Vulnerability Assessment Of

Vulnerability Assessment of

Missouri Drinking Water

to Chemical Contamination

Vulnerability Assessment Report

FY 2002

Prepared for the

Public Drinking Water Program

Water Protection and Soil Conservation Division

Missouri Department of Natural Resources

Submitted by

Stan Wood and Christopher Barnett

Center for Agricultural, Resource and Environmental Systems

Department of Agricultural Economics

College of Agriculture, Food and Natural Resources

University of Missouri – Columbia

December 31, 2001

Table of Contents

Introduction...... 1

Vulnerability Assessment Procedure...... 1

Vulnerability Assessment Results...... 5

Comparing PDWP Monitoring Results with VA Results.....6

GIS Output...... 8

Map Illustrations...... Appendix A

Vulnerability Assessment Results Hardcopy.....Appendix B

Vulnerability Assessment AMLs...... Appendix C

Acronyms...... Appendix D

Note:

ArcInfo coverage names are in BOLD CAPITALS

Coverage items (field names) are in ITALIC CAPITALS

INFO file names are in CAPITALS

dBASE, Excel® and text file names are in lower case italic

AML code, names and commands are in courier new font

Introduction

Objectives

• To conduct a wellhead Vulnerability Assessment (VA) of active public drinking water (PDW) wellheads in the State of Missouri.
• To provide a list of wells determined to be vulnerable to the Missouri Department of Natural Resources (MoDNR).
• Compare VA findings with Public Drinking Water Program (PDWP) monitoring results.

Summary of How the Objectives Were Met

• Conduct a Vulnerability Assessment. This document describes the procedure used to produce the vulnerability list.
• Provide a list of wells determined to be vulnerable. Delivered a list to the Public Drinking Water Program of the MoDNR of wells determined to be vulnerable. The original electronic files are produced in hardcopy in Appendix B.
• Compare monitoring results. The output of the VA was compared to the monitoring results provided by PDWP. Every ground water system that has tested positive for Synthetic Organic Compounds (SOCs) since 1995 was represented on the VA FY 2002 list.

Vulnerability Assessment Procedure

Step One – The most recent SOC contaminant site (SCHEMCOVSCHMPOLY) and pesticide dealership (SDEALCOV) ArcInfo coverages are copied to the VA FY 2002 workspace. The current wells coverage (SPWWELLS) is copied to the workspace and a new coverage is extracted from it. The new layer, ACTWELLS, is composed only of the wells that have a STATUS of “Active”. Wells with a STATUS of “Emergency” are not considered active for the purposes of the Vulnerability Assessment.

Improved locational data acquired by the Source Water Inventory Project (SWIP) was incorporated into this assessment. 141 potential contaminant sites (out of 1,656) were re-located using SWIP Digital Orthophoto Quarter Quadrangle (DOQQ) interpolated coordinates. 98 pesticide dealerships (out of 1,431) were moved as well. Maps #5 and #6 in Appendix A show the location method for dealerships and contaminant sites. Many wells were added or moved in the first two quarters of the FY 2002. 387 wells were re-located using DOQQ coordinates from the SWIP team. 73 wells were moved or added using GPS coordinates from PDWP. Additionally, three wells were added by map interpolation and three by survey interpolation.

Step Two – The AMLs (ARC Macro Language programs) used in previous year’s assessments were copied to the new workspace and checked for correct path and variable names. The INFO file output from the previous assessment, FY01TESTWELLS was imported to the FY 2002 INFO directory. Last year this file was the parent of fy01wells.xls. This year, the file is used to create the data file containing the wells dropped from last year’s VA list. AMLs are included in Appendix C.

Step Three – The main AML, doassess.aml, is run. It first creates the ACTWELLS coverage and then begins to call a series of AMLs that search SCHEMCOV and SDEALCOV for each possible contaminant site within a one-mile (1,609 meters) radius of each wellhead in ACTWELLS.

