Radiation Monitoring Around United States Nuclear Test Areas , Calendar Year 1996

This report describes the Offsite Radiation Safety Program conducted during 1996 by the U. S. Environmental Protection Agency's (EPAs), Off ice of Radiation and Indoor Air-Las Vegas, Radiation Science Laboratory . This laboratory operated an environmental radiation monitoring program in the region surrounding the Nevada Test Site (NTS) and at former test sites in Alaska, Colorado, Mississippi, Nevada, and New Mexico. The surveillance program is designed to measure levels and trends of radioactivii, if present, in the environment surrounding testing areas to ascertain whether current radiation levels and associated doses to the general public are in compliance with existing radiation protection standards. The surveillance program additionally has the responsibility to take action to protect the health and well being of the public in the event of any accidental release of radioactii contaminants. Offsite levels of radiation and radioactivii are assessed by sampling milk, water, and air; by deploying thermoluminescent dosimeters (TLDs); and using pressurized ionization chambers (PICs). No nuclear weapons testing was conducted in 1996 due to the continuing nuclear test moratorium. During this period, R&IE personnel maintained readiness capability to provide direct monitoring support if testing were to be resumed and ascertained compliance with applicable EPA, DOE, state, and federal regulations and guidelines. Comparison of the measurements and sample analysis results with background levels and with appropriate standards and regulations indicated that there was no airborne radioactivii from diffusion or resuspension detected by the various EPA monitoring networks surrounding the NTS. There was no indication of potential migration of radioactivii to the offsite area through groundwater and no radiation exposure above natural background was received by the offsite population. All evaluated data were consistent with previous data history. Using the EPAs CAP88-PC model and NTS radionuclide emissions and environmental monitoring data, the calculated effective dose equivalent (EDE) to the mm'mally exposed individual offsite would have been about 0.1 1 mrem. This value is less than two percent of the Federal dose limit prescribed for radionuclide air emissions. The dose received from natural background radiation was about 144 mrem. The offsite Environmental Monitoring Report: Radiation Monitoring Around United States Nuclear Test Areas, Calendar Year 1994 and 1995 was not and will not be published. Please refer to the 1994 and 1995 Nevada Test Site Annual Site Environmental Report, for data covering that time period.


PREFIXES CONVERSIONS
Assuring the health and safety of the people living near the NTS. active contaminants in the vicinity of past atomic testing areas.
Maintaining readiness to resume nuclear testing at some future date.
Verifying compliance with applicable radiation protection standards, guidelines, and regulations.
Offsite levels of radiation and radioactivity are assessed by gamma-ray measurements using pressurized ion chambers (PICs) and thermoluminescent dosimeters (TLDs); and by sampling air, water, and milk:

Program Summary and Conclusions
The primary functions of the ORSP are to conduct routine environmental monitoring for radioactive materials in areas potentially impacted by nuclear tests and, when necessary, to implement actions to protect the public from radiation exposure. Components of the ORSP include surveillance networks for air, and milk, exposure monitoring by thermoluminescent dosimetry, and pressurized ion chambers, and long-term hydrological monitoring of wells and surface waters. In 1996, data from all networks and monitoring activities indicated no radiation directly attributable to current activities conducted at the NTS. Therefore, protective actions were not required. The following sections summarize the ORSP activities for 1996.

Thermoluminescent Dosimetry Program
In 1996, external exposure was monitored by a network of thermoluminescent dosimeters (TLDs) at 49 fixed locations surrounding the NTS and by TLDs worn by 25 off site residents. No net exposures were related to NTS activities. Neither administrative, ALARA, nor regulatory investigation limits were exceeded for any individual or fixed location cumulative exposure. The range of exposures was similar to those observed in other areas of the United States and were slightly lower than those of the past.
Measuring and documenting levels and trends of environmental radiation or radio- No radioactivity attributable to current NTS operations was detected by any of the monitoring networks. However, based on the releases reported by NTS users, atmospheric dispersion model calculations (CAP88-PC) (EPA 1992) indicated that the maximum potential effective dose equivalent to any offsite individual would have been 0.1 1 mrem (1.1 x mSv), and the dose to the population within 80 kilometers of the emission sites would have been 0.34 person-rem (3.4 x IU3 person-Sv). The hypothetical person receiving this dose was also exposed to 144 mrem from normal background radiation. Details of this program may be found in Section 3 of this Report.

Network
The Pressurized Ionization Chamber (PIC) network measures ambient gamma radiation exposure rates on a near real-time basis. The 26 PlCs deployed around the NTS in 1996 showed no unexplained deviations from background levels. These background exposures, ranging from 71 to 156 mR/yr are within the U.S. background range and are consistent with previous years' trends. Details of this program may be found in Section 3 of this Report.

Air Surveillance Network
In 1996, the Air Surveillance Network (ASN) included 20 continuously operating sampling stations at locations surrounding the NTS. In the majorrty of cases, no gamma emitting radionuclides were detected by gamma spectrometry (i.e., the results were gamma-spectrum negligible). Naturally occurring ' Be was the only radionuclide occasionally detected. As in previous years, the majority of the gross beta results exceeded the MDC. Analysis of air samples for gross alpha showed results to be either below or very slightly above (i.e. statistically indistinguishable from) the MDC. The MDC for 239+240Pu was exceeded for one high volume sample from Rachel, NV. Details of the Atmospheric Monitoring program may be found in Section 4 of this Report.

Milk
Milk samples were collected from 11 Milk Surveillance Network (MSN) stations in 1996. The average total potassium concentration derived from ' OK was consistent with results obtained in previous years. No man-made gamma-emitting radionuclides were detected in any of the milk samples. Results of analyses for OeSr and @ ' S r were similar to those obtained in previous years. Neither increasing nor decreasing trends were evident. Detailed discussion of the collection and analysis of milk may be found in Section 5 of this report.

.I .5 Long-Term Hydrological
Monitoring Program

5.1 Nevada Test Site Monitoring
Nineteen wells on the NTS or immediately outside its borders on federally owned land were sampled. All samples collected during 1996 were analyzed for gamma-emitting radionuclides by gamma spectrometry and for triiium by the conventional andlor the enrichment method. No gamma-emitting radionuclides were detected. The highest triiium level, detected in a sample from Well UE-5n (4.5 x lo4 pCK) , was less than 60% of the defied concentration guide for tritium. There were no indications that migration from any test cavity is affecting any domestic water supply.

3 . 2 Offsite Monitoring in the Vicinity of the Nevada Test Site
These sampling locations represent drinking water sources for rural residents and for communities in the area. Sampling locations include 12 wells, nine springs, and a surface water site. All the locations are sampled quarterly or semiannually. Gamma spectrometric analysis is completed on all samples. No man-made gamma-emitting radionuclides were detected. Triiium analysis is performed on a semiannual basis.
None of the 1996 samples analyzed for triiium using the conventional method had results above the MDC. Two that were analyzed for tritium by the enrichment method showed detectable activity. These results were felt to represent scavenged atmospheric tritium by precipitation. Details of the on-site, near NTS, and off-NTS hydrological monitoring programs may be found in Section 6 of this Report.

.I .6 Dose Assessment
The extensive offsite environmental surveillance system detailed in this report measured no radiation exposures that could be attributed to recent NTS activities. The potential Effective Dose Equivalent (EDE) to the maximally exposed offsite resident was calculated to be 0.015 mrem, using certain assumptions as all data were not available due to a decrease of funding. Background radiation yielded a CEDE of 3,064 person-rem(30.6 person-Sv). Details of the dose assessment calculations may be found in Section 7 of this Report.

.I .7 Hazardous Spill Center
EPA participated on the control board for four series of spill tests using 28 different chemicals conducted at the HSC, located in Area 5 of the NTS. The amounts used in the tests were so small that boundary monitoring was not necessary.
human exposure pathways. Another component is public information.
As a result of the continuing moratorium on nuclear weapons testing, only simulated tests were conducted in 1996. Three simulated nuclear weapons test readiness exercises and one non-proliferation experiment using conventional (non-nuclear) explosives were conducted at the NTS. For each one, R&IE-LV senior personnel served on the Test Controller's Scientific Advisory Panel and on the EPA offsite radiological safety staff. To add as much realism as possible to the exercises, actual meteorological conditions were used and data flow was managed in the same manner as a real test. Routine offsite environmental radiation monitoring continued throughout 1996, as in past years.
Public information presentations provide a forum for increasing public awareness of NTS activities, disseminating radiation monitoring results, and addressing concerns of residents related to environmental radiation and possible health effects. Community Technical Liaison Program (CTLP) stations have been established in prominent locations in a number of offsite communities. The CTLP stations contain samplers for several of the monitoring networks and are managed by local residents. The University of Utah and DRI are cooperators with EPA in the CTLP. The CTLP is discussed in Section 3.
Environmental monitoring networks, described in the following subsections, measure radioactivity in air, milk, and ground water. These networks monitor the major potential pathways of radionuclide transfer to man via inhalation, submersion, and ingestion. Gamma radiation levels are continuously monitored at selected locations using Reuter-Stokes pressurized ion chambers (PICs) and Panasonic TLDs. Atmospheric monitoring equipment includes both high-and low volume air samplers. Milk is sampled and analyzed annually. Ground water on and in the vicinity of the NTS is monitored in the Long-Term Hydrological Monitoring Program (LTHMP). Data from these monitoring networks are used to calculate an annual exposure dose to the offsite residents, as described in Section 7. Detailed discussion of R&IE-LV activities in support of this facility may be found in Section 8 of this Report.

Offsite Radiological 1.2 Offsite Monitoring Quality Assurance
Under the t e n s of an Interagency Agreement The policy of the EPA requires participation in a between DOE and EPA, the EPA R&IE conducts the centrally managed QA Program by all EPA organi-Offsite Radiation Safety Program (ORSP) in the zational units involved in environmental data collecareas surrounding the NTS. The largest component tion. The QA Program developed by the R&IE for the of R&IE's program is routine monitoring of potential Offsite Radiological Safety Program (ORSP) meets all requirements of EPA policy, and also includes applicable elements of the Department of Energy QA requirements and regulations. The ORSP QA program defines data quality objectives (DQOs), which are statements of the quality of data a decision maker needs to ensure that a decision based on those data is defensible. Achieved data quality may then be evaluated against these DQOs. In addition, R&IE meets the EPA policy which states that all decisions which are dependent on environmental data must be supported by data of known quality. EPA policy requires participation in a centrally managed Quality Assurance Program by all EPA elements as well as those moniloring and measurement efforts supported or mandated by contracts, regulations, or other formalized agreements. The R&IE QA policies and requirements are summarized in the "Quality Management Plan" (EPA/R&IE 1996).

Nonradiological Monitoring
R&IE also provides support for the HAZMAT Spill Center( HSC) located at Frenchman Flat in Area 5 of the NTS. The HSC was designed for safe research on the handling, shipping, and storage of liquified gaseous fuels and other hazardous liquids. The R&IE provides a chemist to participate in meetings of the Advisory Panel which reviews and approves all programs prior to testing and maintains readiness for monitoring emissions at the boundary of the NTS.
For those tests requiring monitoring, the R&IE personnel deploy air sampling sensors to detect any offsite releases. No spills required moniloring in 1996 as such small amounts were released that they would be far below the limit of detection for the R&IE monitoring equipment at the edge of the NTS under any reasonable scenario, including catastrophic failure of the container in which the spill material is stored.

