ADKINS, J.

FACTS AND LEGAL PROCEEDINGS

Maryland Department of the Environment ("MDE") issued municipal separate storm sewer system ("MS4") discharge permits ("the Permits") to Anne Arundel County, Baltimore City, Baltimore County, Montgomery County, and Prince George's County ("the Counties"). Multiple organizations argue that the Permits do not comply with federal and state law, and they request that we remand for MDE to correct these legal errors.1

Federal Framework: NPDES Permits and Municipal Stormwater Discharges

Under the Clean Water Act ("CWA"), the discharge of pollutants is illegal. 33 U.S.C. § 1311. Through the National Pollution Discharge Elimination System ("NPDES"),2 33 U.S.C. § 1342, either the Environmental Protection Agency ("EPA") or an EPA-approved state, such as Maryland, may issue permits exempting a discharger from this prohibition. See Piney Run Pres. Ass'n v. Cnty. Comm'rs of Carroll Cnty., Md., 268 F.3d 255, 265 (4th Cir.2001). MDE is the authority in Maryland that administers the NPDES program. Code of Maryland Regulations ("COMAR") 26.08.04.07. An NPDES permit, however, does not give a discharger carte blanche. "Generally speaking, the NPDES requires dischargers to obtain permits that place limits on the type and quantity of pollutants that can be released into the Nation's waters." S. Fla. Water Mgmt. Dist. v. Miccosukee Tribe, 541 U.S. 95, 102, 124 S.Ct. 1537, 158 L.Ed.2d 264 (2004). These limits are called effluent limitations. See 33 U.S.C. § 1362(11) (defining an effluent limitation as "any restriction established by a State or the Administrator on quantities, rates, and concentrations of chemical, physical, biological, and other constituents which are discharged from point sources into navigable waters, the waters of the contiguous zone, or the ocean, including schedules of compliance"). The type of discharge determines the type of limitations the permit must impose on the discharger.

The Permits before us control stormwater pollutant discharge.3 Stormwater consists of the rain and snowmelt that filters through the soil and courses over surfaces — collecting pollutants along the way — before passing through the municipal storm sewer systems4 and into waterbodies. During the development of the Permits, the Water Groups explained the problems that stormwater poses, whether to the surface conditions of Maryland's waters, for humans who recreate and subsist on them, and for wildlife who live in them. See Letter with Comments on Draft MS4 Permit for Baltimore City from Blue Water Baltimore, Inc. and Earthjustice, to Brian Clevenger, MDE (Sept. 21, 2012). In recognition of extensive public commentary on the severity of the problems associated with stormwater, MDE stated: "[i]t becomes fairly easy for all organizations, individuals, and government agencies to agree that urban stormwater is a problem that must be addressed." MDE, Response to Formal Comments for Montgomery County NPDES Permit (2009).

Nevertheless, municipal stormwater discharge is "highly intermittent," "usually characterized by very high flows occurring over relatively short time intervals," and "depend[s] on the activities occurring on the lands." See National Pollutant Discharge Elimination System Permit Application Regulations for Storm Water Discharges, 55 Fed.Reg. 47,990, 48,038 (Nov. 16, 1990) (codified at 40 C.F.R. § 122.26). It is also difficult to discern the amount of pollutant that any one discharger contributes to a waterbody because municipalities have so many outfalls, or discharge points, leading into the waters. See MDE, Montgomery County NPDES Permit Fact Sheet (900 outfalls); MDE, Anne Arundel County NPDES Permit Fact Sheet (nearly 1,000 outfalls); MDE, Baltimore County NPDES Permit Fact Sheet (nearly 700 outfalls.); MDE, Prince George's County NPDES Permit Fact Sheet (more than 4,000 outfalls); MDE, Baltimore City NPDES Permit Fact Sheet (around 350 outfalls.); see also 40 C.F.R. § 122.26(b)(5), (9) (outlining minimum diameters of pipes in major MS4 outfalls).5

Because of the nature of municipal stormwater discharges, Congress adopted a flexible approach to the control of pollutants in MS4s. See 55 Fed.Reg. at 48,038 (The Congressional Record from 1986 stated not only that "an end-of-the-pipe treatment technology is not appropriate for [the MS4] discharge" but also that "[MS4] controls may be different in different permits.").6 Pursuant to 33 U.S.C. § 1342(p)(3)(B)(iii), municipal stormwater permits "shall require controls to reduce the discharge of pollutants to the maximum extent practicable, including management practices, control techniques and system, design and engineering methods...."7

Best management practices ("BMPs") have been a long-standing control or effluent limitation8 in MS4 permits. See 40 C.F.R. § 122.44(k)(2) (BMPs "control or abate the discharge of pollutants when [a]uthorized under [33 U.S.C. § 1342(p)]"); id. § 122.44(k)(3) (BMPs are an appropriate control when "[n]umeric effluent limitations are infeasible"); see also Tualatin Riverkeepers v. Or. Dep't of Envtl. Quality, 235 Or.App. 132, 230 P.3d 559, 564 (2010) ("`Best management practices,' such as those incorporated in the permits at issue in this case, are a type of effluent limitation."). The EPA defined BMPs to mean "schedules of activities, prohibitions of practices, maintenance procedures, and other management practices to prevent or reduce the pollution of `waters of the United States.'" 40 C.F.R. § 122.2; cf. Natural Res. Def. Council, Inc. v. EPA, 808 F.3d 556, 579 (2d Cir. 2015) ("But EPA's narrative WQBEL [water-quality based effluent limitation] does not qualify as a BMP, as it is neither a practice nor a procedure."). Examples of the types of BMPs the Counties might implement pursuant to the Permits are infiltration practices and green roofs.9

Through guidance memos, the EPA has endorsed the use of BMPs in MS4s for decades but has increasingly recommended that, where feasible, such permits include numeric effluent limitations. Interim Permitting Approach for Water-Quality Based Effluent Limitations in Storm Water Permits, 61 Fed.Reg. 43,761 (1996); EPA, Memorandum on Establishing Total Maximum Daily Load (TMDL) Wasteload Allocations (WLAs) for Storm Water Sources and NPDES Permit Requirements Based on Those WLAs § 3 (2002) [hereinafter "2002 Memo"]; EPA, Memorandum on Revisions to the November 22, 2002 Memorandum 4-5 (2010) [hereinafter "2010 Memo"]; EPA, Memorandum on Revisions to the November 22, 2002 Memorandum at 4 n. 5 (2014) [hereinafter "2014 Memo"]. (A "numeric" effluent limitation "refers to [a] limitation[ ] with a quantifiable or measurable parameter related to a pollutant (or pollutants).").

Total Maximum Daily Loads ("TMDLs")

The concept of total maximum daily load ("TMDL") looms large in this case.10 We begin by setting forth its basic purpose, then unpacking its complex formation.

TMDLs inform. See Am. Farm Bureau Fed'n v. EPA, 792 F.3d 281, 291 (3d Cir. 2015) ("Our understanding of [TMDLs] as informational tools is supported by every case and piece of scholarship to consider them as well as the language of the Chesapeake Bay TMDL itself."); see also EPA, Chesapeake Bay TMDL § 1.41, at 1-15 (2010) ("TMDLs are `primarily informational tools' that `serve as a link in an implementation chain....'"), available at http://www.epa.gov/chesapeake-bay-tmdl/ chesapeake-bay-tmdl-document [https:// perma.cc/9R7V-9VHV].

TMDLs arise out of a multi-step process that begins with the establishment of water quality standards ("WQS"). See Am. Farm Bureau Fed'n, 792 F.3d at 289 ("TMDLs happen after a state enacts pursuant to its law (but required by the Clean Water Act) `water quality standards.'"). Because the EPA and the states interact throughout this process, it has been described as one of "cooperative federalism." Id. at 288; see Anacostia Riverkeeper, Inc. v. Jackson, 798 F.Supp.2d 210, 214-17 (D.D.C.2011).

Water quality standards, as the term itself suggests, protect water quality. 40 C.F.R. § 130.2(d); COMAR 26.08.02.01(A). Each state must set water quality standards by assigning a "use" to a water, such as recreation or fishing, then developing criteria to protect those uses, as well as ensuring that higher quality waters do not degrade to the minimally accepted standard (also known as an anti-degradation policy). 33 U.S.C. § 1313; COMAR 26.08.02.01(B)(1). All water quality standards are subject to EPA review, and if the EPA does not approve of them, the EPA will set those standards itself. 33 U.S.C. § 1313.

By way of example, the EPA approved a TMDL that MDE submitted for fecal bacteria for the Non-tidal Cabin John Creek Basin in Montgomery County in 2007. See MDE, Total Maximum Daily Loads of Fecal Bacteria for the Non-tidal Cabin John Creek Basin in Montgomery County, Maryland (Document version: Oct. 13, 2006) [hereinafter "John Creek Basin TMDL"]. These bacteria are microscopic organisms in animal waste. Fecal bacteria in water can raise the risk of illness in humans who recreate there. Id. at § 1.0, at 1. To develop the WQS, MDE selected "water contact recreation and protection of aquatic life and public water supply" as the use of the water and 126 MPN11 per 100 milliliters as the criteria. Id. at § 2.3, at 11 (citing COMAR 26.08.02.08(O) and 26.08.02.03-3). This figure (126) represents a mean density for this pollutant. Id.; see 40 C.F.R. § 130.2(i) ("[M]ass per time, toxicity, or [an]other appropriate measure" may be used to express TMDLs.).

After setting WQSs, the states establish effluent limitations in permits as the primary way to meet the WQSs because, as we have explained, effluent limitations restrict the discharge of pollutants. See 33 U.S.C. § 1362(11). Nevertheless, we note, importantly, that MS4s are not subject to the requirement of imposing effluent limitations "necessary to meet water quality standards." See 33 U.S.C. § 1311(b)(1)(C); see also Defenders of Wildlife v. Browner, 191 F.3d 1159, 1165 (9th Cir.1999); cf. 33 U.S.C. § 1342(p)(3)(A) (Industrial dischargers must comply with 33 U.S.C. § 1311.).12 This important point notwithstanding, Congress requires that "[e]ach State shall identify those waters within its boundaries for which the effluent limitations required by [33 U.S.C. § 1311] are not stringent enough to implement any water quality standard applicable to such waters." 33 U.S.C. § 1313(d)(1)(A).

This is where the TMDL comes into play. The TMDL tells a state what is the threshold amount of a pollutant that a body of water can tolerate before violating the WQS. See In re City of Moscow, Idaho, 10 E.A.D. 135, 2001 WL 988721, at *4 (EAB July 27, 2001) ("A TMDL is a measure of the total amount of a pollutant from point sources, nonpoint sources and natural background, that a water quality limited segment can tolerate without violating the applicable water quality standards."); EPA, Chesapeake Bay TMDL § 1.1, at 1-2 ("A TMDL specifies the maximum amount of a pollutant that a waterbody can receive and still meet applicable WQS.").13

States must establish TMDLs "at a level necessary to implement the applicable water quality standards," 33 U.S.C. § 1313(d)(1)(C), when they identify those waters for which effluent limitations cannot implement the WQSs, 33 U.S.C. § 1313(d)(1)(A).14 As with water quality standards, the states have the obligation of setting TMDLs and submitting them to the EPA for approval. See supra MDE, John Creek Basin TMDL (The EPA approved of MDE's TMDL in March 2007.). If the EPA disapproves of the TMDLs, the EPA will set them itself. 33 U.S.C. § 1313(d)(2).

For this case, wasteload allocations ("WLAs") are the most critical part of the TMDL equation. See 40 C.F.R. § 130.2(i) (A TMDL is "[t]he sum of the individual WLAs for point sources and LAs [load allocations] for nonpoint sources and natural background."). The WLA represents a water's "loading capacity" assigned to its "point sources of pollution." Id. § 130.2(h). Continuing with our example, MDE set the TMDL for fecal bacteria at 176.36 billion MPN/day, the LA at 68.17 billion MPN/day, and the WLA at 108.19 billion MPN/day. MDE, John Creek Basin TMDL.15

Although TMDLs are informational tools, of which WLAs are a part, WLAs are more akin to restrictions. See Am. Farm Bureau Fed'n v. EPA, 984 F.Supp.2d 289, 328 (M.D.Pa.2013) ("WLAs are not permit limits per se; rather they still require translation into permit limits....") (citation omitted) (internal quotation marks omitted) (emphasis in original), aff'd, 792 F.3d 281 (3d Cir.2015). Under 40 C.F.R. § 122.44(d)(1)(vii)(B), permitting authorities must ensure that effluent limitations "are consistent with the assumptions and requirements" of any approved WLA.16

We conclude our introduction of TMDLs by noting that MS4s are subject to the MEP standard under 33 U.S.C. § 1342. MS4s are not, however, required to impose effluent limitations necessary to meet water quality standards. The CWA still requires Maryland to set water quality standards and TMDLs — subject to the EPA's approval. Flowing from this obligation is the requirement that MS4s are subject to effluent limitations that are consistent with WLAs of EPA-approved TMDLs.

The Chesapeake Bay TMDL

As we will discuss in more detail, the Permits require the Counties to take actions to make progress in meeting the WLAs of many EPA-approved TMDLs.17 By far, though, the most critical TMDL in this case is the Chesapeake Bay TMDL ("Bay TMDL").

Regarded as a national treasure,18 the Chesapeake Bay is the largest estuary in the United States, a product of flooding from the Susquehanna River over thousands of years. Alice Jane Lippson, The Chesapeake Bay in Maryland: An Atlas of Natural Resources 2 (Johns Hopkins University Press 1973). Over 2,000 species of animals and plants reside in the Chesapeake Bay. Alice Jane Lippson & Robert L. Lippson, Life in the Chesapeake Bay viii (Johns Hopkins University Press 1984). These include phytoplankton, the blue crab, and striped bass, among many, many others. Lippson, The Chesapeake Bay in Maryland, supra at 14, 26, 36. In addition to housing much wildlife, the Chesapeake Bay is a shipping and commerce hub and a source of recreation. Chesapeake Bay Program, Chesapeake Bay: Introduction to an Ecosystem 2 (2004).

Human activity, however, threatens this complex ecosystem. "Excess sediment and nutrients endanger the Bay's water quality." Id. at 3-4. Such threats include: depriving species of oxygen; delivering chemicals which collect in animal tissue; and even destroying habitats because sunlight cannot reach critical underwater grasses where species reside. Id.

