The Cost of Clean Water in the Delaware River Basin (USA)

water

Article The Cost of Clean Water in the Delaware River Basin (USA)

Gerald J. Kauffman ID Water Resources Center, School of Public Policy and Administration, University of Delaware, Newark, DE 19716, USA; [email protected]; Tel.: +1-302-831-4929

Received: 22 September 2017; Accepted: 8 January 2018; Published: 24 January 2018

Abstract: The Delaware River has made a marked recovery in the half-century since the adoption of the Delaware River Basin Commission (DRBC) Compact in 1961 and passage of the Federal Clean Water Act amendments during the 1970s. During the 1960s, the DRBC set a 3.5 mg/L dissolved oxygen criterion for the river based on an economic analysis that concluded that a waste load abatement program designed to meet fishable water quality goals would generate significant recreational and environmental benefits. Scientists with the Delaware Estuary Program have recently called for raising the 1960s dissolved oxygen criterion along the Delaware River from 3.5 mg/L to 5.0 mg/L to protect anadromous American shad and Atlantic sturgeon, and address the prospect of rising temperatures, sea levels, and salinity in the estuary. This research concludes, through a nitrogen marginal abatement cost (MAC) analysis, that it would be cost-effective to raise dissolved oxygen levels to meet a more stringent standard by prioritizing agricultural conservation and some wastewater treatment investments in the Delaware River watershed to remove 90% of the nitrogen load by 13.6 million kg N/year (30 million lb N/year) for just 35% ($160 million) of the $449 million total cost. The annual least cost to reduce nitrogen loads and raise dissolved oxygen levels to meet more stringent water quality standards in the Delaware River totals $45 million for atmospheric NOX reduction, $130 million for wastewater treatment, $132 million for agriculture conservation, and $141 million for urban stormwater retrofitting. This 21st century least cost analysis estimates that an annual investment of $50 million is needed to reduce pollutant loads in the Delaware River to raise dissolved oxygen levels to 4.0 mg/L, $150 million is needed for dissolved oxygen levels to reach 4.5 mg/L, and $449 million is needed for dissolved oxygen levels to reach 5.0 mg/L.

Keywords: river basin; water quality; water pollution; economic valuation

1. Introduction Nutrient pollution due to high loads of nitrogen and phosphorus causes costly impacts on the tourism, commercial ﬁshing, recreation, hunting, real estate, and water treatment sectors of the economy [1]. Noting that 50% of the nation’s streams have medium to high nutrient levels and 78% of coastal waters experience eutrophication, the Environmental Protection Agency (EPA) has urged states to adopt numeric nutrient criteria to reduce nitrogen and phosphorus loads on U.S. waters [2]. Nutrient load reduction costs in the nation’s waters are significant and range from $35 million/year in the 16,500 km2 Wisconsin Fox-Wolf River watershed [3] to $203 million/year in the 26,200 km2 Connecticut River/Long Island Sound Basin [4]. The Chesapeake Bay Program [5] estimated that restoration of the 166,000 km2 Chesapeake Bay watershed could cost $1 billion/year. Rabotyagov et al. [6] estimated a cost of $1.8 billion/year to reduce nutrient loads and increase dissolved oxygen levels in the 492,000 km2 Upper Mississippi River Basin. Lyon and Farrow [7] reported to the EPA that the Federal Clean Water Act stormwater programs could cost up to $14 billion/year nationwide. The Interstate Commission on the Delaware River Basin [8] once called the Delaware River near Philadelphia “one of the most grossly polluted areas in the United States.” In 1961, President John

Water 2018, 10, 95; doi:10.3390/w10020095 www.mdpi.com/journal/water Water 2018, 10, 95 2 of 21

F. Kennedy and the governors of Delaware, New Jersey, New York, and Pennsylvania signed the Delaware River Basin Commission (DRBC) Compact as one of the first models of Federalism or shared power in water management between the Federal government and the states [9]. For over half a century, the DRBC has been empowered by this compulsory 1961 Federal/state compact to oversee water pollution control programs on the Delaware River [10,11]. Water quality has been impaired by nutrient pollution [12–16] but the estuary has recovered in the last several decades due to restoration efforts by DRBC, EPA, and the states [17–19]. A century-long water quality record reconstructed by Sharp [20] indicates that the tidal Delaware has made one of the most extensive recoveries of any estuary in the world as dissolved oxygen levels declined to zero during the 1950s and 1960s and increased to near 400 µmol/L by the turn of the 21st century. While pollutant loadings have decreased and water quality has measurably improved in the Delaware Estuary since the adoption of the 1961 DRBC Compact, dissolved oxygen levels still do not fully meet the criterion of 3.5 mg/L during the summer when dissolved oxygen saturation declines with warming water temperatures. Scientists with the Delaware Estuary Program and the DRBC have discussed setting more rigorous dissolved oxygen criteria along the tidal Delaware River (to at least 5.0 mg/L) to protect the year-round propagation of anadromous fish such as the American shad and Atlantic sturgeon [21,22]. More stringent dissolved oxygen criteria would also address the prospect of atmospheric warming and rising sea levels that are projected to increase water temperatures, raise salinity, and further depress dissolved oxygen saturation. While dissolved oxygen levels have recovered over the last half-century, little is known about modern costs to restore the Delaware River to meet fishable water quality standards. The objectives of this research are to estimate the costs of investments to reduce pollutant loads and restore the Delaware River to meet more protective year-round fishable dissolved oxygen criteria in accordance with DRBC, EPA, and state water quality standards.