&run ptdist.aml %.activewellcov% chem

&run ptdist.aml %.activewellcov% deal

&run makewellarea.aml

&run polydist.aml %.activewellcov% schmpoly

1. Ptdist.aml – This program is called twice, once for each point potential contaminant site (PCS) layer; SCHEMCOV and SDEALCOV. Using the pointdistance command the code finds each SOC contaminant point within one mile of each wellhead. A record is generated for each site identifying it by the site number and distance from the well. The program then generates an output text file with a record for each site for each unique chemical found and test method required. The same is done for SDEALCOV although the output does not contain METHOD as dealerships are automatically associated with all test methods covering agricultural chemicals.

The text files are read back into INFO where the distance is changed to meters. Chemfile.txt becomes WELLCHEM.DIST and dealfile.txt becomes WELLDEAL.DIST. At this time, DISTANCE is recalculated from meters into feet.

Below is a sample of chemfile.txt (columns are PWSSIDEXT, SITENUM, chemical ID, Chemical name, test METHOD and distance to PCS in feet):

1010609102.000 11365 C08 Carbofuran 531.1 344.449

1010609102.000 11365 C26 Metolachlor 507 344.449

1010609102.000 11365 C42 3-Hydroxycarbofuran 531.1 344.449

1010609102.000 11491 C05 Atrazine 507 396.622

1010609102.000 11491 C08 Carbofuran 531.1 396.622

1010609102.000 11491 C14 Dinoseb 515.1 396.622

1010609102.000 11491 C26 Metolachlor 507 396.622

1010609102.000 11491 C42 3-Hydroxycarbofuran 531.1 396.622

Below is a sample of the output text file (dealfile.txt) when ptdist.aml is run with the type variable set to “deal” (columns are PWSSIDEXT, SITENUM, chemical ID [99 = dealership], Chemical name, test METHOD and distance to PCS in feet):

1010609102.000 20367 C99 Pesticide Dealership 0 349.196

1010609102.000 20368 C99 Pesticide Dealership 0 395.958

1010609101.000 20367 C99 Pesticide Dealership 0 144.879

1010609101.000 20368 C99 Pesticide Dealership 0 174.432

1010098103.000 21929 C99 Pesticide Dealership 0 750.100

1010098103.000 21946 C99 Pesticide Dealership 0 928.027

1010098103.000 20165 C99 Pesticide Dealership 0 1054.750

1010098101.000 21929 C99 Pesticide Dealership 0 640.739

1. Makewellarea.aml - A layer of buffers from SCHMPOLY is created by this AML. There are no pesticide dealership polygons, they were incorporated into the chemical polygons layer a couple of years ago. All contaminant polygons are buffered one-half mile. This layer is called CHEMREG. An intersection command is performed on CHEMREG with ACTWELLS. All the wells that fall within the buffer zones are selected and placed in a new layer WELLSCHEMREG. A frequency file of this layer called WELLAREA.FRQ is created. This file is used by the next AML.

1. Polydist.aml – This AML does with WELLAREA.FRQ what ptdist.aml does with point coverages. The output file is called WELLSCHMPOLY.DIST.

Step Four – The three output files from ptdist.aml and polydist.aml are then combined into NCONT.DIST (number of contaminants by distance).

arcplot

infofile wellchem.dist INFO ncont.dist INIT

infofile welldeal.dist INFO ncont.dist APPEND

infofile wellschmpoly.dist INFO ncont.dist APPEND

Step Five – While still in ArcPlot, the NCONT.DIST file is reselected for distances less than a half-mile (2,640 feet) and the selected records are placed in NCONTHALF.DIST. This file is then run through the frequency command using PWSSIDEXT and METHOD. The output is called NCONTHALF.FRQ.

