Northwest Coalition for Alternatives to Pesticides v. United States Environmental Protection Agency

Before HARRY PREGERSON IKUTA, Judges, SANDRA S. Circuit MOSKOWITZ,* BARRY T. District Judge. PREGERSON;

Opinion Judge Partial Concurrence and Partial Dissent by Judge IKUTA.

PREGERSON, Judge: Circuit environmental Petitioners two groups challenging the Environmental (“EPA”) Agency’s Protection establish- pesticides ment of tolerances for seven mostly vegetable crops. used on fruit and grant petition deny it in part, We part, and remand to the EPA. BACKGROUND Olsen, Colangelo and Erik D. Na- Aaron Council, Washing- Resource Defense tional I. D.C.; Davis, ton, Shelley Farm-worker regulates two pesticides The under D.C.; Fund, Washington, Patti Justice Insecticide, Fungi- statutes: the Federal Seattle, WA, Goldman, Earthjustice, cide, (“FIFRA”), 7 and Rodenticide Act petitioners-appellants Northwest Coalition Food, 136-136y, § U.S.C. and the Federal for Aternatives to Pesticides and National (“FDCA”), 21 Drug, and Cosmetic Act Resources Defense Council. § U.S.C. 346a. Fleuchaus, Environmental Jonathan J. FIFRA, in the Under sold D.C.; Agency, Washington, Protection Sue registered by the United States must be General, Woodridge, Attorney Ellen John may § EPA. 7 The EPA U.S.C. 136a. Cruden, Deputy Attorney Assistant C. register pesticide pesticide unless the Hanson, General, E. and Kent Environ- function perform will intended without Division, and Natural Resources ment causing “any unreasonable adverse effects Justice, Department States United environment.” U.S.C. D.C., Washington, respondent-appellee 136a(e)(5)(C). § Agency. Environmental Protection Weinstein, The authorizes the EPA to set FDCA Kenneth W. Claudia M. O’Brien, Sturkie, in food. pesticide Latham “tolerances” for residues and Cassandra * Moskowitz, Barry sitting by designation. The T. District Honorable California, Judge for the Southern District of 346a(b). II. § “A tolerance is the U.S.C. pesticide maximum allowable amount of Congress amended FDCA commodity.” may remain in or on a by enacting Quality the Food Protection (9th EPA, Nader v. 104-170, (“FQPA”), Pub.L. No. Act Cir.1988). “establish key provisions Stat. 1489. One leave in effect a tolerance for FQPA requires give spe- the EPA to only in or on a food if chemical residue posed cial consideration to risks to infants determines that the tolerance is [the EPA] establishing pesticide and children when 346a(b)(2)(A)(i). safe.” FQPA Specifically, tolerances. directs to mean that “there term “safe” is defined EPA “an additional1 tenfold to use certainty a that no harm will is reasonable margin safety ... to take into account aggregate exposure result from to the potential pre-and post-natal toxicity and residue, all pesticide including chemical respect completeness data with anticipated dietary exposures and all other toxicity and chil- exposure and to infants exposures for which there is reliable infor- 346a(b)(2)(A)(ii). 346a(b)(2)(C). § § mation.” 21 U.S.C. 21 dren.” U.S.C. This Any (or “10x”) tolerance that is not “safe” must be safety tenfold child factor is 21 modified or revoked. U.S.C. all presumptively tolerances. applied 346a(b)(2)(A)(i). § decisions, Thus, making in tolerance EPA must assume risk to children Tolerances are established rulemak- 346a(d). particular pesticide § from the use of a ing. 21 Pesticide man- U.S.C. rulemaking process ufacturers initiate the greater food is ten times than for adults. by petitioning the EPA to a tolerance. set may margin EPA a “use different 346a(d)(l). § pub- 21 U.S.C. The EPA safety for the chemical residue in petitions lishes notices of these the Fed- data, only if, on the basis reliable such 346a(d)(3). § Register. eral 21 U.S.C. margin will be safe for infants and chil- reviewing petition any After a com- added). (emphasis dren.” Id. it, EPA ments received on issue: Unfortunately, FQPA does not de- (1) (2) tolerance, establishing a final rule a dispute fine “reliable data.” The before us (3) rule, proposed denying a an order turns on the definition of this term. 346a(d)(4). petition. § 21 U.S.C. 2002, April Between 2001 and December EPA Once the takes final action on a published regulations seven es- any days petition, party sixty affected has tablishing pesticides tolerances for the ace- objections EPA to file with the and to seek fenhexamid, tamiprid, halosulfuron-methyl, evidentiary objections. hearing on those isoxadifen-ethyl, mepiquat, pymetrozine, § 346a(g)(2). U.S.C. The EPA’s final foods, zetacypermethrin many used on objections in response order to those fruits, nuts, including vegetables, subject judicial cereal review. 346a(h)(l). milk, § grains, eggs.2 regulation Each margin, safety range popula- 1. This is "additional” to two of sensitivities in the human tion). long factors that the has used to account for other a fac- uncertainties: tenfold rulings, 2. For the EPA’s individual tolerance (to interspecies tor for differences account for Tolerance, Acetamiprid; see Pesticide 67 Fed. possibility people suscepti- are more 14,649 27, Fenhexamid; (Mar. 2002); Reg. laboratory experi- ble than in animals studied Tolerance, 19,114 Fed.Reg. (Apr. Pesticide ments), tenfold factor for intra- 18, 2002); Halosulfuron-methyl; Pesticide (to species differences account for the wide Tolerance, 66,333 (Dec. Fed.Reg. objections response in to an indus- NRDC filed administrative promulgated was try petition to establish tolerances. publish- the tolerances in 2002. denying objec- ed its Final Order those presumptive apply The