United States Court of Appeals
                         FOR THE EIGHTH CIRCUIT
                                 ___________

                              Nos. 04-3629/3632
                                ___________

Friends of the Boundary Waters           *
Wilderness; Sierra Club; Superior        *
Wilderness Action Network; American      *
Lands Alliance, Minnesota Canoe          *
Association, American Canoe              *
Association, Minnesotans for             *
Responsible Recreation,                  *
                                         *
              Plaintiffs/Appellees,      * Appeal from the United States
                                         * District Court for the
       v.                                * District of Minnesota.
                                         *
Dale N. Bosworth, Chief of the United *
States Forest Service; Mike Johanns,     *
                          1
Secretary of Agriculture,                *
                                         *
              Defendants/Appellants,     *
________________________________ *
                                         *
Conservationists with Common Sense; *
Ely Outfitters Association; Gunflint     *
Trail Outfitters Association; Seagull-   *
Saganaga Homeowners Association,         *
                                         *
              Intervenor Defendants/     *
              Appellants.                *
                                    ___________


      1
       Pursuant to Federal Rule of Appellate Procedure 43(c)(2), Mike Johanns is
automatically substituted for his predecessor, Ann Veneman, as Secretary of the
Department of Agriculture and appellant in this case.
                              Submitted: September 16, 2005
                                 Filed: February 15, 2006
                                  ___________

Before BYE, HEANEY, and COLLOTON, Circuit Judges.
                           ___________

BYE, Circuit Judge.

       The defendants and intervenor defendants, collectively referred to as the United
States Forest Service (USFS), appeal the district court’s grant of summary judgment
to the plaintiffs, collectively referred to as the Friends of the Boundary Waters
(Friends). Specifically, the USFS appeals the district court’s ruling the USFS did not
have authority to recalculate the average actual annual motorboat use during 1976-78
within certain lake chains of the Boundary Waters Canoe Area Wilderness (BWCAW)
to include uses improperly excluded from its initial base period calculation. The
USFS also appeals the district court’s alternate finding the recalculation was arbitrary
and capricious. We affirm in part and reverse in part.


                                           I


                                           A


       The BWCAW was one of the first wilderness areas recognized under the
Wilderness Act of 1964. 16 U.S.C. §§ 1131-36 & note. As the largest wilderness area
east of the Rocky Mountains and north of Everglades National Park, it comprises
approximately 1,080,300 acres of forest land encompassing over 1,175 lakes
connected by several hundred miles of streams and rivers. See Minnesota v. Block,

, 1247 (8th Cir. 1981). It provides habitat for hundreds of species,
including the gray wolf, pine marten, bald eagle, black bear, moose, and lynx.


                                          -2-
       The Wilderness Act generally prohibits all motorboat use within wilderness
areas protected by the Act. However, the BWCAW was excepted from this general
prohibition insofar as established motorboat use within the BWCAW and other
motorboat use not undermining the ability to maintain the “primitive character of the
area” were permitted. 16 U.S.C. § 1133(d)(5) (1976). In 1978, however, Congress
reconsidered the BWCAW exception to the Wilderness Act. “[I]n reaction to
threatened deterioration of the wilderness from excessive use,” Congress enacted the
Boundary Waters Canoe Area Wilderness Act, Pub. L. No. 95-495, 92 Stat. 1649
(BWCAW Act). See also 

.


       The BWCAW Act prohibited all motorboat use within the BWCAW except on
specifically enumerated lakes comprising approximately one-quarter of its waters.2
The motorboat use allowed by the BWCAW Act is circumscribed: the Secretary of
Agriculture is directed to establish motorboat quotas restricting use to less than or
equal to the “average actual annual motorboat use of the calendar years 1976, 1977,
and 1978.” BWCAW Act § 4(f).3 The quota levels are “based on such criteria as the
size and configuration of each lake, and the amount of use on that lake.” 

the “average actual annual motorboat use,” motorboat use by “lake
homeowners and their guests and resort owners and their guests on that particular
lake” is not counted. 

lakes included Fall Lake, Newton Lake, Moose Lake, Newfound Lake,
Sucker Lake, Snowbank Lake, East Bearskin Lake, South Farm Lake, Trout Lake,
Saganaga Lake, and Basswood Lake. See 

n.9 (citing
BWCAW Act §4(c)). Motorboat use was phased out on other lakes by 1999. See 

“average actual annual motorboat use of the calendar years 1976, 1977,
and 1978,” or base period calculation, establishes the maximum number of permits
USFS may issue. The USFS may establish permit quotas lower than the base period
calculation but may not establish quotas allowing more use than during the base
period.