There is, then, no underestimating the importance of the restoration of the Chesapeake Bay in Maryland. See Am. Farm Bureau, 984 F.Supp.2d at 298 ("[The Bay] has been described as one of the most biologically productive ecosystems in the world," and, along with its watersheds, "add[s] ecological, economic, recreational, historic, and cultural value to the region.").

How to restore the Bay, however, has been a prolonged, frustrated process. See id. (The Bay TMDL "is not a new or recent idea," and thus, "it would be improper to view the Final TMDL in a vacuum as a single, isolated effort to restore water quality to the Chesapeake Bay."). Some of these restoration efforts include the Chesapeake Bay Agreement in 1980, another agreement in 1987, amendments to the agreement in 1992, and the Chesapeake 2000 Agreement. Department of Legislative Services, Office of Policy Analysis, Chesapeake Bay Restoration and the Tributary Strategy: An Analysis of Maryland's Efforts to Meet the Nutrient and Sediment Reduction Goals of the Chesapeake 2000 Agreement 3-4 (2007).

The EPA established the Bay TMDL in December 2010. See Am. Farm Bureau Fed'n, 792 F.3d at 290 ("As noted, for the Chesapeake Bay the relevant states and the EPA agreed that the EPA would draft the TMDL in the first instance."). It has survived legal challenges before the U.S. District Court for the Middle District of Pennsylvania as well as the Third Circuit.19 See Am. Farm Bureau Fed'n, 984 F.Supp.2d at 294; Am. Farm Bureau Fed'n, 792 F.3d at 287. These courts have noted that the efforts to restore the Chesapeake Bay extend back decades, and that the development of the Bay TMDL itself has been a decade-long process. Am. Farm Bureau Fed'n, 984 F.Supp.2d at 299; Am. Farm Bureau Fed'n, 792 F.3d at 291.

The Bay TMDL provides information pertaining to pollution reduction for nitrogen, phosphorus and sediment in the Chesapeake Bay and applies to the District of Columbia and the six "Bay" states, including Maryland. EPA, Chesapeake Bay TMDL at ES-1.20

Modeling

Before delving into Maryland's role in the formation of the Bay TMDL, we must discuss the "critical and valuable" role that modeling played in the Bay TMDL's development. EPA, Chesapeake Bay TMDL at ES-5.21 "Modeling is an approach that uses observed and simulated data to replicate what is occurring in the environment to make future predictions." Id. "A model `is an abstraction from and simplification of the real world.'" Am. Farm Bureau, 984 F.Supp.2d at 340 (citation omitted). Models are essential when one seeks to study "ecosystems that are too large or complex for real-world monitoring," such as the Chesapeake Bay and its watersheds. Chesapeake Bay Program, About the Bay Program: Modeling, available at http:// www.chesapeakebay.net/about/programs/ modeling [https://perma.cc/RDM4-C4XF] (last visited Feb. 9, 2016) [hereinafter CBP: Modeling].

A prominent component in the modeling of the Bay TMDL was the Phase 5.3 Chesapeake Bay Watershed Model ["Phase 5.3 Model"]. EPA, Chesapeake Bay TMDL, at 5-19.22 "The Phase 5.3 Model is the most recent of a series of increasingly refined versions of the Chesapeake Bay Watershed Model." EPA, Chesapeake Bay Phase 5.3 Community Watershed Model § 1.2.1, at 1-13 (2010), available at http://www.chesapeakebay. net/about/programs/modeling/53/ [https:// perma.cc/9YKS-WBZ4].23 The Phase 5.3 Model simulates the "loading and transport of nitrogen, phosphorus, and sediment from pollutant sources throughout the Bay watershed." EPA, Chesapeake Bay TMDL, at 5-20. Additionally, this model provides "estimates of watershed nitrogen, phosphorus, and sediment loads resulting from various management scenarios." Id.

Because models are not "perfect forecasts," however, modeling is "part of a broader toolkit," including monitoring, "to gain the highest possible level of accuracy." CBP: Modeling. As the EPA explained: "The Bay modeling framework takes advantage of decades of atmospheric deposition, streamflow, precipitation, water quality, biological resource, and land cover monitoring data" as well as "tracking and reporting of the implementation of pollution load reduction best management practices." EPA, Chesapeake Bay TMDL, at § 5.1, 5-1-5-2. These resources allowed the EPA to calibrate its models. Id.

Because the Bay TMDL exists in significant part as a result of modeling, and because of how prevalent modeling is in TMDL formulation, MDE incorporated modeling into the Permits.

MDE incorporated by reference a document the agency published, called Accounting for Stormwater Wasteload Allocations and Impervious Acres Treated, Guidance for National Pollutant Discharge Elimination System Stormwater Permits ("the Guidance").24 As the name suggests, the document serves dual purposes: the Counties can assess progress in achieving WLAs and also assess restoration of impervious surface areas through a credits-to-acres approach.25 In the Guidance, MDE sets forth acceptable models that the Counties can use, including, Maryland's Assessment and Scenario Tool ("MAST").26 MDE, Guidance at 2. The Guidance includes the pollutant rates for the Bay TMDLs — Total Nitrogen ("TN"), Total Phosphorus ("TP"), and Total Suspended Sediment ("TSS") — and requires that the Counties use these pollutant rates together with land use data to calculate baseline stormwater loads. Id. at 2-3. As the document explains, "[t]hese pollutant loads are specific to the [] Bay TMDL." Id. at 2. But the Counties may use the principles and methods in the Guidance "for any EPA approved TMDL." Id. at 1.

Maryland's Watershed Implementation Plan ("WIP")

The EPA developed the Bay TMDL to ensure that the Bay jurisdictions would put in place "all pollution control measures needed to fully restore the Bay and its tidal rivers" by 2025. EPA, Chesapeake Bay TMDL at ES-1. The EPA approved the Bay TMDL "only after" determining that each jurisdiction provided "reasonable assurance" that it would meet established pollutant reductions. Am. Farm Bureau Fed'n, 792 F.3d at 291. The Bay jurisdictions set forth their strategies for meeting pollutant reductions in Watershed Implementation Plans ("WIPs"). Id.

WIPs are "roadmaps" setting forth a plan for how and when a jurisdiction will reach the pollution reduction goals in the Bay TMDL. EPA, Chesapeake Bay TMDL, at ES-8. The EPA described these roadmaps as the "cornerstone" that ensured the States were accountable in achieving pollution reductions. Id. Notably, the EPA expressed no concerns about Maryland's Final WIP,27 whereas the EPA had to implement backstop allocations and adjustments in other Bay jurisdictions so that the EPA had reasonable assurance that all jurisdictions would achieve necessary reductions. See EPA, Chesapeake Bay TMDL Executive Summary, ES-10-ES-13 (Dec. 29, 2010) [hereinafter "Bay TMDL ES"].28

Maryland's WIP lists restoration of "twenty percent of the counties' impervious surface area that is not already restored to the maximum extent practicable (MEP)" in the "key elements" supporting the reasonable assurance of the implementation of the WIP. Phase I WIP at 5-30. The elements also include the adaptive management approach whereby additional or alternative practices are implemented if existing programs are not meeting target reductions. Id.29

Maryland's Stormwater Management History

In addition to an explication of the federal permitting system, NPDES, and the complex components arising out of it, such as TMDLs, we also set forth Maryland's stormwater management program, which has evolved since its inception in the 1980s, and which is informative for purposes of analyzing the Permits.

In 1982, the General Assembly enacted laws "to reduce as nearly as possible the adverse effects of stormwater runoff." Maryland Code (1982, 2007 Repl.Vol.), § 4-201 of the Environment Article ("EN"); see H.B. 1091, 1982 Gen. Assemb. Reg. Sess. (Md.1982). As a result, each county and municipality in Maryland was required for the first time to "adopt ordinances necessary to implement a stormwater management program" by July 1, 1984. See EN § 4-202. Then authorized by the General Assembly, the Department of Natural Resources issued regulations setting forth minimum control requirements and design criteria for the counties and municipalities. See 10 Md. Reg. 881, 884-85 (May 13, 1983) (to be codified at COMAR 08.05.05).30 The regulations fostered the "primary goal" of "maintain[ing] after development, as nearly as possible, the predevelopment runoff characteristics" of the land. See EN § 4-203(b)(1); see also EN § 4-204(a) (Development of land is prohibited without submitting a stormwater management plan and obtaining the municipality's or county's approval of the plan.)

Maryland entered a new phase of stormwater management in the early 2000s. Pursuant to EN § 4-203(b), MDE adopted regulations to "rectify the[] programmatic shortcomings" of then-existing regulations that had provided "sparse guidance" on "water quality enhancement." 27 Md. Reg. 1167, 1168 (June 16, 2000) (to be codified at COMAR 26.07.02). Amending the stormwater regulations, MDE intended to "provide water quality treatment of up to 90 percent of the average annual rainfall throughout the State, establish ground water recharge standards, and out-line a channel erosion control strategy," as well as "promote environmentally friendly site design." Id. To fulfill this purpose, MDE incorporated by reference the 2000 Maryland Stormwater Design Manual ("the Manual"). Id. at 1167, 1169.31 MDE required the counties and municipalities to revise their ordinances to incorporate the Manual's policies and practices by July 1, 2001. Id. at 1170.

The Manual "provide[d] designers a general overview on how to size, design, select and locate BMPs at a new development site to comply with State stormwater performance standards." Center for Watershed Protection ("CWP") & MDE, Manual, § 1.3, at 1.16. There are 14 performance standards, including the water quality volume standard ("WQv"). Id. § 1.2, at 1.13.

Another stormwater management phase began when the General Assembly required MDE to mandate the use of environmental site design ("ESD") in 2007. H.B. 786, Gen. Assemb. Reg. Sess. (Md. 2007). ESD is best understood as those practices, such as "small-scale stormwater management practices, nonstructural techniques, and better site planning," that "mimic natural hydrologic runoff characteristics and minimize the impact of land development on water resources." EN § 4-201.1(b); see, e.g., note 9 (green roofs). MDE implemented regulations to this effect and explained that "[t]he goal of the regulations is to maintain after development as nearly as possible, the predevelopment runoff characteristics of the site being developed using ESD to the MEP." 35 Md. Reg. 2191 (Dec. 5, 2008) (to be codified at COMAR 26.17.02).

The Permits

MDE issued several series of MS4 permits to the Counties that preceded the Permits before us today. See MDE, NPDES MS4 Permit Montgomery County Fact Sheet (2008) (The first two permits were issued in 1996 and 2001.); MDE, Basis for Final Determination to Issue Prince George's County's NPDES MS4 Permit (2013) (The first three permits were issued in 1993, 1999, and 2004.); MDE, Basis for Final Determination to Issue Baltimore County's NPDES MS4 Permit (2013) (The first three permits were issued in 1994, 2000, and 2005.); MDE, Basis for Final Determination to Issue Anne Arundel County's NPDES MS4 Permit (2013) (The first three permits were issued in 1993, 1999, and 2004.); MDE, Basis for Final Determination to Issue Baltimore City's NPDES MS4 Permit (2013) (The first three permits were issued in 1993, 1999, and 2005.).32

At issue here are five-year term Permits MDE most recently issued: to Montgomery County in February 2010, to Baltimore County in December 2014, to Baltimore City in December 2013, to Prince George's County in January 2014, and to Anne Arundel County in February 2014.

The Water Groups challenge the Permits in several respects, namely, (1) the requirement to restore impervious surface area, (2) the requirement to submit plans for TMDLs, (3) the monitoring requirements, and (4) the public's ability to participate in the development of the Permits.

Some of these provisions are new and therefore represent an increase in responsibility on the Counties to maintain and improve the quality of their waters. See, e.g., MDE, Basis for Final Determination to Issue Anne Arundel County's NPDES MS4 Permit (2013) ("These meetings resulted in the addition of more stringent conditions to Anne Arundel County's stormwater permit, in large part due to a regional and growing focus on restoring Chesapeake Bay."); id. ("New requirements in the permit will include ... developing restoration plans to meet stormwater WLAs for impaired waters."); MDE, NPDES Montgomery County Stormwater Permit Response to Formal Comments at 2 (2009) ("MDE believes that this current municipal stormwater permit will force Montgomery County to make major strides toward controlling urban runoff better than ever before. New conditions such as trash abatement jurisdiction-wide and requiring an additional 20% of the County's impervious area to be restored are major additions.");33 MDE, Maryland's 2006 TMDL Implementation Guidance for Local Governments i (2006) ("Until recently, Maryland has focused primarily on TMDL development, which establishes limits on pollutant loads. Now the State is moving into the implementation phase....").

Before discussing these Permit provisions, we note additional Permit requirements that illustrate the breadth of the Counties' obligations. The Counties must implement management programs "to control stormwater discharges to the maximum extent practicable." These programs include a stormwater management program ("SWMP") and an erosion and sediment control program in accordance with state law; an "illicit discharge detection and elimination" program; requirements to reduce trash; obligations on the Counties to reduce pollutants associated with maintenance activities and on municipal facilities to submit pollution prevention plans; as well as a requirement to engage in public outreach activities to reduce stormwater pollution.

The Permits also require the Counties to engage in thorough analyses of the water quality of their watersheds. Among other things, the watershed assessments oblige the Counties to identify and prioritize water quality improvement projects.

MDE has also ensured that the Counties cannot use lack of adequate funding as a defense for failure to comply with Permit terms. The Permits explain that "[l]ack of funding does not constitute a justification for noncompliance with the terms of this permit."34 To this end, the General Assembly enacted EN § 4-202.1 in 2012, requiring the Counties "to adopt local laws or ordinances necessary to establish an annual stormwater remediation fee and a local watershed protection and restoration fund to provide financial assistance for the implementation of local stormwater management plans." H.B. 987, 2012 Gen. Assemb. Reg. Sess. (Md.2012); see also EN § 4-202.1(a)(1) (This "section applies to a county or municipality that is subject to a [NPDES Phase I MS4 permit].").35 MDE had investigated the costs of meeting the Bay TMDL and commissioned a study that revealed that "stormwater BMPs likely represent the largest costs to local governments in implementing the TMDL." H.B. 987, 2012 Gen. Assemb. Reg. Sess. (Md. 2012).

The Permits also contain annual reporting requirements for: (1) the components of the stormwater management programs, and (2) data pertinent to the assessment of progress in implementing the Permit requirements, such as impervious surfaces and pollutant load reductions. MDE will review the Counties' reports to assess "progress toward meeting WLAs developed under EPA approved TMDLs" and the effectiveness of the programs in "reducing the discharge of pollutants to the MEP to protect water quality." MDE will require BMP and program modifications if the Counties fail to comply with the Permit or show progress.