2. The Delaware River While just the 33rd largest river in the United States, the Delaware River is the longest undammed river east of the Mississippi, extending 390 mi (628 km) from the 3000 ft (970 m) high Catskill Mountains in New York State to the mouth of the Delaware Bay at Cape May, New Jersey. The river is fed by 216 streams including its two largest tributaries (the Schuylkill and Lehigh River) and drains 13,539 mi2 (35,077 km2) in Delaware, Pennsylvania, New Jersey, New York, and a small part of Maryland. The Delaware Basin covers just 0.4% of the continental U.S., yet supplies drinking water to 5% of the nation’s population and the ﬁrst (New York City) and seventh (Philadelphia) largest metropolitan economies in the nation [23]. Over 16 million people rely on the Delaware Basin for drinking water, including 8.2 million people who live in the watershed and 8 million people who live outside the basin in New York City and central New Jersey. Between 2000 and 2010, the population in the Delaware Basin increased by half a million people, an amount equal to the combined population of the cities of Camden, Trenton, and Wilmington. The Delaware Estuary extends 130 mi (208 km) from the Atlantic Ocean to the head of tide at Trenton [24]. High nutrient loads discharged from tributaries near Philadelphia and rural streams along the bay are diluted by saltwater as the estuary widens toward the mouth of the bay [13]. The Delaware Estuary recirculates every 8 days (Table1) with half mixing with freshwater from the Delaware River at Trenton, Schuylkill, Lehigh, Brandywine, and smaller tributaries and the other half from the Atlantic Ocean [17]. The estuary is relatively turbid with a light extinction coefﬁcient of 0.3–7.0 [25]. Water 2018, 10, 95 3 of 21

Water 2018, 10, x FOR PEER REVIEW 3 of 20 Table 1. Characteristics of the Delaware River [17,25]. Table 1. Characteristics of the Delaware River [17,25].

CharacteristicCharacteristic Value Value DrainageDrainage Area (km Area2) (km2) 35,25235,252 PopulationPopulation (2010) (2010) 8,200,000 8,200,000 Total Length (km) 628 Total Length (km) 628 Tidal Length (km) 155 Watershed/EstuaryTidal Length Ratio (km) 155 18 Estuary RecirculationWatershed/Estuary (days) Ratio 18 8 Light ExtinctionEstuary Recirculation Coefficient (days) 8 0.3–7.0 Light Extinction Coefficient 0.3–7.0 The DRBC [26,27] classifies the Delaware River and Bay according to 10 nontidal and tidal water The DRBC [26,27] classifies the Delaware River and Bay according to 10 nontidal and tidal water quality management zones based on (a) Agricultural, Industrial, and Public Water Supply; (b) Wildlife, quality management zones based on (a) Agricultural, Industrial, and Public Water Supply; (b) Fish and Aquatic Life; (c) Recreation (Swimming, Boating, Fishing, Wading); (d) Navigation; and Wildlife, Fish and Aquatic Life; (c) Recreation (Swimming, Boating, Fishing, Wading); (d) Navigation; (e) Waste Assimilation designated uses (Figure1). In the tidal Delaware, the summer dissolved oxygen and (e) Waste Assimilation designated uses (Figure 1). In the tidal Delaware, the summer dissolved criterionoxygen varies criterion from varies 3.5 mg/L from in3.5 Zones mg/L 3 in and Zones 4 (from 3 and Rancocas 4 (from CreekRancocas past Creek Philadelphia past Philadelphia to Wilmington) to to 4.5Wilmington) mg/L in Zoneto 4.5 5 (frommg/L Wilmingtonin Zone 5 (from to the Wilmington Chesapeake to & the Delaware Chesapeake Canal). & MinimumDelaware dissolvedCanal). oxygenMinimum criteria dissolved are 6.5 mg/Loxygen during criteria spring are 6.5 and mg/L fall during in Zones spring 2 through and fall 5 toin allowZones for2 through seasonal 5 to spawning allow andfor propagation seasonal spawning of resident and andpropagation anadromous of resident fish. and anadromous fish.

