBs, Aroclor 1260 was recovered at a
level less than the lower limit of the acceptance criterion
in the MS sample, but the recovery in the associated MSD
sample was acceptable. Because the investigative samples
contained only Aroclor  1242,  of the  two  spiking
compounds used to prepare the MS/MSD  samples, only
the Aroclor 1016 recoveries were considered to be relevant
based on the  PCB congener distribution; the Aroclor 1260
recoveries were not considered to be relevant. Therefore,
the  low recovery associated with Aroclor 1260 had no
impact on data quality.

In all equipment rinsate samples analyzed for PCBs,
decachlorobiphenyl (the surrogate) was recovered at levels
lower than the  lower limit  of the acceptance criterion,
indicating a negative bias in the PCB results for equipment
                                                   69

-------
rinsate samples.  However, the analytical laboratory had   analysis was attributed to matrix interference. Because the
no problem meeting the acceptance criteria for control   surrogate was recovered at levels lower than the lower
samples such as PE samples and deionized water blanks.   limit of the acceptance  criterion in all equipment rinsate
For this reason, the failure to meet the surrogate recovery   samples associated with both the innovative and reference
acceptance criterion for the equipment  rinsate sample   samplers, the PCB results could still be compared.
                                                   70

-------
                                               Chapter 8
                                         Economic Analysis
As  discussed  throughout this  ITVR,  the  Split Core
Sampler was demonstrated at two sites, each consisting of
two areas. This chapter presents an economic analysis of
sediment sample collection using the Split Core Sampler
in two of the four demonstration areas: (1) a river mouth
contaminated  with  PCBs  (S1A1)  and  (2) a  lake
contaminated with arsenic  (S2A1).  These areas were
selected for the economic  analysis because the varied
sampling conditions in these  areas provide a range of costs
involved in conducting sediment sampling using the Split
Core Sampler.  For example, during the demonstration in
S1A1, the water depth was about 5 to 6 feet, and sediment
samples were collected in two depth intervals: 0 to 4 and
6 to 12 inches bss.  On the other hand, in S2A1, the water
depth was  about 18 feet, and sediment  samples were
collected in two depth  intervals: 0  to  4  and  10  to
30 inches bss.

The purpose of this economic analysis is to estimate the
costs of using the Split Core Sampler to collect sediment
samples in environments similar to S1A1 and S2A1. The
analysis is based on the results of the demonstration, unit
costs in published cost data sources, and costs provided by
the technology developers or equipment vendors.

This chapter provides  information on the issues and
assumptions  involved  in  the  economic  analysis
(Section 8.1), discusses the costs associated with using the
Split Core  Sampler (Section  8.2), discusses  the costs
associated with using the reference samplers (Sections 8.3
and 8.4), and  presents a comparison  of  the  economic
analysis results for the Split Core Sampler and reference
samplers (Section 8.5).
8.1    Issues and Assumptions

Several factors affect sediment sampling costs.  In this
economic analysis, wherever possible, these factors are
identified such that decision-makers can independently
complete  a site-specific  economic  analysis.    Costs
included  in the analysis are divided into four categories:
sampler,  labor,  IDW disposal, and support equipment
costs.  The issues and assumptions associated with these
categories and the costs not included in this analysis are
briefly discussed below.

8.1.1  Sampler Costs

Sampler costs include the costs of samplers and associated
equipment used during  the  demonstration,  such  as
extension rods and core tube liners, as applicable.  These
costs were provided by the technology developers  or
equipment vendors.

8.1.2  Labor Costs

Labor costs cover the time required for sampler setup,
sample collection,   sampler disassembly,  and  sampler
decontamination.  In this analysis, the actual amount of
time required for sample collection activities during the
demonstration is used as the labor requirement,  and all
labor times are  rounded off to  the  nearest  half-hour.
Because  it may not be feasible to hire  sampling
technicians for a fraction of a day, a site-specific analysis
should consider the local availability of such technicians
and  modify labor cost estimates  accordingly.   In this
analysis,  an hourly rate of $13.51 is used for a technician
(R.S.  Means  Company  [Means]  1999),  and   a
multiplication factor of 2.5 is applied to labor costs in
                                                   71

-------
order to account  for  general  and administrative  and
overhead costs. Thus, an hourly rate of $34 is used for a
technician.

8.1.3  WWDisposal Costs

IDW disposal costs cover disposal of unused sediment and
spent core tube liners. Unused sediment was assumed to
be   a   nonhazardous  waste   because   during  the
demonstration, the sediment PCB concentrations in S1A1
did not exceed 3.7 parts per million, and wastes containing
PCB concentrations less than 50 parts per million can be
disposed of as nonhazardous waste (40 Code of Federal
Regulations [CFR] 761). Similarly, arsenic-contaminated
wastes  that   are  not  listed  wastes  with   toxicity
characteristic   leaching  procedure  (TCLP)   extract
concentrations less than 5 milligrams per liter (mg/L) can
be disposed of as nonhazardous  waste (40 CFR 261).
During the demonstration, the maximum and average
arsenic concentrations in sediment in S2A1 were 300 and
70 mg/kg, respectively.  Based on  the  average arsenic
concentration and the dilution factor (20) associated with
the TCLP,  the TCLP extract concentration  for the
sediment waste generated during the demonstration was
estimated to  be about  3.5  mg/L.  Therefore, unused
sediment in S2A1 was also assumed to be a nonhazardous
waste.
During  the demonstration,  insignificant  quantities of
sediment were present on the spent core tube liners.
Therefore, the spent core tube liners were also assumed to
be a nonhazardous waste. Also, as shown in Table 8-1, the
samplers generated different quantities of IDW in each
demonstration  area.   However,  the volume of  IDW
generated by each sampler  in each  area was less than
55 gallons. Because the cost to package, load, transport,
and dispose of smaller containers is generally the same as
the cost to perform these activities for one 5 5-gallon drum,
it is assumed that the IDW in each area would be collected
in a 55-gallon drum.  As a result, the cost for  IDW
disposal is the same for each  sampler. However, if larger
numbers of samples were to be collected and the resulting
IDW volume were  larger, differences in IDW disposal
costs among samplers would become apparent. The cost
to package, load, transport, and dispose of one 55-gallon
drum of nonhazardous waste is $182 (Means 1999).

8.1.4  Support Equipment Costs

Support equipment includes  equipment used for sampler
preparation,  sample  extrusion,  and  other  activities
associated with sample collection.  Examples  of support
equipment are a tripod-mounted winch and an electrical
power generator.
Table 8-1.  Comparison of Investigation-Derived Waste Quantities Generated by Split Core Sampler and Reference Samplers

                                                        Quantity of Investigation-Derived Waste
Demonstration Area
S1A1 (river mouth)
S2A1 (lake)
Sampler
Split Core Sampler
Vibrocorer
Split Core Sampler
Hand Corer
Unused
Sediment
(liters)
3
45
7
12
Number of Core Tubes
Not applicable
5a
Not applicable
Not applicable
Number of Core Tube
Liners
9b
Not applicable
52°
41 d
Number of Core Tube
Liner End Caps
18
Not applicable
104
Not applicable
Notes:

a    10-foot-long, 4-inch-diameter, plastic core tubes

b    18-inch-long, 2-inch-diameter, plastic core tube liners

0    36 6-inch-long, 2-inch-diameter, plastic core tube liners and 16 36-inch-long, 2-inch-diameter, plastic core tube liners

d    36-inch-long, 2-inch-diameter, plastic core tube liners
                                                     72

-------
8.1.5  Costs Not Included

Items whose  costs are not included in this analysis are
identified below along with a rationale for the exclusion of
each.

Oversight of Sampling Activities. A typical user of a
sampler  would not be required to pay  for  customer
oversight of  sample collection.   EPA representatives
audited all  activities associated with  sample collection
during the demonstration, but costs for EPA oversight are
not included  in this analysis because they are project-
specific  and  not sampler-dependent.   In addition, if
physical characterization of sediment samples is required
to be performed  in the field, a soil scientist may be
necessary.  However, costs for such oversight are not
included  in this analysis because they are project-specific
and not sampler-dependent.

Health  and  Safety Personnel.   Health  and safety
personnel are  required to be present during hazardous
waste site operations, but they are not directly involved in
sample collection activities.

Analyses of Samples Collected.  Analytical costs can
vary greatly depending on site-specific contaminants and
are not directly related to sample collection costs.

Personal Protective Equipment.  The type of personal
protective equipment required can vary greatly depending
on site-specific contamination and hazards, and the cost of
such equipment is not sampler-dependent.

Disposal of Decontamination Water. Decontamination
water may  frequently be disposed of without incurring
additional costs  (as  was  the  case  during the demon-
stration).

Travel and Per Diem for the Sampling Team. Members
of the sampling team may be available locally.  For the
demonstration, the sampling team consisted of both local
and nonlocal staff. Because the availability of sampling
team members is a function of the geographic location of
the sampling  site and does not depend on the samplers,
travel and per diem  costs for the sampling team are not
included  in this analysis.

Boat Rental.  A boat may or may not be  necessary for
sediment sampling, depending on site conditions and the
sampler chosen. Because the cost of boat rental is not
included in this analysis, other costs associated with using
a boat, such as fuel costs,  are also not included.

Time Spent in Maneuvering the Sampling Platform.
The time  required to maneuver the sampling platform
varies greatly depending on site conditions such as water
depth and weather. For example, when the wind velocity
was high during the demonstration, a significant amount of
time  was spent  maneuvering  the  EPA  GLNPO's
Mudpuppy (in S1A1) and  the pontoon boat (in S2A1); as
a result, the sampling sometimes had to be discontinued
for the day.   Because these delays were not sampler-
dependent, the time spent in maneuvering the sampling
platforms  is not included in this analysis.

Time Spent  in Managing the Samples.  The time
required to homogenize the sediment, fill and label sample
containers, prepare sample containers for shipment, fill out
chain-of-custody forms,  and ship the  samples  varies
greatly depending on the number of samples collected and
site location.  Therefore, the time spent in managing the
samples is not included  in this  analysis because it is
project-specific and not sampler-dependent.

Mobilization  and Demobilization.  Mobilization and
demobilization costs vary greatly depending on the site
location   and  conditions.    For  the   demonstration,
mobilization and demobilization  activities were  mainly
associated with procuring sampling platforms and setting
up sample management areas.  The sampling platforms
used were selected based on their availability but not
necessarily based on sampler requirements. For example,
in S1A1, the EPA GLNPO's Mudpuppy was used because
it was available free of charge from EPA Region 5. Also,
two tents were set up for sample management in S1A1 and
S2A1 to avoid delays resulting from inclement weather but
not  based  on  sampler requirements.    Therefore,
mobilization and demobilization costs are not included in
this analysis.

Commonly Available Support Equipment. The cost of
support equipment that is  commonly available and likely
would not be purchased specifically for sampling is not
included in  this analysis.   For  example, the cost  of
wrenches  and tape  measures  is  not  included  in this
analysis because it is assumed that a field sampling team
would already have such tools as part of its field sampling
gear.
                                                   73

-------
Support Equipment That Costs Less Than $10. The
cost of inexpensive support equipment, such as stainless-
steel  spoons and mixing bowls used  to homogenize
sediment samples is not included in this analysis.  In
addition, the cost of fuel consumed to  operate support
equipment such as a generator is not included because,
based on the fuel consumed during the demonstration, the
fuel cost was estimated to be less than $10.

8.2     Split Core Sampler Costs

This section presents information on sampler, labor, IDW
disposal, and support equipment costs for the Split Core
Sampler as well as a summary of these costs. Table 8-2
presents these costs.

8.2.1    Sampler Cost

In S1A1, AMS used the Split Core Sampler kit ($524) as
well as  one 6-inch-long core  tube ($211); one core tube
coupling ($72.50); one AMS  rubber-coated cross handle
($31); nine 18-inch-long, plastic core liners ($6 each); nine
pairs of core liner end caps ($0.21 per pair); two 4-foot-
long,  stainless-steel  extension rods  ($69  each); and one
3-foot-long, stainless-steel extension rod ($67).  The
sampler kit contained a 12-inch-long core tube; top cap;
coring tip; basket retainer; 12-inch-long, plastic liner; and
slip wrench.  The  total sampler cost  for S1A1 was
estimated to be $1,097.50. The total sampler cost shown
in Table 8-2 does not include the cost of end caps because
the total cost of the end caps used was less than $10.

In S2A1, AMS used the Split Core Sampler kit ($524) as
well as one 6-inch-long core tube ($211); two additional,
12-inch-long core  tubes  ($229 each);   two  core tube
couplings ($72.50 each);  one AMS rubber-coated cross
handle ($31); 36 6-inch-long, plastic core tube liners ($2
each); 16 36-inch-long, plastic core liners ($9  each);
52 pairs of core liner end caps ($0.21 per pair); one 3-foot-
long, stainless-steel extension rod ($67); four 4-foot-long,
stainless-steel extension rods  ($69 each); and two 4-foot-
long, carbon-steel extension rods ($48.50 each). The total
sampler cost for S2A1 was estimated to be $2,036.

8.2.2    Labor Cost

In S1A1, the time for sampler  setup, sample collection,
sampler disassembly, and sampler decontamination totaled
68 minutes or about 1 hour for one technician.  In this
area, five investigative samples each were collected in the
0- to 4- and 6- to 12-inch bss depth intervals using the
Split  Core  Sampler.   Table 4-3  presents  additional
information on the total number of samples collected. The
labor cost for sampling in S1A1 was estimated to be $34.

In S2A1, the time for sampler setup, sample  collection,
sampler disassembly, and sampler decontamination totaled
639 minutes or about 11 hours for one technician. In this
area, 15 investigative samples each were collected in the
0- to 4- and 10- to 30-inch bss depth intervals using the
Split Core Sampler.  The labor cost for sampling in S2A1
was estimated  to be  $374. When field technicians work
more  than  8 hours  in 1 day, overtime costs  may be
incurred. In this estimate, however, no overtime costs are
included.

8.2.3  IDW Disposal Cost

Sampling in S1A1 generated IDW consisting of 3  L of
unused sediment, 9 core liners, and 18 end caps. The cost
for disposal of one 55-gallon drum of nonhazardous waste
is $182.

Sampling in S2A1 generated IDW consisting of 7  L of
unused sediment, 52 core liners, and 104 end caps.  The
cost for disposal of one 55-gallon drum of nonhazardous
waste is $182.

8.2.4  Support Equipment Cost

Support  equipment used during Split  Core  Sampler
sampling in S1A1 included  one electric hammer (with a
rental cost of $40 per day [Wirtz Rentals Co.  1999]), an
SDS Max self-locking adapter for attaching the electric
hammer to the top extension rod ($90 [Tetra Tech 1999a]),
one sample extrusion rod ($25 [Tetra Tech  1999a]), two
slip wrenches, one carbon-steel bristled brush,  and one
stainless-steel  bristled brush.  The  costs of  the  slip
wrenches and  brushes are not included in  this  analysis
because a field sampling  team would already have  such
tools as part of its field sampling gear. The  total cost for
support equipment for S1A1 is $155.

