dissenting in part and concurring in part:
The U.S. Army Corps of Engineers (Corps) has authorized the filling of twenty-three valleys and more than thirteen miles of headwater streams in Southern West Virginia in connection with four mountaintop removal mining operations. Despite its failure to fully assess the impact that the proposed valley fills will have on the aquatic ecosystem, the Corps claims that, after mitigation measures have been implemented, the valley fills will not significantly degrade the waters of the United States or have a significant adverse impact on the human environment. Because the record in this case does not support the Corps’ claims that the assessments conducted and the mitigation measures imposed were adequate to fulfill the requirements of the Clean Water Act (CWA) and the National Environmental Policy Act (NEPA), I respectfully dissent from part IV.B of the majority’s opinion.
The Corps’ regulations implementing the CWA’s § 404(b) dredge and fill program require the Corps to assess the effect that a proposed fill will have “on the structure and function of the aquatic ecosystem and organisms.” 40 C.F.R. § 230.11(e). In upholding the Corps’ interpretation of its obligations under § 230.11(e), the majority declines to give effect to the unambiguous requirements of the regulations. The majority — agreeing with the Corps — concludes that an evaluation of stream structure may substitute for an evaluation of function. This interpretation, however, is impossible to reconcile with the plain language of the regulations, which clearly mandates that the Corps assess both structure and function. The majority then accepts the Corps’ alternative argument that the (nonfunctional) stream assessment protocols used by the Corps provided sufficient information about stream function to satisfy the demands of § 230.11(e). But because the majority has failed to identify the stream functions to be measured under § 230.11(e), the majority cannot meaningfully evaluate the adequacy of the stream assessment protocols that were used.
The majority’s analysis of the mitigation approved by the Corps also overlooks the plainly stated requirements of § 230.11(e). Rather than basing its decision on the (binding) language of the regulations, the majority focuses instead on the Corps’ compliance with an internal guidance document that is at odds with the regulations’ clear requirements. The effect is to completely undermine the goal of mitigation: replacement of what is being lost. Because the Corps has offered no basis on which to conclude that the environmental impacts of the valley fill projects as mitigated will be insignificant, this court should reject the mitigation as inadequate under the CWA and NEPA.
For these reasons, I would affirm the district court’s judgment rescinding the permits and direct that court to remand the permits to the Corps for further con*218sideration consistent with the requirements of § 230.11(e) and NEPA.
I concur in the parts of the majority opinion upholding the scope (or physical boundary) of the Corps’ NEPA analysis (part IV.A) and the Corps’ interpretation of its regulatory definition of “waters of the United States” (part V).
I.
Before the Corps can issue a dredge or fill permit, it must determine, pursuant to its § 404(b) regulations, the “nature and degree of effect that the proposed discharge will have, both individually and cumulatively, on the structure and function of the aquatic ecosystem and organisms.” 40 C.F.R. § 230.11(e). As the Corps points out, “[t]he Guidelines [or regulations] do not ... define the ‘function of the aquatic ecosystem’ or provide any guidance on how that function is to be measured.” Corps’ Br. at 35. The Corps therefore contends that under its internal guidance documents, its district officials may “discharge their responsibility to analyze stream function by exercising their ‘best professional judgment’ when it is not feasible to conduct a full functional assessment.” Corps’ Br. at 35.
A.