Description of the Nevada Test Site
The NTS, located in southern Nevada, was the primary location for testing of nuclear explosives in the continental U.S. from 1951 until the present moratorium began. Historical testing has included (I) atmospheric testing A the 1950s and early 1960s,

Location
The NTS is located in Nye County, Nevada, with its southeast corner about 54 miles (90 km) northwest of Las Vegas (Figure 2.1). It occupies an area of about 1,350 square miles (3,750 square km), varies from 28 to 35 miles (46 to 58 km) in width (eastwest) and from 49 to 55 miles (82 to 92 km) in length (north-south). This area consists of large basins or flats about 2,970 to 3,900 feet (900 to 1,200 m) above mean sea level (MSL) surrounded by mountain ranges rising from 5,940 to 7,590 feet (1,800 to 2,300 m) above mean sea level (MSL).
The NTS is surrounded on three sides by exclusion areas, collectively named the Nellis Air Force Base Range Complex, which provides a buffer zone between the test areas and privately owned lands. This buffer zone varies from 14 to 62 miles (24 to 104 km) between the test area and land that is open to the public.

Climate
The climate of the NTS and surrounding area is variable, due to its wide range in altitude and its rugged terrain. Most of Nevada has a semi-arid climate characterized as mid-latitude steppe.
Throughout the year, water is insufficient to support the growth of common food crops without irrigation. Climate may be classified by the types of vegetation indigenous to an area. According to Nevada Weather and Climate (Houghton et al., 1975), this method of classlication developed by K6ppen is further subdivided on the basis of "...seasonal distribution of rainfall and the degree of summer heat or winter cold." Table 2.1 summarizes the characteristics of climatic types for Nevada.
According to Quiring (1968), the NTS average annual precipitation ranges from about 4 inches (10 cm) at the lower elevations to around 10 inches (25 cm) at the higher elevations. During the winter months, the plateaus may be snow-covered for a period of several days or weeks. Snow is uncommon on the flats. Temperatures vary considerably with elevation, slope, and local air currents. The average daily temperature ranges at the lower altitudes are around 25 to 50°F (-4 to I 0°C) in January and 55 to 95°F (13 to 35°C) in July, with extremes of -15°F (-26°C) and 120°F (49°C). Corresponding temperatures on the plateaus are 25 to 35°F (-4 to 2%) in January and 65 to 80°F (18 to 27°C) in July with extremes of -30°F (-34°C) and 115°F (46°C).
The wind direction, as measured on a 98 ft (30 m) tower at an observation station approximately 7 miles  Limits of annual precipitation overlap because of variations in temperature which affect the water balance.
(1 1 km) north-northwest of CP-1, is predominantly northerly except during the months of May through August when winds from the south-southwest predominate (Quiring, 1968). Because of the prevalent mountain/valley winds in the basins, south to southwest winds predominate during daylight hours of most months. During the winter months, southerly winds predominate slightly over northerly winds for a few hours during the warmest part of the day. These wind patterns may be quite different at other locations on the NTS because of local terrain effects and differences in elevation.

Hydrology
Two major hydrologic systems shown in Figure 2.2 exist on the NTS (US. Energy Research and Development Administration, 1977). Ground water in the northwestern part of the NTS (the Pahute Mesa area) flows at a rate of 6.6 to 600 feet (2 to 180 m) per year to the south and southwest toward the Ash Meadows discharge area in the Amargosa Desert. Ground water to the east of the NTS moves from north to south at a rate of not less than 6.6 feet (2 m) nor greater than 730 feet (220 m) per year. Carbon-14 analyses of this eastern ground water indicate that the lower velocity is nearer the true value. At Mercuryvalley in the extreme southern part of the NTS, the eastem ground water flow shifts to the southwest, toward the Ash Meadows discharge area. farming of a variety of crops. Grazing is also common in this area, particularly to the northeast. The area north of the NTS is also mid-latitude steppe, where the major agricultural activity is grazing of cattle and sheep. Minor agriculture, primarily the growing of alfalfa hay, is found in this portion of Nevada within 180 miles (300 km) of the CP-1. Many of the residents have access to locally grown fruits and vegetables.

Regional Land Use
Recreational areas lie in all directions around the NTS and are used for such activities as hunting, fishing, and camping. In general, the camping and fishing sites to t'he northwest, north, and northeast of the NTS are closed during winter months. Camping and fishing locations to the southeast, south, and southwest are utilized throughout the year. The peak of the hunting season is from September through January.

Population Distribution
The population of counties surrounding the NTS based on the 1990 Bureau of Census (BOC) count (DOC, 1990) is still fairly accurate although growth has occurred in all parts of the state. Excluding Clark County, which has grown tremendously since the 1990 census and is the major population center (approximately 1,000,000 in 1996), the population density within a 90-mi (1 50-km) radius of the NTS is about 0.9 persons per square mile (0.5 persons per square kilometer  , 1990).
The extreme southwestern region of Utah is more developed than the adjacent part of Nevada. The largest community is St. George, located 132 miles (220 km) east of the NTS, with a 1996 population estimated at 40,000. The next largest town, Cedar City, with a population of over 18,000, is located 168 miles (280 km) east-northeast of the NTS. The extreme northwestern region of Arizona is mostly range land except for that portion in the Lake Mead National Recreation Area. In addition, several small communities lie along the Colorado River.
iDoc; 1 990). jhe estimated average population The largesf towns the area are Bullhead City, 99 for Nevada i n Igg0 was persons per miles (165 km) soufh-southeast of the NTS, with a 'quare (3.1 persons per 'quare 1990 population of 21,951 and Kingman, located 168 (DOC, 1986). miles (280 km) southeast of the NTS, with a popula- The offsite area within 48 miles (80 km) of CP-1 (the tion of 12,722 (DOC, 1990).
primary area in which the dose commitment must be determined for the purpose of this report) is predominantly rural. Several small communities are located in the area, the largest being in Pahrump Valley. Pahrump, a growing rural community with a population of about 23,000 (Pahrump Times) in 1996, is located 48 miles (80 km) south of CP-1. The small residential community of Crystal, Nevada, also located in the Pahrump Valley, is several miles north of the town of Pahrump (Figure 2.2). The Amargosa farm area, which has a population of about 950, is located 30 miles (50 km) southwest of CP-1. The largest town in the near offsite area is Beatty, which has a population of about 1,500 and is located approximately 39 miles (65 km) to the west of CP-1.

External Ambient Gamma Monitoring
External ambient gamma radiation is measured by the Therrnoluminescent Dosimetry (TLD) Network and also by the Pressurized Ion Chamber (PIC) Network. The primary function of the two networks is to detect changes in ambient gamma radiation. In the absence of nuclear testing, ambient gamma radiation rates naturally differ among locations since rates vary with altitude (cosmic radiation) and with radioactii in the soil (terrestrial radiation). Ambient gamma radiation will also vary slightly at a location due to changes in weather patterns and other factors.

Thermoluminescent Dosimetry Network
The primary purpose of the EPA R&IE-LV offsite environmental dosimetry program is to establish dose estimates to populations living in the areas surrounding the NTS. This is accomplished by developing baseline information regarding ambient radiation levels from all radiation sources and looking for any deviations from data trends. In addition to the environmental TLD program, EPA deploys personnel TLDs to Community Technical Liaison Program (CTLP) station managers and their alternates, living in areas surrounding the NTS. Information gathered from this program would help identify possible exposures to residents.

Design
The current EPA TLD program utilizes the Panasonic Model UD-802 TLD for personnel monitoring and the UD-814 TLD for environmental monitoring. Each dosimeter is read using the Panasonic Model UD-71 0A automatic dosimeter reader.
The UD-802 TLD incorporates two elements of Li ,B,O,:Cu and two elements of CaSO,: Tm phosphors. The phosphors are behind approimately 17, 300,300, and 1000 mg/cm2 of attenuation, respectively. With the use of different phosphors and filtrations, a dose algorithm can be applied to ratios of the different element responses. This process defines the radiation type and energy and provides a mechanism for assessing an absorbed dose equivalent.
Environmental monitoring is accomplished using the UD-814 TLD, which is made up of one element of Li2B40,:Cu and three elements of Ca$04:Tm. The CaS0,:Tm elements are behind approximately 1000 mg/cm2 attenuation. An average of the corrected values for elements two through four gives the total exposure for each TLD. Two UD-814 TLDs are deployed at each station per monitoring period.
In general terms, TLDs operate by trapping electrons at an elevated energy state. After the collection period, each TLD element is heated. When heat is applied to the phosphor, the trapped electrons are released and the energy differences between the initial energies of the electrons and the energies at the elevated state are given off in the form of photons. These photons are then collected using a photomultiplier tube. The number of photons emitted, and the resulting electrical signal, is proportional to the initial deposited energy.
New computers and software were installed in 1996 to increase report options, and further hardware upgrades will be completed in 1997.

Results of TLD Monitoring
ENVIRONMENTAL DATA: In 1996, the TLD program consisted of 49 fixed environmental monitoring stations and 25 offsite personnel. Henderson and Boulder City, Nevada, were added to the network in the fourth quarter. Figure 3.1 shows the fixed environmental TLD monitoring stations and the location of personnel monitoring participants. Total annual exposures were calculated by dividing each quarterly result by the number of days representing each deployment period. The quarterly daily rates were averaged to obtain an annual daily average. If a deployment period overlapped the beginning or end of the year a daily rate was calculated, for that deployment period, and multiplied by the number of days that fell within 1996. The total average daily rate was then multiplied by 365.25 to determine the total annual exposure for each station.
There were 49 offsite environmental stations monitored using TLDs.

Offsite Dosimetry Locations
Personnel ( Table 3.1 for 1996 results. These results are consistent with those for 1995.

PERSONNEL DATA:
Twenty-five offsite residents, managers, and alternates for the CTLP, were issued TLDs to monitor their annual dose equivalent. Locations of personnel monitoring participants are also shown in Figure 3.1 Annual whole body dose equivalents ranged from a low of 48 mrem (0.48 mSv) to a high of 125 mrem (1.2 mSv) with a mean of 96 mrem (0.96 mSv) for all monitored personnel during 1996. See Table 3.2 for 1996 results. These results are similar to those for 1995.

Quality Assurance1 Quality Control
The following procedures assure that the TLD data are of acceptable quality: Two calibration instruments were available to support the program. One is a TLD irradiator manufactured by Williston-Elin housing a nominal 1.8 Ci 137Cs source. This irradiator provides for automated irradiations of the TLDs. The second calibration method used a nominal 10 Ci '137Cs well type irradiator. Unlike the Williston-Elin irradiators, this well type does not provide automated capabilities. TLD exposures accomplished with the well type irradiator are monitored using a Viitoreen E-5000 precision electrometer whose calibration is traceable to the National Institute of Standards and Technology (NIST). The exposure rates of both irradiators have been confirmed by measurement using a precision electrometer which has a calibration traceable to NIST. Panasonic UD-802 dosimeters exposed by these irradiators are used to calibrate the TLD readers and to verify TLD reader linearity. Control dosimeters of the same type as field dosimeters (UD-802 or UD-814) are exposed and read together with the field dosimeters. This provides daily on-line process quality control checks in the form of irradiated controls.
For each read-out three irradiated control TLDs are included that have been exposed to a nominal 200 mR. After the irradiated controls have been read, the ratio of recorded exposure to delivered exposure is calculated and recorded for each of the four elements of the dosimeter. This ratio is applied to all raw element readings from field and unirradiated control dosimeters to automatically compensate for reader variations.
Prior to being placed in service, element correction factors are determined for all dosimeters. Whenever a dosimeter is read, the mean of the three most recent correction factor determinations is applied to each element to compensate for normal variability (caused primarily by the TLD manufacturing process) in individual dosimeter response.
In addition to irradiated control dosimeters, each group of TLDs is accompanied by three unirradiated control dosimeters during deployment and during return. These unirradiated controls are evaluated at the dosimetry laboratory to ensure that the TLDs did not receive any excess dose while either in transit or storage. The exposure received while either in storage or transit is typically negligible and thus is not subtracted.
An assessment of TLD data quality is based on the assumption that exposures measured at a fixed location will remain substantially constant over an extended period of time. A number of factors will combine to affect the certainty of measurements. The total uncertainty of the reported exposures is a combination of random and systematic components. The random component is primarily the statistical uncertainty in the reading of the TLD elements themselves. Based on repeated known exposures, this random uncertainty for the calcium sulfate elements used to determine exposure to tied environmental stations is estimated to be approximately * 3 to 5%. There are also several systematic'components of exposure uncertainty, including energy-directional response, fading, calibration, and exposures received while in storage. These uncertainties are estimated according to established statistical methods for propagation of uncertainty.
Accuracy and reproducibility of TLD processing of personnel dosimeters has been evaluated via the Department of Energy Laboratory Accreditation Program (DOELAP). This process concluded that procedures and practices utilized by the EPA R&IE-LV TLD Laboratory comply with standards published by the Department of Energy. This evaluation includes three rounds of blind performance testing over the range of 50 mrem to 500 rem and a comprehensive onsite assessment by DOELAP site assessors.
The DOELAP accreditation process requires a determination of the lower limit of detectability and veriiication that the TLD readers exhibit linear performance over the range included in the performance testing program. The lower limit of detectability (L, ) for the R&IE-LV TLD Laboratory has been calculated to be approximately 3 mrem above background at the 95% confidence level. See Appendbc A for L, calculations.

