by Richard Kelley & Marinea Mehrhoff

Abstract
Radon-222 analyses were completed on samples collected from 153 public water supplies using a single hydrogeologic formation for their source of drinking water. Samples were collected from both the source and the distribution systems. The analyses suggest that over 50% of the ground water sources are naturally high (over 300 pCi/l) in radon-222. Very few sources (4 samples) exceeded 1000 pCi/l. Although the majority of groundwater sources were high in radon, only 28% of the public water supplies in the study exceeded the proposed drinking water standard of 300 pCi/l in samples collected from the distribution systems.

Acknowledgments
The authors would like to express their appreciation to Paul Van Dorpe of the Iowa Department of Natural Resource, Geological Survey Bureau; and, Roy Ney of the Iowa Department of Natural Resources, Drinking Water Section, for their assistance in site selection and review of this report.

This study was funded through a grant from the Iowa Department of Natural Resources and completed under contract, 93-7151-01.

Introduction
The US Environmental Protection Agency has proposed a radon standard for public water supplies of 300 pCi/L. Recent investigations by the U.S. Geological Survey (USGS) and the University Hygienic Laboratory (UHL) suggests that a number of public water supplies may expect to be in violation of the proposed standard if adopted. Further, investigations in Iowa suggest that even water supplies using shallow geologic formations as a source of their water may be expected to violate the proposed radon standard. The University Hygienic Laboratory, working with the Iowa Department of Natural Resources (IDNR) and local public water supplies undertook a study of public water supplies using ground water as their primary source of drinking water to assess the probable extent of elevated radon levels in both the source and finished water of public water supplies in Iowa.

The UHL staff worked with the staff of the IDNR to determine appropriate sampling locations across the state. A total of 150 public water supplies were targeted for inclusion in the study. These supplies were selected to reflect both shallow and deep bedrock sources of drinking water. The UHL provided proper sample collection kits and instructions to the public water supply operators participating in the study. The operators of the public water supplies collected the samples and returned them to the UHL in the prepaid mailers provided by the laboratory. Two samples were collected at each public water supply: one sample from the source (raw water) and one sample from the distribution system (finished water).

Sampling locations
Sampling locations were determined by the staffs of the UHL and the IDNR. Only public water supplies using a single hydrogeologic source were considered for participation. The Iowa Municipal Water-Supply Inventory (MWSI), maintained by the Iowa Department of Natural Resources, was used for selecting the water supplies to be included in the study.

A total of 493 public water supplies were identified as using a single hydrogeological source. These water supplies were then divided by their source into the following categories: Alluvium, Drift, Dakota or Cretaceous, Pennsylvanian, Mississippian, Silurian/Devonian, Ordovician (above the St. Peter), Cambrian/Ordovician, Cambrian (below the St. Lawrence) or Precambrian. The water supplies were then listed alphabetically under the appropriate hydrogeological source.

The number of water supplies selected from each hydrogeological source was proportional to the percentage of the municipal population using that source as their primary source of drinking water. For each hydrogeologic source, a random starting number between one and the number of water supplies using the source was generated. This number represented the position of the first public water supply on the list to be selected for inclusion in the study. A systematic count number (the number of sites divided by the starting number) was then generated. This number represented the number of positions on the list between supplies selected for inclusion in the study. For example, the random starting number identified position 43 on the list of public water supplies using alluvial sources as the first supply to be included in the study. The systematic count number produced an incremental value of two. Thus, supplies listed in positions 45, 47, 49, etc. were selected to be included in the study.

An inventory of public water supply wells was used to select the well to be monitored. The selection of the well was based on the professional judgment of the IDNR Geological Survey Bureau staff and the amount of information available for each well. As a rule, the active well with the most information was selected as the well to be monitored. Wells selected in the study were constructed over a broad period of time. A well monitored at one particular site may not have been the most recently constructed well. Further, because public water supply systems often change the numbering sequence of their wells, the well monitored may not have been the well originally selected in the study. However, since the supplies included in this investigation used only water drawn from a single hydrogeologic source the change in the well being monitored should not have a significant impact on the study.

