KABAM Version 1.0 User's Guide and Technical Documentation

KABAM Version 1.0
(K_ow (based) Aquatic BioAccumulation Model)

April 7, 2009

Environmental Fate and Effects Division
Office of Pesticide Programs
U.S. Environmental Protection Agency
Washington, D.C.

On this page

• Acknowledgements

1. Introduction
   • 1.1 Model Description
   • 1.2 When to Use This Model
   • 1.3 Conceptual Model
   • 1.4 Model Application
2. Input Parameters
   • 2.1 Chemical Specific Inputs
   • 2.2 Ecosystem Inputs
3. Parameters & Calculations
4. Model Results
5. Assessing Pesticide Concentrations in Fish Tissues for Human Consumption
6. Model Assumptions, Limitations, and Uncertainties

Links to appendices

• Appendix A. Description of Bioaccumulation Model
• Appendix B. Explanation of Defaults and Alternative Values Representing Abiotic Characteristics of Aquatic Ecosystem
• Appendix C. Explanation of Default Values Representing Biotic Characteristics of Aquatic Ecosystem, Including Food Web Structure
• Appendix D. Selection of Mammal Species of Concern and Corresponding Biological Parameters
• Appendix E. Selection of Bird Species of Concern and Corresponding Biological Parameters
• Appendix F. Description of Equations Used to Calculate the BCF, BAF, BMF, and BSAF Values
• Appendix G. Description of Equations Used to Calculate Dietary-based and Dose-based EECs, Toxicity Values, and RQs for Mammals and Birds Consuming Contaminated Aquatic Organisms
• Appendix H. Methods for Estimating Metabolism Rate Constant (k_M)
• Appendix I. References Cited

1. Introduction

   • 1.1 Model Description

     KABAM (K_OW (based) Aquatic BioAccumulation Model) is used to estimate potential bioaccumulation of hydrophobic organic pesticides in freshwater aquatic food webs and subsequent risks to mammals and birds via consumption of contaminated aquatic prey. This model can also be used to estimate pesticide concentrations in fish tissues consumed by humans. The model was designed for use by the U. S. Environmental Protection Agency Office of Pesticide Programs' Environmental Fate and Effects Division (EFED) scientists. KABAM is composed of two parts:

     1. a bioaccumulation model estimating pesticide concentrations in aquatic organisms

     2. a risk component translating exposure and toxicological effects of a pesticide into risk estimates for mammals and birds consuming contaminated aquatic prey.

     The bioaccumulation portion of KABAM is based on an aquatic food web bioaccumulation model published by Arnot and Gobas (2004). This model was originally published in 1993 by Gobas (Gobas 1993) and was modified by Arnot and Gobas (2004). The Arnot and Gobas (2004) model was selected for estimating pesticide bioaccumulation based on the following reasons:

     1. the Gobas 1993 model underlying the Arnot and Gobas 2004 version is generally accepted by the scientific community as a reasonable approach for estimating bioaccumulation of persistent hydrophobic organic compounds in aquatic systems (Burkhard 1998)

     2. the 1993 version of the model has been used by EPA for regulatory purposes (USEPA 1995, 2000, 2003)

     3. both Gobas 1993 and Arnot and Gobas 2004 have been published in peer-reviewed literature.

     Although originally developed and applied to the Great Lakes ecosystem for modeling PCBs and selected pesticides, this model has been applied and validated for other ecosystems, including the Hudson river, Fox river/Green Bay and Bayou D'Indie in Louisiana (USEPA 2003, Burkhard 2003). A detailed description of the Arnot and Gobas (2004) model is available in Appendix A.

     The bioaccumulation portion of KABAM relies on a pesticide's octanol-water partition coefficient (K_OW) to estimate uptake and elimination constants through respiration and diet of aquatic organisms in different trophic levels. Pesticide tissue concentrations in aquatic organisms are calculated for different trophic levels of a food web through diet and respiration.

     In the risk component of KABAM, pesticide concentrations in aquatic organisms are used to estimate dose- and dietary-based exposures and associated risk quotients for mammals and birds consuming aquatic organisms. The methods used in the risk component of KABAM are consistent with EFED's current modeling approach for assessing risks to terrestrial mammals and birds described in USEPA 2004a, as implemented in the T-REX model (version 1.4.1; USEPA 2008a). Search EPA Archive

     ---

   • 1.2 When to Use This Model

     KABAM should be used for pesticides having all of the following characteristics:

     • The pesticide is a non-ionic, organic chemical.
     • The Log K_OW value is between 4 and 8.
     • The pesticide has the potential to reach aquatic habitats.

     ---

   • 1.3 Conceptual Model

     Conceptually, KABAM represents a freshwater aquatic ecosystem. This ecosystem receives runoff and spray drift containing pesticides from sites where pesticides are applied. The aquatic ecosystem incorporates seven food web components to describe the trophic transfer of a pesticide in an aquatic food web. These include, in increasing order of trophic level within the food web: phytoplankton, zooplankton, benthic invertebrates, filter feeders, small fish, medium fish and large fish. These components are referred to within this User's Guide as "trophic levels." They are not intended to represent discrete trophic levels, but rather generic levels of an aquatic food web (e.g., primary producers, primary consumers, secondary consumers, and predators). KABAM also evaluates potential exposures and risks to mammals and birds that feed upon aquatic animals containing pesticide residues accumulated through the aquatic food web (Figure I).

     

     KABAM can represent a specific ecosystem, as defined by the model user. The ecosystem can be defined by abiotic (e.g., water temperature, % organic carbon in sediment) and biotic input parameters (e.g., body weights of aquatic animals, feeding preferences of fish, birds, and mammals). The model user can modify these parameters to match the characteristics of ecosystems relevant to a specific mesocosm study or a field study.

     For general use, the default model ecosystem for KABAM is defined as the EFED standard pond scenario for the Exposure Analysis Modeling System (EXAMS). The standard pond has two compartments: a water column and a benthic area. The water column is 20,000,000 liters in volume and the benthic area has a volume of 500,000 liters. The standard pond receives pesticides in runoff (dissolved in water and sorbed onto eroded soil) and spray drift from a 10-ha treatment field that is immediately adjacent to the pond. The treatment field is represented by various scenarios using the Pesticide Root Zone Model (PRZM). The meteorological data corresponding to the selected PRZM scenario can influence the runoff of a pesticide into the standard pond and also the water temperature of the pond environment.