FigureFigure 1. 1.Delaware Delaware RiverRiver waterwater quality management management zones zones [27]. [27 ].

Water 2018, 10, 95 4 of 21

Water 2018, 10, x FOR PEER REVIEW 4 of 20 Despite high nutrient loading, the Delaware Estuary does not exhibit classic eutrophication symptomsDespite of hypoxia high nutrient or algal loading, blooms the as Delaware observed inEstuary the nearby does not Chesapeake exhibit classic Bay. Algaleutrophication blooms are inhibitedsymptoms by theof hypoxia assimilative or algal capacity blooms of as wetlands observed that in rimthe nearby the Delaware Chesapeake Bay and Bay. by Algal low blooms light levels are in theinhibited well-flushed by the and assimilative turbid Delaware capacity of Estuary. wetlands Through that rim the the wide Delaware 17 mi Bay (27 and km) by mouth low light of Delaware levels Bay,in the well-flushed Atlantic Ocean and contributes turbid Delaware significant Estuary. tidal Through flushing, the thus wide limiting 17 mi (27 algal km) blooms mouth of that Delaware cause fish killsBay, except the Atlantic during Ocean an occasional contributes spring significant bloom intidal the flushing, mid estuary thus [ 17limiting]. algal blooms that cause fishDuring kills except the 1960s during when an occasional the river wasspring anoxic bloom and in the a decade mid estuary before [17]. the 1970s Federal Clean Water Act Amendments,During the 1960s the DRBC when imposedthe river was waste anoxic load an allocationsd a decade on before 80 dischargers the 1970s Federal and adopted Clean Water the first interstateAct Amendments, water quality the DRBC standards imposed along waste the load Delaware allocations River. on The80 dischargers Federal Water and adopted Pollution the Control first Administrationinterstate water (FWPCA) quality standards and Harvard along Water the ProgramDelaware [28River.–31] The issued Federal an economic Water Pollution report in Control 1966 that concludedAdministration that water (FWPCA) supply and and Harvard recreation Water benefits Progra duem [28–31] to improved issued wateran economic quality report in the Delawarein 1966 Riverthat wouldconcluded exceed that water water pollution supply and control recreation costs [32 benefits–35]. due to improved water quality in the DelawareThe 1966 River FWPCA would study exceed estimated water pollution pollutant control reduction costs costs [32–35]. ranged from $150 million to achieve a The 1966 FWPCA study estimated pollutant reduction costs ranged from $150 million to achieve dissolved oxygen level of 2.5 mg/L to $490 million to achieve a dissolved oxygen criterion of 4.5 mg/L a dissolved oxygen level of 2.5 mg/L to $490 million to achieve a dissolved oxygen criterion of 4.5 mg/L with diminishing marginal costs of improvement occurring at dissolved oxygen of 3.0 mg/L (Table2). with diminishing marginal costs of improvement occurring at dissolved oxygen of 3.0 mg/L (Table 2). Thomann [33] estimated that shad passage would achieve 80% survival if dissolved oxygen improved Thomann [33] estimated that shad passage would achieve 80% survival if dissolved oxygen improved from 0.5 mg/L in 1964 to a future level of 3.0 mg/L. In 1967, the DRBC considered this economic from 0.5 mg/L in 1964 to a future level of 3.0 mg/L. In 1967, the DRBC considered this economic analysis analysisand set andthe current set the dissolved current dissolvedoxygen standard oxygen of standard3.5 mg/L in of the 3.5 Delaware mg/L in River the near Delaware Philadelphia River to near Philadelphiasupport the tospring support and the fall spring migration and fallof anadro migrationmous of fish. anadromous In 1968, the fish. DRBC In 1968, quite the presciently DRBC quite prescientlyanticipated anticipated that the waste that theload waste abatement load abatementplan would plan remove would 85% remove to 90% 85%of carbonaceous to 90% of carbonaceous BOD and BODboost and dissolved boost dissolved oxygen from oxygen near fromzero to near 4.0 zeromg/L to at4.0 Philadelphia mg/L atPhiladelphia (Figure 2). (Figure2).

FigureFigure 2. 2.Delaware Delaware River River Basin Basin CommissionCommission (DRBC) disso dissolvedlved oxygen oxygen criteria criteria along along the the Delaware Delaware Estuary in 1968 [22]. Estuary in 1968 [22].