Support  equipment used during Split  Core  Sampler
sampling in S2A1 included  one electric hammer (with a
rental cost of $40 per day [Wirtz Rentals Co.  1999]), an
SDS Max self-locking adapter for attaching the electric
hammer to the top extension rod ($90 [Tetra Tech 1999a]),
                                                    74

-------
Table 8-2. Split Core Sampler Cost Summary
 Item
Quantity
    Consumable supplies

    The total dollar amount is rounded to the nearest $10.
Unit Cost ($)
Total Cost ($)
S1A1 (River Mouth) Costs
Sampler
Split Core Sampler kit
6-inch-long core tube
Core tube coupling
AMS rubber-coated cross handle
18-inch-long, plastic core liners3
3-foot-long, stainless-steel extension rod
4-foot-long, stainless-steel extension rods
Labor
I DW disposal
Support equipment
Electric hammer
Electric hammer adapter
Sample extrusion rod
Total"

1 unit
1 unit
1 unit
1 unit
9 units
1 unit
2 units
1 hour
1 55-gallon drum

1 unit for 1 day
1 units
1 unit


524
211
72.50
31
6
67
69
34
182

40
90
25


524
211
72.50
31
54
67
138
34
182

40
90
25
$1,470
S2A1 (Lake) Costs
Sampler
Split Core Sampler kit
6-inch-long core tube
12-inch-long core tubes
Core tube coupling
AMS rubber-coated cross handle
6-inch-long, plastic core liners3
36-inch-long, plastic core liners3
Liner end caps3
3-foot-long, stainless-steel extension rods
4-foot-long, stainless-steel extension rods
4-foot long, carbon-steel extension rods
Labor
I DW disposal
Support equipment
Electric hammer
Electric hammer adapter
Generator
Tripod-mounted winch
Total"
Notes:
AMS = Art's Manufacturing & Supply, Inc.
IDW = Investigation-derived waste

1 unit
1 unit
2 units
2 units
1 unit
36 units
16 units
52 pairs
1 unit
4 units
2 units
1 1 hours
1 55-gallon drum

1 unit for 2 days
1 unit
1 unit for 2 days
1 unit for 3 days





524
211
229
72.50
31
2
9
0.21
67
69
48.50
34
182

40
90
20
28





524
211
458
145
31
72
144
11
67
276
97
374
182

80
90
40
84
$2,890



a generator (with a rental cost of $20 per day), one AMS
tripod-mounted winch (with a rental cost of $28 per day
[TetraTech 1999a]), two slip wrenches, one carbon-steel
bristled brush, and one stainless-steel bristled brush. The
        costs of the slip wrenches and brushes are not included in
        this analysis because a field sampling team would already
        have such tools as part of its field sampling gear. The total
        cost for support equipment for S2A1 is $294.
                                                      75

-------
8.2.5  Summary of Split Core Sampler Costs

In summary, for the Split  Core Sampler, the  costs to
collect the number of samples listed in Table 4-3 for the
top two depth intervals in S1A1  and both depth intervals
in S2A1 were estimated to be $1,470 and $2,890 for S1A1
and S2A1, respectively.  This economic  analysis shows
that most of the total cost (about 70 to 75 percent) was
associated with the purchase of samplers.  The remaining
total cost was associated with labor, IDW disposal, and
support equipment costs.

8.3    Hand Corer Costs

This section  presents information on sampler, labor, IDW
disposal, and support equipment costs for the Hand Corer
as well as a  summary of these costs. Table 8-3  presents
these costs.

8.3.1  Sampler Cost

The Hand Corer purchase cost was approximately $329.
During the  demonstration, 41 core  tube liners  and
five 5-foot-long, galvanized-steel extension rods were used
in S2A1.  Liners were purchased in four packages of 12 at
a cost of $192 per package.  The purchase  cost of
        each extension rod was $93.  The total sampler cost was
        estimated to be $1,562.

        8.3.2  Labor Cost

        In S2A1, the time required  for sampler setup,  sample
        collection,  sampler disassembly,  and sampler  decon-
        tamination totaled 594 minutes or about 10 hours for each
        of two technicians.   In addition, to facilitate  sample
        extrusion, 41 core tube liners were cut at a local machine
        shop at a cost of $3 each, for a total cost of $ 123.  In this
        area, 15 investigative samples each were collected in the
        0- to 4- and 10- to 30-inch bss depth intervals using the
        Hand Corer. Table 4-3 presents additional information on
        the total number of samples collected.  The labor cost for
        sampling was  estimated to be  $803.    When  field
        technicians work more than 8 hours in one day, overtime
        costs may be incurred. This estimate, however, includes
        no overtime costs.

        8.3.3  IDW Disposal Cost

        Sampling in S2A1 generated IDW consisting  of 12 L of
        unused sediment and 41 core tube liners. The total volume
        of IDW generated was less than 55 gallons.  The cost for
        disposal of one 5 5-gallon drum of nonhazardous waste is
        $182.
Table 8-3.  Hand Corer Cost Summary for S2A1 (Lake)
 Item
Quantity
Unit Cost ($)
Total Cost ($)
Sampler
Hand Corer
Core tube liners3
Galvanized-steel extension rods
Labor
Technicians
Cut liners
IDW disposal
Support equipment
Tripod-mounted winch
Total"
1 unit
4 dozen
5 units

20 hours
41 units
1 55-gallon drum
1 unit for 3 days

329
192
93

34
3
182
40

329
768
465

680
123
182
120
$2,670
Notes:

IDW   =   Investigation-derived waste

a    Consumable supplies

b    The total dollar amount is rounded to the nearest $10.
                                                    76

-------
8.3.4  Support Equipment Cost

Support equipment used during Hand  Corer sampling
included an a tripod-mounted winch with telescoping legs.
The tripod-mounted winch was rented for 3 days at a daily
rate of $40 (Hazco 1999). The total cost of the support
equipment was estimated to be $120.

8.3.5  Summary of Hand Corer Costs

In summary, for the Hand Corer, the costs to collect the
number of samples listed in Table 4-3 were estimated to be
$2,670. This economic analysis shows that most of the
total  cost   was  associated  with  sampler  purchase
(59 percent) and labor (30 percent).   The remaining
11 percent was associated with IDW disposal and support
equipment costs.

8.4    Vibrocorer Costs

This section  presents information on sampler, labor, IDW
disposal, and support equipment costs for the Vibrocorer
as well as a  summary of these costs. Table 8-4 presents
these costs.

8.4.1  Sampler Cost

The Vibrocorer purchase cost was approximately $24,500.
Also, 4-inch-diameter,  10-foot-long,  plastic core tubes
          were  required  for  sample  collection.   During  the
          demonstration, five tubes were used, and the purchase cost
          of each tube was $25.  The total  sampler cost was
          estimated  to be  $24,625.   Because  the Vibrocorer's
          purchase cost is relatively high and because the Vibrocorer
          is  not available for rental, the Vibrocorer  should  be
          considered for sediment sampling only when the sampling
          program is expected to be of long duration, which will
          allow recovery of the sampler purchase cost.

          8.4.2  Labor Cost

          The time required for sampler setup, sample  collection,
          sampler disassembly, and sampler decontamination totaled
          142  minutes or  about  2.5  hours   for  each of two
          technicians.  In addition, one technician spent about 1 hour
          preparing core catchers at an off-site location. In  S1A1,
          five investigative samples each were collected in the 0- to
          4- and  6- to 12-inch bss  depth intervals  using the
          Vibrocorer.  Table 4-3 presents additional information on
          the total number of samples  collected. The labor cost for
          sampling was estimated to be $204.

          8.4.3  IDW Disposal Cost

          Sampling in S1A1 generated IDW consisting of 45 L of
          unused sediment  and five plastic core tubes.  The total
          volume of IDW generated was less than 55 gallons. The
          cost for disposal of one 5 5-gallon drum of nonhazardous
          waste is $182.
Table 8-4. Vibrocorer Cost Summary for S1A1 (River Mouth)
 Item
Quantity
Unit Cost ($)
Total Cost ($)
Sampler
Vibrocorer
Core tubes3
Labor
IDW disposal
Support equipment
A-frame and winches
Drill
Saw
Total"
1 unit
5 units
6 hours
1 55-gallon drum
1 unit
1 unit for 1 day
1 unit for 1 day

24,500
25
34
182
3,500
12
15

24,500
125
204
182
3,500
12
15
$28,540
Notes:

IDW   =   Investigation-derived waste

a    Consumable supplies

b    The total dollar amount is rounded to the nearest $10.
                                                    77

-------
8.4.4  Support Equipment Cost

Support equipment costs for the Vibrocorer included a
purchase price of $3,500 for an A-frame and two electric
(12-volt direct current) winches with steel cable for raising
and lowering the sampler; a 1-day rental cost of $12 for
one portable drill (Cincy Tool Rental 1999); and a 1-day
rental cost of $ 15 for one portable circular saw (Falls Tool
Rental 1999). Two 3/4-inch socket wrenches, each costing
less than $ 10, were also used. The total cost of the support
equipment was estimated to be $3,527.

8.4.5  Summary of Vibrocorer Costs

In summary, for the Vibrocorer, the costs to collect the
number of samples listed in Table 4-3 for the top two
depth intervals in S1A1 were estimated to be $28,540.
This economic analysis shows that most of the total cost
was associated with sampler purchase (86 percent). The
remaining 14 percent was associated with labor, IDW
disposal, and support equipment costs.

8.5    Comparison of Economic Analysis Results

The costs for each sampler used in S1A1 and S2A1 are
summarized in Table 8-5. For S1A1, the total costs forthe
Split Core Sampler were about 95 percent less than the
costs for the reference sampler,  the Vibrocorer.  This
difference was due  mainly to the costs  involved in
purchasing the samplers.   However,  costs  that were
dependent on  the number of samples collected or the
amount of time required (which is itself dependent on the
number of samples collected), such as labor and support
equipment costs, were also higher for the Vibrocorer.

For S2A1, the total costs forthe Split Core Sampler were
8 percent higher than the costs for the reference sampler,
the Hand  Corer.  The  sampler cost for the  Split Core
Sampler was about 30  percent higher than that for the
Hand Corer. The labor cost for the Hand Corer was about
twice that for the Split Core Sampler, primarily because
two technicians were used to operate the Hand Corer while
only one technician was used to operate the  Split Core
Sampler.  Finally, the  support equipment cost for the Split
Core Sampler was about two times higher than that for the
Hand Corer, primarily because the Split Core Sampler
required  an electric  generator, electric hammer,  and
electric hammer adapter.
Table 8-5. Comparison of Costs for Split Core Sampler and Reference Samplers

                              S1A1 (River Mouth)
                        S2A1 (Lake)
Item
Sampler
Labor
IDW disposal
Support equipment
Total"
Split Core Sampler
$1 ,097.50
34
182
155
$1,470
Vibrocorer
$24,625
204
182
3,527
$28,540
Split Core Sampler
$2,036
374
182
294
$2,890
Hand Corer
$1 ,562
803
182
120
$2,670
Notes:

IDW   =   Investigation-derived waste

a    Each total dollar amount is rounded to the nearest $10.
                                                   78

-------
                                              Chapter 9
                               Summary of Demonstration Results
As  discussed  throughout  this ITVR, the Split  Core
Sampler was demonstrated at two sites in EPA Regions 1
and 5.  At the Region 1 site, the Split Core Sampler was
demonstrated in two areas: a lake (S2A1) and a wetland
(S2A2). At the Region 5 site, the Split Core Sampler was
also demonstrated in two areas: a river mouth (S1A1) and    PI.
a freshwater bay (S1A2).  Collectively, the four  areas
provided a variety of sampling conditions such as different
water depths, sediment types,  sediment contaminant
characteristics, and  sediment  thicknesses necessary to
properly  evaluate  the  sampler.    Based   on  the
predemonstration  investigation  results,  demonstration
objectives,  and site  support facilities available, (1) the    PI.
Hand Corer was selected as the  reference sampler for
S1A2,  S2A1, and  S2A2, and (2) the Vibrocorer was
selected as the reference sampler for S1A1.

This chapter compares the performance and cost results
for the Split Core Sampler with those for the  reference
samplers.   Tables 9-1 and 9-2 summarize the  demon-
stration results for the primary and secondary objectives,
respectively. As shown in these tables, both the Split Core    P1.
Sampler and the reference samplers were unable to collect
samples in the deep depth interval (4 to 11 feet bss). Key
demonstration findings are summarized below  for the
primary and secondary objectives.

9.1     Primary Objectives

Key  demonstration findings are  summarized below for    P2.
primary objectives PI through P7.

PI.   In the shallow depth interval (0 to 4 inches bss), to
     collect a specified number of samples, the  Split
     Core Sampler required 7 percent more attempts than
expected (46 actual versus 43 expected), whereas
the reference  samplers required 14 percent more
attempts than  expected  (49  actual versus  43
expected).

In the moderate depth interval (4 to 32 inches bss),
the Split Core Sampler required 38 percent more
attempts than  expected  (40  actual versus  29
expected), but  the  reference  samplers required
156 percent more attempts than expected (64 actual
versus 25 expected).

For the shallow depth interval, mean PSRs ranging
from 89 to 100 were achieved by the Split  Core
Sampler, whereas the reference samplers' mean
PSRs ranged from 85 to 100. The variation in PSRs
as measured  by their RSDs ranged from  0 to
26 percent for the Split Core Sampler, whereas the
reference  samplers'  RSDs  ranged from 0  to
33 percent.

For the moderate depth interval, mean PSRs ranging
from 37 to 100 were achieved by the Split  Core
Sampler, whereas the reference samplers' mean
PSRs ranged from 21 to 82. The RSDs for the Split
Core Sampler ranged from 0 to 51 percent, whereas
the reference  samplers' RSDs ranged from 3 to
161 percent.

For the shallow depth interval,  the Split  Core
Sampler's actual core lengths equaled the target
core length in  96 percent of the  total  sampling
attempts.  The reference  samplers' actual  core
lengths equaled the target core length in 94 percent
of the total sampling attempts.
                                                     79

-------
Table 9-1.  Summary of Results for Primary Objectives
Primary Objective
P1 Ability to consistently
collect a specified
volume of sediment
P2 Ability to consistently
collect sediment in a
specified depth interval
P3 Ability to collect
samples with
consistent
characteristics from a
homogenous layer of
sediment
P4 Ability to collect a
representative sample
from a clean sediment
layer below a
contaminated sediment
layer
P5 Ability to be adequately
decontaminated
Evaluation Criterion
Actual versus expected
number of sampling
attempts
Volume of sediment
sampled versus design
volume
Number of sampling
attempts in which
actual core length
equaled target core
length
Variability of sample
characteristics in terms
of PSD
Mean difference
between innovative
and reference sampler
arsenic concentrations
for clean layer is zero
Contaminant
concentrations in
equipment rinsate
samples are below
reporting limits
Sampling Depth Interval/
Demonstration Area3
Shallow (0 to 4 inches bss)/S1 A1 , S1 A2,
and S2A1
Moderate (4 to 32 inches bss)/S1 A1 , S1 A2,
S2A1 , and S2A2
Deep (4 to 11 feet bss)/S1A1 and S2A2
Shallow (0 to 4 inches bss)/S1 A1 , S1 A2,
and S2A1
Moderate (4 to 32 inches bss)/S1 A1 , S1 A2,
S2A1 , and S2A2
Deep (4 to 11 feet bss)/S1A1 and S2A2
Shallow (0 to 4 inches bss)/S1 A1 , S1 A2,
and S2A1
Moderate (4 to 32 inches bss)/S1 A1 , S1 A2,
S2A1 , and S2A2
Deep (4 to 11 feet bss)/S1A1 and S2A2
0 to 4 inches bss/S2A1
1 0 to 30 inches bss/S2A1
12 to 32 inches bss/S1A2
1 0 to 30 inches bss/S2A1
Objective addressed by area: one PCB-
contaminated area (S1A1) and one arsenic-
contaminated area (S2A1)
Performance Results
Split Core Sampler
46 actual attempts versus 43 expected
attempts (7% more than expected)
40 actual attempts versus 29 expected
attempts (38% more than expected)
Not applicable0
Mean PSRs: 89 to 100
RSDs of PSRs: 0 to 26%
Mean PSRs: 37 to 100
RSDs of PSRs: 0 to 51%
Not applicable0
44 of 46 attempts (96%)
14 of 36 attempts (39%)
Not applicable0
Sand: 29 to 42%
Silt: 49 to 59%
Clay: 8 to 10%
Sand: 25 to 42%
Silt: 50 to 67%
Clay: 6 to 10%
Sand: 2 to 5%
Silt: 66 to 74%
Clay: 23 to 31 %
Reference Sampler11
49 actual attempts versus 43 expected
attempts (14% more than expected)
64 actual attempts versus 25 expected
attempts (156% more than expected)
Unable to collect samples0
Mean PSRs: 85 to 100
RSDs of PSRs: 0 to 33%
Mean PSRs: 21 to 82
RSDs of PSRs: 3 to 161%
Unable to collect samples0
46 of 49 attempts (94%)
8 of 64 attempts (13%)
Unable to collect samples0
Sand: 26 to 46%
Silt: 48 to 72%
Clay: 0 to 5%
Sand: 35 to 43%
Silt: 53 to 62%
Clay: 0 to 4%
Sand: 3 to 6%
Silt: 63 to 72%
Clay: 22 to 31 %
According to the Wilcoxon signed rank test, there was a 0.9 percent probability that the
innovative and reference sampler arsenic concentrations were not different; at a
statistical significance level of 0.05, the samples collected by the Split Core Sampler
contained arsenic concentrations lower than those in the samples collected by the
reference sampler (the Hand Corer). Because of the greater opportunity for sample
compaction in the Split Core Sampler core tube, no conclusion could be drawn.
The contaminant concentrations in the equipment rinsate samples for the Split Core
Sampler and reference samplers were below the reporting limits (1 part per billion for
PCBs and 10 parts per billion for arsenic).

