OPINION
MEROW, Judge.
This is a pre-bid protest in which plaintiff Redland Genstar, Inc. (“Redland”) objects to solicitation no. DACW31-97-B-0022 issued by the Baltimore District of the United States Army Corps of Engineers (“Corps”). The solicitation seeks bids for the construction of a stone dike around the historic perimeter of Poplar Island, Maryland. Redland, an operator of several Maryland quarries, alleges that the solicitation unduly restricts competition because it contains unreasonable stone abrasion tests. Redland also asserts that the list of approved stone sources in the solicitation unreasonably excludes its Texas, Maryland quarry. Redland asks the court to declare the abrasion tests invalid and direct the Corps to add the Texas quarry to the approved source list. Redland also asks the court to enjoin the Corps from advertising or receiving bids pursuant to the solicitation.
The matter is currently before the court on the parties’ cross-motions for judgment on the administrative record.1 For the reasons stated below, plaintiffs motion is granted to the extent it seeks a declaration that the abrasion test requirements in the solicitation are invalid. Plaintiffs motion is otherwise denied. Defendant’s motion is denied in its entirety.
BACKGROUND
Poplar Island is located in the Chesapeake Bay thirty-two miles southeast of Baltimore and two miles west of Maryland’s Eastern Shore. The island has severely eroded from nearly 1,100 acres in 1847 to about five acres today, causing the loss of important wildlife habitat. At the same time, the Maryland Port Authority (“MPA”) has run critically short of space to dispose of sediment dredged from shipping channels approaching the Port of Baltimore. In response, Maryland and the Corps have undertaken the “Poplar Island Restoration Project” which calls for the construction of a containment dike around the 1847 perimeter of the island. Dredged sediment will be discharged within the dike to restore the island while simultaneously alleviating the MPA’s shortage of dredged material disposal space. The Corps is responsible for 75% of the dike construction costs and Maryland is responsible for the remainder. Maryland is also exclusively responsible for all of the dike maintenance costs.
A joint venture of engineering firms (“JV”) was hired by the MPA to study the Poplar Island environment and prepare an economical dike design. In its September 28, 1995 report, the JV found that wind-generated waves approach the island from all directions. The most intense waves approach from the north and south and will affect the western perimeter dike, while the eastern perimeter dike will be subjected to considerably less intense conditions. The JV also found that the severity of storms affecting the dike is likely to increase over time. For instance, the maximum breaking wave height for waves approaching from the north is about 3 feet for a 5-year storm event (a [223]*223storm of a particular intensity statistically likely to occur once every five years), 9 feet for a 25-year event, and 10 feet for a 50-year event. Administrative Record (“AR”) Tab 53 at 2-21.
Based on an optimized approach balancing initial construction costs with maintenance costs associated with storm-induced damage, the JV determined that the dike design should be based on a 25-year storm return period even though the Corps typically designs its projects on the basis of a 50-year period. Id. at 6-2 — 3, 10-5. The JV concluded that it was “reasonable to design the Poplar Island containment dikes for a shorter project life because the purpose of the project is to retain uncontaminated dredged sediment rather than serve as flood or shore protection structures for developed real estate.” Id. at 6-3. The design chosen by the JV calls for the construction of a 36,700-foot dike encompassing 1,100 acres. The western perimeter dike is to consist of two layers of 3,000-4,000 pound armor stone overlaying two layers of 250-pound stone protected in front by two additional layers of 2,000 pound toe armor stone. The eastern perimeter dike is to consist of two layers of 250-pound armor stone.
The JV also prepared dike construction specifications stating that all stone “shall consist of fresh, sound, hard, dense, durable stone crystalline igneous or metamorphie rock.” AR Tab 86 at 02486-2. The specifications contain numerous stone quality tests generally following the recommendations in the Corps’ engineering manual, EM 1110-2-2302, “Construction With Large Stone” (1990), which presents extensive guidance for the selection, evaluation and use of large stone materials in construction. AR Tab 21 at 2, 1-1, 6-2 — 4; Tab 86 at 02486-3-4. The manual is included in a list of referenced publications in the JVs specifications. AR Tab 86 at 02486-2.
With respect to stone abrasion resistance, EM 1110-2-2302 recommends American Society for Testing and Materials (“ASTM”) test C 535, “Standard Test Method for Resistance to Degradation of Large Size Coarse Aggregate by Abrasion and Impact in the Los Angeles Machine.” AR Tab 21 at 6-4; Tab 72. The L.A. Machine is essentially a large, rotating steel drum. ASTM C 535 states that stone samples no larger than 3” are to be placed in the L.A. Machine with twelve steel spheres and rotated. The stone samples are then removed and the percentage of abrasion loss calculated. The test itself specifies 1,000 revolutions, AR Tab 19 at 2, but EM 1110-2-2302 recommends 500 and states that, “[rjoughly, losses less than 20 percent for 500 revolutions are generally considered satisfactory while losses exceeding 40 percent suggest probable poor service.” AR Tab 21 at 64. The JV followed the guidance in EM 1110-2-2302 and specified ASTM C 535 at 500 revolutions with a maximum 20% loss as the appropriate abrasion test for the Poplar Island armor stone. AR Tab 86 at 02486-4.
Another abrasion test at issue in this action is ASTM C 131, “Standard Test Method for Resistance to Degradation of Small-Size Coarse Aggregate by Abrasion and Impact in the Los Angeles Machine.” This test applies to smaller stone samples (no greater than 1.5”) which are to be rotated 500 times in the L.A. Machine with between six and twelve steel spheres, depending on the gradation of the stone. AR Tab 18 at 1-2. The percent of abrasion loss determined by ASTM C 131 “has no known consistent relationship to the percent loss for the same material when tested by Test Method C 535.” Id. at 3 n. 7; Tab 19 at 2 n. 4. The Corps’ “Shore Protection Manual” (“SPM”), issued in 1984, contains a paragraph on stone quality requirements which states that stone should conform to ASTM C 131 but does not specify a maximum permissible loss. AR Tab 25 at 6-97.
Finally, the JVs specifications contain a list of stone quarries, including Redland’s Texas quarry, which, according to the JV, were “investigated, tested, and approved as sources of armor stone, underlayer stone, toe armor stone and quarry run.” AR Tab 86 at 02486-4 — 5.
