COASTAL PRODUCTION SERVICES INC. v. Hudson

555 F.3d 426 (2009) COASTAL PRODUCTION SERVICES INC.; Forest Oil Corporation; ACE American Insurance Company, Petitioners, v. Terry W. HUDSON; Director, Office of Worker's Compensation Programs, U.S. Department of Labor, Respondents. No. 06-60766. United States Court of Appeals, Fifth Circuit. January 14, 2009. *428 Christopher Leonard Zaunbrecher (argued), Briney & Foret, Lafayette, LA, for Petitioners. Arthur Joseph Brewster (argued), Law office of Arthur J. Brewster, Metairie, LA, for Terry W. Hudson. Matthew W. Boyle (argued), Mark Ambrose Reinhalter, Donald Shire, Richard Anthony Seid, U.S. Dept. of Labor, Thomas O. Shepherd, Jr., Clerk, Benefits Review Board, Washington, DC, David Duhon, U.S. Dept. of Labor, Employment Standards Admin., New Orleans, LA, for Director, Office of Worker's Compensation Programs. Before WIENER, DeMOSS, and PRADO, Circuit Judges. WIENER, Circuit Judge: Petitioners Coastal Production Services Inc. ("Coastal"), Forest Oil Corporation, and ACE American Insurance Company (collectively, "Petitioners") seek our review of an order of the Benefits Review Board ("BRB") affirming the determination of the Administrative Law Judge ("ALJ") that Respondent Terry Hudson ("Hudson") is entitled to benefits under the Longshore and Harbor Workers' Compensation Act ("LHWCA" or "Act"), 33 U.S.C. § 901 et seq. Perceiving no error in the determinations of the ALJ and the BRB that Hudson was a maritime employee who was injured on a maritime situs, we deny the petition for review. I. FACTS AND PROCEEDINGS In January 2001, Hudson began working as a junior contract platform operator for Coastal. He was assigned to the Saturday Island facility in Barataria Bay, Plaquemines Parish, Louisiana.[1] The Saturday Island field comprises (1) a large platform with living quarters (the "Saturday Island platform"), (2) fourteen satellite wells that are connected to the platform by horizontal subsurface piping, and (3) a sunken oil storage barge ("the Cherokee") that is adjacent to the platform. The satellite wells pipe their production (a mixture of oil, gas, and saltwater) to the Saturday Island platform, where the components are separated. The required amount of the separated gas is re-distributed throughout the oil field for gas-lift purposes, and the remainder is piped to a Southern Natural Gas pipeline for resale. The separated saltwater is injected into disposal wells. The separated oil is first piped into holding tanks on the platform, then is periodically transferred to larger storage tanks on the Cherokee. When the Cherokee's storage tanks are full, the oil is transferred into customers' transport barges for delivery onshore. All oil produced from the Saturday Island field's satellite wells is ultimately shipped ashore by transport barge. Hudson's job entailed several different duties. He (1) checked the satellite wells daily by "jo-boat" and was responsible for their maintenance and upkeep, (2) serviced *429 and maintained the Saturday Island platform, (3) transferred oil from the platform's holding tanks to the storage tanks on the Cherokee approximately three to four times per week, (4) performed daily inspections and maintenance of the Cherokee, including inspecting its lines for leaks, checking its gauges, and maintaining its engines, and (5) assisted in the loading of oil from the Cherokee into the transport barges when they came to pick up the oil. Hudson testified that in performing this last duty, he would place a walk-board between the Cherokee and the arriving transport barge, hook up pipelines and hoses to transfer oil, disconnect and reconnect the hoses and pipelines as the Cherokee's holding tanks emptied, and monitor the tank levels. According to Hudson, he took part in many, if not all, of the oil transfers that occurred while he was working. On August 11, 2001, Hudson was injured when the saltwater disposal pump on which he was working exploded.[2] That pump was located on the Saturday Island platform, which is located approximately 30 to 40 feet from the Cherokee and connected to it by a permanent walkway and oil transfer pipes. Hudson filed a claim for benefits under the LHWCA. Petitioners contested coverage, arguing that pursuant to Herb's Welding, Inc. v. Gray,[3]Thibodeaux v. Grasso Production Management Inc.,[4] and Munguia v. Chevron USA, Inc.,[5] the Saturday Island platform did not qualify as a maritime location. Petitioners also contended that Hudson did not qualify as a maritime employee because his employment activities furthered the non-maritime purpose of oil production; moreover, they argued that any job-related maritime activities performed by Hudson were minimal and not a regular part of his duties. The ALJ ruled that Hudson satisfied both the situs and status requirements of the LHWCA. With respect to situs, the ALJ concluded that the Cherokee's docking facility — which was used to load oil into transport barges — qualified as an "other adjoining area customarily used by an employer in loading" a vessel, and that the Cherokee and the Saturday Island platform were component parts of a single area. The ALJ distinguished the Saturday Island platform from the fixed platforms in Herb's Welding, Thibodeaux, and Munguia on the basis of function: The Saturday Island platform and the Cherokee comprised a single facility, the maritime purpose of which is the loading of cargo (already extracted oil) onto vessels.[6]*430 With regard to status, the ALJ determined that Hudson met the requirements for land-based maritime employment because his assisting in the loading of oil and maintaining the loading equipment was essential to the cargo loading process. Additionally, the ALJ concluded that Hudson's maritime activities were routine and recurring, not episodic, momentary, or incidental to his non-maritime work. On appeal, the BRB affirmed the ALJ's holding as to both situs and status. The BRB adopted the ALJ's reasoning that the Saturday Island platform was a covered situs because (1) it is surrounded by navigable water and (2) functions as a facility for loading cargo (oil) onto vessels. The BRB determined that the Cherokee and the Saturday Island platform could not be classified as separate areas because they are permanently connected to each other through a system of pipes and a walkway. The BRB also concluded that Hudson's work in loading oil into the transport barges and maintaining the pipes and equipment necessary to the loading process constituted maritime employment and that the ALJ rationally computed the amount of time that Hudson spent in the performance of these activities. II. ANALYSIS 1. Standard of Review Our review of the BRB's decision is limited in scope to "considering errors of law, and making certain that the BRB adhered to its statutory standard of review of factual determinations, that is, whether the ALJ's findings of fact are supported by substantial evidence and [are] consistent with the law."[7] "Substantial evidence is that relevant evidence — more than a scintilla but less than a preponderance — that would cause a reasonable person to accept the fact finding."[8] We review the BRB's legal conclusions de novo.[9] Although perhaps somewhat quizzical in light of the typical understanding of the difference between conclusions of law and findings of fact, we decided in Texports Stevedore Co. v. Winchester that the determination of situs by the ALJ is one of fact.[10] Status *431 determinations are also findings of fact, unless made under an erroneous legal standard.[11] 2. Analysis For a claimant to be eligible for benefits under the LHWCA (1) his injury must occur on a maritime situs, and (2) his status must be that of a maritime employee.[12] Both requirements must be met for the claimant to receive benefits under the Act.[13] a. Situs Section 3(a) of the LHWCA states that a claimant is eligible for benefits "only if the disability or death results from an injury occurring upon the navigable waters of the United States (including any adjoining pier, wharf, dry dock, terminal, building way, marine railway, or other adjoining area customarily used by an employer in loading, unloading, repairing, dismantling, *432 or building a vessel)."[14] As the Saturday Island platform does not meet the definition of any of the kinds of facilities specifically listed in the statute, it must qualify as an "other adjoining area" to be considered a maritime situs. To qualify as an "other adjoining area," the situs must be located in proximity to navigable waters (i.e., possess a geographical nexus) and have a maritime nexus — here, "customarily used by an employer in loading . . . a vessel."[15] The situs need not be used exclusively or even primarily for maritime purposes, as long as it is customarily used for significant maritime activity.[16] No party disputes that the Saturday Island platform adjoins navigable waters, so the geographic nexus prong of the test is satisfied; the only open question is whether it is within an area with a maritime nexus.[17] At the outset, we must define the relevant area and then ask whether it is customarily used for loading a vessel or other maritime activities listed in § 3(a) of the LHWCA. There are several levels of generality that could, as a theoretical matter, be used to define the area. We could consider (1) the exact locus of the injury (the saltwater pump), (2) the Saturday Island platform, (3) the Saturday Island facility (i.e., the platform plus the Cherokee), (4) the entire Saturday Island field (i.e., the satellite wells, the gas transportation system, the platform, and the Cherokee), (5) the whole Barataria Bay, and so on. The first option is foreclosed by our opinion in Winchester.[18] The last option and levels of even greater generality are obviously absurd, and we repudiated such extremes in Winchester as well.[19] We assume without deciding that the entire Saturday Island field is also too extreme a level of generality.[20] We are left, therefore, *433 with the second and third options. This case turns on the correctness of the ALJ's determination that both the platform and the Cherokee are part of the same "overall area" that is customarily used for loading cargo (oil). We do not define a covered area for LHWCA purposes according to fence lines and local designations because they "are subject to manipulation for compensation purposes."[21] Instead, Winchester appears to have merged the two questions — definition of the area and connection to enumerated maritime activities — into a simple functional inquiry. The "area" that adjoins navigable waters for purposes of the LHWCA is that area "customarily used by an employer in loading, unloading, repairing, or building a vessel."[22] Yet the Saturday Island platform, except for its storage of fully processed and produced oil (which is arguably part of the process of loading a vessel[23]), is itself not customarily used for loading oil onto the customers' transport barges, which on-load the processed oil from the Cherokee for transport to shore. Instead, at least without the temporary storage tanks, the Saturday Island platform is better thought of as a processing plant functionally connected to oil production and not to the loading or unloading of cargo from vessels.