-------
Table 9-1. Summary of Results for Primary Objectives (Continued)
Primary Objective
P6 Time requirements for
sample collection
activities
P7 Sampling costs
Evaluation Criterion
Total time required for
sampler setup, sample
collection, sampler
disassembly, and
sampler
decontamination
Total cost, including
sampler, labor, IDW
disposal, and support
equipment costs
Sampling Depth Interval/
Demonstration Area3
Objective addressed by area: S1A1
Objective addressed by area: S1A2
Objective addressed by area: S2A1
Objective addressed by area: S2A2
Objective addressed by area: S1A1
Objective addressed by area: S2A1
Performance Results
Split Core Sampler
68 minutes
45 minutes
639 minutes
105 minutes
$1,470
$2,890
Reference Sampler11
142 minutes
53 minutes
594 minutes
165 minutes
$28,540
$2,670
Notes:

bss
IDW
PCB
PSD
PSR
Below sediment surface          RSD
Investigation-derived waste        S1A1
Polychlorinated biphenyl          S1A2
Particle size distribution          S2A1
Percent sample recovery         S2A2
Relative standard deviation
River mouth
Freshwater bay
Lake
Wetland
     Based on the PSD results, the shallow depth interval contained silty sand in S1A1, predominantly sand and silt with some clay in S1A2, and sandy silt in S2A1.  The moderate depth interval
     contained sandy silt in both S1A1 and S2A1, clayey silt in S1A2, and predominantly silt with some sand and clay in S2A2. Also, in S2A2, the (1) shallow and moderate depth intervals contained
     significant amounts of partially decomposed reeds and leaves and live vegetation and (2) deep depth interval contained peat. The sediment in the deep depth interval was not analyzed for
     PSD.

     The Hand Corer was used as the reference sampler in S1A2, S2A1, and S2A2.  The Vibrocorer was used as the reference sampler in S1A1.

     The Split Core Sampler and the Hand Corer are not designed to collect samples in the deep depth intervals in S1A1 and  S2A2. The Vibrocorer was unable to collect samples below 5 feet
     bss because of the presence of clay hardpan in S1A1.

-------
        Table 9-2.  Summary of Results for Secondary Objectives
         Secondary Objective
                                                                                                    Performance Results
         Split Core Sampler
                                                                                                                           Reference Sampler3
             Hand Corer
             Vibrocorer
         S1   Skills and training
              requirements for proper
              sampler operation
Easy to operate; requires minimal skills and
training

Can be operated by one person; a tripod-
mounted winch is recommended when more
than two extension rods are used

For more efficient recovery of samples, an
electric hammer should be used to induce
vibrations in the sampler
Easy to operate; requires minimal skills and
training

Can be operated by one person when up to
two extension rods are used; two persons
and a tripod-mounted winch are
recommended when more extension rods
are used
Relatively complicated to operate; requires
moderate skills and training

Requires two persons and a motor-operated
winch because of the heaviness of the
sampler (about 150 pounds)
         S2  Ability to collect samples under
             a variety of site conditions
oo
to
Collected samples in a river mouth (S1A1),
freshwater bay (S1A2), lake (S2A1), and
wetland (S2A2) where water depths ranged
from 0.5 foot to 18 feet

Collected samples in shallow (0- to 4-inch
bss) and moderate (4- to 32-inch bss) depth
intervals; sampler is not designed to collect
samples in depth intervals below 4 feet bss

Collected samples from a variety of
sampling platforms: wooden planks fastened
to ladders, a Jon boat, a pontoon boat, and
the  EPA GLNPO's Mudpuppy
Collected samples in a freshwater bay
(S1A2), lake (S2A1), and wetland (S2A2)
where water depths ranged from 0.5 foot to
18 feet

Collected samples in shallow (0- to 4-inch
bss) and moderate (4- to 32-inch bss) depth
intervals; sampler is not designed to collect
samples in depth intervals below 3 feet bss

Collected samples from a variety of
sampling platforms: wooden planks fastened
to ladders, a Jon boat, and a pontoon boat

Material caught between core tube and
flutter valve could cause partial or complete
loss of sample
Collected samples in a river mouth (S1A1)
where water depths ranged from 5 to 6 feet

Collected samples in shallow (0- to 4-inch
bss) and moderate (4- to 32-inch bss) depth
intervals but was unable to penetrate clay
hardpan in order to collect samples in 4- to
6-foot bss depth interval

Collected samples from the EPA GLNPO's
Mudpuppy
         S3  Ability to collect an undisturbed
             sample
Collected relatively compressed core
samples of both unconsolidated and
consolidated sediments from the sediment
surface downward, based on visual
observations

Sediment stratification preserved for both
unconsolidated and consolidated sediments
Collected relatively compressed core
samples of both unconsolidated and
consolidated sediments from the sediment
surface downward, based on visual
observations

Sediment stratification preserved for both
unconsolidated and consolidated sediments
Collected relatively compressed core
samples of both unconsolidated and
consolidated sediments from the sediment
surface downward, based on visual
observations

Sediment stratification preserved for both
unconsolidated and consolidated sediments

-------
        Table 9-2.  Summary of Results for Secondary Objectives (Continued)
         Secondary Objective
                                                                                                     Performance Results
                                               Split Core Sampler
                                                                                                                            Reference Sampler3
                                                          Hand Corer
                                                          Vibrocorer
         S3  Ability to collect an undisturbed
             sample (continued)
                                      Samples collected in and below moderate
                                      depth interval may be of questionable
                                      representativeness because of core
                                      shortening and core compression; sampler
                                      is not designed to collect samples in depth
                                      intervals below 4 feet bss
                                             Samples collected in and below moderate
                                             depth interval may be of questionable
                                             representativeness because of core
                                             shortening and core compression; sampler
                                             is not designed to collect samples in depth
                                             intervals below 3 feet bss
                                             Samples collected in moderate and deep
                                             depth intervals may be of questionable
                                             representativeness because of core
                                             shortening and core compression
         S4   Durability based on materials
              of construction and
              engineering design
                                      Sampler is sturdy; its primary components
                                      are made of stainless steel

                                      Both stainless-steel and carbon-steel
                                      extension rods are rigid; significant bowing
                                      was observed when rods were coupled to a
                                      total  length of 27 feet, but the rods were not
                                      damaged
                                             Sampler is sturdy; most of its primary
                                             components are made of stainless steel

                                             Galvanized extension rods are rigid; minimal
                                             bending or bowing was observed when rods
                                             were coupled to a total length of 25 feet

                                             During sample collection in S2A1, where
                                             water depth was about 18 feet, the pontoon
                                             boat drifted; the resulting stress damaged
                                             one extension rod at the threads
                                             Sampler is sturdy; its primary component,
                                             the vibrohead, is made of anodized
                                             aluminum and has a life expectancy of
                                             10,000 operating hours

                                             During sample collection in S1A1, the power
                                             supply for the sampler malfunctioned; the
                                             source of the malfunction was identified and
                                             corrected  by on-site personnel
oo
OJ
S5  Availability of sampler and
    spare parts
Sampler and its support equipment are not
expected to be available in local retail stores
but may be obtained from technology
developer by overnight courier in 2 days or
less, depending on the location of the
sampling site
Primary components of sampler are not
expected to be available in local retail stores
but may be obtained from technology
developer by overnight courier in 2 days or
less, depending on the location of the
sampling site ;  extension rods are expected
to be available  in local retail stores
Primary sampler component, the vibrohead,
is not available in local retail stores;
because the vibrohead is pressure-sealed, if
it malfunctions, it should be packaged and
shipped to the developer for servicing
        Notes:

        bss      =    Below sediment surface
        EPA     =    U.S. Environmental Protection Agency
        GLNPO  =    Great  Lakes National Program Office
             The Hand Corer was used as the reference sampler in S1A2, S2A1, and S2A2.  The Vibrocorer was used as the reference sampler in S1A1.

-------
P2.  For the moderate  depth interval, the Split Core
     Sampler's actual core  lengths equaled the target
     core length in 39  percent of the total  sampling
     attempts.   The  reference samplers' actual  core
     lengths equaled the target core length in 13 percent
     of the total sampling attempts.

P3.  Based on the PSD results, both the Split Core
     Sampler and reference  samplers collected samples
     with  consistent physical  characteristics from  a
     homogenous layer  of sediment.

P4.  In  sampling  a  clean  sediment  layer  below  a
     contaminated sediment layer, the Split Core Sampler
     and reference  sampler  (the Hand Corer)  collected
     samples whose contaminant concentrations were
     statistically  different  from   each   other  at  a
     significance level of 0.05. Arsenic concentrations in
     the samples collected by the Split Core Sampler
     were less than those in the samples collected by the
     Hand  Corer.  However, because of the greater
     opportunity for sample compaction in the Split Core
     Sampler core tube, no  conclusion could be drawn
     regarding the Split  Core Sampler's ability to collect
     representative samples  from a clean  layer below a
     contaminated layer. Explanation of this result was
     beyond the scope of the demonstration.

P5.  Both the Split Core Sampler and reference samplers
     demonstrated   the ability   to  be  adequately
     decontaminated  after  sampling  in  areas
     contaminated with  either PCBs or arsenic.

P6.  Compared to the reference samplers,  the Split Core
     Sampler reduced sampling time by 15 to 52 percent
     in three of the four areas sampled but increased the
     sampling time by 10 percent in the remaining area.

P7.  Sampling costs were estimated for two of the four
     areas sampled. In  one area, the sampling costs for
     the Split Core Sampler were 95 percent less than
     those for the reference sampler (the Vibrocorer); in
     the other area, the sampling costs for  the Split Core
     Sampler were 8 percent greater than those for the
     reference sampler (the Hand Corer).
9.2    Secondary Objectives

Key demonstration findings are summarized below for
secondary objectives SI through S5.

S1.  The Split Core Sampler, like the Hand Corer, is easy
     to operate and requires minimal skills and training.
     However, operation of the Vibrocorer is relatively
     complicated  and  requires  moderate  skills and
     training.

S1.  The Split Core Sampler was operated by one person,
     whereas the Hand Corer was operated by one or two
     persons and the Vibrocorer was operated by two
     persons. When more than two extension rods were
     required, both the Split Core  Sampler and Hand
     Corer were operated using a tripod-mounted winch.
     Vibrocorer operation required a motor-operated
     winch because of the heaviness of the sampler.

S1.  For more efficient recovery of samples, an electric
     hammer should be used to induce vibrations in the
     Split Core Sampler; a 110-volt power supply is
     required to operate the  electric hammer.   The
     Vibrocorer requires athree-phase, 230- or 440-volt,
     50-  to 60-hertz power supply, which  may be  a
     sampler limitation if the power supply fails. The
     Hand Corer does not require any power supply.

S2.  Both the Split Core Sampler and reference samplers
     collected samples  in shallow  and moderate depth
     intervals in all demonstration areas.  No sampler
     was able to collect samples in deep depth intervals
     (4 to 11 feetbss).

S3.  Based on visual observations, both the  Split Core
     Sampler and reference samplers collected partially
     compressed core  samples  of  consolidated and
     unconsolidated sediments from the sediment surface
     downward.  Samples collected by both the Split
     Core Sampler and  reference samplers in moderate
     and deep  depth intervals may be of questionable
     representativeness  because of core shortening and
     core compression.
                                                   84

-------
S3.  Sediment stratification was preserved for both    S5.   The Split Core Sampler and its support equipment
     consolidated and unconsolidated sediments in the          are not expected to be available in local retail stores.
     samples collected by the  Split Core Sampler and          Similarly,  the  primary components of the  Hand
     reference samplers.                                      Corer and Vibrocorer are  not expected to  be
                                                             available in local retail stores; extension rods for the
S4.  Based  on  their materials  of construction and          Hand Corer may be locally available.
     engineering designs, both the Split Core Sampler
     and reference samplers are considered to be sturdy.
                                                    85

-------
                                            Chapter 10
                                            References
Aaby,  B.,  and  G.  Digerfeldt.    1986.    "Sampling
  Techniques for Lakes  and Bogs."   Chapter  8 in
  Handbook  of Holocene  Palaeoecology   and
  Palaeohydrology. B.E.Berglund, Editor. John Wiley &
  Sons, New York.

Analytical Software.   1996.   Statistix®  for Windows.
  Version 2.0. Tallahassee, Florida.

ASTM.   1998. "ASTM Standards in Building Codes."
  Volumes. C 962-D 2940.

Belokopytov, I.E.,  and  V.V. Beresnevich.    1955.
  "Giktorf s Peat Borers."  Torfyanaya Promyflilennost.
  Number 8. Pages 9 and 10.

Blomqvist,  S.  1991.  "Quantitative Sampling of Soft-
  Bottom Sediments: Problems and Solutions." Marine
  Ecology Progress  Series.  Volume 72.   Pages  295
  through 304.

Cincy Tool Rental.  1999.  "Electric Tools."  On-Line
  Address:  http://www.cincytool.com/cgi-bin/
  hackmail4.cgi?electric.  Accessed on August 4.

Downing, J.A. 1984. "Sampling the Benthos of Standing
  Waters."   Chapter 4 in Manual on Methods for the
  Assessment oj'Secondary Productivity in Fresh Waters.
  J.A. Downing  and F.H  Rigler, Editors.  Blackwell
  Scientific Publications.  Oxford, England.

Environment Canada.  1994.  "Guidance Document on
  Collection and  Preparation   of Sediment  for
  Physicochemical   Characterization  and  Biological
  Testing."  Environmental Protection Series. Report EPS
  l/RM/29.  December.
EPA.  1996. "Test Methods for Evaluating Solid Waste."
  Volumes  1A through 1C.   SW-846.   Third Edition.
  Update III. OSWER Washington, DC. December.

