Global SuperTanker Services, LLC (GST), a small business located in Colorado Springs, Colorado, protests the terms of request for proposals (RFP) No. AG-024B-S-16-9014, issued by the Department of Agriculture (USDA), Forest Service, for Next Generation Large Airtanker (LAT) services for wildland firefighting. The protester alleges that the solicitation's limitation on the maximum tank size for retardant is unduly restrictive of competition.
We sustain the protest.
DOCUMENT FOR PUBLIC RELEASE
The decision issued on the date below was subject to a GAO Protective Order. This redacted version has been approved for public release.
Matter of: Global SuperTanker Services, LLC
File: B-414987; B-414987.2
Date: November 6, 2017
Robert K. Magovern, Esq., Marc J. Fink, Esq., and Jeff R. Vogel, Esq., Cozen O'Connor, for the protester.
Antonio Robinson, Esq., Department of Agriculture, for the agency.
Elizabeth Witwer, Esq., and Jennifer D. Westfall-McGrail, Esq., Office of the General Counsel, GAO, participated in the preparation of the decision.
Protest challenging the solicitation's limitation on maximum tank size as unduly restrictive of competition is sustained where the record fails to show that the limitation is reasonably necessary to meet the agency's needs.
Global SuperTanker Services, LLC (GST), a small business located in Colorado Springs, Colorado, protests the terms of request for proposals (RFP) No. AG-024B-S-16-9014, issued by the Department of Agriculture (USDA), Forest Service, for Next Generation Large Airtanker (LAT) services for wildland firefighting. The protester alleges that the solicitation's limitation on the maximum tank size for retardant is unduly restrictive of competition.
We sustain the protest.
The RFP was issued on May 16, 2017, as a set-aside for small businesses. RFP at 6, § B.1. The solicitation contemplated the award of multiple Call When Needed (CWN) basic ordering agreements (BOAs) to obtain Next Generation LAT services for wildland firefighting on a nationwide basis. RFP, § B.1. The solicitation provided that the services were intended to deliver large volumes of retardant to support actions on emerging wildland fires in the "initial" and "extended" attack phases for large fire support. RFP, § C.1(a).
The solicitation provided that the airtankers were "required to have a minimum tank capacity of 3,000 gallons that will deliver long term retardant to active wildland fires." RFP, § C.1(a). See also, RFP, §§ B.3, C.1(b), L.6(b)(3)(ii). This requirement pertaining to minimum tank capacity is consistent with prior solicitations issued by the agency. See e.g., 10 Tanker Statement, Aug. 28, 2017, Exh. 5, Solicitation No. AG-024B-S-015-9002, Feb. 19, 2015, § M-2(a)(1)(iv).
In a footnote, the solicitation also included a new requirement, not previously included in solicitations issued by the agency, pertaining to the maximum tank capacity:
The minimum required volume is 3000 gallons (dispensable) and 27,000 pounds of payload. The maximum allowed volume is 5000 gallons (dispensable) and 45,000 pounds of payload. . . . Aircraft with less than 3000-gallon dispensing capacity or greater than 5000-gallon dispensing capacity will not be considered.
RFP, § B.3, n.1 (emphasis added).
By way of background, GST is the owner and operator of The Spirit of John Muir, which the protester represents to be the world's newest and largest firefighting airtanker. Protest at 2. The B747-400 series aircraft is a Next Generation VLAT and can carry up to 19,200 gallons of retardant. Id.; Decl. of GST President & CEO, July 27, 2017, ¶ 4. GST contends that its aircraft can meet all of the technical requirements in the RFP except for the 5,000-gallon limitation on tank size. Decl. of GST President & CEO at ¶ 4.
GST does not currently provide aerial firefighting services for the Forest Service. However, GST alleges that its VLAT has recently deployed successfully to Israel and Chile to conduct the same type of operations being procured here. Protest at 5; id., Attach. 2, Agency-Level Protest, at 5; Decl. of GST President & CEO at ¶ 7. By contrast, GST's VLAT competitor, 10 Tanker, has provided such services to the
Forest Service for almost a decade. 10 Tanker Statement, Aug. 28, 2017, at 4. According to GST, GST and 10 Tanker are the only two companies that possess aircraft with tanks larger than 5,000 gallons. Id. at 2; Supp. Protest, Sept. 7, 2017, at 3.
On May 24, prior to the date set in the RFP for the submission of questions, GST sent a letter to the agency requesting an explanation of the agency's basis for including the 5,000-gallon maximum restriction. Protest, Attach. 2, Exh. 1, GST RFP Questions. GST did not receive a response. Protest at 3.
On June 13, GST filed a timely pre-award protest with the agency. The agency denied the protest on July 18. In its denial, the agency stated that the solicitation sought services to support initial attack operations, and that VLATs are "not ideal for initial attack purposes." AR, Tab 7, Denial of Agency-Level Protest, at 1. For support, the agency cited four studies performed by the Forest Service in 1995, 1996, 2005, and 2012, respectively. Id. at 2 n.1. In its decision denying the agency-level protest, the Forest Service recognized the need for VLATs in wildland firefighting, however, and informed GST that the agency would be issuing a "call when needed" solicitation for VLATs sometime in the beginning of fiscal year 2018. Id. at 1.
This protest followed on July 27.
The protester challenges the 5,000-gallon maximum restriction as unduly restrictive of competition. GST asserts that, because it and 10 Tanker are the only two companies whose aircraft have tanks larger than 5,000 gallons, the restriction is knowingly targeted at excluding their aircraft. The protester also argues that the agency's stated bases for including the restriction do not withstand logical scrutiny and that, as a result, the agency has failed to meet its burden to demonstrate that the 5,000-gallon maximum restriction is reasonably necessary to meet the agency's needs. Based on the record, we are unable to conclude that the specifications included in the RFP are necessary to meet the agency's needs.
Before addressing the specific contentions raised in this protest, we note that certain unique features of this procurement factor into our conclusions. Critically, under a "call when needed" BOA, the agency is not required to place any orders and incurs no costs for days of inactivity, thus providing the agency the ultimate flexibility to tailor its orders, on a case-by-case basis, to precisely the aviation assets it requires for each individual fire. Similarly, a company that receives a "call when needed" agreement is under no obligation to accept any particular orders. In other words, it costs the Forest Service nothing to permit the inclusion of VLATs in the subject procurement, while simultaneously increasing the agency's available options during fire season. As a result, for the reasons set forth below, we conclude that, in the context of a "call when needed" agreement, the agency's decision to exclude VLATs from the competition altogether is not reasonable.