[24] If Winchester *434 countenanced a strict reading of "area," or if we were to address this issue de novo, we might end the inquiry here and conclude that Hudson was not injured on an LHWCA-covered situs. As neither is the case, however, we continue. Winchester involved a gear room located five blocks from the gate of the nearest dock. The gear room contained "spreader bars, pallets, wire cable slings, tow motors, forklifts, etc." that were "used by stevedores to perform the loading operation."[25] We held that an injury in the area of the gear room occurred on a covered situs. Our holding on those facts teaches that simply because a vessel cannot dock for loading and unloading at a particular area does not mean that the area is not a covered situs. It further teaches that if a particular area is associated with items used as part of the loading process, the area need not itself be directly involved in loading or unloading a vessel or physically connected to the point of loading or unloading. Vessels were not loaded or unloaded directly from the Saturday Island platform, at least not with cargo. Rather, vessels were loaded with cargo (oil) from the Cherokee, which was directly and permanently connected to the platform (by pipes and a walkway), in close proximity to it (30 to 40 feet), and described by the employer in its operations manual as part of the same "gathering and processing" facility as the platform.[26] Further, the platform was functionally integral to the Cherokee's loading and unloading mission because the configuration of the Saturday Island field used the platform as the consolidation point for transport. If the platform were to sink tomorrow, the satellite wells could still produce/extract the oil, water, and gas mixture, but that mixture (whether or not processed and separated) could not be loaded onto vessels. The fact that the Saturday Island field could have been set up differently, perhaps with pipelines (as was done for gas), or with vessels loading the raw, unseparated product at each satellite well, or some other way, is no answer to the fact that as actually set up, the platform's "gathering" function plays an integral role in the loading of vessels, even if its "processing" function might not. As such, the entire platform is "customarily used by an employer in loading. . . a vessel." This conclusion is buttressed by the liberal construction we are instructed to employ in favor of LHWCA coverage,[27] and the fact that whatever argument there might be about the platform's role in loading, the Cherokee is unquestionably involved *435 in loading vessels. Covering the platform as well as the Cherokee, between which Hudson moved on a daily basis in unquestionable furtherance of the loading duties of his job (for example, to inspect the Cherokee each day), "will reduce the number of employees walking in and out of coverage"[28] and further the congressional policy of eliminating "shifting and fortuitous coverage."[29] Winchester also directed us to consider the "general area" and the "overall loading process."[30] The fact that "the specific locus of the injury is not customarily used for maritime purposes even though the general area is so used" is not fatal to a finding of maritime situs.[31] These instructions clearly contemplate some role for geography (i.e., proximity and interconnectedness), independent from function. The dissent implicitly acknowledges this when it repeats six times that the geographical separation of the Cherokee is relevant to the nexus determination. If Winchester only stood for a functional inquiry, the fact that a particular area is not used for maritime activities would be dispositive. But, Winchester does not require every square inch of an area "generally" used for loading and unloading to be so used. If it did, we would have a game of hopscotch: The bathrooms in an otherwise "adjoining area" would not be covered, nor would be pavement that although clearly within the area, had not been walked on by stevedores loading and unloading a vessel. Winchester countenanced defining a general area (a geographic notion) by its function. If a general area is customarily — not necessarily exclusively or predominantly — used for loading and unloading of vessels, all parts within it are a maritime situs. To determine whether it is fair to call a particular part of a facility "within" the "general area" used for loading and unloading, we must look both to its proximity and its interconnectedness to the loading and unloading location, along with its function.[32] Simply because Winchester extended coverage to a non-proximate, non-interconnected location on the basis of function does not mean that proximity and connection play no role in our analysis. Additional support for this proposition comes from Northeast Marine Terminal Co., Inc. v. Caputo. There, the Supreme Court addressed a facility with "two `finger-piers.'"[33] One pier, located on 21st Street, "was used to berth ships for purposes of loading and unloading them."[34] Another pier, on 19th Street, and therefore presumably some distance away, "was used only for stripping and stuffing containers and storage."[35] The Court held *436 that an injury on the 19th Street "finger-pier" nevertheless occurred on a covered situs because "[t]he entire terminal facility adjoined the water and one of its two finger-piers clearly was used for loading and unloading vessels."[36] This is obviously a close analogy to the Cherokee and the platform. Unlike the dissent, we cannot conclude in light of Caputo that there is not more than a scintilla of evidence that supports the situs determination by the ALJ and BRB. Like the BRB itself in this case, we also see the BRB's decision in Gavranovic v. Mobil Mining & Minerals[37] as support for this proposition. There, the BRB concluded that the situs requirement was satisfied because the manufacturing facility where the accident occurred was "used to store finished product and to load rail cars and trucks."[38] Even though Building 10 (the situs of the injury) was not used to load vessels (as the statute requires), it was "in close proximity to the docks and . . . not [a] separate and distinct area[ ]."[39] More to the point, the BRB said: Although the facility at which claimants herein work is a manufacturing operation, and the building in which they were injured is not directly involved with the loading and unloading of vessels, part of employer's business involves sending and receiving goods by barges or vessels — a distinctly maritime activity. Moreover, the geography of . . . the entire facility and the building in question [are such that they] are adjacent to navigable water and to the docks where barges are loaded and unloaded.[40] Again emphasizing that geography plays some role, the BRB said: "In light of the location of employer's facility and because significant maritime activity . . . occurs . . . [there], we affirm . . . that claimants' injuries. . . occurred on a covered situs."[41] *437 On the one hand, Hudson might have been injured while on the Cherokee, which clearly qualifies as a covered situs. On the other hand, Hudson might have been injured while working on the satellite wells, which do not qualify as "adjoining area[s]" under Thibodeaux. Given the close functional and geographic relationship between a clearly maritime situs (the Cherokee) and a somewhat less clearly maritime situs (the platform), several factors we have identified support affirmance on the facts of this case. First, we owe deference to the ALJ/BRB's situs determination, and they concluded that the platform was a maritime situs. Second, the LHWCA is to be construed liberally in favor of compensation. Third, Winchester deliberately selected a very broad and liberal construction of situs (a room used to store gear located blocks away from the nearest gate to a terminal and outside the property line of the port) and counseled us to do the same by inspecting the "overall loading process" and the "general area." Fourth, we effectuate congressional policy to avoid shifting and fortuitous coverage (although, as the dissent points out, this cannot always be avoided because there must be a line somewhere) by finding that the platform, adjoining the Cherokee and functionally related to its loading purpose, is a maritime situs. Considered as a whole, these additional thumbs on the scale tip the balance in favor of affirming the ALJ/BRB's determination that the platform qualifies as a maritime situs. Merely because one discrete portion of an area adjoining navigable waters, and part of the general loading and unloading area, is not directly used for loading and unloading (say, a food cart on a dock, or a scrubber as part of the hose connected to unload the oil), this should not create hopscotching coverage. If, as Petitioners urge, we were surgically to separate the platform from the Cherokee — and classify one as maritime and the other as non-maritime — we would do exactly that which Congress sought to eliminate with the 1972 amendments to the Act.[42] We decline to slice the Saturday Island facility that finely and cause that situation. We cannot say that the ALJ did not have substantial evidence that would permit a reasonable person to conclude that the Cherokee and the platform are in the same general area customarily used for loading vessels. Neither can we say in the alternative that functionally the platform was not actually part of the loading process (even though it also served the arguably non-maritime production purpose of separation). The area need only "customarily" be used for loading, not exclusively or predominantly so. This is why the presence of gas pipelines from the platform to shore does not vitiate the maritime nature of the platform. Piping gas to shore is a non-maritime activity, but, again, the platform need not be used exclusively or even predominantly for maritime activities to qualify as a maritime situs. In fact, oil is the primary product of the platform and is shipped by barge, for which the platform (unlike the satellite wells) is a necessary part of the loading process in the field as configured. So, although it need not be *438 the case, here the platform does appear to have a predominantly maritime use — facilitation of the loading of cargo (oil, the main product of the platform). This conclusion is not foreclosed by Thibodeaux or Herb's Welding. In Herb's Welding, a majority of the Supreme Court expressly declined to decide whether a fixed drilling platform qualifies as a maritime situs.[43] The dissent thought that the situs requirement was satisfied.[44] Yet, although the majority refused to reach the issue, there are hints that even it thought a drilling platform might qualify as a covered situs.[45] In Thibodeaux, we resolved that question for fixed oil "production"[46] platforms in the territorial waters of Louisiana by concluding that the platform was not a maritime situs.[47] The claimant in Thibodeaux was a pumper/gauger who was injured when he attempted to repair a leaking line under the deck of the platform.[48] Relying in part on the Supreme Court's comments in Herb's Welding, we determined that the platform did not serve a maritime purpose.