     The default biotic portions of KABAM are designed to be representative of organisms from the seven trophic levels defined above. Mammals and birds of concern are defined by considering species of mammals and birds relevant to the United States which rely upon aquatic ecosystems for their food sources.

     ---

   • 1.4 Model Application

     The application of KABAM is referred to in this User's Guide as the "KABAM tool." The KABAM tool is implemented in Microsoft^® Excel 2003. This software program was chosen as an operating platform because it is available to EFED users and to the public. Excel is a commonly used spreadsheet program that most scientists are familiar with. Computers suitable for running the software programs necessary for this tool require no additional hardware.

     Once the KABAM tool is opened, the "Model Description" worksheet is displayed. This worksheet contains the version information, a brief model description, and a list of references. Across the bottom of this Excel window are several worksheet tabs indicating the various portions of the KABAM tool, including "Chemical Specific Inputs," "Ecosystem Inputs," "Parameters & Calculations," and "Results." The requirements and functions of these worksheets are explained in more detail below.

     The overall format of the KABAM tool was developed for ease of use. Tables embedded in the worksheets were designed for clarity of information and for eventual cut and paste from Excel into a Microsoft^® Word document containing a risk assessment. Where necessary, comments are provided for guidance on selecting input parameters. For more detailed information than is contained in the spreadsheet concerning the model, input parameters, calculations, and results, this guidance document should be utilized.

   ---

2. Input Parameters

   Two types of input parameters are required to run KABAM: those that are specific to the pesticide and those that define the aquatic ecosystem, including the mammals and birds of concern. These input parameters are distinguished by two worksheets that are titled "Chemical Specific Inputs" and "Ecosystem Inputs."

   To run the model, the user is only required to input chemical specific values since default values are already inserted into the appropriate locations for ecosystem input parameters. These default values allow the user to run KABAM with reasonable and reliable parameters; however, the user can select other parameters to explore bioaccumulation of a chemical and associated potential risk to mammals and birds that consume aquatic animals. Guidance for altering input parameters from the default values is provided in this User's Guide.

   • 2.1 Chemical Specific Inputs

     The "Chemical Specific Inputs" worksheet contains three tables for the user to input data. Tables 1 and 3 require the user to input chemical-specific values. Table 2 contains default values that do not require user inputs, but are designed to allow the user flexibility in the case that chemical-specific data are available for uptake and depuration rate constants in aquatic organisms.

     Table 1

     Table 1 requires inputs related to the chemistry and estimated environmental concentrations (EECs) of the pesticide. Required inputs include:

     1. pesticide name
     2. Log K_OW
     3. organic carbon partition coefficient (K_OC)
     4. sediment pore water concentrations of pesticide residues (Pore water EEC)
     5. aqueous concentration of pesticide residues (water column EEC).

     This table contains no default values. The user should input values for each of these parameters in the "Value" column of Table 1.

     The titles of several tables displayed in the KABAM tool are designed to automatically insert the pesticide name as entered in Table 1.

     Of all parameters incorporated into KABAM, Log K_OW has the greatest influence on estimates of bioaccumulation in aquatic organisms (see section A.7 of Appendix A). As a result, this parameter is the most important for estimating potential exposures of mammals and birds to pesticides through consumption of contaminated aquatic organisms. Estimates of Log K_OW can be obtained for a pesticide from acceptable or supplemental registrant-submitted studies (OPPTS Guidelines 830.7550, 830.7560, 830.7570) and from scientific literature. One useful source for locating Log K_OW data in the scientific literature is Sangster (2007). Before using data from this database, the scientist should review the original citation and determine whether the data are acceptable or supplemental. If no measured values of Log K_OW are available, this value can be estimated using EPI Suite software that includes KOWIN (USEPA 2009), which considers contributions of the molecule's individual fragments to the overall Log K_OW. If a range of Log K_OW values is available, it is suggested that the model user input the high and low Log K_OW values separately in order to bracket the bioaccumulation potential and its associated risks. Bioaccumulation potential increases as Log K_OW increases. General guidance for evaluating measured and estimated Log K_OW data is available in Appendix B of USEPA 2003.

     In Table 1 of the KABAM tool (reproduced below), K_OW is automatically calculated as 10 to the power of the Log K_OW value that is entered by the model user. The K_OW is used to estimate uptake and clearance rate constants that define the concentrations of the pesticide in the tissues of the aquatic organisms.

     K_OC data can be obtained from registrant-submitted studies (OPPTS Guidelines 835.1230, 835.1240) or from the scientific literature. As the K_OC value of a chemical increases, the estimated accumulation of a chemical also increases. The user should select the K_OC value input into PRZM/EXAMS for deriving aquatic and benthic EECs. Input parameter guidance for PRZM/EXAMS indicates that the K_OC parameter value should be calculated as "the average K_OC from batch experiments" (USEPA 2002). If no scientifically valid estimates of K_OC are available, this parameter value can be estimated as 0.35*K_OW (USEPA 2004b).

     Table 1
     Chemical Characteristics of Pesticide X

     Characteristic	Value	Guidance
     Pesticide Name	Pesticide X	Required input
     Log KOW	5	Required input Enter value from acceptable or supplemental study submitted by registrant or available in scientific literature.
     KOW	100000	No input necessary. This value is calculated automatically from the Log KOW value entered above.
     KOC(L/kg OC)	25000	Required input Input value used in PRZM/EXAMS to derive EECs. Follow input parameter guidance for deriving this parameter value (USEPA 2002).
     Time to steady state (TS; days)	30	No input necessary. This value is calculated automatically from the Log KOW value entered above.
     Pore water EEC (µg/L)	5	Required input Enter value generated by PRZM/EXAMS benthic file. PRZM/EXAMS EEC represents the freely dissolved concentration of the pesticide in the pore water of the sediment. The appropriate averaging period of the EEC is dependent on the specific pesticide being modeled and is based on the time it takes for the chemical to reach steady state. Select the EEC generated by PRZM/EXAMS which has an averaging period closest to the time to steady state calculated above. In cases where the time to steady state exceeds 365 days, the user should select the EEC representing the average of yearly averages. The peak EEC should not be used.
     Water Column EEC (µg/L)	6	Required input Enter value generated by PRZM/EXAMS water column file. PRZM/EXAMS EEC represents the freely dissolved concentration of the pesticide in the water column. The appropriate averaging period of the EEC is dependent on the specific pesticide being modeled and is based on the time it takes for the chemical to reach steady state. The averaging period used for the water column EEC should be the same as the one selected for the pore water EEC (discussed above).