Data Management
The TLD data base resides on a Digital Equipment Corporation MicroVAX II directly connected to the two ~anasonic TLD readers. samples are tracked using field data cards and an issue data base tracking system incorporated into the reader control software. Two major software packages are utilized by the TLD network. The first, a proprietary package written and supported by International Science Associates, controls the TLD readers, tracks dosimeter performance, completes necessary calculations to determine absotbed dose equivalent, performs automated QAIQC functions, and generates raw data files and re~orts. The second software package, locally d&eloped, maintains privacy act infirmation and the identifying data, generates reports in a number of predefined formats, and provides archival storage of TLD results.

Pressurized Ion Chambers
The Pressurized Ion Chamber (PIC) Network continuously measures ambient gamma radiation exposure rates, and because of its sensitivii, may detect lowlevel exposures not detected by other monitoring methods. The primary function of the PIC network is to detect changes in ambient gamma radiation due to anthropogenic activities. In the absence of anthropogenic activities, ambient gamma radiation rates naturally dlfer among locations as rates vary with altitude (cosmic radiation) and with radioactivii in the soil (terrestrial radiation). Ambient gamma radiation also varies slightly within a location due to weather patterns, i.e., snow changes the amount of radonthoron released by the soil and detected by the PICs.

Network Design
There are 26 PlCs located in communities around the NTS and one in Mississippi, which provide near real-time estimates of gamma exposure rates. Two new stations were added to the network in the fourth quarter of 1996. They were Henderson and Boulder City, Nevada. The PIC at Boulder City was vandal-ized after only five days of data collection. Another site in Boulder City is being proposed to prevent future incidents. The locations of the PlCs for stations around the NTS are shown in Figure 3.2.
Because of the successful experience with the Citizen's Monitoring Program during the purging of the Three Mile Island containment in 1980, the Community Radiation Monitoring Program (CRMP) was begun. Because of reductions in the scope of monitoring, the CRMP was changed to the CTLP. It now consists of stations located in the states of Nevada and Utah. In 1996, there were 15 stations located in these two states. The CTLP is a cooperative project of the DOE, EPA, and DRI.
The DOUNV sponsors the program. The EPA provides technical and scientific direction, maintains the instrumentation and sampling equipment, analyzes the collected samples, and interprets and reports the data. The DRI administers the program by hiring the local station managers and alternates, securing rights-of-way, providing utilities, and performing additional q u a l i assurance checks of the data. Shown in Figure 3.2 are the locations of the CTLP stations.
Each station is operated by a local resident. In most cases, this resident is a high-school science teacher. Samples are analyzed at the R&IE Laboratory. Thirteen of the 15 CTLP stations have a low volume air sampler, a triiium and noble gas sampler on standby, and a TLD. The two stations recently setup have no tritium or noble gas sampler. In addition, a PIC and recorder for immediate readout of external gamma exposure and a recording barograph are located at the station. All of the equipment is mounted on a stand at a prominent location in each community. Residents may visit the stations and if interested, they can check the data. Also, computergenerated reports of the PIC data are issued monthly by EPA for each station.

Procedures
The PIC Network utilizes Reuter-Stokes models 1011,1012, and 1013 PICs. The PIC is a spherical shell filled with argon gas to a pressure 25 times that of atmospheric. In the center of the chamber is a spherical electrode with a charge opposite to the outer shell. When gamma radiation penetrates the sphere, ionization of the gas occurs and the ions are collected by the center electrode. The electrical current generated is measured, and the intensity of the radiation field is determined from the magnitude of this current.

Community Technical Liaison Program (CTLP) (15) Other PIC Locations) (1 1)
Data are retrieved from the PlCs shortly after measurements are made. The near real-time telemetrybased data retrieval is achieved by the connection of each PIC to a data collection platform which collects and transmits the data. Gamma exposure measurements are transmitted via the Geostationary Operational Environmental Satellite (GOES) directly to a receiver earth station at the NTS and from there to the R&IE-LV by dedicated telephone line. Each station routinely transmits data every four hours (i.e., 4-hour average, l-minute maximum, and l-minute minimum values) unless the gamma exposure rate exceeds the currently established alarm threshold. When the threshold is exceeded for two consecutive l-minute intervals, the system goes into the alarm mode and transmits a string of nine consecutive 1minute values every 2 to 15 minutes. Additionally, the location and status (i.e.,routine or alarm mode) of each station are shown on a map in the control room at the NTS and at R&IE-LV. Thus, the PIC Network is able to provide immediate documentation of radioactive cloud passage in the event of an accidental release of radioact~w.
The threshold limits are established at approximately two times background for each station location.
These threshold values range from 16 pFUh for Pahrump, Nevada, to 35 pWh for Milford, Utah, and Stone Cabin Ranch, Nevada. A significant improvement was made to the network in 1993. In previous years, 4-hour average, l-minute minimum, and 1minute maximum values were the only values transmitted every four hours. In 1993, the software at the stations was upgraded to allow a string of 48 fiveminute averages to be transmitted every four hours.
In addition to telemetry retrieval, PIC data are also recorded on magnetic tapes at 24 of the 27 EPA stations and on magnetic cards for the other three EPA stations. The magnetic tapes and cards, which are collected monthly, provide a backup to the telemetry data and are also useful for investigating anomalies in the data are recorded in smaller increments of time (5-minute averages). The PlCs also contain a liquid crystal display, permitting interested persons to monitor current readings.
The data are evaluated daily by R&IE-LV personnel. Trends and anomalies are investigated and equipment problems are identified and referred to field personnel for correction. Monthly averages are stored in Lotus files on a personal computer. These monthly averages are compiled from the 4-hour averages from the telemetry data, when available, and from the 5-minute averages from the magnetic tapes or cards when the telemetry data are unavailable. Computer-generated reports of the PIC monthly average data are issued monthly for posting at each station. These reports indicate the current month's average gamma exposure rate, the previous month's averages, and the maximum and minimum background levels in the U.S. Table 3.3 contains the number of monthly averages available from each station and the maximum, minimum, mean, standard deviation, and median of the monthly averages. The mean ranged from 8.0 ~R h r at Pahrump, Nevada, to 17.7 pRhr at Tonopah, Nevada, or annual exposures from 71 to 156 mR (18 to 40 @/kg). The table shows the total m Wyr (calculation based on the mean of the monthly averages) and the average gamma exposure rate for each station. Background levels of environmental gamma exposure rates in the U.S. (from the combined effects of terrestrial and cosmic sources) vary between 49 and 247 mFUyr (I3 to 64 pC/kg-yr) (BEIR 111, 1980). The annual exposure levels observed at each PIC station are well within these U.S. background levels. Figure 3.3 shows the distribution of the monthly averages from each PIC station. The horizontal lines extend from the mean value (+) to the minimum and maximum values. The vertical lines are the approximate U. S. background range.

Results
The data from Milford, Rachel, Twin Springs, and Uhalde's Ranch stations show the greatest range and the most variability. These data are within a few tenths pWhr from those of last year.

Quality AssuranceIQuality Control
General QAIQC guidelines for the PlCs follow the Quality Management Plan referenced on page 66 and are summarized as follows: Procedures for the operation, maintenance, and calibration, of PIC equipment and the data review, statistical analysis and records are documented in approved SOPS.
Radiation monitoring specialists place a radioactive source of a known exposure on the PlCs monthly to check the performance of the units.
Source check calibration and background exposure rate data are evaluated monthly and compared to historical values.
Data not transmitted via the telemetry system due to equipment failure are retrived by reading mag tapes.
A data quality assessment of the PIC data is given in Section 11, Quality Assurance.    Mean of total exposure is: 96.3mrem Total data completeness: 98%

Atmospheric Monitoring
The inhalation of radioactive airborne particles can be a major pathway for human exposure to radiation. The atmospheric monitoring networks are designed to detect environmental radiation from NTS and non-NTS activities. Data from atmospheric monitoring can determine the concentration and source of airborne radioactivity and can project the fallout patterns and durations of exposure to man. The only atmospheric monitoring network still operating is the Air Surveillance Network (ASN). The ASN was designed to monitor the areas within 350 kilometers (220 miles) of the NTS.
Most of the data collected from the ASN fall below the minimum detectable concentration (MDC).
Averages of data presented in this chapter were calculated including measured results below MDCs. All of the data collected from the atmospheric monitoring network reside on a VAX computer in the Sample Tracking Data Management System (STDMS).

Design
During 1996 the ASN consisted of 18 continuously operating sampling stations. Two stations were added during the fourth quarter to bring the total number of stations to 20 (see Figure 4.1 for these locations).
Each station is equipped with a low volume air sampler to collect particulate radionuclides on fiber filters and gaseous radioiodines in charcoal cartridges. The filters and charcoal cartridges receive complete analyses at the R&IE-LV Radioanalysis Laboratory. Duplicate air samples are collected from two ASN stations each week. The duplicate samplers operate at randomly selected stations continuously for three months and are then moved to a new location.
Six of the air sampling stations are equipped with high volume air samplers that collect particulate radionuclides on glass fiber filters. The filters are analyzed by gamma spectrometry in the R&IE-LV Radioanalysis Laboratory. The filters are then composited by month and analyzed for plutonium isotopes by wet chemistry methods. One duplicate high volume sampler is co-located at a randomly selected high volume sampling station and is moved to a new location at the beginning of each quarter. Duplicate samples are collected and analyzed by the same methods as the routine samples.

Procedures
Low volume samplers collect airborne particulates at each ASN station. The samples are collected as air is drawn through 5 cm (2.1 in) diameter, glass-fiber filters (prefilters) at a flow rate of about 100 m3 (2800 fl per day. Activated charcoal cartridges are placed directly behind the filters to collect gaseous radioiodines. Filters and cartridges are exchanged after sampler operation periods of about one week (approximately 560 m3 or 20,000 ff). High volume (hi-vol) samplers are located at selected stations within the ASN. The hi-vol samplers collect airborne particulates as air is drawn through an eight inch by ten inch glass fiber filter at a rate of about 2000m3 (58,000 ft3) per day. The hi-vol filters are collected monthly with a total volume of approximately 60,000 m3 (1,700,000 ft3).
Duplicate air samples are obtained weekly from selected stations. Two low volume air samplers and one high volume air sampler, which are identical to the ASN station samplers, are rotated between ASN stations quarterly. The results of the duplicate field sample analyses are given in Chapter 8 as part of the data quality assessment.
Prefilters and charcoal cartridges from low volume samplers, and high volume filters are initially analyzed by high resolution gamma spectrometry. The low volume prefilters are then analyzed for gross alpha and gross beta activity. Analysis is performed 7 to 14 days after sample collection to allow time for the decay of naturally occurring radon-thoron daughter products. Gross alphabeta analysis is used to detect trends in atmospheric radioactivity since it is more sensitive than gamma spectrometry for this purpose. High volume filters are submitted for wet chemistry analysis for plutonium isotopes upon completion of gamma spectrometry. Additional information on the analytical procedures is provided in Chapter 10.