The Corps, purporting to exercise its “best professional judgment,” claims initially that stream structure can be measured as a surrogate for function. Corps’ Br. at 36. The majority accepts the Corps’ argument, stating that “[i]n this case, the Corps, using its best professional judgment, used stream structure as a surrogate for assessing stream function”. Ante at 199. Neither the Corps nor the majority explains how the assessment of structure as a surrogate for function can be squared with the plain language of the regulations. If stream structure were truly an adequate surrogate for stream function, the Corps and the majority should offer some explanation as to why § 230.11(e) explicitly requires assessments of the effects of proposed fills on both the structure and function of the aquatic ecosystem and organisms. One of the most basic rules of statutory interpretation is that all of the words in a statute must be given effect. PSINet, Inc. v. Chapman, 362 F.3d 227, 232 (4th Cir.2004) (“General principles of statutory construction require a court to construe all parts to have meaning and to reject constructions that render a term redundant.”); Reiter v. Sonotone Corp., 442 U.S. 330, 339, 99 S.Ct. 2326, 60 L.Ed.2d 931 (1979) (“In construing a statute we are obliged to give effect, if possible, to every word Congress used.”). It is not within the bounds of permissible interpretation to say that the word “function” as used in § 230.11(e) is merely a redundancy for “structure.”1
*219The argument that “structure” is a surrogate for “function” is further undermined by the fact that a separate functional stream assessment protocol is currently being developed by the Environmental Protection Agency (EPA). Once the protocol is completed, the Corps will use it to conduct § 230.11(e) functional analyses for new permit applications. This functional stream protocol will direct the assessment of numerous stream functions (including nutrient uptake and transport, organic matter retention, and downstream export of organic matter) that the Corps did not measure for any of the permit applications at issue in this appeal. If existing schemes used by the Corps to assess stream structure were adequate surrogates for a functional assessment protocol, it would make little sense for the EPA to be currently undertaking the expensive and time-consuming effort of developing an independent functional protocol.
The Corps’ determination that stream structure can be used as a surrogate for function under § 230.11(e) constitutes a clear abuse of discretion. This court should not uphold a construction that gives no effect to a central term in the controlling regulations.
B.
The Corps, perhaps recognizing the weakness of its position, does not rely entirely on the argument that structure is a surrogate for function. It contends at times in its opening brief that the stream assessments conducted here were sufficient to satisfy both the structure and function inquiries mandated by § 230.11(e). The majority attempts to uphold the Corps on this alternative ground as well. Unfortunately, because the majority does not come to grips with the plain language of § 230.11(e), its review lacks both a defined scope and a clear legal standard.
The majority credits the Corps for using “detailed measurements provided by [permit applicants] on the benthic macroinver-tebrate population to draw conclusions about the level of stream function at the proposed fill sites.” Ante at 200. The majority also credits the Corps for making use of the EPA’s Rapid Bioassessment Protocol (RBP) and the West Virginia Stream Condition Index (WVSCI or Index). But because the majority has not identified the relevant stream functions to be measured, it is not possible to say whether these assessment protocols provide relevant information. Further, by failing to acknowledge or employ in its review the relevant language from the regulations that requires the Corps to assess the “nature and degree of effect” that the proposed fills will have on stream function, the majority affords itself no legal basis for testing the sufficiency of any assessment of stream function.
To the extent that the record provides some indication of the appropriate meaning of “function” as used in § 230.11(e), the Corps’ assessment of function was demonstrably inadequate. There is some dispute among the parties as to which stream functions are appropriately covered by § 230.11(e). However, as mentioned above, the EPA (with Corps’ approval) is in the process of developing a functional assessment protocol for streams in West Virginia’s Huntington District that will serve as the standard for future § 230.11(e) functional assessments. Hence, the most logical place to begin an inquiry into the meaning of the term “function” in § 230.11(e) is with EPA’s *220proposed (or draft) stream assessment protocol.2 Under the proposed protocol:
Contractor activities would involve field and laboratory studies aimed at conventional measurement of headwater stream functional processes in mined and un-mined watersheds. These functional processes can include, but are not limited to: 1) organic matter decomposition rates; 2) nutrient transport and uptake; 3) primary production and metabolism; 4) secondary production; and 5) organic matter retention and transport.
J.A. 1836.3
A look at only those functions listed by the EPA in its proposed functional stream *221assessment protocol reveals that the assessments carried out by the Corps were deficient. The data provided by the EPA’s RBP and the WVSCI, the protocols used by the Corps in this case, are insufficient to assess the bulk of the functions listed by the EPA to a degree that satisfies the requirements of § 230.11(e).