     Note: Table 1 of this User's Guide contains example data for chemical specific characteristics.

     The time to steady state (T_S; in days) is also calculated automatically by the KABAM tool according to Equation 1 (Hawker and Connell 1988). This equation is consistent with recommendations provided in USEPA and OECD guidelines for fish BCF studies for determining the time to reach steady state (USEPA 1996, OECD 1996). It should be noted that there is uncertainty in using this equation for chemicals with Log K_OW > 6, since this falls outside of the range of data used to derive this relationship. Alternatively, the time to steady state can be defined using empirical data from available BCF studies that were sufficient to define steady state. This information can be used to supplement the calculated T_S value.

     Equation 1

     T_S = [(6.54 x 10^-3)(K_OW) + 55.31] / 24

     EECs from PRZM (v3.12.2, May 2005) and EXAMS (v2.98.4.6, April 2005) (coupled with the input shell pe5.pl, dated Aug 2007) are used in the KABAM tool. EECs generated by PRZM/EXAMS represent the freely dissolved concentration of the pesticide in the surface and pore water of the standard pond. The bioaccumulation portion of KABAM assumes that the aquatic environment is at steady state. Because the time to reach steady state is pesticide specific, the appropriate averaging period of the EEC should be determined on a chemical by chemical basis. Generally, the time to reach steady state can be related to the Log K_OW of a chemical, with increasing time required as the Log K_OW increases. Therefore, it is not relevant to use short-term (peak) estimates of pesticides in the aquatic environment. The EEC used to represent the concentration of the pesticide in the pore and surface waters of the aquatic habitat should be selected so that the averaging period (i.e., 4-d, 21-d, 90-d, 1 year), is consistent with the time to steady state estimated for that chemical. For example, a chemical with a Log K_OW = 5 would have an estimated time to steady state value of 30 days. Since the standard output file from PRZM/EXAMS does not include a 30-d average, the next closest averaging period would be selected (either 21 or 60 days). Therefore, the EEC represented by the 21-day average would be selected for this chemical. In cases where the time to steady state exceeds 365 days, the user should select the EEC representing the yearly EEC.

     In cases where multiple uses of a single pesticide are possible (e.g., cotton, corn, apples), EECs from the different uses can be modeled to allow for an understanding of the bioaccumulation and associated risks associated with different uses.

     Table 2

     KABAM automatically calculates uptake and elimination constants through respiration (k₁ and k₂, respectively) and diet (k_D and k_E, respectively). In using the model with its default parameters in place, it is assumed that the elimination of the pesticide from aquatic organisms through metabolism does not occur (i.e., metabolism rate constant k_M = 0).

     In Table 2 (reproduced below) of the KABAM tool, the model user can enter measured rate constants for uptake and elimination constants. These data can be obtained from acceptable or supplemental studies submitted by the registrant or from the literature. For example, k₁ and k₂ rate constants for fish can be obtained from pesticide BCF studies submitted for the fish (OPPTS Guideline 850.1730). However, caution should be used when altering rate constants. For example, the k₂ from a bioconcentration study typically represents a total elimination half life. However, the k₂ in KABAM represents elimination from the gills. Therefore, incorporation of a measured k₂ into KABAM without consideration of other elimination pathways may result in erroneous results. In order to run the model, it is not necessary for the user to alter the default values inserted into Table 2. If the model user alters the parameters in Table 2 of the KABAM tool, they will be highlighted yellow.

     Table 2
     Input Parameters for Rate Constants
     ("calculated" indicates that model will calculate rate constant)

     Trophic level	k1 (L/kg*d)	k2 (d-1)	kD (kg-food/kg-org/d)	kE (d-1)	kM* (d-1)
     phytoplankton	calculated	calculated	0*	0*	0
     zooplankton	calculated	calculated	calculated	calculated	0
     benthic invertebrates	calculated	calculated	calculated	calculated	0
     filter feeders	calculated	calculated	calculated	calculated	0
     small fish	calculated	calculated	calculated	calculated	0
     medium fish	calculated	calculated	calculated	calculated	0
     large fish	calculated	calculated	calculated	calculated	0

     ^*Default value is 0.
     k₁ and k₂ represent the uptake and elimination constants respectively, through respiration.
     k_D and k_E represent the uptake and elimination constants, respectively, through diet.
     k_M represents the metabolism rate constant.

     The model user should exercise caution when using a value of k_M > 0, as this approach will decrease predicted EECs and RQs for mammals and birds. Initially, k_M should be set to 0 as a screen. The assumption that there is no metabolism of the pesticide within aquatic organisms is conservative. If no metabolism is observed in available fish BCF studies, then k_M should not be altered. In cases where metabolism occurs, this assumption can result in overestimates of pesticide accumulation in tissues of aquatic organisms. In cases where the model user has evidence to indicate that metabolism may occur in fish (i.e., from BCF studies) and RQ values exceed LOCs, then the user can estimate k_M using approaches described in Appendix H. This will allow the model user to characterize effects of metabolism on bioaccumulation in aquatic ecosystems and associated risks to mammals and birds consuming aquatic organisms.

     Table 3

     To calculate risk quotients, user-supplied avian and mammalian toxicity endpoints should be entered into Table 3 (reproduced below). Acceptable or supplemental registrant-submitted or open literature studies should be used to define the effects of the pesticide on birds and mammals. Required input data include: avian acute oral LD₅₀ (OPPTS Guideline 850.2100), avian subacute dietary LC₅₀ (OPPTS Guideline 850.2200), avian reproduction (expressed as a NOAEC or a NOAEL) (OPPTS Guideline 850.2300), mammalian acute oral LD₅₀ (OPPTS Guideline 870.1100), or subacute dietary LC₅₀ (if available), and mammalian reproduction NOAEC or NOAEL (OPPTS Guideline 870.3800).