The determination of the government's needs and the best method of accommodating them is primarily the responsibility of the procuring agency, since its contracting officials are most familiar with the conditions under which supplies, equipment, and services have been employed in the past and will be utilized in the future. Pitney Bowes, Inc., B-413876.2, Feb. 13, 2017, 2017 CPD ¶ 56 at 3. In preparing a solicitation, a procuring agency is required to specify its needs in a manner designed to achieve full and open competition, and may include restrictive requirements only to the extent they are necessary to satisfy its legitimate needs. Parcel 49C Ltd. P'ship, B-412552 et al., Mar. 23, 2016, 2016 CPD ¶ 95 at 11. In this respect, solicitations should be written in as non-restrictive a manner as possible in order to enhance competition.
To the extent a protester challenges a specification as unduly restrictive, the procuring agency has the burden to establish that the specification is reasonably necessary to meet its needs. Pitney Bowes, Inc., supra; Smith and Nephew, Inc., B-410453, Jan. 2, 2015, 2015 CPD ¶ 90 at 5. We review the agency's explanation for reasonableness, that is, whether it can withstand logical scrutiny. Pitney Bowes, Inc., supra; Parcel 49C Ltd. P'ship, supra.
Here, the agency contends that the solicitation is intended to procure services to support initial attack operations and that VLATs are not suitable for such operations. COS at ¶ 12. For support, the agency relies upon the findings contained in previously referenced studies from 1995, 1996, 2005, and 2012. Id. at ¶¶ 12, 17; AR, Tab 7, Denial of Agency-Level Protest, at 2 n.1. Additionally, the agency explains that VLATs are not suitable for initial attack operations for the following reasons:
- There are limited bases from which VLATs can operate, which hinders their ability to conduct initial attack operations.
- VLATs require an aerial supervision module/lead plane (ASM/LP); the supply of such planes is limited, which could adversely impact a VLAT's ability to conduct initial attack operations.
- Additional government personnel are needed to staff a VLAT.
- VLATs require more fuel and retardant, which can be limited at some bases.
- VLATs require extensive time to fuel, which interferes with the ability of other aircraft to operate and adversely impacts the agency's ability to conduct initial attack operations.
- A VLAT damaged a hanger while taxiing at a base, and, at another base, a VLAT's weight partially damaged the airport ramp.
COS at 2-3; Memorandum of Law (MOL), Aug. 28, 2017, at 3; AR, Tab 7, Denial of Agency-Level Protest. The agency further contends that the solicitation cannot be considered unduly restrictive of competition because the agency intends to issue a subsequent "call when needed" solicitation that will be open to VLATs. MOL at 11; AR, Tab 7, Denial of Agency-Level Protest. Based on these reasons, the agency argues that its decision to limit the maximum capacity of retardant tank size to exclude VLATs is reasonable.
In response, GST contends that the RFP did not specify that the operations were limited to initial attack operations. In any event, GST contends that the agency's claims with respect to the suitability of VLATs to perform initial attack operations are not supported by the record. GST also argues that the agency failed to consider a number of relevant factors, including the benefits offered by a larger aircraft.
In support of its arguments, the agency has provided statements from the contracting officer and the Branch Chief of Aviation Operations, who generally assert that VLATs are not suited to initial attack operations. Notwithstanding these general assertions, we conclude that the majority of the agency's arguments are based upon anecdotal evidence unsupported by the record. We address the agency's arguments in the discussion that follows.
The RFP Contemplates Both Initial and Extend Attack Operations
In its denial of the protester's agency-level protest, the crux of the agency's argument was that the solicitation is intended to support initial attack operations and that, for various reasons, VLATs are not suited for such operations. AR, Tab 7, Denial of Agency-Level Protest, at 1. See also COS at ¶¶ 12, 15, 17 (focusing exclusively on the unsuitability of VLATs for initial attack operations). As GST correctly notes, however, the RFP seeks to procure aerial firefighting services to support both initial and extended attack operations. Protest at 5 (citing RFP, § C.1(a)). See also, AR, Tab 5, Acquisition Plan, at 3 (reiterating the agency's intent to procure services for both phases of attack). GST argues that, because the RFP seeks to procure services for both initial and extended attack operations, the contention that LATs are better suited than VLATs for initial attack operations does not provide adequate justification for excluding VLATs from this competition for a "call when needed" BOA.
In the agency's legal memorandum, the Forest Service concedes that the RFP contemplates "both initial attack and extended phases of the firefighting effort." MOL at 4; id. at 5. The agency argues, however, that it excluded VLATs on the basis that they do not meet the agency's needs for both phases of attack. Id. at 4. Alternatively, the Forest Service argues that, even if the record does not demonstrate that the agency reached this conclusion with respect to extended attack operations, the agency could have reached this conclusion. Supp. MOL at 5 (Although "[t]he agency-level protest may have focused primarily on initial attacks,  that does not mean its arguments do not hold true for a solicitation that contemplates airtankers in support of initial and extended attack.").
The record furnishes no support for the agency's arguments. Here, agency officials expressly indicated that the 5,000-gallon limitation was based upon the conclusion that VLATs were not suited to perform initial attack operations, omitting any discussion of extended attack operations. As a result, the agency's argument represents a post hoc attempt to justify the 5,000-gallon restriction.
Given the nature of the "call when needed" agreement the agency contemplates here, we agree with GST that the conclusion that VLATs are not suited for initial attack operations does not provide a reasonable justification for excluding VLATs from the competition where the solicitation also seeks services for extended attack operations. More importantly, as we explain below, there is also no support for the agency's contention that VLATs are not suited for performing initial attack operations.
Origin of Requirement
As an initial matter, we note that, despite multiple document productions, there is nothing in the record showing that the 5,000-gallon restriction on maximum tank size was ever discussed, considered, or recommended by any agency official in the context of this solicitation. Indeed, the record is completely silent regarding who, if anyone, at the agency made the decision to include the restriction, when the decision was made, and why the decision was made.
Notably, none of the pre-solicitation documents contain any reference to a 5,000-gallon maximum restriction. For example, although the agency's source selection plan, acquisition plan, and request for information (RFI) discuss the 3,000-gallon minimum restriction, these documents do not mention any maximum restriction or the rationale for such a restriction. See generally, AR, Tab 4, Source Selection Plan; Tab 5, Acquisition Plan; Tab (unnumbered), RFI. Moreover, the agency has failed to produce any emails, memoranda, reports, analyses, or any other communications in which it addresses the decision to include a maximum restriction in this solicitation.