In April 1995, while the JV was evaluating potential dike designs, the Corps solicited bids for the construction of a Poplar Island test dike section. The specifications stated that stone used in construction had to be [224]*224hard, durable, “resistant to abrasion and not subject to disintegration when exposed to air or sea water or to freezing and thawing,” and “of such quality that its permanence is assured under all conditions to which it will be subjected.” AR Tab 60 at 02486-3. The specifications also referenced and followed the recommendations in EM 1110-2-2302 with respect to stone quality tests. Id. at 02486-1 — 3. The abrasion test specified was ASTM C 535 at 500 revolutions with a maximum 20% loss. Id. The specifications also contained a list of sources, including the Texas quarry, which the Corps determined could “produce the quality of stone for the Armored Dike section for this job.” Id. App. D. Plaintiff did not supply stone for the project, however. Oral Argument Transcript (“Tr.”) at 52.
In May 1996, the Corps assumed responsibility for finalizing the Poplar Island dike design and specifications. Sometime in early 1997, the Corps prepared a draft list of approved stone sources. The list was not distributed to the public, but Redland somehow became aware of it and learned that it did not include the Texas quarry. Redland contacted the Corps and requested that the stone at the Texas quarry be examined so that the quarry could be added to the list. Tr. at 57.
On April 11, 1997, Michael Saint-Clair, Corps geologist, and two other Corps representatives visited the Texas quarry. They met with Page Herbert, Redland geologist, and William Blanton, president of Chesapeake Materials, Inc., a company hoping to work with Redland in connection with the project, and were shown a 700,000-ton stone stockpile. Although he did not comment on the perceived quality of the stockpile at the time, in an April 24, 1997 memorandum, Mr. Saint-Clair describes it as “a rubble pile or stone trash pile.” AR Tab 24 at 1. He wrote that the stone, known as Cockeysville Marble, showed numerous signs of weaknesses and that “Los Angeles Abrasion tests indicate a very friable material (40% loss)” which may be “subject to break down in a high energy, dynamic environment such as the Poplar Island project site.” Id. at 2-3. He also states that the same type of stone from other Redland quarries had been used in previous Corps projects and “included as much as 60% by volume of ... questionable material.” Id. at 2. Mr. Saint-Clair states that the Texas quarry was not added to the approved source list “[djue to the lack of service record, lack of internal records showing the source of material in the stockpile, inherent zones of weakness and cracking in the tale-schist marble, and high abrasion percentages of the Cockeysville Marble.” Id. at 3.
On April 16,1997, five days after the Texas quarry visit, an unidentified Corps employee prepared a memorandum entitled, “Poplar Island Habitat Restoration — Sources of Stone for Armor,” the first document in the record in which the Corps discusses stone quality requirements for the Poplar Island project. Section 4.1 states that “[sjpecifica-tions for the armor stone for the Poplar Island project were developed considering guidelines presented in [EM 1110-2-2302], conditions specific to the Poplar Island site, and available commercial sources.” AR Tab 38 at 1. With respect to abrasion resistance, § 4.1.4 provides:
EM 1110-2-2302 recommends that the loss from 500 cycles of the LA abrasion test (method CRD-C 148)[2] not exceed 20 percent and states that losses exceeding 40 percent suggest probable poor service. In place of the CRD-C 148 test method, the CRD-C 535 test method applicable to larger sample particle size, will be used for the Poplar Island specifications. The final specification limits the losses from the LA abrasion test to 30 percent.
Id. at 2. An unidentified Corps employee on an unspecified date crossed out the “CRD-C” preceding “535” and handwrote “ASTM-C.” As stated above, ASTM C 535 is indeed the abrasion test applicable to large size stone samples. Also handwritten in the margin next to § 4.1.4 is a reference to ASTM C [225]*225131. The purpose of the reference is indeterminable.
Section 4.3 discusses previous Chesapeake Bay projects involving “large rock,” including Tedious Creek, Punch Island, Chesapeake Beach, and Island Creek. Id. at 2-3. The record does not contain the specifications for the Chesapeake Beach project but indicates that the other projects involved armor stone and followed the recommendations in EM 1110-2-2302 with respect to stone quality tests. AR Tab 37. The Tedious Creek, Punch Island, and Chesapeake Beach projects used the ASTM C 535 abrasion test at 500 revolutions with a maximum 40% loss. Island Creek used the same test but specified a maximum 20% loss. Id.; AR Tab 38 at 2. The memorandum states that Tedious Creek, Punch Island, and Chesapeake Beach consisted of small quantities of rock and were located in a lower energy environment than Poplar Island. Id. As a result, the author determined that higher quality stone was needed for Poplar Island. Id. at 2-3.
A chart attached to the April 16, 1997 memorandum discusses the advantages and disadvantages of permitting a maximum abrasion loss of 20%, 30%, or 40%. The author does not explicitly state which abrasion test (ASTM C 131 or ASTM C 535) is under consideration. However, the chart does state that a maximum loss of 20% would comply with the “strictest EM criteria,” a loss factor of 30% “[rjemains within EM recommended range of LA abrasion of 20 to 40%,” and that a loss of 40% “[plushes to upper limit of EM recommended range of LA abrasion.” Since the only “EM” referenced anywhere in the memorandum is EM 1110-2-2302, and since the only “EM” containing a discussion of a permissible abrasion loss range of 20%-40% is EM 1110-2-2302, it must be concluded that the abrasion test being discussed is the abrasion test recommended in EM 1110-2-2302-ASTM C 535 at 500 revolutions.
The chart states that a maximum 30% loss under ASTM C 535 at 500 revolutions would ensure “reliable good quality stone with probable satisfactory performance” and require “[l]ow to moderate frequency of maintenance.” The author also states that the Texas quarry stone suffers 40% abrasion loss and concludes that a 30% loss requirement would probably result in a protest from Red-land. The chart states that if 40% loss were permitted, it might allow the use of the Texas quarry’s stone with an estimated savings of $1.5 to $3 million. Redland’s Texas quarry is the only quarry specifically referenced anywhere in the chart or the memorandum.
