California Environmental Protection Agency

Department of Pesticide Regulation

Environmental Monitoring and Pest Management

1020 N Street, Room 161

Sacramento, California 95814-5624

November 15, 1996

Study #156: Protocol for Obtaining a Preliminary Characterization of Intralaboratory Precision of Acute Toxicity Tests Performed on Ambient Water Samples


The Department of Pesticide Regulation (DPR) will increasingly rely on the results of aquatic toxicity tests, in addition to chemical analysis, from ambient surface water monitoring studies to direct surface water regulatory programs. Specifically, aquatic toxicity tests will be the primary monitoring tool in the dormant spray studies #154 and #155 scheduled to be conducted in the Fall/Winter of 1996/97 (Nordmark, 1996; Bennett, 1996). Aquatic toxicity tests are a test method similar to chemical analysis methods. Test methods should always be accompanied by a statement of the precision of the method. Typically the precision statements for aquatic toxicity tests are made in terms of the precision of the LC50 (the concentration of a chemical that causes 50% mortality in exposed test organisms) value for various reference toxicants (U.S. EPA, 1993). A reference toxicant is a standard compound , such as sodium chloride (NaCl) or potassium dichromate (K2Cr2O7), approved by U.S. EPA for use in confirming the general health of the test organism colony and to document that the laboratory aquatic toxicity tests are in a state of statistical control. The LC50 precision is obtained by conducting studies on a dilution series including various concentrations of the chemical or effluent of interest. The precision of the LC50 is relevant to point source effluents because it is important to know the degree of dilution required before the effluent can be safely discharged into ambient receiving waters. However, the LC50 precision is not the correct precision estimate for the measurement endpoint used in aquatic toxicity tests employed in DPR monitoring studies.

DPR is exclusively monitoring ambient waters, therefore, a dilution series to estimate the LC50 is not performed. The U.S. Environmental Protection Agency (U.S. EPA) Acute Toxicity Manual (U.S. EPA, 1993) states that, although a dilution series may be performed, ambient water tests are typically conducted at one concentration - 100% (undiluted) ambient water. This is because the toxicity of the water as it is collected in the natural environment is the property of interest. The question is simply whether or not the ambient sample shows a significantly higher proportion mortality than a control sample. The actual proportion mortality of the sample is not directly of interest because the sample is not considered toxic unless there is a significantly higher proportion mortality in the sample than in the control. For example, 30% mortality in the ambient sample is not relevant unless this level of proportion mortality is statistically significantly higher than that observed in the control. Otherwise this is simply an anecdotal observation of 30% mortality in that single sample that may or may not be observed in a second toxicity test performed on a split of that same ambient sample. The proportion mortality is, however, used to perform the required statistical test. Furthermore, the precision of the proportion mortality measurement endpoint determines the sensitivity of the statistical test. (either a t-test or an appropriate nonparametric test depending upon the toxicity test protocol used).

The precision of the proportion mortality measurement endpoint for toxicity tests conducted on ambient waters has not been established. This may be because aquatic toxicity tests are not commonly used to monitor toxicity associated with non-point source pollution in ambient waters. DPR issued a request for information about the use of aquatic toxicity tests to monitor ambient waters for non-point source pollution to all 10 U.S. EPA regions (Stoddard, 1996). Responses were received from 2 regions (Kanetsky, per. comm.; Howell, 1996). Only Region 6, Dallas, Texas, indicated that they specifically " the use of ambient toxicity testing to investigate potential toxicity originating from either (or both) point and nonpoint sources" (Howell, 1996). U.S. EPA Region 9, which contains California, did not respond. Nevertheless, Region 5 of the California Regional Water Quality Control Board is employing aquatic toxicity tests to screen ambient waters for violations of the narrative toxicity objective (Foe and Sheipline, 1993; Foe, 1995). Agricultural pesticides have been implicated in the cause of observed toxicity in some of these ambient water samples. However, the studies implicating agricultural pesticides in causing the observed toxicity did so by using the results of single toxicity tests with no accompanying measure of test precision. Essentially, single positive tests (i.e. those with statistically significant proportion mortality)were linked with the detection of an agricultural chemical to prove causation. These studies did not quantify the uncertainty in the results obtained from a single toxicity test. Nor was any detailed statistical analysis conducted on the data to establish a statistically significant link between occurrence of significant toxicity and detected concentrations of agricultural pesticides..

Because aquatic toxicity test are playing an ever larger role in DPR surface water monitoring studies conducted, it is important to characterize the precision of the aquatic toxicity test measurement endpoint actually used to detect ambient toxicity. Quality Assurance/Quality Control (QA/QC) programs for all aquatic toxicity monitoring studies must be required to aid in quantifying the precision of aquatic toxicity tests. This will allow DPR to place greater confidence in the aquatic toxicity test results. This requirement would bring aquatic toxicity testing requirements in line with DPR's chemical analysis requirements. All chemical analyses conducted for DPR monitoring studies by the California Department of Food and Agriculture (CDFA) lab have QA/QC requirements specified by DPR. The historical data set assembled from these requirements have provided DPR with an assessment of the reliability of the chemical results obtained in DPR studies. There is no companion data set for aquatic toxicity tests. DPR has no basis from which to draw conclusions about the reliability of the results of aquatic toxicity tests. Of particular interest is the intra-laboratory precision at the California Department of Fish and Game (CDFG) Aquatic Toxicity Laboratory (ATL) because ATL is the primary aquatic toxicity testing facility for DPR monitoring studies. This role is similar to that of the CDFA lab who is the primary chemical analysis lab employed in DPR monitoring studies. The protocol for aquatic toxicity described herein will provide an initial basis from which to assess the confidence that can be place in results of aquatic toxicity tests. These results can also be used to develop a general DPR specified QA/QC plan for future monitoring studies employing aquatic toxicity tests. It is further assumed that more detailed studies beyond this protocol may required. It is anticipated that a standard QA/QC protocol for ambient aquatic toxicity testing will be developed from the data generated according to this protocol. The data from this study will also be used for quality control information for the dormant spray studies #154 and #155.