EPA.  1999. "Sediment Sampling Technologies Demon-
  stration Plan." ORD. Washington, DC. April.

Faegri, K., and J. Iversen.  1989.  Textbook of Pollen
  Analysis.  Knut Faegri, Peter Emil Kaland, and Knut
  Krzywinski,  Editors.   Fourth  Edition.   Pages  60
  through 63.

Falls Tool Rental.  1999.  "Power Tool  Rental Rates."
  On-Line  Address:  http://www.fallstoolrental.com/
  powertools.html. Accessed on August 4.

Gilbert, R. 1987. Statistical Methods for Environmental
  Pollution  Monitoring.    Van  Nostrand  Reinhold
  Company, Inc.  New York.

Hazco.  1999.  Equipment Management Program Price
  List.  Dayton, Ohio.

Jowsey, P.C. 1966.  "An Improved Peat Sampler." New
  Phytologist. Volume 65. Pages 245 through 248.

Means.  1999. Environmental Remediation Cost Data—
  Unit Price. Kingston, Massachusetts.

Tetra Tech. 1999a.  Record of Telephone Conversation
  Regarding Sampler Costs. Between Kumar Topudurti,
  Environmental  Engineer, and Jim Egbert, Purchasing
  Manager, AMS. August.
                                                  86

-------
Tetra Tech.  1999b.  "Sediment Sampling Technologies    Wirtz Rentals Co.  1999. "Electric Hammers and Drills."
  Data Evaluation Report."  Prepared for ORD, EPA.      On-Line   Address:  http://www.wirtzrentals.com/
  October.                                              hammers. Accessed on August 4.
                                                  87

-------
                                             Appendix A
                                    Developer's Claims for the
                      AMS Split Core Sampler for Submerged Sediments
The product used in the SITE demonstration described in
this report was a modified AMS Split Core Sampler for
Submerged Sediments. The unmodified sampler has been
used successfully to sample surface and subsurface soils
for over 10 years at environmental sites throughout the
Untied States and around the world.  Prior to the SITE
demonstration,  AMS  supplied  about five  Split  Core
Samplers modified to incorporate a ball check valve in the
sampler top cap.

The Split Core Sampler incorporating a ball check valve in
the sampler top cap allows air and water to escape from
the sampler as it is lowered through the water column and
pushed  into  submerged sediment.   During  sampler
retrieval, the valve prevents backwash and subsequent loss
of sample.

The modification to incorporate a catcher in the sampler
tip has been found to assist in preventing loss of sample
when dry-flowing  soil is  sampled.  During the SITE
demonstration, the catcher was too stiff to be effective for
the submerged sediment that was sampled.   AMS will
continue its search for a catcher made from a softer plastic
material.

The sampler's use  of multiple body sections joined by
couplings provides versatility in that the sampler can be
deployed in lengths from 6 to about 48 inches for practical
applications.  Use of liners within the sampler allows the
collected sample to be capped in the field and subsampled
or composited later in a sample  handling area.   This
feature represents a significant advantage when multiple
samples are being collected from a boat or platform with
limited space.

The  stainless-steel  construction  of  the  sampler  and
deployment accessories allows decontamination using all
available methods. The ability to completely disassemble
the sampler  also supports an efficient decontamination
process.

The rugged design of the sampler provides the durability
needed for unknown  field  situations.  There were no
equipment failures during the demonstration.

Updates  or  improvements to the Split Core Sampler
(Section A. 1), prior deployment of the Split Core Sampler
(Section  A.2), and  developer comments  on the SITE
demonstration (Section A.3) are presented below.

A.I    Updates or Improvements to the Split
       Core Sampler

In S2A1, it was necessary  on  several  occasions to
disassemble  the ball check valve in order to clear it of
obstructions. The ball check valve is  being modified to
allow increased clearance around the ball. Consideration
  Appendix A was written solely by AMS. The statements presented in this appendix represent the developer's point of view and summarize
  the claims made by the developer regarding the AMS Split Core Sampler for Submerged Sediments. Publication of this material does not
  represent the EPA's approval or endorsement of the statements made in this appendix; performance assessment and economic analysis results
  for the Split Core Sampler are discussed in the body of this ITVR.

-------
is also being given to use of a larger ball with a larger vent
hole.

A.2   Prior Deployment of the Split Core
       Sampler

AMS made a modified Split Core Sampler with a ball
check valve in the top cap for the Texas A&M Research
Department at College Station about 3 years ago; the
sampler was used successfully in wetland studies.  Since
then, at least four other companies have successfully used
a vented sampler for wetland and underwater sampling.

A.3   Developer Comments on the SITE
       Demonstration

First, AMS compliments the EPA SITE Program staff and
Tetra Tech team on conducting an efficient demonstration.
It was a real pleasure to work with them.

The problems experienced  centered on  determining
exactly where the sediment surface was located in relation
to the water surface.  For the Split Core Sampler, AMS
determined  that  location to  be where the  sampling
technician  was  just  able  to feel resistance.   Upon
reflection, this point may not have represented the true
location of the sediment surface, particularly where the
surface was made up of soft, gelatinous materials that may
or may not be considered "sediment."

After several unsuccessful trial attempts to collect samples
at the first sampling location in S2A1 using hand- or slide-
hammer-assisted driving, a decision was made to use an
electric impact hammer.  This device was used to provide
vibrations primarily  to  the  sampler  at  all  sampling
locations where  110-volt power was available.  Its use
resulted in more  efficient recovery of samples.

The Split Core Sampler had to be overpushed in order to
recover the required sample  volumes at many sampling
locations.   This need may  have been avoided had a
sampler with a larger inside diameter been available. Such
a sampler would also have reduced the number of pushes
needed to recover the required sample volumes.
  Appendix A was written solely by AMS. The statements presented in this appendix represent the developer's point of view and summarize
  the claims made by the developer regarding the AMS Split Core Sampler for Submerged Sediments. Publication of this material does not
  represent the EPA's approval or endorsement of the statements made in this appendix; performance assessment and economic analysis results
  for the Split Core Sampler are discussed in the body of this ITVR.
                                                     89

-------
                                             Appendix B
                           Performance and Cost of the Ekman Grab
The EPA  conducted a demonstration of an innovative
sediment sampler known as the Split Core Sampler, a core
sampler designed and fabricated by AMS of American
Falls, Idaho. The demonstration was conducted under the
EPA SITE Program at two sites during the last week of
April and  first week of May 1999. The purpose of this
demonstration was to obtain reliable performance and cost
data on the Split Core Sampler in order to (1) achieve a
better understanding of the sampler's capabilities relative
to conventional sediment samplers and (2) provide an
opportunity for the sampler to enter the marketplace and
compete with conventional samplers without long delays.

In addition to the Split Core Sampler and the reference
samplers, a conventional grab sampler was included in the
demonstration because grab samplers are commonly used
to collect  surficial sediment  in  order  to  assess the
horizontal distribution of sediment characteristics. The
Ekman Grab, a commonly used sampler, was chosen for
the demonstration.  Performance and cost data collected
for the Ekman Grab are not intended to be compared to
those for the Split Core Sampler but rather are presented
in this appendix as supplemental information.

Specifically, this appendix describes the Ekman Grab that
was demonstrated (Section B.I), two demonstration sites
(Section B.2), demonstration  approach (Section B.3),
performance  of the  Ekman  Grab (Section  B.4),  and
references used to prepare this appendix (Section B.5).

B.I    Description of the Ekman Grab

The Ekman Grab is a "box" sampler whose bottom end
collects sediment as the sampler penetrates the sediment.
The sampler is designed to collect samples of soft, finely
divided sediment that is free of vegetation, stones, and
other coarse debris.  A technical description,  general
operating procedures, and advantages and limitations of
the Ekman Grab are presented below.

B. 1.1  Sampler Description

Components  of the Ekman Grab selected  for  the
demonstration included (1) two stainless-steel  scoops;
(2) two stainless-steel springs attached to four scoop
buttons;  (3)  two  stainless-steel  scoop cables; (4) a
stainless-steel  messenger;  (5)  a  3/16-inch-diameter,
braided, polyester  line or  5-foot-long,  galvanized-steel
extension handle; (6) a release mechanism consisting of a
stainless-steel strike pad and two stainless-steel pins; and
(7) two hinged, overlapping,  stainless-steel lids  (see
Figure B-l).

Optional accessories include a  10-foot-long  extension
handle and weights that can be fastened to either side of
the Ekman Grab.   Top  screens designed to  prevent
sediment from escaping from the top of the Ekman Grab
are also  available.

In water depths up to 10 feet, the Ekman Grab can be
manually deployed using the extension  handle. In water
up to 60 feet deep and with low velocity, the sampler can
be deployed  using the polyester line  and messenger.
During sampler deployment, the two lids at the top of the
sampler open to allow water to pass  through the  sampler
in order to minimize bow wave formation, thus minimizing
disturbance of the sediment. Once the sampler is deployed
to the desired sampling location, the release mechanism is
actuated using the extension handle  or the messenger on
the polyester line. Once actuated, the mechanism  releases
the scoop cables, allowing the springs to close the scoops
and   collect  a   sediment  sample.      During
                                                   90

-------
 Scoop cable
                                   Messenger
                                    3/16-inch-diameter,
                                        braided line
                                               Lids
 Scoop
                                             Spring
    Scoop button
Figure B-1.  Ekman Grab.
sampler retrieval, the lids automatically close to minimize
sample washout.

The Ekman Grab is available in many sizes; however, for
this demonstration, the standard-size  Ekman Grab was
chosen because of its ability to collect a sample volume
that met the demonstration objectives while generating
relatively little IDW.   The standard-size Ekman Grab
contains a 6-inch-long, 6-inch-wide, and 6-inch-high
sample chamber with a volume of 3,460 mL.  The area
below the chamber created by the two scoops when closed
constitutes  an  additional 630 mL.  The fully assembled
Ekman Grab, not including the extension handle, weighs
about 10 Ib.
B. 1.2  General Operating Procedures

The Ekman Grab can be manually operated by one person
from a sampling platform or while wading in shallow
water.  Prior to sampler deployment, each of the two
springs must be manually  attached to the two  scoop
buttons on either side of the sampler.  Also, before the
Ekman Grab is lowered into the water, each scoop cable
must be manually hooked to one of the two  pins in order
to hold the sampler in an open position. During and after
sampler preparation for deployment, care must be used to
avoid catching  any body parts such as fingers or feet
between the scoops.

The sampler can be manually  lowered to the sediment
surface using the extension handle or polyester line.   In
either case, the speed of sampler deployment needs to be
controlled in order to avoid bow wave formation.  If the
polyester line is used, the sampler should not be allowed
to fall freely for  a significant distance.  The  sampler
should be manually lowered to the sediment surface and
then slightly raised before it is released; this procedure
allows the weight of the sampler to  control  sediment
penetration.

Once  the  sampler  penetrates the sediment, the release
mechanism is actuated using the extension handle or by
placing the messenger on the polyester line and allowing
it to slide down the line to the strike pad. When the strike
pad is depressed, the pins are lowered, the scoop cables
are released, and the springs close the scoops to  collect a
sediment sample.  After the scoops are fully closed, the
Ekman Grab should be raised slowly from the sediment
and then raised steadily to the water surface.

There are several ways to process grab samples collected
using the Ekman Grab. Upon removal of the sampler from
the water, the grab sample may be discharged into a bucket
or bowl. Another way of processing the sample is to keep
the scoops closed and open the lids on top of the  sampler;
then small-diameter tubes can be inserted  into the  top
portion of the sampler to collect subsamples.

B.1.3  Advantages and Limitations

An advantage of the Ekman Grab is that  it is easy to
operate, requiring minimal skills and training.  Sampler
assembly and collection procedures can be learned in the
                                                   91

-------
field with a few practice attempts.  In addition, a written
SOP typically  accompanies the  sampler  when it  is
procured. The sampler can be operated by one person in
shallow (wading) and deep water depths because of its
lightness (10 Ib, not including the weight of the extension
handle). Sampler operation is simple because the sampler
does not require complete disassembly and reassembly
after each sampling attempt.  Only the scoops have to be
opened in order to retrieve the  sediment sample.  The
sampler also requires no support  equipment.

Another advantage of the Ekman Grab  is that  during
sampler deployment, the two lids  at the top of the sampler
open to allow water to pass through the sampler and  to
minimize the  bow  wave  formation, thus minimizing
disturbance of the sediment.  The sampler's scoops are
designed to  overlap  in the closed position in order  to
minimize sample loss during  sampler  retrieval.   In
addition, the release mechanism and pivoting scoops are
designed to minimize sediment disturbance when a sample
is collected.

A limitation of the Ekman Grab is that  because of its
lightness, the  sampler may  not be  able to penetrate
consolidated sediment if the sampler is deployed by
gravity penetration with a polyester line.  In addition,
small stones or vegetation may become caught between the
scoops, causing the scoops to remain in the open position
during  sampler retrieval, resulting in partial or complete
loss  of the  sample.   Also,  during  and after sampler
preparation for deployment, care must be used to avoid
catching any body parts such as fingers or feet between the
scoops.

B.2     Description of the Demonstration Sites

The Ekman Grab was demonstrated at two sites in EPA
Regions 1 and 5.  At the Region 1  site, Ekman Grab
sampling was conducted in one sampling area (S2A1) that
represented lake conditions and had a water depth of about
18 feet. At the Region 5 site, Ekman Grab sampling was
conducted in two areas. One area (S1A1) was in a river
mouth  and had a water depth of about 5 to 6 feet.  The
other area (S1A2) was in a freshwater bay along a river
and had a water depth of about 2  feet.

Additional information on demonstration site and area
characteristics   and  the sampling  platforms used  is
provided in Chapter 3 of the ITVR.
B.3    Demonstration Approach

This section presents the demonstration objectives, design,
and field sampling and measurement procedures, for the
Ekman Grab.

B. 3.1  Demonstration Objectives

The  demonstration  had both  primary  and secondary
objectives.   Primary  objectives were  critical to  the
technology evaluation and were intended to produce
quantitative  results regarding technology performance.
Secondary obj ectives provided information that was useful
but did  not necessarily produce  quantitative results
regarding technology performance.

As  stated in Section 4.1 of the  ITVR, the primary
objectives for the demonstration were as follows:

PI.  Evaluate whether  the  sampler can  consistently
     collect a specified volume of sediment

P2.  Determine  whether the sampler can consistently
     collect samples in a specified depth interval

P3.  Assess the  sampler's  ability to  collect multiple
     samples  with consistent  physical or chemical
     characteristics, or both, from a homogenous layer of
     sediment

P4.  Evaluate  whether  the sampler  can  collect a
     representative sample from a "clean" sediment layer
     that is  below a contaminated sediment layer

P5.  Assess the  sampler's  ability to   be  adequately
     decontaminated between sampling areas

P6.  Measure the  time  required  for  each  activity
     associated with sample collection  (sampler setup,
     sample collection,  sampler  disassembly,  and
     sampler decontamination)

P7.  Estimate costs associated with sample collection
     activities (sampler, labor, IDW disposal, and support
     equipment costs)

Primary objective P4 was not addressed  for the Ekman
Grab because this sampler is not designed in such a way
                                                   92

-------
that  it can  be evaluated under P4.   The  secondary
objectives for the demonstration were as follows:

SI.   Document the  skills  and training  required  to
      properly operate the sampler

S2.   Evaluate the sampler's ability to collect samples
      under a variety of site conditions

S3.   Assess   the   sampler's  ability   to  collect  an
      undisturbed sample

S4.   Evaluate  the  sampler's  durability based on  its
      materials of construction and engineering design

S5.   Document the availability of the sampler and spare
      parts

B.3.2  Demonstration Design

Samples were collected using the Ekman Grab to obtain
supplemental performance and cost data.   Table B-l
summarizes the demonstration design for collecting grab
samples.   Sediment samples were collected using the
Ekman Grab only in the 0- to 4-inch bss  depth intervals in
                                             S1A1, S1A2, and S2A1. The Ekman Grab is designed to
                                             collect surficial sediment samples in areas that are largely
                                             free  of  vegetation.   According to the  findings of the
                                             predemonstration investigation, most of the  surficial
                                             material in S2A2 was composed of decomposed leaves and
                                             wood chips. Therefore, grab samples were not collected
                                             in S2A2.   The  approach for addressing the  primary
                                             objectives using the Ekman Grab was generally the same
                                             as that for the Split Core Sampler presented in Section 4.2
                                             of the ITVR. Differences in the approach for the Ekman
                                             Grab are discussed below.