Results
The ASN measures the major radionuclides which could potentially be emitted from activities on the NTS. Data from the ASN represents Afile of calibration records, control charts, I exposure to the general public. and log books is maintained.
Gamma spectrometry was performed on all ASN low and high volume samples. The majority of the samples were gamma-spectrum negligible (i.e., no gamma-emitting radionuclides detected). Naturally occurring ' Be, averaging 2.5 x 10-l3 pCVmL, was detected ocasionally by the low volume network of samplers. Beryllium-7 was detected consistently by the high volume sample method with average activity of 2.4 x l W 3 pCimL. Alpha and beta low volume air sample results were not included in data analysis 1 they met one or more of the following criteria: sampling duration of greater than 14 days, total volume of less than 400 m3, average flow rate less than 2.9 m3/hr or greater than 4.0 m3/hr, or power outage lasting more than one-third of sampling interval length. All remaining results were used in data analysis, including preparation of tables.
As in previous years, the gross beta results from the low volume sampling network consistently exceeded the analysis minimum detectable concentration (MDC). The annual average gross beta activity was 1.42 x l@14 PCifrnL. Summary gross beta results for the ASN are shown in Table 4.1.
Gross alpha analysis was performed on all low volume network samples. The average annual gross alpha activities were 1.32 x 10""Ci/mL. Summary gross alpha results for the ASN are presented in Table 4.2.
During the first three quarters of 1996, samples collected at 'high volume sampling sites were cornposited by month and analyzed for plutonium isotopes. Starting with the last quarter of 1996, the hi-vol samples were collected on a monthly basis to improve the ease of sample preparation in the laboratory. Due to a lower limit of detection for high volume sampling and analysis methods, environmental levels of plutonium were occasionally detected at all six of the sampling sites. Plutonium results for the high volume air sampling network are presented in Table 4.3.

Quality Assurance1 Quality Control
General QAIQC guidelines for the ASN are as follows: a Unique sample numbers are assigned.
The laboratory supervisor approves all analytical results before they are entered into the permanent data base. Gross beta analysis should be within k 20%.
Plutonium analysis of internal spikes should produce results within -+ 20% of the known value.
a The combined error due to both sampling and analytical technique is estimated by using replicates.
An estimate of bias (the dlference between the value obtained and the true or reference value) is determined by participating in intercomparison studies.
Further discussion of the QA program and the data quality assessment is given in Chapter 11. a All field sampling and laboratory instruments are calibrated and the date of calibration is marked on a decal affixed to the equipment.

Milk
Ingestion is one of the critical exposure pathways for radionuclides to humans. Food crops may absorb radionuclides from the soil in which they are grown. Radionuclides may be found on the surface of fruits, vegetables, or food crops. The source of these radionuclides may be atmospheric deposition, resuspension, or adhering particles of soil. Weather patterns, especially precipitation, can affect soil inventories of radionuclides. Grazing animals ingest radionuclides which may have been deposited on forage grasses and, while grazing, ingest soil which could contain radionuclides.
These radionuclides may be transferred to milk. Water is another significant ingestion transport pathway of radionuclides to humans.

Milk Surveillance Network
Milk is an important source for evaluating potential human exposures to radioactive material. It is one of the most universally consumed foodstuffs and certain radionuclides are readily traceable through the chain from feed or forage to the consumer. This is particularly true of radioiodine isotopes which, when consumed by children, can cause significant impairment of thyroid function. Because dairy animals consume vegetation representing a large area of ground cover and because many radionuclides are transferred to milk, analysis of milk samples may yield information on the deposition of small amounts of radionuclides over a relatively large area. The samples collected in July are from animals consuming local feed. Accordingly, milk is monitored by R&IE-LV through the Milk Surveillance Network (MSN).

Design
The MSN includes commercial dairies and familyowned milk cows and goats representing the major milksheds within 186 mi (300 km) of the NTS. The 11 locations comprising the MSN at the beginning of 1996 and any changes are shown in Figure 5.1. Samples were collected from only ten of these locations because the Hafen Ranch in Ivins, Utah, was not milking during the collection period.

Procedures
formaldehyde. This network was designed to monitor areas adjacent to the NTS, which could be affected by a release of activity, as well as from areas unlikely to be so affected.
All milk samples are analyzed by high-resolution gamma-ray spectroscopy to detect gamma-ray emitting radionuclides. These samples are also analyzed for "Sr andBOSr by radiochemical separation and beta counting (see Table 5.1).

Results
The average total potassium concentration derived from naturally occurring 40K activity was 1.5 g/L for samples analyzed by gamma spectrometry. All MSN milk samples were analyzed for "' Sr andBOSr, and the results are similar to those obtained in previous years. The MSN network average values are shown in Table 5.2 for "Sr and ' OSr.
In conclusion, the MSN data is consistent with previous years and is not indicative of increasing or decreasing trends. No radioactivity directly related to current NTS activiies was evident.

Quality Assurance/Control
General QAIQC guidelines for the MSN are as follows: Procedures for the operation, maintenance and calibration of laboratory counting equipment, the control and statistical analysis of the samples and the data review and records are documented in approved SOPS.
External and internal comparison studies were performed and field and internal duplicate samples were obtained for precision and accuracy assessments.
~nalytical results are reviewed for completeness and comparability.
Trends are identified and potential risks to humans and the environment are determined based on the data. The data quality assessment is given in Chapter 10.
Raw milk was collected in 3.8-L (1-gal) cubitainers from each MSN location in July and preserved with

Network Design
The LTHMP was instituted because AEC (later DOUNV) acknowledged its responsibility for obtaining and for disseminating data acquired from all locations where nuclear devices have been tested. The three objectives originally established for the LTHMP were to: Laboratory formerly the Environmental Monitoring Systems Laboratory in Las Vegas, Nevada (EMSL-Assure public safety.

LV). The LTHMP is only one component of the total
Inform the public, news media, and scientific community about any radiological contamination.
Document compliance with existing federal, state, and local antipollution requirements.
Another objective which has been incorporated into the LTHMP is to, where possible, detect trends in radionuclide activities which may be indicative of migration from test cavities.
The primary radionuclide analyzed in the LTHMP is tritium. As a product of nuclear weapons testing, high levels of tritium are found in test cavities. Because tritium can be incorporated into water molecules, it is expected to be the first radionuclide to migrate from a test cavity. Therefore, tritium serves as an indicator of radionuclide migration. Atmospheric tritium may also be deposited into water, primarily by precipitation. Tritium from this source is primarily found in surface waters, surficial aquifers, and springs closely connected to surficial aquifers.

.I Sampling Locations
In order to meet the objective of ensuring public safety, R&IE-LV monitors drinking water supply wells and springs around the NTS and in the vicinity of surface ground zero (SGZ) at the other locations. The majority of these sampling sites are privately owned and participation in the LTHMP is voluntary. Municipal drinking water supplies are also represented.
Regardless of the number of individuals served by a particular water supply, the National Primary Drinking Water Regulation1 (NPDWR) pertaining to radioactivity is used as the compliance standard.* All of the nuclear weapons tested at locations other than the NTS were emplaced at depths of greater than 1200 feet. Nuclear weapons tested on the NTS are also emplaced at great depths, with the exception of some shallow underground tests conducted in the early 1960s. The drinking water supply wells tap shallow aquifers and, consequently, do not represent groundwater in the geologic strata containing the test cavities. Therefore, wherever possible, deep wells are included in the monitoring program. These wells include some which were drilled soon after a nuclear test specifically to monitor activities in or near the test cavity and others which can be considered only as "targets of opportunity;" e.g., existing wells for which sampling permission has been obtained. Most of the deep wells tap nonpotable water sources. Monitoring design standards, such as those in the Resource Conservation and Recovery Act (RCRA), di d not become available until long after the LTHMP deep wells had been drilled. Cost has delayed emplacement of new wells, although a program to drill more than 90 new wells on the NTS was initiated in 1990. The sampling locations not associated with the NTS are defined by DOE as inactive hazardous waste sites and are exempt from the RCRA monitoring design requirements. Table 6.1 is a listing of routine sampling locations, on and offsite, where well water samples contained tritium concentrations greater than 0.2 * percent of the National Primary Drinking Water Standards.

Sampling and Analysis Procedures
The procedures for the analysis of samples collected for this report were described by Johns,et al. (1 979) and are summarized in Table 6.2 (see Table 6.3 for Typical MDA Values for Gamma Spectroscopy). These procedures include gamma spectral analysis and radiochemical analysis for tritium. The procedures were based on standard methodology. Two methods for tritium analysis were performed: conventional and electrolytic enrichment. The samples are initially analyzed by the conventional method. If the tritium result is less than 700 pCVL, selected samples are analyzed by the electrolytic enrichment method which lowers the minimum detectable concentration (MDC) from approximately 300 pCK to 5 pCilL. An upper level of 700 pCVL has been established for use of the tritium enrichment method. Sample cross contamination becomes a problem at higher ranges.
For wells with operating pumps, the samples are collected at the nearest convenient outlet. If the well has no pump, a truck-mounted sampling unit is used. With this unit it is possible to collect three-liter samples from wells as deep as 1,800 meters (5,900 ft). At the normal sample collection sites, the pH, conductivity, water temperature, and sampling depth is measured and recorded when the sample is collected.
The first time samples are collected from a well, 3H, 8 9 . 9 0~~ , 2 23g+240P~, and uranium isotopes are determined. At least one of the one gallon samples from each site is analyzed by gamma spectrometry. In late 1995, because there was no indication of migration and because of funding cutbacks, it was decided that only 25% of tritium samples collected would be analyzed by the enrichment method. Sampling locations in a position to show migration are usually selected.

Quality Assurance/Quality Control Samples
Sample collection and analysis procedures are described in standard operating procedures-(SOPS).
Data base management and data analysis activities are described in the Quality Assurance Program Plan (EPA, QAPP 1992).
Use of standardized procedures ensures comparability.of operations and data among monitoring locations and across temporal intervals.
Annual data quality assessments of precision, accuracy, and comparability are based on the results of quality assurancetquality control samples. The data quality assessment results for 1996 are given in Section 8.0.
Overall system precision is estimated from the results of field duplicates. A field duplicate is a second sample collected from a sampling location immediately following collection of the routine sample using identical procedures. Comparison of the R&IE-LV Radioanalysis Laboratory results to the mean for all laboratories provides an estimate of the comparability of results.
In addition to the above described QNQC samples which are used in annual data quality assessments, the Radioanalysis Laboratory employs a number of intemal QC samples and procedures to ensure data quality on a day-to-day basis. Internal QC samples include blanks, regular calibrations, matrix spike samples, and duplicate analyses (ga'mma spectroscopy only). If results of these internal QC samples fall outside prescribed control limits, analysis is stopped until the cause of the discrepant data is found and resolved and corrective actions are implemented.

Data Management and Analysis
A bar code pilot program for the LTHMP was completed in 1991, and was extremely successful. It has been expanded to other monitoring networks. Bar code labels are prepared prior to each sampling excursion, based on the sampling schedule prepared by the LTHMP Technical Leader. Upon receipt of samples m Sample Control, the bar code label was read and the information transferred into the Sample Tracking Data Management System (STDMS), along with information from the field data card.
Analysis data are entered into STDMS after they have been generated and reviewed by the analyst and Group Leader. Special software written in Fortran (referred to as "Chemistry Programs") is used for a majority of the radiochemical data reduction. The Chemistry Programs are used for calculating final data such as activii per unit volume, MDC, and 2-sigma error terms. All hand-entered data are checked for transcription errors. Once data is entered and checked, they are transferred from a "review" data base to a permanent data base, where further changes may be made only by authorized personnel.
Periodically, the assigned media expert reviews the data base and checks for completeness of sample collection, transcription errors, completion of sample analysis and QAlQC samples, and accuracy of information input. All discrepancies are resolved and corrected. Once the data base is complete for a given location, time series plots were generated.