Indeed, the Corps, in its CDD for the Republic No. 2 mine, acknowledges the limitations of using the EPA’s RBP to assess stream function:
While the rapid bioassessment protocol does not provide a detailed analysis of nutrient cycling, organic matter dynamics, respiration, measurement of primary/secondary production, as is typically found in scientific research analysis[,] it does provide baseline data that can be used to analyze chemical, physical, and biological conditions of the stream channel.
J.A. 3570. The functions about which the RBP fails to provide detailed analysis are the very functions the EPA’s proposed functional analysis would evaluate. And the plain language of § 230.11(e) — requiring an assessment of the “nature and degree of effect that the proposed discharge will have ... on the structure and function of the aquatic ecosystem and organisms”— makes clear that stream function must be assessed in some detail. The Corps does not explain how the baseline data on stream condition generated by the RBP will assist in any way in measuring the (actual) stream functions for which the RBP provides no detailed analysis. Accordingly, the RBP appears to be an inadequate substitute for a functional assessment protocol.
The WVSCI fares only slightly better. The Index purports to measure stream “quality” and notes that its surveys “are used to measure the attainment of biological integrity.” A Stream Condition Index for West Virginia Wadeable Streams, at 3 (July 21, 2000), available at http://www. wvdep.org/Does/536_WV-Index.pdf. The Index asserts that it is “an appropriate indicator of ecological quality, reflecting biological responses to changes in physical habitat quality, the integrity of soil and water chemistry, geologic processes, and land use changes (to the degree that they affect the sampled habitat).” Id. at 4. It makes no mention of organic matter processing or retention, primary or secondary production, nutrient retention, cycling, transport or uptake, or respiration. And nowhere does it claim to be a functional assessment protocol.
It is not enough that the Corps’ expert Dr. Mindy Armstead testified that the WVSCI’s EPT Index “was a good surrogate for the functional measurement of secondary biomass,” J.A. 4424-25, one of the functions to be covered by the EPA’s functional assessment protocol. Measurement of a single function does not make the Index an adequate replacement for the required functional assessment. For decomposition and primary production, two other functions to be measured under the EPA’s proposed functional assessment protocol, the Corps’ experts claim only that the WVSCI provides information about their presence or absence. The mere ability of an assessment protocol to detect the presence or absence of a stream function is insufficient to fulfill the more exacting “[determine the nature and de*222gree of effect” language of § 230.11(e). And for the remaining functions slated to be measured under the EPA’s functional stream assessment protocol — nutrient uptake and processing, organic matter retention, and downstream export of organic matter — the Corps makes no claim that the WVSCI provides any relevant information at all.
The majority opinion considers only one stream function — nutrient cycling' — and concedes that the Corps’ assessment of that function was deficient: “The Corps’ CDDs themselves acknowledge this shortcoming, noting that the effects of filling ephemeral streams on nutrient cycling are difficult to measure and that there is a lack of consensus among the relevant agencies about how best to collect quantitative evidence regarding these functions.” Ante at 200. The majority excuses this deficiency, noting that: “[t]o compensate for these effects ... the Corps’ permitting decisions call for limiting impacts to channels that do not sustain long periods of flow and for establishing a riparian buffer around mitigation sites.” Ante at 200. Unfortunately, the Corps’ attempt to minimize impacts to stream functions that it has failed to assess sufficiently has no bearing on whether it has met its obligations under § 230.11(e) to “[determine the nature and degree of effect” on stream function.
Finally, and most fundamentally, in asserting that the proposed mitigation measures meet regulatory requirements, the Corps implicitly concedes that the stream assessment protocols used in this case failed to sufficiently assess stream function. The Corps’ Regulatory Guidance Letter (RGL) provides that “Districts should require compensatory mitigation projects for streams to replace stream functions where sufficient functional assessment is feasible.” J.A. 1174. Where sufficient functional assessment is not feasible, the RGL permits the Corps to rely on a substitute one-to-one linear stream foot mitigation that does not specifically account for lost stream function. In this case, the Corps does not claim that its mitigation measures will replace lost stream functions. The Corps asserts in its opening brief that its “approach is not arbitrary and capricious just because the precise functions of ephemeral or intermittent streams are not being replaced” by the mitigation measures required in this case. Corps’ Br. at 47. Instead, the Corps elects to rely on the substitute one-to-one mitigation ratio. By choosing to rely on one-to-one mitigation rather than attempting to replace lost function, the Corps implicitly but clearly concedes that the EPA’s RBP and the WVSCI did not provide for an adequate assessment of stream function. If these protocols had generated the requisite data for a sufficient functional assessment, the Corps could not have logically invoked the RGL’s one-to-one mitigation provisions. Since the Corps’ actions show that it does not believe its own contention that it has sufficiently assessed stream function, this court is under no obligation to believe it either.