Section 5 of the memorandum, “Conclusions,” states that, “[biased on the recommendations in EM 1110-2-2302 and the experience on previous Corps projects, it is considered appropriate to require higher quality stone for all armor stone to assure satisfactory performance. The armor stone should have a maximum LA abrasion value of 30 percent after 500 revolutions.” AR Tab 38 at 3. Given the content of the memorandum and attached chart discussed above and the fact that the only “reeommendation[ ] in EM 1110-2-2302” with respect to abrasion is to use ASTM C 535 at 500 revolutions, the April 16, 1997 memorandum concludes that ASTM C 535 at 500 revolutions with a maximum 30% loss is the appropriate abrasion test for the Poplar Island armor stone.
Also attached to the memorandum is a list of potential sources of stone containing abrasion test results obtained from the Maryland State Highway Administration Aggregate Bulletin (“MSHAAB”). The MSHAAB is apparently the source of the author’s determination that the Texas quarry stone suffers 40% loss. AR Tab 77 at 10. However, the MSHAAB results are based on ASTM C 131, not ASTM C 535. AR Tab 73. Relying on this fact, defendant asserts that the April 16 memorandum actually concludes that ASTM C 131 at 500 revolutions with a maximum 30% loss is the appropriate abrasion test for Poplar Island. DSOF Iff 54-55; Def. Reply at 10; Tr. at 20-22. However, this assertion is not supported by the content of the memorandum or any other document in the record. The explanation which is consistent with the memorandum and attached chart is that the author was under the mistaken assumption that the MSHAAB abrasion test data was based on ASTM C 535. This explanation is further supported by an April 10, 1997 memorandum prepared by Dennis E. Webb, Chief [226]*226Engineer for the Baltimore District’s Geo-technical and Water Resources Branch, which references a meeting with Corps personnel, including Mr. Saint-Clair, and states that abrasion test data based on ASTM C 535 can be obtained from the MSHAAB:
LA Abrasion calculations are preformed [sic] by others and reported in the Aggregate Bulletin published annually by the Maryland State Highway Administration, Materials Testing Division. LA Abrasion calculations are obtained from laboratory testing of stone source materials in accordance with ASTM C 535, Test Method for Resistance to Degradation of Large-Size Coarse Aggregate by Abrasion and Impact in the Los Angeles Machine. The Aggregate Bulletin can be obtained free of charge ...
AR Tab 37.
On April 25, 1997, the April 16 memorandum was revised to conclude that, “[b]ased on the design conditions at the Poplar Island site and recommendations in EM 1110-2-2302, ... [e]xposed stone should have a maximum LA abrasion value of 20% percent [sic] after 500 revolutions.” AR Tab 39 at 4.3 To support this conclusion, an expanded comparison of Poplar Island to other Corps projects was added in § 4.1 and § 4.2. It was noted that most Baltimore District projects “are exposed to a narrow range of storms since they have a single dominant direction of exposure compared to Poplar Island which is exposed to storms from three different directions” and that “the size of the armor protection for the Poplar Island project was based on a 25-year storm event in contrast to some other [Baltimore District] projects which have been based on a 50-year event.” Id. at 2. The author also states that the dike would “be subject to storm events exceeding the design event more frequently than most other projects” and that dike failure “could result in extensive erosion of dike core materials and potential breach of the containment dike requiring costly emergency reconstruction of the damaged dike seetion(s).” Id. For these reasons, § 4.1 concludes that it is “appropriate to maintain stringent requirements for armor stone quality to assure satisfactory performance.” Id.
In § 4.2, the author discusses the wave conditions at the Island Creek project site and states that, because of the “higher energy environment,” it was “appropriate to require high quality rock capable of resisting deterioration.” Id. The author concludes that, based on the “similarity in exposure and design wave heights” between Poplar Island and Island Creek and the Corps’ need for “more durable stone,” the abrasion test used for Island Creek — ASTM C 535 at 500 revolutions with a maximum 20% loss — is also appropriate for Poplar Island and “consistent with the recommendations of EM 1110-2-2302.” Id.
Section 4.3 discusses the availability of sources which could meet the specifications and states that “10 sources of stone satisfying the maximum LA abrasion value of 20 percent as recommended in EM 1110-2-2302” will be able to compete. Id. at 3 (emphasis added). Likewise, § 4.4 provides that “[specifications for the armor stone for the Poplar Island project were developed considering guidelines presented in [EM 1110-2-2302], design conditions specific to the Poplar Island site, and available commercial sources.” Id. However, § 4.4.4, “Abrasion,” states:
EM 1110-2-2302 recommends that the loss from 500 cycles of the LA abrasion test (method CRD-C 148) not exceed 20 percent. In past [Corps] projects LA Abrasion test method CDR-C 117 (based on ASTM C 131) was used verses method CRD-C 148 (based on ASTM C 535). LA Abrasion (CRD-C 117) testing results are also obtained from the MD SHA Aggregate Bulletin. While no direct correlation exists between methods CDR-C 117 and CDR 148, it is known that CRD-C 117 is the more rigorous of the two tests. As noted from previous Aggregate Bulletins a variation of ± 1 to 2% exists between tests conducted on similar quarry material in [227]*227prior years. The final specification limits the losses from the LA abrasion test to 20 percent and includes quarries within 2% points.
Id, at 4. This section indicates the author had learned that the MSHAAB data was based on ASTM C 131. Defendant asserts that this paragraph and the source list attached to the memorandum (based on the MSHAAB) show that the author actually concluded that ASTM C 131 with a maximum 20% loss is the appropriate abrasion test for Poplar Island. DSOF 1156; Def. Reply at 10-11, 13-14; Tr. at 23-24, 84. Section 4.4.4, although somewhat ambiguous, appears to support defendant’s assertion. However, defendant’s reading is inconsistent with the analysis in § 4.2, which concludes that the test used for Island Creek — ASTM C 535 at 500 revolutions with a maximum 20% loss — should be used for Poplar Island, and with § 4.3 and § 5, which state that the abrasion test chosen was based on the recommendations in EM 1110-2-2302, which specifies ASTM C 535 at 500 revolutions. Defendant’s reading is also inconsistent with the fact that on July 15, 1997, the Corps again changed the memorandum to conclude that ASTM C 131 with a maximum 20% loss was appropriate. AR Tab 20 at 2, 5. Obviously, this change would not have been necessary had the memorandum already reached that conclusion.