     Table 3
     Mammalian and Avian Toxicity Data for Pesticide X
     (These are required inputs.)

     Animal	Measure of effect (units)	Value	Species	If selected species is "other," enter body weight (in kg) here.
     Avian	LD50 (mg/kg-bw)	50	mallard duck
     	LC50 (mg/kg-diet)	500	Northern bobwhite quail
     	NAOEC (mg/kg-diet)	10	mallard duck
     	Mineau Scaling Factor	1.15	Default value for all species is 1.15 (for chemical specific values, see Mineau et al. 1996).
     Mammalian	LD50 (mg/kg-bw)	50	other	1.2
     	LC50 (mg/kg-diet)	N/A	other
     	Chronic Endpoint	10	laboratory rat
     	units of chronic endpoint*	ppm

     ^*ppm = mg/kg-diet
     Note: Table 3 of this User's Guide contains example data for chemical specific characteristics.

     In the appropriate cell under the "value" column of Table 3, the user should input the lowest (most sensitive) available toxicity data for each toxicity endpoint. If an endpoint value is not discrete (i.e., contains a ">" symbol), the whole number should be entered as a discrete value, keeping in mind that all resulting risk quotient (RQ) values derived using this endpoint are "<". For the chronic mammalian data, the user must also select the units of the value. The user should select units from the drop down menu as either "ppm" or "mg/kg-bw."

     Under the "species" column, the user should use the drop down menu to select the appropriate test species associated with the toxicity value entered in the adjacent cell in the "value" column. If the test species is not one of the options available in the drop down list, the model user should select "other" as the test species. If "other" is selected, the user must enter the body weight (in kg) of the test species. In the case that "other" is selected as the test species, a message will appear in the spreadsheet below Table 3 to alert the user of the need to enter the body weight of the test species. These data should be obtained from the study report if possible (time weighted average of control animals). Alternatively, reference body weight values may be obtained from a variety of sources, including U.S. EPA 1993 and Dunning 1984. Failure to enter the body weight of the test species when it is entered as "other" will prevent calculation of risk quotients that correspond to that endpoint.

     If available, the model user should enter chemical-specific data to represent the avian scaling factor (see Mineau et al. 1996). If no chemical specific data are available, the default value of 1.15 should be entered. This value is used to adjust avian dose-based toxicity values based on the weight of the species of concern (e.g., herons) as described in the T-REX User's Guide (USEPA 2008a) and in Appendix G.

     ---

   • 2.2 Ecosystem Inputs

     In order to estimate pesticide concentrations in tissues of aquatic organisms, biotic and abiotic characteristics of the model aquatic ecosystem must be defined. In addition, the mammals and birds consuming aquatic organisms are also defined as ecosystem inputs.

     To run KABAM, it is not necessary to alter any of the default parameters in the "Ecosystem Inputs" worksheet.

     If the model user alters default parameter values, they will be highlighted yellow in the KABAM tool.

     It may be necessary for the model user to incorporate alternate ecosystem input values if the modeling incorporates EECs from a source other than PRZM/EXAMS (e.g., from a mesocosm study). In that case, the model user should enter parameter values that correspond to the specific water body used.

     Table 4

     Abiotic characteristics of the aquatic ecosystem that are necessary for KABAM are defined in Table 4 (reproduced below) of the model tool. These characteristics include: concentrations of particulate organic carbon (X_POC), dissolved organic carbon (X_DOC), dissolved oxygen (C_OX) and suspended solids (C_SS), water temperature (T), and % organic carbon (OC) content of the sediment. The model tool is populated with default values for these parameters, which can be altered based on the needs of the model user. Default values relevant to the abiotic characteristics of the aquatic ecosystem are designed to be consistent with the OPP standard pond scenario used in EXAMS. Brief explanations for these default values as well as guidance on selecting alternative values are provided in Appendix B.

     Table 4
     Abiotic Characteristics of the Model Aquatic Ecosystem

     Characteristic (symbol; units)	Value	Guidance*
     Concentration of Particulate Organic Carbon (XPOC; kg OC/ L)	0	When using EECs generated by PRZM/EXAMS, use a value of "0" for both POC and DOC.
     Concentration of Dissolved Organic Carbon (XDOC; kg OC/L)	0
     Concentration of Dissolved Oxygen (COX; mg O2/L)	5.0	Default value is 5.0 mg O2/L when using EECs generated by PRZM/EXAMS.
     Water Temperature (T; °C)	15	Value is defined by the average water temperature of the EXAMS pond when using EECs generated by PRZM/EXAMS. Model user should consult output file of EXAMS to define this value.
     Concentration of Suspended Solids (CSS; kg/L)	3.00E-05	Default value is 3.00x10-5 kg/L when using EECs generated by PRZM/EXAMS.
     Sediment Organic Carbon (OC; %)	4.0%	Default value is 4.0% when using EECs generated by PRZM/EXAMS.

     ^*When using pesticide concentrations from monitoring data or mesocosm studies, consult Appendix B of the User's Guide for specific guidance on selecting values for these parameters.

     Table 5

     Necessary biotic components of the aquatic ecosystem define characteristics of the sediment and water column biota. These include body weights and body compositions, specifically % lipids, % NLOM (non-lipid organic matter), and % water. These values are defined for the seven trophic levels of the aquatic ecosystem (phytoplankton, zooplankton, benthic invertebrates, filter feeders, small fish, medium fish and large fish) modeled by KABAM in Table 5 (reproduced below) of the tool. Default values for these biotic parameters are displayed in Table 5 below. A description of how these default parameters were selected is available in Appendix C. In addition, Table 5 allows the model user to define whether organisms within each trophic level respire pore water. If yes, it is assumed that 5% of the total respired water is from pore water. Default assumptions related to respiration of pore water for each trophic level are depicted in Table 5 below.