Nevada Test Site Monitoring
The present sample locations on the NTS, or immediately outside its borders on federally owned land are shown in Figure 6.2. All sampling locations are selected by DOE and primarily represent potable water supplies. In 1995, sampling on the NTS was modified so that EPA only samples wells without pumps and, for Quality Assurance purposes, collects samples from 10 percent of the potable wells sampled by Bechtel Nevada. A total of 21 wells was scheduled to be sampled, but only 19 wells were sampled because the pumps were not working.
All samples were analyzed by gamma spectrometry and for tritium. No gamma-emitting radionuclides were detected in any of the NTS samples collected in 1996. Summary results of tritium analyses are given in Table 6.4. The highest average tritium activity was 4.5 x 104 pCYL (1,700 BqIL) in a sample from Well UE-5n. This activii is less than 60 percent of the DCG for tritium established in DOE Order 5400.5 for comparison with the dose limit (4 mrem) in the National Primary Drinking Water Regulations. Eight of the wells sampled yielded tritium results greater than the minimum detectable concentration (MDC). Well UE-7ns was routinely sampled between 1978 and 1987 and sampling began again in 1992. An increasing trend in tritium activity was evident at the time sampling ceased in 1987. Recent results have shown a decrease from those previous results, although the present result is higher than results for 1995.

Offsite Monitoring In The Vicinity Of The Nevada Test Site
The monitoring sites in the area around the NTS are shown h Figure 6.3. Most of the sampling locations represent drinking water sources for rural residents or public drinking water supplies for the communities in the area. The sampling locations include 12 wells, nine springs, and a surface water site. All of the locations are sampled quarterly or semiannually.
Gamma spectrometric analyses are performed on the samples when collected. No man-made gamma-emitting radionuclides were detected in any sample. Tritium analyses are performed on a semiannual basis. Adaven Spring is the only site which consistently shows detectable tritium activity. The triiium activity in this spring represents environmental levels that have been decreasing over time. All results for this project for 1996 are shown in Table 6.5. The sampling procedure is the same as that used for sites on the NTS and offsite areas (described in Section 6.1.2), with the exception that two 3.8-L samples are collected in cubitainers. The second sample serves as a backup or as a duplicate sample.

Hydrological Monitoring At Other United States Nuclear Device Testing Locations
Because of the variability noted in past years in samples obtained from the shallow monitoring wells near Project DRIBBLE ground zero (GZ), the sampling procedure was modlied several years ago. A second sample is taken after pumping for a specified p e t i d of time or after the well has been pumped dry and permitted to recharge. These second samples may be more representative of formation water, whereas the first samples may be more indicative of recent area rainfall.

Project FAULTLESS
Project FAULTLESS was a "calibration test" conducted on January 19, 1968, in a sparsely populated area near Blue Jay Maintenance Station, Nevada. The test had a yield of less than 1 Mt and was designed to test the behavior of seismic waves and to determine the usefulness of the site for highyield tests. The emplacement depth was 975 m (3,199 ft). A surface crater was created, but as an irregular block along local faults rather than as a saucer-shaped depression. The area is characteriied by basin and range topography, with alluvium overlying tuffaceous sediments. The working point of the test was in tuff. The groundwater flow is generally from the highlands to the valley and through the valley to Twin Springs Ranch and Railroad Valley (Chapman and Hokett, 1991).
Sampling was conducted on March 6 and 7,1996, at locations shown in Figure 6.4. Routine sampling locations include one spring and five wells of varying depths. The Bias Well was not sampled because the ranch was closed and Si-Mile Well was not sampled because the pump was removed. A new sampling location (site C Complex) was established to replace the Bias Ranch Well. This site is approximately 8 mi from Blue Jay Maintenance Station and is approximately 20 mi from surface ground zero (SGZ) .
At least two wells (HTH-1 and HTH-2) are positioned to intercept migration from the test cavity, should it occur (Chapman and Hokett, 1991). All samples yielded negligible gamma activity.
Tritium concentrations were less than the MDC. These results are all consistent with results obtained in previous years. The results for tritium indicate that, to date, migration into the sampled wells has not taken place and no event-related radioactivii has entered area drinking water supplies. Samples were collected on March 4 and 5, 1996. The sampling locations are shown in Figure 6.5. Only five of the seven routine wells were sampled. No sample was collected from Spring Windmill because the pump was removed. No sample was collected from Well H-2 because the well was locked and no key was available to EPA until after sampling was completed. This well will be sampled in the 1997 annual sampling. The routine sampling locations include one spring, one windmill, and five wells of varying depths. At least one location, Well HS-1, should intercept radioactivii migrating from the test cavity, should it occur (Chapman and Hokett, 1991).

Project SHOAL
Gamma-ray spectral analysis results indicated that no man-made gamma-emitting radionuclides were present in any samples above the MDC. All tritium results were also below the MDC. Sampling was conducted June 4-7, 1996, with collection of samples from eight out of nine wells in the area of Grand Valley and Rulison, Colorado. The spring 300 yards from SGZ was dry. Routine sampling locations are shown in Figure 6.6, including the Grand Valley municipal drinking water supply springs, water supply wells for five local ranches, and three sites in the vicinity of SGZ, including one test well, a surface-discharge spring which was dry, and a surface sampling location on Battlement Creek. Seven new monitoring wells were completed at the RULISON Site in 1995 as part of the Remedial Investigation and Feasibility Study. These wells will be added to the LTHMP in 1998.

Project RULISON
Triiium has never been observed in measurable concentrations in the Grand Valley City Springs. All of the remaining sampling sites show detectable levels of tritium, which have generally exhibited a stable or decreasing trend over the last two decades.

Scale in Kilometers
The range of tritium activity in 1996 was from 242 * 140 pCK (9 Bq/L) at Battlement Creek, to 11 2 * 6.9 pCVL (4.1 BqIL) at Lee Hayward Ranch. All values were less than one percent of the DCG. The detectable tritium activities were probably a result of the high natural background in the area. This was supported by the DRI analysis, which indicated that most of the sampling locations were shallow, drawing water from the surficial aquifer which was unlikely to become contaminated by any radionuclides arising from the Project RULISON cavity (Chapman and Hokett, 1991). All samples were analyzed for presence of gamma-ray emitting radiinuclides. None were detected above the MDC.

Project RIO BLANCO
Like Project RULISON, Project RIO BLANCO was a joint government-industry test designed to stimulate natural gas flow and was conducted under the Plowshare Program. The test was conducted on May 17, 1973, at a location between Rifle and Meeker, Colorado. Three explosives with a total yield of 90 kt were emplaced at 1780-, 1920-, and 2040-m (5838-, 6229-, and 6689-11) depths in the Ft. Union and Mesa Verde formations. Production testing continued to 1976 when cleanup and restoration activities were completed. Tritiated water produced during testing was injected to 1710 m (5610 ft) in a nearby gas well.
Samples were collected June 6 and 7, 1996, from the sampling sites shown in Figure 6.7. Only 13 of the 14 routine wells were sampled. No sample was collected from Brennan Windmill because the pump was inoperable. The sample taken from CER # l ; was lost in transit. The routine sampling locations included three springs and six wells. Three of the wells are located near the cavh and at least two of

Project GNOME
Project GNOME, conducted on December 10,1961, near Carlsbad, New Mexko, was a multipurpose test performed in a salt formation. A slightly more than 3-M nuclear explosive was emplaced at 371 m (1217 ft) depth in the Salado salt formation. Radioactive gases were unexpectedly vented during the test. The USGS conducted a tracer study in 1963, involving injection of 20 Ci 3H, 10 Ci 13' Cs, 10 Ci @%r , and 4Ci l3'I (740, 370, 370 and 150 GBq, respectively) into Well USGS-8 and pumping water from Well USGS-4. During cleanup activities in 1968-69, contaminated material was placed in the test cavity access well. More material was slurried into the cavity and driits in 1979.
Sampling at Project GNOME was conducted June 22-25, 1996. The routine sampling sites, depicted in Figure 6.8, include nine monitoring wells in the vicinity of GZ and the municipal supplies at Loving and Carlsbad, New Mexico. Stock tanks at wells PHs 8, PHs 9, and PHs 10, were sampled at the request of DOE. Tritium results from stock tank PHs 8 was greater than the MDC. The remaining two were below the MDC.
Tritium results greater than the MDC were detected in water samples from seven of the nine sampling locations in the immediate vicinity of GZ. Triiium activities in Wells DD-1, LRL-7, USGS-4, and USGS-8 rangedfrom 5x103 pCVL (185 Bq/L) in Well LRL-7 to 6.8 x lo7 pCin (2.5 MBqIL) in Well DD-1. Well DD-1 collects water from the test cavity; Well LRL-7 collects water from a sidedriit; and Wells USGS-4 and -8 were used in the radionuclide tracer study conducted'by the USGS. None of these wells are sources of potable water. the wells (Wells RB-D-01 and RB-D-03) were addition to tritium, concentrations were suitable for monitoring possible migration of i n samples from Wells DD-l (7.29 1 0~ radioactivity from the cavity. collected were above the MDC detectable tritium concentrations were PHs4 tritium and the rest were less than the MDC' The and -8, with results less than 0.02 percent of the tritium concentrations are well below 20.000 pCi5 D C G No tritium was detected in the remaining level defined in the EPA National Primary Drinking sampling locations, including Well USGS-l , which Water Regulations (40 C.F'R. 14'). All the DRI analysis (Chapman and Hoken, 1991) were for presence of gamma-ray indicated is to detect any migration of radionuclides, and none were detected. The tritium radioactivity from the cavity. concentrations were consistent with those collected previously at this site.  The principal aquifers near the test site are the Ojo Alamo sandstone, an aquifer containing nonpotable water located above the test cavity and the San Jose formation and Nacimiento formation, both surficial aquifers containing potable water. The flow regime of the San Juan Basin is not well known, although it is likely that the Ojo Alamo sandstone discharges to the San Juan River 50 mi northwest of the GASBUGGY site. Hydrologic gradients in the vicinity are downward, but upward gas migration is possible (Chapman and Hokett, 1991).
Sampling at GASBUGGY was conducted during June 1996. Only ten samples were collected at the designated sampling locations shown in Figure 6.9. The Bider Ranch has been sealed up and the pond north of Well 30.3.32.343N was dry.
The three springs sampling sites yielded tritium activities of 26 * 4.3 pCVL for Bubbling Springs, 43 * 4.0 pCVL for Cedar Springs, and 54 * 6.2 pCVL for Cave Springs (0.96, 1.6, and 2.0 BqIL, respectively), which were less than 0.2 percent of the DCG and similar to the range seen in previous years. Tritium samples from the three shallow wells were all below the average MDC.
Well EPNG 10-36, a gas well located 132 m (435 ft) northwest of the test cavity, with a sampling depth of approximately 1,100 m (3,600 ft), has yielded detectable tritium activities since 1984. The sample collected in June 1996 contained tritium at a concentration of 130 k 5.2 pCVL (4.8 BqR). The migration mechanism and route is not currently known, although an analysis by DRI indicated two feasible routes, one through the Printed Cliffs sandstones and the other one through the Ojo Alamo sandstone, one of the principal aquifers in the region. In either case, fractures extending from the cavity may be the primary or a contributing mechanism.
All gamma-ray spectral analysis results indicated that no man-made gamma-emitting radionuclides were present in any offsite samples. Tritium concentrations of water samples collected onsite and offsite are consistent with those of past studies at the GASBUGGY site.  (on April 19,1970). The ground surface and shallow groundwater aqulers were contaminated by disposal of drilling muds and fluids in surface pits. The radioactive contamination was primarily limited to the unsaturated zone and upper, nonpotable aquifers. Shallow wells, labeled HMH wells on Figure 6.10, have been added to the area near surface GZ to monitor this contamination. In addition to the monitoring wells near GZ, extensive sampling of water wells is conducted in the nearby offsite area as shown in Figure 6.1 1.