Hence, either the author of the April 25 memorandum intended to conclude that ASTM C 535 at 500 revolutions with a maximum 20% loss is the appropriate test for the Poplar Island armor stone, as the brunt of the memorandum and the facts in the record suggest, or, if defendant’s reading is correct, the document concludes that ASTM C 131 is appropriate based largely on an analysis justifying the use of ASTM C 535.
On June 6, 1997, the Corps issued the solicitation and specifications for Phase I of the Poplar Island Restoration Project. Phase I involves the construction of a segment of the perimeter dike enclosing 640 acres and consisting of over 500,000 tons of 250 to 4,000 pound armor stone. The stonework specifications reference EM 1110-2-2302 and most of the stone quality tests follow the recommendations therein. AR Tab 1A at 02486-1, 4; Tab 21 at 6-2. However, with respect to abrasion resistance, the Corps departed from the guidance in EM 1110-2-2302 and prohibited the use of stone which suffered more than 20% loss under ASTM C 131. Id. at 024864.4
The specifications also contain a list of “Sources of Approved Stone” which includes ten of the fifteen potential sources identified by the Corps. Id. at 02486-5; AR Tab 39. Redland’s Texas quarry is not included on the list but its Churchville, Maryland and Frederick, Maryland quarries are. However, the MSHAAB indicates that the stone from these quarries suffers 22% and 24% loss, respectively, under ASTM C 131, AR Tab 77 at 9-10, and Redland has stated that neither quarry can produce stone passing the specified abrasion test. Hence, the abrasion test requirement chosen by the Corps eliminates at least seven of the fifteen potential stone sources.
The due date for bids was set for July 15, 1997.
Following the issuance of the construction specifications, Mr. Herbert and Mr. Blanton sent various letters to the Corps attacking its decision to use the ASTM C 131 test, asserting that ASTM C 535 as recommended in EM 1110-2-2302 is the appropriate test, and asking that the Texas quarry be added to the approved source list. AR Tab 7 Exs. 2-4.
On July 3, 1997, Mr. Saint-Clair telephoned Mike Hart of the MPA to discuss the differences between ASTM C 131 and ASTM C 535 at 500 and 1,000 revolutions. According to Mr. Saint-Clair’s memorandum of the conversation, Mr. Hart “agreed that [abrasion] tests represent an indexing tool” and that the use of ASTM C 131 “with the MD SHA history of testing was sound.” He also stated that he felt the “best quality and durable stone should be placed during construction as to allow minimal repair or maintenance of the dike,” for which Maryland is [228]*228responsible. He also pointed out that a pier in Los Angeles used ASTM C 535 at 1,000 revolutions with a maximum 50% loss. AR Tab 44.
On July 14,1997, the day before bids were due, Redland filed a protest with the General Accounting Office (“GAO”) alleging that the use of ASTM C 131 as opposed to ASTM C 535 at 500 revolutions with a maximum 20% loss rendered the solicitation unduly restrictive because it was unnecessary, eliminated the Texas quarry, and effectively allowed only one or two quarries to compete. AR Tab 7. Mr. Blanton asserted in an attached affidavit that the use of ASTM C 131 could increase stone costs by $3 million to $10 million. Id. Ex. 1 at 2. The Corps immediately postponed bid opening pending resolution of the protest.
The next day, the Corps revised the April 25, 1997 memorandum. The Corps changed the conclusion in § 4.2 containing the discussion of the similarities between Poplar Island and Island Creek to read that ASTM C IS 1 at 500 revolutions with a maximum 20% loss was the appropriate abrasion test for Poplar Island. AR Tab 20 at 2. The Corps also added a reference to the “multiple directions of exposure at the Poplar Island site,” id., but this information was already contained in § 4.1 of the April 25 analysis. AR Tab 39 at 2. In § 4.3 and § 5, the Corps deleted the reference to EM 1110-2-2302 which, as stated, recommends ASTM C 535 at 500 revolutions. AR Tab 20 at 2, 4. The Corps also added a reference to ASTM C 131 in § 5. Id. at 4. No analyses or other substantive information were added to the document to support these changes.
Aso on July 15, 1997, Mr. Saint-Clair telephoned Richard Lutton, a former Corps employee who is now working for the agency on a consulting basis. Mr. Lutton wrote the sections of EM 1110-2-2302 discussing stone testing criteria and has published several papers on the subject. In an article he wrote in 1988, he included ASTM C 535 at 1,000 revolutions with a maximum 25% loss in a table of “suggested criteria of suitability.” AR Tab 22 at 1855. Mr. Saint-Clair asked Mr. Lutton to explain the disparity between his suggestion in the article and his recommendation in EM 1110-2-2302 to apply ASTM C 535 at only 500 revolutions with a maximum 20% loss. According to Mr. Saint-Clair’s memorandum of the conversation, Mr. Lutton “indicated that he may have lowered the standard because the test was too severe and the producers in geographical areas of this country (i.e., North-Central, Mid-West) would not be able to meet this standard.” Mr. Lutton also stated that ASTM C 535 “would be preferred” over ASTM C 131, although is not clear if he was referring to ASTM C 535 as specified in the manual or the article. He.added that the function of the abrasion test “was as an indexing tool and may be the most doubtful of all tests.” Finally, he “felt that [Mr. Saint-Clair] used sound engineering judgment using the 131 as just an index tool.” AR Tab 51.