     Table 5
     Characteristics of Aquatic Biota of the Model Ecosystem

     Trophic level	Wet Weight (kg)	% lipids	% NLOM	% Water	Do organisms in trophic level respire some pore water?
     Sediment*	N/A	0.0%	4.0%	96.0%	N/A
     Phytoplankton	N/A	2.0%	8.0%	90.0%	no
     Zooplankton	1.0E-07	3.0%	12.0%	85.0%	no
     Benthic invertebrates	1.0E-04	3.0%	21.0%	76.0%	yes
     Filter feeders	1.0E-03	2.0%	13.0%	85.0%	yes
     Small fish	1.0E-02	4.0%	23.0%	73.0%	yes
     Medium fish	1.0E-01	4.0%	23.0%	73.0%	yes
     Large fish	1.0E+00	4.0%	23.0%	73.0%	no

     ^*Note that sediment is not a trophic level. It is included in this table because it is consumed by aquatic organisms of the KABAM food web.
     N/A = not applicable

     Table 6

     Table 6 (reproduced below) of the KABAM tool allows the model user to define the diet composition of each of the trophic levels of the aquatic ecosystem. The aquatic trophic levels are assigned a hierarchy, which is relevant to the assignment of diet composition. The order of the trophic levels, in increasing hierarchy, is as follows: phytoplankton, zooplankton, benthic invertebrates, filter feeders, small fish, medium fish, and large fish. The diet of each aquatic trophic level is composed of sediment or water column biota from lower trophic levels. The KABAM tool does not allow the model user to assign a portion of the diet of one organism to its own trophic level or to trophic levels that are higher. The default values defining the diet of each trophic level are in Table 6 below. An explanation of how these default parameters were determined is available in Appendix C .

     Note that the total diet of each organism within the aquatic food web should equal 100%. If the total diet ≠ 100%, an error message will appear under Table 6.

     Table 6
     Diets of Aquatic Biota of the Model Ecosystem

     Trophic Level in diet	Diet for:
     	Zoo plankton	Benthic Invertebrates	Filter Feeder	Small Fish	Medium Fish	Large Fish
     Sediment*	0.0%	34.0%	34.0%	0.0%	0.0%	0.0%
     Phytoplankton	100.0%	33.0%	33.0%	0.0%	0.0%	0.0%
     Zooplankton		33.0%	33.0%	50.0%	0.0%	0.0%
     Benthic invertebrates			0.0%	50.0%	50.0%	0.0%
     Filter feeders				0%	0%	0.0%
     Small fish					50.0%	0.0%
     Medium fish						100.0%
     Total	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

     ^*Note that sediment is not a trophic level. It is included in this table because it is consumed by aquatic organisms of the KABAM food web.

     Table 7

     Table 7 (reproduced below) of the KABAM tool allows the model user to define the mammalian and avian species of concern, as well as their body weights. Species are considered to be of concern for pesticide exposures through consumption of residues in freshwater aquatic animals that serve as prey.

     For mammals, default species include the fog shrew (Sorex sonomae), the water shrew (S. palustris), the rice rat (Oryzomys palustris), the star-nosed mole (Condylura cristata), the American mink (Neovison vison), and the Northern river otter (Lontra canadensis). For birds, default species include sandpipers, rails, herons, kingfisher, ducks, grebes, ibis, rails, cormorants, osprey, cranes, bald eagles (Haliaeetus leucocephalus) and pelicans. Descriptions of how mammalian and avian species were selected, including their body weights, are provided in Appendices D and E, respectively. These appendices also provide descriptions of the species themselves as well as justifications for default parameters used to represent the species in KABAM (i.e., body weight and diet).

     The selected body weight value influences estimates of pesticide exposure through differential consumption of contaminated food items, as well as dose-based toxicity values. Therefore, the magnitude of the body weight parameter has an effect on the magnitude of the dose-based RQ. For mammals, higher body weight values result in higher dose-based RQs (keeping the diet constant). As a result, default body weight values for the fog shrew, water shrew, rice rat, and star-nosed mole were selected as higher values of relevant ranges in order to represent size classes that would be most vulnerable to exposures through bioaccumulation. In order to bound the risk of accumulated residues to mink and river otter, the lowest and highest body weights of these species were selected as defaults. For birds, higher body weight results in lower RQs. In order to bound the risk of accumulated residues to birds, the lowest and highest body weights of birds with the same diet were selected as defaults. The user can alter the assigned body weights to represent the low and high end of possible weights in order to bound the potential RQs for a particular species. Additional data on body weights of species of mammals and birds are provided in Appendices D and E, respectively.

     Table 7
     Identification of Mammals and Birds Feeding on Aquatic Biota of the Model Ecosystem

     Mammal/Bird #	Name	Body weight (kg)
     Mammal 1	Fog/Water shrew	0.018
     Mammal 2	Rice Rat/Star-nosed mole	0.085
     Mammal 3	Small mink	0.450
     Mammal 4	Large mink	1.800
     Mammal 5	Small river otter	5.000
     Mammal 6	Large river otter	15.000
     Bird 1	Sandpipers	0.02
     Bird 2	Cranes	6.7
     Bird 3	Rails	0.07
     Bird 4	Herons	2.90
     Bird 5	Small osprey	1.25
     Bird 6	White pelican	7.50

     Tables 8 and 9

     Tables 8 and 9 (reproduced below) of the KABAM tool allow the model user to define the diet composition of the mammals and birds of concern that are defined in Table 7. The animal names entered in Table 7 will appear at the heads of the columns of Tables 8 and 9. The diet of each mammal and bird species is attributed to a portion of each trophic level of the aquatic ecosystem. Justifications for the default diets for each mammal and bird species are provided in Appendices D and E, respectively. Note that the total diet of each mammal and bird should equal 100%. If not, an error message will appear under Table 8 or 9.