EPA did not August days, tions on Within any lOx child factor to of these petition filed a NRDC review The EPA reduced the pesticides. seven Circuit, petition Second and NCAP filed a lOx child factor to 3x for four of the fenhexamid, review this court. The Judicial Pan- (acetamiprid, isoxa- pesticides difen-ethyl, pymetrozine). Litigation For the re- el on Multidistrict transferred (halosulfuron, me- maining three petition for NRDC’s this court piquat, zeta-cypermethrin), pursuant to 28 U.S.C. and NRDC at all. apply did not joint appeal. and NCAP filed briefs on Natural Resources Defense STANDARD OF REVIEW (“NRDC”) Council, objections Inc. filed each of the tolerances based on the EPA’s does not The FDCA establish stan- to reduce or remove the child decision reviewing pesticide dard for tolerances safety argued factor.3 NRDC public that are without a established evi- allowing EPA not have “reliable data” did 21 dentiary hearing. See U.S.C. it to deviate from the child 346a(h)(2). § In the absence of such a 10, 2005, EPA August On issued its standard, governed by our review is sec- objections. final rejecting order NRDC’s tion 706 of the Administrative Procedure Denying Objections to Issu- See Order (“APA”). 706; § Act 5 City See U.S.C. Tolerances, 46,706 Fed.Reg. 70 ances of O’Neill, Sausalito v. 1205- 2005) (hereinafter (Aug. “Final Or- (9th Cir.2004). 06 der”). upheld The Final the 3x Order APA, Under the we must set aside acetamiprid, factor for fenhex- agency’s “arbitrary, an decision if it is amid, isoxadifenethyl, pymetrozine, discretion, capricious, an abuse of or other (i.e., and a lx child no wise not in accordance with law.” 5 U.S.C. factor) halosulfuron, mepiquat, and ze- 706(2)(A). agency § An decision would 46,711-13, 46,736. taeypermethrin. Id. at if normally arbitrary capricious be NRDC and the Northwest Coalition for has relied on factors which agency “the (“NCAP”) Alternatives to Pesticides each consider, Congress has not intended it to petitioned for review of the Final Order. entirely important failed to consider an JURISDICTION aspect problem, explana of the offered an decision that runs counter to tion for its FDCA, jurisdiction have under the We or is so agency, the evidence before the “any provides person which who will that it not implausible could be ascribed adversely be affected” an order product a difference in view or the denying objections may tolerance file a agency expertise.” Ass’n v. State petition Appeals for review in the Court of Mfrs’ Co., 29, 43, Mut. Auto. Ins. 463 U.S. 21 Farm days within 60 of the EPA’s order. (1983). 346a(h)(l). § 77 L.Ed.2d 103 S.Ct. 443 Tolerance, 6,422 Tolerance, ers; 2001); Fed.Reg. Isoxadifenethyl; Pesticide Pesticide 12,875 (Mar. 2002); 15, 2002). Fed.Reg. Mepi- (Feb. Tolerance, 3,113 quat; Fed.Reg. Pesticide (Jan. 23, 2002); Pymetrozine; Pesticide Tol- request evidentiary hear- 3. NRDC did not erance, 66,786 27, 2001); (Dec. Fed.Reg. ing objections. on its Zeta-cypermethrin and R-isonom- its Inactive FQPA, required under the As scope “The of review standard is nar will ‘arbitrary capricious’ whether infants and children monitors a court is not to substitute row and in unsafe levels of exposed be agency.” Id. judgment for that of However, the EPA lacks drinking water. “Nevertheless, agency must examine data for drinking exposure water sufficient and articulate a satisfac the relevant data including pesticides, several the ones including its action tory explanation for Therefore, issue this case. the facts ‘rational connection between computer modeling to determine used ” (cita made.’ Id. found and the choice pes- drinking exposure water seven omitted). explana “In reviewing tion modeling results re- Because the ticides. tion, must consider whether court] [the pesticide expo- little-to-no risk of vealed decision was based on consideration water, drinking reduced sure and whether there has the relevant factors safety factor for each or removed the child (cita judgment.” Id. been a clear error pesticide tolerances. omitted). tion argue that because the EPA Petitioners “uphold a decision Although we will modeling sampling instead of actual used clarity agency’s than if the of less ideal drinking data to determine discerned,” reasonably path be the EPA does not have pesticides, to these “may supply not a reasoned basis for justifying its use of a lower “reliable data” agency’s action that the itself has margin than lOx for the (citations omitted). given.” Id. 346a(b)(2)(C). tolerances. See U.S.C. short, argue modeling In Petitioners ANALYSIS can never constitute reliable data results challenge Petitioners seven tolerances FQPA. meaning within the fenhexamid, (acetamiprid, halosulfuron- methyl, isoxadifen-ethyl, mepiquat, pyme- disagree. have We Petitioners trozine, zeta-cypermethrin) based presented modeling no evidence that does gather drinking the EPA’s failure to water nothing yield not reliable data. There is exposure analysis. Peti- part data as its inherently unreliable about the use of mod challenge tioners also three tolerances See, e.g., els in scientific assessments. (acetamiprid, mepiquat, pymetrozine) Small Lead Phase-Down Task Refiner toxicity the absence of certain based on (D.C.Cir. Force v. EPA studies, and based on the EPA’s failure to 1983) (“[Administrative agencies have un explain why it had reliable data in the models.”). power predictive doubted to use toxicity absence of those studies.4 We ad- Moreover, difficulty in because of the sam challenges in turn. dress these pling supply, the entire nation’s water