On July 17, 1997, an unidentified Corps employee prepared a two-page document entitled, “Poplar Island Habitat Restoration Amor Stone Deterioration Aalysis,” which concludes that increasing the permissible abrasion loss under ASTM C 131 from 20% to 40% will increase maintenance costs by $9.6 million over a 23-year period. AR Tab 23. The author determined that if stone passing ASTM C 131 with a maximum 20% loss is used, 4,000 tons of stone will be lost each year. Over a 23-year period, this amounts to 92,000 tons or 17% of the total initial armor stone tonnage (539,100). This figure is broken down as follows: 10% “[disintegration of stone by cracking and splitting due to freeze-thaw action”; 1% “[deterioration of stone by chemical weathering resulting in spalling”; and 6% “[deterioration of stone by mechanical breakdown due to abrasion at points of contact between stones.” Id. at 1. The basis of this breakdown is not provided. The document then states that the amount lost will increase to 7,300 tons per year, or 31% of the initial armor stone tonnage over a 23-year period, if stone suffering 40% loss under ASTM C 131 is used. The 31% figure is broken down as follows: 15% due to cracking and splitting; 5%' due to spalling; and 11% due to abrasion. Id. The basis of this asserted increase is not provided, nor is there any indication as to how a decrease in abrasion resistance will produce [229]*229an increase in damage due to freeze-thaw or chemical weathering. The document concludes that the cost of this additional 3,300 tons of stone per year will amount to $9.6 million over 23 years. Defendant has determined that it cannot disclose the basis of this purported cost increase “because of the sensitive nature of the cost data which, if disclosed, would cause the Government irreparable harm.” PI. Counter-Statement of Facts and Proposed Additional Facts (“PCS & PAF”) Ex. A at 2.
Although not provided in the July 17 analysis, defendant has stated in subsequent filings that the basis of the Corps’ determination that 4,000 tons of stone will have to be replaced each year if stone passing ASTM C 131 with a maximum 20% loss is used is contained in §§ 1.4.1 and 2.2.1 of Appendix D of the JV report. Both sections are entitled, “Damage Assessment for Wave Impact Force,” and contain detailed mathematical equations. Defendant asserts that these equations reveal that 4,533 tons of armor stone (which the Corps rounded off to 4,000) will have to be replaced each year due to wave impact. None of cited equations contain abrasion resistance as a variable. In addition, the JV never referenced the ASTM C 131 abrasion test anywhere in its voluminous report or in its construction specifications.
On July 18, 1997, the Corps amended the solicitation to provide that stone passing ASTM C 131 with a maximum 20% loss or ASTM C 535 at 1,000 revolutions with no more than 25% loss could be used the project. AR Tab 15. The effect of this addition on competition is indeterminable because the record does not contain any data indicating whether a quarry exists which can meet this test.
On July 21,1997, the Corps requested that GAO summax-ily dismiss Redland’s protest in light of this “cori’ective action.” AR Tab 16. Redland objected to the request, asserting that the addition did not alleviate the unduly restrictive nature of the specifications. AR Tab 17. On July 29,1997, representatives of GAO, Redland, and the Corps participated in a conference aimed at settling the protest but their efforts were unsuccessful. On July 31, 1997, apparently assuming that GAO would not sustain Redland’s protest, Tr. at 108, the Corps amended the solicitation to set August 12, 1997, as the due date for bids. AR Tab 31.
On August 8, 1997, the contracting officer (“CO”) prepared a report for submission to GAO stating that the Corps appropriately determined its minimum needs and that any restriction on competition resulting from its choice of abrasion tests is reasonable. AR Tab 34 at 1. In support of the decision to use ASTM C 131, the CO cites the SPM, id. at 5 K11, which, as stated above, contains a pai’a-graph on stone testing which references the test but does not specify a maximum permissible loss. AR Tab 25 at 6-97. The CO also relies on the fact that other Corps districts have used the test and cites the Huntington District’s Dry Run Levee project located in Columbus, Ohio, which applied ASTM C 131 with a maximum 30% loss, and the Norfolk District’s James R. Olin Flood Control Project located in Buena Vista, Virginia, which applied ASTM C 131 with a maximum 20% loss. AR Tab 34 at 5 1110; Tab 30 at 02225-1; Tab 26 at 02200-2.
The CO also states that the Corps’ decision to add ASTM C 535 at 1,000 revolutions with a maximum 25% loss “was based on information available about ASTM C 535” in Mr. Lutton’s 1988 article and in a 1993 publication by Mr. Lutton and Sam Wong. AR Tab 34 at 8 n. 10. The 1993 article compares stone quality criteria applied by states to those applied by federal agencies, including the Coxps, and focuses primarily on riprap weighing less than 500 pounds. AR Tab 32. The authors state that a discussion of larger stone is contained in two other publications not included in the record. Id. at 131. The article states that five stone quality tests typically employed by states, including ASTM C 131, are “primarily methods for evaluating aggregate for concrete.” Id. at 132-33. The authors then discuss the stone quality criteria employed by the Corps, including ASTM C 535 with a maximum 20% loss. Id. at 135, 137, 139. The number of revolutions is not explicitly stated. The CO apparently assumed that 1,000 revolutions was intended. AR Tab 34 at 8 n. 10. This [230]*230assumption is incorrect, however, because the authors make clear that the Corps’ criteria were obtained from EM 1110-2-2302 which, as stated above, recommends ASTM C 535 at 500 revolutions. AR Tab 32 at 135, 137, 139. The article states that the tests recommended in EM 1110-2-2302 are “broader and more demanding of stone quality” than the tests employed by states, are “intended as expected properties of only the best stone,” and are “biased in application toward very large, high-cost stone materials such as for armor.” Id. at 138.
Although the CO’s report appears to have been filed with GAO, AR Tab 35, on August 7, 1997, Redland withdrew its protest, apparently in response to the Corps’ decision to set August 12,1997 as the due date for bids. Tr. at 104. On August 8, 1997, Redland filed a complaint in this court seeking a temporary restraining order and declaratory and permanent injunctive relief. During an informal conference call on August 8, 1997, counsel reported that the Corps had agreed to defer the due date for bids for a reasonable period pending resolution of this action. By order dated August 8,1997, defendant was directed to provide ten days advance notice if the Corps decided to proceed with bid opening. To date, defendant has not provided such notice and so it has not been necessary to rule on plaintiffs request for a preliminary injunction.
On August 15, 1997, defendant submitted an “administrative record” which was supplemented by agreement of the parties on August 25, 1997. Cross-motions for judgment on the administrative record were subsequently filed. The central issue addressed in the briefs is whether the record supports the Corps’ choice of abrasion tests.
Oral argument was held on October 6, 1997. The court asked if supplementing the record with the testimony of Mr. Lutton was a possibility since both parties rely on his work (and hearsay attributed to him) and since the court perceived a deficit in the record regarding the rationale for the Corps’ choice of abrasion tests. Tr. at 82, 107-13. The court also indicated it would consider supplementing the record with affidavits of other Corps personnel as well as testimony from one of plaintiffs experts. Tr. at 112-14. Proceedings were adjourned so that counsel could discuss the matter.