Additionally, GST contends that it discussed the subject requirement with the agency and that the agency led GST to believe that it would be permitted to compete for the contract. Protest at 2; Decl. of GST President & CEO at ¶ 5. GST also responded to the agency's RFI, noting that its aircraft had a tank larger than 19,000 gallons. AR, Tab (unnumbered), GST Resp. to RFI, at 2. GST contends that the agency did not inform GST that it would be ineligible to compete. Supp. Comments, Sept. 25, 2017, at 2 n.3.
Not only is there no explanation in the contemporaneous record supporting the limitation on the maximum tank size, but, as we explain below, the limitation is also contrary to the agency's well-established, longstanding practice of allowing VLATs to compete for contracts having the primary purpose of initial attack.
Historical Use of VLATs
The record reflects that the agency's first "call when needed" agreements for airtanker services were awarded in 2009 and included initial attack operations. AR, Tab 5, Acquisition Plan, at 3. These agreements were reserved exclusively for VLATs and contained a minimum tank capacity restriction of 8,000 gallons. Id. Moreover, the agency acknowledged that, since 2009, it has never issued a solicitation containing a restriction on the maximum tank capacity. Agency Email, Sept. 15, 2017.
The record further reflects that 10 Tanker, with aircraft certified at 11,600 gallons, has recently been awarded three airtanker services contracts in which it performs initial attack operations with VLATs. 10 Tanker Statement, Aug. 28, 2017, at 2; Decl. of 10 Tanker President at ¶ 9 ("Throughout 10 Tanker's performance, it has conducted initial attack operations for the Forest Service."); id. ("10 Tanker is regularly part of the first wave of firefighting resources performing  missions on the first day of the effort to engage the fire."). In fact, 10 Tanker was recently awarded a "call when needed" agreement for the exact same services being procured here. 10 Tanker Statement, Aug. 28, 2017, at 2. The agency does not dispute this information. Thus, the record shows that, until very recently, the agency considered 10 Tanker's VLATs to be capable of performing initial attack operations.
Moreover, information submitted by GST shows that 10 Tanker has seen a steady increase in the volume of missions it has been called upon to fly since 2012, from 264 in 2012 to 512 in 2016. Decl. of 10 Tanker President at ¶ 5. Thus far, in 2017, 10 Tanker has been called upon to perform 700 missions. Id. at ¶ 6. Accordingly, not only has the agency not previously favored aircraft with tank capacities under 5,000 gallons, but it has, in fact, used aircraft with larger tank capacities for initial attack operations on an ever-increasing basis.
Furthermore, the record here shows that the agency has extended both the annual duration and potential frequency of services of all three of 10 Tanker's contracts in order to use each of 10 Tanker's VLATs more often and to retain the VLATs on call for increasingly longer periods of time. 10 Tanker Statement, Aug. 28, 2017, at 3. In 2017, at the agency's request, 10 Tanker agreed to modify all of its contracts to allow the Forest Service to order 10 Tanker's VLATs daily, id., thus demonstrating a current need for these capabilities.
In response, the agency, for the most part, does not dispute these claims. Instead, the agency contends that what it has done in the past is irrelevant. MOL at 11. In this respect, the agency argues, "[a]lthough an agency has procured something one way under a previous solicitation does not mean that the current solicitation, after an agency's careful analysis of past experience, has to be identical." Id.
Although we agree with the Forest Service that an agency may elect to deviate from its prior practices, the agency's basis for its requirements must be reasonable where such requirements allegedly restrict competition. See e.g., WingGate Travel, Inc., et al., B-405007.9, Nov. 29, 2011, 2011 CPD ¶ 260 at 4 (finding that an agency's deviation from former practices was reasonable where the agency demonstrated that it had access to new data that supported those changes). Here, the Forest Service primarily relies upon four studies from 1995, 1996, 2005, and 2012 to justify its position. COS at ¶¶ 12, 17; AR, Tab 7, Denial of Agency-Level Protest, at 1, 2. As explained below, we find that these studies do not support the agency's position.
Agency's Reliance on Four Studies
Based upon our review, we find that there are several problems with the agency's reliance upon the cited studies. First, the studies do not support the agency's conclusion that VLATs are unsuited for initial attack operations. Second, as the protester points out, other independent wildfire experts have questioned the completeness and the accuracy of these studies. Finally, the agency has not reasonably explained why it is now introducing a limitation on maximum tank size based upon studies performed in 2012 and earlier.
The Cited Studies Do Not Support Exclusion of VLATs
Although the agency initially cited to the studies from 1995, 1996, and 2005 as a basis for its restriction, it did not identify any language in these studies to support the restriction at issue here. See AR, Tab 7, Denial of Agency-Level Protest, at 1-2; COS at ¶ 12, MOL at 6-7. Further, in answer to GST's assertion that the agency abandoned its argument that the 1995, 1996 and 2005 studies provide a basis for the restriction, see Comments, Sept. 11, 2017, at 6, the agency identified one page in the 1996 study and one page in the 2005 study to support its position. Supp. MOL at 2 (citing AR, Tab 21, 1996 Study, at 84; Tab 22, 2005 Study, at 4).
The cited pages do not lend support to the agency's position. As an example, the page in the 2005 study merely indicates that the agency prefers larger aircraft over smaller aircraft, not that VLATs are somehow less desirable for initial attack operations. AR, Tab 22, 2005 Study, at iii ("[A]irtanker platforms of 3,000 to 5,000 gallons continue to show significantly greater economic benefit over smaller capacity platforms.").
The agency appears to pin its defense of the limitation on maximum tank size on the 2012 study. See MOL at 4-6; Supp. MOL at 12-13. That study, however, does not recommend that the agency include a 5,000-gallon maximum restriction in future solicitations. See Protest, Attach. 2, Exh. 6, 2012 Study. Indeed, the 2012 study could be construed to support the protester's arguments.
In this respect, the study recommends that the wildland firefighting aircraft fleet be composed of a mix of aircraft, including "Very Large Airtankers (>8000 gallons)." Id. at 3. In discussing tank sizes, the study recommends a minimum capacity, not a maximum capacity, and reflects a preference for larger retardant tank capacities.
Specifically, the study recommends as follows:
Minimum capacity should be at least 2000 gallons of retardant, 3000 gallons or more would be preferred. For the specific aerial firefighting missions in dense forests, the need for canopy penetration is particularly acute and larger loads and higher speeds are preferred.