I. necessary modeling is to determine wheth drinking er water has been contaminated FQPA EPA directs the to use an Topography, geology, and pesticides. margin additional tenfold to take hydrology greatly differ across the nation completeness into account of data with re- cases, constantly change. many spect pesticide “exposure to infants and 346a(b)(2)(C) (em- accurately computer modeling can more children.” 21 added). phasis incorporate provide these elements and however, knowledge, subsequent agency initially challenged 4. Petitioners also the tol- halosulfuronmethyl zeta-cy- erances for challenges moot. actions have rendered those permethrin grounds. They on these now ac- reliable data than actual water pesticides more sam- from likely to result in pling data can provide.5 exposures. such accomplish To this goal, the models are based on data from

Therefore, unwilling adopt we are studies at highly sites that are vulnera- narrow construction of “reliable data” that pesticides ble runoff of to surface Although FQPA Petitioners advance. data,” water leaching does not define “reliable we are ground modeling satisfy confident that results can water. If a fails this conserva- statutory requirement. (health-protective) screen, tive EPA would investigate whether the model is Furthermore, case, in this significantly overstating the residue lev- provided specific, explanation detailed els that actually occur. drinking water models do yield reliable developed data. The Final Order has models for esti- states: mating exposure in both surface water ground Lack comprehensive drinking wa- water. EPA uses two- (DW) monitoring

ter data. NRDC con- tiered approach to modeling pesticide that, tends because used a model exposure in water. In the initial surface for calculating drinking exposure, tier, EPA FQPA uses the Index Reser- have, EPA does not aas definitional (FIRST) voir Screening Tool model. matter, choosing “reliable data” for replaces FIRST the Generic Estimated factor different than the 10X default val- (GEN- Environmental Concentrations ue. Similar comments were made dur- EEC) model that was used as the first ing development of EPA’s Children’s tier screen EPA from 1995-1999. If Safety Policy. This issue was addressed suggests first tier model pesti- length response to the imida- cide levels water unacceptably be objections. eloprid That in- response is high, a more refined model is used as a corporated herein and is summarized be- second tier assessment. The second tier low. actually model is a combination of the Although availability drinking models, Pesticide Root Zone Model monitoring water data has increased (PRZM) Exposure Analysis and the dramatically in years, the last several (EXAMS). System Model For estimat- EPA still it necessary rely finds for ing pesticide in groundwater, residues most pesticides upon exposure various EPA uses the Screening Concentration models to exposure estimate levels in (SCI-GROW) In Ground Water model. drinking water. These models are Currently, EPA has no second tier based on generic data regarding fate groundwater model. transport and in pesticides the envi- Whether EPA pesticide expo- assesses ronment, they operate by combining drinking sure in water through moni- generic data with pesticide-specific toring modeling, data or EPA uses the data on chemical properties to estimate higher the two values exposure. primarily EPA has used its from surface ground drinking assessing water in overall water models to “screen” those exposure pose unacceptable pesticide. to the In most exposures cases, risks due to in drinking water pesticide residues in wa- surface Panel, Assessment, (Dec. Advisory 1997), See FIFRA etary Scientific A Set Risk 4-7 avail- Being of Scientific Issues Considered http://epa.gov/scipoly/sap/meetings/ able at Agency Estimating in Connection with Drink- 1997/december/finaldec.pdf. ing Exposure Component Water as a of Di- II. higher in significantly are than ter ground water. challenge to the to NRDC’s addition data, exposure Petitioners completeness of Order, EPA ana- Imidacloprid

In the wheth- arguments regarding also raise two models extensive- lyzed each its water toxicity supported data a reduc- er reliable design Based on the results of char- ly. tion in factor. models, peer outside acteristics of the models, validation of A. models, comparison between the challenges first the EPA’s NRDC water predictions models’ and extensive failure for the of certain to wait results data, monitoring EPA concluded that developmental neurotoxicity studies on reliable data the models are based studies”) (“DNT establishing before toler that are un- produce and will estimates py acetamiprid, mepiquat, ances for likely exposure pes- underestimate specifically required metrozine. The EPA drinking water. ticides pesticides to con registrants of these Accordingly, EPA reaffirms its earlier However, EPA did duct DNT studies. drinking water mod- conclusion that its studies not wait for the results the DNT finding provide reliable basis els making its determina before tolerance exposure residues tions. are not underestimated. contend that DNT studies Petitioners (internal 46,726-29 Final Order citations ef- assessing pesticide are essential for added). omitted) Thus, (emphasis explain fects. that DNT stud- Petitioners length reliability EPA addressed at ies are best available studies for ex- explained models. that the amining neurological in children effects yield data because the models conservative sensitive because DNT studies more incorporate higher of the two models studies, that are than other address areas ground from water in values surface studies, not covered other and some- assessing risk of overall developmental times harm that reveal pesticides. Petitioners have not estab- require significant enough to more strin- explanation regard- lished EPA’s gent argue that regulation. Petitioners reliability faulty models is ing the of its arbitrarily capriciously the EPA acted suspect. by requesting failing DNT studies but wait the results of those studies before Accordingly, we conclude the com- removing the child modeling by the EPA puter used to calcu- safety however, drinking late the water was nei- position, Petitioners’ is incon contrary arbitrary Supreme ther to law nor and decision sistent with the Court’s capricious.6 aspect We affirm this of the in Ass’n Home Builders v. National — U.S. -, Wildlife, Final Order. Defenders of course, conclusion, 6. Our is not ensure that the first result was accurate. The meant however, test, suggest that are the EPA’s models flawless. EPA run a if does not second Petitioners have raised some concerns about expo- acceptable the initial test results reveal may exposure the EPA's models that Although required we de- sure are levels. give scrutiny. further exam- warrant To one expertise fer to in this the EPA’s technical ple, performs the EPA first model-based area, EPA, see, Ctr., e.g., Envtl Inc. v. Def. exposure using water ria; test conservative crite- 832, (9th Cir.2003), F.3d we that the note unacceptable the reveal if test results methodology. EPA to revisit wish levels, exposure EPA runs a second test to the (2007). B. 168 L.Ed.2d 467 S.Ct. Builders, rejected a final our court