A status conference was held on October 8, 1997. Counsel for defendant indicated that defendant did not intend to present testimony from Mr. Lutton or anyone else. Counsel for plaintiff felt that an evidentiary hearing was necessary. It was decided that an order would be issued on the present submissions indicating whether further proceedings were necessary.
The court has jurisdiction over Redland’s objections to the solicitation pursuant to 28 U.S.C. § 1491(b)(1) and is authorized to grant an appropriate remedy, “including declaratory and injunctive relief,” if those objections prove well founded. Id. § 1491(b)(2).5
DISCUSSION
A. Standard of Review
Cross-motions for judgment on the administrative record are treated as motions for summary judgment under RCFC 56(a). Nickerson v. United States, 35 Fed.Cl. 581, 588 (1996), aff'd, 113 F.3d 1255 (Fed.Cir. 1997). Each motion must be evaluated on its own merits and will be granted only if there are no genuine issues of material fact and the movant is entitled to judgment as a matter of law. Mingus Constructors, Inc. v. United States, 812 F.2d 1387, 1390 (Fed.Cir.1987).
The Corps’ choice of abrasion test requirements will be upheld unless it is “ar[231]*231bitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A) (1994); 28 U.S.C.A. § 1491(b)(4) (West Supp.1997). The relevant law is the Competition in Contracting Act (“CICA”) which imposes a duty on the Corps to “obtain full and open competition.” 41 U.S.C. § 253(a)(1)(A) (1994); 10 U.S.C. § 2304(a)(1)(A) (1994). A consequence of this duty is that solicitation provisions which restrict competition may be used only “to the extent necessary to satisfy the needs of the agency or as authorized by law.” 48 C.F.R. § 11.002(a)(l)(ii) (1997); 41 U.S.C. § 253a(a)(2)(B); 10 U.S.C. § 2305(a)(l)(B)(ii). “Unnecessarily restrictive specifications or requirements that might unduly limit the number of bidders are prohibited.” 48 C.F.R. § 14.101(a). Both parties appear to agree that the abrasion test requirements chosen by the Corps restrict competition.6 The issue, then, is whether the restriction is reasonably necessary to meet the Corps’ needs with respect to the Poplar Island project.
In resolving this issue under the arbitrary and capricious standard, the court should not substitute its judgment for the agency’s. Cincom Systems, Inc. v. United States, 37 Fed.Cl. 663, 672 (1997). The agency’s decision is entitled to a “presumption of regularity,” Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 415, 91 5. Ct. 814, 823, 28 L.Ed.2d 136 (1971), although that presumption does not shield it from a “thorough, probing, in-depth review.” Id. While the court “recognize[s] the relevant agency’s technical expertise and experience, and defer[s] to its analysis unless it is without substantial basis in fact,” Federal Power Comm’n v. Florida Power & Light Co., 404 U.S. 453, 463, 92 S.Ct. 637, 644, 30 L.Ed.2d 600 (1972), the court must also perform an informed review of even technical decisions in order to meaningfully exercise its jurisdiction. Prineville Sawmill Co., Inc. v. United States, 859 F.2d 905, 910-11 (Fed.Cir.1988). Furthermore, “[ejxpertise is a rational process and a rational process implies expressed reasons for judgment.” Mid-State Fertilizer v. Exchange Nat’l Bank, 877 F.2d 1333,1339 (7th Cir.1989) (quoting Federal Power Comm’n v. Hope Natural Gas Co., 320 U.S. 591, 627, 64 S.Ct. 281, 299-300, 88 L.Ed. 333 (1944) (Frankfurter, J., dissenting)). The court “must ensure that the agency has ‘examined] the relevant data and articulat[ed] a satisfactory explanation for its action including a rational connection between the facts found and the choice made.’” Rainbow Navigation, Inc. v. Department of Navy, 783 F.2d 1072, 1080 (D.C.Cir.1986) (Scalia, J.) (quoting Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Automobile Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2866-67, 77 L.Ed.2d 443 (1983) (citation omitted)).
B. Merits
The Corps did not articulate a rational connection between its needs with respect to the Poplar Island project and its decision to use ASTM C 131 with a maximum 20% loss prior to issuing the construction specifications on June 6, 1997. The only relevant analyses conducted prior to that date are contained in the April 16 and April 25, 1997 memoranda. The Corps described in detail the site conditions at Poplar Island, compared the project to previous Chesapeake Bay armor stone projects, and explained why durable stone passing stringent quality tests recommended in EM 1110-2-2302 was needed. The April 16 memorandum concludes [232]*232that stone passing ASTM C 535 at 500 revolutions with a máximum 30% loss would satisfy the Corps’ needs. The April 25 memorandum either concludes that ASTM C 535 at 500 revolutions with a maximum 20% loss is appropriate, as the brunt of the analysis indicates, or, when considered in the light most favorable to defendant, it is internally inconsistent, concluding that ASTM C 131 is appropriate based on an analysis largely justifying the use of ASTM C 535 at 500 revolutions. Either way, neither it nor the April 16 memorandum provides a logical link between the Corps’ stated needs and its decision to depart from the guidance in EM 1110-2-2302 and require stone passing ASTM C 131 with a maximum 20% loss.7
Following the issuance of the specifications, the Corps made numerous efforts to retroactively justify its decision to use ASTM C 131. Defendant notes that during their July 3, 1997 phone conversation, Mr. Hart advised Mr. Saint-Clair that he agreed that abrasion tests “represent an index tool” and that the use of ASTM C 131 “with the MD SHA history of testing was sound.” AR Tab 44. However, neither the meaning of Mr. Hart’s understanding of abrasion tests as “an index tool” nor the basis for his conclusory statement is provided. Even assuming Mr. Hart is an expert on armor stone quality criteria, “[a]n expert who supplies nothing but a bottom line supplies nothing of value to the judicial process.” Mid-State, 877 F.2d at 1339; cf. J.F. Allen Co. v. United States, 25 Cl.Ct. 312, 325 (1992) (“Conclusory statements, without factual support, are not admissible and do not comply with rules of this court”).