Id. at 7. Additionally, the study states:
The effectiveness of different airtanker types is often dependent on the fuels being treated. It can take a large quantity of retardant to penetrate dense forest canopies and large airtankers are more capable of being effective in thick forest canopies and areas with dense brush. The larger load capacity also allows large airtankers to split their loads to provide critically needed support on different parts of a fire without delay.
Id. at 5. In short, the study does not recommend a particular cap on the tank capacity, nor does it support the conclusion that VLATs are not suited for initial attack operations. This is because, rather than analyzing whether VLATs are unsuited for initial attack operations, the study focuses on the suitability of LATs for such operations. These are different questions. The Forest Service has not identified any study or analysis, upon which it relied to develop the RFP requirements, that has considered the question presented here: whether VLATs are unsuited for initial attack operations. In sum, the studies relied upon by the agency do not provide a reasonable basis to restrict competition.
Other Studies Raise Concerns
As noted above, the protester also points out that numerous other independent wildfire experts have questioned the completeness and the accuracy of the studies relied upon by the agency. Protest at 6. We discuss two examples here.
First, GST submits an article from Fire Aviation, dated April 2017, entitled "Does Anyone Know How Many Large Air Tankers We Need?" in which the Forest Services' National Fire Director was interviewed. Id. (citing Protest, Attach. 2, Exh. 4, Fire Aviation Article. The article specifically discusses the 1996 study relied upon by the agency and argues that "[m]uch has changed in the world of aerial firefighting" in the 21 years since the study was written[.]" Id. at 2-3. Moreover, the article maintains that "[t]here has not been in the United States a thorough, well designed analysis of the effectiveness of aerial attack, exactly how much retardant is needed in a certain time frame, where aircraft need to be based, and how many and what type of aircraft are required." Id. at 3. The article further notes that, in 2012, under pressure from Congress and GAO, the Forest Service "began a program to develop metrics and collect data to document and quantify the effectiveness of aircraft in assisting firefighters on the ground[,]" but that it will be several more years before the Forest Service releases its findings. Id at 3-4.
The referenced program is called the Aerial Firefighting Use and Effectiveness (AFUE) program. See https://www.fs.fed.us/managing-land/fire/aviation/afue (last visited Nov. 3, 2017). The Forest Service's website states that the program was "initiated in 2012 to answer a general, but complex question: 'What are the best mixes of aircraft to do any fire suppression job?'" Id. The website further states that "[d]ata collected from 2012 through 2014 is not sufficient to provide statistically defensible analysis and results" but that the agency plans to begin releasing aircraft use summaries in 2017. Id. Neither the contemporaneous record nor the agency's protest filings make reference to the agency's own internal program, which is designed to collect data directly relevant to the agency's decision in this matter.
This lack of data is echoed in a second study/report identified by the protester--an August 2013 report from GAO. See Wildland Fire Management: Improvements Needed in Information, Collaboration, and Planning to Enhance Federal Fire Aviation Program Success, GAO-13-684 (Aug. 2013). In the report, GAO found that the Forest Service and the Department of Interior have produced nine major studies and strategy documents related to their firefighting aviation needs, including the four studies cited by the agency in the instant protest. GAO Report at 11-12. After reviewing these studies, GAO determined that the studies did not consider all relevant factors. Id. at 12. GAO further determined that the Forest Service's efforts to identify the number and type of aircraft it needs have been hampered by limited information and collaboration. Id. at 11. Moreover, GAO found that the Forest Service has collected limited data on LATs, but not on any other aircraft.
In general, the agency does not address these arguments other than to contend that the agency cannot be faulted for failing to conduct more studies to evaluate the effectiveness of VLATs. MOL at 6. The agency misses the point. The role of GAO's bid protest function is limited to assessing whether the agency's justification for including a particular solicitation term is reasonable; our Office does not opine on whether the agency should conduct more studies or recommend methodologies for such studies.
Here, our discussion of study data (or lack thereof) is included because the agency itself offered study data in an attempt to support its decision to exclude VLATs from the competition. Accordingly, we merely find that, although the agency has reached conclusions regarding the technical limitations of VLATs, and is excluding VLATs from competition based upon such conclusions, the record does not demonstrate that the offered studies support the agency's conclusions. For this reason, we are unable to find that the agency's asserted justification for the exclusion of VLATs is reasonable. NCS Techs., Inc., B-403435, Nov. 8, 2010, 2010 CPD ¶ 281 at 4 (agency's justification for restriction is not adequate where the justification consists of unsubstantiated factual assertions and conclusory statements); Total Health Resources, B-403209, Oct. 4, 2010, 2010 CPD ¶ 226 at 3 (same); Navajo Nation Oil & Gas Co., B-261329, Sept. 14, 1995, 95-2 CPD ¶ 133 at 6-7 (same).
The Agency Did Not Implement the Studies
Finally, the protester points out that the agency continued to use VLATs for initial attack operations years, even decades, after the publication of the studies. Therefore, the protester takes issue with the agency's explanation for now introducing a restriction on maximum tank size based on the cited studies.
The agency explains that, for 20 years, the 5,000 gallon restriction has been a "goal range" and that the agency has contemplated including the restriction in prior solicitations. Supp. MOL at 2. According to the agency, it has not been able to do so previously because there have not been sufficient LAT assets, but there are sufficient LATs available now. Id.; COS at ¶ 14; Decl. of Branch Chief of Aviation Operations, Sept. 11, 2017, at ¶ 10.
In response, GST contends that the agency's claim regarding current LAT availability is not true. GST presents data showing that there has been a 10-15 percent shortfall in available inventory over the last two years, and that the agency has consistently been unable to fill large percentages of requests for aerial firefighting services from 2002-2016, including a high of almost 50 percent of unfilled orders in 2012. Protest, Attach. 2, Agency-Level Protest, at 4. GST also argues that a number of aircraft in the Forest Service's fleet are "phantom aircraft," meaning that they are not operational. Protest at 9.
GST also presents evidence from 10 Tanker, explaining that, in 2017, the agency modified all three of 10 Tanker's contracts to permit the Forest Service to use its aircraft daily. 10 Tanker Statement, Aug. 28, 2017, at 8; Decl. of 10 Tanker President at ¶ 7. Critically, 10 Tanker is the only contractor whose contracts have been so modified this fire season. Id. The protester argues that the agency's reliance upon 10 Tanker's VLATs in this manner demonstrates that the agency needs and wants such capabilities and undercuts the agency's argument that it has enough LATs to fulfill its needs.