Table 2. Costs to meet water quality objectives in the Delaware Estuary [28]. Table 2. Costs to meet water quality objectives in the Delaware Estuary [28]. Dissolved BOD/COD Total Marginal % Objective Oxygen Residual % BOD/COD Costs Costs Survival Dissolved BOD/COD Residual % BOD/COD Total Costs Marginal % Survival ObjectiveSet Set CriteriaOxygen Criteria PollutionPollution Removal ($1964)($1964) Costs($1964) ($1964) Shad (lb/Day)(lb/Day) (kg/Day) (kg/Day) Shad Passage (mg/L)(mg/L) Removal ($M/Year)($M/Year) ($M/Year)($M/Year) Passage I. 4.5 100,000 45,360 92–98% 490 160–260 I.II. 4.5 4.0100,000 200,000 45,360 90,72092–98% 90% 230–330 490 100–150160–260 90% II.III. 4.0 3.0200,000 500,000 90,720 226,800 90% 75% 130–180 230–330 100–150 30–30 90% 80% III.IV. 3.0 2.5500,000 800,000 226,800 362,880 75% 50% 100–150 130–180 70–12030–30 80% V. 0.5 status quo 30 0 20% IV. 2.5 800,000 362,880 50% 100–150 70–120 V. 0.5 status quo 30 0 20% Dissolved oxygen levels in the Delaware Estuary vary by water temperature, sunlight, winds, and pollutantDissolved loads oxygen [1]. By levels 2010, in dissolved the Delaware oxygen Estuary levels vary in the by Delawarewater temperature, River at Ben sunlight, Franklin winds, Bridge and pollutant loads [1]. By 2010, dissolved oxygen levels in the Delaware River at Ben Franklin Bridge

Water 2018, 10, 95 5 of 21

at PhiladelphiaWater 2018, 10, x mostly FOR PEER exceeded REVIEW the criterion except for violations below 3.5 mg/L during5 of the 20 hot summerWater months2018, 10, x FOR of June PEER throughREVIEW August (Figure3). During warm summers, 0.5% of readings5 of 20 since 2000at did Philadelphia not meet the mostly 3.5 mg/L exceeded criterion. the criterion In July exce andpt August, for violations dissolved below oxygen 3.5 mg/L in theduring Delaware the hot River at Philadelphia mostly exceeded the criterion except for violations below 3.5 mg/L during the hot at Philadelphiasummer months occasionally of June through declined Au belowgust (Figure the 3.5 3). mg/L During criterion warm summers, (46% dissolved 0.5% of oxygenreadings saturation) since summer months of June through August (Figure 3). During warm summers, 0.5% of readings since when2000 water did temperaturesnot meet the 3.5 approached mg/L criterion. 30 In◦C July or and 86 ◦ FAugust, (Figure dissolved4). At 30 oxygen◦C, dissolved in the Delaware oxygen River is 100% at2000 Philadelphia did not meet occasionally the 3.5 mg/L declined criterion. below In July the 3.5and mg/L August, criterion dissolved (46% oxygen dissolved in theoxygen Delaware saturation) River saturated at 7.54 mg/L and 80% saturated at 6 mg/L; therefore, when water temperatures rise to 30 ◦C, whenat Philadelphia water temperatures occasionally approached declined below 30 °C the or 3.586 °Fmg/L (Figure criterion 4). At (46% 30 °C, dissolved dissolved oxygen oxygen saturation) is 100% a future DRBC dissolved oxygen standard higher than 5 mg/L (66% saturation) may prove difficult to saturatedwhen water at 7.54temperatures mg/L and approached 80% saturated 30 °C at or6 mg/L; 86 °F (Figuretherefore, 4). whenAt 30 water°C, dissolved temperatures oxygen rise is 100%to 30 achieve°C,saturated givena future at the 7.54DRBC warm mg/L dissolved water and 80% temperatures oxygen saturated standard at that 6 mg/L; occurhigher therefore, during than 5 summer.whenmg/L water(66% temperaturessaturation) may rise prove to 30 difficult°C,Scientists a future to achieve on DRBC the Delawaregiven dissolved the warm Estuaryoxygen water standard Program temperatures higher Science that than and occur 5 Technicalmg/L during (66% summer. Advisory saturation) Committee may prove have recommendeddifficultScientists to achieve that on the the given DRBCDelaware the warm raise Estuary thewater fishable Program temperatures dissolved Science that and occur oxygen Technical during standard Advisory summer. from Committee the current have level of 3.5recommended mg/LScientists to at leastonthat the the 5.0 Delaware DRBC mg/L raise inEstuary Zones the fishable Program 3 and diss 4 Science fromolved Philadelphia oxygenand Technical standard to Advisory Wilmington,from the Committee current given level have that of the literature3.5recommended mg/L suggests to at thatleast that the the5.0 DRBC mg/L current raisein dissolvedZones the fishable 3 and oxygen 4diss fromolved criterion Philadelphia oxygen of standard 3.5 to mg/L Wilmington, from is toothe low currentgiven to supportthatlevel the of the 3.5 mg/L to at least 5.0 mg/L in Zones 3 and 4 from Philadelphia to Wilmington, given that the year-roundliterature survival suggests of that anadromous the current shaddissolved and sturgeonoxygen criterion [36,37]. of Secor 3.5 mg/L and is Gunderson too low to support [38] found the that literature suggests that the current dissolved oxygen criterion of 3.5 mg/L is too low to support the juvenileyear-round Atlantic survival sturgeon of anadromous may suffer shad over and 50% sturgeon mortality [36,37]. at 25 Secor◦C (77and◦ F)Gunderson when dissolved [38] found oxygen that is year-round survival of anadromous shad and sturgeon [36,37]. Secor and Gunderson [38] found that 3.5 mg/L.juvenile Juvenile Atlantic shortnosesturgeon may sturgeon suffer over areprone 50% mortality to 50% mortalityat 25 °C (77 when °F) when dissolved dissolved oxygen oxygen declines is 3.5juvenile mg/L. Atlantic Juvenile sturgeon shortnose may sturgeon suffer over are prone50% mortality to 50% mortality at 25 °C (77when °F) dissolvedwhen dissolved oxygen oxygen declines is below 3.0 mg/L at 25 ◦C[39]. In 2017, the DRBC passed a resolution authorizing basin commission below3.5 mg/L. 3.0 Juvenilemg/L at 25shortnose °C [39]. sturgeonIn 2017, theare DRBCprone passedto 50% amortality resolution when authorizing dissolved basin oxygen commission declines scientists to begin reviewing water quality regulations to determine whether dissolved oxygen criteria scientistsbelow 3.0 tomg/L begin at 25reviewing °C [39]. Inwater 2017, quality the DRBC regulations passed a toresolution determine authorizing whether basindissolved commission oxygen shouldcriteriascientists be increasedshould to begin be increased fromreviewing the from current water the currentquality 3.5 mg/L 3.5 re gulationsmg/L to ato highera higherto determine level level to whetherprovide provide moredissolved more protection protection oxygen of of anadromousanadromouscriteria should fish fish spawning be spawningincreased and andfrom year-round year-round the current propagation propagation 3.5 mg/L to aof higher the the fishery. fishery. level to provide more protection of anadromous fish spawning and year-round propagation of the fishery.