Home also that disputes NRDC the “arbitrary capri- and order EPA as EPA’s decision deviate from the lOx EPA relied on the had “le- cious” because imposed presumptive factor contradictory before gally positions” arriv- Congress supported by reliable was data decision. ing at its final administrative establishing, certainty, with reasonable Court, Supreme 2529. The 127 S.Ct. at margin that the for aceta lower however, opinion. our The Court reversed miprid, pymetrozine and would mepiquat, mere fact that held that the not harm infants and children. “internally adopted inconsis- might have (C). 346a(b)(2)(A), though Even throughout the decision- positions tent” EPA necessarily required was not to wait making process did not render EPA’s studies, the results the DNT arbitrary capricious. final decision and decision to reduce the lOx child explained 127 S.Ct. at 2530. The Court for these three must still agencies proper as follow the long as Moreover, be based on data. reliable they administrative procedures, have the public why EPA must its data tell are authority change their minds before is- reliable. suing a order. Id. final in problem this case is that Thus, appears Home Builders fore- vague, making EPA’s Final is it Order though Even argument. close Petitioners’ impossible for us to determine whether the initially thought DNT EPA’s deviations from the lOx child needed, EPA later deter- studies were acetamiprid, factor for mepiquat, py- mined that the DNT studies were not in supported metrozine were fact reli- Builders, needed. Under Home data, able of the DNT even absence not EPA’s later determination does auto- example, respect py- studies. with For Final matically the EPA’s Order render metrozine, the Final Order states: Notably, arbitrary capricious. Peti- challenged a NRDC December any procedures not identified tioners have establishing pyme- action tolerances for when the decided were violated seed, gin by- trozine on cotton cotton requested DNT not to wait for studies products, fruiting vegetables crop Moreover, setting before the tolerances. group, vegetables crop the cucurbit provided the Final a reasoned ex- Order leafy crop group, vegetables group planation of DNT studies Brassiea), (except head stem actually every case where required Brassiea, Brassiea, leafy turnip greens, there of the lOx child reduction (66 pecans. FR hops, dried 46,711-12, 46,723- factor. Final Order at 2001). pymetro- Given December pattern toxicological zine’s characteristics, Therefore, EPA determined that reject argu- Petitioner’s acute, potentially presented arbitrarily pymetrozine ment acted short-term, chronic, and cancer risks capriciously by establishing these mepiquat, quantitatively and EPA assessed acetamiprid, tolerances for and making safety determination. pymetrozine receiving before the DNT risks (66 66791-66792). of FR at All these Petitioners’ concerns about studies. While Agen- to the of heart risks were found be below apparent change the EPA’s are (Id.). understandable, Although a cy’s level of concern. questions on such scien- EPA de- study outstanding, DNT was judgment, required tific we are to defer to 10X chil- termined that the additional the EPA. 1052 sum, pro- does not be the Final Order safety generally factor could

dren’s toxicological the information to demonstrate enough to 3X because vide reduced sen- greater no evidence of data showed connection the factors rational between was no the and there sitivity young to EPA examined and the conclu- that in the devel- of abnormalities Farm, evidence 463 it reached. See State sions system. the fetal nervous opment 43, are there- at 103 S.Ct. 2856. We U.S. 29, (64 52438, 52444, September FR there unable to determine whether fore 1999). data the EPA’s supporting was reliable 46,711. at Final Order safety lOx child reductions may be able Although remand that the “toxico- Final Order recites The appro- of in- lx is showed no evidence 3x or logical explain data” to sensitivity developing fetuses appropriate creased cannot conclude it is priate, we However, Final Order “young.” or the To record. do so would abdicate on this explain the connection between does not compliance to ensure responsibility our margin of toxicological data and the 3x to have Congress’s expressed desire with safety place the EPA in selected lOx factor. Ac- presumptive critique applies This lOx standard. same regula- we cordingly, remand the tolerance reasoning with re- the Final Order’s to acetamiprid, pyme- and mepiquat, tions for As a acetamiprid mepiquat. spect to See, Farm, e.g., to the EPA. State trozine result, entirely why the EPA it is unclear 57, (remanding 463 U.S. S.Ct. pymetrozine factors of 3x for chose agency supply req- failed to where as acetamiprid, mepiquat, and lx for Food analysis”); “reasoned & uisite Safe As far as to 4x or 5x 8x or 9x. opposed EPA, v. F.3d Fertilizer record, appears can tell from the it we (D.C.Cir.2003) (remanding for the to these levels EPA chose lower an explanation” exemp- “further provide arbitrarily acknowledge certain con- —to in zinc fertil- tion level for chromium found with no pesticide, about each but cerns izers).7 these lower specific evidence argue EPA and the dissent actually the risks factors would account for argument did not raise this infants and children. Petitioners engage scope merely perfunctory. aware of our We are to 7. While are that the narrow, a’searching inquiry, key careful' under the APA we are also review is [agency] of which is to ensure that the stone the re- that we should not endorse mindful decisionmaking." engaged in reasoned statutorily presumed child moval of the points an This formulation acknowl- we cannot that the EPA's factor if determine edged disparity depth between the of our rationally supported. As conclusions are scope and the of that re- review ultimate Skelly Wright Circuit Judge J. of the D.C. Although scope may view: the ultimate be elucidated, important difference there narrow, depth must be sufficient us depth agen- review of an between our agency’s comprehend to be able han- scope cy’s and the action of that review: dling upon. of the evidence cited or relied requires review of action Judicial purpose in-depth review is reviewing whether consider court properly can educate ourselves that we so [agency’s] based on a con decision was perform reviewing our function: determin- sideration of relevant factors ing agency's whether the conclusions are For, there has a clear error of whether been although rationally supported. data judgment. noted in Inter As this court analysis interpretation functions v. national Ladies Garment Workers Union agency's that often lie within an realm of Donovan, (D.C.Cir.1983), expertise, duty it is our those denied, agency's cert. 469 U.S. 105 S.Ct. functions ascertain whether reasoned, (1984), complete, actions were ade- L.Ed.2d 39 such review "is and it below therefore cannot assert now. Each side to bear its own ap- costs on disagree. objec- We In its administrative peal.

tions, challenged the NRDC basis for the PART, GRANTED IN DENIED IN safety reduced inadequate. levels as PART, AND REMANDED. particular, that, argued NRDC absent the studies, DNT the EPA had no reliable IKUTA, Circuit Judge, concurring safety data to reduce the lOx child factor. part dissenting in part: objection NRDC’s challenge was direct I agree to the lower child with by majority’s levels set conclusion justify EPA. To that “the computer its decision to set a 3x or modeling used lx EPA to factor without calculate the safety DNT of drinking studies, the EPA needed to was neither explain why contrary to law nor arbitrary possession the data and capricious.” Maj. was reliable and Op. at 1050; supported such see also reductions. As noted The Lands Council v. McNair, above, (9th the EPA explained why the 537 F.3d absence 989-94 Cir. 2008) (en banc). of the DNT necessarily studies does not I agree also with the majority’s mean that holding there is a lack of reliable data National Ass’n of justifying a Home downward deviation from Builders v. Wildlife, Defenders of — U.S. -, However, lOx child 2529-30, 127 S.Ct. (2007), explain why, EPA failed to L.Ed.2d 467 in the petitioners’ case of forecloses issue, argument the three “that the EPA arbitrarily available acted justified capriciously information specific safety establishing pesti re- cide ductions. Because the tolerances for challenged acetamiprid, mepiquat, NRDC pymetrozine provide that explanation, receiving before ... DNT Maj. EPA’s failure to do so studies.” properly Op. is before at 1051. us for review. however, I disagree, majority’s with the conclusion that regulations the tolerance

