Modeling to Support Monitoring Waiver for Selected Wells in Hawaii
Hawaii Department of Health, USEPA
03/14/11 – 09/30/12
Quarterly monitoring of source water from drinking water wells in Hawaii has been going on for decades. However, many of these wells show no evidence of contamination. For a contaminant to appear in a well, it must be present in the capture zone of the well. Processes that contribute to point or non-point source pollution include leaching of chemicals or pathogens through soil, transport in aquifers, and eventual appearance in well waters. Hawaii, has a unique hydrogeologic setting unlike any in the mainland United States. It is also one of the few states that has delineated the capture zones of all municipal water supply wells for travel times ranging from 2 years to 25 years in its Source Water Assessment Program (SWAP). The University of Hawaii (UI-I) worked with the Hawaii Department of Health (DOH) to develop vulnerability ratings for each drinking water source through ground water modeling and using weighting criteria based on potential contaminating activities (PCAs) within the capture zones of the wells. This effort included a detailed land use survey, delineation of PCAs, and assignment of appropriate scores based on the type of PCA.
The Hawaii Department of Agriculture (HDOA) developed the Comprehensive Leaching Risk Assessment System (CLEARS) for land-applied chemicals to examine their leachability to ground water based on soil and chemical properties, recharge, and travel distance (Li et al., 1998; Stenemo et al., 2007). The CLEARS model is a Tier-l model (a screening model) that is currently used by the HDOA to rank the leaching potential of new chemicals being imported into the state. It uses a modified attenuation factor approach to evaluate leaching (see Rao et al., 1985). The more a chemical is attenuated, the less chance it will have to leach. For a given chemical entering the state, its leaching behavior is compared against a chemical known to leach (such as atrazine or bromacil) and a known non-leaching chemical (such as endosulfan). Uncertainties in soil and pesticide properties are accounted for in the process. Depending upon its modified attenuation factor score (which is determined from the soil and pesticide interaction accounting for the annual recharge and depth at which the leaching is evaluated, see Stenemo et al., 2007), a chemical is ranked either having high potential to leach down the soil profile to ground water (a “leacher”), or with low potential to leach (a “non-leacher”), or with uncertain leaching potential. Recently, a field leaching evaluation (Dusek et al., 2010) was carried out for the validation of this model.
As some priority pollutants had not been detected in Hawaiian municipal water supply wells over decades, and since there was adequate information about land use (past and current) and PCAs, it was feasible to justify the waiving of regular monitoring for chemicals that show low risk of appearance in ground water. This will save the tax payers money.
Objectives of the Research:
- Obtain physico-chemical and toxicological properties of monitored chemicals from appropriate databases,
- Obtain information about the primary modes of release of these chemicals from anthropogenic sources and include the volatility component in the modified CLEARS model,
- Update the soil and recharge data to the CLEARS model using recently available information from state and federal agencies,
- Run a process based model such as HYDRUS- 1 D for the chosen chemicals in selected soil types to estimate concentrations passing through a reference plane, and
- Develop an overall leachability potential for these compounds based on CLEARS values, HYDRUS-1D model predictions, and PCA information present in the capture zone areas of the municipal water systems requesting monitoring waivers.
Task 1 – Physico-chemical property and toxicity evaluation
For the monitored chemicals, an appropriate database maintained by EPA (e.g., IRIS data base of USEPA, http://www.epa.gov/iris/) was consulted for toxicological properties. Other data bases were searched for appropriate physico-chemical parameters such as half-life, sorption coefficient, volatility (Henry’s constant), etc. As the wells were routinely monitored for the priority pollutants, EPA sources were contacted to obtain more detailed information on these compounds.
Task 2 – Modes of release and volatility component
For volatile chemicals, the mode of release is important for determining its transport. If the release occurs on the land surface, only a fraction of the released mass will enter the soil and leach down the profile. On the other hand, if the release is subsurface (e.g., tank leak or leak from pipes), then a larger mass can move downward. Thus, it may be problematic to estimate the mass of released chemical without knowing the release history. The CLEARS model is a screening model and does not account for actual mass release. An “effective” volatility component was included in the CLEARS database to account for some loss of compounds to the atmosphere. However, for other process-based models, mass release information was needed so actual concentrations at given depths were calculated.
The effective volatility was a “reduced” volatility assuming a fraction of the mass of the chemical did not get fully volatilized as would happen in contact with atmosphere. Thus, it was a somewhat conservative approach.
Task 3 – Updating recharge and soil database
The CLEARS databases for soils and recharge were last updated in 2006 (Stenemo et a!., 2007). The soils database was obtained from the Natural Resources and Conservation Service (NRCS). The recharge database was obtained from parallel efforts of a project that was doing Source Water Assessment and Protection (SWAP) for drinking water supplies for wells and surface water sources in Hawaii. The recharge data were compiled from studies conducted by the US Geological Survey and other agencies.
In this task, we conducted a thorough search of literature on recharge and contact personnel from appropriate agencies for data. In addition, updates to NRCS databases were searched through contacts at the NRCS. Updated data on hydraulic properties, field capacity water content, and organic carbon content were added the soils database. Additionally, spatial variability in recharge was built into the database. In the old CLEARS database, recharge was usually estimated as a fraction of rainfall. In 2006, recharge was estimated from various USGS water balance studies. As land use is changing with time, we feel the spatial variability in recharge will be changing. The CLEARS model was rerun with these updated parameters
Task 4 – Running Process Based Model HYDRUS-1D
The CLEARS model provides the relative leachability of a compound compared to a known leaching chemical and a non-leaching chemical. In order to obtain the relative persistence of a chemical and its transport in the soil profile, the HYDRUS-1D model (Simunek et al., 2005) was run for given soil types. The most recent version of the software was used for this purpose. This was done particularly for soil types within the capture zones of the wells. The susceptibility of each chemical to leach in a given soil type was identified and the profile concentration over different travel times were shown on a GIS platform. Also, the time to reach 5 or 10% of the initial concentration passing a reference plane was used as a secondary indicator of its potential to reach ground water.
Task 5 – Aggregate Leachability Potential based on CLEARS and HYDRUS runs and PCA scores
Based on the leachability index of the chemical determined from CLEARS runs and the profile concentration information from HYDRUS-1D, we examined the chemicals that have high potential to leach to ground water. From the PCA data, we determined if these chemicals were present in the capture zone of the wells. If a chemical was not present in the capture zone, then a request was made to waive monitoring of that chemical. If a chemical was present, the overall leaching potential of the chemical was determined based on the PCA score and its leachability potential from CLEARS and HYDRUS. If the overall risk seemed to be low, a request was made to drop the chemical from monitoring, or initially to reduce the frequency of monitoring until more data was obtained.
Dr. Chittaranjan Ray (WRRC)
Clinton Williams (USDA – Arid-Land Agricultural Research Center)
Philip Moravcik (WRRC)