                                         -3-
      The BWCAW Act was passed to ensure the BWCAW’s wilderness character
would be preserved. See 

(“Congress passed the BWCAW
Act with the clear intent of insuring that the area would remain as a wilderness and
could be enjoyed as such.”). Limiting motorboat use is integral to preserving the
wilderness values and primitive character of the area. See United States v. Gotchnik,

, 804 (D. Minn. 1999), aff’d, 

(8th Cir. 2000).


                                          B


       In 1981, the USFS initially calculated the “average actual annual motorboat
use.” In doing so, the USFS considered computer data and analyses, wilderness
permit data, records of motorboat use during 1976-78, public comments, and
interviews. For three chains of lakes–the Moose Lake Chain, the Saganaga Lake
Chain, and the Farm Lake Chain–the USFS deemed use by home and resort owners
(as well as their guests) within the lake chain encompassing their property to be
exempt from the base period use calculation. The USFS determined the Moose Lake
Chain encompassed Moose, Sucker, Newfound, and Birch Lakes; the Saganaga Lake
Chain included Saganaga Lake, Seagull River, and Gull Lake; and the Farm Lake
Chain included White Iron, Farm, Garden, and South Farm Lakes.4 Accordingly, the
USFS concluded motorboat use by homeowners, resort owners, and their guests did
not affect the base period use calculation or quota system and did not require permits
when such use was limited to their lake chain. The USFS calculated the base period




      4
         Only five of these lakes are located at least partially within the BWCAW and
still allow motorboat use: Moose, Newfound, Sucker, Saganaga, and South Farm
Lakes.

                                         -4-
day use – as opposed to overnight use – as 14,925 day trips for the BWCAW. The
base period use calculated for the three lake chains at issue was 3,205 day trips.5


       In 1993, the USFS determined allowing motorboat use to the maximum extent
possible under the statute was “strain[ing] the wilderness environment and [was]
tending to degrade the intended primitive and unconfined recreation experience” of
the BWCAW. It therefore established the Boundary Waters Canoe Area Wilderness
Management Plan and Implementation Schedule of 1993 (1993 BWCA Plan) to set
the motorboat day-use quotas at approximately seventy-five percent of the base period
use calculation. The motorboat day-use permit quota was set at 7,902 permits for the
entire BWACW. The day-use permit quota for the three lake chains was set at 2,376
day-use permits.6 One day-use permit can accommodate a group of up to four boats
or multiple trips in one day. Admin. R. (A.R.) Exh. 14, at 82.


       In 1999, this court concluded the USFS’s interpretation that homeowner, resort,
and guest lake chain use did not require a permit was contrary to the plain language
of the BWCAW Act. Friends of the Boundary Waters Wilderness v. Dombeck, 

, 1124-25 (8th Cir. 1999). The court concluded the term “particular lake”
in the BWCAW Act did not mean a lake chain but referred to each individually-named
lake. 

held use by homeowners, resort owners, and their




      5
       This includes 837 for Moose Lake Chain, 345 for Farm Lake Chain, and 2,023
for Saganaga Lake Chain.
      6
         This figure is lower than seventy-five percent of the base period use
calculation because it also reflects the reduction in permits resulting from phase outs
of motorboat use on various lakes within the lake chains.

                                         -5-
guests is exempted from a permit requirement only when such use is limited to the
lake adjoining the owners’ properties. Use on any other lake requires a permit.7


       Because the USFS did not include homeowner, resort, or guest non-exempt lake
chain use when initially calculating the base period use, the quotas established by the
USFS did not account for non-exempt lake chain use. Given the fixed quotas, the
Dombeck ruling increased demand for permits as homeowners, resorts, and guests
were required to obtain permits for non-exempt lake chain use. In 2002, the USFS
responded to this increased demand by recalculating the base period use and
correlated quotas to include non-exempt lake chain use. The recalculated base period
use figures for the entire BWCAW establish a maximum quota of 15,999 day-use
permits.8 The USFS applied the 1993 BWCA Plan to establish quotas at seventy-five
percent of the recalculated base period use. Accordingly, pursuant to the 1993 BWCA
Plan, the recalculated total quota for the three lake chains at issue is 6,892 day-use
permits.9


       Friends challenged the recalculated base period use figure, arguing the USFS
lacks authority to recalculate the base period use and the recalculation was arbitrary