Generally speaking, the agency does not dispute this information. Importantly, neither the contemporaneous record nor the agency's protest filings provide any explanation or analysis of the number of LATs and the ability of these LATs to adequately respond to the agency's aerial firefighting needs. Accordingly, we find the agency's explanation for its current requirement to be unsupported.
Other Bases to Support the Exclusion of VLATs
As noted above, in addition to its reliance upon the four studies, the agency also lists several other reasons why, in its opinion, VLATs are not suited to initial attack operations. Although we do not discuss each basis, or even each argument to support each basis, we have thoroughly considered all of the agency's arguments and conclude that they do not provide adequate justification for excluding VLATs. We address a few representative arguments below.
Limited Number of Bases
The agency contends that, because of their larger size, VLATs can operate only out of "a small number" of bases around the country. COS at ¶ 12. Specifically, the agency represents that there are only three existing bases where VLATs can be loaded without restricting the operations of LATs. Id. at ¶¶ 12, 16; MOL at 10. According to the agency, this "severely hinders" the effectiveness of VLATs because they cannot be prepositioned close to known fire activity. COS at ¶ 12. It also increases the cost to the agency because the agency is required to pay for additional flight time necessary for the VLATs to reach the fire activity. Id.
GST disputes the agency's claims. GST contends it can operate from 16 bases and identifies those bases by name. Comments at 3. GST argues that its contention is supported by the National Wildfire Coordinating Group Interagency Airtanker Base Directory, published in June 2017. Id.
The agency responds unequivocally that GST's contention "isn't true." Decl. of Branch Chief of Aviation Operations at ¶ 5. In the very next sentence, however, the agency concedes that it does not possess any knowledge about the capacity of GST's aircraft to operate from the bases identified. Id. ("DC-10s can be loaded at those bases, but we don't know about the 747."). The agency explains that the directory only shows what bases are VLAT-capable and that, unless GST has verified that its 747 aircraft can actually operate out of a base, GST "can't be sure." Id.
In response, GST provides a detailed explanation of how it determines which bases can support its VLAT. Supp. Comments at 8-9. In summary, it consults the directory; evaluates published airfield strength data and confirms operating weights against each airfield's maximum weight limits; verifies this data with available Federal Aviation Administration data; contacts the individual airfield managers for each base to validate the feasibility of GST's aircraft operations at that particular airfield; verifies taxiway, runway, parking ramp dimensions, aircraft classification number/pavement classification number (ACN/PCN) strength, the availability of water, and location of existing Retardant Loading facilities; evaluates where it could operate portable retardant plants if required; and provides a list of necessary equipment to the airfield managers to verify that equipment is on location and in good working condition. Id. at 9. GST contends that it has verified with each of the 16 designated VLAT bases listed that it can operate out of each base. Id.
Furthermore, GST explains that the 16 identified bases are not the maximum number of bases from which GST could operate. If, for example, the operation in question is a water operation, GST alleges that its aircraft can operate out of any of 44 major airports located in every region of the United States. Id.; Protest at 7. GST also discusses the option of setting up portable bases at any airport in less than 24 hours. Protest at 7; Decl. of GST President & CEO, July 27, 2017, at ¶ 9. If fixed and portable retardant options are considered, GST contends that it can operate out of 29 bases and has contacted the airfield managers to confirm this statement. Supp. Comments at 9.
Based upon the record before us, we conclude that the agency has not reasonably demonstrated that VLATs can only operate out of a small number of bases.
Requirement for Lead Planes
The agency also contends that VLATs are not suited for initial attack operations because they require a lead plane, the availability of which is limited and, therefore, could adversely impact a VLAT's ability to conduct initial attack operations. COS at ¶ 12. GST acknowledges the current need for lead planes, but notes that lead planes are necessary for LATs as well. Protest at 8; Comments at 5. Thus, if lead planes are indeed in short supply, such a shortage would impact both LATs and VLATs equally. Id. GST also argues that, in large wildfires, a lead plane is used for all aircraft and is not purely dedicated to VLATs. Decl. of GST President & CEO at ¶ 18.
In this respect, GST explains that the need for a lead plane is not based upon the type of aircraft, but is based, instead, upon pilot qualifications. Comments at 5. If the pilot possesses the necessary qualifications and/or certifications, the pilot and, as a result his or her aircraft, are deemed initial attack (IA) rated and do not require the assistance of a lead plane. GST contends that the only reason its pilots are not IA rated by the Forest Service is that GST does not currently possess a Forest Service contract. Supp. Comments at 11. If awarded a contract, GST plans to undergo the required pilot testing. Id.
The agency concedes that the requirement for a lead plane is contingent upon pilot qualifications, not upon the type of aircraft. Agency Statement, Oct. 24, 2017, at 2. The agency explains that, once a pilot completes the required training and examination, he or she is determined to be IA rated. Id. The agency asserts that all of its contract LAT pilots are IA rated and, therefore, do not require a lead plane. Supp. MOL at 11. The agency further asserts that, by policy, this same standard is not applied to VLAT pilots. Agency Statement, Oct. 24, 2017, at 2. Accordingly, lead planes are required for all VLATs. Id.
For support, the agency cites page 42 of its Interagency Aerial Supervision Guide (AR, Tab 9). This page supports the agency's argument that lead planes are required for all VLATs. However, it also shows that lead planes are required for LATs in certain circumstances, even if the pilot is IA rated. Accordingly, the benefits of using a LAT versus a VLAT do not appear as pronounced as the agency contends with respect to the need for a lead plane. In fact, in circumstances where a lead plane is required for LATs, such as a large wildfire, GST contends that the use of VLATs would reduce the number of lead planes needed. Protest at 8. This is due to the fact that one VLAT can perform the same amount of work as five to six LATs. Id. For these reasons, we do not find that the agency has sufficiently demonstrated the need to exclude VLATs from competition on the basis that they require a lead plane.
Requirement for Additional Personnel
The agency contends that, due to their size, VLATs require additional government personnel to staff. COS at ¶ 16. The agency has yet to quantify the number of additional personnel. Id.; MOL at 8, 10.
By contrast, GST contends that the use of one VLAT versus one LAT would likely require a maximum of one additional person. Protest at 7; Decl. of GST President & CEO at ¶ 11. Even assuming that a VLAT requires one additional person, GST contends that the use of VLATs would likely reduce the overall personnel requirements. Protest at 7; Decl. of GST President & CEO at ¶ 12. This is because the use of a VLAT replaces the need to use several smaller aircraft. Id.