The Permits also contain provisions setting forth sanctions for the violation of Permit conditions, including civil and criminal penalties. See, e.g., Montgomery County NPDES Permit Part VI.C ("Failure to comply with a permit provision constitutes a violation of the CWA and is grounds for enforcement action; permit termination, revocation, or modification; or denial of a permit renewal application.").

Circuit Court and Court of Special Appeals Opinions

EN § 1-601 provides for judicial review of MDE's final determination to issue a permit.36 The Water Groups37 challenged the Permits in the various counties where MDE issued them.

The Circuit Court for Montgomery County remanded for MDE to revise the Permit in accordance with its opinion and order. In a reported opinion, the Court of Special Appeals affirmed. Md. Dep't of the Env't v. Anacostia Riverkeeper, 222 Md.App. 153, 157, 112 A.3d 979, 981 (2015), cert. granted, 443 Md. 734, 118 A.3d 861 (2015). MDE filed a petition for writ of certiorari, which we granted.

The Circuit Court for Baltimore County, the Circuit Court for Anne Arundel County, and the Circuit Court for Prince George's County affirmed MDE's decision to issue those Permits. The Water Groups filed notices of appeal to the Court of Special Appeals and, upon MDE's motion, the Court of Special Appeals consolidated these three cases. MDE then filed a petition for writ of certiorari to this Court with questions nearly identical to those MDE submitted in its petition for writ of certiorari with respect to the Montgomery County Permit.

Finally, the Circuit Court for Baltimore City also affirmed MDE's decision to issue the Baltimore City Permit. The Water Groups filed a notice of appeal, and the Mayor & City Council of Baltimore ("Baltimore City") filed a petition for writ of certiorari with a request that we consider this petition in conjunction with MDE's petitions. We granted the City's petition.

As the Water Groups state in their brief, "the underlying Permits are substantively identical" and "are affected by the same legal defects." We agree that the Permits are so substantively similar that we will analyze the agreed upon questions brought before the Court with respect to all the challenged Permits. We have slightly rephrased the questions:

We uphold MDE's decision to issue the Permits on all grounds. Thus, we reverse the judgment of the Court of Special Appeals, which did not uphold the Montgomery County Permit, and we affirm the judgments of the circuit courts, which upheld the Permits in Anne Arundel County, Baltimore City, Baltimore County, and Prince George's County.

STANDARD OF REVIEW

Before 2009, challenges to the issuance or denial of a discharge permit were subject to a contested case hearing. Md.Code (1984, 2014 Repl.Vol.), § 10-222 of the State Government Article ("SG"), which is part of Maryland's Administrative Procedure Act, delineates the grounds upon which a court can reverse an agency decision in a contested case. Specifically, SG § 10-222 provides that a court can reverse an agency decision in a contested case that "is unsupported by competent, material, and substantial evidence in light of the entire record as submitted" or that "is arbitrary or capricious."

In 2009, the General Assembly changed the procedures for challenging a discharge permit. EN § 1-601 now permits direct judicial review of agency permitting decisions without a contested case hearing. Although this statute does not set forth a standard of review, the substantial evidence and arbitrary and capricious standards apply where an "organic statute" authorizes judicial review without a contested case hearing and does not set forth a standard of review. See Supervisor of Assessments of Carroll Cnty. v. Peter & John Radio Fellowship, Inc., 274 Md. 353, 355, 335 A.2d 93, 94 (1975) ("Our cases have held that where no scope of review is thus provided, decisions of an administrative body will not be disturbed on appeal unless they are not supported by substantial evidence or are arbitrary, capricious or unreasonable.") (citations omitted); Med. Waste Assocs., Inc. v. Md. Waste Coal., Inc., 327 Md. 596, 621, 612 A.2d 241, 253 (1992) ("In an action for judicial review of an administrative decision granting a permit, however, the court determines not only whether the agency's decision to issue the permit was in accordance with law, but also whether the particular administrative decision was arbitrary, capricious or unsupported by substantial evidence in light of the record as a whole."). Thus, even though all challenges going to the merits of the Permits in these consolidated cases originated in the Circuit Courts,38 we will review MDE's decision to issue the Permits under the substantial evidence and arbitrary and capricious standards of review.

Applying the substantial evidence standard of review to a case where no contested case hearing took place may seem anomalous because there is no formal record that was presented before an administrative law judge. EN § 1-606, however, expressly details the documents that can be included in a record. EN § 1-606(c)(1)-(9).39 For example, EN § 1-606 stipulates that any draft permit, comments submitted to MDE during the public comment period, transcripts of public hearings on the permit application, and responses to submitted comments constitute part of the administrative record. Thus, we are essentially reviewing the same record that we would have examined, excluding the administrative law judge's decision, had the merits of this case been subject to a contested case proceeding. Accordingly, our review of the issuance of the Permits fits within the substantial evidence standard of review contemplated by SG § 10-222.

In a review for substantial evidence, we ask "whether a reasoning mind reasonably could have reached the factual conclusion the agency reached." Najafi v. Motor Vehicle Admin., 418 Md. 164, 173, 12 A.3d 1255, 1261 (2011). We should accord deference "`to the agency's fact-finding and drawing of inferences'" when the record supports them. Id. (citation omitted); see Mayor & Aldermen of City of Annapolis v. Annapolis Waterfront Co., 284 Md. 383, 399, 396 A.2d 1080, 1089 (1979) ("The court may not substitute its judgment on the question whether the inference drawn is the right one or whether a different inference would be better supported. The test is reasonableness, not rightness.") (citation and internal quotation marks omitted). Moreover, we shall review the agency's decision "`in the light most favorable to it.'" Najafi, 418 Md. at 173, 12 A.3d at 1261. Finally, we must accord an agency great deference regarding factual questions involving scientific matters in its area of technical expertise. Bd. of Physician Quality Assurance v. Banks, 354 Md. 59, 69, 729 A.2d 376, 381 (1999) ("[T]he expertise of the agency in its own field should be respected.").

We have characterized the arbitrary and capricious standard of review as one that is "extremely deferential." Harvey v. Marshall, 389 Md. 243, 299, 884 A.2d 1171, 1205 (2005). In reviewing the issuance of an NPDES permit, the U.S. Court of Appeals for the Second Circuit quoted language derived from Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983) — the United States Supreme Court's leading case on the arbitrary and capricious standard:

Natural Res. Def. Council, 808 F.3d at 569 (citations and quotation marks omitted). The court also elaborated that "[w]e must be `satisfied from the record that the agency... examine[d] the relevant data and articulate[d] a satisfactory explanation for its action'" and that it "afford[ed] the agency's decision greater deference regarding factual questions involving scientific matters in its area of technical expertise." Id. (quoting State Farm, 463 U.S. at 43, 103 S.Ct. 2856) (citations and internal quotation marks omitted). The Second Circuit's articulation of the arbitrary and capricious standard is in accord with Maryland's treatment of this standard as one that is highly deferential. See Harvey, 389 Md. at 299, 884 A.2d at 1205. We are therefore mindful of the Second Circuit's explanation of the principles underlying the arbitrary and capricious standard when applying that standard to this case.40

In addition, we will review an agency decision for an error of law. When our review concerns a legal question, we apply less deference to the agency's conclusions. HNS Dev., LLC v. People's Counsel for Balt. Cnty., 425 Md. 436, 449, 42 A.3d 12, 20 (2012). We refuse to uphold an agency decision "premised solely upon an erroneous conclusion of law." Id. (citation omitted) (internal quotation marks omitted). Otherwise, we ordinarily should give "considerable weight" to an agency's "interpretation and application of the statute which [it] administers." W.R. Grace & Co. v. Swedo, 439 Md. 441, 453, 96 A.3d 210, 217 (2014); John A. v. Bd. of Educ. for Howard Cnty., 400 Md. 363, 381-82, 929 A.2d 136, 147 (2007) ("In reviewing an agency's legal conclusions, it is a fundamental principle of administrative law that a reviewing court should not substitute its judgment for the expertise of those persons who constitute the administrative agency.").

DISCUSSION

Part I: The 20% Restoration Requirement

The Permits require, by the end of the five-year term, that the Counties restore 20% of the impervious surface areas in their watersheds that have not been restored to the MEP. This requirement "uses percent impervious cover in a watershed as a surrogate TMDL target." ENSR, Pilot TMDL Applications Using the Impervious Cover Method § 1.0, at 1-1 (2005). Like so much of this case, we must unpack the science before we analyze the parties' arguments.

As we develop on land, science has shown us that we profoundly impact our waters. Consider, for example, when "[t]rees, meadow grasses, and agricultural crops that had intercepted and absorbed rainfall are removed...." CWP & MDE, Manual, § 1.1, at 1.3. Problematically, "[c]leared and graded sites erode, are often severely compacted, and can no longer prevent rainfall from being rapidly converted into stormwater runoff." Id. These kinds of sites are known as impervious surfaces, surfaces "that do[] not allow stormwater to infiltrate into the ground," such as "rooftops, driveways, sidewalks, or pavement." EN § 4-201.1(d). "Impervious surfaces accumulate pollutants deposited from the atmosphere," pollutants which are "rapidly delivered to downstream waters" during storms. CWP & MDE, Manual, § 1.1.1, at 1.5. The purpose of the 20% restoration requirement, then, is to use stormwater management practices to restore the natural, beneficial processes in our environment that we have changed by developing impervious surfaces.

In other words, the 20% restoration requirement is a surrogate because the requirement does not control pollution reduction directly. See ENSR, Pilot TMDL Applications Using the Impervious Cover Method § 1.0, at 1-1. Rather, it is through restoring impervious surfaces with management practices that the Counties will reduce pollution. See, e.g., CWP & MDE, Manual, § 1.2, at 1.13 ("[Management practices] shall be designed to remove 80% of the average annual post development total suspended solids load (TSS) and 40% of the average annual post development total phosphorus load (TP).").

A. Maximum Extent Practicable

The Water Groups argue that the 20% restoration requirement is too opaque to comply with 33 U.S.C. § 1342(p)(3)(B)(iii), the MEP standard. They so argue because, they contend, MDE "failed to provide a specific performance standard for restoration activities" or a "numeric limitation ... for what pollution reductions must be accomplished by the permittees' twenty-percent restoration efforts." They also argue that MDE failed to explain what impervious surface is "not restored to the MEP."

We disagree because (1) the applicable law does not impose a specific performance standard on MS4s and (2) MDE did actually select a performance standard for the Counties to adhere to. 33 U.S.C. § 1342(p)(3)(B)(iii) states:

From the text, we discern that Congress established a broad requirement for MS4s. The list of required controls is not exclusive. See 55 Fed.Reg. at 48,038 ("[MS4] controls may be different in different permits."). And the purpose of the controls — reducing the discharge of pollutants — exists alongside the flexible, undefined standard "to the maximum extent practicable." See City of Abilene v. EPA, 325 F.3d 657, 659-60 (5th Cir.2003); Natural Res. Def. Council, Inc. v. N.Y. State Dep't of Envtl. Conservation, 25 N.Y.3d 373, 406, 13 N.Y.S.3d 272, 34 N.E.3d 782 (N.Y.2015) (Rivera, J., dissenting in part) ("The CWA does not define the maximum extent practicable standard. However, it appears to provide broad authority to agencies to control stormwater pollution."); 55 Fed.Reg. at 48,038 ("In enacting section 405 of the WQA [Water Quality Act], Congress recognized that permit requirements for [MS4s] should be developed in a flexible manner to allow site-specific permit conditions to reflect the wide range of impacts that can be associated with these discharges.").

33 U.S.C. § 1342(p)(3)(B)(iii) imposes no minimum standard or requirement on MDE other than to establish controls for MS4s to reduce the discharge of pollutants. See Natural Res. Def. Council, Inc. v. EPA, 966 F.2d 1292, 1308 (9th Cir.1992) ("Congress did not mandate [in § 1342(p)(3)(B)(iii)] a minimum standards approach or specify that EPA develop minimal performance requirements. .... Congress could have written a statute requiring stricter standards, and it did not."). Thus, we reject the Water Groups' argument that MDE committed legal error by "fail[ing] to provide a specific performance standard for restoration activities."41

Moreover, MDE tethered restoration to the practices in the Manual, which MDE incorporates into the Permits by reference. See Part III.E ("These management programs are designed to control stormwater discharges to the maximum extent practicable...."); Part III.E.1 ("At a minimum, the County shall ... [i]mplement the stormwater management ... practices found in the [Manual]...."). The Manual explains that the list of acceptable stormwater management practices is tied to the WQv. "The Water Quality Volume (denoted as the WQv) is the storage needed to capture and treat the runoff from 90% of the average annual rainfall." CWP & MDE, Manual, § 2.1, at 2.2.42 The Manual further explains that the "WQv is directly related to the amount of impervious cover created at a site." Id. In other words, MDE chose a standard that relates to the very problem the 20% restoration requirement serves to abate: the increase in stormwater runoff and the discharge of pollutants because of the increase in impervious surfaces. See CWP & MDE, Manual, § 1.1., at 1.4 ("As can be seen, the volume of stormwater runoff increases sharply with impervious cover."). Thus, the record reflects that MDE has established a performance standard, WQv, that defines as acceptable those practices the Counties may choose from to fulfill the 20% restoration requirement. See id. § 2.7 (Acceptable Urban BMP Options).

Moreover, our discussion of restoration is instructive as to why, despite the Water Groups' contention, the "impervious surface area that is not restored to the MEP" is sufficiently clear and measurable. (Emphasis added.) The area that is not restored to the MEP is the area without the restoration controls described in the Manual. Moreover, the Manual explains that impervious area refers to an area "that does not have vegetative or permeable cover." CWP & MDE, Manual, § 2.1, at 2.4. Put together, the "impervious surface area that is not restored to the MEP" refers to a defined type of area (impervious surface) lacking a type of stormwater management control (the BMPs in the Manual). By way of example, previous MS4 reports delineate these criteria so that MDE can evaluate whether Montgomery County (in this example) installed the required controls. See Montgomery County Department of Environmental Protection, Annual Report for 2006 NPDES MS4 Permit F2, at III-64; see also 33 U.S.C. § 1342(p)(3)(B)(iii) (MS4 permits "shall require controls" such as management practices.).