                                             •   Primary objective  PI  was generally addressed as
                                                 described for the Split Core Sampler. The volume of
                                                 sediment collected was noted. However, measurement
                                                 of core lengths was not appropriate for the Ekman
                                                 Grab and was not conducted.

                                             •   Primary objective  P2 was generally addressed as
                                                 described for the Split Core Sampler. The volume of
                                                 sediment collected  and the  approximate  sampler
                                                 penetration depth were noted. However, measurement
                                                 of core lengths was not  appropriate for the Ekman
                                                 Grab and was not conducted.
Table B-1. Ekman Grab Demonstration Design
Target Sampling
Depth Interval
Demonstration Area (inches bss)
S1A1 (river mouth) Oto4
S1 A2 (freshwater bay) 0 to 4
S2A1 (lake) 0 to 4
P1
P2
P3
P6
P1
P2
P5
P6
P7
P1
P2
P3
P5
P6
P7
Primary Objective
Volume
Depth interval
Consistent samples from a homogenous layer
Sample collection time
Volume
Depth interval
Decontamination
Sample collection time
Cost
Volume
Depth interval
Consistent samples from a homogenous layer
Decontamination
Sample collection time
Cost
Sampling Volume Required
Parameter (Matrix) per Sample
PSD and volume
(sediment)
PCBs and volume
(sediment)
PCBs (final rinsate)
Arsenic, PSD, and
volume (sediment)
Arsenic (final
rinsate)
250 ml
250 ml
1 L
250 ml
500ml
Notes:

bss   =
L
ml
PCB   =
PSD   =
Below sediment surface
Liter
Milliliter
Polychlorinated biphenyl
Particle size distribution
                                                   93

-------
•   Primary objective P3 was addressed as described for
    the Split Core Sampler except that sample collection
    was limited to the 0- to 4-inch bss depth intervals in
    SlAlandS2Al.

•   Primary objectives P5, P6, and P7 were addressed in
    the 0- to 4-inch bss depth intervals in S1A2 and S2A1
    as described for the Split Core Sampler. P6 was also
    addressed in the  0- to 4-inch  bss depth interval in
    S1A1.

Secondary objectives SI,  S2, and S3 were addressed for
the Ekman Grab in all three demonstration areas because
no  additional sampling was  required to address  them.
Secondary objectives S4 and S5 were not area-dependent;
they were  addressed  for the  Ekman Grab based on
information provided by the sampling technician as well
as  observations of sampler  performance  during  the
demonstration.    The  approach for addressing each
secondary obj ective was the same as that for the Split Core
Sampler presented in Section 4.2 of the ITVR.

B. 3.3  Field Sampling  and Measurement
       Procedures

Using the Ekman Grab, sediment samples were collected
in S1A1 for PSD analysis, in S1A2 for PCB analysis, and
in S2A1  for PSD and arsenic analyses.  The sampling
locations in each of  these  demonstration  areas  are
presented in Figure B-2. Additional information  on these
areas and the sampling platforms used is presented in
Chapter 3 of the ITVR. Table B-2 lists the target sampling
depth interval, planned numbers of investigative samples,
and analytical parameters for each demonstration  area and
provides  the rationale for  their selection.  In general, the
rationale for choosing the number of samples  to be
collected in each area was based on the objectives to be
addressed, the analyses to  be conducted to address one or
more objectives, the time required to collect samples, and
the cost of each analysis.  When five samples were to be
collected in a sampling area, samples were collected in the
four corners and center of the area; when ten samples were
to be collected  in a sampling area, the additional five
samples were collected at locations randomly distributed
throughout the area.

Many of the field measurements  made to support the
primary objectives were simple, standard measurements
and  do  not  require  additional explanation.   These
measurements included the volume of IDW generated,
number of sampling technicians, number  of  sampling
attempts per location, volume of sediment collected, time
required  for sample collection activities, sampling area
grid  size, and water velocity.   However, several  field
measurements were  made to address  demonstration-
specific requirements, and additional explanation of these
measurements is warranted to enhance understanding of
the sampler performance results presented in Section B.4.
Information  regarding  sample  preparation,   sampler
decontamination, and measurement of the time required to
conduct  sample collection  activities (sampler setup,
sample collection, sampler  disassembly, and  sampler
decontamination) is presented in Section 4.3 of the ITVR.

The depth of Ekman Grab deployment was measured after
the sampling  technician had  lowered the sampler to the
sediment surface.  Once the technician identified the
location of the sediment surface using the sampler, a mark
was made on the extension handle or polyester line with
reference to a fixed point (the boat side or floor).  For
extension handle applications, another mark was  made
higher on the extension handle indicating the depth to
which the sampler should be  pushed in order to collect a
sediment sample in the target sampling  depth interval.
The sampler was pushed to this depth, and a sample was
collected. For polyester line applications, the depth of
sampler deployment was dictated by gravity penetration.
Once the  sampling technician had lowered the sampler to
the sediment surface using the polyester line, he allowed
the sampler to penetrate the sediment by its own weight.
The depth of sampler deployment was then measured by
making another mark on the polyester line with reference
to the fixed point.

Field and laboratory QC checks for the demonstration are
discussed in  Sections  4.3   and 4.4  of the  ITVR,
respectively.  Section 4.4 of the ITVR also presents the
laboratory  sample  preparation  and  analysis  methods.
Table B-3 identifies the planned numbers of sediment and
equipment  rinsate  samples.   Acceptance  criteria  and
associated corrective actions for field  QC checks are
presented in  the demonstration plan  (EPA 1999).  A
summary discussion of whether the field and laboratory
QC procedures generated scientifically valid and legally
defensible data that met the demonstration objectives is
presented in Section B.4.3.
                                                   94

-------
                        S1A1 (river mouth)
                  Target sampling depth interval:
               0 to 4 inches below sediment surface

                     S1A2 (freshwater bay)
                  Target sampling depth interval:
               0 to 4 inches below sediment surface

                           S2A1 (lake)
                 Target sampling depth interval:
              0 to 4 inches below sediment surface
2
3
4
5
"V

1
2
3
4
5
X

*



*






A B
o



o







•


10 feet
C


O







*



*

D E





o



o
L— 	 infect -J
1
1

1
1

                                                                              -10 feet.
A
A
A
A

A

A


A

A
A
A
                                                                                                    I
                                                                                                    1
 Legend
   O    Polychlorinated biphenyls
   •    Particle size distribution
  A   Arsenic
  ^   Arsenic and particle size distribution
 	^-  Flow direction
Note: Approximate  scale:  1 inch  = 1,200 feet
Figure B-2. Sampling locations for Ekman Grab demonstration.
                                                             95

-------
Table B-2. Rationale for Sampling Approach
Demonstration Area
         Target Sampling        Number of
          Depth Interval     Investigative Samples3
           (inches bss)     (Analytical Parameter)
Matrix
Rationale
S1A1 (river mouth)         Oto4
S1A2 (freshwater bay)      0 to 4
S2A1 (lake)              0 to 4
                              5 (PSD)          Sediment   Determine whether an Ekman Grab could collect multiple
                                                        samples from a homogenous layer of sediment (primary
                                                        objective P3) with consistent characteristics
                              5 (PCBs)         Sediment   Determine whether an Ekman Grab could be adequately
                                                        decontaminated (primary objective P5)
                              1 (PCBs)         Equipment  Determine whether an Ekman Grab could be adequately
                                              rinsate     decontaminated (primary objective P5)
                             10 (Arsenic)        Sediment   Determine whether an Ekman Grab could collect multiple
                              5 (PSD)                    samples from a homogenous layer of sediment (primary
                                                        objective P3) with consistent characteristics
                              1 (Arsenic)        Equipment  Determine whether an Ekman Grab could be adequately
                                              rinsate     decontaminated (primary objective P5)
Notes:

bss   =
PCB  =
PSD  =
Below sediment surface
Polychlorinated biphenyl
Particle size distribution
    The number of investigative samples varied depending on the analytical parameters and the objectives addressed in each demonstration area.
    Ten investigative samples were collected and analyzed for arsenic to address primary objective P3. However, only five investigative samples were
    collected and analyzed for PSD to address primary objective P3 because the variability associated with PSD is less than that associated with
    arsenic concentrations.
B.4    Performance of the Ekman Grab

This section describes the performance of the Ekman Grab
based  on the  primary objectives (Section B.4.1)  and
secondary objectives  (Section B.4.2); this section also
discusses the data quality of the demonstration results for
the Ekman Grab (Section B.4.3).

B.4.1   Primary  Objectives

This section discusses the performance results  for  the
Ekman Grab based on the primary objectives specified in
Section  B.3.1.   To  address these primary objectives,
samples were collected in three different areas: (1) S1A1,
a river mouth; (2) S1A2, a small, freshwater bay;  and
(3) S2A1, a lake.  Samples were collected only in the 0- to
4-inch bss depth interval in these areas because the Ekman
Grab is capable of collecting surficial sediment only. The
numbers of investigative and QC samples collected in each
area, sediment sample volumes required, and  sample
analytical parameters are presented in Table B-3.
                                                During the demonstration, because the water depth  in
                                                S1A1  and S2A1 exceeded the length of the extension
                                                handle (5  feet),  the sampling technician  deployed the
                                                Ekman Grab by gravity penetration using a polyester line.
                                                In S1A2, where  the water depth was about 2 feet, the
                                                sampler was deployed with the 5-foot-long extension
                                                handle.   The sampling technician  was  provided  an
                                                opportunity  to  practice  sample  collection  at  each
                                                demonstration area until  he  felt  confident enough  to
                                                initiate demonstration sampling.

                                                The demonstration results for the Ekman Grab under
                                                primary objectives PI and  P2 were evaluated using the
                                                Wilk-Shapiro test to determine whether the results were
                                                normally distributed. Because most of the data sets were
                                                not normally distributed, the Wilk-Shapiro test was used
                                                in an attempt to evaluate whether the results followed a
                                                lognormal distribution.  The test revealed that the results
                                                either were not lognormally distributed or could not be
                                                tested for lognormality because the results contained
                                                values  equal to zero.  For these reasons,  the Student's
                                                t-test,  a parametric test, was  not used to perform the
                                                hypothesis testing; the Wilcoxon  signed rank test, a
                                                nonparametric test, was used as  an alternative to the
                                                       96

-------
Table B-3. Ekman Grab Sample Matrix
Demonstration
Area
S1A1
(river mouth)
(SHA2
(freshwater
bay)
S2A1
(lake)
Target Sampling
Depth Interval
(inches bss)
Oto4

0 o

Oto4

Analytical
Parameter
PSD

PCBs

Arsenic
PSD

Investi-
gative
Samples
5

5

10
5
Sediment
MS/MSD
Samples3
NA

1

2
NA
Samples
Field
Triplicate
Samples"
1

2

3
1

Laboratory
Analyses
7

11

20
7
Equipment Rinsate Samples
Equipment
Rinsate
Samples
NA

1

1
NA
Field
Duplicate
Samples0
NA

1

1
NA
Laboratory
Analyses
0

2

2
0
Notes:

bss      =
MS/MSD  =
NA
PCB
PSD
Below sediment surface
Matrix spike/matrix spike duplicate
Not applicable
Polychlorinated biphenyl
Particle size distribution
    MS/MSD samples were collected for PCB and arsenic analyses and were designated in the field. MS/MSD samples were not collected for
    equipment rinsate samples because the additional volume required for the analysis may have diluted any contamination present to concentrations
    below laboratory detection limits.  Sediment MS/MSD samples did not require additional sample volume.

    Field triplicate sediment samples were collected by filling three sample containers with homogenized sediment.  A sufficient volume of sediment
    for field triplicate samples was collected as described in the approach for addressing primary objective P1 in Section 4.2 of the innovative
    technology verification report. Field triplicate samples were submitted for analysis as blind samples.

    Field duplicate equipment rinsate samples were collected by filling one additional container for PCB or arsenic analysis.  Field duplicate samples
    were submitted for analysis as blind samples.
Student's t-test. As described in Section 6.1 of the ITVR,
Statistix® was used to perform statistical evaluations of the
demonstration   results  (Analytical  Software   1996).
Appendix C provides details on the statistical methods
used for data evaluation.

B.4.1.1 Ability to Consistently Collect a Specified
        Volume of Sediment

Primary objective  PI involved evaluating the Ekman
Grab's ability to consistently collect a specified volume of
sediment.  This objective was addressed by comparing
(1) the  actual number of sampling  attempts required to
collect  a specified volume of sediment to  the expected
number of attempts (rounded to the nearest higher integer)
at each sampling location and (2) the actual volume of
sediment  collected  in each attempt  to  the calculated
sampler volume (design volume). The expected number of
attempts was determined by dividing the specified sample
volume by  the design volume.  The results  of these
comparisons are summarized below.
                                             Number of Sampling Attempts Required

                                             Tables B-4 and  B-5 present the expected and  actual
                                             number of sampling attempts for the Ekman Grab in S1A1
                                             and  S1A2 and  in S2A1,  respectively.   Initially, the
                                             Wilcoxon signed rank test was used to determine whether
                                             the difference between the expected and actual number of
                                             attempts was statistically significant. However, in two of
                                             the three areas, there were  too few  locations where the
                                             expected number of attempts differed from the  actual
                                             number to perform the test.

                                             Regarding the number of sampling attempts required to
                                             collect the specified volume, the Ekman Grab performed
                                             well  in all three areas. As shown in Tables B-4 and B-5,
                                             the  actual number  of attempts equaled  the expected
                                             number of attempts  at 15  of 20 locations.  In  S1A1,
                                             Location  1A was the  only location  where  the  actual
                                             number of attempts (four) exceeded the expected number
                                             (one) by more than  one  attempt.   In two of the  four
                                                      97

-------
Table B-4.  Comparison of Expected and Actual Number of Sampling Attempts for Ekman Grab at Site 1
Location
1A
1E
3C
5A
5E
Total
Number of Attempts in S1A1
Expected
1
1
1
1
1
5
(River Mouth)
Actual
4
1
1
1
2
9
Number of Attempts
Expected
1
1
1
1
1
5
in S1A2 (Freshwater Bay)
Actual
1
1
1
1
1
5
Table B-5.  Comparison of Expected and Actual Number of
         Sampling Attempts for Ekman Grab in S2A1 (Lake)

                     Number of Attempts in S2A1
Location
1A
1B
1E
2A
2C
2D
2E
3A
3B
3E
Total
Expected
1
1
1
1
1
1
1
1
1
1
10
Actual
1
2
2
1
1
1
2
1
1
1
13
attempts at Location 1A, only one scoop was closed after
the messenger was released, and the sediment sample was
lost through the open scoop.