      7
         For purposes of this opinion, any lake adjoining the property owned by a
homeowner or resort on which use by a homeowner, resort, or guest is considered
exempt under Dombeck shall be referred to as the “particular lake” or “adjoining
lake.” Use on any other lake within a lake chain encompassing such “particular lake”
shall be referred to as “non-exempt lake chain use.”
      8
      This is an increase of 5,830, or 3,045 for the Moose Lake Chain; 1,393 for the
Saganaga Lake Chain; and 1,392 for the Farm Lake Chain.
      9
        This constitutes an increase of 4,516, or a 290% increase over the previously
calculated 1993 quota levels for the three lake chains. The resulting quotas are: 2,895
for the Moose Lake Chain; 2,653 for the Saganaga Lake Chain; and 1,344 for the
Farm Lake Chain.

                                         -6-
and capricious. Friends also challenged USFS’s recalculated quotas as exceeding the
base period use in violation of the BWCAW Act. The district court agreed with
Friends, ruling at summary judgment the USFS was without authority to recalculate
the base period use and the recalculation was arbitrary and capricious. The district
court did not reach Friends’ challenges to the quotas established by the USFS.


                                         II


                                         A


      We review de novo a district court’s grant of summary judgment, applying the
same legal standards used by the district court. Voyageurs Nat’l Park Ass’n v.
Norton, 

, 763 (8th Cir. 2004). Summary judgment is proper when there
is “no genuine issue as to any material fact and the moving party is entitled to
judgment as a matter of law.” Fed. R. Civ. P. 56(c); Employers Mut. Cas. Co. v.
Wendland, 

, 893 (8th Cir. 2003). “We may affirm the district court’s
grant of summary judgment on any ground supported by the record.” Woods v.
DaimlerChrysler Corp., 

, 990 (8th Cir. 2005).


       We view the evidence and all reasonable inferences drawn therefrom in the
light most favorable to the nonmoving party. In re Derailment Cases, 

,
792 (8th Cir. 2005). An issue of fact is genuine when a “reasonable jury could return
a verdict for the nonmoving party.” Anderson v. Liberty Lobby, Inc., 

,
248 (1986). However, “[t]he nonmoving party may not rest on mere allegations or
denials, but must show a genuine issue of material fact (or that the movant is not
entitled to judgment).” Wenzel v. Mo.-Am. Water Co., 

, 1039 (8th Cir.
2005) (internal quotation omitted).



                                         -7-
                                             B


       Judicial review of federal agency administrative decisions is, unless expressly
stated otherwise, governed by the Administrative Procedures Act (APA). 5 U.S.C.
§ 706; In re Sac & Fox Tribe of Miss. in Ia./Meskwaki Casino Litig., 

,
755 (8th Cir. 2003). Under the APA, an agency administrative decision may be set
aside only if it is “arbitrary, capricious, and an abuse of discretion, or otherwise not
in accordance with law,” 5 U.S.C. § 706(2)(A), “in excess of statutory . . . authority,”

or “without observance of procedure required by law.” 

However, we are “not obliged to stand aside and rubberstamp [the
district court’s] affirmance of administrative decisions that [we] deem inconsistent
with a statutory mandate or that frustrate the congressional policy underlying a
statute.” Nat’l Labor Relations Bd. v. Brown, 

, 291 (1965).


      “When reviewing an agency’s construction of a statute, the court first considers
whether the intent of Congress is clear; if so, the court’s inquiry is over, ‘for the court,
as well as the agency, must give effect to the unambiguously expressed intent of
Congress.’” 

(quoting Chevron U.S.A. Inc. v. Natural Res.
Def. Council, 

, 842-43 (1984)); see also Cuadra v. Gonzales, 

, 950 (8th Cir. 2005).


         Where Congress has explicitly or implicitly left a gap in a statute to be filled by
a particular agency, the agency’s interpretations of the statute having the force of law
are entitled to substantial deference under Chevron. United States v. Mead Corp., 

, 230 (2001). Chevron deference requires courts to give “considerable weight
. . . to an executive department’s construction of a statutory scheme it is entrusted to
administer.” 

. This considerable weight has been interpreted
by the Eighth Circuit to mean controlling weight unless “arbitrary, capricious, or


                                            -8-
manifestly contrary to the statute.” In re Old Fashioned Enters., Inc., 

,
425 (8th Cir. 2001).