resel ncont.dist INFO distance le 2640

infofile ncont.dist INFO nconthalf.dist

quit

frequency nconthalf.dist nconthalf.frq

pwssidext

method

end

end

Step Six – This code calls the dealmeth.aml. Dealmeth.aml adds the ten test methods recommended for wells with pesticide dealerships in their buffer zone to the records in NCONTHALF.FRQ. This file is copied to NCONTHALF.FRQ2 preserving NCONTHALF.FRQ. Frequency is run on NCONTHALF.FRQ2 (using PWSSIDEXT and METHOD) to produce NCONTHALF.FRQ3 and then again (using just PWSSIDEXT) to create NCONTHALF.FRQ4. This last file contains the wells that will be on the VA list. There are more data to be added before the file is in its final form.

ap

&run dealmeth.aml /* adds 10 test methods

quit

frequency nconthalf.frq2 nconthalf.frq3

pwssidext

method

end

end

frequency nconthalf.frq3 nconthalf.frq4

pwssidext

end

end

Step Seven – In this section of the program the code calls maketable.aml which adds several fields to NCONTHALF.FRQ4. These fields include IPWS, PWSSNAME, LOCALNAME and WELLNUM. ACTWELLS.PAT is related by PWSSIDEXT and the values are copied over to NCONTHALF.FRQ4.

&run maketable.aml

Step Eight – At this point we are ready to add STATUS and CODE to the VA list file, NCONTHALF.FRQ4. Once these are added, NCONTHALF.FRQ4 is copied to FY02TESTWELLS to preserve the former file for future reference. In INFO, FY02TESTWELLS is related to SPWWELLS.PAT by PWSSIDEXT and the well status is copied to the new field. Then FY02TESTWELLS is related to FY01TESTWELLS by PWSSIDEXT and the code A (well added) is applied to all records which are not in the previous year’s file. Last year’s file, FY01TESTWELLS, is then selected and related to the new VA list file FY02TESTWELLS, and all records not found in the new file are given a code of D (well dropped). These codes are used to select the added and dropped records so they can be placed in their own files for ease of reference.

additem nconthalf.frq4 nconthalf.frq4 code 1 1 c

additem nconthalf.frq4 nconthalf.frq4 status 16 16 c

copyinfo nconthalf.frq4 FY02TESTWELLS

&data arc info;arc

sel FY02TESTWELLS

relate SPWWELLS.PAT BY PWSSIDEXT

move $1STATUS TO STATUS

sel FY02TESTWELLS

relate FY01TESTWELLS by PWSSIDEXT

res PWSSIDEXT ne $1PWSSIDEXT

move 'A' to CODE

sel FY01TESTWELLS

relate FY02TESTWELLS by PWSSIDEXT

res PWSSIDEXT ne $1PWSSIDEXT

move 'D' to CODE

move $1PWSSNAME TO PWSSNAME

sel FY01TESTWELLS

relate SPWWELLS.PAT BY PWSSIDEXT

res PWSSNAME = ''

move $1PWSSNAME TO PWSSNAME

as

move $1STATUS to STATUS

q stop

&end

Step Nine – The final GIS outputs were exported to dBASE IV format files. These files were then imported to Microsoft Excel® and used to create fy02wells.xlsdroppedwells.xls, and addedwells.xls

infodbase FY02TESTWELLS fy02wells.dbf

infodbase FY01TESTWELLS fy01wells.dbf

Vulnerability Assessment Results

611 wells appear on the FY 2002 VA list. In FY 2001 there were 613 on the list. In the current VA, thirty-four wells have been dropped and thirty-two have been added. (In FY 2001 twenty-six wells were dropped and thirty added.) The wells on the VA list represent 16.5% of the 3,714 active public drinking water wells considered for the analysis. The FY 2002 VA list is provided in Appendix B.

Wells Dropped This Year

There were thirty-four wells dropped from the VA list that had appeared on the list created in FY 2001. Table 1 shows the reasons for the exclusion and the number of wells in each category. Note that a status of “Inactive” includes “Emergency”, “Plugged” and “Abandoned” wells. Examples of the “well moved”, “contaminant site moved” and “well did not exist” are shown on the map series in Appendix A and described in the section Map Illustrations. A copy of the table of dropped wells is provided In Appendix B.