Because 33 U.S.C. § 1342(p)(3)(B)(iii) does not require a specific performance standard, and because the concepts of restoration and impervious surface "not restored to the MEP" are sufficiently clear as to the controls that the Counties must install, the 20% restoration requirement in the Permits complies with the MEP standard. See 33 U.S.C. § 1342(p)(3)(B)(iii).

B. Substantial Evidence and Arbitrary and Capricious

The Water Groups also argue that MDE has not explained why it selected 20% as the restoration goal or how this Permit provision will promote necessary pollution reduction. The Water Groups contend that MDE ineffectively justifies its choice based on the Bay TMDL because the Permits do not assure that the Counties will achieve the Bay TMDL's objectives or reductions. Even accepting a connection between the 20% restoration requirement and the Bay TMDL, the Water Groups argue that MDE still failed to take into account numerous other TMDLs related to the Counties' waters.

We disagree with the Water Groups' position because (1) the applicable law affords permitting authorities flexibility in establishing controls for MS4s and (2) MDE has justified its decision based on a well-developed and vetted strategy. Natural Res. Def. Council, 808 F.3d at 569 (citation and internal quotation marks omitted). ("We must be `satisfied from the record that the agency ... examine[d] the relevant data and articulate[d] a satisfactory explanation for its action.'").

Congress established a flexible framework in 33 U.S.C. § 1342(p)(3)(B)(iii). As the text states, MS4s shall require controls "and such other provisions as the Administrator or the State determines appropriate for the control of such pollutants." 33 U.S.C. § 1342(p)(3)(B)(iii) (emphasis added). MDE has determined that the 20% restoration requirement is a key strategy in restoring the Chesapeake Bay. See Part VI.A (Chesapeake Bay Restoration by 2025.). The Permits in particular state that the Counties will carry out the 20% restoration requirement in order to comply with the Bay TMDL. Id.

A review of Maryland's Watershed Implementation Plan ("WIP") is instructive as to why the 20% restoration requirement will help to restore the Chesapeake Bay. MDE, Basis for Final Determination. As we have discussed, WIPs are "roadmaps" setting forth a plan for how and when a jurisdiction will reach the pollution reduction goals in the Bay TMDL. EPA, Bay TMDL ES at ES-8. The EPA developed the Bay TMDL "in reliance on" the WIPs that each affected jurisdiction submitted to the EPA. Am. Farm Bureau Fed'n, 792 F.3d at 291 (emphasis added). Moreover, the EPA approved the Bay TMDL "only after" determining that each jurisdiction provided "reasonable assurance" that it would meet the pollutant reductions in its WIP. Id. (emphasis added); see EPA, Bay TMDL ES at ES-8 (The WIPs are the "cornerstone" that ensures accountability to achieve pollution reductions).43

Importantly, the strategies in Maryland's WIP for urban stormwater include, among other things, the reduction of impervious surface area by 20% over a five-year period, just as the Permits do. Compare Maryland Phase I Watershed Implementation Plan: Executive Summary ES-15 (Dec. 3, 2010) [hereinafter Phase I WIP ES], with Part III.G.2. In fact, restoration of "twenty percent of the counties' impervious surface area that is not already restored to the maximum extent practicable (MEP)" is listed in the "key elements" that provide reasonable assurance of the implementation of Maryland's WIP. Maryland Phase I Watershed Implementation Plan: § 5.2.2, at 5-30 (Dec. 3, 2010).

MDE incorporated another important element from the WIP to bolster compliance with the 20% restoration requirement. The EPA explained that the WIPs were to include "mechanisms to track and report implementation activities" and "provide alternative approaches." EPA, Chesapeake Bay TMDL ES at ES-8-ES-9. Maryland's WIP thus includes what we refer to as the adaptive management approach, whereby additional or alternative practices are implemented if existing programs are not meeting target reductions. Phase I WIP at 5-30.44 MDE incorporated this approach from the WIP into the Permits as part of the agency's review of the Counties' annual reports: "BMP and program modifications shall be made" if the Counties fail to comply with the Permits or fail to show progress in meeting WLAs of EPA-approved TMDLs.45 We are satisfied from the record that MDE "examine[d] the relevant data and articulate[d] a satisfactory explanation for its action." Natural Res. Def. Council, 808 F.3d at 569 (citations and internal quotation marks omitted). We conclude, then, that MDE's decision to include the 20% restoration requirement in the Permits was supported by substantial evidence and is not arbitrary and capricious.46

Although we recognize that MDE issued the Montgomery County Permit before Maryland's WIP and the Bay TMDL were prepared, the Bay TMDL was not an isolated event. See Am. Farm Bureau, 984 F.Supp.2d at 298 (The Bay TMDL "is not a new or recent idea," and thus, "it would be improper to view the Final TMDL in a vacuum as a single, isolated effort to restore water quality to the Chesapeake Bay."). As we have discussed, an important prior effort to restore the Chesapeake Bay was the Chesapeake 2000 Agreement. To achieve the restoration goals of that agreement, the Governor of Maryland's Chesapeake Bay Cabinet prepared a draft of the State's plan. Department of Legislative Services, Office of Policy Analysis, Chesapeake Bay Restoration and the Tributary Strategy: An Analysis of Maryland's Efforts to Meet the Nutrient and Sediment Reduction Goals of the Chesapeake 2000 Agreement iii (2007). The strategy for stormwater included restoration of impervious surface in Montgomery County, the first County to successfully implement this program. Id. at 13. The State conceived of this strategy as an effective state-wide method of improving the Chesapeake Bay. See Maryland's Chesapeake Bay Tributary Strategy Statewide Implementation Plan 16 (2008) (stormwater strategy). Thus, MDE's decision to include the 20% restoration requirement in the Montgomery County Permit was also supported by substantial evidence and was not arbitrary and capricious.

We further disagree with the Water Groups' position that the 20% restoration requirement is insufficient because it does not relate to other TMDLs. Indeed, the Permits incorporate a requirement to submit plans regarding WLAs for all EPA-approved TMDLs. That requirement ensures the Permits address all applicable TMDLs.47

C. The Methodology in the Guidance

Finally, with respect to the 20% restoration requirement, the Water Groups object to MDE's method of calculating impervious surface area not restored to the MEP. They assert that the Guidance, which MDE incorporated into these Permits by reference, is flawed because MDE arbitrarily selected 2002 as the baseline for measuring impervious surface area. For example, the Water Groups allege that MDE has ignored the current definition for MEP in the agency's own regulations.

We disagree with the Water Groups because MDE reasonably justifies its decision based on the accurate determination that 2002 marked a significant milestone in the State's treatment of water quality. See Najafi, 418 Md. at 173, 12 A.3d at 1261 (discussing the substantial evidence test); Natural Res. Def. Council, 808 F.3d at 569 (discussing the arbitrary and capricious test).

MDE explained its selection of 2002 in the Guidance: "Maryland regulations and local ordinances began requiring BMPs [around this time] to address a specific suite of volumes [recharge (Rev), water quality (WQv), and channel protection (Cpv)] and it can therefore be justified that water quality treatment has been provided to the MEP." MDE, Guidance at 4. In proposing new regulations for stormwater management in 2000, MDE explained: "The resulting program will provide water quality treatment of up to 90 percent of the average annual rainfall throughout the State, establish ground water recharge standards, and outline a channel erosion control strategy." 27 Md. Reg. at 1168. Additionally, MDE explained that prior regulations had provided "sparse guidance" on "water quality enhancement." Id. To bolster MDE's focus on water quality, the agency incorporated the Manual by reference into the 2000 regulations. See id. at 1170; see also CWP & MDE, Manual § 2.1, at 2.2 ("The WQv is directly related to the amount of impervious cover created at a site.").

The Water Groups reject MDE's decision by arguing that Maryland's regulations currently define "maximum extent practicable" based on a different requirement the State imposed after 2002. That is, COMAR 26.17.02.02 B(22) defines MEP as the design of "stormwater management systems so that all reasonable opportunities for using ESD planning techniques and treatment practices are exhausted and, only where absolutely necessary, a structural BMP is implemented." (Emphasis added.)

MDE has explained, however, that pursuant to the Stormwater Management Act of 2007, which precipitated the regulatory change the Water Groups refer to,48 the performance requirement ESD pertains to "future development," not to restoration. MDE, Basis for Final Determination. As set forth in EN § 4-203(b)(1), ("[t]he rules and regulations shall [i]ndicate that the primary goal of the State and local programs will be to maintain after development, as nearly as possible, the predevelopment runoff characteristics.") (emphasis added). See also COMAR 26.17.02.01 B ("These regulations for stormwater management apply to the development or redevelopment of land...."). MDE has reached a correct legal conclusion, and one regarding the law in which the agency has expertise. See John A., 400 Md. at 381-82, 929 A.2d at 147 ("In reviewing an agency's legal conclusions, it is a fundamental principle of administrative law that a reviewing court should not substitute its judgment for the expertise of those persons who constitute the administrative agency."). We find no error here.

Although the agency selected ESD as the MEP via regulation in 2009,49 MDE chose 2002 because that year marked the beginning of the stormwater management era when BMPs were subjected to performance standards to better treat water quality. 27 Md. Reg. at 1168. WQv is one such standard. CWP & MDE, Manual § 1.2, at 1.13. The Manual shows a direct relationship between the amount of impervious surface and WQv. Id. § 2.1, at 2.3 (Figure 2.2). Thus, beginning in 2002, Maryland connected stormwater management practices to the restoration of impervious surface area, which is the Permit requirement at issue. MDE's conclusion that the Counties should calculate impervious surface area using 2002 as a baseline as part of the 20% restoration requirement was supported by substantial evidence and was not arbitrary and capricious. See Banks, 354 Md. at 69, 729 A.2d at 381 ("[T]he expertise of the agency in its own field should be respected.").50

Because we uphold the Guidance as a component the Counties may legally use to achieve the 20% restoration requirement, we must address whether Montgomery County can use this methodology, too. We recognize that MDE issued the Montgomery County Permit before MDE prepared the Guidance. Compare Montgomery County Permit: Part I: C (Effective Date: February 16, 2010), with MDE, Basis for Final Determination at 10 (MDE held meetings with the Counties about the Guidance from August 2010 to February 2011 before completing it in June 2011.). As we have discussed, MDE limited the practices the County could implement by connecting restoration to the Manual. But, importantly, MDE did not limit the manner in which Montgomery County could show its compliance with the 20% restoration requirement. See, e.g., Montgomery County Permit: Part G.2 (Watershed Restoration), Part IV.A (Annual Reporting). Because the Permit does not restrict how the County must show its progress in achieving the 20% restoration requirement and because of the important links we have already analyzed between the Guidance and the Manual (i.e., the WQv standard), we conclude that Montgomery County can rely on the Guidance in showing its compliance with this requirement.

Part II: TMDL Plans

As we discussed at the outset of this opinion, TMDLs are informational tools, of which WLAs — wasteload allocations — are critical. As the District Court for the Middle District of Pennsylvania aptly noted, "WLAs are not permit limits per se; rather they still require translation into permit limits...." Am. Farm Bureau Fed'n, 984 F.Supp.2d at 328 (citation and internal quotation marks omitted) (emphasis in original). The EPA requires such translation pursuant to 40 C.F.R. § 122.44(d)(1)(vii)(B), which the Water Groups argue MDE failed to comply with.51

Pursuant to 40 C.F.R. § 122.44(d)(1)(vii)(B), MDE included a provision that requires the Counties to submit plans for all EPA-approved TMDLs one year after the issuance of the Permits (or alternatively, one year after the EPA approves the TMDL).52 In creating these restoration plans, the Counties must select actions (including cost estimates) and set forth a schedule (including deadlines and pollution reduction benchmarks) to meet WLAs. Because MDE foresaw the process of meeting WLAs as "iterative," the Counties must also describe what actions they will take when they fail to make progress in meeting WLAs.53

Effluent Limitations Must Be Consistent with WLAs

40 C.F.R. § 122.44(d)(1)(vii)(B) requires MDE to establish effluent limitations that take into account WLAs:

(emphasis added).

The Water Groups construe the contents of the TMDL implementation plans, that is, the actions and deadlines discussed supra, as effluent limitations. As they put it, nothing else in the Permits is "remotely capable of ... impos[ing] effluent limitations that are consistent with the assumptions and requirements of TMDLs." Thus, the Permits can only be in compliance with 40 C.F.R. § 122.44(d)(1)(vii)(B) when MDE has reviewed the contents of the TMDL implementation plans. The problem for MDE, then, is one of timing. Because MDE will review the contents one year after issuing the Permits, the Water Groups conclude that MDE cannot know whether the effluent limitations in the Permits are consistent with the WLAs.

We disagree with the Water Groups because (1) 40 C.F.R. § 122.44(d)(1)(vii)(B) is, like the MEP standard, flexible as to how a permitting authority complies with this regulation and (2) MDE has complied with the regulation by incorporating the WLAs (the most critical element of the regulation) into the Permits and by using an "iterative" process of agency review and program change to ensure progress in meeting the WLAs.

Before we analyze MDE's "iterative" process, we must take issue with the Water Groups' position that the contents of the restoration plans are effluent limits. We understand the effluent limits to be best management practices ("BMPs").54 But, importantly, the Permits do not incorporate these effluent limits in the restoration plans but in another section of the Permits — Part III.E. That is, the Counties must implement a stormwater management program ("SWMP"), including the "practices found in the 2000 Maryland Stormwater Design Manual."55 Importantly, the Permits instruct that "these [management] programs shall be integrated with other permit requirements to promote a comprehensive adaptive approach toward solving water quality problems." In this light, we understand the TMDL implementation plans as roadmaps because they describe a plan for achieving a goal (using BMPs to meet WLAs). See, e.g., Montgomery County Permit Part III.J (In the TMDL implementation plans, the County must "describe those actions necessary to meet the storm drain system's share of WLAs in EPA approved TMDLs."). The requirement to implement BMPs, however, exists independent of the requirement to submit the TMDL implementation plans.