[49] Specifically, we noted that Thibodeaux has pointed to no connection Garden Island Bay platform No. 276 has with maritime commerce that distinguishes it from the platforms in [Herb's Welding or Munguia]. Oil is not shipped from the platform. Although personal gear and occasionally supplies are unloaded at docking areas on the platform, the purpose of the platform is to further drilling for oil and gas, which is not a maritime purpose.[50] Because the oil production platform in Thibodeaux did not serve any maritime purpose, we held that it was neither a "pier" nor an "other adjoining area" as defined by the Act.[51] The dissent questions our reliance on Thibodeaux's statement that the platform in that case was not involved in oil shipment (i.e., the loading or unloading of cargo). Far from being an isolated or superfluous statement, that observation is the linchpin of the distinction between Thibodeaux and Caputo: The platform served no maritime purpose precisely because it was in no way involved in loading or unloading a vessel. The structure of the quoted material makes that clear: Drilling is contrasted with shipment (which we construe to mean loading or unloading). Unlike the platform in Thibodeaux, there is no evidence that the Saturday Island platform is now or ever was involved in drilling or extracting hydrocarbons,[52]*439 and it has an independent connection to maritime commerce — as we have excruciatingly demonstrated above, the loading of cargo (oil) from the Cherokee, with the necessary intermediation of the platform given the configuration of the Saturday Island field, is a maritime purpose. Although there are no docking facilities on the Saturday Island platform per se, it is directly and permanently connected to the Cherokee, and it serves as a temporary holding station for the already-produced and separated oil until it is further transferred the 30 to 40 feet to the Cherokee where it is held until it can be shipped ashore by barge. Thus, the Saturday Island platform is part of the "general area" used as part of the "overall loading process" adjoining navigable waters, clearly distinguishable from the drilling (but not shipping) platform in Thibodeaux. Based on the foregoing analysis, we conclude that Hudson was injured on a maritime situs. As the LHWCA requires both maritime situs and status, however, our analysis does not end here. We must also determine whether Hudson's status was that of a maritime employee. b. Status The Act confers maritime status on "any person engaged in maritime employment, including any longshoreman or other person engaged in longshoring operations, and any harbor-worker including a ship repairman, shipbuilder, and ship-breaker."[53] An employee may qualify for maritime status based on either (1) the nature of the activity in which he is engaged at the time of the injury or (2) the nature of his employment as a whole.[54] Although the Act does not define "maritime employment," the Supreme Court has instructed that occupations in addition to those enumerated in the statute will be covered as maritime employment if the occupation entails activities that are an integral or essential part of the loading, unloading, building, or repairing of a vessel.[55] In addition, the employee's maritime activities must be more than episodic, momentary, or incidental to his non-maritime work.[56] As Hudson was not employed in any of the occupations enumerated in the statute, his work must have been integral or essential to the loading, unloading, building, or repairing of a vessel to be covered under the LHWCA. The ALJ and the BRB determined that Hudson was a maritime employee based on his responsibility for the general upkeep of the Cherokee and his facilitation of the loading of cargo (oil) into transport barges. The ALJ and the BRB also determined that these activities were not merely episodic, momentary, or incidental to non-maritime work, because they were regularly assigned to platform operators. *440 Petitioners assert that the BRB erred by concluding that Hudson was a maritime employee because the majority of his daily activities were related to servicing and maintaining oil and gas production facilities. Petitioners claim that, under the Supreme Court's holding in Herb's Welding, such activities are not maritime in nature and, therefore, do not qualify Hudson as a maritime employee. In Herb's Welding, the Supreme Court addressed whether an employee who was injured while welding a gas-flow line on a fixed offshore oil drilling platform was a "maritime employee."[57] The Court explained that "the maritime employment requirement is an occupational test that focuses on loading and unloading."[58] After determining that the employee's work as a welder was wholly unrelated to any loading or unloading process, the Court held that he was not a maritime employee.[59] We do not disagree with Petitioners' insistence that those of Hudson's activities that related solely to the production of oil and gas from the fourteen remote production facilities do not imbue him with maritime status. But, there was more: Unlike the welder in Herb's Welding, Hudson was directly involved in the loading of cargo into transport barges for shipment to shore — a distinctly maritime activity. Specifically, Hudson transferred previously produced oil that had come to rest in the platform's storage tanks to the larger tanks on the Cherokee where it awaited transport by barge (not by pipeline); he checked the Cherokee's cargo loading lines for leaks; and he maintained its engine. Hudson also hooked up lines for transferring the oil from the Cherokee to the customer transport barges, manned the emergency shutoff during such transfers, and boarded customers' barges to witness gauge readings. These routine, non-episodic activities are all maritime in nature and support the conclusion that Hudson's status was that of a maritime employee. Petitioners nevertheless contend that even if Hudson was involved in maritime activities, he failed to dedicate a sufficient percentage of his work to such activities to bestow maritime status under the LHWCA. We disagree. The ALJ determined that, as part of his regularly assigned duties, Hudson spent approximately 9.7% of his time engaged in the above-described maritime activities.[60] To arrive at this figure, the ALJ calculated the amount of time Hudson spent (1) loading oil into transport barges and (2) servicing the equipment necessary to load the oil.[61] Crediting Hudson's testimony regarding *441 his participation in the oil transfers, the ALJ determined that Hudson spent a total of 22 hours transferring cargo. The ALJ also determined that, according to time sheets entered in the record, Hudson spent a total of approximately 122 hours maintaining the Cherokee and its equipment. As each of these calculations is supported by substantial evidence, we perceive no error in the ALJ's determinations. As for the fact that Hudson spent approximately 9.7% of his time performing maritime activities, we have never set a minimum amount of time required to qualify as a maritime employee under the Act. We have, however, upheld maritime status for an employee who spent only 2.5-5% of his employment engaged in maritime activities when those activities were part of his regularly assigned duties, as were Hudson's.[62] Thus, Hudson easily spent more than any minimum amount of time that an employee must be engaged in maritime activities to be considered a maritime employee under the Act. III. CONCLUSION We hold that, under the discrete facts of this case, the fixed platform in question is inseparable from the Cherokee and that together they constitute a loading facility for the transhipment of cargo by vessel. As such, the platform qualifies as a maritime situs, being an element that is essential to the maritime activity of loading cargo for transport. We also hold that Hudson regularly engaged in sufficient maritime activities to meet the status requirement of the Act. It follows that, as Hudson was injured on a maritime location and qualifies as a maritime employee, neither the ALJ nor the BRB erred in determining that he is eligible for benefits under the LHWCA. The petition for review is DENIED. DeMOSS, Circuit Judge, dissenting: Terry Hudson was injured while repairing a saltwater disposal pump on a fixed oil and gas production platform located in the state territorial waters of Louisiana. In my opinion, Hudson's injury did not occur on a maritime situs. By treating the Cherokee loading barge and the production platform as a single LHWCA situs, the majority fails to apportion different functional areas within the same facility into covered and non-covered areas. I believe that the Cherokee has a functional nexus with maritime activities, but the production platform does not. The petition should be granted, and the decision of the BRB awarding LHWCA benefits should be reversed. Because the majority relies on an elastic definition of "loading" and allows the tail to wag the dog in its situs analysis, I respectfully dissent. I. In order to recover benefits under the LHWCA, the employee must satisfy both the situs and status test. Herb's Welding, Inc. v. Gray, 470 U.S. 414, 415-16, 105 S.Ct. 1421, 84 L.Ed.2d 406 (1985). "Situs and status play equal, coordinate roles in determining coverage." Thibodeaux v. Grasso Prod. Mgmt., Inc., 370 F.3d 486, 493 (5th Cir.2004). The situs test is derived from 33 U.S.C. § 903(a), which states, in relevant part: [C]ompensation shall be payable under this chapter in respect of disability or death of an employee, but only if the disability or death results from an injury occurring upon the navigable waters of *442 the United States (including any adjoining pier, wharf, dry dock, terminal, building way, marine railway, or other adjoining area customarily used by an employer in loading, unloading, repairing, dismantling, or building a vessel). The majority correctly observes that the production platform where Hudson was injured does not meet the definition of any of the seven sites specifically listed in § 903(a), so it must qualify as an "adjoining area" to be considered a maritime situs. Therefore, Hudson can recover benefits under the LHWCA "only if [his] disability . . . results from an injury occurring upon . . . [an] adjoining area customarily used by an employer in loading. . . a vessel." Id. Hudson must demonstrate that he was located on a covered situs at the approximate time he sustained his injuries. See Thibodeaux, 370 F.3d at 488. II. The "adjoining area" must have both a geographical and functional nexus with maritime activities to qualify as a covered situs. See Texports Stevedore Co. v. Winchester, 632 F.2d 504, 514 (5th Cir.1980) (en banc). Because the production platform and the Cherokee are in direct contact with navigable waters, I agree with the majority that the geographical nexus prong of the situs test is satisfied. See id.; see also Thibodeaux, 370 F.3d at 494. My dissent argues that the production platform where Hudson was injured does not satisfy the functional nexus prong of the situs test. The functional nexus inquiry is fact-intensive. See Winchester, 632 F.2d at 513, 515. The scope of the covered situs cannot be determined solely through reference to fence lines and employers' designations, which are subject to manipulation. Id. at 515. Rather, "[t]he perimeter of an area is defined by function." Id. An adjoining area must be "customarily used for significant maritime activity." Id. In my opinion, the BRB's conclusion that Hudson was injured on a maritime situs is not supported by substantial evidence.