Much of the sampler's overall success in terms of number
of sampling  attempts required can be attributed  to the
design volume for the Ekman Grab (about 2,900 mL for
the 0- to 4-inch bss depth interval, including the volume of
the scoops) being much greater than the specified sediment
sample volumes, which ranged from 250 to  1,000 mL.
Consequently,  a sampling attempt with low recovery
compared to the design volume could still collect the
specified volume of sediment.

Volume of Sediment Collected

The volume of sediment collected by the Ekman Grab in
each sampling  attempt was divided by the corresponding
design volume, and the resulting ratio was multiplied by
100 to estimate the PSR.  The RSD of the PSRs was
calculated to evaluate the ability of the Ekman Grab to
consistently collect a specified volume of sediment; if the
sampler were to consistently recover an identical volume
of sediment in every attempt, the RSD would equal zero.
Both  PSR and RSD results should be considered to
properly evaluate the sampler's performance because a
low RSD, which indicates that the sampler's performance
was consistent, may be based on consistently low PSRs.
Table B-6 presents the  PSR summary statistics (range,
mean, and RSD) for all three areas. Figure B-3 presents
PSRs for the Ekman Grab in S1A1, S1A2, and S2A1.

The  Ekman Grab  performed well in S1A2  but had
difficulty in S1A1 and S2A1. As shown in Table B-6, for
S1A2, PSRs ranged from 100 to 145 with a mean PSR of
127.  The RSD of the PSRs  for S1A2  (13  percent)
compares  favorably to  the  30  percent RSD guideline
discussed  in Section 6.1.1  of the ITVR.  On the other
hand, as shown in Figure B-3, 5 of 9 attempts in S1A1 and
3 of 13 attempts in  S2A1 had PSRs in the 0 to 20 range.
These low recoveries were due to the failure of one or both
scoops to  close after the messenger was released or to
incomplete sampler penetration of the specified depth
interval. Unlike S1A2, where the  sampler was deployed
with an extension handle, the sampler was deployed by
gravity penetration using a polyester line in S1A1 and
S2A1. As a result, the sampling technician had relatively
poor control of the depth of sampler penetration.  As
shown in Table B-6, RSDs  of 103 and 65 percent that
exceeded the 30 percent RSD guideline were observed for
S1A1 and S2A1, respectively, indicating that the Ekman
Grab did not consistently collect its design volume.

In summary, the Ekman Grab performed well with regard
to the number of attempts required, but did not perform
well with regard to consistently  collecting its  design
                                                   98

-------
Table B-6. Percent Sample Recovery Summary Statistics for Ekman Grab

Demonstration Area      Actual Number of Attempts          PSR Range3
                 Mean PSR
RSD (%)
S1A1 (river mouth)
S1A2 (freshwater bay)
S2A1 (lake)
9
5
13
Oto40
100 to 145
Oto71
16
127
38
103
13
65
Notes:

PSR   =  Percent sample recovery
RSD   =  Relative standard deviation

a    PSRs exceeding 100 resulted from pushing the sampler beyond the specified depth interval because of difficulty in accurately assessing the
    location of the sediment, the volumetric measurement error associated with the presence of void spaces when the sediment was transferred to
    a graduated container, or both.
volume.   The actual  number of attempts equaled the
expected number of  attempts at  15  of 20  locations.
However, for S1A1 and S2A1, low mean PSRs (16 and 38,
respectively) and high RSDs (103  and 65 percent,
respectively)  were   observed,  indicating   low  and
inconsistent recoveries.  For S1A2, a much lower RSD
(13 percent) was observed.  In addition, all the PSRs for
sampling attempts in S1A2 were 100 or greater.

B.4.1.2 Ability to Consistently Collect Sediment in a
       Specified Depth Interval

Primary  objective P2 involved  evaluating the Ekman
Grab's ability to  consistently collect  sediment in  a
specified depth interval by comparing actual and target
core lengths for each attempt. The Ekman Grab does not
collect a core, but to facilitate its comparison to the other
samplers, the actual depth interval sampled in a given
attempt was calculated based on the Ekman Grab's design
volume and the volume of sediment collected.

The Ekman Grab's box chamber is 6 inches tall and can
hold about 580 mL of sediment per inch. The scoop
chamber  has a triangular cross section; is approximately
1.5 inches tall in the middle; and can hold about 105 mL
in the bottom one-third, about 210 mL in the middle
one-third, and about 315 mL in the top one-third, which
amounts to a total volume of about 630 mL. Therefore, if
the Ekman Grab collected 2,100 mL of sediment in a given
attempt, the sampling  depth interval is 0 to 4 inches bss
because the 1.5-inch-tall scoop chamber holds 630 mL,
and  the  remaining 1,470 mL would fill  approximately
2.5 inches of the box chamber at 580 mL  per inch.
However, the height  of the  scoop chamber was not
accounted for during  demonstration sampling, and the
sampling technician tried to push the box chamber to a
depth of 4 inches bss in each attempt.  Consequently, the
target  sediment thickness  in  each area was actually
5.5 inches instead of 4 inches, which corresponds to a
sample volume of approximately 2,900 mL.

Table B-7 presents the number of attempts in which the
actual  sediment thickness equaled the target sediment
thickness, target sediment thicknesses, and mean actual
sediment thicknesses. Initially, the Wilcoxon signed rank
test was to  be  used to  determine whether differences
between the actual and target sediment thicknesses were
statistically significant. However, the Wilcoxon  signed
rank test revealed that the  test results for many  of the
primary objective P2 data sets were inconsistent with the
conclusions  reached in comparing the actual and target
sediment thicknesses for  the  reasons  described in
Section 6.1 of the ITVR.  Therefore, P2 was addressed by
evaluating (1) the number of attempts in which the actual
sediment thickness equaled the target sediment thickness
and  (2) the  difference between the target sediment
thickness and the mean actual sediment thickness.

The Ekman Grab did not perform well in any of the three
areas.   As  shown in Table B-7, sediment thicknesses
collected by the Ekman Grab equaled the target sediment
thicknesses in only 1 of 27 attempts.  Attempts in S1A1
and S2A1 generally had low recoveries (0 to 40 percent in
S1A1 and 0 to 71 percent in S2A1), which resulted in low
mean actual  sediment thicknesses of 1.0 and 2.5 inches,
respectively. In S1A2, the mean actual sediment thickness
of 6.5 inches exceeded the  target sediment thickness of
5.5 inches. Although the Ekman Grab sampled the entire
target sediment thickness in all 5 attempts, in 4  of the
5  attempts  in  S1A2,  the  actual sediment thickness
                                                    99

-------
              S1A1, 0-to 4-inch bss depth interval
                                 I Total number of attempts: 9
        Oto20    >20to40   >40to60   >60to80   >80to100    >100
                 Percent sampterecovery



             S1A2, 0-to 4-inch bss depth interval
       Total number of attempts: 5
        Oto20    >20to40   >40to60   >60to80   >80to100
                  Percent sample recovery



            S2A1, 0- to 4-inch bss depth interval
                                              >100
                                  Total number of attempts: 13
        Oto20    >20to40   >40to60   >60to80   >80to100    >100
                  Percent sample recovery

Notes:

bss  =   Below sediment surface

Percent sample recoveries exceeding 100 resulted from pushing the
sampler beyond the specified depth interval because of difficulty in
accurately assessing the location of the sediment surface, the volumetric
measurement error associated with the presence of void spaces when
the sediment was transferred to a graduated container, or both.

Figure B-3. Percent sample recoveries for Ekman Grab in S1A1
          (river mouth), S1A2  (freshwater bay), and S2A1 (lake).
exceeded the target  sediment thickness by 1 inch on
average. Because of the nature of the sampler, the portion
of the  sediment sample corresponding  to  the 5.5- to
6.5-inch bss  depth interval could not be  separated from
that corresponding to the target depth interval.  Based on
demonstration   results,   the  Ekman  Grab   did   not
demonstrate an ability to consistently collect sediment in
the specified depth interval.

B.4.1.3 Ability to Collect Multiple Samples with
        Consistent Physical or Chemical
        Characteristics, or Both, from a Homogenous
        Layer of Sediment

Primary objective  P3 involved evaluating  the Ekman
Grab's ability to collect multiple samples  with consistent
physical or  chemical characteristics,  or both, from  a
homogenous layer of  sediment.   This  objective  was
addressed by calculating the RSD values for the S1A1 and
S2A1   sample  analytical  results.    Based  on   the
predemonstration investigation results, the 0-to 4-inch bss
depth intervals  in  these areas  were determined to be
homogenous in terms of their physical characteristics, and
the S2A1 depth interval was determined to be homogenous
in terms of its chemical characteristics.

Figure B-4 presents the Ekman Grab sample  analytical
results for S1A1 and S2A1.   Although no outliers were
found in the arsenic and PSD results for the samples
collected by the Ekman Grab, the sampler was evaluated
only on its  ability  to  collect multiple  samples with
consistent physical characteristics; this approach was used
to be consistent with the evaluations of the innovative and
reference samplers discussed in  Sections 6.1.3 and 7.1.3,
respectively.  Also,  the Ekman  Grab sample arsenic
concentrations for S2A1 varied  over a wide range (53 to
240 mg/kg), indicating that the area may not be chemically
homogenous despite the lack  of statistical outliers.  The
RSDs were calculated based on the PSD analytical results
for all locations  sampled in S1A1 and S2A1.

Table B-8 presents PSD summary statistics (range, mean,
and RSD)  calculated for the  Ekman Grab samples and
field triplicates relevant to primary obj ective P3. As stated
in Section 6.1.3 of the ITVR, RSDs calculated for the PSD
results  were compared  to the laboratory acceptance
criterion of 15 percent for field triplicates.  When the RSD
for all  samples  was greater than  15 percent, it  was
compared to the measured RSD  for the
                                                      100

-------
Table B-7. Comparison of Target and Actual Sediment Thickness Data for Ekman Grab
Demonstration Area
Number of Attempts in Which Actual Sediment
    Thickness Equaled Target Sediment
        Thickness/Total Attempts
Target Sediment Thickness
       (inches)
Mean Actual Sediment
 Thickness (inches)
S1A1 (river mouth)
S1A2 (freshwater bay)
S2A1 (lake)
0/9
1/5
0/13
5.5
5.5
5.5
1.0
6.5
2.5
field  triplicates,  which  were   prepared  by   first
homogenizing  and  then subsampling the  sediment
collected in a given location and area.  An RSD for all
samples that is less than the RSD for field triplicates may
be  more  attributable  to the  laboratory's analytical
procedure  or the sample homogenization procedure
implemented in the  field, or both,  for the sediment
sampled than to the sampler's ability to collect physically
consistent samples. However, PSD parameters with means
less than 10 percent  were not evaluated in this  manner
because at low levels, the analytical method is not as
precise; as a result, it will generate high RSD values and
may not reveal whether multiple samples with consistent
physical characteristics have been collected.

As shown in Table B-8, the RSDs for silt results for both
S1A1 and S2A1 were below the  15 percent laboratory
acceptance criterion.   The RSD for the sand result for
S1A1 was also below the laboratory acceptance criterion,
but the sand result RSD for S2A1 (17 percent) was slightly
above the laboratory  acceptance criterion and above the
measured RSD for field triplicates (8 percent). Therefore,
some  of the variation  in the  sand  results may be
attributable to the Ekman Grab's ability to collect samples
with consistent physical  characteristics. However, the
variation in the sand results for S2A1 was not considered
to be significant because it was only 2 percentage points
greater than the laboratory acceptance criterion. The mean
clay results for samples collected in both S1A1 and  S2A1
were less than 10 percent and were not evaluated using the
criterion. However, the clay results fell in a tight range (0
to 2 and 0 to 5 percent in S1A1  and S2A1, respectively).

In summary, the Ekman Grab met primary  objective P3
criteria except for a 2 percentage point exceedance  in the
RSD  for sand  results for S2A1.  Therefore,  it was
concluded  that the Ekman  Grab  was  able to  collect
multiple samples with consistent physical characteristics.
                                    B.4.1.4 Ability to be Adequately Decontaminated

                                    Primary objective P5  involved evaluating the Ekman
                                    Grab's ability to be adequately decontaminated.   This
                                    objective was addressed by collecting equipment rinsate
                                    samples after sampler decontamination activities in S1A2
                                    and S2A1.  These areas were  chosen because  they
                                    contained  high  concentrations of PCBs and arsenic,
                                    respectively.   If the  Ekman Grab were  adequately
                                    decontaminated, the  analytical results for the equipment
                                    rinsate samples would be below the laboratory's reporting
                                    limits. To ensure that the water used to decontaminate the
                                    sampler was not contaminated, decontamination  water
                                    blanks were also analyzed. Contaminant concentrations in
                                    both the equipment rinsate samples and decontamination
                                    water blanks were below the laboratory reporting limits for
                                    PCBs (1 part per billion) and arsenic (10 parts per billion).
                                    Thus,  the  Ekman Grab demonstrated the ability  to  be
                                    adequately decontaminated.

                                    B.4.1.5 Time Requirements for Sample Collection
                                           Activities

                                    Primary objective P6  involved evaluating the Ekman
                                    Grab's time requirements for sample collection activities.
                                    These requirements were evaluated in S1A1, S1A2, and
                                    S2A1.  One technician conducted sampler setup, sample
                                    collection,  sampler  disassembly,  and   sampler
                                    decontamination in each of the three demonstration areas.
                                    The amounts of time required to complete these activities
                                    are shown in Table B-9.  The time measured for sample
                                    collection activities  did not include the time taken for
                                    mobilization,  demobilization,  and   maneuvering  the
                                    sampling platforms to sampling locations because these
                                    activities were not sampler-specific; they were either site-
                                    or weather-related.
                                                    101

-------
                  S1A1
Location 1A
Sand: 83%
Silt: 17%
Clay: 0%



Location 5A
Sand: 85%
Silt: 15%
Clay: 0%







Location 3C
Sand: 84%
Silt: 15%
Clay: 0%







Location 1 E
Sand: 84%
Silt: 14%
Clay: 2%



Location 5E
Sand: 86%
Silt: 13%
Clay: 1%
                  S2A1
Location 1A
Arsenic: 110 mg/kg
Sand: 39%
Silt: 55%
Clay: 5%
Location 2A
Arsenic: 240 mg/kg

Location 3A
Arsenic: 53 mg/kg
Sand: 52%
Silt: 46%
Clay: 0%
Location 1 B
Arsenic: 87 mg/kg




Location 3B
Arsenic: 110 mg/kg




Location 2C
Arsenic: 110 mg/kg
Sand: 42%
Silt: 53%
Clay: 4%






Location 2D
Arsenic: 160 mg/kg




Location 1 E
Arsenic: 200 mg/kg
Sand: 33%
Silt: 57%
Clay: 2%
Location 2E
Arsenic: 130 mg/kg

Location 3E
Arsenic: 89 mg/kg
Sand: 48%
Silt: 51%
Clay: 0%
Notes:

mg/kg   =   Milligram per kilogram

The particle size distribution results for a given sample may not total 100 percent because of rounding or because some sediment did not pass through
a U.S. Standard No. 4 sieve and was classified as gravel rather than sand, silt, or clay.
Figure B-4.  Ekman Grab sample analytical results for S1A1 (river mouth) and S2A1 (lake).
                                                              102

-------
Table B-8. Particle Size Distribution Summary Statistics for Ekman Grab
Demonstration Area
S1A1 (river mouth)

S2A1 (lake)

Parameter
Sand
Silt
Clay
Sand
Silt
Clay
Number of
Samples
5
5
5
5
5
5
Range (%)
83 to 86
13 to 17
Oto2
33 to 52
46 to 57
Oto5
Mean (%)
84
15
1
43
53
2
RSD (%)
(All Samples)
1
10
128
17
8
104
RSD (%)
(Field Triplicates)
1
8
173
8
4
35
Note:

RSD   =  Relative standard deviation
Sampler setup times for the Ekman Grab ranged from
1 minute in S1A1 and S2A1 to 4 minutes in S1A2.  The
Ekman Grab was operated using a polyester line in S1A1
and S2A1  because the water depth was greater than the
length of the extension handle available during the
demonstration.  In  S2A1, the sampler arrived with the
polyester line used to lower the sampler already attached;
therefore, the setup time for S2A1 was estimated to be
equal to the setup time for S1A1.  An extension handle
was used instead of the  polyester line in  S1A2, which
required additional sampler setup time.