CONCLUSION acetamiprid, for mepiquat, and pymetro- Because the EPA failed to adequately zine must be remanded to be- explain the basis for its deviations from cause “it entirely is unclear the lOx child factor for acetamiprid, chose factors of 3x pymetrozine for mepiquat, pymetrozine, grant we and acetamiprid, lx mepiquat,” petitions for part review in and remand to and therefore “the Final Order [did] the EPA for further proceedings provide enough consis- information to demon- tent opinion. issues, with this allOn other strate a rational connection between the deny we petitions for review. factors the EPA examined and the eonclu- quately explained. (internal The mere fact that an quotation citations and marks omit operating expertise is in a field ted). of its Riverkeeper, See also Inc. v. United customary does not excuse us from our Agency, States Environmental Protection And, responsibilities. where the (2d Cir.2007) ("In F.3d 103-04 a technical agency's reasoning, although complex, is sort, area judges of this it is difficult for rational, clear, complete, we must af- parties propriety interested to determine the Contrarily, agency's firm. where the rea- Agency's justification action without a irrational, unclear, soning sup- or not supported by clearly for the action identified ported by purports interpret, the data it import substantial evidence whose is ex disapprove agency’s must action. plained.”). Peck, Safety Center Auto v. (D.C.Cir. 1985) J., (Wright, dissenting) 1054 A The Maj. at 1052. Op. it reached.”

sions argument raise this failed to petitioners tolerance majority, the noted the As to the prior objections to their have substan- appeal in this rules at issue De Order Final Rule. See issuance the devel- While tially similar histories.2 Toler Objections to Issuances nying inwas policy factor of its opment 2005) 46,706 (Aug. ances, Fed.Reg. 70 pesticide peti- the EPA received process, Rule], therefore the [hereinafter, Final of tolerances issuance requesting tions the explain no given opportunity was EPA lx pesticides 3x and three adopting the for the rationale for under the FFDCA its may FFDCA, safety factors. Because the required in the As issue. EPA, the raised before objections only the and other the data EPA reviewed scientific 346a(g)-(h); § 40 C.F.R. 21 U.S.C. see information, completeness their evaluated 180.30(b), 178.25(a)(2), is not the issue §§ “the rela- reliability, considered petition if the Even before us. properly of such studies tionship of the results objection, there is no raised this ers had risk.” human infor provided the EPA could have doubt 346a(b)(2)(D)(iii); see, e.g., Pymetrozine; a rational sufficient mation demonstrate 66,786, Tolerance, Fed.Reg. Pesticide safety factors. selecting these for basis 2001). (Dec. 27, EPA 66,787 deter- Indeed, the record establishes safety factor was not that the child mined expertise to select upon drew 3x, for necessary, or be reduced could principles on uncertainty factors based analysis based on its pesticides each of the communi in the scientific established well given exposure levels for potential majority’s no basis for ty. There is extent to which pesticide, the not path is agency’s conclusion effects, and the development, affected fetal See, case. reasonably discernable in multi- infants as observed any, if Ass’n v. State Motor Vehicle e.g., Mfrs. mammals. In each generational studies Co., 29, 43, Ins. 463 U.S. Mut. Auto. Farm case, expla- a reasoned provided (1983). 77 L.Ed.2d 103 S.Ct. eliminating reducing or nation for IIB the I from Part Accordingly, dissent factor.3 majority opinion.1 Tolerance, 27, 2002); Mepiquat; Pesticide majori- 6 of the 1. I also dissent from footnote 3,113 (Jan. 2002); Pymetrozine; Fed.Reg. majority’s suggestion that the ty opinion. 66,786 (Dec. Tolerance, Fed.Reg. Pesticide methodology” revisit[its] wish to "EPA 27, 2001). exposure regarding its "model-based teaching directly with the conflicts testfs]” establishing tolerance for 3.In its order Lands panel in Council. en banc Lands our explained: acetamiprid, proper not a emphasized it "is Council appellate to "act as a a federal court” for role to 3x could be reduced [T]he [agency] of scientists that instructs panel toxicology acetamiprid da- because hypotheses,” "chooses to validate its how quantitative complete; there is no tabase is studies,” requires among scientific suscep- of increased qualitative evidence every possible "explain scientific following rat tibility in útero majority uncertainty.” 537 at 988. The F.3d fetuses; (food dietary rabbit by doing that which Lands in this case errs water) exposure assess- and residential expressly forbids. Council potential ments will underestimate infants, children, exposures and/or childbearing age; and the re- are acetami- at issue women of 2. The three neurotoxicity developmental pymetrozine. quirement For the of a mepiquat, and prid, reflecting Acetamiprid; study on criteria rulings, is not based see tolerance individual 14,649 Tolerance, (Mar. developing fetuses Fed.Reg. special concern for Pesticide objections study pesticide, In NRDC filed to for each and the because the for each EPA drinking exposure the final rules relied on water to reduce than based on the EPA’s decision the models rather on pesticide-specific or lx. safety drinking monitoring lOx to 3x The data.4 Notably, on the the objections object NRDC’s were based the- NRDC did not the ground on that ory justify that did not have “reliable the EPA failed to its choice of a safety data” which allow it to reduce the factor of 3x or lx would relative to some factor, other safety child factor because the had such or 5x as “4x or 8x a or yet Maj. Op. received the results of DNT 9x.” at 1052.5 The EPA generally establishing young pymetro- or which are used for re- for tolerance zine, quiring study a DNT and a Tolerance, the EPA stated: Acetamiprid; complete Pesticide 67 Fed. toxicity There is a database for 14,655. Reg. at pymetrozine exposure and data are com- The EPA removed additional safe- plete or are estimated on data based lx) (reducing ty mepiquat it to factor for reasonably potential expo- accounts for already itself of because it had satisfied sures. EPA determined that 10X mepiquat "salt” that it of another of protect factor to infants and children could was identical in its envi- determined terms of FQPA be reduced to 3. The factor is re- toxicology. Mepi- ronmental distribution assessing potential duced after for addi- Tolerance, quat; Fed.Reg. Pesticide sensitivity of tional infants and children to 3,115-16. explained: It pymetrozine following residues of published Agency The a assessment risk for developmental toxicity studies: studies in 12, 2000, mepiquat January chloride on two-generation repro- rabbit and rat and use on as well as all discusses cotton study duction in the rat. There was no registered mepiquat other uses of chloride. susceptibility evidence of increased in these analysis exposure In that risk for estimates FQPA safety studies. The factor was not Agen- mepiquat chloride were below reduced due to one to the need for devel- cy’s Agency level concern. The has re- of opmental neurotoxicity study. mepiquat study viewed a dissociation for Tolerance, Pymetrozine; Pesticide 66 Fed. pentaborate mepi- that demonstrates that 66,791. Reg. at quat pentaborate in an dissociates identical Council, Objec- 4. Natural Resources Defense physical mepiquat manner to chloride in Therefore, tions to the Establishment performed Tolerances analysis water. of for Pesticide Chemical Residues: mepiquat for salt,” chloride the "chloride or Halosulfuron- (filed methyl Pymetrozine mepiquat "penta- pertains also to this Tolerances rate, 25, 2002); borate use use Feb. salt” because the maxi- Natural Resources Defense Council, pertinent Objections mum seasonal use rate and other to the Establishment mepiquat use remain the for factors same Tolerances Pesticide Chemical Residues: for chloride the "chloride ... or salt.” Imidacloprid, Mepiquat, Bifenazate, Zetacy- Margins incorporated into permetkrin, and Tolerances Diflubenzuron directly EPA risk assessments either (filed 19, 2002); May Natural De- Resources (MOE) through margin aof use Council, Objections fense Establish- (safe- analysis through using uncertainty ment Tolerances Pesticide Chemical for ty) calculating a dose level that factors in Isoxadifen-ethyl, Acetamiprid, Residues: Pro- poses appreciable no risk to humans. Furilazole, Fenhexamid, piconazole, Agency has determined that (filed 2002). May Tolerances Fluazinam FQPA safety mepiquat factor for is IX. See objected 5. The three NRDC also to the toler- Agency’s mepiquat risk assessment for grounds, ances a number of other includ- on January chloride dated for details. ing the ad- failure consider farm children pentaborate mepiquat are that facts equately, worker risk failure consider mepiquat mepi- another “salt” adequately, inadequate aggregate an risk as- quat pentaborate mepiquat disassociates to sessment, improper on a toxicology reliance "lowest the basic data base therefore (LOAEL) mepiquat pertains mepiquat observed adverse effect level” rath- chloride pentaborate. er than a "no observed adverse effect level” (NOAEL), assumption "percentile” Id. a residue agency to an give notice quirement is to Final objections NRDC’s rejected issues, can so the errors or specific explanation Rule, detailed provided experience bear Agency’s “bring had reli- that it concluding of its basis EPA, 859 Nader v. question,” a contested absence of notwithstanding the data able (9th Cir.1988), either and F.2d expo- the use of and DNT studies certain a reasoned provide the error or correct reliable data models, why such and sure explanation. it that conclusion the EPA’s supported child the safely reduce could EPA’s deci- objected to the The NRDC 46,706. The Fed.Reg. generally re- See was not factor that a lOx sion safety fac- child 3x upheld Rule the Final toler- three each of the quired for and pymetrozine, and acetamiprid objection for on the ances, tor based mepiquat. for factor data the lx child lacked reliable the EPA ground that 46,736. 46,711-13, a DNT at to obtain Id. it had failed because drinking water study and had relied