     Table 8
     Diets of Mammals Feeding on Aquatic Biota of the Model Ecosystem

     Trophic level in diet	Diet for:
     	Fog/Water Shrew	Rice Rat/Star-nosed mole	Small Mink	Large Mink	Small River Otter	Large River Otter
     Phytoplankton	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%
     Zooplankton	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%
     Benthic invertebrates	100.0%	34.0%	0.0%	0.0%	0.0%	0.0%
     Filter feeders	0.0%	33.0%	0.0%	0.0%	0.0%	0.0%
     Small fish	0.0%	33.0%	0.0%	0.0%	0.0%	0.0%
     Medium fish	0.0%	0.0%	100.0%	100.0%	100.0%	0.0%
     Large fish	0.0%	0.0%	0.0%	0.0%	0.0%	100.0%
     Total	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

     Table 9
     Diets of Birds Feeding on Aquatic Biota of the Model Ecosystem

     Trophic level in diet	Diet for:
     	Sandpipers	Cranes	Rails	Herons	Small Osprey	White pelican
     Phytoplankton	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%
     Zooplankton	0.0%	0.0%	0.0%	0.0%	0.0%	0.0%
     Benthic invertebrates	33.0%	33.0%	50.0%	50.0%	0.0%	0.0%
     Filter feeders	33.0%	33.0%	0.0%	0.0%	0.0%	0.0%
     Small fish	34.0%	0.0%	50.0%	0.0%	0.0%	0.0%
     Medium fish	0.0%	34.0%	0.0%	50.0%	100.0%	0.0%
     Large fish	0.0%	0.0%	0.0%	0.0%	0.0%	100.0%
     Total	100.0%	100.0%	100.0%	100.0%	100.0%	100.0%

     If the model user chooses to alter the default diet of a mammal or bird, the model user should consider the daily food intake for determining appropriate aquatic trophic levels to include within an animal's diet. The user should verify that the weight of an individual dietary item does not greatly exceed the daily food intake of the mammal or bird. This will prevent the user from simulating a bird or mammal that consumes prey that are much larger than could be reasonably consumed. This can be determined using allometric equations for estimating daily food intake, as described in Appendices D and E. In addition, these appendices contain data defining the daily food intake for several species of birds and mammals.

     Pesticide exposures to mammals and birds through consumption of contaminated aquatic organisms are determined by weighing the exposure concentration by the contribution of each food item to the total diet. While this approach is reasonable for chronic exposures, it may underestimate acute exposures resulting from consumption of larger trophic level organisms within short periods of time. In order to explore high-end exposure concentrations and subsequent risks resulting from acute exposures, the model user can set the highest aquatic trophic level consumed by a bird or mammal to 100%. For example, high-end acute exposures of cranes (which consume benthic invertebrates, filter feeders, and medium fish) to a pesticide could be assessed by setting the crane diet to 100% of medium fish.

   ---

3. Parameters & Calculations

   Also included in the KABAM tool is a tabularized summary of the relevant parameters for the bioaccumulation portion of KABAM. This summary is included in a separate worksheet, titled "Parameters & Calculations" (Table 10 of the KABAM tool) and represents values used to calculate pesticide tissue concentrations for the trophic levels of the aquatic ecosystem. This worksheet is locked (read only) in the KABAM tool and cannot be altered by the model user; however, this worksheet can be printed by the model user or copied into a risk assessment as a model output. A full description of the parameters contained in Table 10 of the KABAM tool as well as the equations used to calculate these parameters can be found in Appendix A.

   ---

4. Model Results

   The final outputs of KABAM include Bioconcentration Factors (BCFs), Bioaccumulation Factors (BAFs), Biomagnification Factors (BMFs), Biota-Sediment Accumulation Factors (BSAFs), estimates of pesticide concentrations in tissues of aquatic organisms, and RQ values for mammals and birds consuming contaminated aquatic organisms.

   Note that the "results" worksheet of KABAM is locked (read only) and cannot be altered by the model user, with the exception of format changes (e.g., number of decimal places). Also, the KABAM tool does not automatically account for significant figures. The format of numerical values in the Tool can be altered by the user to increase or decrease the number of decimal places.

   Table 11 and Figure 1

   Table 11 (reproduced below) of the KABAM tool reports pesticide concentrations in tissues of aquatic organisms on both a total body weight and lipid normalized basis. The table also reports contributions of the pesticide concentration in tissue from respiration and from diet. These values are useful for understanding the dominant uptake route of the pesticide that influences bioaccumulation. Figure 1 (reproduced below) of the KABAM tool graphically represents the relative contributions of pesticide uptake through diet and through respiration to the overall concentrations of the pesticide in the tissues of the different aquatic animals.

   Table 11
   Estimated Concentrations of Pesticide X in Ecosystem Components

   Ecosystem Component	Total concentration (µg/kg-ww)	Lipid normalized concentration (µg/kg-lipid)	Contribution due to diet (µg/kg-ww)	Contribution due to respiration (µg/kg-ww)
   Water (total)*	6	N/A	N/A	N/A
   Water (freely dissolved)*	6	N/A	N/A	N/A
   Sediment (pore water)*	5	N/A	N/A	N/A
   Sediment (in solid)**	5,000	N/A	N/A	N/A
   Phytoplankton	27,298	1364913	N/A	27,298.25
   Zooplankton	21,065	702157	651.72	20,412.98
   Benthic Invertebrates	23,678	789265	1,812.95	21,865.01
   Filter Feeders	15,549	777440	1,167.92	14,380.88
   Small Fish	34,713	867830	7,246.79	27,466.40
   Medium Fish	41,050	1026242	14,492.66	26,557.01
   Large Fish	56,332	1408297	30,795.48	25,536.39

   ^* Units: µg/L; ^**Units: µg/kg-dw
   Note: Table 11 of this User's Guide contains example results based on example chemical specific data entered in Tables 1 and 3.

   

   Note: Figure 1 of this User's Guide contains example results based on example chemical-specific data entered in Tables 1 and 3.

   Tables 12 and 13

   BCF, BAF, BMF and BSAF are calculated by KABAM (Tables 12 and 13). These terms are intended to provide a relative measure of the pesticide concentration in an organism to the pesticide concentration in sources (i.e., the environment and the diet) of that pesticide. Appendix F contains the equations used to calculate BCF, BAF, BMF and BSAF.

   Table 12
   Total BCF and BAF Values of Pesticide X in Aquatic Trophic Levels

   Trophic Level	Total BCF (µg/kg-ww)/(µg/L)	Total BAF (µg/kg-ww)/(µg/L)
   Phytoplankton	4801	4550
   Zooplankton	3421	3511
   Benthic Invertebrates	3705	3946
   Filter Feeders	2435	2591
   Small Fish	4766	5786
   Medium Fish	4766	6842
   Large Fish	4806	9389

   Note: Table 12 of this User's Guide contains example results based on example chemical specific data entered in Tables 1 and 3.