Having clarified this matter, we turn to the regulation to better understand MDE's obligation with respect to effluent limits and WLAs. We note that the plain text imposes a duty on the permitting authority to ensure that effluent limits are consistent with WLA assumptions and requirements. See In re City of Moscow, Idaho, 2001 WL 988721, at *9 ("[T]he governing regulations require consistency....") (emphasis in original). But the text does not instruct the permitting authority as to how it must ensure this consistency. See id. at *8 (noting the lack of detail in the regulation). Instead, the EPA set a minimal, flexible requirement in which the permitting authority is to design a scheme where effluent limits are compatible or in agreement with WLAs. See National Pollution Discharge Elimination System, 54 Fed.Reg. 23,868, 23,870, 23,879 (June 2, 1989) (codified at 40 C.F.R. § 122.44) ("Subparagraph (vii) does not prescribe detailed procedures for developing water quality-based effluent limits. Rather, the regulation prescribes minimum requirements for developing water quality-based effluent limits, and at the same time, gives the permitting authority the flexibility to determine the appropriate procedures for developing water quality-based effluent limits."); The American Heritage Dictionary of the English Language 392 (4th ed. 2006) (Consistent means in agreement or compatible.)56; cf. Am. Farm Bureau Fed'n, 984 F.Supp.2d at 328 ("Accordingly, in some circumstances, a state may write a NPDES permit limit that is different from the WLA, provided that it is consistent with the operative assumptions underlying the WLA.").

In re City of Moscow, Idaho is illustrative of the flexibility the regulation affords MDE. There, the Environmental Appeals Board rejected the City of Moscow's petition for review of an EPA-issued NPDES permit. 2001 WL 988721, at *1. In pertinent part, the Board concluded that the EPA (Region X) did not err in creating permit limits although the EPA did not incorporate the design flow rate of an applicable TMDL. Id. at *8-9. After reviewing 40 C.F.R. § 122.44(d)(1)(vii)(B) and the regulatory history, see 54 Fed.Reg. 23,868, 23,879, the Board explained that there was no law or rule prescribing how the EPA was to select a flow rate to create effluent limits, In re City of Moscow, Idaho, 2001 WL 988721, at *8. The Board concluded that the agency acted "well within the discretion accorded [it] under the applicable regulatory scheme." Id. at *9.

So too here. No law or regulation specifies how or burdens MDE in undertaking the process of complying with 40 C.F.R. § 122.44(d)(1)(vii)(B). In fact, as we have discussed, the overarching federal law for MS4s — 33 U.S.C. § 1342(p)(3)(B)(iii) — is broad and flexible, too. See Natural Res. Def. Council, 966 F.2d at 1308.

Under the Permits, the Counties must include in the TMDL implementation plans the actions needed to meet "WLAs in EPA approved TMDLs." Part III.J.1. Thus, the Permit ties the activities to WLAs, not to some other benchmark that could be inconsistent with the WLA. As the Permits make clear, "[a]ll EPA approved TMDL's that establish WLA's applicable to the County's storm drain system are incorporated by reference into this permit." Id.

The Counties must also submit "documentation of progress toward meeting applicable WLAs developed under EPA approved TMDLs." Reporting on WLA progress is reinforced through watershed assessment and restoration requirements. See Part III.F ("[W]atershed assessments shall [s]pecify how restoration efforts will increase progress toward meeting any applicable WLAs included in EPA approved TMDLs."); Part III.G (The Counties shall annually report on progress toward meeting WLAs.).

Moreover, "MDE will review program implementation, annual reports, and periodic data submittal on an annual basis." The Permit requires "BMP and program modifications" if this report fails to "show progress toward meeting WLAs." The modifications are a product of the "iterative approach" the Counties must follow if they do not make progress in achieving WLAs. That is, the Counties must propose "additional or alternate stormwater controls" to meet WLAs, which MDE will review and approve, if adequate.

Environmental Defense Center, Inc. v. United States Environmental Protection Agency ("EDC") is instructive on the importance of permittee reporting and agency review to ensure compliance with the law. 344 F.3d 832, 858 (9th Cir.2003). In EDC, the Ninth Circuit vacated a portion of an EPA rule that permitted small MS4s to discharge under a general permitting scheme, and remanded for the EPA to revise the rule to comply with the CWA. Id. at 858.57 The Ninth Circuit stated that the rule did not require the permitting authority to review the MS4 operator's stormwater management program "to ensure" that the operator's program would "in fact" comply with the relevant law. Id. at 855 (emphasis in original). Thus, the Ninth Circuit concluded that "nothing prevents the operator of a small MS4 from misunderstanding or misrepresenting its own stormwater situation." Id.

The scheme the Ninth Circuit rejected in EDC is materially different from the scheme in the Permits here. MDE requires reporting, assessment, and adaptation to ensure that the Counties' BMPs will make progress to achieve WLAs. We find the effluent limits under this scheme to be consistent with approved WLAs. See id. at 856 ("However, stormwater management programs that are designed by regulated parties must, in every instance, be subject to meaningful review by an appropriate regulating entity...."). Thus, the Permits comply with 40 C.F.R. § 122.44(d)(1)(vii)(B).

Part III: Monitoring

Here, the Water Groups raise two arguments: the Permits do not include federally mandated monitoring to (1) produce representative data in the MS4 jurisdictions and (2) assure compliance with Permit requirements. They contend that MDE elected not to comply with federal regulations that impose the above requirements and that the record does not otherwise show that MDE complied with these laws. MDE responds by defending its monitoring scheme as sufficient to comply with the applicable federal regulations. Moreover, MDE argues that some of the federal provisions are inapposite because they only pertained to the initial application for MS4 permits in the 1990s.

A. The Permit's Monitoring Requirements

We begin by noting what the Permits require. Under Part III.F, each County must conduct chemical, biological, and physical monitoring at one outfall and an associated in-stream station.58 The Permits set forth the number of required monitoring events, sampling methods, pollutants, and locations.59 The Counties must also conduct monitoring of stream channel protection.60 These two levels of monitoring occur at different locations. Compare note 58, with note 60.

B. Applicable Law

We now consider what federal monitoring regulations apply to the Permits. Under 40 C.F.R. § 122.48(b), "[a]ll permits shall specify [r]equired monitoring including type, intervals, and frequency sufficient to yield data which are representative of the monitored activity including, when appropriate, continuous monitoring." (Emphasis added.) Moreover, monitoring requirements must "assure compliance with permit limitations." 40 C.F.R. § 122.44(i)(1); see also In re Gov't of the D.C. Mun. Separate Storm Sewer Sys., 10 E.A.D. 323, 2002 WL 257698, at *20 (Feb. 20, 2002) ("First, both section 122.48(b) and section 122.44(i) would appear to require that certain monitoring conditions be included in all permits."). The Clean Water Act makes plain the EPA's broad authority to set these requirements. 33 U.S.C. § 1318(a); Natural Res. Def. Council, Inc. v. EPA, 863 F.2d 1420, 1434 (9th Cir.1988) (The EPA "has wide discretion and authority to determine monitoring requirements in NPDES permits...."). These requirements apply to state programs. 40 C.F.R. § 123.25(15), (19).

The Water Groups point us to an additional regulation as part of their argument that the Permits are not capable of producing representative data. Specifically, pursuant to 40 C.F.R. § 122.26(d)(2)(iii)(D), applicants for large MS4s must submit a "proposed monitoring program for representative data collection for the term of the permit that describes the location of outfalls or field screening points to be sampled (or the location of instream stations)," and, among other things, "why the location is representative." But MDE disputes whether this regulation is even applicable to our analysis of the agency's decision to issue the Permits when the regulation, in its view, pertains to applicants (i.e., permittees). MDE argues that we should read 40 C.F.R. § 122.26(d)(2)(iii)(D) as relevant only to a permit application — not a permit — and, moreover, only to the first cycle of permit applications for MS4s (i.e., the early 1990s). In MDE's view, the agency had "the flexibility to establish requirements [it] deem[ed] appropriate" in subsequent MS4 permits.

Recognizing that 40 C.F.R. § 122.26(d) concerns "[a]pplication requirements," we nevertheless agree with the Water Groups that the regulation is relevant to our review of MDE's decision to issue the Permits. Although state and federal permitting laws are unhelpful, an EPA Policy Memorandum interpreting 40 C.F.R. § 122.26(d) ("the Policy Memo")61 is instructive. In the Policy Memo, the EPA instructs that "reapplications should focus on maintenance and improvement of [MS4 storm water management] programs" that were required in the initial applications. EPA, Interpretive Policy Memorandum on Reapplication Requirements for Municipal Separate Sewer Systems, 61 Fed.Reg. 41,698, 41,698 (Aug. 9, 1996) [hereinafter Policy Memo].62 According to the Policy Memo, MDE can make changes to the MS4 monitoring program during the reapplication period, but such changes must be "appropriate and useful." Id. at 41,699. In other words, when a permittee reapplies for a discharge permit, and if the permitting authority reissues the permit, it is the permitting authority's responsibility to ensure that the reissued permit contains programs that are adequate in light of the initial application requirements in 40 C.F.R. § 122.26(d). We rely, then, on 40 C.F.R. § 122.26(d) as a baseline to review the monitoring program in the Permits to ensure that MDE only made program changes that were "appropriate and useful." Id. at 41,699.

C. The EPA Policy Memo

The EPA published the Policy Memo to "respond[] to requests from municipalities and NPDES permit writers for clarification about regulations which do not appear to address reapplication requirements, i.e., permit reissuance." 61 Fed.Reg. at 41,698.

In the Policy Memo, the EPA first explains that it does not require an applicant to repeat in full the process in 40 C.F.R. § 122.26(d)(1)-(2). Id. The EPA explained that it would be redundant to "request the same information again [including characterization of data at § 122.26(d)(2)(iii)], where it has already been provided and has not changed." Id. But an applicant "should identify any proposed changes or improvements" to "monitoring activities." Id.

If MS4 applicants and permit writers wish to change programs in future permits, the EPA explains, they may not only "de-emphasize" but also propose to eliminate a program component. Id. at 41,699; see id. at 41,698 ("MS4 permit applicants and NPDES permit writers have considerable discretion to customize appropriate and streamlined reapplication requirements on a case-by-case basis."). De-emphasis may be justified based on the discharger's experience during the first permit term. Id. By way of example, the EPA stated that "new information on the relative magnitude of a problem" and "new data on water quality impacts of storm water discharges" could justify program changes. Id. Moreover, elimination could be justified "when a different water quality program would serve the same goals." Id. We construe the phrase "same goals" as referring to the CWA's objective in Chapter 26 of Title 33 of "restor[ing] and maintain[ing] the chemical, physical, and biological integrity of the Nation's waters." 33 U.S.C. § 1251(a); see 33 U.S.C. § 1318(a)(1)(A)(i) ("Whenever required to carry out the objective of this chapter [26: Water Pollution Prevention and Control],.... the Administrator shall require the owner or operator of any point source to establish and maintain such records ... [and use such monitoring methods] ... as he may reasonably require.") (emphasis added).

If MS4 applicants and permit writers wish to change monitoring programs, the EPA encouraged them "to determine if storm water monitoring efforts are appropriate and useful." 61 Fed.Reg. 41,699. The EPA then suggested alternative techniques: "municipalities may wish to consider using monitoring techniques other than end-of-the pipe chemical-specific monitoring, including habitat assessments, bioassessments, and/or other biological methods." Id. (emphasis added).

D. Representative Monitoring

MDE argues, somewhat correctly, that the Policy Memo grants it "the flexibility to establish requirements [it] deem[s] appropriate." The Water Groups only repeat the language of the Policy Memo, arguing that MDE can change programs to make them "more appropriate and useful." They argue that MDE's decision was not supported by substantial evidence. Although the Policy Memo does grant NPDES permit writers, like MDE, the flexibility to modify program components, including monitoring, 61 Fed.Reg. at 41,699, the EPA does not allow permit writers to reissue permits and abdicate their responsibility in "restor[ing] and maintain[ing] the chemical, physical, and biological integrity of the Nation's waters." 33 U.S.C. § 1251(a). 40 C.F.R. 122.26(d) required, among other things, a "proposed monitoring program for representative data collection for the term of the permit that describes the location of outfalls or field screening points to be sampled (or the location of instream stations)," and, among other things, "why the location is representative." And as we also discussed earlier, the Permits must include monitoring "sufficient to yield data which are representative of the monitored activity." 40 C.F.R. § 122.48(b).

We conclude that MDE's monitoring program in the Permits will produce representative data because MDE has (1) ensured that the Counties monitor stormwater discharges at monitoring locations that represent an adequate range of land uses statewide, and (2) increased the frequency of monitoring to yield more representative information at the County level.

MDE revised the MS4 monitoring program in the late 1990s to implement a monitoring approach called the "three legged stool," so named because it incorporated not only "[c]hemical testing" but also "biological criteria" and "physical assessments." MDE, Maryland's NPDES Municipal Stormwater Monitoring at § 4.2 (1997); see also MDE, Basis for Final Determination to Issue Anne Arundel County's NPDES MS4 Permit. The agency articulated this approach as being "more aligned with the CWA's goal to `restore and maintain the chemical, physical, and biological integrity of the nation's waters.'" MDE, Maryland's NPDES Municipal Stormwater Monitoring at § 4.2 (emphasis added); see also MDE, Basis for Final Determination to Issue Anne Arundel County's NPDES MS4 Permit. MDE also increased the frequency of monitoring so that the Counties needed to monitor twelve storm events instead of three. MDE, Maryland's NPDES Municipal Stormwater Monitoring at § 4.0. Finally, in light of the increase in monitoring type, MDE decreased the number of monitoring sites to one outfall and one associated in-stream station in a watershed. Id. at § 4.3.

The agency explained that it adopted the "three legged stool" approach after an analysis of the MS4s' concerns in implementing the first permits: "local governments argued that in many instances, biological and physical monitoring results are better indicators of small stream health." Maryland's NPDES Municipal Stormwater Monitoring § 4.1; MDE, Basis for Final Determination to Issue Anne Arundel County's NPDES MS4 Permit (2013) (emphasis added). MDE acknowledged the local jurisdictions' concern that "water chemistry data are of little value" in discerning the effects of stormwater in local streams when there is no assessment of stormwater as to "physical stream habitat and biological health." MDE, Maryland's NPDES Municipal Stormwater Monitoring at § 4.1. Nevertheless, MDE explained in its analysis that "[t]here is still a need to continue water chemistry work for the far field objectives of nutrient reduction in the Bay and accurate loading estimates for NPDES." Id. at § 4.2.

In explaining its decision to increase the number of monitored storm events, MDE stated that "more frequent sampling" would "improve pollutant load estimates." MDE, Maryland's NPDES Municipal Stormwater Monitoring at § 4.0.