Project DRIBBLE
Of the twenty-eight wells that are sampled on the SALMON test site, five regularly have tritium values above those expected in surface water samples. In the 52 samples collected from offsite sampling locations, tritium activities ranged from less than the MDC to 28 pCVL (1.0 BqlL), 0.02 percent of the DCG. These results do not exceed the natural tritium activi expected in rainwater in the area. In general, results for each location were similar to results obtained. in previous years. Long-term decreasing trends in tritium concentrations are evident only for those locations that had detectable tritium activi at the beginning of the LTHMP, such as in the samples from the Baxterville City Well depicted in Figure 6.12 and Well HM-S shown in Figure 6.13.
Due to the high rainfall in the area, the normal sampling procedure is modiied for the shallow onsite wells as described in Section 6.4. Of the 32 locations sampled onsite (20 sites sampled twice), 14 yielded tritium activities greater than the MDC in either the first or second sample. Of these, eight yielded results higher than normal background (approximately 60 pCVL [2.2 BqIL]) as shown in  activities were measured generally correspond to areas of known contamination. Decreasing trends are evident for the wells where high tritium activiiies have been found, such as Well HM-S depicted in Figure 6.13. No tritium concentrations above normal background values were detected in any offsite samples. Man-made gamma-ray emitting radionuclides were not detected in any sample collected in this study.
Results of sampling related to Project DRIBBLE are discussed in greater detail in Onsite and Offsite Environmental Monitoring Report, "Radiation Monitoring around SALMON Test Site," Lamar County, Mississippi, April 1996(Davis 1996, available from R&IE-LV).

AMCHITKA ISLAND, ALASKA
Sampling is normally conducted biannually on odd years. The next sampling is scheduled for 1997.

Summary
None of the domestic water supplies monitored in the LTHMP in 1996 yielded tritium activiiies of any health concern. The greatest tritium activii measured in any water body which has potential to be a drinking water supply was less than one percent of the limit prescribed by the NPDWRs. In general, surface water and spring samples yielded tritium activities greater than those observed in shallow domestic wells in the same area. This is probably due to scavenging of atmospheric tritium by precipitation. Where suitable monitoring wells exist, there were no indications that migration from any test cavity is affecting any domestic water supply.
In most cases, monitoring wells also yielded no radionuclide activii above the MDC. Exceptions include wells into test cavities, wells monitoring known areas of contamination, and one well at GASBUGGY. Known areas of contamination exist at Project GNOME where USGS conducted a tracer study experiment, some areas onsite at Project DRIBBLE, and a few surface areas near Project LONG SHOT. The 1996 results for these monitoring wells are consistent with decreasing trends observed over time.
1. The NPDWR states that the sum of all betalgamma emitter concentrations in drinking water cannot lead to a dose exceeding 4 mremlyear, assuming a person were to drink two liters per day for a year (40 CFR 141). Assuming tritium to be the only radioactive contaminant yields a maximum allowable concentration of

HpGe
HpGe detector 100 Radionuclide concen-3.5L Varies with radio-Gamma" calibrated at 0.5 keV/ tration quantified from nuclides and detector channel (0.04 to 2 gamma spectral data used, see Table 6.3 MeV range) individual by online computer below. detector efficiencies program. ranging from 15 to 35 percent.
(a) The detection limit is defined as the smallest amount of radioactivity that can be reliably detected, i.e., probability of Type I and Type II error at 5 percent each (DOE 1981).
(b) Gamma spectrometry using a high purity intrinsic germanium (HpGe) detector.  Conventional andlor enrichment *tritium analysis techniques were used for the samples summarized in this table.
DCG Derived Concentration Guide; established by DOE Order as 90,000 pCi/L for water. NA Not applicable; percent of concentration guide is not applicable as the tritium result is less than the MDC or the water is known to be nonpotable. (a) For each sample: 1st row is from enrichment analysis, 2nd row from conventional analysis. Derived Concentration Guide (DCG) established by DOE Order as 90,000 pCi/L. NIA Not applicable. Percent of concentration guide is not applicable because the result is less than the MDC or the water is known to be nonpotable. (a) For each sample: 1st row is from enrichment analysis, 2nd row from conventional analysis. Derived Concentration Guide (DCG) established by DOE Order as 90,000 pCi1L. NIA Not applicable. Percent of concentration guide is not applicable because the result is less than the MDC or the water is known to be nonpotable.      (<MDC) Indicates results are less than MDC. ND Non-detected, no gamma radionuclides detected above MDC.

(<MDC) Indicates results are less than MDC (enriched and conventional method). ND
Non-detected, no gamma radionuclides detected above MDC.

Dose Assessment
There are several sources of possible radiation exposure to the population of Nevada which were monitored by EPA's offsite monitoring networks during 1996. The pathways are: Background radiation due to natural sources such as cosmic radiation, natural radioactivity in soil, and ' Be in air, anb H in water.
Worldwide distributions of radioactivity, such as eoSr in milk, "Kr in air, and plutonium in soil.
Airborne emissions and radioactive liquid discharges to onsite containment ponds. measured no radiation exposures attributed to recent NTS operations. However, using onsite emission measurements, estimates provided by U.S. Department of Energy (DOE) and calculated resuspension data as input to the EPA's CAP88-PC model, a potential effective dose equivalent (EDE) to the maximally exposed individual (MEI) was calculated to be 0. -mSv). These maximum dose estimates, excluding background, are less than one percent of The potential EDE to the offsite population due to NTS activities is estimated annually. Two methods are used to estimate the EDE to residents in the offsite area in order to determine the community potentially most impacted by airborne releases of radioactivity from the NTS. In the first method, effluent release estimates, based on monitoring data or calculated resuspension of deposited radioactivity, and meteorological data are used as inputs to EPA's CAP88-PC model which then produces estimated EDEs. The second method entails using data from the Offsite Radiological Safety Program (ORSP) with documented assumptions and conversion factors to calculate the committed effective dose equivalent (CEDE). The latter method provides an estimate of the EDE to a hypothetical individual continuously present outdoors at the location of interest that includes both NTS emissions and worldwide fallout. In addition, a collective EDE is calculated by the first method for the total offsite population residing within 80 km (50 mi) of each of the NTS emission sources. Background radiation measurements are used to provide a comparison with the calculated EDEs. In the absence of detectable releases of radiation from the NTS, the Pressurized Ion Chamber (PIC) network provides a measurement of background gamma radiation in the offsite area.
The extensive offsite environmental surveillance system operated around the NTS by EPA R&IE-LV the most restrictive standard.
Onsite source emission measurements, as provided by DOE, are listed in Table 7.1, and include tritium, radioactive noble gases, and plutonium. These are estimates of releases made at the point of origin. Meteorological data collected by the Air Resources Laboratory Special Operations and Research Division, (ARUSORD) were used to construct wind roses and stability arrays for the following areas: Mercury, Area 12, Area 20, Yucca Flat, and the Radioactive Waste Management Site (RWMS) in Area 5. A calculation of estimated dose from NTS effluents was performed using EPA's CAP88-PC model (EPA 1992). The results of the model indicated that the hypothetical individual with the maximum calculated dose from airborne NTS radioactivity would reside at Springdale, Nevada, 58 km (36 mi) west-northwest of CP-1. The maximum dose to that individual could have been 0.1 mrem (1 x 105 mSv). For comparison, data from the PIC monitoring network indicated a 1996 dose of 144 mrem (1.44 mSv) from background gamma radiation occurring in that area. The population living within a radius of 80 km (50 mi) from the airborne sources on the NTS was estimated to be 32,210 individuals, based on 1995 population data. The collective population dose within 80 km (50 mi) from each of these sources was calculated to be 0.3 person-rem (3 x 10" person-Sv). Activity concentrations in air that would cause these calculated doses are much higher than actually detected by the offsite monitoring network. For example, 0.107 mrem of thecalculated EDE to the ME1 is due to plutonium. The annual average plutonium concentration in air that would cause this EDE is 4.1 x IOl7 pCVmL. This is about 20 times the annual average plutonium in air measured in Goldfield (nearest community) of 0.19 x I0'l7 pCVmL (Chapter 4, Table 4.3). Table 7.2 summarizes the annual contributions to the EDEs due to 1996 NTS operations as calculated by use of CAP88-PC and the radionuclides listed in Table 7.1.
Input data for the CAP88-PC model included meteorological data from ARUSORD and effluent release data calculated from monitoring results and from resuspension estimates. These release data are known to be estimates and the meteorological data are mesoscale; e.g., representative of an area approximately 40 km (25 mi) or less around the point of collection. However, these data are considered sufficient for model input, primarily because the model itself is not designed for complex terrain such as that on and around the NTS. Errors introduced by the use of the effluent and meteorological data are small compared to the errors inherent in the model. The model results are considered overestimates of the dose to offsite residents. This has been confirmed by comparison with the offsite monitoring results.

Estimated Dose From ORSP Monitoring Network
Data Potential CEDEs to individuals may be estimated from the concentrations of radioactivity, as measured by the EPA monitoring networks during 1996. Actual results obtained in analysis are used; the majority of which are less than the reported minimum detectable concentration (MDC). No krypton or triiium in air data were collected offsite, so the onsite krypton for this year, and an average value for previous year's offsite tritium were used. No vegetable or animal samples were collected in 1996, so calculations for these intakes were not done.
The concentrations given in Table 7.3 are expressed i n terms of activii per unit volume. These concentrations are converted to a dose by using the assumptions and dose conversion factors described below. The dose conversion factors assume continuous presence at a fixed location and no loss of radioactivity in storage or handling of ingested materiils.
The algorithm for the dose calculation is: (concentration) x (assumption in volume/unit time) x (CEDE conversion factors) = CEDE As an example calculation, the following is the result of breathing tritium in air concentration of 0.2 pCVm3: Data quality objectives for precision and accuracy (2 x 10 " pCVm3) x (8400 m3/yr) x (6.4 x 10'' are, by necess'ky, less stringent for values near the mremIpCi) = 1 .I x mrem/yr MDC, so confidence intervals around the input data are broad. The concentrations of radioactivity However, in calculating the inhalation CEDE from 3H, detected by the monitoring networks and used in the the value is increased by 50 percent to account for calculation of potential CEDEs are shown in Table  absorption through the skin (ICRP, 1975). The total 7 n dose in one year, therefore, is 1 .l x 10.' x 1.5 = 2.4 x lo4 mreml yr. Dose calculations from ORSP data are summarized in Table 7.3.
The individual CEDES, from the various pathways,

Summary
added together give a total of 0.015 mrem/yr. Total

EDEs can be based On
The offsite environmental surveillance system combinations of data. If the interest was in just one area, for example, the concentrations from those operated around the NTS by EPA's R&IE-LV stations closest to that area could be substituted into detected no radiological exposures that could be the equations used here. attributed to recent NTS operations, but a calculated EDE of 0.015 mrem can be obtained, if certain In 1996, because of budget cuts and the standby status of nuclear device testing, samples of game animals and garden vegetables were not collected. Also, the noble gas and tritium sampling network was discontinued in the offsite locations, and the air sampling network was reduced. In order to calculate an EDE for a resident of Springdale, the ME1 from the CAP88-PC operation, it is necessary to make some assumptions. The NTS average krypton-85 concentration is representative of statewide levels; tritium in air does not change significantly from year to year; and, because Goldfield has the nearest air sampler to Springdale, its plutonium concentration is used to calculate the EDE.