II.
On the issue of the adequacy of the Corps’ proposed mitigation measures, the majority again errs in overlooking the plain language of the relevant regulations. The majority chooses to rely on internal Corps guidance documents that are inconsistent with, and must therefore yield to, the clear requirements of the regulations.
A.
The valley fills will bury more than 68,-000 feet of intermittent and ephemeral headwater streams. The Corps does not dispute that, absent mitigation measures, the adverse impacts of the proposed pro-*223jeets would be significant, and an Environmental Impact Statement (EIS) would therefore be required under NEPA. The Corps instead asserts that the mitigation measures it has approved are sufficient to reduce adverse impacts of the fills below the threshold of significance and avoid significant degradation of waters of the United States.4 The Corps bases these assertions on the anticipated success of a combination of stream creation and enhancement of existing stream channels. The Corps does not claim that the proposed mitigation will replace lost stream function; rather the mitigation will replace or enhance at least as many stream feet as the valley fills will bury.
To justify allowing the Corps to approve this so-called one-to-one mitigation in lieu of requiring the replacement of lost stream function, the majority relies on the language of a Memorandum of Agreement (MOA) between the Corps and the EPA and an internal Corps’ Regulatory Guidance Letter (RGL).
From a legal standpoint, the majority’s discussion of the MOA and RGL is largely beside the point. To the extent that the MOA and RGL are inconsistent with the plain language of the regulations, the regulations control. The MOA itself states that it “does not change the substantive requirements of the Guidelines [regulations]. It is intended to provide guidance regarding the exercise of discretion under the Guidelines.” J.A. 1165. Similarly, RGLs are “ ‘issued without notice and comment and do not purport to change or interpret the regulations applicable to the section 404 program ... [and] are not binding, either upon permit applicants or Corps District Engineers.” Northwest Bypass Group v. U.S. Army Corps of Eng’rs, 470 F.Supp.2d 30, 51 (D.N.H.2007) (quoting Envtl. Def. v. U.S. Army Corps of Eng’rs, No. 04-1575(JR), 2006 WL 1992626 at *7 (D.D.C. July 14, 2006)).
An analysis of the sufficiency of the mitigation in this case must begin with the provisions that are truly mandatory: those *224in the regulations. Compliance with 40 C.F.R. § 230.11(e) is non-discretionary. The Corps cannot issue a § 404(b)(1) permit without first assessing the “nature and degree of effect that the proposed discharge will have, both individually and cumulatively, on the structure and function of the aquatic ecosystem and organisms.” Thus, when confronted with a decision about the appropriate mitigation measures to require, the Corps should never find itself in a position where it has failed to sufficiently assess stream function; § 230.11(e) always requires the Corps to conduct this assessment. Considered in light of the clear requirement of § 230.11(e), the provision in the Corps’ RGL that purports to permit one-to-one mitigation where a sufficient stream functional assessment is not feasible can never be properly triggered. Simply put, the Corps cannot rely on an illegal provision in its RGL to justify a failure to mitigate for lost stream functions.