Figure 3. Mean Daily Dissolved Oxygen (DO) at Ben Franklin Bridge along Delaware River. FigureFigure 3. Mean 3. Mean Daily Daily Dissolved Dissolved Oxygen Oxygen (DO)(DO) at Ben Ben Franklin Franklin Bridge Bridge along along Delaware Delaware River. River.

Figure 4. Water temperature/dissolved oxygen along the Delaware River. FigureFigure 4. Water4. Water temperature/dissolved temperature/dissolved oxygen oxygen along along the the Delaware Delaware River. River.

Water 2018, 10, 95 6 of 21

3. Research Objectives While dissolved oxygen levels have recovered over the last half-century, little is known about modern costs to restore the Delaware River to meet ﬁshable water quality standards. The objectives of this research are to estimate the costs of investments to reduce pollutant loads and restore the Delaware River to meet more protective year-round ﬁshable dissolved oxygen criteria in accordance with the Delaware River Basin Commission, Environmental Protection Agency, and state water quality standards.

4. Methods With the following research, we estimate the modern costs of the nitrogen pollutant load reductions necessary to increase dissolved oxygen from the current criterion (3.5 mg/L) to a future, more stringent water quality standard (5.0 mg/L) in the Delaware River. To estimate the most cost-effective combination of nitrogen load reductions, we (1) quantified nitrogen loads in the Delaware Basin from atmospheric, urban/suburban, wastewater, and agricultural sources and estimated the pollutant load reductions needed to improve dissolved oxygen in the Delaware River from the current 3.5 mg/L to a future, more protective standard; (2) estimated the costs of nitrogen load reductions to improve the dissolved oxygen levels in the tidal Delaware River for various best management practice scenarios; and (3) constructed marginal abatement cost curves to define the annual least costs to raise the dissolved oxygen levels to more stringent fishable criteria. Nitrogen Loads: We estimated annual nitrogen loads in the Delaware Basin in Delaware, New Jersey, New York, and Pennsylvania using the USGS SPAtially Referenced Regressions on Watershed (SPARROW) model [40]. The SPARROW model has been calibrated by the USGS to estimate nitrogen loads for the base year 2002 from point sources (wastewater discharges) and nonpoint sources (atmospheric deposition, agriculture fertilizer/manure, and urban/suburban land) and accounts for watershed characteristics such as precipitation, temperature, soil permeability, stream density, flow rate, velocity, and lake/reservoir hydraulics [41]. The USGS SPARROW model simulates nitrogen removal based on hydrological processes such as denitrification, particulate settling, and water velocity [42]. SPARROW is a nonlinear least squares regression model where the mean annual N load, as the dependent variable, is weighted by land-to-water movement, instream transport, and assimilation of nitrogen as the explanatory variable (Table3). Since the USGS SPARROW model calibrates the nitrogen load estimates with EPA STORET water quality monitoring data, the model is well correlated as coefficients of determination (r2) are 0.83 for yield and 0.97 for load, which explains 83% to 97% of the variance between the predictive model and observed water quality data.