Sample collection times for the Ekman Grab ranged from
8 to  40 minutes during the demonstration.   Sample
collection required 0.5 to 2 minutes per attempt in S1A1
and S1A2 but 1.5 to 3.5 minutes per attempt in S2A1.
Additional time was required in S2A1 because it was the
first  area  sampled and  because  the  water  depth  was
18 feet.

Sampler disassembly times for the Ekman Grab ranged
from  1 to 3 minutes during the demonstration.  Sampler
disassembly required 3 minutes in S1A2.  In S1A1, the
disassembly time was estimated to be equal to the sampler
setup time.  Because  a  disassembly time of less than
1 minute was  recorded in S2A1, the time for  sampler
disassembly in this area was conservatively rounded up to
1 minute.

Decontamination of the Ekman Grab required 22 minutes
in  S1A2  and   13   minutes   in  S2A1;  sampler
decontamination  time was  not evaluated  in  S1A1.
Decontamination of the extension handle used in S1A2
accounts for the difference in  decontamination times
between this area and S2A1.
A technician familiar with the Ekman Grab would be
expected to require 1 to 4 minutes for sampler setup, 0.5
to 2 minutes per attempt for sample collection, about
3 minutes for sampler disassembly, and 15 to 20 minutes
for sampler decontamination.  However, these activities
might take longer, depending on the number of extension
handles used at  a given location.  Furthermore, when
sediment sampling activities are planned, the time required
for mobilization, demobilization,  and setting up  and
positioning the sampling platform would have to be
considered in addition to the times presented above.

B.4.1.6 Costs Associated with Sample Collection
       Activities

Primary objective P7 involved estimating costs associated
with Ekman Grab sample collection activities in S1A2 and
S2A1.  Because characteristics of these two areas are
different, the sampling  activities  in these areas  were
expected to  provide  a range  of costs  involved in
conducting sediment sampling using the Ekman Grab. For
example, during the demonstration in S1A2, the average
PCB concentration was about 310 parts per billion, and the
water depth was about 2 feet. On the other hand, in S2A1,
the average arsenic concentration was 120 mg/kg, and the
water depth was about 18 feet.

The issues and assumptions discussed in Section 8.1 of the
ITVR apply to this section as well except that unused
sediment in S2A1 was assumed to be a hazardous waste.
During  the  demonstration,  the  average  arsenic
concentration in the samples collected using the Ekman
Grab was 120 mg/kg.  Arsenic-contaminated wastes  with
TCLP extract concentrations greater than 5 mg/L must be
disposed of as hazardous waste (40 CFR 261). Based on
                                                   103

-------
Table B-9. Time Required to Complete Sampling Activities for Ekman Grab
                                                             Time Required (minutes)
Activity
Sampler setup
Sample collection
Sampler disassembly
Sampler decontamination
Total
S1A1 (River Mouth)
1
10
1
Not evaluated
12
S1A2 (Freshwater Bay)
4
8
3
22
37
S2A1 (Lake)
1
40
1
13
55
the average arsenic concentration and the dilution factor
(20)  associated  with the TCLP,  the  TCLP  extract
concentration for the sediment waste generated during the
demonstration  was estimated  to  be about 6  mg/L.
Therefore, unused sediment in S2A1 was assumed to be a
hazardous waste.

This section presents information on sampler, labor, IDW
disposal, and support equipment costs for the Ekman Grab
as well as a summary of these costs. Table B-10 presents
these costs.

Sampler Cost

In S1A2,  the Ekman Grab was used with  one 5-foot
extension handle.  The Ekman Grab and extension handle
costs were $304 and $131, respectively. The total sampler
cost for S1A2 was estimated to be $435.

In S2A1, the Ekman Grab was used with one messenger
and polyester line. The Ekman Grab and messenger costs
were $304 and  $52, respectively. The polyester line cost
less than  $10  and therefore was  not included in the
estimate. The total sampler cost for S2A1 was estimated
to be $356.

Labor Cost

In S1A2, the time for sampler setup, sample collection,
sampler disassembly, and sampler decontamination totaled
37 minutes or  about 1 hour for one technician.  In this
area, five  investigative samples for PCB analysis were
collected using the Ekman Grab.  Table  B-3  presents
additional information on the total number of samples
collected.   The labor cost for sampling in S1A2 was
estimated to be $34.
In S2A1, the time for sampler setup, sample collection,
sampler disassembly, and sampler decontamination totaled
55 minutes or about 1 hour for one technician.  In this
area, ten investigative  samples for arsenic analysis and
five investigative samples for PSD analysis were collected
using the Ekman Grab.  Table B-3 presents additional
information on the total number of samples collected. The
labor cost for sampling in S2A1 was estimated to be $34.

IDW Disposal Cost

Sampling in S1A2 generated IDW consisting of 15 L of
unused sediment. The cost for disposal of one 55-gallon
drum of nonhazardous waste is $182.

Sampling in S2A1 generated IDW consisting of 7 L of
unused sediment. The cost for disposal of one 55-gallon
drum of hazardous waste is $196.

Support Equipment Cost

Support equipment used during Ekman Grab sampling in
S1A2 and S2A1 included a crescent wrench and a Phillips-
head screwdriver.  The costs of these items were not
included in the estimate because a field sampling team
would already have such tools as part of its field sampling
gear.

Summary of Ekman Grab Costs

In summary, for the Ekman Grab, the costs to collect the
numbers of samples listed in Table B-3 were estimated to
be $650 and $590 for S1A2 and S2A1, respectively.  Most
of the   costs  were associated  with the purchase of
samplers(about 67 percent  for S1A2  and 60 percent for
S2A1) and IDW disposal (about 28 percent for S1A2 and
33 percent for S2A1).
                                                  104

-------
Table B-10. Ekman Grab Cost Summary
 Item
Quantity
Unit Cost ($)
Total Cost ($)
 S1A2 (Freshwater Bay) Costs
Sampler
Ekman Grab
Extension handle
Labor
I DW disposal
Support equipment
Total"
1 unit
1 unit
1 hour
1 55-gallon drum
Not applicable
304
131
34
182
Not applicable
304
131
34
182
0
$650
S2A1 (Lake) Costs
Sampler
Ekman Grab
Messenger
Labor
I DW disposal
Support equipment
Totala
1 unit
1 unit
1 hour
1 55-gallon drum
Not applicable
304
52
34
196
Not applicable
304
52
34
196
0
$590
Notes:

IDW   =   Investigation-derived waste

a    The total dollar amount is rounded to the nearest $10.
B.4.2  Secondary Objectives

This section describes the performance results  for the
Ekman Grab based on the secondary objectives specified
in  Section  B.3.1.   The secondary  objectives  were
addressed  based  on  observations of Ekman  Grab
performance during the demonstration.

B.4.2.1 Skill and Training Requirements for Proper
       Sampler Operation

The Ekman Grab is easy to operate, requiring minimal
skills  and  training.    Sampler  assembly  and  sample
collection procedures can be learned in the field with a
few practice attempts.   In  addition,  a written  SOP
accompanies the sampler when it is procured. The sampler
can be operated by one person in shallow (wading) and
deep water depths because of its lightness (10 Ib, not
including the weight of the extension handle).  Sampler
operation is simple because the sampler does not require
complete disassembly and reassembly after each sampling
attempt.  Only the scoops or lids have to be opened in
order to retrieve the sediment sample. The sampler
                  requires no support equipment unless a sampling platform
                  is needed.

                  During the demonstration, minimal strength and stamina
                  were required to deploy the sampler into and retrieve it
                  from the 0- to 4-inch bss depth interval in S1A1, S1A2,
                  and S2A1.  Previous sediment sampling experience is
                  beneficial  in  selecting  the most appropriate  optional
                  accessories (such as the extension handle length or weight
                  attachments)  for  a  given Ekman  Grab application.
                  Previous sediment sampling experience is also beneficial
                  for accurately assessing the  location of the sediment
                  surface using the sampler, as is  the case using other
                  samplers.

                  B.4.2.2 Ability to Collect Samples Under a Variety of
                         Site Conditions

                  The Ekman  Grab demonstrated the ability to collect
                  sediment samples under all conditions encountered during
                  the demonstration, which included a variety of sampling
                  platforms, water depths, sediment depths, and sediment
                  compositions. During the demonstration, the range of
                                                   105

-------
platforms used included an 18-foot-long, 4-foot-wide Jon
boat in S1A2;  a sturdier, 30-foot-long,  8-foot-wide
pontoon boat in S2A1; and the EPA GLNPO Mudpuppy in
S1A1.  Because the sampler does not require electricity or
a tripod-mounted winch for deployment, sampler operation
was feasible from any location on the sampling platforms
used.

Because of the lightness of the sampler and extension
handle (when  needed), water depth had no significant
impact on the sampling technician's ability to deploy and
retrieve the sampler. In S1A2, the sampler was deployed
and retrieved using a 5-foot-long extension handle because
the water depth was about 2 feet. In S1A1 and S2A1,
where water depths were about 6 and 18 feet, respectively,
the sampler was deployed and retrieved using the polyester
line and messenger. As with other samplers, the Ekman
Grab's ability  to accurately assess the location of the
sediment surface decreases with increasing water depth
and turbidity. Because of the significant water depth and
turbidity in S1A1 and  S2A1 and the significant turbidity
in  S1A2,  the  sampling technician  could not see the
sediment surface from  the sampling  platforms.   An
underwater video camera may have enabled the sampling
technician to accurately assess the location of the sediment
surface in these areas (Blomqvist 1991).

Water velocity had an impact on the sampling technician' s
ability to deploy the sampler when gravity penetration was
used.  As mentioned above, the sampler was deployed in
S1A1 and S2A1  using the polyester line and messenger.
During a few sampling attempts in each area, the current
carried the sampler at least 1 foot beyond the desired
sampling location near the sediment surface. The average
water velocity in S1A1 and S2A1 was ^0.07 ft/s and
<0.05 ft/s, respectively. In S1A2, where the average water
velocity was less than 0.05 ft/s, the sampler was deployed
with a 5-foot-long extension handle because the water
depth was  only  2  feet.  Because use of the extension
handle provided  more control during positioning of the
sampler, water velocity had no significant impact on the
sampling technician's ability  to  properly  deploy the
sampler.

The sampler was able to collect surficial sediment samples
in  all three demonstration  areas. However,  the sampler
exhibited   a  few  limitations related  to  sediment
composition. As discussed in Section B.4.1.2 for primary
objective P2, the Ekman Grab performed poorly in terms
of its ability to consistently collect sediment samples in a
specified depth interval. InSlAl and S2A1, low sediment
recoveries were attributed to the failure of one or both of
the  scoops to  close after the messenger  was sent.  In
addition, in these  areas, the sampling technician  was
unable to push the sampler into the sediment because the
sampler  was  attached only  to  the  polyester  line.
Therefore, the  sampler may not have fully  penetrated the
target depth interval because the weight of the sampler
may not have been adequate to overcome the degree of
sediment compaction in these areas. In S1A2, the mean
actual sediment sample thickness  exceeded the  target
sediment  sample thickness.    The  excessive sample
thickness can be attributed to the sampling technician's
pushing the sampler beyond the specified  depth interval
because  of  his  difficulty in accurately  assessing the
location of the sediment surface using the sampler.

B.4.2.3 Ability to Collect an Undisturbed Sample

During the demonstration, as  was expected given the
nature of the sampler, the Ekman Grab did not consistently
collect  sediment  samples  in  which  the  sediment
stratification was preserved.  Specifically, in S1A1 and
S2A1, sediment stratification was not preserved. When
the  samples collected in these areas  were discharged into
stainless-steel bowls, the samples were unable to  retain
their form because of their high water content; as a result,
sediment from different layers  was  allowed  to  mix.
However, in S1A2, where the water content in the 0- to
4-inch  bss depth interval was relatively low  and the
sediment contained a relatively high clay content, the
samples were able to retain their form after discharge, and
the  sediment stratification was preserved.

The disturbance associated with bow wave formation near
the  water-sediment  interface  was not  likely to  be
significant  in   S1A2  because  the  speed of  sampler
deployment was controlled by  the  use of the extension
handle.  However, in S1A1 and S2A1, the sampler was
deployed using the polyester line; the sampler had to be
dropped in order to allow gravity  penetration  into the
target depth interval. As a result, the opportunity for bow
wave formation was greater in these areas.  However,
because of the water depth and turbidity in both areas, the
sampling technician was unable to observe whether bow
wave formation occurred.
                                                   106

-------
B.4.2.4 Durability Based on Materials of
       Construction and Engineering Design

As  described in  Section   B.I.I,  the  Ekman  Grab
components  are  made  of  either  stainless steel or
galvanized steel.  Based on observations made during the
demonstration, the Ekman Grab is a sturdy sampler; none
of the sampler components  was damaged or required
repair or replacement during the demonstration.

B.4.2.5 Availability of Sampler and Spare Parts

As mentioned above, no primary component of the Ekman
Grab was damaged or required replacement during the
demonstration.   Had a  primary  sampler  component
(excluding  the polyester line)  required replacement, it
would not  have  been available in  a  local retail store.
Replacement components may be  obtained from the
developer by overnight  courier in  2 days  or  less,
depending on the  location of  the sampling site.  The
polyester line, which may need occasional replacement,
should be available locally.

B.4.3  Data Quality

The overall QA objective for the demonstration was to
produce well-documented data of known quality.  The
TS As conducted to evaluate data quality did not reveal any
problems that  would make the   demonstration  data
unusable.   The scope of these TSAs is described in
Sections 4.3 and 4.4 of this ITVR.

This  section  briefly  discusses  the   data  quality of
demonstration results for the Ekman Grab; more  detailed
information is provided in the DER  (Tetra Tech 1999a).
Specifically, the data quality associated with the field
measurement activities is discussed first, followed by the
data quality  associated  with  the  laboratory  analysis
activities.

B.4.3.1 Field Measurement Activities

Field  measurement  activities  conducted  during the
demonstration  included  measurement  of  the  time
associated  with  sample  collection  activities, water
velocity, water depth, volume of IDW, volume of sediment
collected in a given sampling attempt, and depth of
sampler  deployment.     Of these  measurement
parameters, specific acceptance criteria were set  for the
precision associated with the time and water velocity
measurements  only (EPA  1999).  All time and water
velocity measurements made during the demonstration met
their respective criteria (see Table 6-7). Of the remaining
parameters,  some difficulties  were  encountered  in
measuring the  volume of sediment collected in a given
sampling attempt and the depth of sampler deployment,
which are discussed below.

To measure the volume of sediment collected in a given
sampling attempt, the  sediment  sample was  transferred
into a 2-L container graduated in increments of 20 mL.
The container was tapped on a hard surface to minimize
the presence of void spaces in the sample, the sample
surface was made  even using a spoon, and the volume of
the sample was measured.  However, because the  void
spaces could not be completely eliminated, the volumetric
measurements  are believed to have a positive bias that
resulted in overestimation  of PSRs.  Because  the  total
volume of the void spaces could not be measured, its
impact on the PSR results could not be quantified.