On July 15, 1997 — one day after Red-land filed its protest arguing that ASTM C 535 at 500 revolutions was appropriate — the Corps changed the conclusion in § 4.2 and § 5 of the April 25 memorandum to state that ASTM C 131 was required for Poplar Island. The Corps also deleted the references in § 4.3 and § 5 to EM 1110-2-2302, which recommends ASTM C 535 at 500 revolutions. Since no additional analyses or information were added to justify these charges, the changes are arbitrary and do not retroactively rationalize the Corps’ choice of ASTM C 131.
Also on July 15, 1997, Mr. Saint-Clair telephoned Mr. Lutton who, according to Mr. Saint-Clair’s memorandum of the conversation, advised him that he felt the Corps “used sound engineering judgment using the 131 as just an index tool.” AR Tab 51. It is unclear if Mr. Lutton intended this statement as an endorsement of the Corps’ decision to exclude any quarry not capable of producing stone passing ASTM C 131 with a maximum 20% loss. It is doubtful that he did since in the same conversation he stated that ASTM C 535 “would be preferred” over ASTM C 131 and that abrasion tests “may be the most doubtful” of all stone quality tests. Id. Furthermore, his 1993 publication co-authored with Mr. Wong characterizes ASTM C 131 as primarily a method “for evaluating aggregate for concrete” but states that the criteria in EM 1110-2-2302, including ASTM C 535 at 500 revolutions with a maximum 20% loss, are “intended as expected properties of only the best stone” and “biased in application toward very large, high-cost stone materials such as for armor.” AR Tab 32 at 138. Furthermore, even if Mr. Lutton did intend to endorse the Corps’ use of ASTM C 131, his endorsement does not create a genuine issue of fact because the basis for it is not provided. Mid-State, 877 F.2d at 1339; J.F. Allen Co., 25 Cl.Ct. at 325.
[233]*233Mr. Saint-Clair also states that he asked Mr. Lutton about the disparity between his 1988 article, which references ASTM C 535 at 1,000 revolutions with a máximum 25% loss, and EM 1110-2-2302, which recommends 500 revolutions with a maximum 20% loss. According to Mr. Saint-Clair, Mr. Lut-ton responded that he “may” have relaxed the test for inclusion in EM 1110-2-2302 because it was “too severe,” and because producers in the north-central and mid-west-em United States could not meet it. AR Tab 51. These facts appear to be the sole basis for the Corps’ decision to add ASTM C 535 at 1,000 revolutions with a maximum 25% loss to the specifications. However, they are not indicative of a rational link between the conditions at Poplar Island and the need for the test; the record shows that the Baltimore District has consistently followed EM 1110-2-2302 and used ASTM C 535 at 500 revolutions for armor stone projects in the Chesapeake Bay, even though the Baltimore District is not located in the north-central or mid-western United States. AR Tabs 37-39.
It must be remembered that defendant was given an opportunity to present Mr. Lutton, who is currently working for the Corps, so that he could explain to the court and to plaintiff why the conditions at Poplar Island provide a reasonable basis for the Corps’ choice of abrasion tests. Defendant declined to take this opportunity. As they stand on the record, Mr. Lutton’s statements and the statements attributed to him simply do not provide a rationál basis for the Corps’ decision.
The Corps’ July 17, 1997 cost analysis, concluding that maintenance costs will increase by $9.6 million if ASTM C 131 is relaxed to permit a 40% loss, also does not retroactively justify the Corps’ choice of ASTM C 131 with a maximum 20% loss. As stated above, the premise of the analysis— that ASTM C 131 with a maximum 20% corresponds to the JVs determination that 4,000 tons of stone will be lost per year due to wave impact — is unfounded because the JVs wave impact calculations do not indicate that stone abrasion resistance was a variable and because the JV never referenced the ASTM C 131 abrasion test anywhere in its voluminous report. See Tr. at 32-36. Instead, the JV determined that ASTM C 535 at 500 revolutions with a maximum 20% loss was the appropriate abrasion test for the Poplar Island armor stone. AR Tab 86 at 02486-4. Since the premise of the July 17 analysis is unfounded, the conclusions therein are not reasonable.8
The CO’s reference to previous Corps projects in other districts, including the Huntington District’s Dry Run Levee project, which used ASTM C 131 with a maximum 30% loss, and the Norfolk District’s James R. Olin project, which used ASTM C 131 with a maximum 20% loss, as part of the Corps’ response to Redland’s GAO protest also does not retroactively justify the Corps’ decision to use ASTM C 131 for Poplar Island. The Dry Run project required “crushed limestone or dolomite” having “less than 5% passing a 2” sieve,” not large armor stone. AR Tab 30 at 02225-3, 6. In addition, the stone was to be used for “slope protection” along a river able to withstand “processing, transport, placement, and subsequent weathering” rather than wave impact. Id. at 02225-3, 6; Tr. at 81. The James R. Olin project involved rip-rap along a canal and along a levee or flood-wall near a river and there is no indication that one of the purposes of the project was to withstand wave attack. AR Tab 26 at 02200-2 — 6. Furthermore, defendant has repeatedly argued that riprap and armor stone projects are quite different. Def. Mot. at 5, 20; DSOF 111117-19; Tr. at 89-90; AR Tab 34 11114-5.
[234]*234In contrast, the Island Creek project is located in the Chesapeake Bay and required large armor stone between 1,700 and 3,200 pounds, similar to the stone required for Poplar Island. PCS & PAF Ex. D at 02486-4; AR Tab 1A at 02486-3. Likewise, the Corps determined — before any formal dispute with Redland arose — that Island Creek was “similar[ ] in exposure and design wave height” to Poplar Island. AR Tab 39 at 2. Yet Island Creek used ASTM C 535 at 500 revolutions with a maximum 20% loss, AR Tab 37, while the Corps chose ASTM C 131 for Poplar Island. Given the stated similarity between the two projects, some rational basis is needed to explain the Corps’ decision to apply different abrasion tests. Trans-active Corp. v. United States, 91 F.3d 232, 237 (D.C.Cir.1996) (“A long line of precedent has established that an agency action is arbitrary when the agency offered insufficient reasons for treating similar situations differently”). Dissimilar projects in other Corps districts certainly do not provide a reasonable basis for the disparate treatment.