The majority’s analysis skirts the requirements of § 230.11(e). First, the majority observes that “a full functional assessment protocol is not yet available to the Corps.” Ante at 204. It then concludes that under the Corps’ RGL, “where a full functional assessment is not feasible, the only compensatory mitigation measure the Corps must require in a permitting decision is stream replacement on a one-to-one basis.” Ante at 204. The majority’s approach is unsupportable. Whatever it means to sufficiently assess stream function, it should mean the same thing under both § 230.11(e) of the regulations and the Corps’ MOA and RGL. To allow the Corps to interpret sufficiency of assessment differently for purposes of measuring function and determining appropriate mitigation wholly undermines the purposes of mitigation. Indeed, the MOA itself states that “[t]he determination of what level of mitigation constitutes ‘appropriate’ mitigation is based solely on the values and functions of the aquatic resource that will be impacted.” J.A. 1166. It is paradoxical to conclude that the (largely structural) assessments carried out as part of the Corps’ § 230.11(e) analysis adequately measured stream function and to simultaneously conclude that these same assessments provided insufficient data on stream function to require mitigation to replace lost function. The majority’s construction of the Corps’ functional assessment and mitigation requirements defeats the basic goal of the MOA, the RGL, and compensatory mitigation.
B.
Since the Corps cannot properly rely on its RGL to avoid mitigating for lost stream function, we would ordinarily look next at whether the mitigation measures required by the Corps will adequately replace lost function. This inquiry is premature in the present instance, however, because the Corps did not engage in the functional analysis required by 40 C.F.R. § 230.11(e). Thus, even if we credited the Corps’ almost wholly unsubstantiated assertion that the new stream creation projects required in the CDDs will create working streams, the CMPs and CDDs offer no guarantees that the newly created streams will replace lost headwater stream functions the Corps has failed to quantify. The Corps therefore acted in direct contravention of the applicable regulations.5
*225Furthermore, even under the majority’s construction of the Corps’ mitigation duties (allowing for one-to-one mitigation where sufficient functional assessment is not feasible), there is good reason based on the record before us to question whether the mitigation will prevent significant degradation of waters of the United States. As the majority concedes, ante at 205, the Corps offers virtually no scientific support for the viability of creating working streams from scratch, particularly headwa-ter streams. The Corps provides evidence of a single successful stream creation project in Kentucky, but this was not a head-water stream. Here, the bulk of the proposed stream creation in the mitigation plans is to take place in the sediment ditches on the valley fills, where the former headwater streams were located. According to the Draft EIS for Mountaintop Removal Mining and Valley Fills, a document jointly authored by the Corps and other agencies:
to date functioning headwater streams have not been re-created on mined or filled areas as part of mine restoration or planned stream mitigation efforts. Most on-site mitigation construction projects have resulted in the creation of palustrine wetlands that resembled ponds.
J.A. 862. The Draft EIS adds that “it is not known whether the organic matter processing that occurs in created wetlands would mimic the processing found in a natural stream system.” J.A. 863.
In addition to the lack of evidence about the viability of stream creation, the U.S. Fish and Wildlife Service (USFWS) and its West Virginia Field Office (WVFO) submitted a joint comment on the Laxare East permit expressing a continued belief that it is not possible to fully replace the critical aquatic and terrestrial ecosystem functions of healthy headwater streams. These agencies also commented that they were unaware of any scientific support for the concept that on-bench sediment ditches can be considered biologically equivalent to, or even rough approximations of, flowing streams. These comments undercut the Corps’ contention that the mitigation will produce its advertised results.
In sum, the regulations in 40 C.F.R. Part 230 do not allow the Corps to engage in one-to-one mitigation in this case to claim it has achieved no significant degradation of waters of the United States. Even if they did, the Corps has not provided sufficient evidence to allow this court to conclude that the impacts of the fills will truly be insignificant.
C.
Pursuant to NEPA an agency engaging in a major federal action may decline to issue a comprehensive environmental impact statement (EIS) only if, after mitigation measures are accounted for, the agency concludes the action (or project) will result in no significant adverse impact to the human environment. See 42 U.S.C. § 4332(2)(C), 40 C.F.R. § 1501.4(e). Because, as I have explained, the Corps has failed to establish that the permitted valley fill projects will not significantly degrade the waters of the United States, I must also conclude that the Corps has likewise failed to establish that the projects will have no significant adverse environmental impact. Consequently, the Corps has not justified its decision to decline to issue an *226EIS for the fill projects, and NEPA’s requirements have not been satisfied.