The  depth of  sampler deployment was measured  with
reference to the sediment surface. To identify the location
of the sediment surface, the sampling technician lowered
the sampler into the water and used the bottom end of the
sampler to feel the sediment surface. Subsequently, the
technician used an extension rod to drive the sampler into
the sediment to a depth that he estimated to be appropriate
to collect a sediment sample or used a polyester line to
allow the sampler to penetrate the sediment  by gravity.
Regardless of the method used to deploy the sampler, the
technician could  not  control  the  depth  of  sampler
deployment precisely; when the extension rod was used,
the actual depth of sampler deployment exceeded the
target depth of deployment by up to 2 inches, and when
the polyester line was  used, the sampler  did not  fully
penetrate the target depth interval. Because of the nature
of the Ekman Grab, when the actual depth of penetration
was more than the target depth of penetration (as indicated
by the volume of sediment sampled), the portion of the
sediment sample associated with the excessive depth  of
penetration could not be removed from the sampler before
the sediment volume was measured; consequently, the
PSR results had a positive  bias  that could  not be
quantified.
                                                   107

-------
B.4.3.2 Laboratory Analysis Activities

The laboratory analyses conducted for the demonstration
included  the following:  (1) PCB,  arsenic,  and PSD
analyses of sediment samples and (2) PCB and arsenic
analyses of equipment rinsate samples. To evaluate the
data quality of the laboratory  analysis  results,  field-
generated  QC samples, PE samples, and  laboratory  QC
check samples were analyzed.  The  field-generated  QC
samples included the  field replicates and temperature
blanks described in Section 4.3  of this ITVR. The PE
samples and laboratory QC check samples are described in
Section 4.4.  The acceptance criteria for the QC samples
are presented in Table 6-7.

All temperature blanks and field replicates subjected to
PCB and  arsenic  analyses met the acceptance criteria,
indicating that the sample homogenization procedure (field
replicates)  and   sample  preservation  procedure
(temperature blanks) implemented in the field met the
demonstration  requirements.   However, as  stated  in
Section B.4.1.3, in a few cases the results of field triplicate
sample analyses  for PSD did not meet the acceptance
criterion.  Despite the  failures to meet the acceptance
criterion, the PSD results are considered to be valid for the
reasons detailed in Section B.4.1.3.

The  PE sample results for both the PCB and arsenic
analyses met the acceptance  criteria, indicating that the
analytical laboratory accurately measured  both PCBs  and
arsenic.

The analytical results for all laboratory QC check samples
except the  following met  the  acceptance criteria:
(1) MS/MSD  samples for analysis for  PCBs  in  the
sediment matrix and (2) equipment rinsate samples for
PCB analysis. These issues and their likely impact on data
quality are discussed below.

For the sediment matrix, in all MS/MSD samples analyzed
for PCBs, Aroclor 1016 was recovered at levels higher
than the upper limit of the acceptance criterion, indicating
a positive  bias in the PCB results for sediment samples.
However,  the  analytical laboratory had no problem
meeting the acceptance criteria for control samples such as
BS/BSDs.   For this reason, the failure to  meet  the
acceptance criterion for MS/MSD sample analysis was
attributed to matrix interference.  The MS/MSD spiking
compounds (Aroclors 1016 and 1260) were selected based
on the Aroclors detected during the predemonstration
investigation  and   as   recommended   in  SW-846
Method 8082.

Also for the sediment matrix, in one out of three MS/MSD
pairs analyzed for PCBs, Aroclor 1260 was recovered at a
level less than the lower limit of the acceptance criterion
in the MS sample, but the recovery in the associated MSD
sample was acceptable. Because the investigative samples
contained only  Aroclor 1242,  of the  two  spiking
compounds used to prepare the MS/MSD samples, only
the Aroclor 1016  recoveries were considered to be
relevant based on the PCB  congener distribution; the
Aroclor  1260  recoveries were not considered to be
relevant.  Therefore, the low recovery associated  with
Aroclor 1260 had no impact on data quality.

In all equipment rinsate samples  analyzed for PCBs,
decachlorobiphenyl (the surrogate) was recovered at levels
lower than the lower limit of the acceptance criterion,
indicating a negative bias in the PCB results for equipment
rinsate samples.  However, the analytical laboratory had
no problem meeting the acceptance criteria for control
samples such as PE samples and deionized water blanks.
For this reason, the failure to meet the surrogate recovery
acceptance criterion  for the  equipment rinsate sample
analysis was attributed to matrix interference.
B.5    References

Analytical  Software.  1996.  Statistix11
   Version 2.0.  Tallahassee, Florida.
for Windows.
Blomqvist, S.   1991.  "Quantitative Sampling of Soft-
    Bottom Sediments: Problems and Solutions." Marine
    Ecology Progress Series.  Volume  72.  Pages 295
    through 304.

EPA. 1999. "Sediment Sampling Technologies Demon-
    stration Plan."  ORD.  Washington, DC. April.

Tetra Tech.  1999a.  "Sediment  Sampling Technologies
    Data Evaluation Report."  Prepared  for ORD, EPA.
    October.
                                                   108

-------
                                              Appendix C
                                          Statistical Methods
This appendix summarizes two statistical methods used in
evaluating the Split Core Sampler demonstration results:
the Wilk-Shapiro test for evaluating whether data are
normally or lognormally distributed (Section C.I) and the
Wilcoxon signed rank test for evaluating whether two data
sets are statistically different (Section C.2).  Section C.3
lists references used to prepare this appendix.  Examples
of the use of the two  tests  are included in each  test
description.  Both tests  were  performed using Statistix®
developed by Analytical Software of Tallahassee, Florida
(Analytical Software 1996).

C.I     Wilk-Shapiro Test

The Wilk-Shapiro test is an effective method for testing
whether a data set has been drawn from an underlying
normal distribution. Furthermore, by conducting the test
on the logarithms of the data, it is an equally effective way
of evaluating the hypothesis of a lognormal distribution.
This test was used to determine whether the demonstration
results  followed  either  the  normal   or  lognormal
distribution in order to use a parametric test, such as the
Student's t-test, for  evaluating the results  for primary
objectives PI, P2, and P4.  The Wilk-Shapiro test results
indicated that the data sets  for PI, P2, and  P4 were
generally not normally distributed or  could not be tested
for lognormality because the results contained values that
were equal to zero. Therefore, the Wilcoxon signed rank
test, a nonparametric test for paired samples that makes no
assumptions regarding the distribution, was used  as an
alternative to the Student's t-test.

For a given data set, the Statistix® software package first
counts the  number of values in the data set  and then
generates the same number of expected values as if the
data were perfectly, normally distributed. The expected
values are generated using a standard normal distribution
function (a standard normal distribution has a mean of 0
and a variance of 1). Both the actual and expected values
are ranked in numerical order and plotted;  the actual
values (ordered data)  are plotted on the y-axis, and the
expected values (rankits) are plotted on the x-axis. The
package  performs  a  linear  regression analysis and
calculates the square of the correlation coefficient, also
known as the approximate Wilk-Shapiro normality statistic
(W). The W values can range from 0 to 1; 0 indicates no
correlation  between actual  and expected values, and  1
indicates perfect correlation between actual and expected
values.

The W values calculated for each data set were compared
to critical W values corresponding to various significance
levels (a) and sample sizes (Gilbert 1987). If the W value
for a given data set was greater than the critical value
listed for the corresponding sample size at a= 0.05, the
data were  assumed to be  normally distributed.   The
examples discussed below illustrate this test.

Table  C-l presents two example data sets for primary
objective P2 that were tested for normality.  Figures C-l
and C-2 provide Statistix® Wilk-Shapiro test outputs for
these data sets. The calculated W values for S1A2 and
S2A2 were 0.9509 and 0.6740, respectively.  At a=0.05,
the critical W values for S1A2 and S2A2 were 0.818 and
0.905, respectively.  Because the calculated W value for
S1A2 was greater than the critical W value, the S1A2 data
for primary objective P2 were considered to be normally
distributed. The opposite was true for S2A2 data.
                                                    109

-------
Table C-1. Data Sets for Example Wilk-Shapiro Test Calculations

Demonstration Area
( S1A2 freshwater ay)
S2A2 (wetland)
Depth Interval
(inches bss)
t 3 12 o 2
4 to 12


7
1.5


3
8


8
0


7
1


8
0


5 11
0 0

Core Length (inches)
6
000 2.5 00000748



8
Note:

bss  =   Below sediment surface
    11 -
     9-
     7 -
     5-
     3-
Approximate Wilk-Shapiro
normality statistic:           0.9509

Number of cases:           8
        -2                       -1                        0

                                                      Rankits

Figure C-1. Wilk-Shapiro test plot for core length measurements in S1A2 (freshwater bay).
C.2    Wilcoxon Signed Rank Test

The Wilcoxon signed rank test is a nonparametric test for
paired samples that makes no assumptions regarding the
distribution of data. This test was selected to evaluate the
demonstration results for primary objectives PI, P2, and
P4 as an alternative to the paired Student's t-test, which
was originally prescribed in the demonstration plan under
the assumption that the demonstration results would be
normally  or  lognormally distributed.   The  Wilcoxon
signed rank test was selected  for evaluating the project
data because the Wilk-Shapiro test indicated that most of
the data sets were  neither normally nor  lognormally
distributed.
                                           The primary limitation of the Wilcoxon signed rank test is
                                           that it lacks the power of the Student's t-test because it
                                           does not consider the magnitude of the difference between
                                           sample  pair  results.   For example,  the test  cannot
                                           distinguish the difference between one pair in which the
                                           expected core length was 8 inches and the actual core
                                           length was 7.5 inches and another pair  in which the
                                           expected and actual core lengths were 8  and  0  inches,
                                           respectively.  Instead, the test first evaluates how many
                                           pairs in  a given data set have positive, negative,  or zero
                                           differences and then uses  this information to test the
                                           hypothesis. In addition, the test ignores cases in which the
                                           expected and actual core lengths are the same.
                                                     110

-------
     4 -
     2 -
     0 -
              Approximate Wilk-Shapiro
              normality statistic:           0.6740

              Number of cases:           20
                                                         0

                                                       Ran kits
Figure C-2. Wilk-Shapiro test plot for core length measurements in S2A2 (wetland).
The Wilcoxon signed rank test was performed using the
Statistix®  software  package,  which   calculated  the
probability value (p-value) at which the null hypothesis
was true.  The p-value was compared to an a of 0.05 to
determine whether the null hypothesis should be accepted
or rejected.  If the p-value exceeded a, it was concluded
that the mean difference for the paired results was not
statistically significant; otherwise, it was concluded that
the difference was statistically significant.

Several conclusions drawn from the Wilcoxon signed rank
test results for primary objectives PI and P2 did not seem
to be correct based on the  magnitude of the differences
observed for sample pairs in a given data set. However,
the results for primary objective P4 were evaluated using
this test  because no such  problem  was observed.  To
illustrate this point, example calculations are presented
below.

Table C-2 and Figure C-3 provide the primary objective
PI  Hand Corer sample data set for the 4- to  12-inch bss
depth interval in S2A2 and the corresponding Statistix®
output for the Wilcoxon signed rank test, respectively.
The test calculated a one-tailed p-value of 0.0625,
indicating that the difference between the expected and
actual number of attempts was not statistically significant
(the null hypothesis was that the mean difference between
the expected and actual values equals zero).  Because the
expected and  actual values differed for four of the five
sample  pairs and  particularly for the second pair, the
difference  was  in fact  considerable.   Therefore, the
conclusion drawn from the Wilcoxon signed rank test
appears to be incorrect.
Table C-2. Hand  Corer Sample  Data  for 4-  to  12-Inch Below
         Sediment Surface Depth Interval in S2A2 (Wetland)
Expected Number of Attempts
Actual Number of Attempts
                                      2
                                      12
                                      3
                                      2
                                      1
Table C-3 and Figure C-4 provide the primary objective
P4 Hand Corer and Split Core Sampler sample data for the
                                                     111

-------
4- to 12-inch below sediment surface depth interval
 STATISTIX FOR  WINDOWS

 WILCOXON SIGNED RANK TEST FOR S2A2_4_12 -  G

 SUM  OF NEGATIVE RANKS
 SUM  OF POSITIVE RANKS

 EXACT PROBABILITY OF A  RESULT AS  OR MORE
 EXTREME THAN THE OBSERVED RANKS  (1  TAILED  P-VALUE)

 NORMAL APPROXIMATION WITH CONTINUITY CORRECTION
 TWO-TAILED P-VALUE for  NORMAL APPROXIMATION

 TOTAL NUMBER OF VALUES  THAT WERE  TIED         2
 NUMBER OF ZERO DIFFERENCES DROPPED             1
 MAX.  DIFF. ALLOWED BETWEEN TIES          0.00001
                                                 8/3/99, 4:44:45 PM
                                  0.0000
                                  10.000
                                  0.0625

                                   1.643
                                  0.1003
 CASES  INCLUDED
MISSING CASES  6
Figure C-3. Statistix output for Hand Corer sample data for S2A2 (wetland).
Table C-3. Hand Corer and Split Core Sampler Sample Data for 10- to 30-inch Below Sediment Surface Depth Interval in S2A1 (Lake)

 Sampler                                Arsenic Concentration (milligrams per kilogram)
Hand Corer
Split Core Sampler
24
5.0
8.5
5.3
16
4.7
8.3
4.6
9.7
5.4
7.2
4.7
7.2
5.3
8.2
5.0
52
5.2
10-to 30-inch below sediment surface depth interval
 STATISTIX FOR  WINDOWS

 WILCOXON SIGNED RANK TEST FOR REFERENCE -  IS2

 SUM  OF NEGATIVE RANKS
 SUM  OF POSITIVE RANKS

 EXACT PROBABILITY OF A  RESULT AS  OR MORE
 EXTREME THAN THE OBSERVED RANKS  (1  TAILED  P-VALUE)

 NORMAL APPROXIMATION WITH CONTINUITY CORRECTION
 TWO-TAILED P-VALUE for  NORMAL APPROXIMATION

 TOTAL NUMBER OF VALUES  THAT WERE  TIED         2
 NUMBER OF ZERO DIFFERENCES DROPPED             0
 MAX.  DIFF. ALLOWED BETWEEN TIES          0.00001
                                                8/19/99, 3:19:04 PM
                                  0.0000
                                  45.000
                                  0.0020

                                   2.606
                                  0.0092
 CASES  INCLUDED  9
MISSING CASES  0
Figure C-4. Statistix output for Hand Corer and Split Core Sampler sample data for S2A1 (lake).
                                               112

-------
10- to  30-inch  bss depth interval  in  S2A1 and the   C.3   References
corresponding Statistix® output for the Wilcoxon signed
rank test, respectively. The test calculated a two-tailed   Analytical Software.  1996.   Statistix® for Windows.
p-value of 0.0092, indicating that the difference between       Version 2.0. Tallahassee, Florida.
the two sets of arsenic results was statistically significant
(the null hypothesis was that the mean difference between   Gilbert, R. 1987. Statistical Methods for Environmental
the innovative and reference  sampler sample analytical       Pollution  Monitoring.   Van  Nostrand  Reinhold
results for  the clean  layer equals zero).  Because the       Company, Inc. New York.
arsenic results for the Hand Corer  samples were greater
than those for the Split Core Sampler samples in each of
the nine pairs, the conclusion drawn from the Wilcoxon
signed rank test appears to be correct.
                                                    113

-------