In response, the agency does not dispute these arguments, but contends that they represent mere disagreement with the agency's conclusions. MOL at 8. We disagree. The record does not reflect that the agency conducted any analysis to quantify the number of additional personnel who would be required to operate a VLAT, nor does the record reflect that the agency considered the economies of scale associated with using a VLAT as opposed to multiple LATs. In other words, although there may be pros and cons to using a VLAT, the agency has not adequately explained whether the cons outweigh the pros. Accordingly, we see no basis in the record to support the agency's conclusions with regard to these considerations.
Requirement For Additional Fuel And Retardant
The agency also contends that VLATs require additional fuel and retardant. COS at ¶ 16. Here, too, we find that the agency failed to weigh the pros and cons of using VLATs. We discuss a few of these considerations below.
GST acknowledges that a VLAT requires more fuel and carries more retardant than a single LAT. Protest at 7. GST contends, however, that the agency fails to account for the fact that one VLAT will carry more retardant than five to six LATs. Id.; Decl. of GST President & CEO at ¶ 14. Thus, GST argues that "prudent use of VLATs will in the aggregate require less fuel than a comparable amount of LATs," Protest at 7. There is no evidence in the contemporaneous record that the agency considered such economies of scale.
The agency also argues that VLATs take longer to fuel--time that could be better spent attacking a fire during initial and extended attack. MOL at 9. GST explains that it can fully fuel its VLAT in under 30 minutes and that the VLAT can remain in the air significantly longer than a LAT. Decl. of GST President & CEO at ¶ 17. There is nothing in the record demonstrating that the agency compared the fueling times of the two aircraft. In fact, at no time did the agency identify how long each aircraft takes to fuel. There is also no analysis of the impact of any alleged delay as a result of fueling a VLAT. In other words, there is no analysis or explanation of the impact on aerial firefighting of a 5 or 10 or 20-minute delay. As a result, the agency has not adequately shown why a VLAT's fueling time justifies the exclusion of VLATs from competition.
The agency also claims that VLATs can negatively impact the operations of other aircraft because other aircraft will have to wait while the VLAT is fueled and loaded with retardant. COS at ¶¶ 16, 17; AR, Tab 7, Denial of Agency-Level Protest at 1. GST disputes this claim and points out that there is no operational data or anecdotal evidence that the use of VLATs can negatively impact the operations of other aircraft. Protest at 8.
We agree that the record is lacking in this respect. In responding to a document request to produce any document demonstrating how often, since 2012, LATs and VLATs have both been dispatched by the agency to address a fire, the agency responded that no documents exist. Agency Email, Sept. 15, 2017. Additionally, in responding to a document request to produce any documents discussing problems associated with dispatching both LATs and VLATs to address a fire, including any documents discussing the merits and/or difficulties of using both LATS and VLATs at the same time, the agency responded that no documents exist. Id.
Finally, the agency asserts that fuel and retardant can be limited at some bases. COS at ¶ 16; MOL at 9. Although this might be a justification to not order the services of a VLAT for that particular fire, it is not a justification to exclude VLATs from competition altogether. The agency also claims that the VLAT's larger payload may not be needed for small initial attack fires. MOL at 9. Even if true, the agency is not required, under a "call when needed" agreement, to use a VLAT.
Isolated Incidents of Damage
To further support its claim that VLATs can negatively impact the operations of other aircraft, the agency points to two instances in which VLATs caused damage to airbases. MOL at 9 ("Contrary to the protester's allegations, the agency does have anecdotal evidence that use of VLATs can negatively impact the operations of other aircraft."). Specifically, the agency references an incident in which a VLAT hit a hanger while taxiing and another incident in which a VLAT partially damaged an airport ramp. COS at ¶ 16; MOL at 9. As a result, the agency contends that VLATs negatively affect the operations of other aircraft. MOL at 9.
In its pleadings, the agency omits any mention of the cause of these accidents, leaving our Office to assume that the damage was due to operational issues related to VLATs. The record, however, demonstrates that the accidents were solely or primarily attributable to the actions of Forest Service personnel, not to the size of the aircraft. AR, Tab 14, Damage Report. Moreover, there is no indication in the record that the accidents impacted the operations of other aircraft.
With regard to the hanger strike, the contracting officer concluded as follows:
The [Forest Service] marshal and wing walkers should have made sure that the directed route was safe and free of obstructions before signaling [the aircraft] to continue its movement. Based on a review of all the witness accounts and interviews I find that the Fire Services is 100% at fault for the damages to [the aircraft] and the hanger.
Id. at 8. As a result, the agency "agreed to cover 100% of the costs associated with the damages to the aircraft." Id.
With regard to the second incident, the contracting officer found:
The [Forest Service] marshal should have made sure that the new route requested was safe and free of obstructions before signaling the [aircraft] to move forward. Based on a review of all witness accounts and interviews, I find that the Forest Service is 75% at fault for the damages to the [aircraft] and the stair car and tug.
Id. at 1. As a result, the agency "agreed to cover 75% of the costs associated with the damages to the aircraft and the stair car equipment." Id.
In summary, the underlying record does not support the agency's allegations that VLATs were the source of these problems.
In addition to disputing the agency's stated bases for excluding VLATs, GST contends that the agency failed to consider a number of relevant factors. We highlight a few.
First, due to economies of scale, GST alleges that VLATs present a significantly more cost-competitive operation when compared to the use of LATs. Protest at 8. Using the agency's own data, GST claims that the average cost per gallon of retardant dropped is $5.55 for a VLAT and $6.96 for a LAT. Decl. of GST President & CEO at ¶ 15. GST explains that, under the current "call when needed" contract, the agency spends approximately $54,710 per day and $10,640 per hour to use each of the six 3,000 gallon aircraft. Id. at ¶ 16. GST contends that the agency could use GST's VLAT--with over six times the capacity--for $55,000 per day and $16,500 per hour. Id. Despite GST's request for any documents evidencing that the agency considered the cost advantages of VLATs, none were produced.
Second, GST argues that the restriction and the agency's rationale for including the restriction are simply not logical in the context of a "call when needed" agreement, where the placement of orders is not guaranteed and where the agency possesses ultimate flexibility to tailor its orders to precisely the aviation assets it requires for each individual fire on a case-by-case basis. Protest at 6; Comments at 2. GST states that if, for example, the agency has a particular need that cannot be supported by a VLAT, the government has the option to not call upon VLATs for that opportunity. Protest at 7. In the meantime, the government does not incur any costs to include VLATs on a "call when needed" agreement. Id.