Table 3. Mid-Atlantic SPAtially Referenced Regressions on Watershed (SPARROW) model coefficients [40].

Parameter Coefficient Unit Model Coefficient Nitrogen Sources Developed land (km2) kg/km2/year 1422 Wastewater discharge (kg/year) 1.16 Fertilizer and fixation from agriculture in corn, soybeans, alfalfa (kg/year) 0.310 Fertilizer to agriculture in other crops (kg/year) 0.186 Manure from livestock (kg/year) 0.090 Land to Water Delivery Mean annual temperature ln deg C −0.864 Average overland flow distance to stream (km) km−1 −0.190 ln ratio of nitrate to inorganic N wet deposition 2.56 Aquatic Decay Time of travel in stream reach where mean discharge <2.83 m3/s (days) per day 0.224 Statistics Root Mean Square Error (RSME) 0.35 r2 load 0.97 r2 yield 0.83 Water 2018, 10, 95 7 of 21

Water 2018, 10, x FOR PEER REVIEW 7 of 20 Nitrogen Load Reduction Costs: We estimated the N load reductions needed to improve water qualityNitrogen to Load meet Reduction a future 5.0 Costs: mg/L We dissolved estimated oxygen the N load standard reduct inions the needed Delaware to improve River betweenwater Philadelphiaquality to meet and Wilmington.a future 5.0 Wemg/L examined dissolved the ox EPAygen Water standard Quality in the Analysis Delaware Simulation River between Program (WASP),Philadelphia USGS Hydrologicaland Wilmington. Simulation We examined Program-Fortran the EPA Water (HSPF), Quality and GeneralizedAnalysis Simulation Watershed Program Loading Function(WASP), (GWLF) USGS toHydrological estimate Total Simulation Maximum Progra Dailym-Fortran Load (TMDL) (HSPF), pollutant and Generalized load reductions Watershed in the lowerLoading Delaware Function Basin (GWLF) [43]. These to estimate hydrodynamic Total Maximu modelsm Daily suggest Load “better-than-secondary” (TMDL) pollutant load reductions treatment is in the lower Delaware Basin [43]. These hydrodynamic models suggest “better-than-secondary” needed to meet a more stringent dissolved oxygen water quality standard of 5 mg/L in the Delaware treatment is needed to meet a more stringent dissolved oxygen water quality standard of 5 mg/L in the River at Philadelphia. Delaware River at Philadelphia. We reviewed a survey of 15 TMDL models by Scatena et al. [44] in the lower Delaware River that We reviewed a survey of 15 TMDL models by Scatena et al. [44] in the lower Delaware River suggests that achieving a dissolved oxygen target of 5.0 mg/L would require a 32% (median) reduction that suggests that achieving a dissolved oxygen target of 5.0 mg/L would require a 32% (median) in nitrogenreduction loading in nitrogen to water loading bodies, to water within bodies, a range with fromin a range 20% from (25th 20% percentile) (25th percentile) to 48% (75th to 48% percentile) (75th reductionpercentile) (Figure reduction5). Similarly, (Figure the5). Brandywine-ChristinaSimilarly, the Brandywine-Christina watershed TMDL watershed model TMDL estimated model that a 38%estimated reduction that a in 38% nitrogen reduction loads in nitrogen is needed loads to meetis needed a dissolved to meet a oxygen dissolved water oxygen quality water criterion quality of 5 mg/Lcriterion in theof 5watershed mg/L in the that watershed contributes that contribute 8% of thes N 8% load of the to theN load Delaware to the Delaware Estuary [45 Estuary]. [45].

Nitrogen Reduction from TMDLs Lower Delaware River 100

40 % N Reduction % N 20

0 1

FigureFigure 5. 5.Nitrogen Nitrogen load load reductions reductions fromfrom TotalTotal MaximumMaximum Daily Daily Load Load (TMDL) (TMDL) models models for for the the lower lower Delaware River [44]. Delaware River [44].