[1] The production platform is not customarily used for any maritime activity, let alone significant maritime activity. In affirming the decision of the BRB, the majority disregards binding Fifth Circuit precedent holding that a "fixed oil and gas production platform," which is connected to satellite wells and located in state territorial waters, does not qualify as a maritime situs under the LHWCA. See Thibodeaux, 370 F.3d at 487-88. The majority purports to distinguish Thibodeaux from this case by relying on a single sentence within that eight-page opinion: "Oil is not shipped from the platform." Id. at 494. Even assuming that this observation is the lynchpin of the Thibodeaux court's situs analysis, it is undisputed that oil was not shipped from the production platform where Hudson was actually injured. Rather, oil was shipped from the Cherokee, which is geographically *443 separate and functionally distinct from the production platform.[2]Thibodeaux controls this case because Hudson's injury occurred on the production platform, not the loading barge. Under my analysis, "transportation" occurs at a maritime situs; "production" does not.[3] I define "production" to include both extraction (at the satellite wells) and separation (at the production platform); I define "transportation" to include the loading of cargo in maritime commerce (at the Cherokee). The production platform serves the non-maritime purpose of oil and gas extraction and separation, while the Cherokee serves the maritime purpose of a storage and loading facility for marketable crude oil. In addition to extraction and separation, the production platform also pipes marketable natural gas to shore, which even the majority acknowledges is a non-maritime activity. Because the perimeter of an adjoining area is determined by its function, the production platform should have been distinguished from the Cherokee in the majority's situs analysis. III. The fourteen satellite wells feed an unmarketable mixture of oil, gas, and salt water, which is extracted from the subterranean strata beneath Barataria Bay, to the production platform. The production platform separates that mixture into two marketable products—crude oil and natural gas—and one waste product—salt water. At his deposition, Hudson testified that the production platform contains saltwater pumps, heater treaters, compressors, and dehydrators. Like many manufacturing plants, the production platform converts raw materials into marketable products. The total flow from the satellite wells is gathered by a manifold on the production platform. Crude oil and salt water from the separation process flows to the heater treaters on the production platform, where marketable crude oil is temporarily stored. Gas from the separation process is metered for sale into the Southern Natural Gas pipeline, and reinjected throughout the field for gas-lift purposes. The salt water is stored in tanks on the production platform, and saltwater disposal pumps reinject the salt water into the ground via a satellite well. The separation process that occurs on the production platform is a discrete stage of oil and gas production; it has nothing to do with the loading or unloading of a vessel. From the 4,000 barrel-capacity storage tanks on the production platform, marketable crude oil is gravity fed via hard piping and flexible hose to the Cherokee loading barge, which has a 8,900 barrel-capacity. The production platform and the Cherokee are approximately thirty to forty feet apart and separated by open water. Marketable crude oil is pumped from the loading barge to transport barges via a pump located on the loading barge. The loading barge serves as a docking and loading facility, which can accommodate one 10,500 *444 barrel-capacity transport barge. During the eight months of Hudson's employment, a transport barge would come once or twice a week. Hudson was injured while attempting to repair a saltwater disposal pump on the production platform. The reinjection of salt water, which is one of the final stages of the production process, prevents contamination of the oil storage tanks on the production platform. When Hudson cranked the pump, the starter exploded and caught on fire, resulting in serious burns and shrapnel wounds on his body. At the time of his injury, no transport barge was being loaded with crude oil from the Cherokee. Hudson admitted that the vast majority of his activities as a contract operator were related to the servicing and maintenance of equipment on the production platform. No cargo or other materials are ever loaded or unloaded onto a vessel in the area of the platform where Hudson was injured. No pipeline connects the saltwater disposal pump to the Cherokee. The equipment used to produce oil and gas on the production platform is the same equipment used to produce oil and gas on the land. All of the loading equipment is located on the Cherokee; Hudson did not venture onto the production platform during the loading process. IV. The crux of my argument is simple: the production platform, which extracts hydrocarbons and converts them into a marketable form, is functionally distinct from the loading barge, which facilitates the loading of marketable crude oil onto mobile transport barges. The production of a commodity at a non-maritime situs is different from the transportation of that commodity at a maritime situs. Although both the production platform and the Cherokee have oil storage tanks, only the Cherokee stores crude oil in a manner that facilitates the loading of a vessel. The majority fails to grasp the following fact: oil and gas production is the only activity that occurred on the platform where Hudson was injured. In an effort to avoid the import of this inconvenient truth, the majority makes two flawed arguments: (1) the production platform is a covered situs because it is connected to the Cherokee by pipes and a permanent walkway, and (2) the production platform is a covered situs because it stores marketable crude oil in temporary storage tanks, which is part of the loading process. The first argument fails because it impermissibly bootstraps the loading activities that occur on a separate structure, and the second arguments fails because it relies on an over-inclusive definition of "loading." In holding that the production platform and the Cherokee constitute a single LHWCA situs, the majority allows proximity to trump functionality. The central issue is whether Hudson was injured on an "adjoining area customarily used by an employer in loading . . . a vessel." 33 U.S.C. § 903(a). When defining the scope of the LHWCA situs, "[t]he perimeter of an area is defined by function." Winchester, 632 F.2d at 515. Under the functional approach, the production platform is inextricably intertwined with the satellite wells, not the Cherokee loading barge. Without one, the other is useless. The satellite wells and the production platform are involved in two distinct stages of the production process: the satellite wells extract an unmarketable mixture of oil, gas, and salt water from the subterranean strata, and the production platform separates that mixture into marketable products and disposes of the waste. The activities occurring on the production platform have nothing to do with cargo transportation or *445 maritime commerce. That happens later, at the loading barge.[4] Nevertheless, the majority's analysis combines two structures that lack a functional nexus (the production platform and the loading barge) and separates two structures that share a functional nexus (the production platform and the satellite wells). Production does not cease until you have a marketable product. In this case, you do not have a marketable product until the total flow from the satellite wells is separated using the saltwater pumps, heater treaters, compressors, and dehydrators on the production platform. Transportation begins when the crude oil is transferred from the temporary storage tanks on the production platform to the loading barge, at which point it becomes cargo awaiting transportation in maritime commerce. The ALJ, the BRB, and the majority all focus on the fact that the production platform and the loading barge are connected via pipes and a permanent walkway. This observation does not alter the reality that production occurred on one functionally integrated structure (the satellite wells and the production platform) and transportation occurred on another (the Cherokee). Many land-based oil and gas production sites are connected, whether by pipeline or asphalt, to various maritime sites used for shipment of cargo by vessel. To further add to the confusion, the majority insists that the satellite wells and the production platform are not functionally integrated despite the fact that the satellite wells are connected to the production platform via gulf-floor transmission lines. If the determinative factor of the situs test is interconnectedness, why is the production platform inextricably intertwined with the loading barge but not with the satellite wells? The majority refers to the close proximity and permanent attachment of the production platform and the Cherokee to support its argument that these two functionally distinct areas constituted a single, inseparable facility. We have previously held that function, not proximity or physical connectivity, is the determinative factor regarding situs. See Winchester, 632 F.2d at 515. While proximity and physical connectivity might provide some evidence of common function, they are not a substitute for it. According to the majority, Hudson spent 90.2% of his time performing non-maritime activities on the production platform and 9.8% of his time performing maritime activities on the loading barge. In holding that the loading activities occurring on the Cherokee control the situs designation of the production platform, the majority is allowing the tail to wag the dog. The production platform where Hudson was injured is geographically separate and functionally distinct from the Cherokee. Perhaps there will be cases where the production and transportation areas in a single facility are situated over navigable waters and are not geographically separate and functionally distinct. This case does not fall within that hypothetical category. The majority's second argument relies on an expansive definition of "loading" that includes incidental storage at the site of production, regardless of whether that storage actually facilitates the loading of a vessel. After observing that marketable crude oil is stored in temporary storage tanks on the production platform, the majority argues that this storage is part of the loading process because "[i]f no oil flows from the platform's storage tanks to the Cherokee, no oil is available for a *446 vessel to load." Because part of the "general area" is used for loading, the majority concludes that the entire production platform is a covered situs. The majority's reliance on the attenuated connection between