       Even where an agency is accorded deference, the “agency must provide a
satisfactory explanation for its actions based on relevant data.” Niobrara River Ranch,
L.L.C. v. Huber, 

, 884 (8th Cir. 2004). This requires an analysis of
whether the decision was “based upon consideration of the relevant factors and
whether there has been a clear error of judgment.” Citizens to Preserve Overton Park,
Inc. v. Volpe, 

, 416 (1971). If an agency’s decision to which deference
is afforded may be supported on any rational basis, we must uphold it. Voyageurs
Nat’l Park 

; Sw. Bell Tel. Co. v. Fed. Comm. Comm’n, 

, 554 (8th Cir. 1998) (“If an agency . . . does not attempt either to close itself
off from informed opinion or to extend its reach beyond the scope of permissible
authority, then it is our duty to accept that judgment if it is rational and not
unreasonable.”). Therefore, even if the agency’s underlying data are flawed,
substantial deference requires the ruling be reversed only if “‘there is a significant
chance that but for the errors the agency might have reached a different result.’” Cent.
S.D. Co-op. Grazing Dist. v. Sec’y of the United States Dep’t of Ag., 

,
899 (8th Cir. 2001) (quoting 

).


       Because an agency’s choice of methodology is typically borne out of the
agency’s expertise, we defer to an agency’s choice of methodology so long as it is not
arbitrary or without foundation. See 

(citing Minn. Pub.
Interest Res. Group v. Butz, 

, 1302 (8th Cir. 1976)). A decision is
arbitrary or capricious if


      the agency relied on factors which Congress has not intended it to
      consider, entirely failed to consider an important aspect of the problem,
      offered an explanation for its decision that runs contrary to the evidence

                                           -9-
      before the agency, or is so implausible that it could not be ascribed to a
      difference in view or the product of agency expertise.


Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 

, 43 (1983).


                                          III


                                          A


        The USFS appeals the district court’s ruling it had no authority to recalculate
the base period use. The USFS argues Dombeck requires the USFS to recalculate the
base period use to ensure Congress’s intent in passing the BWCAW Act is fulfilled.
The USFS interprets the BWCAW Act to give it the authority to correct a major error
in its initial base period calculation made evident by a court ruling. Friends argues
Dombeck does not require the USFS to recalculate the base period use and also argues
that any recalculation of the base period use calculation effectively undermines the
fixed nature of the quota maximum.


       The USFS’s interpretation of Dombeck is not entitled to deference. See Ass’n
of Civilian Technicians v. Fed. Labor Rel. Auth., 

, 50 (D.C. Cir. 2004);
Donovan v. Anheuser-Busch, Inc., 

, 326 (8th Cir. 1981). We agree with
Friends Dombeck does not mandate the USFS to recalculate the base period use, since
its holding was limited to requiring a permit for homeowners and resorts when they
engage in non-exempt lake chain use. 

. However, our
analysis does not end there. The USFS also interprets the BWCAW Act as
authorizing it to recalculate the base period use. Accordingly, we must analyze this
agency interpretation for reasonableness.


                                         -10-
      In determining whether to afford deference to the USFS’s interpretation of the
BWCAW Act, we must first determine whether the congressional intent is clear.
Specifically, we must determine whether Congress intended the USFS to have the
authority to recalculate the statutory cap based upon legal error. See 

.


       The BWCAW Act prohibits the use of motorboats within the BWCA except as
specifically authorized by the Act. BWCAW Act § 4(c), 92 Stat. at 1650. The Act
does not directly address whether the USFS has authority to recalculate the “average
actual annual motorboat use of the calendar years 1976, 1977, and 1978 for each
lake.” We have not been directed to any legislative history materials directly
addressing the issue.


        Agencies given the authority to promulgate a quota are presumed to have the
authority to adjust that quota. See United Gas Improvement Co. v. Callery Props.,
Inc., 

, 229 (1965) (“An agency, like a court, can undo what is wrongfully
done by virtue of its order.”); see also Regions Hosp. v. Shalala, 

, 457-58
(1998) (concluding an agency decision to recalculate a base amount in a fixed
reimbursement system to prevent the distortion of future reimbursements was
reasonable); Dun & Bradstreet Corp. Found. v. United States Postal Serv., 

, 193 (2d Cir. 1991) (“It is widely accepted that an agency may, on its own
initiative, reconsider its interim or even its final decisions, regardless of whether the
applicable statute and agency regulations expressly provide for such review.”);.