One hundred and fifty (150) public water supplies were initially targeted for sampling in the study. To insure that at least 150 systems were included in the study additional systems were selected. Approximately 168 public water supplies were sent sample collection kits. One hundred and fifty three (153) public water supplies responded by submitting samples for analysis.

Sample Collection
The UHL provided instructions and guidance to the public water supply operators for the collection of samples. The operators collected the samples and returned them to the laboratory for analyses. Sample collection kits were mailed to public water supply operators over a two-month period beginning August 12, 1992.

One sample was collected from the raw water source and one sample from the finished water at each sample location in the study. Samples were collected in accordance with the instructions provided by the UHL, and in containers provided the laboratory. A single sample consisted of two 40 mL vial, filled slowly to avoid aeration or agitation of the sample. A total of four 40 mL vials (2 raw and 2 finished) were collected from each sampling location The samples were transported to the UHL in Iowa City within 48 hours of collection.

Analysis
All samples were analyzed for the determination of radon at the UHL facility in Iowa City, Iowa. Analysis was conducted following EPA methodology defined in EPA 600/ 2-87/082, Appendix B (proposed 913). Water from each of the two vials that constituted a sample was mixed with a two phase liquid scintillation cocktail and radon was then counted on a liquid scintillation counter. The analytical results derived for the sample had to be within 25% relative difference to be valid. Maximum holding time for samples in the laboratory was 3.8 days (one half life) from the time of collection.

The UHL followed strict quality assurance/quality control practices. Each set of analyses was preceded with low, mid and high level National Bureau of Standards' (NBS) traceable standards. In addition, blank and background vials were run through the analysis. Daily instrument performance and periodic efficiency checks were routine practice.

Results and Discussion
One hundred and fifty three public water supplies were included in the study (Figure 1). Of the 153 supplies included in the study, 60 used the Alluvium as their source of drinking water; 15 used Drift; 13 Dakota or Cretaceous; 1 Pennsylvanian; 7 Mississippian; 26 Silurian/Devonian; 4 Ordovician (above the St. Peter); 22 Cambrian/Ordovician; and, 5 Cambrian (below the St. Lawrence) or Precambrian.

Figure 1. Sampling Sites

Analyses of the raw water from these supplies found 79, or 52%, of the samples exceeded the 300 pCi/l level. Concentrations ranged from below detection level to 2546.4 pCi/l (Table 1). Statistical analysis found no correlation between well depth and the concentration of radon in the raw water samples. The mean concentration of observed values was 371.7 pCi/l.

The mean concentration of radon exceeded 300 pCi/l in all hydrogeologic sources above the Cambrian/Ordovician, except the Pennsylvanian. Only the Pennsylvanian and Cambrian (below the St. Lawrence) or Precambrian supplies showed no raw water samples in excess of the 300 pCi/l level. At the other extreme, 44 of the 60 Alluvium raw water samples (73%), exceeded the 300 pCi/l level. High radon concentrations were observed in other hydrogeologic sources as well. Sixty percent, or 9 of the 15, Drift samples; 7 of the 13 Dakota/Cretaceous (54%); 4 of the 7 Mississippian (57%); 10 of 26 Silurian/Devonian (38%); 2 of the 4 (50%) of the Ordovician (above the St. Peter); and 3 of the 22 Cambrian/Ordovician (14%) exceed the 300 pCi/l level.

Only four public water supplies exceeded 1000 pCi/l in either the raw or the finished water. All observed values greater than 1000 pCi/l occurred in samples taken from bedrock formations above the Ordovician.

Mean radon concentrations were aggregated by county and various regional configurations to identify possible geographic patterns. With one exception, no clearly discernible geographic patterns were evident in these data. In part, the lack of apparent geographic patterns may have been due to the relatively small sample size. Generally, mean concentrations of radon increased from east to west and south to north. All concentrations above 700 pCi/l were found in the northwestern half of the state (Figure 2).