Finally, defendant argues that the Corps’ decision to use ASTM C 131 cannot be arbitrary because it was based on established guidance in the SPM. DSOF 111157-58; Def. Mot. 9-10; Def. Reply at 9. However, the first document in the record referencing the SPM as justification for the use of ASTM C 131 is the CO’s August 8, 1997 report, written to defend against Redland’s protest over two months after the issuance of the specifications. There is simply no evidence in the record that the SPM was ever consulted by anyone in the Corps prior to Redland’s protest or that the single paragraph it contains on stone testing is even applicable to the Poplar Island project.
On the other hand, the record overwhelmingly demonstrates that EM 1110-22-302 contains engineering guidance which the Corps in fact consulted and which is valid with respect to the project. Defendant is correct that the recommendations in the manual are “broad generalizations still needing verification or adjustments for local experience” and “provide little more than first estimates of stone performance that may or may not prove valid within a region.” AR Tab 21 at 6-1, 7. However, the Baltimore District decided that the guidance in the manual was appropriate in the Chesapeake Bay region and applied it to all of the armor stone projects discussed in the April 16, April 25, and July 15, 1997 memoranda. In addition, all three memoranda repeatedly state that the Corps’ choice of stone quality tests for Poplar Island was based on the recommendations in EM 1110-2-2302; the SPM is not referenced in any of these documents. The Corps also referenced and followed the recommendations in EM 11 10-2-2302 in preparing the Poplar Island test dike stone quality specifications. AR Tab 60 at 02486-1- —3. Likewise, the JV was aware of the SPM and cited it in its report, e.g., AR Tab 53 at 2-14, R-3, but referenced and followed the guidance in EM 1110-2-2302 in its stone quality specifications. AR Tab 86 at 02486-2— 3. The final Poplar Island specifications themselves reference and follow the recommendations in EM 1110-2-2302 — except, of course, with respect to the abrasion tests. AR Tab 1A at 02486-1, 4. In contrast, there is no evidence in the record that the Baltimore District has ever departed from EM 1110-2-2302 and used the ASTM C 131 abrasion test for armor stone projects in the Chesapeake Bay (or any other project), and there is no evidence that any Corps district has ever used the ASTM C 535 test at 1,000 revolutions with a maximum 25% loss.
The Corps is certainly entitled to depart from previous practice and the nonbinding engineering guidance 'contained in EM 1110— 2-2302, but it must provide a rational basis for doing so. Motor Vehicle, 463 U.S. at 57, 103 S.Ct. at 2874 (“ ‘an agency changing its course must supply a reasoned analysis’”) (quoting Greater Boston Television Corp. v. FCC, 444 F.2d 841, 852 (D.C.Cir.1970), cert, denied, 403 U.S. 923, 91 S.Ct. 2233, 29 L.Ed.2d 701 (1971)). This is especially true in this case since the departure restricts competition and may result in at least several million dollars of increased costs. AR Tab 38; AR Tab 7 Ex. 1 at 2. Since the record does not reveal a rational connection between the facts associated with the Poplar Island project and the Corps’ choice of abrasion tests, the departure in this case is invalid. Plaintiff is entitled to a declaration that the [235]*235abrasion test requirements included in the specifications lack a rational basis and that, as a result, the solicitation unduly restricts competition in violation of CICA.
Plaintiff also asks the court to direct the Corps to add its Texas quarry to the approved source list. In light of the finding that the abrasion tests included in the solicitation are invalid, this request is not ripe. Whether the Texas quarry should be approved depends, at least in part, on what valid abrasion test the Corps includes in the specifications, if any.
In addition, Mr. Saint-Clair characterizes the stockpiled stone at the Texas quarry as “a rubble pile or stone trash pile.” AR Tab 24 at 1. Among other things, he observed “[e]ompositional banding, seams, and structural planes of weaknesses, cracking and spalling.” Id, at 2. These observations suggest it may be perfectly appropriate for the Corps to conduct additional tests or even to reject the stockpiled stone at the Texas quarry regardless of what abrasion test is chosen: the Corps reserves the right in the solicitation to conduct more tests if “the acceptability of the stone from the section of the quarry proposed for this work is questionable on the basis of geologic examination by the [CO]’s representative,” AR Tab 1A at 02486-6, and may “reject certain localized areas, strata, or channels within the approved source when in the opinion of the [CO], the stone is disintegrated, badly weathered, contains incipient planes of weakness or hidden joints/fractures, or is otherwise unsatisfactory for use in the work.” Id. at 02486-7. In short, the observation that the stockpile at the Texas quarry might be a “stone trash pile” does not provide the rational basis required for the Corps’ choice of abrasion tests which exclude not only the Texas quarry, but six additional quarries. Conversely, though, it does not follow that the Corps must approve the Texas quarry from the fact that its choice of abrasion tests was arbitrary.
Furthermore, plaintiff is not necessarily harmed if the Texas quarry is not added to the list because the solicitation permits the use of unlisted quarries, subject to the CO’s approval. Id. at 02486-4 — 6. If Redland’s stone meets all of the final quality criteria included in the solicitation, there will be no basis for excluding it from the project regardless of whether the quarry is on the approved source list. For these reasons, plaintiffs request that the judgment direct the Corps to add the Texas quarry to the approved source list is denied.
Plaintiff’s request that the court enjoin the Corps from advertising or receiving bids pursuant to the solicitation is also denied. Athough the abrasion test requirements lack validity, there is no need to enjoin the procurement: the Corps can either proceed on the basis of the current solicitation without the arbitrary abrasion tests or it can continue to postpone bid opening pending the addition of a test having a rational basis in light of CICA requirements.
CONCLUSION
Since there are no genuine issues of material fact, further proceedings in this matter are unnecessary. For the reasons stated above, plaintiff’s motion for judgment on the administrative record is granted insofar as it seeks a declaration that, based on the record before the court, the abrasion test requirements contained in the solicitation are invalid and not in compliance with CICA Plaintiffs motion is otherwise denied. Defendant’s motion is denied.
Accordingly, it is hereby ORDERED that:
(1) Final Judgment be entered declaring the abrasion test requirements in § 2.1.5 of solicitation no. DACW31-97-B-0022 null and void as lacking a rational basis and contrary to CICA requirements; and
(2) Except as GRANTED in (1), all other relief sought in this matter is DENIED. No costs.