       The BWCAW Act creates a gap which must be filled by the USFS (or, more
precisely, by the Secretary of Agriculture): the determination of the “average actual
motorboat use of the calendar years 1976, 1977, and 1978.” The USFS is given no
significant direction regarding this calculation. See 

.


                                          -11-
Accordingly, the statute is silent on the issue and we must accord the USFS deference
to its interpretation of the statute.


        USFS’s interpretation that the BWCAW Act allows it to recalculate the base
period use to correct a major error made manifest by court opinion is reasonable and
not contrary to the purposes of the BWCAW Act. The USFS recalculated the base
period use as a means of addressing the increased permit demand after the Dombeck
ruling. The purpose of the BWCAW Act is to protect the wilderness values of the
BWCAW while maintaining limited motorized access not inconsistent with those
values. See 

-23. Indeed, the purpose of the BWCAW Act
is to establish a maximum level of motorboat use based upon use patterns in 1976-78.
The USFS recalculation, if performed accurately, would not result in a base period use
calculation higher than the “average actual annual motorboat use of the calendar years
1976, 1977, and 1978.” Rather, if performed accurately, it would produce the actual
use figure contemplated by the legislature in passing the BWCAW Act. The USFS
therefore acted reasonably and not contrary to the purposes of the statute when it
recalculated the base period use to include uses previously excluded because of
significant legal error made apparent by a court ruling. As such, the USFS’s
reasonable interpretation must be upheld whether afforded deference under Skidmore
v. Swift & Co., 

, 139-40 (1944), or Chevron. Accordingly, the judgment
of the district court as to the issue of the USFS’s authority to recalculate the base
period use is reversed.


                                          B


       Having determined the USFS has authority to recalculate the base period use
to correct a significant legal error made manifest by court order, we now must
determine whether the USFS recalculated the base period homeowner and resort use
arbitrarily, capriciously, or in a manner contrary to its statutory authority.

                                        -12-
         The USFS claims “the specific means of implementing motorboat use quotas
is left to the discretion of the Secretary.” Appellants’ Reply at 7 (citing 

). The USFS also argues because “[t]he methodology was tailored to the
USFS’s knowledge of, and the best available information regarding, varying patterns
of motorboat use within the three lake-chains, . . . that methodology is entitled to
deference.” Appellants’ Br. at 31.


        While the methodology used by an agency is generally entitled to deference,
this is only true where the methodology is not arbitrary, without foundation, or “so
implausible that it could not be ascribed to a difference in view or the product of
agency expertise.” See 

; State Farm Mut. Auto. Ins. 

. It is our view that some of the data relied upon and calculations
performed by the USFS are so unreliable or inadequately explained as to make
reliance on them arbitrary and capricious.


      The USFS, in recalculating the base period use, sought to determine the amount
of non-exempt lake chain use. To do this, the USFS first estimated total homeowner
and resort lake chain use. The USFS then estimated what percentage of such use was
non-exempt lake chain use. Because the USFS relied upon different data and
methodologies for each lake chain recalculation, we analyze the USFS’s calculations
with respect to each lake chain separately.


                                           1


       We begin by analyzing the recalculated base period use on Moose Lake Chain.
We conclude the USFS recalculated the Moose Lake Chain base period use in an
arbitrary and capricious manner.



                                         -13-
       In calculating what percentage of use during the base period was non-exempt
lake chain use, the USFS relied upon deficient homeowner and resort owner surveys.
The USFS acknowledges the surveys were not conducted in a statistically valid
manner. However, it claims the surveys were adequate for the purposes for which
they were used: to supplement other data or to reduce use estimates from other data
sources. Accordingly, we must determine whether the USFS’s reliance upon the
survey results was reasonable. We conclude it was not.


       Based upon aerial photographs from 1982, the USFS concluded sixty-two
residences (sixteen of which were resort residences or staff cabins) existed on Moose
Lake during the base period. The survey data relied upon by the USFS was obtained
from a 1999 survey completed by thirteen homeowners on Moose Lake. The USFS
limited its survey to current owners of Moose Lake properties who maintained
ownership since the base period. The survey respondents were aware the survey was
related to the USFS’s recalculation of the base period use. The survey asked two
questions of property owners. The survey asked of 1998 use: “How many days did
you use your stickered motorized watercraft in the BWCAW on Moose, Sucker, and
Newfound [Lakes]?” Of 1978 use, the questionnaire asked: “What percentage of your
motorized watercraft use in the BWCAW was for accessing Basswood, Birch, and/or
Knife [Lakes]?” Only five of the thirteen survey respondents provided answers to the
survey question pertaining to 1978 motorboat use. Those responses varied greatly by
individual. We conclude the survey design and implementation was so inadequate as
to make any reliance upon the data obtained from the survey unreasonable.