Best Management Practice (BMP) Installation Costs: We derived the unit costs of N load reductionsBest Management ($/lb N reduced) Practice for point (BMP) source Installation wastewater Costs: treatmentWe derivedBMPs and the nonpoint unit costs source of BMPs N load reductionssuch as atmospheric ($/lb N reduced) controls for (v pointehicle source exhaust wastewater and industrial treatment plant BMPs scrubbers), and nonpoint urban stormwater source BMPs suchretrofitting, as atmospheric stream restoration, controls (vehicle wetlands, exhaust and agricult and industrialural practices plant such scrubbers), as no till, cover urban crops, stormwater forest retroﬁtting,buffers, and stream animal restoration, waste management wetlands, (Table and 4). agricultural practices such as no till, cover crops, forest buffers, and animal waste management (Table4). Table 4. Nitrogen reduction best management practices. Table 4. Nitrogen reduction best management practices. Nitrogen Source Best Management Practice (BMP) Wastewater Treatment Plant Nutrient Reduction Technology Nitrogen Source Best Management Practice (BMP) Motor vehicle exhaust controls Atmospheric Deposition Wastewater Treatment PlantPower/industrial Nutrient Reduction plant scrubbers Technology Motor vehicle exhaust controls Atmospheric Deposition Ag Nutrient Management Plans Power/industrialConservation Tillage plant scrubbers Ag NutrientCover Crops Management Plans Conservation Tillage Agricultural Conservation Diversions Cover Crops Agricultural Conservation ForestDiversions Buffers GrassForest Buffers Buffers TerracesGrass Buffers Terraces Wet Detention Pond GrassWet Detention Swale Pond Grass Swale InfiltrationInﬁltration Basin Basin Urban/SuburbanUrban/Suburban Stormwater Stormwater SepticSeptic System System Replacement Replacement StormwaterStormwater Wetland Wetland Vegetated Filter Strip Vegetated Filter Strip

Water 2018, 10, x FOR PEER REVIEW 8 of 20 Water 2018, 10, 95 8 of 21 We calculated the costs to reduce nitrogen loads by 20% (25th percentile), by 32% (median), and by 48% (75th percentile) to meet a more stringent DRBC dissolved oxygen standard by multiplying N loadWe calculatedreduction therates costs (kg/year) to reduce by nitrogenthe unit loadscost ($/kg) by 20% (in (25th 2010 percentile), dollars) for by 32%atmospheric (median), NOX and byreduction, 48% (75th wastewater percentile) totreatment, meet a more agricultur stringente DRBCconservation, dissolved and oxygen urba standardn/suburban by multiplying BMPs. The N loadcombined reduction Total rates Maximum (kg/year) Daily by Load the unit (TMDL) cost ($/kg) models (in for 2010 the dollars) Delaware for River atmospheric required NOX nitrogen reduction, load wastewaterreductions of treatment, 32% (median) agriculture within conservation,a range of 20% and (25th urban/suburban percentile) and BMPs.48% (75th The percentile). combined Using Total Maximumbenefit (value) Daily transfer Load (TMDL) principles, models we reviewed for the Delaware the literature River [46–49] required and nitrogen translated load nitrogen reductions load of 32%reduction (median) costs within ($/kg) a from range nearby of 20% watersheds (25th percentile) (such andas the 48% Chesapeake (75th percentile). Bay) to Using the Delaware benefit (value) River transferwatershed principles, since the weadjacent reviewed watersheds the literature share similar [46–49] climate, and translated soils, topography, nitrogen load physiography, reduction costs and ($/kg)hydrogeology. from nearby watersheds (such as the Chesapeake Bay) to the Delaware River watershed since the adjacentTotal Nitrogen watersheds Load share Reduction similar climate,Costs: We soils, estimated topography, the costs physiography, to reduce nitrogen and hydrogeology. loads by the medianTotal 32% Nitrogen by maximizing Load Reduction load reductions Costs: We from estimated least cost the agriculture costs to reduce and wastewater nitrogen loads sources by thefor medianthe following 32% by five maximizing options. load We reductionsestimated fromthe nitrogen least cost agricultureload reduction and wastewatercosts by multiplying sources for the followingnecessary fivenitrogen options. load We reduction estimated from the nitrogenthe SPARROW load reduction model costs(kg N/year) by multiplying by the nitrogen the necessary load nitrogenreduction load costs reduction ($/kg N). fromOption the 1 would SPARROW reduce model nitrogen (kg loads N/year) equally by the by nitrogenthe median load 32% reduction from all costssources ($/kg (agriculture, N). Option wastewater, 1 would reduce atmospheric, nitrogen an loadsd urban/suburban equally by the stormw medianater). 32% fromOption all 2 sources would (agriculture,reduce nitrogen wastewater, loads from atmospheric, agriculture andby 32%, urban/suburban wastewater by stormwater). 47%, atmospheric Option deposition 2 would reduceby 5%, nitrogenand urban/suburban loads from agriculturestormwater byby 32%,5%. Option wastewater 3 would by 47%,reduce atmospheric nitrogen loads deposition from agriculture by 5%, and by urban/suburban60%, wastewater stormwaterby 29%, atmospheric by 5%. Option deposition 3 would by reduce 5%, and nitrogen urban/suburban loads from agriculturestormwater byby 60%, 5%. wastewaterOption 4 would by 29%, reduce atmospheric nitrogen deposition loads from by agriculture 5%, and urban/suburban by 75%, wastewater stormwater by 20%, by 5%.atmospheric Option 4 woulddeposition reduce by nitrogen5%, and loadsurban/suburban from agriculture stormwater by 75%, by wastewater 5%. Options by 20%,5 would atmospheric reduce nitrogen deposition loads by 5%,from and agriculture urban/suburban by 90%, stormwaterwastewater by by 5%. 10%, Options atmospheric 5 would deposition reduce nitrogen by 5%, loads and fromurban/suburban agriculture bystormwater 90%, wastewater by 5%. by 10%, atmospheric deposition by 5%, and urban/suburban stormwater by 5%. MarginalMarginal AbatementAbatement Costs: We constructed nitrogen marginalmarginal abatementabatement costcost (MAC)(MAC) curvescurves toto determinedetermine the mostmost cost-effectivecost-effective N loadload reductionsreductions toto improveimprove waterwater qualityquality byby raisingraising dissolveddissolved oxygenoxygen inin thethe DelawareDelaware RiverRiver toto moremore stringentstringent fishablefishable criteria.criteria. TheThe marginalmarginal costcost curvescurves showshow thethe changechange inin costcost comparedcompared withwith thethe changechange inin reducedreduced pollutantpollutant loads.loads. TheThe MACMAC curvecurve isis constructedconstructed byby plottingplotting pollutantpollutant loadload reductionsreductions byby percentagepercentageor or in in kg/year kg/year (lb/year)(lb/year) for the practices and by thethe annualannual costscosts ofof thesethese measures.measures. The MAC curvescurves are constructedconstructed by plottingplotting NN loadload reductionreduction costscosts ($/year)($/year) on thethe horizontalhorizontal axisaxis and and 25th, 25th, 50th 50th (median), (median), and and 75th 75th percentile percentile N loadN load reductions reductions on theon verticalthe vertical axis. axis. Typical Typical marginal marginal abatement abatement cost curves cost [50curves] depict [50] relatively depict relatively inexpensive inexpensive measures onmeasures the left on and the more left expensiveand more expens measuresive tomeasures the right to (Figure the right6). (Figure 6).