post-production storage on the platform and loading on the Cherokee is suspiciously similar to the "but for" logic that it repudiates in a footnote. Not all storage is related to loading. Most if not all production facilities contain temporary storage areas for finished products. In the absence of a pipeline, any land-based oil and gas production site must possess temporary storage tanks for marketable crude oil that has recently been extracted and separated. In my opinion, the statutory term "loading" does not encompass temporary storage at the production platform because that storage facilitates the production of crude oil, not the loading of a vessel. Only the Cherokee stores crude oil in a manner that facilitates the loading of a vessel. Unlike the crude oil stored on the production platform, the crude oil stored in the Cherokee is awaiting shipment by transport barge. The "hopscotch" hypothetical posed by the majority is a red herring because no part of the production platform, including the temporary storage tanks, is integral to the loading of the transport barges. The majority declares that the production platform was the "consolidation point for transport." I beg to differ. The 8,900 barrel-capacity storage tank in the Cherokee serves that "gathering function," not the temporary storage tanks on the production platform. The production platform gathers the total flow from the satellite wells in order to separate it, and the Cherokee gathers the crude oil from the production platform in order to transport it. Over twenty years ago, the Supreme Court declared that oil and gas production is "not even suggestive of traditional maritime affairs." Herb's Welding, 470 U.S. at 422, 105 S.Ct. 1421 (citing Rodrigue v. Aetna Cas. & Sur. Co., 395 U.S. 352, 360-61, 89 S.Ct. 1835, 23 L.Ed.2d 360 (1969)). The majority should have avoided expanding LHWCA coverage into this uncharted territory. See Herb's Welding, 470 U.S. at 421, 105 S.Ct. 1421 ("[The plaintiff's] approach would extend coverage to virtually everyone on the stationary [drilling] platform. We think this construction of the [LHWCA] is untenable.") (status case); Thibodeaux, 370 F.3d at 488 ("The sole issue for our review is whether a fixed oil production platform built on pilings over marsh and water and inaccessible from land constitutes either a `pier' or an `other adjoining area' within the meaning of § 903(a). We hold the platform in question is neither. . . .") (situs case). V. Precedent from the Supreme Court, the BRB, the Fifth Circuit, and the Eleventh Circuit supports the conclusion that the production platform where Hudson was injured does not qualify as a covered situs. The cases cited by the majority are distinguishable. The similarities between this case and Thibodeaux are striking. Both Hudson and Thibodeaux were injured while attempting to repair equipment on fixed oil and gas production platforms, which were connected to satellite wells. 370 F.3d at 487. The platform where Hudson was injured served no greater "maritime purpose" than the platform where Thibodeaux was injured. Id. at 490. These production platforms were not "customarily used by an employer in loading . . . a vessel." See 33 U.S.C. § 903(a). The only significant factual difference between the two cases is that the production platform in this case was attached by pipeline and a permanent *447 walkway to a separate structure used for loading marketable crude oil onto a transport barge. In describing the Fifth Circuit's "functional approach," Thibodeaux reaffirmed the holding of Jacksonville Shipyards, Inc. v. Perdue, 539 F.2d 533 (5th Cir.1976), that "a putative situs actually be used for loading, unloading, or one of the other functions specified in the Act." Id. at 489 & n. 3. It is undisputed that Hudson's injury did not occur on the Cherokee or a docked transport barge. Thibodeaux also indicated that the "adjoining area" should have some resemblance to the other parenthetically enumerated structures in the statute. Id. at 491 n. 5. Although the loading barge, which has docking facilities, resembles a pier, we have previously held that an oil and gas production platform does not fall within that enumerated category because it does not have a connection to maritime commerce. See id. at 491, 493. Like any other land-based oil and gas production site, the platform in this case was solely used for a non-maritime function: the exploration and discovery of hydrocarbons, the drilling of test and production wells, the extraction of those hydrocarbons from subterranean strata, the separation of those hydrocarbons into marketable products, and the disposal of the production waste products.[5] Both the Supreme Court and this Court have declared that the oil and gas production process has absolutely nothing to do with maritime commerce. See id. at 491 ("[I]t would be incongruous to extend [the LHWCA] to cover accidents on structures serving no maritime purpose."); see also Herb's Welding, 470 U.S. at 425, 105 S.Ct. 1421. Although Herb's Welding is often pigeonholed as a "status" case, "[its] description of the work performed on fixed oil production platforms as non-maritime is highly relevant to the issue of whether the oil production platform has a connection to maritime commerce." Thibodeaux, 370 F.3d at 494. Importantly, Herb's Welding reaffirmed the classification of off-shore production platforms as "islands," which reinforces the conclusion that the island where Hudson was injured was geographically separate and functionally distinct from the Cherokee, which was used to load cargo that was produced on that island. See 470 U.S. at 422 n. 6, 105 S.Ct. 1421 (citing Rodrigue, 395 U.S. at 360, 89 S.Ct. 1835). VI. I agree with Coastal's argument that we must apportion different functional areas within the same facility into covered and non-covered areas. The apportionment principle naturally flows from the requirement that "[t]he perimeter of an area is defined by function." Winchester, 632 F.2d at 515. Because this circuit has not been presented with a case requiring application of the apportionment principle, it is appropriate to look to persuasive precedent from the BRB and the Eleventh Circuit for guidance. In my opinion, the BRB unsuccessfully distinguishes its own precedent indicating that the production platform is not a covered situs. In Jones v. Aluminum Co. of America, the decedent's widow sued ALCOA under the LHWCA, claiming that her husband died of cancer due to exposure to asbestos on a covered situs. 35 BRBS 37, 2001 WL 467885, at *1 (April 9, 2001). The decedent worked as a millwright welder *448 and later as a general mechanic, which required him to construct, repair, and maintain all types of equipment in the ALCOA facility, which was adjacent to the navigable waters of the Mobile River. Id. at *1, *6. The decedent's regularly-assigned duties included repairing and maintaining the bauxite conveyor belt system, which was used to unload raw bauxite for the production of aluminum. Id. at *6. The conveyor belts "transported bauxite from the ships to [ALCOA's] storage facility for later use in the manufacturing process." Jones v. Aluminum Co. of Am., BRB No. 97-287, 31 BRBS 130 (Oct. 16, 1997). In addition to his work on the conveyor belt system, the decedent worked the vast majority of his time in the aluminum manufacturing plant, which was located near the docks and conveyor belts in the same ALCOA facility. Regarding the situs issue, the BRB stated the following: [T]hat portion of employer's facility where loading and unloading occur constitutes a maritime situs. It is also clear, however, that employer's manufacturing plant is not a covered situs. The Fifth Circuit's decision in Winchester recognizes that the "function" of an adjoining area must be one that is used for the loading, unloading, repairing or building of vessels. A plant that manufactures aluminum oxide is not engaged in these functions. In Stroup [v. Bayou Steel Corp., BRB No. 97-1406, 32 BRBS 151 (BRB July 2, 1998)], the Board recognized that there is a point at which the maritime process ceases, and the manufacturing process begins, and vice versa. This statement is consistent with cases holding that employees, whose duties are integral to a manufacturing process rather than to a longshoring process, are not engaged in maritime employment pursuant to Section 2(3) of the Act. As employer's operation contains manufacturing facilities as well as areas used in maritime work, the entire site is not covered under Section 3(a); the plant itself lacks the functional nexus to be considered a covered area, and it cannot be brought into coverage simply because goods are shipped by water from another portion of the facility. Jones, 2001 WL 467885, at *6 (internal citations omitted). In this case, the BRB held that the production platform where Hudson was injured was more analogous to the covered situs in Gavranovic v. Mobil Mining & Materials, 33 BRBS 1, 1999 WL 122921 (Feb. 23, 1999), than the non-covered situs in Jones. In Gavranovic, the employee's injury occurred in a cargo storage building that was separate from the production area and was connected to another building that stored cargo awaiting transport by vessel. See Gavranovic, 1999 WL 122921, at *4. The BRB found that the building where the injury occurred was adjacent to navigable water, was in close proximity to the docks, and was not a separate and distinct area. Id. In both Gavranovic and Jones, the BRB recognized that the entire area used for loading and unloading, including the areas containing conveyor belts and cargo storage facilities, qualified as a covered situs. Jones, 2001 WL 467885, at *6; Gavranovic, 1999 WL 122921, at *4. Thus, if the employee is injured in an area used as part of the overall shipment process, then he is injured on a covered situs, but if the employee is injured in an area "where only manufacturing took place," then he is not.[6]Jones, 2001 WL 467885, at *7. *449 In my opinion, Jones is much more closely analogous to this case than Gavranovic, upon which the BRB relied. Both Hudson and Jones performed work in the loading/unloading area and the manufacturing/production area of a facility that was adjacent to navigable waters. In Jones, the area where the bauxite was unloaded from vessels and moved by conveyor belt was geographically separate and functionally distinct from the area where the aluminum was manufactured. Similarly, in this case, the area where the oil was stored for future shipment and loaded onto vessels was geographically separate and functionally distinct from the area where the hydrocarbons were extracted and separated into marketable form. In this case, like Jones, there was a point where the non-maritime production process ceased and the maritime loading process began. Both the production platform in this case and the aluminum manufacturing plant in Jones were part of larger facilities that were adjacent to navigable waters. After observing that an injury is covered under the LHWCA only if it occurs on a covered situs, the BRB remanded the Jones case to the ALJ for a determination of "whether decedent was exposed to asbestos on a covered situs," i.e. while he worked on the conveyor belt system used to unload raw materials from vessels. Jones, 2001 WL 467885, at *8. In this case, we know that Hudson's injury occurred while repairing a saltwater disposal pump on the production platform. Similarly, in Melerine v. Harbor Construction Co., the BRB stated that "[t]he fact that the [steel] mill receives raw materials and ships its products by water, utilizing its dock area for loading and unloading, cannot in and of itself convert the site of the mill into a covered situs." 