1、Designation: D5956 15Standard Guide forSampling Strategies for Heterogeneous Wastes1This standard is issued under the fixed designation D5956; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision, the year of last revision. A number in p
2、arentheses indicates the year of last reapproval. Asuperscript epsilon () indicates an editorial change since the last revision or reapproval.1. Scope1.1 This guide is a practical, nonmathematical discussionfor heterogeneous waste sampling strategies. This guide isconsistent with the particulate mat
3、erial sampling theory, aswell as inferential statistics, and may serve as an introductionto the statistical treatment of sampling issues.1.2 This guide does not provide comprehensive samplingprocedures, nor does it serve as a guide to any specification. Itis the responsibility of the user to ensure
4、appropriate proce-dures are used.1.3 This standard does not purport to address all of thesafety concerns, if any, associated with its use. It is theresponsibility of the user of this standard to establish appro-priate safety and health practices and determine the applica-bility of regulatory limitat
5、ions prior to use.2. Terminology2.1 Definitions of Terms Specific to This Standard:2.1.1 attribute, na quality of samples or a population.2.1.1.1 DiscussionHomogeneity, heterogeneity, and prac-tical homogeneity are population attributes. Representativenessand intersample variance are sample attribut
6、es.2.1.2 characteristic, na property of items, a sample orpopulation that can be measured, counted, or otherwise ob-served.2.1.2.1 DiscussionA characteristic of interest may be thecadmium concentration or ignitability of a population.2.1.3 component, nan easily identified item such as a largecrystal
7、, an agglomerate, rod, container, block, glove, piece ofwood, or concrete.2.1.4 composite sample, na combination of two or moresamples.2.1.4.1 DiscussionWhen compositing samples to detecthot spots or whenever there may be a reason to determinewhich of the component samples that constitute the compos
8、iteare the source of the detected contaminant, it can be helpful tocomposite only portions of the component samples. Theremainders of the component samples then can be archived forfuture reference and analysis. This approach is particularlyhelpful when sampling is expensive, hazardous, or difficult.
9、2.1.5 correlation, nthe mutual relation of two or morethings.2.1.6 database, na comprehensive collection of relateddata organized for quick access.2.1.6.1 DiscussionDatabase as used in this guide refers toa collection of data generated by the collection and analysis ofmore than one physical sample.2
10、.1.7 data quality objectives (DQO), nDQOs are qualita-tive and quantitative statements derived from the DQO processdescribing the decision rules and the uncertainties of thedecision(s) within the context of the problem(s).2.1.8 data quality objective process, na quality manage-ment tool based on the
11、 scientific method and developed by theU.S. Environmental Protection Agency to facilitate the plan-ning of environmental data collection activities.2.1.8.1 DiscussionThe DQO process enables planners tofocus their planning efforts by specifying the use of the data(the decision), the decision criteria
12、 (action level) and thedecision makers acceptable decision error rates. The productsof the DQO process are the DQOs.2.1.9 heterogeneity, nthe condition of the population un-der which items of the population are not identical with respectto the characteristic of interest.2.1.10 homogeneity, nthe cond
13、ition of the populationunder which all items of the population are identical withrespect to the characteristic of interest.2.1.10.1 DiscussionHomogeneity is a word that has morethan one meaning. In statistics, a population may be consideredhomogeneous when it has one distribution (for example, if th
14、econcentration of lead varies between the different items thatconstitute a population and the varying concentrations can bedescribed by a single distribution and mean value, then thepopulation would be considered homogeneous). A populationcontaining different strata would not have a single distribut
15、ionthroughout, and in statistics, may be considered to be hetero-geneous. The terms homogeneity and heterogeneity as used inthis guide, however, reflect the understanding more common to1This guide is under the jurisdiction of ASTM Committee D34 on WasteManagement and is the direct responsibility of
16、Subcommittee D34.01.01 onPlanning for Sampling.Current edition approved May 1, 2015. Published May 2015. Originallyapproved in 1996. Last previous edition approved in 2006 as D5956 96 (2006),which was withdrawn in January 2015 and reinstated in May 2015. DOI:10.1520/D5956-15.Copyright ASTM Internati
17、onal, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States1chemists, geologists, and engineers. The terms are used asdescribed in the previous definitions and refer to the similarityor dissimilarity of items that constitute the population. Accord-ing to this guide, a p