The objective of this study is to obtain a preliminary characterization of intralaboratory precision of acute toxicity tests performed on ambient water samples.


This study will be conducted under the general direction of Randy Segawa, Senior Environmental Research Scientist (supervisor) by Environmental Hazard Assessment Program (EHAP) personnel as follows:

Project Leader: Terri Barry

Field Coordinators: Kevin Bennett

Craig Nordmark

Senior Scientist: Lisa Ross

Principal Investigator (Toxicity) California Department of Fish and Game

Agency & Public Contact: Pat Dunn

All questions concerning this study should be directed to Pat Dunn at: (916) 324-4077; FAX (916) 324-4088

Study Plan and Sampling Methods

This study plan is to be coordinated with Dormant Spray Studies #154 and #155. The acute toxicity tests conducted for this study will be in the form of duplicate toxicity tests performed on split samples taken at the Karnak pumping station in the Sacramento River watershed (study #154) and Orestimba Creek in the San Joaquin River watershed (study # 155). In those studies, ambient water samples will be collected for acute toxicity testing twice weekly during a three month period. The initial characterization of the precision of the ambient water acute toxicity tests will be from results of the duplicate toxicity testing performed on the split samples. The protocols for study #154 and study #155 should be consulted for a complete description of each study plan and the sampling methods.

To assess intralaboratory performance, all field samples will be split and submitted for duplicate toxicity testing at ATL. The split samples will not be labeled as such and will be submitted as field samples. All field samples from both study #154 and study #155 will be submitted together. No samples submitted to ATL will be labeled to identify location of collection. It is anticipated that each site will yield 24 acute toxicity ambient water samples that will be split for duplicate analysis.

Toxicity Testing

Acute toxicity testing conducted by the CDFG ATL will follow current U.S. EPA procedures for the cladoceran Ceriodaphnia dubia (U.S. EPA, 1993). Acute toxicity will be determined using a 96-hour, static-renewal bioassay in undiluted ambient water. All bioassays must commence within 36 hours of sample collection. CDFG will also analyze ambient water samples for total alkalinity, ammonia, hardness, and specific conductivity.

ATL will be requested to submit raw data results for all toxicity tests conducted. In addition, quality control charts for reference toxicant tests during the course of the study will also be requested. Reference toxicant tests should be conducted on a regular schedule according to normal ATL operations, but at a minimum of once a month. These data will be used to ensure that the toxicity testing procedures were in a state of statistical control during theses studies.

Number of Toxicity Tests

2 acute toxicity tests for 2 sites each week for 12 weeks 48

Total number of acute toxicity tests 48

Data Analysis

Toxicity data will be analyzed for consistency within the CDFG ATL laboratory. Two aspects of the toxicity test results are of interest: 1) the agreement between the determination of toxicity in split samples (i.e. if one split is declared toxic, it the second split also declared toxic?) and 2) the agreement in the measurement of proportion mortality in split samples. Nonparametric statistical methods suitable for data measured as counts will be used to examine the agreement between of the determination of toxicity in split samples. Linear regression analysis will be used to examine the agreement in proportion mortality between the split samples. Analysis of these data will provide an initial characterization of the level of confidence DPR can place in the results from any single toxicity test. Further studies may be required to specifically quantify the precision of the proportion mortality measurement endpoint. These further studies may involve ambient water samples spiked at various levels of pesticides of interest.

Time Table

Begin Sampling December 1996

Finish Sampling March 1997

Draft Report August 1997

Literature Cited

Bennett, K. 1996. Protocol for monitoring acute and chronic toxicity in the San Joaquin River watershed: Winter, 1996-97. California EPA/Department of Pesticide Regulation, Environmental Hazards Assessment Program. Draft, November 1, 1996.

Foe, C. 1995. Insecticide concentrations and invertebrate bioassay mortality in agricultural return water form the San Joaquin Basin. Regional Water Quality Control Board, Central Valley Region, Sacramento, Ca.

Foe, C. and R. Sheipline. 1993. Pesticides in surface water from applications on orchards and alfalfa during the winter and spring of 1991-92. Regional Water Quality Control Board, Central Valley Region, Sacramento, Ca.

Howell, C. 1996. Letter to P. Stoddard from U.S. EPA Region 6, Dallas, Texas, concerning use of aquatic toxicity tests for non-point source pollution monitoring. U.S. EPA letter dated July 12, 1996.

Nordmark, C. 1996. Protocol for monitoring acute and chronic toxicity in the Sacramento River watershed: Winter, 1996-97. California EPA/Department of Pesticide Regulation, Environmental Hazards Assessment Program. Draft, November 1996.

Stoddard, P. 1996. Letter to U.S. EPA, Regions 1-10, concerning use of aquatic toxicity tests to monitor non-point source pollution. DPR letter dated July 1, 1996.

U.S. EPA. 1993. Methods for measuring the acute toxicity of effluents and receiving waters to freshwater and marine organisms. 3rd edition. EPA/600/4-90-027F. August 1993.