III.
Today’s decision will have far-reaching consequences for the environment of Appalachia. It is not disputed that the impact of filling valleys and headwater streams is irreversible or that headwater streams provide crucial ecosystem functions. Further, the cumulative effects of the permitted fill activities on local streams and watersheds are considerable. By failing to require the Corps to undertake a meaningful assessment of the functions of the aquatic resources being destroyed and by allowing the Corps to proceed instead with a one-to-one mitigation that takes no account of lost stream function, this court risks significant harm to the affected watersheds and water resources. We should rescind the four permits at issue in this case until the Corps complies with the clear mandates of the regulations. First, the Corps must adequately determine the effect that the valley fills will have on the function of the aquatic ecosystem. Second, based on this determination, the Corps must certify that the fills, after mitigation is taken into account, will result in no significant degradation of waters of the United States and no significant adverse impact to the human environment.
. The majority contends that I have not afforded the Corps the appropriate level of deference in interpreting its regulations. Ante at 199 n. 16. The majority does acknowledge, however, that the Corps may not proffer an interpretation that is "plainly erroneous or inconsistent with the regulation.” Auer v. Robbins, 519 U.S. 452, 461, 117 S.Ct. 905, 137 L.Ed.2d 79 (1997). Of course, "[t]he agency’s interpretation need not be the best or most natural one by grammatical or other standards. Rather, it need only be a reasonable construction of the regulatory language.” Dist. Mem. 7 Hosp. of Southwestern North Carolina, Inc. v. Thompson, 364 F.3d 513, 518 (4th Cir.2004) (internal quotations and citations removed).
As I have explained, the Corps’ contention that a structural assessment may substitute for a functional one is wholly inconsistent with § 230.1 l(e)’s clear mandate that the Corps assess both structure and function. Further, in light of the Corps’ failure to define "function” for purposes of § 230.11(e), its unsubstantiated assertion that the structural assessments used in this case provide ade*219quate information about stream function cannot be upheld as a reasonable construction of its regulations. See Dist. Mem’l Hosp., 364 F.3d at 518.
. The majority points out that any attempt by this court "to define stream function beyond these guidelines [that is, the language in § 230.11(e) itself] would certainly be inappropriate judicial intrusion into the Corps and EPA’s sphere of authority.” Ante at 200 n. 19. I do not disagree. The majority’s observation underscores precisely why the proper judicial resolution of this issue would have been to conclude that the Corps abused its discretion in failing to offer an interpretation of the term "function” as used in § 230.11(e) that is not "plainly erroneous or inconsistent with the regulation.” Auer, 519 U.S. at 461, 117 S.Ct. 905. So long as the Corps declines to provide a construction that gives meaning to the operative terms in its regulations, this court cannot uphold the Corps’ actions.
I have engaged in what the majority calls an "inappropriate judicial intrusion” only to point out the shortcomings of the majority’s analysis. The majority has determined that the Corps did not abuse its discretion in approving the stream assessments conducted in this case. The majority upholds the Corps in spite of the agency’s unwillingness to define the term "function.” Any evaluation of the majority’s (or the Corps’) determination that the assessment of stream function was adequate, however, necessarily requires giving some meaning to the term "function.”
According to the majority, ante at 200 n. 19, "the only clues § 230.11(e) offers regarding the stream functions to be measured are the ... factors” listed in one sentence in § 230.11(e). The sentence provides in full: "Consideration shall be given to the effect at the proposed disposal site of potential changes in substrate characteristics and elevation, water or substrate chemistry, nutrients, currents, circulation, fluctuation, and salinity, on the recolonization and existence of indigenous aquatic organisms and communities.” 40 C.F.R. § 230.11(e) (2006). The majority mistakes the significance of these "factors.” They are not stream functions. The term "function” refers to the "role, duty, work” or "purpose” of a thing. Webster's Third New Int’l Dictionary 920 (2002). The factors listed in § 230.11(e) are merely stream characteristics (or attributes), and § 230.11(e) requires the Corps to analyze the effects that changes to these stream characteristics will have on the "recolonization and existence of indigenous aquatic organisms and communities.” Compliance with this sentence cannot be determinative of whether the Corps has adequately analyzed the effects of the proposed fills on stream function.