18、opulation that has dissimilar items wouldbe considered heterogeneous regardless of the type of distri-bution.2.1.11 item, na distinct part of a population (for example,microscopic particles, macroscopic particles, and 20-ft longsteel beams).2.1.11.1 DiscussionThe term component defines a subsetof it
19、ems. Components are those items that are easily identifiedas being different from the remainder of items that constitutethe population. The identification of components may facilitatethe stratification and sampling of a highly stratified populationwhen the presence of the characteristic of interest
20、is correlatedwith a specific component.2.1.12 population, nthe totality of items or units underconsideration.2.1.13 practical homogeneity, nthe condition of the popu-lation under which all items of the population are not identical.For the characteristic of interest, however, the differencesbetween i
21、ndividual physical samples are not measurable orsignificant relative to project objectives.2.1.13.1 DiscussionFor practical purposes, the populationis homogeneous.2.1.14 random, nlack of order or patterns in a populationwhose items have an equal probability of occurring.2.1.14.1 DiscussionThe word r
22、andom is used in twodifferent contexts in this guide. In relation to sampling, randommeans that all items of a population have an equal probabilityof being sampled. In relation to the distribution of a populationcharacteristic, random means that the characteristic has anequal probability of occurrin
23、g in any and all items of thepopulation.2.1.15 representative sample, na sample collected in sucha manner that it reflects one or more characteristics of interest(as defined by the project objectives) of a population fromwhich it was collected.2.1.15.1 DiscussionA representative sample can be (1)asi
24、ngle sample, (2) a set of samples, or (3) one or morecomposite samples.2.1.16 sample, na portion of material that is taken fortesting or for record purposes.2.1.16.1 DiscussionSample is a term with numerousmeanings. The scientist collecting physical samples (forexample, from a landfill, drum, or was
25、te pipe) or analyzingsamples, considers a sample to be that unit of the populationcollected and placed in a container. In statistics, a sample isconsidered to be a subset of the population, and this subset mayconsist of one or more physical samples. To minimize confu-sion the term physical sample is
26、 a reference to the sample heldin a sample container or that portion of the population that issubjected to in situ measurements. One or more physicalsamples, discrete samples, or aliquots are combined to form acomposite sample. The term sample size has more than onemeaning and may mean different thi
27、ngs to the scientist and thestatistician. To avoid confusion, terms such as sample mass orsample volume and number of samples are used instead ofsample size.2.1.17 sample variance, na measure of the dispersion of aset of results. Variance is the sum of the squares of theindividual deviations from th
28、e sample mean divided by oneless than the number of results involved. It may be expressedas s25( xi2x!2/n21!.2.1.18 sampling, nobtaining a portion of the materialconcerned.2.1.19 stratum, na subgroup of a population separated inspace or time, or both, from the remainder of the population,being inter
29、nally consistent with respect to a target constituentor property of interest, and different from adjacent portions ofthe population.2.1.19.1 DiscussionA landfill may display spatially sepa-rated strata since old cells may contain different wastes thannew cells. A waste pipe may discharge temporally
30、separatedstrata if night-shift production varies from the day shift. Also,a waste may have a contaminant of interest associated with aparticular component in the population, such as lead exclu-sively associated with a certain particle size.2.1.19.2 DiscussionHighly stratified populations consistof s
31、uch a large number of strata that it is not practical oreffective to employ conventional sampling approaches, norwould the mean concentration of a highly stratified populationbe a useful predictor (that is, the level of uncertainty is toogreat) for an individual subset that may be subjected toevalua
32、tion, handling, storage, treatment, or disposal. Highlystratified is a relative term used to identify certain types ofnonrandom heterogeneous populations. Classifying a popula-tion according to its level of stratification is relative to thepersons planning and performing the sampling, theirexperienc
33、e, available equipment, budgets, and sampling objec-tives. Under one set of circumstances a population could beconsidered highly stratified, while under a different context thesame population may be considered stratified.2.1.19.3 DiscussionThe terms stratum and strata are usedin two different contex