GST's argument is supported by the RFP, which provides:
After BOAs are awarded, the CO will order a CWN Airtanker based on operational needs, which is determined to be the best value to the Government from all aircraft that are awarded a BOA. . . . To determine best value for each initial order, the CO will consider the following factors: aircraft capability, aircraft location, availability and cost at the time of its activation. . . . [T]he relative importance of these factors will vary depending upon the requirements of each fire incident. The Government's urgency in acquiring services may be a factor and override any other criteria identified above.
RFP, § C.20(a), (b). See also, AR, Tab 5, Acquisition Plan, at 5 (containing identical language). The Branch Chief of Aviation Operations also acknowledges that, if an incident commander does not want such a large payload of retardant, the incident commander "can request a different airtanker." COS at ¶ 17.
Furthermore, by precluding the participation of VLATs, GST argues that the agency curtails its ability to use VLATs when they could be of great assistance. Protest at 7. GST argues that "[i]ncreasing the Government's available options during fire season should be a priority because, depending on the location, resources may not always be available." Id. Here, too, the agency's internal documents support GST's argument. AR, Tab 5, Acquisition Plan, at 5 ("The risk of not having all qualified airtankers available under agreement for immediate dispatch for the Government's use puts the public in danger during heavy wildfire activity.").
In short, the RFP and the Acquisition Plan demonstrate the agency's preference for "having all qualified airtankers available" and for maintaining flexibility in determining what assets to order. As a result, the agency's decision to restrict those assets at this time does not withstand logical scrutiny.
We recognize that the agency is entitled to great discretion in establishing its needs. Here, however, we find that the agency has failed to provide reasonable justifications for the challenged specification, such that we are unable to conclude that the challenged specification is reasonably necessary for the agency to meet its needs.
We recommend that the agency make a documented determination of its needs. Once the agency identifies its needs, the agency should revise its solicitation to include specifications that are reasonably necessary to meet those needs. We also recommend that the protester be reimbursed the costs of filing and pursuing the protest, including reasonable attorneys' fees. 4 C.F.R. § 21.8(d). The protester's certified claim for costs, detailing the time expended and costs incurred, must be submitted to the agency within 60 days after the receipt of this decision. 4 C.F.R. § 21.8(f).
The protest is sustained.
Susan A. Poling
 The agency uses both "call when needed" agreements and exclusive use indefinite-delivery, indefinite-quantity (IDIQ) contracts to obtain airtanker services. Pursuant to an exclusive use IDIQ contract, the contractor is required to make its services exclusively available to the Forest Service and/or agency partners during a specific period of time referred to as the Mandatory Availability Period (MAP). Agency Statement, Oct. 24, 2017, at 1. As a result, the contractor is guaranteed a daily availability rate for the duration of the MAP, as well as payment for actual use of the aircraft by the agency to conduct firefighting missions. Id. Conversely, pursuant to a "call when needed" agreement, the contractor is not required to make its services exclusively available to the Forest Service and there is no MAP. Id. Moreover, the contractor is not required to accept any orders placed by the agency, and the agency is not required to place any orders. Id.; RFP, §§ B.1, C.20(b) ("The Government does not guarantee the placement of any orders for service under the BOA and the Contractor is not obligated to accept any orders."). Under a "call when needed" agreement, there is no cost to the government for days of inactivity.
 Initial attack refers to those actions taken by the first resources to arrive at a wildfire, and extended attack refers to those actions conducted when a fire cannot be controlled by initial attack resources within a reasonable period of time. Agency Report (AR), Tab 5, Acquisition Plan, at 3; Forest Service Fire Terminology Glossary, available at https://www.fs.fed.us/nwacfire/home/terminology.html (last visited Nov. 3, 2017). The agency explains that, generally, the extended attack phase commences when the fire has not been contained or controlled within 24 hours. Agency Statement, Oct. 24, 2017, at 1-2. In such situations, more firefighting resources are ordered by the initial attack incident commander. See Forest Service Fire Terminology Glossary (definition of "extended attack").
 GST's Very Large Airtanker (VLAT) competitor, 10 Tanker Air Carrier, LLC, sought permission to intervene in this protest, or in the alternative, to participate as an amicus curiae in support of GST's protest. 10 Tanker Statement, Aug. 28, 2017, at 1. 10 Tanker asserts that it has provided aerial firefighting services to the Forest Service for almost a decade. Id. at 4; Decl. of 10 Tanker President, Sept. 11, 2017, ¶ 9. Although the company represented that it did not intend to bid in response to the "call when needed" solicitation challenged by GST (due to the fact that it had recently been awarded a contract by the Forest Service to provide the exact same services solicited here), it did intend to bid in the billion-dollar Next Generation 3 exclusive use procurement for which a draft RFP has been issued that includes the same restriction at issue here. 10 Tanker Supp. Statement, Aug. 30, 2017, at 1-2 & 2 n.1. 10 Tanker expressed concern that, were our Office to deny GST's protest, the Forest Service could view our decision as sanctioning the same restriction in the upcoming exclusive use procurement. 10 Tanker Statement, Aug. 28, 2017, at 2. We denied 10 Tanker's request to intervene or submit an amicus curiae pleading, concluding that the company did not meet the definition of either a protester or an intervenor under our Bid Protest Regulations. 4 C.F.R. § 21.0(a)(1), (b)(1). Moreover, we concluded that 10 Tanker would have an opportunity to challenge the terms of any subsequent solicitation, which our Office would review at that time. Although we denied 10 Tanker's request to participate formally in the subject protest, GST was not precluded from consulting with 10 Tanker on matters not subject to the protective order. As a result, in its filings, GST submitted statements, affidavits, and documents obtained from 10 Tanker.
 For the purposes of this decision, we use the term VLAT to refer to those aircraft with tank capacities greater than 5,000 gallons. We use the term LAT to refer to those aircraft with tank capacities between 3,000 and 5,000 gallons. This is consistent with the agency's use of the terms in this protest. See e.g., Contracting Officer's Statement (COS), Aug. 28, 2017, at ¶ 12.