Dose from Background
assumptions are made, as shown in Table 7.2. Calculatiori with the CAP88-PC model, using estimated or calculated effluents from the NTS during 1996, resulted in a maximum dose of 0.11 mrem (1.1 x 10" mSv) to a hypothetical resident of Springdale, Nevada, 14 km (9 mi) west of the NTS boundary. Based on monitoring network data, this dose is calculated to be 0.005 mrem. This latter EDE is about 5 percent of the dose obtained from CAP88-PC calculation. This maximum dose estimate is less than one percent of the International Commission on Radiological Protection (ICRP) recommendation that an annual EDE for the general public not exceed 100 mremtyr (ICRP 1985). The calculated population dose (collective EDE) to the approximately 32,210 residents living within 80 km (50 mi) of each of the NTS airborne emission -Radiation sources was 0.34 person-rem (3.4 x lo9 person-Sv).
Background radiation yielded a CEDE of 3,064 person-rem (30.6 person-Sv). In addition to external radiation exposure due to cosmic rays and gamma radiation'from naturally occurring radionuclides in soil (e.g., 'OK, U, and Th and their progeny), there is a contribution from 7Be that is formed in the atmosphere by cosmic ray interactions with oxygen and nitrogen. The annual average 7Be concentration measured by the offsite surveillance network was 0.24 pCiJm3. With a dose conversion factor for inhalation of 2.6 x mremtpci, and a breathing volume of 8,400 m3/yr, this equates to a dose of 5.2 x l o 4 mrem as calculated in Table 7.3. This is a negligible quantity when compared with the PIC network measurements that vary from 73 to 156 mWyear, depending on location.
Data from the PIC gamma monitoring indicated a 1996 dose of 144 mrem from background gamma radiation measured in the Springdale area. The CEDE calculated from the monitoring networks or the model, as discussed above, is a negligible amount by comparison. The uncertainty (20) for the PIC measurement at the 144 mrem exposure level is approximately five percent. Extrapolating to the calculated annual exposure at Springdale, Nevada, yields a total uncertainty of approximately 7 mrem which is greater than either of the calculated EDEs. Because the estimated dose from NTS activities is less than 1 mrem (the lowest level for which Data Quality Objectives (DQOs) are defined, as given in Chapter 10) no conclusions can be made regarding the achieved data quality as compared to the DQOs for this insignificant dose.

Training Program
Proper and efficient performance of radiological health functions by qualified personnel is required to ensure protection from radiological hazards. The purpose of the training program is to provide welltrained, qualified personnel to safely and efficiently perform their assigned duties at a predetermined level of expertise.

Emergency Response
Training Program , Emergency response training is essential to maintain a cadre of personnel who are qualified to perform approved radiological health and field monitoring practices.  The HAZMAT Spill Center has the advantages of being located far from populated areas, inside of a secure facility, and subjected to well characterized and predictable meteorological conditions. The EPA provides a chemist to participate in meetings of the Advisory Panel which reviews and approves all programs prior to testing and maintains a readiness for monitoring emissions at the boundary of the NTS. Recent spills have involved such small amounts of material that monitoring at the boundary was not justified. Dispersion models show that even a catastrophic release of the entire supply of the test materials would not be measurable at the test site boundary. Four spill programs were conducted in 1996. These included a U.S. Navy incinerator test, testing a variety of laser sensors, and a spill mitigation workshop conducted by an industrial user.

Sample Analysis Procedures
The procedures for analyzing samples collected for this report are described in Radiochemical and Analytical Procedures for Analysis of Environmental Samples (Johns, 1979) and are summarized in Table 9.1 and (see Table 6.2 page 52). These include gamma analysis, gross beta on air filters, strontium, triiium, plutonium, and noble gas analyses. These procedures outline standard methods used to perform given analytical procedures.  Table 6.3, page 52.
' The detection limit is defined as the smallest amount of radioactivity that can be reliably detected, i.e., probability of Type I and Type II error at 5 percent each (DOE81). Gamma spectrometry using a high purity intrinsic germanium (HpGe) detector. Depending on sample type.

Policy
One of the major goals of the EPA is to ensure that all agency decisions which are dependent on environmental data are supported by data of known quality. EPA Order 5360.1, "Policy and Program Requirements to Implement the Quality Assurance Program" requires participation in a QA Program by all EPA organizational units involved in environmental data collection. This policy further requires participation in a centrally managed QA Program by all EPA Laboratories, Program Offices, Regional Offices, and those monitoring and measurement efforts supported or mandated through contracts, regulations, or other formalized agreements.
The QA policies and requirements of EPA's R&IE-LV are summarized in the Quality Management Plan (R&IE, draft 1997). Policies and requirements specific to the ORSP are documented in the Quality

Assurance Program Plan for the Nuclear Radiation Assessment Division Offsite Radiation Safety
Program (EPA, 1992, under revision). The requirements of these documents establish a framework for consistency in the continuing application of quality assurance standards and procedures in support of the ORSP. Administrative and technical procedures based on these QA requirements are maintained in appropriate manuals or are described in SOPs. It is R&IE policy that personnel adhere to the requirements of the QA Plan and all SOPs applicable to their duties to ensure that all environmental radiation monitoring data collected by R&IE in support of the ORSP are of adequate quality and properly documented for use by the DOE, EPA, and other interested parties.

Data Quality Objectives
Achieved data quality is monitored continuously through internal QC checks and procedures. In addition to the internal QC procedures, R&IE participates in external intercomparison programs. One such intercomparison program is managed and operated by a group within EPAICRD-LV. These extemal performance audits are conducted as described in and according to the schedule contained in "Environmental Radioactivii Laboratory Intercomparison Studies Program" (EPA, 1992a). The analytical laborato,ry also participates in the DOE Environmental Measurements Laboratory (EML) Quality Assurance Program in which real or synthetic environmental samples that have been prepared and thoroughly analyzed are distributed to participating laboratories. The R&IE laboratory also began participation in the DOE Mixed Analyte Performance Evaluation Program (MAPEP) during 1996. External Dosimetry is accredited every two years. In 1996 the program was accredited under the Department of Energy Accreditation Program (DOELAP). Accreditation includes performance testing as well as an on-site assessment.
The R&IE External Dosimetry Program is currently seeking National Voluntary Laboratory Accreditation Program (NVLAP) accreditation, which will also include performance testing and an on-site assessment.

Representativeness, Comparability, and Completeness Objectives
Representativeness is defined as "the degree to which the data accurately and precisely represent a characteristic of a parameter, variation of a property, a process characteristic, or an operation condition" (Stanley and Vemer, 1985). In the ORSP, representativeness may be considered to be the Data quality are statements of the degree to:,hich the collected samples represent the quality of data a'decision maker needs to ensure that radionuclide act Mi concentrations i n the offsite a decision based on that data is defensible. Data environment. Collection of samples representative quality objectiies are defined in terms of representof pathways to human exposure as well as direct ativenessl comparability, completeness, precision, measurement of offsite resident exposure through and accuracy. and ' Omparahe TLD monitohg programs provides assurance of bilit~ are generally qualitative assessments h e representativeness of h e calculated exposures. completeness, precision, and accuracy may be quantitatively assessed. In the ORSP, represent-Comparability i s defined as ,,the confidence wfih ativeness, comparability, and completeness objecwhich one data set can be to are defined for each monitoring (StanleyandVemer, 1985). Comparabilityof data is Precision and accuracyare defined for each analysis assured by use of for sample collection, type or radionuclide. handling, and analysis; use of standard reporting units; and use of standardized procedures for data analysis and interpretation. In addition, another aspect of comparability is examined through longterm comparison and trend analysis of various radionuclide activity concentrations, and TLD, and PIC data. Use of SOPs, maintained under a document control system, is an inportant component of comparability, ensuring that all personnel conform to a unified, consistent set of procedures.
Completeness is defined as "a measure of the amount of data collected from a measurement process compared to the amount that was expected to be obtained under the conditions of measurement" (Stanley and Vemer, 1985). Data may be lost due to instrument malfunction, sample destruction, loss in shipping or analysis, analytical error, or unavailability of samples. Additional data values may be deleted due to unacceptable precision, accuracy, or detection limit or as the result of application of statistical outlier tests. The completeness objective for all networks except the LTHMP is 90%. The completeness objective for the LTHMP is 80%; a lower objective has been established because dry wells or access restrictions occasionally preclude sample collection.

Precision and Accuracy Objectives of Radioanalytical Analyses
Measurements of sample volumes should be accurate to k 5% for aqueous samples (water and milk) and to * 10% for air and soil samples. The sensitivity of radiochemical and gamma spectrometric analyses must allow no more than a 5% risk of either a false negative or false positive value. Precision to a 95% confidence interval, monitored through analysis of duplicate and blind samples, must be within * 10% for activities greater than 10 times the minimum detectable concentration (MDC) and * 30% for activities greater than the MDC but less than 10 times the MDC. There are no precision requirements for activity concentrations below the MDC, which by definition cannot be distinguished from background at the 95% confidence level. Control limits for accuracy, monitored with matru spike samples, are required to be no greater than * 20% for all gross alpha, gross beta, and gamma spectrometric analyses, depending upon the media type.
At concentrations greater than 10 times the MDC, precision is required to be within * 10% for: Conventional Tritium Analyses Uranium Thorium (all media) Strontium and within * 20% for: Enriched Tritium Analyses Strontium (in milk) Plutonium.
At concentrations less than 10 times the MDC, both precision and accuracy are expressed in absolute units, not to exceed 30% of the MDC for all analyses and all media types.

Quality of Dose Estimates
The allowable uncertainty of the effective dose equivalent to any human receptor is * 0.1 mrem annually. This uncertainty objective is based solely upon the precision and accuracy of the data produced from the surveillance networks and parameter uncertainties does not apply to uncertainties in the model used, effluent release data received from DOE, or dose conversion factors. Generally, effective dose equivalents must have an accuracy (bias) of no greater than 50% for annual doses greater than or equal to 1 mrem but less than 5 mrem and no greater than 10% for annual doses greater than or equal to 5 mrem.

Data Validation
Data validation is defined as "A systematic process for reviewing a body of data against a set of criteria to provide assurance that the data are adequate for their intended use." Data validation consists of data editing, screening, checking, auditing, verification, certification, and review 1983). Data validation procedures are documented in SOPs. All data are reviewed and checked at various steps in the collection, analysis, and reporting processes. 1. Completenessall samples scheduled to be collected have, in fact, been collected and analyzed or the data base contains documentation explaining the reasons for noncollection or nonanalysis.
2. Transcription errorschecks are made of all manually entered information to ensure that the information contained in the data base is accurate.
3. Quality control datafield and analytical duplicate, audit sample, and matrix blank data are checked to ensure that the collection and analytical processes are within specified QC tolerances.
4. Analysis scheduleslists of samples awaiting analysis are, generated and checked against normal analysis schedules to idently backlogs in analysis or data entry.

Unidentified malfunctionssample results
and diagnostic graphics of sample results are reviewed for reasonableness. Conditions indicatiie of instrument malfunction are reported to CERMEWCRQA.
Once the data base has been validated, the data are compared to the DQOs. Completeness, accuracy, and precision statistics are calculated. The achieved quality of the data is reported at least annually. If data fail to meet one or more of the established DQOs, the data may still be used in data analysis; however, the data and any interpretive results are to be qualified.
All sample results exceeding the natural background activi range are investigated. If data are found to be associated with a non-environmental condition, such as a check of the instrument using a calibration source, the data are flagged and are not included in calculations. Only data verified to be associated with a non-environmental condition are flagged; all other data are used in calculation of averages and other statistics, even if the condition is traced to a source other than the NTS (for example, higher-than-normal activities were observed for several radionuclides following the Chemobyl accident). When activities exceeding the expected range are observed for one network, the data for the other networks at the same location are checked. For example, higher-than-normal-range PIC values are compared to data obtained by the air or TLD samplers at the same location.
Data are also compared to previous years' data for the same location using trend analysis techniques.
Other statistical procedures may be employed as warranted to permit interpretation of current data as compared to past data. Trend analysis is made possible due to the length of the sampling history, which in some cases is 30 years or longer.