   Table 13
   Lipid-normalized BCF, BAF, BMF and BSAF Values of Pesticide X in Aquatic Trophic Levels

   Trophic Level	BCF (µg/kg-lipid)/(µg/L)	BAF (µg/kg-lipid)/(µg/L)	BMF (µg/kg-lipid)/(µg/kg-lipid)	BSAF (µg/kg-lipid)/(µg/kg-OC)
   Phytoplankton	240045	227485	N/A	11
   Zooplankton	114028	117026	0.51	6
   Benthic Invertebrates	123488	131544	1.16	6
   Filter Feeders	121769	129573	1.14	6
   Small Fish	119142	144638	1.16	7
   Medium Fish	119142	171040	1.24	8
   Large Fish	120143	234716	1.37	11

   Note: Table 13 of this User's Guide contains example results based on example chemical specific data entered in Tables 1 and 3.

   Tables 14, 15, and 16

   Tables 14, 15, and 16 (reproduced below) of the KABAM tool summarize the estimated exposure values, mammal and bird toxicity values and resulting RQ values, respectively, used to estimate potential risks to mammals and birds that consume aquatic animals contaminated with pesticides accumulated through the aquatic food chain.

   Table 14 uses the mammalian and avian body weights (entered by the model user) to calculate the dry food ingestion and drinking water intake rates according to allometric equations specific to mammals and birds. The wet food intake is calculated using the dry food intake and the % water of the diet. Dose-based EECs represent the sum of pesticide intake through diet and through drinking water, accounting for pesticide concentrations in diet items and in water and food and water intake rates. Dietary-based EECs represent the sum of pesticide intake through diet only, without consideration of species specific intake rates or body weights. Descriptions of the equations used to calculate food intake rates, water intake rates, dose-based EECs, and dietary-based EECs are available in Appendix G.

   Table 14
   Calculation of EECs for Mammals and Birds Consuming Fish Contaminated by Pesticide X

   Wildlife Species		Biological Parameters				EECs (pesticide intake)
   		Body Weight (kg)	Dry Food Ingestion Rate (kg-dry food/kg-bw/day)	Wet Food Ingestion Rate (kg-wet food/kg-bw/day)	Drinking Water Intake (L/d)	Dose Based (mg/kg-bw/d)	Dietary Based (ppm)
   Mammalian	Fog/water shrew	0.02	0.140	0.585	0.003	13.857	23.68
   	Rice rat/star-nosed mole	0.1	0.107	0.484	0.011	11.921	24.64
   	Small mink	0.5	0.079	0.293	0.048	12.041	41.05
   	Large mink	1.8	0.062	0.229	0.168	9.408	41.05
   	Small river otter	5.0	0.052	0.191	0.421	7.844	41.05
   	Large river otter	15.0	0.042	0.157	1.133	8.852	56.33
   Avian	Sandpipers	0.0	0.228	1.034	0.004	25.5861	24.75
   	Cranes	6.7	0.030	0.136	0.211	3.6561	26.90
   	Rails	0.1	0.147	0.577	0.010	16.8571	29.20
   	Herons	2.9	0.040	0.157	0.120	5.0943	32.36
   	Small osprey	1.3	0.054	0.199	0.069	8.1859	41.05
   	White pelican	7.5	0.029	0.107	0.228	6.0108	56.33

   Note: Table 14 of this User's Guide contains example results based on example chemical specific data entered in Tables 1 and 3.

   Table 15 (reproduced below) of the KABAM tool summarizes the acute and chronic, dose-based and dietary-based toxicity values representing effects of a pesticide to mammals and birds. Dietary-based toxicity values are taken directly from user inputs in Table 3, without adjustment. Available dose-based toxicity values are adjusted for the weights of the animal tested (e.g., laboratory rat, mallard duck) and of the animal for which the risks are being assessed (e.g., mink, bald eagle). Methods for adjusting toxicity values are consistent with those used by T-REX (USEPA 2008a). A full description of the methodology for adjusting dose-based toxicity values is provided in Appendix G.

   Table 15
   Calculation of Toxicity Values for Mammals and Birds Consuming Fish Contaminated by Pesticide X

   Wildlife Species		Toxicity Values
   		Acute		Chronic
   		Dose Based (mg/kg-bw)	Dietary Based (mg/kg-diet)	Dose Based (mg/kg-bw)	Dietary Based (mg/kg-diet)
   Mammalian	Fog/water shrew	142.87	N/A	1.05	10
   	Rice rat/star-nosed mole	96.92	N/A	0.71	10
   	Small mink	63.89	N/A	0.47	10
   	Large mink	45.18	N/A	0.33	10
   	Small river otter	35.00	N/A	0.26	10
   	Large river otter	26.59	N/A	0.20	10
   Avian	Sandpipers	25.96	500.00	N/A	100
   	Cranes	62.10	500.00	N/A	100
   	Rails	31.33	500.00	N/A	100
   	Herons	54.77	500.00	N/A	100
   	Small osprey	48.27	500.00	N/A	100
   	White pelican	63.16	500.00	N/A	100

   Note: Table 15 of this User's Guide contains example results based on example chemical specific data entered in Tables 1 and 3.

   Table 16 (reproduced below) of the KABAM tool presents RQs, which are the ratio of exposure concentrates to effects values. RQ values are then compared to Agency levels of concern (LOCs) for non-listed and listed mammals and birds. For acute exposures, the LOC is 0.5 for (non-listed) birds and mammals and 0.1 for federally-listed threatened and endangered (listed) species of mammals and birds. For chronic risk, the LOC is 1.0 for all species (non-listed and listed) mammals and birds (USEPA 2004). RQ values that exceed their respective LOC values appear in RED in Table 16 in the KABAM tool and in BOLD here.

   Dose-based and dietary-based RQs are not equivalent. Dietary-based RQs are calculated by directly comparing the concentration of a pesticide administered to experimental animals in the diet in a toxicity study to the concentration estimated in selected food items. These RQs do not account for the fact that smaller-sized animals need to consume more food relative to their body weight than larger animals. The dose-based RQs account for these factors by incorporating the ingestion rate-adjusted exposure from the various food items to the different weight classes of assessed animals and the weight class-scaled toxicity endpoints.