 The cited studies are as follows: (1) USDA Forest Service National Study of Airtankers to Support Initial Attack and Large Fire Suppression Phase 1, 1995, see AR, Tab 20; (2) USDA Forest Service National Study of (Large) Airtankers to Support Initial Attack and Large Fire Suppression Phase 2, Nov. 1996, see Tab 21; (3) USDA Forest Service Wildland Fire Management Aerial Application Study, Oct. 17, 2005, see AR, Tab 22; and (4) USDA Forest Service Large Airtanker Modernization Strategy, Jan. 17, 2012, see Protest, Attach. 2, Exh. 6.
 The agency subsequently issued a draft solicitation for this requirement on July 25, but later "suspended" the procurement. Agency Statement, Oct. 24, 2017, at 1; Protester's Email, Oct. 24, 2017.
 The contracting officer's statement included comments submitted by the agency's Branch Chief of Aviation Operations. COS at ¶¶ 14-17.
 Because the referenced procurement has since been suspended, we do not discuss what impact, if any, this draft RFP for VLATs may have had on the issues presented here.
 The agency claims that the 3,000-5,000 gallon range "has been included in [Forest Service] document[s] previous to the publishing of the Solicitation," MOL at 2, but the agency did not produce any supporting documents or show that they were considered in drafting the solicitation here. See also Supp. MOL at 2 (contending that, for 20 years, the agency had contemplated including such a restriction in solicitations).
 In its filings, GST submitted a statement from 10 Tanker in which 10 Tanker represents that agency also told 10 Tanker its VLATs would be eligible for the procurement. 10 Tanker Statement, Aug. 28, 2017, at 4.
 10 Tanker explains that a substantial number of the missions it flew qualified as initial attack operations, thereby refuting the agency's contentions that VLATs are not suited for initial attack operations. 10 Tanker Statement, Aug. 28, 2017, at 8 n.4.
 There is still no indication of how the 1995 study supports the agency's position.
 The Branch Chief of Aviation Operations indicates that the agency expects six LATs to be available for this procurement. COS at ¶ 14. However, he also states that "[a]ll of the vendors have at least one or more airtankers under construction or close to being finished." Id. The Branch Chief does not specify what number of LATs the agency believes to be sufficient. It is also unclear whether the same six LATs referenced by the agency would also be competing for contracts under the upcoming exclusive use procurement and, if so, whether the agency believes six LATs to be sufficient to support both contracts.
 The agency responds that the protester's calculations are pure speculation and have no bearing on the agency's determinations of its needs. MOL at 12.
 The directory is published by the National Wildfire Coordinating Group, which is comprised of federal, state, local, tribal, and territorial entities, including the Forest Service. See https://www.nwcg.gov/ (last visited Nov. 3, 2017). The group provides national leadership to enable interoperable wildland fire operations among the various entities. Id. This directory shows which bases are VLAT-capable based on taxiway and ramp capabilities. The bases identified by GST are listed in the directory as capable of supporting VLATs. Comments at 3; id., Exh. B, Interagency Airtanker Base Directory at 18, 23, 26, 27, 30, 34, 40, 45, 48, 60, 62, 64, 67, 75, 76, 77.
 GST contends that, on July 18, 2017, it verified this assertion with ICL, the sole source provider of retardant used by the agency. Decl. of GST President & CEO at ¶ 9.
 Moreover, the record does not reflect any consideration of some of the other issues raised by GST. For instance, in response to the protest allegations, the agency responds only that it objects to the protester's reference to portable bases because "[t]he protest does not say who will pay for these portable base setups or detail how feasible they really are." MOL at 7 n.4.
 GST correctly notes that the record does not contain any support for the agency's claim that lead planes are in short supply. Comments at 5. In response, the agency does not provide any support or further explanation. See COS at ¶ 12; MOL at 10-11. Rather, the agency makes the same claim and contends that GST's protest represents mere disagreement with the agency's conclusion. MOL at 11.
 Despite this assertion, the RFP states that the agency will provide a lead plane for all LATs that are not IA qualified. RFP, § C.4. Accordingly, it appears that the agency is willing to accept proposals from companies whose LATs/pilots are not IA rated, but is unwilling to make the same concession for VLATs vendors.
 In the same document, however, the agency suggests that the reason VLAT pilots are not IA rated is unrelated to policy. Agency Statement, Oct. 24, 2017, at 2 ("Given the expense and duration of the training needed to obtain an initial attack rating, VLAT pilots typically have not received it and are not rated to fly initial attack missions.").
 GST also claims that the agency likely overestimates the additional work required of the agency. Protest at 7. GST claims that it arranges for all necessary support equipment not carried onboard the aircraft. Decl. of GST President & CEO at ¶ 13. The agency provides no substantive response to this allegation, contending only that it is "mere disagreement" and "should be denied." MOL at 8.
 GST also highlights an additional benefit, which is that "[f]ewer aircraft in the Fire Traffic Area (FTA) means less congestion, less potential conflict, and a greater margin for safe operations." Protest at 7.
 GST contends that it was loaded with 19,000 gallons of water and 6,000 gallons of fuel in 23 minutes while deployed to Chile. Decl. of GST President & CEO at ¶ 17. 10 Tanker represents that the Forest Service "has efficiently loaded the 5000+ [gallon] tank system of the DC-10 to bring more retardant sooner to the initial attack effort than could have been possible with multiple smaller 3000-5000 gallon LATs." Decl. of 10 Tanker President at ¶ 9. We also note that one of the agency's internal documents included in the record provides that "normally" a DC-10 VLAT "will sit loaded with 5,000 gallons in order to be able to respond most efficiently to fires." AR, Tab 10, Briefing Paper, at 1. Additionally, should the agency wish to "top off" the DC-10 to its full 11,600-gallon capacity, it "will take approximately 13 minutes." Id.
 Instead, the agency claims--without support-that "[t]ime is of the essence" and that "[i]n the past, the performance of [VLATs] with maximum gallons greater than 5,000 has had a negative impact [on] firefighting efforts." MOL at 10.
 The Forest Service found the vendor to be 25 percent at fault for requesting a change to the standard route. AR, Tab 14, Damage Report, at 1. Despite this request by the vendor, we note that the Forest Service acknowledges that the new route was "authorized by the airplane marshal." Id.
 Whether the agency considered cost concerns is unclear from the record. In one instance, the agency asserts that its reasoning was not based upon cost concerns. MOL at 9. In another instance, the agency states that it crafted its solicitation in a manner that provides the "most . . . cost effective support[.]" Supp. MOL at 17. In any event, the record does not indicate that the agency performed any cost analysis.