26 BRBS 97, 1992 WL 368658, at *4 (Sept. 25, 1992). The BRB noted that the steel mill and the loading dock were geographically separate and functionally distinct. See id. at *3. Because "the line between [the employer's] manufacturing and loading operations is clearly drawn," the BRB held that it was appropriate "to treat the mill as one functional area and the dock as another." Id. at *4. Importantly, the BRB held that "finished products enter the stream of maritime commerce only after delivery by truck to the dock area." Id. Like Melerine, the line between production and loading operations is "clearly drawn" in this case. The marketable oil enters the stream of maritime commerce only after gravity-fed delivery of the cargo from the production platform to the loading barge. At that point, like the situation in Gavranovic, the crude oil is stored in the Cherokee in a manner that facilitates the loading of a vessel. Finally, in Dickerson v. Mississippi Phosphates Corp., the BRB held that the employer's phosphoric acid plant, which was located about 100 feet from a navigable waterway, was not a covered situs. 37 BRBS 58, 2003 WL 21041375, at *6 (April 29, 2003). The facts of Dickerson closely resemble the facts of Jones and Melerine. See id. at *5. The BRB in Dickerson rejected the employee's argument that the plant was a covered situs because it abutted navigable waters and was part of a larger manufacturing facility that included a dock area. Id. at *5. No loading or unloading of a vessel occurred in the area where the employee was injured, and the plant was not used for any intermediate steps in the loading process. Id. Because the plant did not have "a functional nexus to maritime *450 activity," the employee could not satisfy the situs test. Id. at *6. The situs designation of the phosphoric acid plant was not altered by the presence of a pipeline that connected the phosphoric acid plant to the fertilizer plant, or the presence of a conveyor belt that connected the fertilizer plant to the dock. Id. at *5. Dickerson demonstrates that the production platform's situs designation is not affected by (1) the geographical proximity of the platform to either navigable waters or the loading barge; or (2) the pipes connecting the production platform to the Cherokee. Jones, Melerine, and Dickerson support the conclusion that we must apportion different functional areas within the same facility into covered and non-covered areas. I believe that the majority's "all or nothing" approach fails to recognize this basic principle. VII. Eleventh Circuit precedent supports the conclusion that the production platform where Hudson was injured was not a covered situs. In Bianco v. Georgia Pacific Corp., the Eleventh Circuit addressed whether the sheetrock production department of a gypsum products facility qualified an "adjoining area" under § 903(a). 304 F.3d 1053, 1054 (11th Cir.2002). In Bianco, the gypsum products facility lied on the banks of the Turtle and East Rivers. Id. Raw gypsum was unloaded from vessels onto a series of conveyor belts that moved the gypsum to the employer's "rock shed" at its production plant. Id. at 1054-55. From the rock shed, the gypsum was processed and then transported to the sheetrock production department. Id. The finished product was eventually transported to market by truck. Id. The employee in Bianco sustained an injury in the sheetrock production department Id. During the course of his employment, the employee participated in both the production and the unloading process. Id. Relying in part on the principles contained in Winchester, the Eleventh Circuit held that the sheetrock production department where the injury occurred did not satisfy the situs test.[7]Id. at 1058. Because this area was used solely for manufacturing sheetrock and was not part of the "on-going overall process of unloading raw gypsum," the Eleventh Circuit concluded that it was functionally distinct from the unloading area contained in the same facility. Id. Significantly, the Eleventh Circuit responded to the employee's argument that "[because] a portion of the [employer's] facility is maritime, the entire facility must be, because to hold otherwise would result in workers walking in and out of coverage." Id. at 1059. Based on the Supreme Court's statements in Schwalb[8] and Herb's Welding, the Eleventh Circuit determined that this principle "was more concerned with workers engaged in maritime activity walking in and out of coverage at or near the water's edge." Id. Most importantly, the Eleventh Circuit noted that congressional intent that workers not walk in and out of coverage "does not give a court the license to reach out and expand coverage beyond the terms of [33 U.S.C. § 903(a)]." Id. at 1060. *451 Like the Eleventh Circuit in Bianco, I believe that the majority's expansive definition of covered situs "would effectively be writing out of the statute the requirement that the adjoining area `be customarily used by an employer in loading . . . a vessel.'" Id. "[The] broad interpretation of `area' [contained in Winchester] is different from one that ignores other language in the statute indicating that a functional nexus to maritime activity must nonetheless exist." Id. at 1060 n. 10 (emphasis in original). This case is distinguishable from the facts of Caputo, where one employee was injured on a pier while stripping containerized cargo and another employee was injured on an adjoining area while loading cargo from a vessel onto a truck. Ne. Marine Terminal Co. v. Caputo, 432 U.S. 249, 253, 255, 97 S.Ct. 2348, 53 L.Ed.2d 320 (1977). It is also distinguishable from the facts of Winchester, where the injury occurred in a "gear room" that contained "equipment used by stevedores to perform the loading operation."[9] 632 F.2d at 507 & n. 1. The production platform in this case has a functional nexus to oil and gas production, not to the loading of cargo. I disagree with the majority's invocation of the "walking in and out of coverage" principle to justify its situs holding. First, the Supreme Court stated in Herb's Welding that "there will always be a boundary to coverage, and there will always be people who cross it during their employment." 470 U.S. at 426, 105 S.Ct. 1421; see also Bianco, 304 F.3d at 1059-60. In Jones, Melerine, and Dickerson, the BRB recognized that those employees spent at least some of their time performing maritime work on a covered situs; however, the employees were unable to establish that they were located on a covered situs at the approximate time they were injured. See Thibodeaux, 370 F.3d at 488. Second, I believe that the text of § 903(a) is not ambiguous regarding whether the production platform is a covered situs. The production platform was not "customarily used by an employer in loading . . . a vessel." See 33 U.S.C. § 903(a). "The starting point in discerning congressional intent is the existing statutory text." Lamie v. United States Tr., 540 U.S. 526, 534, 124 S.Ct. 1023, 157 L.Ed.2d 1024 (2004). Only after application of the principles of statutory construction, including the canons of construction, and after a conclusion that the statute is ambiguous may the court turn to the legislative history. Carrieri v. Jobs.com, Inc., 393 F.3d 508, 518-19 (5th Cir.2004). Third, even assuming that this principle should guide our situs analysis, the facts of this case do not present the situation that Congress sought to eliminate with the 1972 amendments to the LHWCA. In support of this argument, the majority cites to Caputo and Winchester, which I have previously demonstrated are factually distinguishable from this case. The "walking in and out of coverage" principle reflects Congress's "undoubted desire to treat equally all workers engaged in loading or unloading a ship, *452 whether they were injured on the ship or on an adjoining pier or dock." Herb's Welding, 470 U.S. at 426, 105 S.Ct. 1421; see also Sidwell v. Express Container Servs., Inc., 71 F.3d 1134, 1135 (4th Cir. 1995). This principle merely reminds us that Hudson would have been covered if he were injured on the loading barge, the transport barge, or the walkboard in-between. All three of these locations would qualify as a maritime situs under the LHWCA. The majority makes an unwarranted logical leap when it relies on this principle to transform a non-maritime oil and gas production platform into a covered situs. Hudson "is a far cry from the paradigmatic longshoreman who walked in and out of coverage during his workday and spent substantial amounts of his time `on navigable waters.'" Herb's Welding, 470 U.S. at 427 n. 13, 105 S.Ct. 1421. VIII. The regulation of oil and gas production in state territorial waters has traditionally been an area of state concern. This is not a case where Hudson would be left without a remedy if the LHWCA does not apply. Hudson was injured while performing a job required at many land-based oil and gas production sites located throughout Louisiana, and he is currently receiving state worker's compensation benefits. There is ample precedent supporting the conclusion that the production platform where Hudson was injured was not a covered situs. If there is any doubt as to whether the LHWCA applies in this case, I believe that principles of federalism counsel towards a more conservative approach. The majority and dissenting opinions from our en banc decision in Bienvenu underscore the difficulties arising from application of the LHWCA to the oil and gas production industry based in state territorial waters. The additional ambiguity injected by the majority into this area of law "is plainly in conflict with the policy favoring expeditious but limited compensation to injured workers, that underlies all programs of workers' compensation, whether at the federal or state level." Bienvenu v. Texaco, Inc., 164 F.3d 901, 923 (5th Cir.1999) (en banc) (DeMoss, J., dissenting). For these reasons, I respectfully dissent. NOTES [1] The facility is owned by Forest Oil Corporation. [2] The saltwater pump injects water into disposal wells and helps to prevent water from entering the oil storage tanks on the platform. There is no direct line from the saltwater pump to the Cherokee. [3] 470 U.S. 414, 425, 105 S.Ct. 1421, 84 L.Ed.2d 406 (1985) (holding that a welder responsible for general maintenance on a fixed oil platform in state territorial waters is not engaged in maritime employment). [4] 370 F.3d 486, 494 (5th Cir.2004) (holding that a fixed oil drilling platform is neither a "pier" nor an "adjoining area" because the platform's sole purpose is to further gas and oil production, a non-maritime activity). [5] 999 F.2d 808, 813 (5th Cir.1993) (discussing that the purpose of an oil platform is to drill for oil and gas, which is a non-maritime activity). [6] The platform also served at least one non-maritime purpose, i.e., piping natural