B rests its majority models. The required remand is concluding that for acetami- that the tolerances decision why it explain the EPA failed must be pymetrozine mepiquat, because prid, acetami- uncertainty factor for a different EPA on chose a 3x to the remanded uncertainty failed to and a lx namely that pymetrozine ground, prid uncertainty “4x or fac- or 5x it chose a 3x mepiquat instead explain factor for and a 1052, majority pymetrozine 9x,” acetamiprid the tor Maj. Op. at or 8x instead uncertainty mepiquat lx factor properly before argument an reaching 9x,” Maj. Op. at 1052. or “4x or 5x or 8x to the applicable regulations The us. object to the did not Because the NRDC of administra- exhaustion require FFDCA that the ground the 180.30(b) EPA’s decision on § 40 C.F.R. tive remedies. See explanation a clear give EPA had failed to is not avail- “judicial that (providing “4x in lieu of or lx of the 3x adversely party affected an unless able id., 9x,” identified the issue or 5x or 8x or objection procedures, the[] exhausts exhausted, and we majority by the is not there- procedures preliminary any petition reviewing it. from precluded to”). objection the In order to exhaust FFDCA, peti- by the provided procedure justify ad- majority attempts The to the objections must raise their tioners did not the NRDC dressing issues that See specificity. with adminis- by turning raise to the objections to an (requiring 346a(g)(2)(A) § on its head. requirement trative exhaustion establishing a tolerance to above, regulations re- EPA order As noted under provisions reme- particularity “with administrative specify quiring exhaustion of objectiona- dies, proposed regula- order deemed regulation objecting or parties 178.25(a)(2) (“To burden, be and must ble”); 40 C.F.R. tions bear the initial Administrator, objec- objections agen- to the by specific considered raise their are waived and particularity cy; any issues not raised Specify tion must: with order, agency or considered regulation, need not be provision(s) of appeal. 40 C.F.R. to, objee- by the court on objected basis for denial Here, 178.25(a)(2), majori- pur- §§ The 180.30. tion(s), sought.”). the relief EPA to identi- to the ty re- shifts burden exhaustion of this administrative pose to us on these issues did not raise inadequately protected infants and chil- NRDC Rule, 4; supra Final dren, appeal. See Note improper assessment of chron- and an 46,717-36. 46,713-14, Fed.Reg. at among dietary exposure, others. ic raise, fy, and resolve additional ly” issues to the agency.). Nor would it be rea- all, petitioners failed raise at let alone sonable or consistent with the purpose of requisite “particularity.” frame with the the administrative requirement exhaustion 178.25(a)(2). §Id. Here, to do so. responded to the NRDC’s specific objection by carefully ex- majority first specific objection recharacterizes short, as we have no authority to re- being general a “challenge more to the verse an agency decision based on an ob- lower child levels jection set the EPA.” that was not raised. Supreme Maj. Op. at generaliza 1053.6 From this Court long has warned us not to invent tion, majority leaps to the conclusion new issues part as of our objec- review of specific that the objection places NRDC’s tions raised to an administrative agency. a burden on the EPA to “explain why the “A reviewing usurps court agency’s possession” data in its supported its deci function when it sets aside the administra- sion to set a 3x or lx factor for tive determination upon ground a not opposed issue “as 4x to or presented theretofore and deprives the 5x Maj. 1052, or 8x or 9x.” Op. at agency] [administrative an opportunity of There is no basis in our case law for taking matter, to consider the make ruling, such a leap placing or such a burden on and state the reasons for its action.” Un- the EPA. We have never required an employment Comp. Com’n v. Aragon, 329 interpret a specific objection 143, 155, 245, as U.S. 67 S.Ct. 91 L.Ed. 136 being, at generality, (1946); some level of a differ see also Ngo, v. 548 U.S. Woodford ent, objection, broader 81, 89, and responding 2378, 126 S.Ct. 165 L.Ed.2d 368 See, (2006). objection. different e.g., High addressed, The EPA with ade- Country FERC, Resources v. 255 F.3d quate thoroughness detail, see, e.g., (9th Cir.2001) 745-46 (stating that peti Co., a Balt. Gas and Elec. 462 U.S. at 105- tioner “implicitly” cannot raise an issue to objections S.Ct. each of the a regulatory agency in order to exhaust filed objections the NRDC. These cre- remedies, administrative provide but must compass ate the of our limited review. 21 “much specificity more § the statement 346a(h)(l), (5); of U.S.C. 40 C.F.R. objection” present 178.25(a)(2).7 “square- § the issue however, supra, 6. As noted objection a tantly, valid majority’s mistake in this case general; cannot be this "[sjpecify it must with when it warns that “we implacably must be order, particularity provision(s) reg- skeptical of the recognition of appel- belated at the to, objected ulation stage or denial the basis for the late analysis that elements of scientific objection(s), sought.” unchallenged and the during relief proceeding contested C.F.R. incomprehensible 178.25. expla- without further nation. post-appeal pleas To credit such inadequate information is to threaten the in- majority 7. The part upon also relies in tegrity rulemaking beyond all in fields our Judge Wright dissent of in Center Auto own limited scientific ken.” Id. at 1361. The Peck, Safety (D.C.Cir. v. entirety passage of the repeating: bears 1985) J., (Wright, dissenting). Maj. Op. at However, (now Justice) Judge 1052-53 n. 7. response equally The dissent’s is unsatisfac- opinion majority Scalia’s for the tory, in that consisting essentially case of the assertion error, and, captures the dissent’s presentation concomi- that late of the issue must be an raised, provides the record an issue