Moreover, in explaining why it pared the number of chemical monitoring sites, MDE stated that the addition of biological and physical monitoring required the agency to reapportion resources. Id. at § 4.3. To compensate for the reduced number of sites, MDE explained that: (1) "[a] major goal for future monitoring will be to integrate water chemistry, biology, and physical assessments in one suitable watershed" to ensure "all three techniques [are] analyzed holistically"; and that (2) "site selection will be orchestrated at the state level" to "maintain an adequate number of residential, commercial, and industrial sites for State water chemistry needs." MDE, Maryland's NPDES Municipal Stormwater Monitoring at § 4.3.63

MDE's decision-making illustrates that, as the Policy Memo stated, the permitting authority and permittees worked together to "determine if storm water monitoring efforts [were] appropriate and useful." 61 Fed.Reg. at 41,699. MDE did not rubber stamp the jurisdictions' proposal but affirmed that "[t]here is still a need to continue water chemistry work for the far field objectives of nutrient reduction in the Bay and accurate loading estimates for NPDES." MDE, Maryland's NPDES Municipal Stormwater Monitoring at § 4.2.64

Moreover, biological and physical monitoring are within the scope of the EPA's suggestions for alternative techniques, that is, "monitoring techniques other than end-of-the pipe chemical-specific monitoring, including habitat assessments, bioassessments, and/or other biological methods." 61 Fed.Reg. at 41,699 (emphasis added).

By increasing the number of monitored storm events,65 MDE intended for the Counties to acquire more representative data about pollutant loads. Compare 61 Fed.Reg. at 41,699 (The EPA encouraged permit writers "to determine if storm water monitoring efforts are appropriate and useful."), with 40 C.F.R. § 122.62(d)(2)(iii)(D) (The initial application called for "[a] proposed monitoring program for representative data collection.").66

The Water Groups insist on interpreting representative in terms of the boundaries of the political subdivisions. In its analysis of the Counties' initial monitoring experiences, MDE explained that site selection was a challenging undertaking, and one that it had to interpret flexibly so that the Counties could reasonably carry out their monitoring duties. That is, "monitoring only five sites did not allow any single jurisdiction to monitor all possible land uses." Maryland's NPDES Municipal Stormwater Monitoring at § 2.3. From MDE's perspective, however, the state-wide program to improve water quality was maintained by, "in the aggregate," monitoring "the continuum from low density residential to heavy industrial sites." Id. (emphasis added). The watersheds MDE has approved extend from low-density residential land use to commercial use and light industrial use.67 Thus, we conclude MDE has approved of monitoring locations that adequately represent a continuum of land uses for the agency and the Counties to collect data. See MDE, Maryland's NPDES Municipal Stormwater Monitoring at § 4.3 ("[S]ite selection will be orchestrated at the state level" to "maintain an adequate number of residential, commercial, and industrial sites for State water chemistry needs.") (emphasis added).

MDE amended the program with the CWA's objective at the forefront of its mind: "restor[ing] and maintain[ing] the chemical, physical, and biological integrity of the Nation's waters." Id.; see 33 U.S.C. § 1251(a). In accordance with the Policy Memo, the agency's changes and reasoning strike us as thorough efforts to make the monitoring program "appropriate and useful." 61 Fed.Reg. at 41,699. We conclude that MDE acted within the discretion the EPA accorded it in the Policy Memo. Thus, we also conclude that MDE has committed no legal errors pertaining to 40 C.F.R. § 122.26(d)(2)(iii)(D).

For many of the same reasons, we believe that the Permits contain monitoring "sufficient to yield data which are representative of the monitored activity." 40 C.F.R. § 122.48(b). The Permits require monitoring of many pollutants. Part IV. F.1.a-d.68 Following the Counties' first experience implementing MS4 permits, MDE increased the frequency of monitoring to "improve pollutant load estimates." MDE, Maryland's NPDES Municipal Stormwater Monitoring § 4.0. The agency also decided to orchestrate site selection at the state level to "maintain an adequate number of residential, commercial, and industrial sites for State water chemistry needs." Id. § 4.3. By taking control of the selection of monitored sites, MDE could better ensure monitoring of a continuum of activities (industrial, residential, and commercial). Additionally, the Counties will have to explain their monitoring process to ensure that the agency can evaluate whether they have complied with MDE's requirements. Part V.A.1.b. (The Counties shall submit annual reports and include "[a] narrative summary describing the results and analyses of data, including monitoring data that is accumulated throughout the reporting year.").69

MDE also requires the Counties to assess WLAs, which represent pollutant loads. Part IV.E.2.b.iii. The Counties must record continuous flow measurements to obtain data to estimate pollutant reductions. Part IV.F.1.a.iv. Because the Counties are to achieve WLAs over time through restoration activities, MDE requires the Counties to conduct such monitoring "where the cumulative effects of watershed restoration activities can be assessed." Part IV.F.1 (emphasis added).

MDE thus structured the Permits and imposed requirements on the Counties to ensure that they will monitor the discharge of pollutants to yield representative data. We conclude that the Permits comply with 40 C.F.R. § 122.48(b).

E. Permit Limitations

40 C.F.R. § 122.44(i) requires monitoring "to assure compliance with permit limitations." We begin by noting the two applicable limitations: (1) controls to reduce the discharge of pollutants70 and (2) the restoration of 20% of impervious surface area not restored to the MEP. See Part III.D; Part III.E.2.71

1. Reduction of Pollutant Discharge

The Permits include two important components that inform our analysis of whether MDE's monitoring program ensures reductions in the discharge of pollutants: (a) the monitoring and modeling scheme and (b) adaptive management. As we explain, our initial discussion of monitoring and modeling is not sufficient to measure MDE's compliance with 40 C.F.R. § 122.44(i). Rather, it is a necessary component of what we consider a two-part scheme MDE designed to ensure the Counties would implement controls to reduce stormwater discharges.

a. Monitoring and Modeling

MDE requires the Counties to implement a stormwater management program ("SWMP"), which consists of BMPs. Part IV.D.1.a. The Counties must integrate SWMPs with other permit requirements. Part IV.D. Aside from their role in the 20% restoration requirement, BMPs are most critical in the restoration plans that the Counties must implement to achieve progress in meeting WLAs. Part IV.E.2, 4; see supra note 17 (The WLAs are derived from the EPA approved TMDLs for each Permit.).

To achieve progress in meeting WLAs, the Counties must first set a baseline for stormwater pollutant loads. MDE, Guidance, at 3.72 To set this baseline, the Counties must apply TMDL pollutant loading rates to urban land use data. Id. An MDE-approved model, such as MAST, will automatically calculate the cost and load estimates for Counties to inform their decision-making. See About MAST, Maryland Assessment Scenario Tool, available at http://www.mastonline.org/About.aspx [https://perma.cc/7NRV-QNXX]. The Counties can compare different scenarios to determine what BMPs they can implement to reduce pollution. Id.; see also MDE, Guidance, at 1 ("Local governments can weigh the cost associated with implementing different practices and choose the most efficient option for meeting pollutant load reductions."). The Counties select and begin implementing BMPs in accordance with their restoration plans. Part IV.E.2. Then they can monitor the effectiveness of the activities, Part IV.F.1.a.iv, and use the monitoring information to determine if the activities have helped make progress in achieving WLAs, Part IV.E.2, 4.

But the efficiencies for BMPs that the Counties select are based in part on estimates. See, e.g., MDE, Guidance, at 10 (Table 4).73 As our discussion of modeling illustrates, a model predicts the reductions a BMP can achieve.74 But it does not tell the Counties exactly what the BMP will do in terms of pollution reduction. Indeed, many variables impact this process.75 See Chesapeake Bay Program, Phase 5.3 Community Watershed Model, § 6.1.2, at 6-4 ("Uncertainty in estimates of BMP effectiveness is due to factors including (1) variability in precipitation, hydrology, soils, and geology; (2) variable performance of land management practices; [and] (3) lag time between implementing a practice and full performance and observed water quality benefits...."); see also id. at § 6.2.2, at 6-8 ("Virtually all research data are generated under controlled management conditions" that, among other things, are designed to minimize hazards; thus, "the research estimates are more representative of a best-case scenario.").

But this should not raise a red flag about the legal correctness of the Permits. Indeed, the inherent shortcoming in estimating BMP effectiveness explains MDE's decision to require the Counties to adapt their practices based on improved knowledge over time. See infra Discussion: Adaptive Management. Rather, because the Counties must implement BMPs to make progress in achieving WLAs, estimation and prediction are necessary evils in this context. Thus, high quality assessments of BMPs are particularly important in the Permits so that MDE and the Counties can understand which practices are most effective to meet the many WLAs incorporated into the Permits. See infra note 17.

As we have discussed, MDE requires the Counties to assess BMPs through a focused monitoring approach. Part IV. F.1.a.iv. In the Basis for Final Determination, MDE explained that pollution reduction research supported the use of focused monitoring to better understand and, thus, improve BMPs. MDE, Basis for Final Determination to Issue Anne Arundel County's NPDES MS4 Permit (2013).76 The research includes an analysis of how states, including Maryland, were "tracking nutrient and sediment control practices" to restore the Chesapeake Bay. See Committee on the Evaluation of Chesapeake Bay Program Implementation for Nutrient Reduction to Improve Water Quality, Achieving Nutrient and Sediment Reduction Goals in the Chesapeake Bay vii (National Academies Press 2011) [hereinafter Achieving Nutrient and Sediment Reduction Goals in the Bay]. The National Research Council ("NRC")77 established a "multidisciplinary committee of experts to provide advice to the EPA, the six states in the Chesapeake Bay watershed," and "other interested parties." Achieving Nutrient and Sediment Reduction Goals in the Bay at viii.

Importantly, "[t]he EPA specifically directed the NRC to evaluate the tracking of best management practice implementation...." Id. The NRC's appointed committee stated that "[t]argeted monitoring programs ... would provide valuable data to refine BMP efficiency estimates...." Id. at 5.78

Consistent with the NRC committee report, MDE's approach to monitoring requires the Counties to assess the effectiveness of BMPs. Part IV.F.1 ("Monitoring activities shall occur where the cumulative effects of watershed restoration activities can be assessed."). Indeed, MDE designed the second component of the monitoring program with the same purpose in mind: monitoring "for determining the effectiveness of stormwater management practices for stream channel protection." Part IV.F.2.

MDE then uses the monitoring data to calibrate its models. Part IV.F.1.a.iv ("Data collected shall be used to estimate annual and seasonal pollutant loads and reductions, and for the calibration of watershed assessment models."). That is, MDE's approach yields more accurate data on the efficacy of BMPs, which increases certainty for all interested parties in understanding what effects restoration activities will have on the State's waters. MDE, Guidance, at 1 ("The data gathered may be used to update and improve Maryland's stormwater management matrix of options for achieving water quality."). Our review of the first component in the Permits — the monitoring and modeling scheme — illustrates that it is necessary to ensure the Counties are implementing BMPs to the maximum extent practicable.

b. Adaptive Management

In their annual report, the Counties must submit the pollutant load reductions determined through monitoring. Part V.A.2.g; see also Attachment A: Annual Report Databases (G).

The required measurement is pounds per year, which is consistent with the federal requirement that a discharger monitors "mass (or other measurement specified in the permit) for each pollutant" or "volume of effluent discharged from each outfall." 40 C.F.R. § 122.44(i)(1)(i)-(ii). As discussed above, these reductions pertain to the pollutants which the Counties must monitor. Part IV.1.a.iv; see Attachment A: Annual Report Databases (G) (The pollutants arise from the Bay TMDL and local TMDLs.).

MDE will review the annual reports, Part V.B, and will require program modifications "according to needed program improvements identified as a result of [MDE's] periodic evaluations," Part IV.D.79 "Failure to comply with a [P]ermit provision," such as the stormwater management or reporting requirements, "constitutes a violation of the CWA and is grounds for enforcement action; permit termination, revocation, or modification; or denial of a permit renewal application." Part VII.C.

As we have discussed,80 MDE has flexibility in setting controls in MS4s: "Congress did not mandate [in 33 U.S.C. § 1342(p)(3)(B)(iii)] a minimum standards approach or specify that the EPA develop minimal performance requirements." Natural Res. Def. Council, Inc., 966 F.2d at 1308; see also 54 Fed.Reg. at 23,870, 23,879 (codified at 40 C.F.R. § 122.44) ("Subparagraph (vii) does not prescribe detailed procedures for developing water quality-based effluent limits. Rather, the regulation prescribes minimum requirements for developing water quality-based effluent limits, and at the same time, gives the permitting authority the flexibility to determine the appropriate procedures for developing water quality-based effluent limits."). We discern that the law requires a regulating authority such as MDE to review the entity's actions (or non-action, as the case may be) to ensure accountability in implementing stormwater controls: "stormwater management programs that are designed by regulated parties must, in every instance, be subject to meaningful review by an appropriate regulating entity to ensure that each such program reduces the discharge of pollutants to the maximum extent practicable." EDC, 344 F.3d at 856.81

MDE's adaptive management approach includes a requirement to impose program changes based on annual report data obtained from monitoring. Where, for example, the Counties' current strategy is not reducing discharges, the Permits allow MDE to force the Counties to implement BMPs that will, or at least are an improvement. Because the Counties will implement controls to reduce discharges under MDE's oversight, the monitoring provisions comply with 40 C.F.R. § 122.44(i).82 See Natural Res. Def. Council, Inc. v. Cnty. of L.A., 673 F.3d 880, 897 (9th Cir. 2011) ("As opposed to absolving noncompliance or exclusively adopting the MEP standard, the iterative process ensures that if water quality exceedances `persist,' despite prior abatement efforts, a process will commence whereby a responsible Permittee amends its [stormwater quality management program]."), rev'd on other grounds sub nom. L.A. Cnty. Flood Control Dist. v. Natural Res. Def. Council, Inc., ___ U.S. ___, 133 S.Ct. 710, 713, 184 L.Ed.2d 547 (2013); cf. Port of Seattle v. Pollution Control Hearings Bd., 151 Wn.2d 568, 90 P.3d 659, 679 (2004) (en banc) ("Monitoring and adaptive management provide a mechanism through which [the Washington State Department of Ecology] can mitigate [an] inherent uncertainty" in the question of whether there was a reasonable assurance that an airport runway project would not violate applicable water quality standards.); University of Maryland/Mid-Atlantic Water Program, BMP Assessment Final Report 25-26 (2009) ("Utilizing an adaptive management approach recognizes uncertainty and limitations in science, but does not impede implementation of management actions.").83

2. 20% Restoration Requirement

As to the second limitation, it is true that the Permits do not require monitoring of impervious surfaces in the section entitled "Assessment of Controls," which contains the focused monitoring requirements. See Part IV.F. But, as we explain, because (1) 40 C.F.R. § 122.44(i) contemplates a flexible approach to monitoring and (2) MDE incorporated a clear evaluation tool into the Permits to assess restoration of impervious surfaces, MDE has assured compliance with the 20% restoration requirement under 40 C.F.R. § 122.44(i).