Table 1. Reasons for Wells Dropped From VA List in FY 2002

Wells Added This Year

Thirty-two wellheads were added to the VA list in FY 2002. In at least one case (Liberty, Well #7) the well was moved closer to a pesticide dealership, and in another case (Steelville, Well #3) the well was moved closer to a known contaminant site. Both of the cases are documented on maps included in Appendix A and described in the section Map Illustrations.

In addition to the file of added wells, a second Excel® file was produced showing which chemicals were present at the potential contaminant sites within the wellhead buffer zone. This file is called addedwellschems.xls and a copy is included in Appendix B.

Comparing PDWP Monitoring Results with VA Results

Public Drinking Water Program SOC monitoring data was received from PDWP in the form of a Microsoft Access® (.mdb) file. The original file contained 1,074 records for 108 public water supply systems and covered the period from February 1995 to September 2001. This file was exported into dBASE IV (.dbf) format and imported into INFO (SOC.FULL).

First, we dropped the records relating to surface water systems and any system which had no active wells. This was done by relating the file SOC.FULL to ACTWELLS.PAT by IPWS. Records were dropped that had no corresponding IPWS in ACTWELLS.PAT, as they were either surface water systems or had no active well, leaving 148 records. To eliminate records where a system had multiple detections of the same chemical we then ran a frequency on the INFO file by IPWS, ANALYTNAME and SUPPLYNAME. Now there were only 50 records for 33 systems. There is one record for each system by each chemical detected. This file (SOC-frequency.xls) is included in Appendix B with the title SOC Monitoring Detects on Systems with Active Wells. This approach ended up with only systems rather than individual wells so another path was taken.

The next method started again with the original file (in INFO format). Within INFO we selected SOC.FULL and related it to FY02TESTWELLS by IPWS. A reselect (IPWS = $1IPWS) was performed. This resulted in 100 records in SOC.FULL selected. All of the detect records selected had a corresponding system in the FY 2002 VA list. The INFO file was output to match.txt and imported to Excel® where duplicate (system, analyte and sample location) entries were deleted. There were now 41 records for 19 systems in the resulting file (match.xls, see Appendix B). Because sample location was included, we were able to tell in several cases when a specific well had a detect. This file was compared with fy02wells.xls to determine what relationship could be found between the two lists.

Comparison Findings

All of the systems represented in match.xls are found on the FY 2002 VA list. In other words, all systems that had detections were on the list to be tested. In most cases the wells with detected contaminants were the same wells on the VA list. See Table 2 below for more details. The records in italic were matched to wells on the VA list. Wells with a status in the VA/GIS of inactive, emergency or plugged do not show up on the VA list. These records are shown with a gray background. Some records cannot be matched to any well as the sample location is either blank or after treatment.

In the table below, nineteen wells with detections were also on the VA list for FY 2002. Five wells were inactive, and ten records could not be matched to any one wellhead.

Table 2. Monitoring Results versus VA List Results.

*Corrected from 206, a typo in the monitoring results file.

Conclusion

Conservative Approach

The VA procedure has always used a conservative approach to the problem of determining required testing methods. This approach helps insure that wells deemed vulnerable will be on the VA list. The history of past detections shows that in every case where a detection record provides an active well as a sample location, that well was on the FY 2002 VA list. Conversely, out of 611 wells the procedure found to be vulnerable, only nineteen (3.1%) have had detections since 1995.

Number of Tests

If every active well was tested quarterly, the number of tests required would be 207,984 (3,714 * 14 * 4). The FY 2002 VA list contains 611 wells (16.5% of all active wells) and recommends 18,456 tests (11.3% of 207,984). This means a reduction in the number of required tests by 88.7%.