C for us to conclude basis adequate us to address proper were it Even uncertainty lx a 3x EPA’s selection of EPA’s consideration of the adequacy IIHS, Though Farm and State be used?” validity [data] disregarded because others, responded with substantial agency’s among validity of the curves essential case, suggestion how- not even simply not the there was isIt comments conclusion. ever, represented a postulates for curves the essential effectiveness that all of methodology. fundamentally must be contained agency rule invalid any why peti- conse- Every judgment of of reasons provides record. list dissent upon an infinitude objec- quence is constructed these might not have raised tioners greater lesser certi- judgments, rulemaking proceeding. other during the tions tude, logical dependen- progression of in a down all of which boil at 1387-88 Dissent equiv- principle the cy in a first terminating upset probable petitioners' reluctance They possibly all cannot 1 + alent of 1=2. Even if determinations. favorable in the statement basis be included consideration, would be of only it were the rulemaking. have We do not purpose for a tacti- reward their questionable wisdom to under- authority require that all elements agency in the to leave the cal decision “un- forth in a fashion lying a rule set be agen- encourage benighted thus dark —and layman,” Vermont a see to derstandable event, any in future. But cy in the action Corp. v. Natural Re- Power Yankee Nuclear explain speculations do not the dissent's Inc., 519, Council, 435 U.S. sources Defense mph standard proponents of the 2.5 1218, 1197, 557, 460 55 L.Ed.2d S.Ct. 98 fact, arguments. these did not raise (1978). well be disclosed And it as June 1979 Final on the commenters some judges we fields ourselves that in scientific (which mph supported the 5.0 Assessment equipped to laymen, ill determine are standard) object "effectiveness did that the expla- requisite where the falls between line systems to mph are overstated for 5.0 ratios analysis problematic useless nation of extent,” objection that but considerable cognoscenti replication of what for correspondence upon the lack was of based physics. basic textbook on amounts to a repair data than theoretical with cost rather Thus, adequacy of an for the our "feel” curves. inadequacy of the effectiveness necessarily gov- analysis in cases is such were consid- conclude curves We from that it elicits erned the reaction ground expert of part of the common ered knowledgeable commenters. From explana- required analysis no further that ground point at which it reaches common objections on this score Petitioners' tion. reasonably further assume that no we can Supreme Court’s put us in mind required. We will hear on explication is adminis- injunction in Vermont Yankee appeal that needful assertions elaborations game proceedings not be a should trative requested provided; but fairly were engage unjustified obstruc- in a forum skeptical rec- implacably of belated must be making cryptic ref- and obscure tionism appellate stage that ele- ognition at the "ought to be” consid- to matters erence analysis unchallenged ments scientific then, failing to do more after ered proceeding are incom- during a contested attention, agency’s the matter to the bring explanation. further prehensible without agency seeking have that determination post-appeal pleas inade- To credit such ground on the vacated integri- quate is to threaten the information "forcefully pre- to consider matters failed beyond rulemaking our ty in fields of all sented.” present ken. The limited scientific own 553-54, S.Ct. at 1217. U.S. at pres- challenge curves to the effectiveness appropri- more instructions even These flagrant particularly in a ents the threat reference to cryptic and ate when obscure peti- specifically asked the NHTSA form. fact, even less inadequacy theoretical (and rulemaking participants) tioners other —or that, explana- lack of than mere assertion of existing analyses repre- in 1979: "Do adequacy first made theoretical ap- tion of appropriate sent the most methods —is on appeal. study bumper at proaching a standards original, (emphases internal at Id. 1361-62 impact speeds and levels of dam- different omitted). not, the record age If what should citations resistance? method arbitrary, majority factor was not as the ance documents selecting FQPA erroneously Maj. Op. concludes. at factor different from the default lOx FQPA See, safety factor. e.g., Fenari- The record in this case contains the mol; Tolerance, Pesticide 32,- 71 Fed.Reg. published guidance EPA’s for determina- 841, 32,844-45 (June 2006) (noting that an appropriate FQPA safety tion of the EPA originally FQPA reduced the EPA, Programs, See Office of Pesticide Safety factor from lOx to subject 3x Appropriate FQPA Determination receiving data, additional subsequent- Factor(s) Safety in Tolerance Assessment ly recommended reducing the FQPA 3x (Feb. 2002). explained in As Safety factor lx when it obtained ade- guidance document, “when evaluating data); quate Thiacloprid; Pesticide Toler- whether reliable data are available to sup- ances, Fed.Reg. 55,503, 55,509 (Sept. port FQPA safety factor different from 2003) (“Although the lack morpho- FQPA default lOx factor and metric assessments in the DNT raised FQPA what level of ‘different’ safety fac- some uncertainty, EPA determined that children,” tor would be safe for infants and there were sufficient reliable data id. at to select inmay, the exercise of additional factor of 3X instead of judgment, scientific consider the 3x un- 10X.”). certainty relied on this traditionally long-es- factor that has been approach tablished in setting evaluating posed by used pesti- risks factors for the three at cide in a related issue context. See id. here, 26; supra see at nn. & 3. see also id. at (noting 9-10 that in evaluating the posed by pesticide risks ex- Because petitioners did not chal- posure, a “often value of 3x is used to lenge the absence an explanation of as to ”). address database deficiencies.... why picked a 3x or lx factor However, in “the optimal case” in “which rather than a “4x or factor, 5x or 8x or 9x” high is a there level of that confidence the we are deprived of the EPA’s more de- hazard and assessments are suf- tailed explanation toas the scientific ficiently conservative!,] and there are no community relies uncertainty on the 3x assessment,” residual uncertainties in the factor when database deficiencies are iden- factor, i.e., EPA will remove the Nevertheless, tified. the record indicates apply “safety IX,” it a will factor of deem- that use of uncertainty the 3x factor is existing ing threshold “sufficient to generally accepted by the scientific com- protect infants and children.” at Id. 51- munity. al., See L. Michael Dourson et 52. See also Office of Programs, Pesticide Evolution Uncertainty Science-Based of EPA, Programs’ Policy Pesticide Office Assessment, Factors in Noncancer Risk on The Appropriate Determination 24 Regulatory Toxicology and Pharmacolo- FQPA Factor(s) Safety For Use in The (1996) gy (generally discussing the sci- Tolerance-setting Response Process: to ence underlying the use of lOx and 3x Public Comments OPP Docket OPP-00610 factors). Indeed, default uncertainty (Feb. 2002) (internal citation omit- any lack of controversy regarding ra- ted) (further explaining long-standing tionale underlying the selection aof 3x for “imposing basis uncertainty database safety factor rather “4x than a or 5x or 8x factor of 3X if study one is key missing or 9x” may explain why the NRDC from the database and a factor of 10X if argue objections did not in its to the EPA than more one is missing.”). agency justifica- a special needed EPA tolerance-setting factor,