   Table 16
   Calculation of RQ Values for Mammals and Birds Consuming Fish Contaminated by Pesticide X

   Wildlife Species		Acute		Chronic
   		Dose Based	Dietary Based	Dose Based	Dietary Based
   Mammalian	Fog/water shrew	0.097	N/A	13.198	2.368
   	Rice rat/star-nosed mole	0.123	N/A	16.737	2.464
   	Small mink	0.188	N/A	25.643	4.105
   	Large mink	0.208	N/A	28.335	4.105
   	Small river otter	0.224	N/A	30.498	4.105
   	Large river otter	0.333	N/A	45.296	5.633
   Avian	Sandpipers	0.986	0.049	N/A	0.247
   	Cranes	0.059	0.054	N/A	0.269
   	Rails	0.538	0.058	N/A	0.292
   	Herons	0.093	0.065	N/A	0.324
   	Small osprey	0.170	0.082	N/A	0.410
   	White pelican	0.095	0.113	N/A	0.563

   Note: Table 16 of this User's Guide contains example results based on example chemical specific data entered in Tables 1 and 3.

   EECs and RQs for birds are based on the selected body weight of the bird as well as its diet. Default values for birds were designed to represent birds on the low and high end of weights with three different diets. Birds consuming benthic invertebrates, filter feeders, and fish include sandpipers, ducks and cranes (default birds 1 and 2, which are named sandpipers and cranes, respectively). Birds consuming benthic invertebrates and fish include belted kingfisher, rails, ibis, grebes, double-breasted cormorants, bitterns, egrets, and herons (default birds 3 and 4, which are named rails and herons, respectively). Birds consuming fish include osprey, bald eagles, and the white pelican (default birds 5 and 6, which are named small osprey and white pelican, respectively). In the case that RQs exceed the LOC for both birds within a feeding group, then it can be assumed that RQs would exceed the LOC for all of the birds within the feeding category since birds on the low and high end of the weight ranges have RQs of concern. In the case that RQs exceed the LOC for the default bird with the high body weight of a feeding category (i.e., birds 2, 4, and 6), the model user can refine the EECs and RQs to be representative of specific bird species within a feeding category by entering specific body weights of individual species of concern. Appendix E contains species specific data on feeding habits and body weights of over 40 species of birds, including some listed species, which consume aquatic animals from freshwater habitats.

   ---

5. Assessing Pesticide Concentration in Fish Tissues for Human Consumption

   It is possible to use KABAM to derive pesticide concentrations in edible tissues of fish that are relevant to assessments of pesticide risks to human health. Current default values described above for % lipid content of fish applies to the whole fish; however, not all fish tissues are consumed by humans. Therefore, it is necessary to modify the output of the pesticide tissue concentration to account for a lower % lipid composition of edible tissues. This can be accomplished by entering in all the relevant default input parameters for KABAM as defined above. It may be necessary to explore different body weights of the large fish, based on those that would be expected to be consumed by humans.

   The relevant output is the lipid normalized concentration of the pesticide in the large fish (Table 11). This value can be converted to the total pesticide concentration in edible tissues by multiplying by the % lipid content of the edible tissues. The default value for lipid content in edible tissue of the large fish is 3%, based on USEPA 2003. The resulting value represents the concentration of pesticide in fish tissue (in µg/kg-ww) potentially consumed by humans. This value can then be used in conjunction with fish consumption rates to characterize risks of a pesticide to humans consuming contaminated fish.

   ---

6. Model Assumptions, Limitations, and Uncertainties

   There are several key assumptions and resulting uncertainties associated with modeling pesticide concentrations in tissues of aquatic organisms. The assumptions involve the equations of the model itself and the parameterization of those underlying equations. Appendix A describes the assumptions associated with the equations of the bioaccumulation model. In order to explore uncertainties associated with specific parameters and their influences on model outputs, a sensitivity analysis was conducted (see section A7 of Appendix A). This was used to define the parameters that have the greatest influence on model outputs (e.g., K_OW, water column, and pore water EECs). Appendices B and C describe the parameterization of the model, including the associated assumptions.

   In addition, the use of PRZM/EXAMS for deriving EECs in the surface and pore waters of the aquatic ecosystem introduces the assumptions and uncertainties associated with PRZM and EXAMS to KABAM.

   One major assumption associated with KABAM concerns the model's assumed steady state. Given the episodic nature of pesticide applications, sporadic peak exposures to aquatic organisms would be expected. For a chemical with a Log K_OW of approximately 5, comparison of the fish tissue EECs predicted using the steady state and dynamic bioaccumulation modeling with PRZM/EXAMS/Arnot and Gobas indicates predictions are similar (USEPA 2008b) when a 60-d average was selected for water and sediment concentrations as input to the steady state model. This result suggests that steady-state bioaccumulation modeling can provide useful predictions of bioaccumulation potential even with highly dynamic exposures, provided proper consideration of the averaging period associated with water and sediment concentrations is made.

   As discussed above, in using KABAM with default settings, it is assumed that the elimination of the pesticide from aquatic organisms through metabolism does not occur, i.e., the metabolism rate constant (k_M) = 0. In cases where pesticide metabolism does occur, this could overestimate pesticide bioaccumulation. Appendix H of this guide provides methods for estimating k_M for fish using empirical data provided for specific chemicals (from BCF studies). This approach can be used to characterize effects of metabolism, but should be used with caution.

   The Arnot and Gobas (2004) model is generally appropriate for chemicals with Log K_OW value ≥ 4 to ≤ 8. Uncertainty increases as the value increases above 8 because the model has generally been validated using chemicals with Log K_OW values within the range of 4 - 8. Making predictions for a chemical with a Log K_OW > 8 leads to uncertainty in model outputs because predictions are based upon extrapolations in its subroutines.

   For chemicals with Log K_ow < 4, exposure from food becomes insignificant because uptake and depuration across the gills controls the residue in the organism. Thus, there is no need to run a food web model for these chemicals. In these cases, available BCF data are sufficient to predict residues in the aquatic species.

   It is assumed that there is no predation within a trophic level of the aquatic food web (e.g., medium fish cannot prey upon medium fish). It is also assumed that mammals and birds only consume organisms from the aquatic system.