Quality Assessment Of 1996 Data
Data quality assessment is associated with the regular QA and QC practices within the radioanalytical laboratory. The analytical QC plan, documented in SOPS, describes specific procedures used to demonstrate that data are within prescribed requirements for accuracy and precision. Duplicate samples are collected or prepared and analyzed in the exact manner as the regular samples for that particular type .of analysis. Data obtained from duplicate analyses are used for determining the degree of precision for each individual analysis. Accuracy is assessed by comparison of data from spiked samples with the "true" or accepted values. Spiked samples are either in-house laboratory blanks spiked with known amounts of radionuclides, or QC samples prepared by other organizations in which data are compared between several laboratories and assessed for accuracy. Achieved data quality statistics are compiled on a quarterly and annual basis. This data quality assessment is performed as part of the process of data validation, described in Section 10.3. The following subsections describe the achieved data quality for 1996.

Completeness
Completeness is calculated as: where: %C = percent completeness V = number of measurementsjudgedvalid n = total number of measurements The percent completeness of the 1996 data is given in Ninety-one percent of the total possible number of samples were collected from ten ranches (see Figure 4.1). Annual means for these locations, individually, cannot be considered to be representative of the year. However, milk collected in July is representative of cows grazing on pasture or fed green chop which represent the typical food chain for those areas. The Hafen Ranch in Ivins, UT was not sampled as they were not milking during the collection period and there was no alternate sampling site in the area.
The achieved completeness of over 76 percent for the PIC Network fails to meet the DQO of 90 percent. This completeness value represents satellite telemetry data only, which is used for reporting purposes.
Gaps in the satellite transmissions are filled by data from the magnetic tape or card media. The redundant data systems used in the PIC Network (i.e., magnetic tape or card data acquisition systems) are responsible for high rates of recovery of the collected data, and are stored electronically for reference.

Precision
Precision is monitored through analysis of duplicate samples. Field duplicates (i.e., a second sample collected at the same place and time and under the same conditions as the routme sample) are collected in the ASN, LTHMP, and MSN. For the ASN, a duplicate sampler is collocated with the routine sampler at randomly selected sites for a period of. three months to provide the field duplicate. A total of two samplers are used for low volume sample duplicates and one sampler is used for a duplicate high volume sample. The duplicate samplers are moved to randomly selected sampling sites throughout the year. Approximately ten percent of samples submitted to the laboratory are analyzed twice for intra laboratory comparison whenever possible. In lieu of field duplicates, precision for the PlCs is determined by the variance of measurements over a specific time interval when only background activities are being measured. Precision may also be determined from repeated analyses of routine or laboratory spiked samples. The spiked QC'samples are generally not blind to the analyst; i.e., the analyst both recognizes the sample as a QC sample and knows the expected (theoretical) actiiity of the sample.
Precision is expressed as percent relative standard deviation (%RSD), also known as coefficient of variation, and is calculated by: %RSD = ( std. dev.) , o0 mean The precision or %RSD (also called Coefficient of Variation) is not reported for duplicate pairs in which one or both results are less than the MDC of the analysis. For most analyses, the Measurement Quality Objectives (MQOs) for precision are defined for two ranges: values greater than or equal to the MDC but less than ten times the MDC and values equal to or greater than ten times the MDC. The %RSDs is partially dependent on statistical counting uncertainty so it is expected to be more variable for duplicate analyses of samples with low activities.
From duplicate samples collected and analyzed throughout the year, the %RSD was calculated for varioustypes of analyses and sampling media. The results of these calculations are shown in Table 10.2. Samples not meeting the precision MQO were low activity, air particulate samples in which ' Be was detected. The precision data for all other analyses were well within their respective MQOs.' The R&IE data presented in Table 10.2 includes only those duplicate pairs that exceeded the minimum detectable concentration (MDC).
A total of 161 low volume duplicate pairs was analyzed for gross alpha and gross beta. Field duplicates account for sixty-nine of the samples and ninety-two were laboratory duplicates.
A total of 84 duplicate pairs exceeded the analysis MDC for gross alpha. Twenty-six of these were field duplicates and fifty-eight were laboratory duplicates. Of the field duplicates, ten of the twenty-six exceeded the MQO of 30 percent for samples greater than MDC but less than ten times MDC. One of the field duplicate samples exceeded ten times the MDC and the RSD for that,sample was zero percent. Of the fifty-eight laboratory duplicates, nineteen exceeded the MQO of thirty percent. None of the laboratory duplicates were greater than ten times the MDC. Sixty-seven of the sixty-nine field duplicates exceeded the analysis MDC for gross beta. Of these, three were greater than ten times the MDC. The average RSD for the pairs greater than ten times MDC was 13.2 percent, exceeding the MQO of 10 percent for samples greater than ten times MDC. All three samples had RSDs of less than 15 percent.

Estimated Precision, %RSD
The average RSD for the sixty-four pairs greater than MDC but less than ten times MDC was 19.9 percent, well below the MQO of thirty percent for the analysis.
Ten of the sixty-four samples exceeded the MQO. Of ninety-two laboratory duplicate pairs, five were greater than ten times the MDC. The average RSD for these five samples was 3.5 percent with all samples less than the MQO of 10 percent. Eightyfour samples were greater than the MDC but less than ten times MDC for the analysis. The average RSD for this group of samples was 15.5 percent, well below the MQO of 30 percent. Eight of the samples exceeded the MQO value.'Be was detectable on 25 low volume duplicate pairs. Eleven were field duplicates and 14 were laboratory duplicates. The average RSD of 31.4 percent is above the precision MQO of 30 percent for samples above MDC and less than ten times MDC. Of the eleven field duplicates, the average RSD was 29.8 percent which meets the MQO. The average RSD for the laboratory duplicates was 32.7 percent. Eight duplicate pairs from the field samples and 11 of the duplicate pairs from the laboratory samples were less than the MQO of 30 percent. High volume duplicate pairs where ' Be was detected did not meet the MQO. The average of 11 samples was 46.8 percent. Four of the eleven samples met the MQO of 30 percent.
Forty-two duplicate pairs were analyzed for tritium using the unenriched method. Of the 42 samples analyzed, two were above the MDC for the analysis. The average RSD for these two samples was 26.2 percent which meets the MQO for this type of analysis. A total of 25 samples was analyzed for tritium using the enrichment method. Five of the duplicate pairs were above ten times MDC for the analysis with an average RSD of 7.1 percent, within the MQO of 10 percent for the analysis. Seven duplicate pairs were greater than MDC and less than ten times MDC, wlh RSD of 8.6 percent which is well within the MQO of 20 percent for this type of analysis.

Accuracy
The accuracy of all analyses is controlled through the use of NIST-traceable standards for instrument calibrations. Internal checks of instrument accuracy may be periodically performed, using spiked matrix samples. These internal QC procedures are the only control of accuracy for Pressurized ton Chambers. For spectroscopic and radiochemical analyses, an independent measurement of accuracy is provided by participation in htercomparison studies using samples of known activities. The EPA R&IE-LV Radioanalysis Laboratory participates in three such intercomparison studies.
In the EPA CRDmADQA Intercomparison Study program, samples of known activities of selected radionuclides are sent to participating laboratories on a set schedule throughout the year. Water, milk, and air filters are used as the matrices for these samples. Results from all participating laboratories are compiled and statistics computed comparing each laboratory's results to the known value and to the mean of all laboratories. The comparison to the known value provides an independent assessment of accuracy for each participating laboratory. Approxjmately 20 laboratories participate in the EML performance evaluation program. The MAPEP program evaluates the performance of approximately forty laboratories. Sample matrices for both of these programs include water, air filters, vegetation, and soil. Results for these performance audit samples are given in Tables 10.5 and 10.6. One of the two EML studies for 1996 was reported outside of acceptable limits for gamma spectroscopy in both air and water matrices.
Follow-up investigation established a volume data entry error in both cases. Corrective actions were implemented.
In addition to use of irradiated control samples in the processing of TLDs, DOELAP and NVLAP both monitor accuracy as part of their accreditation program. As with the intercomparison studies, samples of known activii are submitted as single blind samples. The designation "single blind" indicates the analyst recognizes the sample as being other than a routine sample, but does not know the concentration or activity contained in the sample. Individual results are not provided to the participant laboratories by DOELAP or NVLAP; issuance of the accreditation certificate indicates that acceptable accuracy reproducibility has been achieved as part of the performance testing process and that an onsite independent review has indicated conformance with established accreditation standards.

Comparability
The EPA Performance Evaluation Program provides results to each laboratory participating in each study that includes a grand average for all values, excluding outliers. of this statistic (in multiples of standard normal deviate, unlless) lies between control limits of -3 and +3, the accuracy (deviation from known value) or comparability (deviation from grand average) is within nonnal statistical variation. Table 10.4 displays data from the 1996 intercomparison studies for all variables measured. There were five instances in which the EPA R&IE-LV Radioanalysis Laboratory results deviated from the grand average by more than three standard normal deviate units. All of these were gamma spectrometry analyses of the June gamma in water intercomparison study sample.
After investigation of the error in the reported data for this sample, it was found that the initial dilution of the sample had been improperly performed. The sample data was recalculated by the laboratory using the proper values for dilution which provided satisfactory analytical results for 4 of the 5 analytes in question. All other analyses were within three standard normal deviate units of the grand mean. This indicates acceptable comparability of the Radioanalysis Laboratory with the 70 to 290 laboratories participating in the EPA lntercomparison Study Program.

Representativeness
Representativeness cannot be evaluated quantitatively. Rather, it is a qualitative assessment of the ability of the sample to model the objectives of the program. The primary objective of the ORSP is to protect the health and safety of the offsite residents. Therefore, the DQO of representativeness is met if the samples are representative of the radiation exposure of the resident population. Monitoring stations are located in population centers. Siting criteria speclic to radiation sensors are not available for many of the instruments used. Existing siting criierii developed for other pollutants are applied to the ORSP sensors as available. For example, siting criteria for the placement of air sampler inlets are contained in Prevention of Significant Deterioration guidance documents (EPA, 1976). Inlets for the air samplers at the ORSP stations have been evaluated against these criteria and, in most cases, meet the siting requirements. Guidance or requirements for handling, shipping, and storage of radioactkity samples are followed in program operations and documented in SOPs. Standard analytical methodology is used and guidance on the holding times for samples, sample processing, and results calculations are followed and documented in SOPs.
A normalized deviation statistic compares each laboratory's result (mean of three replicates) to the known value and to the grand average. If the value In the LTHMP, the primary objectives are protection of drinking water supplies and monitoring of any potential cavity migration. Sampling locations are primary "targets of opportunity", i.e., the sampling locations are primarily wells developed for purposes other than radioactiviiy monitoring. Guidance or requirements developed for Comprehensive Environmental Response, Compensation, and Liability Act and Resource Conservation Recovery Act regarding the number and location of monitoring ' wells have not been applied to the LTHMP sampling sites. In spite of these limitations, the samples are representative of the first objective, protectbn of drinking water supplies. At all of the LTHMP monitoring areas, on and around the NTS, all potentially impacted drinking water supplies are monitored, as are many supply sources with virtually no potential to be impacted by radioactivity resulting from past or present nuclear weapons testing. The sampling network at some locations is not optimal for achieving the second objective, monitoring of any migration of radionuclides from the test cavities. An evaluation conducted by DRI describes, in detail, the monitoring locations for each LTHMP location and the strengths and weaknesses of each monitoring network (Chapman and Hokett, 1991). Corrective actions are dependent upon DOE funding of new wells. This evaluation is cited in the discussion of the LTHMP data in Section 6.