Other regulations selecting tion for 3x once the approach reflect guid- described determined additional light required. factor was not lOx BARRETT, Henry Jacob long-established aon EPA’s reliance Plaintiff-Appellant, majori protocol, widely accepted these “the EPA chose ty’s statement v. arbitrarily” sup is not safety levels lower record, contrary to and is Nofziger; BELLEQUE; Rebec- ported J. Brian the scientific Taylor; defer to obligation Russell; Prinslow; Gary our J. ca agen made judgments analysis Defendants-Appellees. Parker, Kent special ex its area of within cy operating No. 06-35667. Co., 462 U.S. and Elec. Balt. Gas pertise. 103, 103 2246. S.Ct. Appeals, States Court United Ninth Circuit. recently emphasized, “[w]e we have

As agen- when most are to be deferential March and Submitted Argued within its area cy making predictions, sci- the frontiers of expertise, at special 22, 2008. Sept. Filed Council, at 993 Lands ence.”

(internal marks alterations quotation

omitted). from prohibits us Lands Council panel “act temptation to as

indulging the Nor can we at 988. Id.

of scientists.” always “to demonstrate

require methodology its scientific reliability underlying [agen- hypotheses

or the (internal Id. at 990

cy’s] methodology.” omitted). Lands Council marks

quotation simply role is proper that our

teaches expertise, in its agency,

ensure rendering judgment

made no clear error id. arbitrary capricious. See

its action case The EPA’s choices

at 992-94. arbitrary capricious nor neither

were judgment. See must defer to 706(2)(A); Elec. Balt. Gas and 103-05,

Co., 103 S.Ct. 462 U.S. deny petition

Accordingly, I would entirety. in its notes “the plaining why the absence of DNT studies that, argued NRDC absent the DNT stud did not require the retention of the lOx ies, had no reliable data to reduce safety factor. The EPA satisfied its bur- the lOx Maj. Op. factor.” den, and had no obligation additional specific objection 1053. This is indeed the make a further response. raised majority NRDC. The then

Case Details

Case Name: Northwest Coalition for Alternatives to Pesticides v. United States Environmental Protection Agency

Court Name: Court of Appeals for the Ninth Circuit

Date Published: Sep 19, 2008

Citation: 544 F.3d 1043

Docket Number: 05-75255, 05-76807

Court Abbreviation: 9th Cir.