      To establish the trustworthiness of a survey, it must be shown: (1) that
      a proper “universe” was examined and a representative sample was
      chosen; (2) that the persons conducting the survey were experts; (3) that
      the data were properly gathered and accurately reported; [and] (4) that
      the sample design, the questionnaires, and interviewers, as well as the
      respondents, were unaware of the purpose of the survey.

                                        -14-
Lutheran Mut. Life Ins. Co. v. United States, 

, 378 (8th Cir. 1987).
Although mechanical application of these factors may not be appropriate in every
case, see 

of the opinion the uses to which the survey results were put by the
USFS counsel in favor of a more trustworthy survey.


       The design of the survey was flawed for a number of reasons. First, there is no
record evidence the USFS considered various sample populations to determine which
respondents, if any, would provide accurate and representative results. There is no
record evidence the USFS contemplated surveying property owners during the base
period who have since transferred title or conducting other surveys to obtain more
robust survey data. Given the significant variation in the survey responses, we find
it unreasonable for the USFS to rely upon the responses of only five current
homeowners without considering other means, including performing additional
interviews or surveys, to validate the responses. Although the USFS attempts to
justify its reliance by suggesting its approach is similar to statistically valid surveys
involving a small percentage of the total population, this argument lacks merit. There
is a substantial difference between relying upon survey data results from thousands
of responses and relying upon five responses. Indeed, the USFS acknowledges its
survey was not statistically valid. While this does not invalidate the possibility of
using survey data, it cautions the agency to identify the potential weaknesses in the
data and to seek validation of the results obtained through other means or to apply
correction factors to the survey data. See Menorah Med. Ctr. v. Heckler, 

, 295-96 (8th Cir. 1985) (noting the failure to respond to criticisms that a survey
was untrustworthy makes reliance upon the survey arbitrary and capricious); see also
St. James Hosp. v. Heckler, 

, 1467 n.5 (7th Cir. 1985) (“[I]t is an
agency’s duty to establish the statistical validity of the evidence before it prior to
reaching conclusions based upon that evidence.”).




                                          -15-
        Second, neither Basswood nor Knife Lake was considered part of the Moose
Lake Chain in the USFS’s initial base period calculation; the lakes had their own
quotas. Birch Lake is considered part of Moose Lake, but motorized access to Birch
Lake ended in 1984 and the Moose Lake Chain quota was reduced accordingly.
Accordingly, the percentage of motorized use on Basswood, Birch, and Knife lakes
is irrelevant to the recalculation of the base period use on the Moose Lake Chain, as
adjusted for closures to motorized access pursuant to the BWCAW Act. Nevertheless,
based upon answers to the second question, the USFS concluded thirty-one percent
of BWCAW use was located outside the Moose Lake Chain. This conclusion cannot
be reasonably drawn from the survey results because it is dependent upon an incorrect
determination Birch Lake exists outside the Moose Lake Chain. Further, the question
does not consider the possibility use occurred outside the Moose Lake Chain to lakes
other than Basswood, Birch, or Knife. We therefore find the USFS’s conclusion that
thirty-one percent of BWCAW use by Moose Lake Chain resort and home owners was
outside the Moose Lake Chain to be arbitrary and capricious.


       Similarly, the USFS’s correlated conclusion sixty-nine percent of BWCAW use
by Moose Lake Chain resort and home owners during the base period was within the
Moose Lake Chain must be re-examined. The conclusion is questionable because,
assuming arguendo the accuracy of the figure, it may not be reasonable to conclude
all use on the Moose Lake Chain was non-exempt lake chain use. The record
evidence clearly establishes some Moose Lake Chain homeowner use was limited to
Moose Lake. A.R. Exh. 61, at 3. The USFS did reduce by thirty-one percent its
estimated actual use figure for the Moose Lake Chain, but that reduction was based
on survey results for use on Bass, Birchwood, and Knife lakes, not based on use of
Moose Lake only. Thus, the USFS’s conclusion all use on the Moose Lake Chain was
non-exempt lake chain use is unsupported by the record.