34、ts in this guide. In relation to thepopulation of interest, stratum refers to the actual subgroup ofthe population (for example, a single truck load of lead-acidbatteries dumped in the northeast corner of a landfill cell). Inrelation to sampling, stratum or strata refers to the subgroupsor divisions
35、 of the population as assigned by the samplingteam. When assigning sampling strata, the sampling teamshould maximize the correlation between the boundaries of theassigned sampling strata and the actual strata that exist withinthe population. To minimize confusion in this guide, thosestrata assigned
36、by the sampling team will be referred to assampling strata.3. Significance and Use3.1 This guide is suitable for sampling heterogeneouswastes.3.2 The focus of this guidance is on wastes; however, theapproach described in this guide may be applicable to non-waste populations, as well.D5956 1523.3 Sec
37、tions 49describe a guide for the sampling ofheterogeneous waste according to project objectives. AppendixX1 describes an application of the guide to heterogeneouswastes. The user is strongly advised to read Annex A1 prior toreading and employing Sections 49of this guide.3.4 Annex A1 contains an intr
38、oductory discussion ofheterogeneity, stratification, and the relationship of samplesand populations.3.5 This guide is intended for those who manage, design, orimplement sampling and analytical plans for the characteriza-tion of heterogeneous wastes.4. Sampling Difficulties4.1 There are numerous diff
39、iculties that can complicateefforts to sample a population. These difficulties can beclassified into four general categories:4.1.1 Population access problems making it difficult tosample all or portions of the population;4.1.2 Sample collection difficulties due to physical proper-ties of the populat
40、ion (for example, unwieldy large items orhigh viscosity);4.1.3 Planning difficulties caused by insufficient knowledgeregarding population size, heterogeneity of the contaminant ofinterest, or item size, or a combination thereof; and,4.1.4 Budget problems that prevent implementation of aworkable, but
41、 too costly, sampling design.4.2 The difficulties included in the first three categories area function of the physical properties of the population beingsampled. The last sampling difficulty category is a function ofbudget restraints that dictate a less-costly sampling approachthat often results in
42、a reduced number of samples and a reducedcertainty in the estimates of population characteristics. Budgetrestraints can make it difficult to balance costs with the levelsof confidence needed in decision making. These difficultiesmay be resolved by changing the objectives or sampling/analytical plans
43、 since population attributes or physical proper-ties of the population can seldom be altered. Documents onDQOs discuss a process for balancing budgets with neededlevels of confidence.4.3 Population access and sample collection difficultiesoften are obvious, and therefore, more likely either to beadd
44、ressed or the resulting limitations well-documented. A fieldnotebook is likely to describe difficulties in collecting largeitems or the fact that the center of a waste pile could not beaccessed.4.4 Population size, heterogeneity, and item size have asubstantial impact on sampling. The cost and diffi
45、culty ofaccurately sampling a population usually is correlated with theknowledge of these population attributes and characteristics.The least understood population attribute is heterogeneity ofthe characteristic of interest. If heterogeneity is not knownthrough process knowledge, then some level of
46、preliminarysampling or field analysis is often required prior to samplingdesign.4.5 Sampling of any population may be difficult. However,with all other variables being the same, nonrandom heteroge-neous populations are usually more difficult to sample. Theincreased difficulty in sampling nonrandom h
47、eterogeneouspopulations is due to the existence of unidentified or numerousstrata, or both. If the existence of strata are not consideredwhen sampling a nonrandom heterogeneous population, theresulting data will average the measured characteristics of theindividual strata over the entire population.
48、 If the differentstrata are relatively similar in composition, then the meancharacteristic of the population may be a good predictor forportions of the population and will often allow the project-specific objectives to be achieved. As the difference in com-position between different strata increases
49、, average populationcharacteristics become less useful in predicting composition orproperties of individual portions of the population. In this lattercase, when possible, it is advantageous to sample the individualstrata separately, and if an overall average of a populationcharacteristic is needed, it can be calculated mathematicallyusing the weighted averages of the sampling stratum means(1).5. Stratification5.1 Strata can be thought of as different portions of apopulation, which may be separated in time or space with eachportion having internal