gas to shore, and the arguably non-maritime purpose of processing the mixture received from the satellite wells. Processing is probably part of "production" in the technical sense, see Production, THE DICTIONARY OF PETROLEUM TERMS (2001) ("[T]he phase of the petroleum industry that deals with bringing the well fluids to the surface and separating them and storing, gauging, and otherwise preparing the product for delivery."), and is probably a non-maritime activity because it frequently occurs onshore and has no peculiar connection to maritime activities. We flesh out a possible counterargument in note 24 infra. It is worth noting that Herb's Welding and Thibodeaux appear to have used "production" to mean drilling and extraction (i.e., vertical removal of the product), even though the technical definition of production seems to include separation, processing, and some storage. See Herb's Welding, 470 U.S. at 416-23, 425, 105 S.Ct. 1421 (repeatedly describing the platform as a drilling platform, but once describing the platform at issue as a production platform); Thibodeaux, 370 F.3d at 490, 492, 494 (noting that drilling for oil and gas is not a maritime activity, but also referring to the platform as a production platform). Munguia, 999 F.2d at 809-10, also appears to have focused on platforms involved in drilling and extraction. The progenitor of all of these cases, Rodrigue v. Aetna Casualty & Surety Co., 395 U.S. 352, 352, 360-61, 89 S.Ct. 1835, 23 L.Ed.2d 360 (1969), described the platform at issue as a drilling platform and did not use the word production. It thus does not appear that mere classification of a platform as involved in "production" in the technical sense forecloses a finding of maritime situs, at least not as a matter of stare decisis. In any event, as we discuss infra notes 14-42 and accompanying text, the Saturday Island platform is also involved in the loading of vessels — a traditional maritime activity — so we need not reach this issue. [7] Miller v. Cent. Dispatch, Inc., 673 F.2d 773, 778 (5th Cir.1982). [8] Dir., OWCP v. Ingalls Shipbuilding, Inc., 125 F.3d 303, 305 (5th Cir. 1997). [9] B & D Contracting v. Pearley, 548 F.3d 338, 339-40 (5th Cir.2008). [10] 632 F.2d 504, 515 (5th Cir. 1980) (en banc) ("In LHWA cases, this determination of whether a site is an `adjoining area' is handled by the ALJ and reviewed by the Board. . . . The ALJ is guided in his factual determination by the section stating that in the absence of substantial evidence to the contrary, LHWA coverage is presumed. If the situs determination is supported by substantial evidence on the record as a whole, it will not be set aside by this court." (emphases added and internal citations omitted)). The references to the situs determination and a factual determination make clear that the Winchester court was speaking of the ultimate conclusion about situs, even though that may seem like a legal conclusion. This matter is not free from doubt, as we have said that "[w]e review the determination of LHWCA coverage by either an ALJ or the BRB as a question of law." Boomtown Belle Casino v. Bazor, 313 F.3d 300, 302 (5th Cir. 2002) (citing Hullinghorst Indus., Inc. v. Carroll, 650 F.2d 750, 753 (5th Cir. Unit A 1981) ("[W]e reject at the outset [the] contention that the Board exceeded its statutory authority in reversing the determination of the ALJ that Carroll was not a covered employee within the meaning of the Act. Under the uncontested facts, the ALJ's determination in this regard was neither a finding of fact nor a factual inference drawn from the evidence. It was a conclusion of law.")). There appears to be a circuit split on this question. Compare Winchester, 632 F.2d at 515, and Brooker v. Durocher Dock & Dredge, 133 F.3d 1390, 1392 (11th Cir.1998) ("Although this court reviews the ALJ's interpretation of the LHWCA de novo, it will not set aside the ALJ's findings of fact, including its situs determination, if substantial evidence supports them." (citing Winchester, 632 F.2d at 515)), with Jonathan Corp. v. Brickhouse, 142 F.3d 217, 221 (4th Cir.1998) ("Because the facts relating to the resolution of the situs issue are not substantially in dispute, coverage becomes a question of law which we determine de novo." (citing Potomac Elec. Power Co. v. Dir., OWCP, 449 U.S. 268, 278-79 & n. 18, 101 S.Ct. 509, 66 L.Ed.2d 446 (1980) and Pittman Mech. Contractors v. Dir., OWCP, 35 F.3d 122, 125 (4th Cir.1994))). The cases on which the Fourth Circuit relied deal with the deference owed to statutory constructions of the LHWCA by the BRB. In particular, Potomac Electricity Power Co. v. Director, OWCP, 449 U.S. at 278 n. 18, 101 S.Ct. 509, stands for the proposition that no deference is owed to BRB statutory constructions because the BRB is not a policymaking agency. This decision by the Supreme Court is not, however, in direct conflict with our classification of the situs determination as one of fact, not law. The Court simply concluded that the BRB's statutory interpretations of the LHWCA are owed no deference. Were we writing on a clean slate, it is unclear how the ultimate conclusion by an ALJ or the BRB about whether a locus satisfies the situs requirement of the LHWCA is a question of fact, particularly when the facts underlying the situs determination are not in dispute. We are nevertheless bound by Winchester, which was a unanimous en banc determination of this precise issue by this court, there being no contrary intervening precedent on this exact issue by either the Supreme Court or this court en banc. [11] Boatel, Inc. v. Delamore, 379 F.2d 850, 857 (5th Cir. 1967). [12] See 33 U.S.C. §§ 902(3), 903(a) (2006). [13] Herb's Welding, Inc. v. Gray, 470 U.S. 414, 415-16, 105 S.Ct. 1421, 84 L.Ed.2d 406 (1985). [14] 33 U.S.C. § 903(a) (emphasis added). [15] Id.; see Winchester, 632 F.2d at 514-15. [16] Winchester, 632 F.2d at 514-15. [17] Although "adjoining" connotes "lying next to" or "beside," we observed in Winchester that "adjoining" can also be defined as "neighboring." 632 F.2d at 514. Additionally, WEBSTER'S DICTIONARY defines "adjoining" as "touching or bounding at some point or on some line: near in space." WEBSTER'S THIRD NEW INTERNATIONAL DICTIONARY 27 (2002). Adopting the broad definition of the word "adjoining," we conclude that the Saturday Island platform adjoins navigable waters because is it located in and has direct contact with navigable waters. This interpretation conforms with our conclusion in Winchester that "[t]o instill in the term [adjoining] its broader meaning is in keeping with the spirit of the congressional purposes [of the Act]." 632 F.2d at 514. For these same reasons, the Cherokee too adjoins navigable waters. In fact, the platform and the Cherokee are in navigable waters. We nevertheless treat fixed platforms as islands for almost all purposes. See, e.g., Rodrigue v. Aetna Cas. & Sur. Co., 395 U.S. 352, 360, 89 S.Ct. 1835, 23 L.Ed.2d 360 (1969). Visually, it is best to conceive of the facility (the platform and the Cherokee) as onshore because we treat the Saturday Island facility as "land." This helps to understand how the facility "adjoins" navigable waters. [18] 632 F.2d at 516. [19] Id. at 514 ("Nevertheless, outer limits of the maritime area will not be extended to extremes. We would not extend coverage in this case to downtown Houston."). [20] We may assume this without deciding it because to pick this level of generality reduces to the choice between options two and three. If the "overall area," Winchester, 632 F.2d at 516, includes the satellite wells and the gas transportation system, we would end up with what might be termed a "mixed coverage situs" under the apportionment principle found in cases like Bianco v. Georgia Pacific Corp., 304 F.3d 1053 (11th Cir.2002). Cases like Maraney v. Consolidation Coal Co., 37 BRBS 97 (2003) and Jones v. Aluminum Co. of America, 35 BRBS 37 (2001), deal with situations in which there are clearly covered and clearly non-covered parts of an overall area. The satellite wells and the gas transportation system are, after Thibodeaux, clearly not covered sections of the overall area. But, Hudson's injury did not occur at either of those places. Accordingly, we would still have to answer whether only the Cherokee or the platform plus the Cherokee are covered parts of the overall area. [21] Winchester, 632 F.2d at 515. [22] 33 U.S.C. § 903(a). [23] We note that storage of fully processed oil in the storage tanks on the platform can quite plausibly be understood to be part of the loading process. The storage tanks hold finished cargo and adjoin navigable waters; they are directly and permanently connected by pipes and a walkway to the point at which the vessel actually receives the cargo (which also adjoins navigable waters and is no more than 30 to 40 feet away); and, operation of the storage tanks is necessary for a vessel to be loaded, given the cargo storage configuration of the Saturday Island facility: If no oil flows from the platform's storage tanks to the Cherokee, no oil is available for a vessel to load. This is quite distinguishable from the dissent's example of storage facilities at most production plants. Those storage facilities (presumably) lack a geographical nexus to navigable waters, (presumably) are not directly and permanently connected to the shipping terminal, and (presumably) operation of the storage facilities is not the only way to get products to the shipping terminal and then loaded onto a vessel given the set-up of the loading system. If those presumptions are wrong, it is unclear why it would be absurd to treat such storage facilities as part of the loading process. We also note that the satellite wells, inasmuch as they are unarguably involved in extraction/production only (and not in processing and storage), do not serve a loading purpose (unlike storage and transfer), and are therefore not part of the loading process, even though they may be necessary to loading in the "but for" sense. [24] Although we do not base our decision on it, it is at least arguable that a facility integral to the conversion of material into cargo suitable for maritime transport, such as the platform's separators, has a role in loading a vessel. It would, of course, be possible to ship the mixture of oil, gas, and water on vessels without first processing it, but that would likely prove to be more expensive (the yield of marketable product per shipment would dramatically decrease because each shipment would include useless saltwater) and would also deprive the producer of an economical method of gas-lifting the field (instead, gas would have to be transported out to the field after processing onshore). We seriously doubt whether a particular process used to facilitate the economical loading of vessels must be necessary to loading a vessel before it may be considered part of the loading process (that conclusion would lead to absurd results wholly inconsistent with our cases), but the conceptual line between production/processing of hydrocarbons and loading/unloading a vessel is somewhat blurry in the case of the separation facility on the Saturday Island platform. As we do not need to resolve this issue today, we do not. [25] 632 F.2d at 507 n. 1. [26] Forest Oil Corporation, Facility Operations Manual: Saturday Island Field: Plaquemines Parish, Louisiana, at 2 (1998) (emphasis added) ("The central gathering and processing facility for the Saturday Island Field consists of one sunken barge used for storage (Cherokee), one raised barge, the S-91 barge, that is used for processing and storage and ten wooden-pile structures which are interconnected by walkways."). Although we shied away from employers' descriptions of the relevant area in Winchester, we did so to prevent employers from manipulating boundaries in a unilateral effort to determine the coverage area. When an employer has defined an area as part of a facility also used for loading and unloading, however, that concern is less pressing and we have no trouble taking account of that description. In any event, Winchester only said that "fencelines and employers' designations will not . . . end the factual inquiry." 