                                        -16-
       Aside from the survey design flaws and the arbitrary conclusions resulting
therefrom, the survey was also performed in a manner to make the results unreliable.
In particular, we find it unreasonable to rely exclusively upon survey results, without
analyzing the potential for bias or adjusting the data based upon any bias found, when
the respondents are aware of the purpose of the survey.10 See Lutheran Mut. Life Ins.

; Menorah Med. 

; St. James 

n.5. Further, it is unreasonable to rely upon survey responses regarding
behavior twenty years earlier without considering the possibility such responses are
not wholly accurate and either applying correction factors to the responses determined
to be less than wholly accurate or attempting to verify the responses through other
means. We are especially troubled by the USFS’s complete reliance upon the survey
data, without even studying the results for possible inaccuracies, in spite of the fact
only five of the thirteen respondents answered the survey question regarding use
patterns in 1978.


       We agree with the district court that “[o]ne does not need to be a statistician to
apprehend the numerous flaws in this sample.” Friends of the Boundary Waters
Wilderness v. Bosworth, No. Civ. 03-624, 

, at *10 (D. Minn. Aug.
26, 2004). Sample size, potential for bias, interviewing techniques, reliability of
extrapolating data, and poor recollection are all relevant factors the agency failed to
properly consider in analyzing the survey results, making the resulting estimates
arbitrary and capricious. We believe these survey deficiencies are so obvious the


      10
          The USFS argues “[i]t seems unlikely that better results would be obtained
by going back now to the same sources with a different interview form, after years of
litigation and explication of the methodology.” Although the history and prominence
of this litigation within the BWCAW and knowledge of the surveys previously
conducted might affect the reliability of future surveys, this concern neither eliminates
the possibility of estimating base period use through means other than surveys,
additional interviews, or both. Accordingly, such a concern is insufficient to support
the USFS’s reliance upon the flawed survey data.

                                          -17-
USFS should have considered the potential impact these deficiencies might have on
the trustworthiness of the surveys. Its failure to do so renders reliance upon the survey
results arbitrary and capricious. See Menorah Med. 

. Further,
the survey results did not support the conclusions the USFS drew from them. We
therefore conclude the USFS’s recalculation of non-exempt Moose Lake Chain use
during the base period was arbitrary and capricious.


                                           2


      The USFS non-exempt Saganaga Lake Chain use calculations are also arbitrary
and capricous or otherwise lack foundation. The USFS determined, based primarily
upon Cook County records, there were thirty-four residences on Seagull River and
Gull Lake within the Saganaga Lake Chain in 1976-78. Eight of these residences
were resort residences or resort crew cabins.


      In 1981, the USFS assumed 512 day-use permits were attributable to exempt
use. In its recalculation, the USFS does not use the figure from 1981, but arrives at
a new estimate by relying on USFS records showing an average of fifteen trips per
household per season, multiplying that number by a total of thirty-four homeowners,
and then adding an estimated 883 trips by resort guests. It is unclear whether the
figure 512 applied in 1981 represented use by homeowners only or use by a
combination of homeowners and resort guests. We conclude that before the USFS
may discard the apparent assumption from 1981 that exempt use amounted to 512
day-use permits, it must explain adequately how the 512 figure relates to its current
analysis or why the original figure should be discarded.


        Even more troubling is the fact the USFS justifies its recalculation by stating:
“it is unsurprising that the estimated use of Saganaga Lake by the affected parties of

                                          -18-
1,622 is much greater than the number of permits attributed to them because most
exempt users did not get day-use permits.” The USFS points to no record evidence
suggesting the USFS attempted to ascertain what percentage of homeowners or other
exempt users did not obtain day-use permits when traveling either on their particular
lake or when engaged in non-exempt lake chain use. Accordingly, the USFS claim
most “exempt users did not get day-use permits” presently lacks foundation.


        The USFS’s recalculated resort non-exempt lake chain use must also be set
aside because it relies upon unreliable and faulty survey data. The USFS validated its
estimate that resort guests and staff used Saganaga Lake six days per week with
interviews of one former resort owner, two current resort owners, and one former
neighbor. Viewing this data in the light most favorable to the USFS, we conclude the
data is insufficiently reliable to justify disregarding, without adequate explanation, its
prior calculation, which was made closer in time to the base period and formed the
basis for the Saganaga Lake Chain quotas for twenty years, that approximately forty
percent of occupied resort sites used Saganaga Lake daily. A.R. Exh. 86, at 25; see
Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Svcs., 

, 2699
(2005) (holding a reversal of a longstanding agency position requires adequate
justification to avoid being arbitrary and capricious); Good Samaritan Hosp. v.
Shalala, 

, 417 (1993) (providing considerably less deference to agency
reversals of position than to longstanding agency views). For the reasons stated
above, we conclude the USFS’s recalculation of homeowner and resort non-exempt
Saganaga Lake Chain use was arbitrary and capricious.