Figure 2. Sampling Sites With Concentrations of 700 pCi/l or Greater. [o& 700pCi/l l>700pCi/l]

Although over half of the public water supplies included in the study were using sources of water high in radon, only 43 (28%) of the supplies in the study exceeded the proposed drinking water standard of 300 pCi/l in samples collected from the supply's distribution systems. With the exception of the Pennsylvanian formation (only one water system was monitored from this formation), each hydrogeologic formation included in the study had one or more public water supplies with raw water radon concentrations above 300 pCi/l.

High radon concentrations in the finished water were observed in 19 of the 60 Alluvium finished water samples (32%);. 4 of the 15, Drift samples (27%); 3 of the 13 Dakota or Cretaceous (23%); 2 of the 7 Mississippian (29%); 7 of 26 Silurian/Devonian (27%); 2 of the 4 (50%) Ordovician (above the St. Peter); 2 of the 22 Cambrian/Ordovician (18%); and, 2 of the 5 Cambrian (below the St. Lawrence) or Precambrian (40%) samples exceed the 300 pCi/l level. Records on treatment technology in place and used at individual public water supplies are incomplete. For the majority of the public water supplies included in this study, treatment technology in use was unknown. However, the information that is available suggests that aeration is an effective treatment technology for the reduction of radon-222.

Of the 153 public water supplies included in the study, the data base suggests that 27 make use of aeration at some point in their treatment. Thirteen of the public water supplies using aeration technology had radon concentrations in their source water in excess of 300 pCi/l. Of these 13 supplies, ten (10) had radon concentrations less than 300 pCi/l in the distribution system. The average reduction in concentration between in the source water and the water in the distribution system in these 10 supplies was 351.0 pCi/l. Radon concentrations were found to increase in two of the 13 supplies and remain unchanged in one supply. It is important to note that treatment of water may vary from well to well, and the use of a particular type of treatment at the time of sample collection is unknown.

The concentration of radon in 42 of the 153 supplies ( 27% ) included in the study was found to be higher in the distribution system than in the raw water. However, only 16 of these sites (10%) were higher than could be accounted for by considering counting error. In six (6) of these supplies the increase in concentration resulted in the supply exceeding the 300 pCi/l proposed standard, even though the source water was below the proposed standard. The differences observed between the source water and the distribution system water ranged up to 811.9 pCi/l.

Recent investigations at the University of Iowa, College of Engineering, suggest that mineral deposits in the distribution systems of public water supplies may be a source of radon in some supplies. If true, the release of radon from these deposits could result in higher concentrations of radon in the finished water than in the source water. This may explain the phenomenon observed in the 6 supplies in this study where concentrations were higher in the finished water than in the source. However, since the collection of samples in this study was not controlled by a single sample collector, nor were the analyses rechecked, sampling error cannot be completely ruled out as a possible explanation.

Fourteen (14) of the supplies in this study were monitored in 1991 and 1992 by the staff of the USGS with sample analyses performed by the UHL in Iowa City. The USGS focused primarily on raw water monitoring. Although, in a few cases the finished water was sampled. In many cases the well that was sampled in 1991/'92 was different from the well monitored in this study. However, because these public water supplies use a single hydrogeologic source of water, the results were compared. The authors recognize that this type of comparison assumes a degree of uniformity within the aquifer that may not exist. No significant differences were found in the concentrations observed in the USGS monitoring and this study.

Public water supplies that use more than one hydrogeologic source for their drinking water will probably have little trouble meeting the proposed Safe Drinking Water Act (SDWA) radon standard of 300 pCi/l, provided that one of the sources that they are using is naturally low in radon. In Iowa, the Silurian/Devonian, Pennsylvanian, Cambrian/Ordovician , Cambrian (below St. Lawrence) or Precambrian and surface water have the potential to supply water with low concentrations of radon-222. The findings of this study suggest that approximately one quarter (123) of the 493 public water supplies that use a sole hydrogeologic source for their drinking water are likely to exceed the radon standard in Iowa. Public water supplies that must rely upon source waters high in radon can significantly reduce the concentration of radon entering the distribution system by using aeration treatment technology. Regardless of the groundwater source, further investigation into the relationship between mineral deposits and radon concentrations in the distribution system should be supported.