632 F.2d at 515 (emphasis added). [27] Ne. Marine Terminal Co., Inc. v. Caputo, 432 U.S. 249, 268, 97 S.Ct. 2348, 53 L.Ed.2d 320 (1977). [28] Winchester, 632 F.2d at 516. [29] Caputo, 432 U.S. at 274, 97 S.Ct. 2348. [30] Winchester, 632 F.2d at 515-16. [31] Id. at 516. [32] The dissent questions how we can conclude that the Cherokee and the platform constitute a single situs, but also conclude that the satellite wells and the platform, which have pipes connecting them, are not. In the first instance, we note that even if the wells are interconnected with the platform, they are not proximate to it. Second, the satellite wells are not interconnected to the platform in the relevant way. The Cherokee and the platform have a walkway that allows, with a few steps, an employee to move between them, bringing concerns about shifting and fortuitous coverage into play. The satellite wells have no such connection to the platform; they must be accessed by boat. Accordingly, although physically connected, they are not connected in the relevant manner. [33] 432 U.S. at 280, 97 S.Ct. 2348. [34] Id. [35] Id. [36] Id. at 281, 97 S.Ct. 2348 (emphasis added). [37] 33 BRBS 1, *1 (1999). [38] Id. at *3. [39] Id. at *4. [40] Id. at *5 (emphasis added). [41] Id. (emphasis added). Although not precedent with which we are forced to grapple as a formal matter, we also note the BRB's decisions in Dickerson v. Mississippi Phosphates Corp., 37 BRBS 58 (2003), Jones v. Aluminum Co. of America, 35 BRBS 37 (2001), Stroup v. Bayou Steel Corp., 1998 WL 461480 (BRB, July 2, 1998), and Melerine v. Harbor Construction Co., 26 BRBS 97 (1992). In Melerine, the BRB noted that the dock was separate and distinct from the mill. The steel mill is located one-quarter mile from the dock area, separated by a road and a levee, and materials are transported by truck from the mill to the dock. The [ALJ] found that no conveyors or other loading apparatus linked the two areas. . . . The mill is used solely for the manufacturing of steel products[,] a clearly non-maritime activity. 26 BRBS at *3. In Stroup, the shipping bay at issue (attached to the structure at issue in Melerine) was "1/4 to 1/2 [of a] mile from the docks and the Mississippi River. Further, the `A' bay is separated from the River by a public road and a levee and is located in a building which also houses the melt shop, the roll mill, and another shipping bay." 1998 WL 461480, at *1. In Dickerson, the site of the accident, a phosphoric acid plant[, had] nothing to do with maritime activity and its sole purpose [was] to convert the liquid sulfuric acid into phosphoric acid and store it. It has no connection to the docks by way of conveyor belt or other means. It is geographically and functionally separate from the docks. 37 BRBS at *5 (emphasis added) (internal citations omitted); see also Maraney v. Consol. Coal Co., 37 BRBS 97, at *6 (2003) ("In the instant case, [the location of the injury] is functionally and geographically separate from employer's unloading/loading operations." (emphasis added)). Jones simply recited the unsurprising proposition that "a plant that manufactures aluminum oxide is not engaged in [loading, unloading, repairing or building of vessels]." 35 BRBS at *6. Jones is perhaps the best support for the dissent's conclusion, because it is not clear how geographically distinct the manufacturing and loading sites of the facility were. In light of Caputo and Winchester, along with the bulk of the BRB's other cases, however, we cannot justify using that ambiguity in Jones to reach a contrary result in this case. [42] See Ne. Marine Terminal Co., Inc. v. Caputo, 432 U.S. 249, 274, 97 S.Ct. 2348, 53 L.Ed.2d 320 (1977). [43] See Herb's Welding, Inc. v. Gray, 470 U.S. 414, 427, 105 S.Ct. 1421, 84 L.Ed.2d 406 (1985). [44] See id. at 446-47, 105 S.Ct. 1421 (Marshall, J., dissenting). [45] See id. at 425, 105 S.Ct. 1421 ("The dissent emphasizes that Gray was generally on or near the water and faced maritime hazards. To the extent this is so, it is relevant to `situs,' not `status.' To hold that Gray was necessarily engaged in maritime employment because he was on a drilling platform would ignore Congress' admonition that not everyone on a covered situs automatically satisfies the status test.") (emphasis added and internal citation omitted). But see id. at 422 n. 6, 105 S.Ct. 1421 ("Rodrigue did observe that offshore platforms are like piers, [but] its holding was that they are islands."). [46] See supra note 6 and accompanying text. [47] 370 F.3d 486, 487 (5th Cir.2004). [48] Id. at 487-88. [49] Id. at 494 (citing Herb's Welding, 470 U.S. at 423-24, 105 S.Ct. 1421). [50] Id. (footnote omitted) (emphases added). [51] Id. at 493-94. [52] In fact, the "platform" appears to be a barge itself. See Facility Operations Manual: Saturday Island Field: Plaquemines Parish, Louisiana, supra note 26, at 2 (noting that storage and processing occur on the "S-91 barge," although the structure of the sentence makes the object of the appositional "used for processing and storage" somewhat unclear). [53] 33 U.S.C. § 902(3) (2006). Certain exclusions, none of which are applicable here, do apply. [54] Universal Fabricators, Inc. v. Smith, 878 F.2d 843, 845 (5th Cir. 1989) (citing Hullinghorst Indus., Inc. v. Carroll, 650 F.2d 750, 754 (5th Cir. Unit A 1981)). [55] Chesapeake & Ohio Ry. Co. v. Schwalb, 493 U.S. 40, 47, 110 S.Ct. 381, 107 L.Ed.2d 278 (1989). [56] See Boudloche v. Howard Trucking Co., 632 F.2d 1346, 1348 (5th Cir. 1980). [57] 470 U.S. 414, 415, 105 S.Ct. 1421, 84 L.Ed.2d 406 (1985). [58] Id. at 424, 105 S.Ct. 1421 (internal quotation marks omitted). [59] Id. at 424-25, 105 S.Ct. 1421. [60] 142/1465 = 0.097. The ALJ determined that Hudson worked a total of 1465 hours during his employment with Coastal. Because of an arithmetical error, the ALJ mistakenly credited Hudson with 142 total hours of maritime-related activity instead of 144. Under the correct mathematical calculation, Hudson actually spent 9.8% of his time engaged in maritime activities. 144/1465 = 0.098. [61] Petitioners also advance that the ALJ erred by including the amount of time Hudson spent maintaining and servicing the equipment on the Cherokee. The ALJ determined that the maintenance of the Cherokee was necessary and essential for the proper loading and transfer of the oil into transport barges. As employees who are injured while maintaining or repairing equipment essential to the loading or unloading process are covered by the LHWCA, the ALJ did not err by including this time in its calculation. See Chesapeake & Ohio Ry. Co. v. Schwalb, 493 U.S. 40, 47, 110 S.Ct. 381, 107 L.Ed.2d 278 (1989) ("Someone who repairs or maintains a piece of loading equipment is just as vital to and an integral part of the loading process as the operator of the equipment."). [62] Boudloche v. Howard Trucking Co., 632 F.2d 1346, 1348 (5th Cir.1980). [1] The panel is bound by Winchester's holding that the BRB's situs determination is a question of fact that we review under the substantial evidence standard. See Winchester, 632 F.2d at 515. When the material facts are undisputed, as they are in this case, I believe that the BRB's situs determination should be reviewed under the de novo standard reserved for questions of law. See B&D Contracting v. Pearley, 548 F.3d 338, 340 (5th Cir.2008). Perhaps the en banc court will have an opportunity to revisit this issue in the future. Nevertheless, even under the more deferential substantial evidence standard, the record does not contain more than a scintilla of evidence that the production platform where Hudson was injured was "customarily used for significant maritime activity." See Dir., OWCP v. Ingalls Shipbuilding, Inc., 125 F.3d 303, 305 (5th Cir. 1997). [2] The geographical nexus prong of the situs test focuses on the production platform's geographical nexus with navigable waters. In contrast, the fact that the production platform is "geographically separate" from the loading barge is relevant to the functional nexus inquiry. [3] "Function" is defined as "the action for which a . . . thing is specially fitted, used, . . . or for which a thing exists." WEBSTER'S THIRD NEW INTERNATIONAL DICTIONARY 920 (16th ed.1971). The function of the satellite wells and the platform is to produce marketable crude oil. The function of the loading barge is to transport that crude oil in maritime commerce. Only the latter function occurs on a LHWCA situs. [4] In support of its argument that the production platform qualifies as an "adjoining area," the majority repeatedly refers to loading activities that occurred on the Cherokee. [5] Winchester and Thibodeaux suggest that the satellite wells (where extraction occurred) and the production platform (where separation occurred) constitute a single functional entity. [6] At most, Gavranovic stands for the proposition that an interconnected cargo storage facility qualifies as a covered situs if it is distinct from the production facility, adjoins navigable waters, and contains some cargo awaiting transport by vessel. To that extent, it is relevant to the situs designation of the Cherokee, not the production platform. [7] The Eleventh Circuit has adopted all Fifth Circuit decisions issued before October 1, 1981, including Winchester, as binding precedent. Bonner v. City of Prichard, Ala., 661 F.2d 1206, 1209 (11th Cir. 1981) (en banc). Because the Eleventh Circuit relied on Winchester when it applied the apportionment principle in Bianco, I find its reasoning particularly persuasive. [8] Chesapeake & Ohio Ry. Co. v. Schwalb, 493 U.S. 40, 110 S.Ct. 381, 107 L.Ed.2d 278 (1989). [9] Unlike the production platform where Hudson was injured, the "gear room" where Winchester was injured was an essential part of the loading and unloading process: [Winchester's] duties as a gear man included supplying and repairing the tools and machinery used by stevedores in loading and unloading ships. His work was performed at the dockside, on board ships, and at each of the gear rooms, including those of other stevedores. When Texports was loading and unloading cargo, Winchester would service several ships, travelling over the public streets connecting the gear rooms and docks. Even when Texports had no ships to load or unload, the gear rooms operated repairing and maintaining gear for the next loading and unloading operation. Winchester, 632 F.2d at 507.