                                            3


       The USFS’s recalculated base period use for the Farm Lake Chain is similarly
arbitrary and capricious and must be set aside. Although the USFS claims it
“considered use patterns particular to home and resort owners in the different lake

                                          -19-
areas in making its estimates,” the USFS did not adequately explain its extrapolation
of Moose Lake Chain use data to the Farm Lake Chain. Indeed, the record establishes
South Farm Lake was fished approximately one fourth as often as Moose Lake. A.R.
Tab C.


      Further, it is unclear from the record whether Farm Lake Chain homeowners
located outside the BWCAW (only South Farm Lake is located within the BWCAW)
were included in the USFS’s initial base period use calculation for the Farm Lake
Chain. That is, we cannot determine from the record whether the USFS, in its initial
calculation, considered Farm Lake Chain homeowners located outside the BWCAW
exempt from the quota system.


        Similarly, the USFS’s recalculation of non-exempt resort Farm Lake Chain use
is arbitrary and capricious. The USFS concluded resorts in the Farm Lake Chain were
not entirely focused on BWCAW use. Accordingly, the USFS interviewed three
resorts and extrapolated the results obtained to the four resorts having exempt permits
in 1998. The three interviewed resorts had capacities of 4 rental units, 12 rental units,
and 13 rental units, and estimated motorboat use per week of 2 days, 7 days, and 39
days, respectively.


       The complete reliance upon such limited survey data to determine resort use on
the Farm Lake Chain, without study to determine the reliability of the results obtained,
is arbitrary and capricious. Given the wide variety of responses, we find it troubling
the USFS did not interview the fourth resort regarding its use of South Farm Lake or
attempt to validate the responses in any way. For the foregoing reasons, we conclude
the USFS’s recalculation of the base period motorboat use within the Farm Lake
Chain was arbitrary and capricious.




                                          -20-
       Because we conclude the USFS recalculated the base period homeowner and
resort motorboat use on the Moose, Saganaga, and Farm Lake Chains in an arbitrary
and capricious manner, we affirm the district court on these issues.


                                          C


       Friends also challenges the USFS’s recalculation of the towboat quotas. The
towboat quotas are separate from the homeowner and resort motorboat quotas, though
total motorboat use by homeowners, resorts, their guests, and towboats cannot exceed
the base period use. See 

-22. The USFS has the
responsibility of allocating motorboat use among homeowners, resorts, guests, and
towboats in a manner consistent with the BWCAW Act. Currently, the towboat quota
occupies the space between the maximum homeowner and resort quota and the base
period quota maximum. See A.R. Exh. 47, at 63. That is, towboats are allowed to the
extent their use, when added to the homeowner, resort, and guest use, does not exceed
the base period use.


      The record is not clear as to whether towboats were included in the original
base period use. See A.R. Exh. 53, at 2; A.R. Exh. 81, at 3. The USFS must explain
adequately why it concludes towboat use was exempted or otherwise not counted
during the 1981 calculation of actual use before it undertakes any future recalculation
of towboat use.


                                          D


      Although the district court did not reach this issue, both parties agree judicial
economy is best served by our ruling on the USFS’s adjusted motorboat quotas based
upon its recalculated base period use. We agree there is no need to remand this issue

                                         -21-
to the district court for initial determination. Having determined the USFS improperly
recalculated the base period use, we conclude the USFS must recalculate the quotas
consistent with BWCAW Act and the views expressed herein.


                                           IV


       We reverse the district court insofar as we conclude the USFS has authority to
recalculate the base period use to correct a significant legal error made manifest by the
Dombeck ruling. We affirm the district court by concluding the USFS’s homeowner
and resort non-exempt lake chain use recalculation for the Moose, Saganaga, and
Farm Lake Chains was arbitrary and capricious. We remand to the district court with
directions to remand to the USFS for a recalculation of the base period use and
motorboat quotas consistent with the BWCAW Act and the opinion. The agency
should do so at the earliest practicable time and include the Friends, Conservationists
with Common Sense, and other affected parties in the rule-making process.
                        ______________________________




                                          -22-