Under the restoration provision, MDE's approach requires the Counties first to submit an assessment of their impervious surface area within a year of the issuance of the Permits. Part IV.E.2.a. This Permit part incorporates by reference the Guidance, which contains the methodology we discuss herein. In the assessment, the Counties must delineate those portions in their jurisdictions "that are either treated to the [MEP], partially treated, or untreated and available for retrofit."84 MDE, Guidance at III.4, at 4.

To ensure that the Counties implement satisfactory BMPs on their untreated surfaces, MDE requires the Counties to translate the activities into credits. Id. at IV, at 8. This credit system is tethered to the performance standard through which MDE determined which BMPs are satisfactory: the WQv. Id. "An acre for acre impervious credit will be given when a structural BMP is specifically designed to provide treatment for the full WQv (one inch), or [a] proportional acreage of credit will be given when less than the WQv is provided." Id. Moreover, for BMPs that "provide greater than one inch of volume control," the activities "receive additional credit." Id.85

Because the Counties must adhere to the credit system, MDE can evaluate the jurisdictions' performances uniformly. The Guidance "standardizes procedures for the reporting of traditional, new, and alternative [BMPs] and the impervious area they control." MDE, Guidance at 1; id. ("By developing a comprehensive matrix of practices and consistent accounting measures, [MDE] brings greater certainty to the local planning and budgeting process."). This accounting system is also flexible enough to accommodate more non-traditional activities for which restoration credits are still available. See MDE, Guidance at V, at 11 ("This section presents alternative BMPs that will give jurisdictions greater flexibility toward meeting stormwater permit requirements.").

Moreover, the Counties must report annually on their progress in achieving the 20% restoration requirement. Part V.A.1.a.ix.86 Failure to comply with these requirements — submission of the impervious surface area assessment; implementation of restoration practices on impervious surfaces not controlled to the MEP; and submission of reports on the Counties' activities according to the credit system — is a violation of the CWA and grounds for agency action. Part VII.C.

We find instructive the Second Circuit's reasoning in upholding, in part, the EPA's monitoring provisions in permits to regulate the discharge of ballast water from ships. Natural Res. Def. Council, Inc., 808 F.3d at 562.87 As to certain effluent limitations in these permits,88 the EPA required the monitoring of "functionality" of a vessel's ballast water treatment system, and of the concentrations of two specific bacteria. Id. at 581.89 The EPA did not, however, require vessel owners "to take any measurement of pollutants or significant categories of living organisms in ballast water being discharged." Id.

Although environmental organizations argued that the EPA should have required monitoring of the concentrations of living organisms, the Second Circuit concluded that the EPA's provisions complied with 40 C.F.R. § 122.44(i). Id. at 582-83. Because the regulation sanctioned the monitoring of some "other measurement specified in the permit," the functionality and indicator organism monitoring "qualif[ied] as such `other measurement.'" Id. at 582. Although the environmental organizations had advocated for the alternative of "direct monitoring," the Second Circuit reasoned that dischargers did not have the capacity to quantify living organisms of certain size classes. Id. Moreover, the current technology would require an analysis that was "prohibitively expensive and impractical." Id. There was, in essence, no "feasible" alternative the EPA could have established under these permits. Id. at 582-83. Thus, the Second Circuit deferred to the EPA's decision. Id. at 583.

From NRDC, we discern that a permitting authority has flexibility in how it sets monitoring requirements. As 40 C.F.R. § 122.44(i) reveals, the EPA wrote the regulation with the understanding that not every permit limitation could be measured in terms of mass or volume. See 40 C.F.R. § 122.44(i)(1)(i) (The regulation requires monitoring of mass "or other measurement specified in the permit."). In this case, those measurements would not aid the Counties or MDE in evaluating progress toward restoring impervious surface area. Rather, MDE requires the translation of restoration practices, implemented on impervious surface areas, into credits to make restoration of those areas understandable. The agency can also monitor restoration in a uniform manner.

From NRDC, we also discern that 40 C.F.R. § 122.44(i)(1)(i) requires feasible monitoring. Here, the Guidance not only promotes accountability in a uniform manner but also gives MS4 permittees the flexibility to choose from among approved restoration practices to address local conditions. Flexibility is a hallmark in designing MS4 permits. See 55 Fed.Reg. at 48,038 ("In enacting section 405 of the [Water Quality Act], Congress recognized that permit requirements for [MS4s] should be developed in a flexible manner to allow site-specific permit conditions to reflect the wide range of impacts that can be associated with these discharges."). The Water Groups have not presented a feasible alternative that contains both the flexibility and the accountability to assure compliance with the permit limitation that Counties restore 20% of their impervious surface areas. MDE has complied with 40 C.F.R. § 122.44(i) by incorporating the Guidance by reference for the Counties to use to meet the 20% restoration requirement.

We recognize that MDE did not incorporate the Guidance into the Montgomery County Permit. But that fact is not fatal to that Permit's ability to comply with 40 C.F.R. § 122.44(i). When the County submits its annual report, the County must include databases of impervious acreage and specify those areas treated by BMPs, proposed for restoration, under construction for restoration, and completed. Montgomery County Permit Part IV:A.2, Attachment A: (C). The County must also submit descriptions of the type of BMP it used at specific locations. Id. at Part IV:A.2, Attachment A: (D). As we have discussed, those BMPs are the stormwater practices in the Manual. These reporting requirements will allow MDE to evaluate on a yearly basis the Counties' compliance with the 20% restoration requirement. We recognize that this reporting system is not as detailed as the method in the Guidance, but 40 C.F.R. § 122.44(i) does not include requirements as to the specificity of the measurement the permitting authority selects. See, e.g., Natural Res. Def. Council, Inc., 808 F.3d at 581 (As to monitoring of a vessel's ballast water treatment system, the EPA required ship owners to measure system functionality, "such as how much chlorine the system is using each month.").

The Permit also states that the County must merely describe "the results and analysis of data." Part IV.A.1.b. Because the Permit does not prohibit the type of analysis that the County may use to evaluate the 20% restoration requirement, we conclude that the County may use the Guidance. Thus, we conclude that the Montgomery County Permit complies with 40 C.F.R. § 122.44(i) too.

Part IV: Public Participation

"Many tensions exist between the democratic aspiration of government of the people, by the people, and for the people and modern representative government with its mass electorate and elaborate bureaucracy for carrying out government functions. Nowhere are these tensions more acute than in the domain of environmental policy." National Research Council, Public Participation in Environmental Assessment and Decision, National Academies Press 7 (2008). The dispute between the Water Groups and MDE over whether the Permits satisfy public participation requirements exemplifies this tension. The Water Groups frame their argument by highlighting the ways in which the Permits' two most critical elements — the TMDL Plans and the 20% restoration requirement — fail to satisfy public participation requirements.

A. TMDL Plans

We have discussed at length that the Permits require the Counties to submit TMDL plans and that, among other things, the Permits satisfy 40 C.F.R. § 122.44(d)(1)(vii)(B), which requires that effluent limits be consistent with WLAs.

Nevertheless, the Water Groups contend that the TMDL plans create another problem. The Permits require the Counties to design and submit these plans after the agency approves of the Permits. See, e.g., Montgomery County Permit Part III.J.2 ("Within one year of the effective date of this permit or the approval of an applicable TMDL by EPA, whichever is later, the County shall submit to MDE for review and approval a TMDL implementation plan...."). The Water Groups contend that MDE has "unlawfully circumvent[ed]" federal and state procedural requirements because the restoration plans, which include significant new requirements, come into existence more than one year after the Permits are issued, without providing for public notice and comment.90 The Water Groups characterize this scheme as an "end-run around mandatory permitting requirements" and as an approach that "thwart[s] accountability." Put simply, the essence of the Water Groups' argument is that the public cannot comment about decisions that have yet to be made.

1. Permit Modification

Specifically, the Water Groups view the TMDL plans that must be submitted to MDE as a modification of the Permits. Federal and state regulations stipulate that a permitting authority must provide the public an opportunity to be heard prior to a modification of a permit authorizing discharges. 40 C.F.R. §§ 122.62, 124.10; EN §§ 1-601, 9-324.91 Courts across the country have cited these federal regulations as well as related state regulations in recognizing the necessity of public participation when a permit has been modified. See, e.g., United States v. Smithfield Foods, 191 F.3d 516 (4th Cir.1999); Citizens for a Better Env't — California v. Union Oil Co. of California, 83 F.3d 1111 (9th Cir.1996). Thus, the Water Groups' argument requires us to determine whether the restoration plans constitute a modification of the Permits triggering an obligation upon MDE to implement the public participation provisions in those regulations.

The Permits direct the Counties to develop restoration plans using BMPs that are found in the Manual and discussed again in the Guidance. These documents, which MDE incorporated into the Permits by reference, provide a "menu of options" for the Counties to utilize. By incorporating these documents, MDE made the BMPs available to the Counties at the time the agency issued the Permits. When the Counties submitted (or will submit) restoration plans using these BMPs, no modification will have occurred because the Counties will merely have drawn from the same pool of BMPs that the agency had previously analyzed and approved of for restoration purposes. CWP & MDE, Manual § 2.0 ("This chapter also presents a list of acceptable BMP options that can be used to comply with the sizing criteria," including, the WQv.); see 27 Md. Reg. at 1168 (to be codified at COMAR 26.07.02) (MDE incorporated the Manual as part of the agency's effort to enhance its regulations which had until then provided "sparse guidance" on "water quality enhancement."). Although we recognize that the Counties have selected specific BMPs to implement from among a larger group of BMPs, the larger group still satisfies MDE's specific performance standard, WQv. Thus, no "material and substantial alterations or additions" will occur after Permit issuance "justify[ing] the application of [] conditions that are different or absent in the existing [P]ermit." 40 C.F.R. § 122.62(a)(1).

2. Opportunity to Comment

The Water Groups also argue that the restoration plans violate federal and state laws on public notice and comment because they are unable to meaningfully comment about decisions that have yet to be made. This argument is unavailing because, as we have explained, the BMPs were previously able for the public to comment on. Indeed, many of the same Water Groups challenging the Permits in this case submitted comments on these BMPs.

For example, in the Basis for Final Determination, MDE noted that it "received many, varied and often conflicting comments regarding the Guidance," which referred to the BMPs, and that the Guidance was "widely distributed and commented on." One salient BMP that was objected to by some Water Groups was the use of detention facilities, such as detention ponds.92 For example, Natural Resources Defense Council submitted a comment in which they criticized the Guidance as "flawed" because it "overestimate[d] the efficacy of detention ponds" and "overstate[d] the channel protection benefits of detention practices." Chesapeake Bay Foundation, also a party, commented that detention systems "provid[e] very little water quality benefits."93

On the other hand, the Water Groups strongly supported the use of ESD in their submitted comments.94 Natural Resources Defense Council proffered that the draft Guidance "ignore[d] the Maryland statute establishing ESD as the preferred Maryland approach."95 The Chesapeake Bay Foundation stated that the Permits "should impose a higher performance standard" and recommended that this "be done primarily through Environmental Site Design."96 In its Basis for Final Determination, MDE responded to the Water Groups' comments by highlighting its credit approach to improving water quality. MDE explicated that the Guidance incentivized the utilization of ESD practices by granting greater pollution credit for ESD and less credit for structural treatment BMPs such as detention facilities. Furthermore, MDE justified the use of detention ponds and other types of structural BMPs pursuant to the Manual to afford the Counties the "flexibility to implement various strategies based on site specific opportunities." The comments and MDE's substantive responses thereto, illustrate that there was a vigorous and pellucid public discussion regarding the practices set forth in the Manual and Guidance, and that the agency defended its approach to water improvement strategies as flexible but effective.97

Thus, the record in this case clearly supports that the public had the opportunity and, in fact, actually commented on which BMPs would best meet the MEP standard for restoration purposes.98 In arguing that they cannot comment about decisions that have yet to be made, the Water Groups overlook the fact that the public was able to comment on BMPs — the core component of the TMDL restoration plans.

Moreover, even though the Counties create the restoration plans after the Permits are approved, the public is still able to participate in the formulation of the plans. The Permits explicitly provide for "[a] minimum 30 day comment period" before finalization of the restoration plans. In addition, the Counties must give notice in a local newspaper and on their website outlining how the public may obtain information and provide comments on the restoration plans. Critically, the Counties must also include a summary in their annual reports of how they "addressed or will address any material public comments received." MDE reviews these reports. The Counties are "responsible for complying with all conditions of" the Permits, and "[f]ailure to comply with a [P]ermit provision constitutes a violation of the CWA and is grounds for enforcement action." Thus, contrary to the Water Groups' assertion that the Permits "thwart accountability," the Permits afford the opportunity for robust public involvement in the attempt to abate stormwater pollution and improve water quality.

That the Permits provide for public participation at the County level is not enough for the Water Groups. The Water Groups fault the Permits for not providing for public participation at the state level and proffer that the Permits "prevent[] members of the public from commenting on ... MDE's decision to approve the restoration plans." Because the public had the opportunity to comment on the "menu" of BMPs in the Manual and the Permits unambiguously mandate public participation at the County level, the restoration plans have been subject to a "double layer" of public participation. We, therefore, are not persuaded by the Water Groups' arguments.

Finally, the Water Groups maintain that MDE's claims of "potential administrative burdens are irrelevant and specious." We must, however, give weight to the notion that MDE permitted the Counties to draw from a broad group of BMPs in the Manual in light of the highly variable nature of stormwater discharges. MDE possesses a comparative advantage in addressing environmental problems affecting our State. It is axiomatic that a reviewing court should respect "the expertise of an agency in its own field." Banks, 354 Md. at 69, 729 A.2d at 381 (citations omitted). Moreover, the Counties have expertise in water quality assessments of waterbodies found within the boundaries of their political subdivisions. As the Permits require, the Counties "shall complete detailed watershed assessments" for their entire jurisdiction. Thus, there is great value in deferring to the Counties' choice of BMPs, and because the public had an opportunity to comment on the BMPs during the drafting of the Permits as well as on the select BMPs in the restoration plans the Counties submit to MDE, we conclude that the public has not been deprived of notice and comment.