Since the factors identified by the majority cannot constitute the stream functions contemplated by § 230.11(e), there is no reason to discount the relevance of the proposed functional protocol currently under development for the Huntington District of West Virginia. Until the Corps itself identifies a specific set of functions to be measured, nothing could be more relevant to determining the meaning of “function” in § 230.11(e) than the list of the stream functions proposed to be measured in future § 230.11(e) functional analyses. Consequently, nothing could be more useful in reviewing the adequacy of the assessment protocols used in this case than this proposed list of stream functions. The Corps should not be rewarded for its recalcitrance in defining function and for its lengthy delay in developing a usable functional assessment protocol for West Virginia.
. If we pieced together the stream functions that the Corps itself identifies in the Combined Decision Documents (CDDs) in this case, the list is, if anything, more inclusive than the preliminary list proposed to be measured by the EPA. Specifically, the Corps in the Black Castle CDD states that: "Some important functions of ... headwater streams include the maintenance of natural discharge regimes, the regulation of sediment export, the retention of nutrients, the processing of *221terrestrial organic matter, and the exportation of water nutrients and organic matter to downstream areas.” J.A. 1823; see also, Camp Branch CDD, J.A. 1319 (same language). The Corps also observes in its § 230.11(e) analysis in the Republic No. 2 CDD that nutrient cycling, organic matter dynamics, respiration, and primary and secondary production are functions "typically found in scientific research analysis.” J.A. 3570.
. The majority quotes 40 C.F.R. § 230.10(d)— providing that a § 404 permit cannot issue "unless appropriate and practicable steps have been taken which will minimize potential adverse impacts of the discharge [of fill material] on the aquatic ecosystem” — but does not mention other relevant provisions of section 230.10 that should also inform its analysis. See ante at 201-02.
Read in its entirety, 40 C.F.R. § 230.10 requires that a permitted dredge or fill activity not only include appropriate and practicable steps to minimize potential adverse impacts, but also that the discharge result in no significant degradation of waters of the United States, taking into account required mitigation.
Section 230.10 describes four restrictions on discharge, all of which must be met before a dredge or fill permit can issue. Section 230.10(c) provides that "no discharge of dredged or fill material shall be permitted which will cause or contribute to significant degradation of the waters of the United States.” 40 C.F.R. § 230.10(c) (2006).
Section 230.12(a)(2) (2006) allows the Corps to permit a dredge or fill activity as “complying with the requirements of these Guidelines with the inclusion of appropriate and practicable discharge conditions (see Subpart H) to minimize pollution or adverse effects to the affected aquatic ecosystems.” But paragraph (a)(3) makes clear that a project must be "[sjpecified as failing to comply with the requirements of these Guidelines where ... (ii)[t]he proposed discharge will result in significant degradation of the aquatic ecosystem under § 230.10(b) or (c)_”
Consequently, when § 230.10 is taken as a whole, it is apparent that for a discharge of dredge or fill material into waters of the United States to be permitted, not only must appropriate steps have been taken to minimize adverse impacts, but the Corps must also find that, once appropriate mitigation measures are accounted for, the discharge will not significantly degrade the waters of the United States.
. The Corps does in various places make claims that the mitigation it requires will replace lost stream functions. However, the functions that the Corps claims will be replaced — movement of water, movement of sediment, nutrient cycling, and organic matter retention — only partially coincide with the headwater functions that the Corps admits will be lost under the valley fills, namely, maintenance of natural discharge regimes, *225regulation of sediment export, retention of nutrients, processing of terrestrial organic matter, and exportation of water nutrients and organic matter to downstream areas. Further, the Corps has offered virtually no evidence to support its claim that the mitigation measures will replace lost stream functions.