3. Case Law

The Seventh Circuit's decision in Texas Independent Producers and Royalty Owners Assoc. v. EPA, 410 F.3d 964 (7th Cir. 2005), is also instructive on this point. In Texas Independent Producers, the EPA issued a general permit for stormwater discharges from construction sites that required operators to file a Notice of Intent ("NOI") stating that they planned to operate under the general permit, rather than apply for an individual permit. 410 F.3d at 968. Absent a negative ruling by the EPA, stormwater discharges that complied with the terms of the general permit were automatically authorized. Id. The permit also required that operators "create, maintain, and implement a site-specific Storm Water Pollution Prevention Plan (`SWPPP')." Id. An environmental organization challenged the Permit and asserted that it violated the Clean Water Act because it failed to satisfy public participation requirements regarding the NOI and SWPPP.

After conducting a Chevron analysis, the court concluded that EPA's reasoning as to why it issued the general permit without providing for public review of NOIs and SWPPPs was "eminently reasonable." Id. at 978. Notably, the court accepted EPA's argument that because the public had the opportunity to comment on the proposed general permits whose terms governed the NOIs and SWPPPs, there was "no need for additional public comment or a notice period [regarding NOIs and SWPPPs themselves]." Id. In addition, the court credited the EPA's contention "that requiring `an additional public hearing on each individual NOI and SWPPP would eviscerate the administrative efficiency inherent in the general permitting concept,' in effect making the general permit scheme no different from the process of obtaining individual permits." Id.

The Water Groups point out that the permit in Texas Independent Producers was a general permit whereas the Permits at issue here are "not eligible for general permits." The Water Groups rightly note that the "existence of different permitting regimes for large and small municipal storm sewer systems reflects Congressional policy judgment" that large MS4s be regulated differently than small systems. Even if the permits are not the same, the Permits at issue in this case are conceptually analogous to a general permit because all Counties create their restoration plans based upon the same document setting forth the same groups of BMPs — the Manual. Accordingly, we find the Seventh Circuit's determination regarding public participation in Texas Independent Producers persuasive, and conclude that the Permits' TMDL planning requirement does not violate notice and comment mandates.

The Water Groups rely on the Second Circuit's decision in Waterkeeper Alliance, Inc. v. EPA, 399 F.3d 486 (2d Cir.2005), in support of their argument that the contents of the TMDL plans are subject to public participation requirements. In Waterkeeper Alliance, the court examined the regulation of water pollutants contained in the runoff from concentrated animal feeding operations ("CAFOs"). In pertinent part, the court held that the nutrient management plans the CAFOs were required to develop were effluent limits but that the EPA violated public participation requirements because it "fail[ed] to require that the terms of the nutrient management plans be included in the NPDES permits." Waterkeeper Alliance, 399 F.3d at 502-04. The court stated that the plans were effluent limits because the CAFOs were required to set waste application rates in the plans. Id. at 502; see 33 U.S.C. § 1362(11) (An effluent limit includes a restriction on a rate.). In contrast to the nutritional plans discussed in Waterkeeper Alliance, the most critical element of the restoration plans — the BMPs — is already included in the Permits because the Permits incorporate the Manual and Guidance, which set forth those practices. Thus, the Water Groups' reliance on Waterkeeper Alliance is misplaced. See Part III.G, Part IV.E.2.a.

A Ninth Circuit decision addressing public participation is also distinguishable. In Environmental Defense Center v. EPA, the Ninth Circuit remanded an EPA regulation that required operators of MS4s to develop and implement individualized pollution control programs that were not subject to agency review or public participation. 344 F.3d at 853-58. The court found that the rule violated 33 U.S.C. § 1342(p)(3)(B)(iii), which is the same statutory provision that is at the center of this case. The court held that MS4 pollution control programs must be reviewed by the permitting agency and were subject to public comment, in part because those permittee-prepared programs — not the general permits issued to MS4 operators — would contain the substantive requirements that operators must implement to reduce discharges to the "maximum extent practicable." Id. at 855, 857.

Here, the Counties must submit restoration plans derived from specified best management practices, including ponds, wetlands, infiltration practices, filtering systems, open channels, and other structural water quality practices, that were previously available for public comment. In contrast to the specific practices that form the basis of the restoration plans in the Permits, all that was available at the time of permit issuance in Environmental Defense Center were "six general criteria" designed to protect water quality that were never reviewed by the public. Id. at 845, 853 (emphasis added).99 Consequently, the court's holding in Environmental Defense Center rested, at least in part, on facts that are distinguishable from the circumstances in this case.

4. Incorporation By Reference And Deference

To be sure, the Permits rely heavily on incorporation by reference. But such incorporation by reference, even of important documents, does not contravene public participation requirements because the Manual and Guidance are readily accessible to the public. Cf. Proposed Action on Regulations, 35:25 Md. Reg. 2193 (Dec. 5, 2008) ("Pursuant to State Government Article § 7-207, Annotated Code of Maryland, the 2000 Maryland Stormwater Design Manual ... has been declared a document generally available to the public and appropriate for incorporation by reference. For this reason, it will not be printed in the Maryland Register or [COMAR].").100 Additionally, including the best management practices in the Permits would significantly lengthen the document, which would obfuscate other requirements in the Permit. The Manual has also been part of Maryland regulatory law for two decades. It is so entrenched in the permitting process that we see no reason to detail the content of the Manual in the Permits.

5. Compliance Schedules

The Water Groups further argue that MDE must formally incorporate the restoration plans through modification procedures because the plans contain compliance schedules. They allege that compliance schedules are effluent limitations under the CWA, see 33 U.S.C. § 1362(11), (17), and they must be "included in a `permit,'" 40 C.F.R. § 122.2. Ergo, MDE must formally incorporate the Counties' restoration plans — containing the compliance schedules — after the Counties submit their plans to MDE. The problem, however, is that the Water Groups do not explain why the schedules in the restoration plans constitute compliance schedules.

The Permits state that the Counties must include in their restoration plans "the final date for meeting applicable WLAs and a detailed schedule for implementing all structural and nonstructural water quality projects, enhanced stormwater management programs, and alternative stormwater control initiatives necessary for meeting applicable WLAs." Part IV. E.2.b.i. Quite obviously, the restoration plans contain schedules. Id. But a schedule of compliance, which is an effluent limitation, is "any restriction established by a State." 33 U.S.C. § 1362(11) (emphasis added); see In re Alexandria Lake Area Sanitary Dist. NPDES/SDS Permit No. MN0040738, 763 N.W.2d 303, 318 (Minn.2009) (holding that even though the TMDL was not yet complete, a schedule of compliance existed where the state pollution control agency required a facility "to meet an effluent limit set upon completion of the TMDL process"). Here, however, the schedule in the restoration plan is set by the Counties. States are not required to set compliance schedules, 40 C.F.R. § 122.47, and MDE has not exercised its discretion to do so.

We recognize that a schedule of compliance is "an enforceable sequence of actions or operations leading to compliance with an effluent limitation, other limitation, prohibition, or standard." 33 U.S.C. § 1362(17). But the "detailed schedule" the Counties must set is subject to change. The Counties may substitute activities where the initially scheduled projects are not making progress in achieving WLAs. Part IV.E.2.b.iv. Consequently, the County-set schedules in the restoration plans are not "enforceable" sequences of actions or operations. What forces the Counties to comply with the WLAs is the annual reporting requirement, which explains that "BMP and program modifications shall be made within 12 months if the [Counties'] annual report[s] do[] not ... show progress toward meeting WLAs developed under EPA approved TMDLs." Part V.A.3. As we have discussed,101 this adaptive management approach is the true enforcement mechanism that "lead[s] to compliance with an effluent limitation [or] other limitation." 33 U.S.C. § 1362(17). Thus, we conclude that the nature of the schedules in the restoration plans does not require MDE to incorporate those plans into the Permits by modification.102

B. 20% Restoration Requirement

The second key component of the Permits that the Water Groups focus on is the 20% restoration requirement.103 This part of the Permits requires the Counties to complete restoration efforts for 20% of the Counties' impervious surface area that is not restored to the maximum extent practicable. Part III.G.2; Part IV.E.2.a. Although not required to do so, see 33 U.S.C. 1342(p), MDE decided to impose this numeric effluent limitation.104

The Water Groups allege that this provision "is not specific, measurable, or enforceable." Thus, they argue, MDE has created a requirement that the Water Groups cannot comment on or seek judicial review of.

Because of our earlier analysis, in part, the Water Groups' arguments are unavailing. Restoration is not "undefined" as the Water Groups argue because MDE anchored restoration in the universe of practices in the Manual. See Part III.E.1 ("At a minimum, the County shall ... [i]mplement the stormwater management ... practices found in the [Manual]....").105 Although the Water Groups do not know in advance which specific practices the Counties will select to restore their impervious surfaces, MDE has permitted the Counties to select from among practices that satisfied a specific performance standard, WQv. CWP & MDE, Manual § 1.2 (General Performance Standards for Stormwater Management in Maryland).

Moreover, it is inaccurate for the Water Groups to allege that MDE can approve activities "that are known to be ineffectual without ever being required to articulate its rationale for doing so, or being held accountable." MDE articulated a response to the Water Groups' criticism of detention practices: "Maryland's Manual for stormwater BMP design and MDE's approach to retrofitting under the municipal permit program are completely aligned with the National Research Council report [stating that detention ponds fail to meet the full range of urban stream and watershed restoration objectives]." MDE, Basis for Final Determination to Issue NPDES Permit. MDE has responded to comments adverse to the draft permit in accordance with Maryland law and its response is adequate. See EN § 1-604(b) ("The Department shall prepare a final determination if" the Department received comments adverse to the tentative determination.). Thus, MDE's approach has not shielded its decision to approve the Counties' restoration efforts from comment or judicial review.

The Water Groups criticize MDE's use of the Guidance because, in their view, the Guidance provides assumptions about stormwater practices rather than an enforceable standard. According to the Water Groups, MDE could not know whether the Counties' efforts would be adequate when it issued the Permits because "whether the chosen practices actually meet these [pollution reduction] expectations depends entirely on the details of a permittee's restoration plans."

This argument, however, must fail because it overlooks the nature of the 20% restoration requirement: a surrogate. ENSR, Pilot TMDL Applications Using the Impervious Cover Method § 1.0, at 1-1 ("The IC [Impervious Cover] method uses percent impervious cover in a watershed as a surrogate TMDL target.").106 Because the 20% restoration requirement is a surrogate for reducing pollution, MDE has logically created an accountability system based on an assessment of compliance with the surrogate, not on assessment of pollution reduction in fact.107 See 33 U.S.C. § 1342(p)(3)(B)(iii) (This law requires "controls to reduce the discharge of pollutants" to the MEP, not "reductions to the MEP."); EDC, 344 F.3d at 855 (The permitting authority must "review these Minimum Measures to ensure that the measures that any given operator of a small MS4 has decided to undertake will in fact reduce discharges to the maximum extent practicable.") (emphasis in original).

To ensure that the 20% requirement is met, MDE requires the Counties to translate the BMPs they implement on impervious surfaces into credits. As we previously explained, the credit system is related to the performance standard in the Manual: the WQv. "An acre for acre impervious credit will be given when a structural BMP is specifically designed to provide treatment for the full WQv (one inch), or [a] proportional acreage of credit will be given when less than the WQv is provided." Moreover, for BMPs that "provide greater than one inch of volume control," the activities "receive additional credit."108 As a result, the Counties must choose the appropriate mix of BMPs to obtain enough credits to satisfy the 20% requirement. If the Counties ignore the credit system and the consequences of selecting BMPs with efficiency estimates below the WQv standard, then the Counties risk failing to fulfill the 20% restoration requirement and confronting an enforcement action by MDE. Part VI.C. The Counties must complete restoration efforts for 20% of their impervious surface area within five years under the express terms of the permit according to the credit system we have explained. The 20% restoration requirement is thus also measurable and enforceable.109 See also EPA, 2014 Memo at 10 (Box 1: Examples of WQBELs in MS4 Permits) ("The MS4 Permit includes a specific, quantifiable performance requirement that must be achieved within a set timeframe. For example: Restore within the 5-year permit term 20 percent of the previously developed impervious land (2014 Prince George's County, MD MS4 permit).").

CONCLUSION

Upholding MDE's decision to issue the Permits, we are guided first and foremost by the law. That law, 33 U.S.C. § 1342, provides a flexible approach to MS4 permitting. Moreover, this approach contemplates that states shall set controls they deem necessary to reduce the discharge of pollutants into their waters. See 33 U.S.C. § 1342(p)(3)(B)(iii). The Water Groups insist that the Permits may not contain effluent limitations "to allow maximum flexibility." But even the EPA, to whom Congress delegated authority to issue regulations pertaining to TMDLs, has afforded permitting agencies such as MDE the flexibility to develop effluent limitations. See 40 C.F.R. § 122.44(d)(1)(vii)(B).

We recognize that the CWA sets forth an important goal of public participation in the "development, revision, and enforcement" of these Permits. 33 U.S.C. § 1251(e). In our consideration of (1) the opportunity to comment in writing and at a public hearing on the draft Permits, (2) MDE's responses to the public's comments, and (3) the opportunity for the public to participate further in the Counties' efforts to meet WLAs, MDE has upheld its part of the bargain. That is, the agency has "provided for, encouraged, and assisted" public participation as Congress envisioned. 33 U.S.C. § 1251(e). Therefore, we conclude that MDE's decision was lawful and that the Permits are valid.

IN CASE NO. 42, JUDGMENT OF THE COURT OF SPECIAL APPEALS REVERSED. CASE REMANDED TO THAT COURT WITH DIRECTIONS TO REVERSE THE JUDGMENT OF THE CIRCUIT COURT FOR MONTGOMERY COUNTY AND AFFIRM MDE'S DECISION TO ISSUE THE MONTGOMERY COUNTY PERMIT. COSTS TO BE PAID BY RESPONDENT.

IN CASE NOS. 43 & 44, JUDGMENTS OF THE CIRCUIT COURTS FOR ANNE ARUNDEL COUNTY, BALTIMORE CITY, BALTIMORE COUNTY, AND PRINCE GEORGE'S COUNTY AFFIRMED. COSTS TO BE PAID BY PETITIONER.