Figure 6. Nitrogen marginal abatement costcost (MAC)(MAC) curvecurve inin thethe NetherlandsNetherlands [[50].50].

5. Results 5. Results Results from our use of the USGS SPARROW model indicates that the highest nitrogen loads in Results from our use of the USGS SPARROW model indicates that the highest nitrogen loads the Delaware River watershed are delivered in the Schuylkill River and Brandywine-Christina in the Delaware River watershed are delivered in the Schuylkill River and Brandywine-Christina watersheds in southeastern Pennsylvania (Figure 7). The Delaware River receives the second-highest watersheds in southeastern Pennsylvania (Figure7). The Delaware River receives the second-highest nitrogen load of any river basin along the Atlantic Coast of the USA (Table 5). The SPARROW model

Water 2018, 10, 95 9 of 21

Water 2018, 10, x FOR PEER REVIEW 9 of 20 nitrogen load of any river basin along the Atlantic Coast of the USA (Table5). The SPARROW model utilizesutilizes land cover cover data data to to predict predict the the nitrogen nitrogen load loadss from from urban/suburban urban/suburban and andagricultural agricultural runoff. runoff. Our OurGIS analysis GIS analysis utilizing utilizing 2006 NOAA 2006 NOAACoastal Services Coastal Center Services (CSC) Center land (CSC) use/land land cover use/land data indicates cover data that indicatesthe Delaware that the River Delaware Basin Riverwas Basincovered was by covered 63% byforest/wetlands, 63% forest/wetlands, 20% agriculture, 20% agriculture, and and17% 17%urban/suburban urban/suburban land. land. Pennsylvania Pennsylvania covers covers 51%, 51%, New New Jersey Jersey and and New New York York each each cover cover 21%, andand DelawareDelaware coverscovers 8%8% ofof thethe DelawareDelaware RiverRiver Basin.Basin. TheThe spatialspatial distributiondistribution andand fragmentationfragmentation ofof landland covercover typestypes affectaffect pollutantpollutant loadsloads toto waterways;waterways; forfor example,example, farmedfarmed areasareas nextnext toto streamsstreams willwill deliverdeliver aa higherhigher nitrogennitrogen yieldyield thatthat cancan bebe ﬁlteredfiltered byby riparianriparian forestedforested bufferbuffer areas.areas. BasedBased onon thethe deliverydelivery fractionfraction ofof nitrogennitrogen (proportion(proportion ofof NN loadload delivereddelivered toto thethe outlet),outlet), BMPsBMPs implementedimplemented inin watershedswatersheds closestclosest to the Delaware Estuary provide provide the the most most immediate immediate improvements improvements in in water water quality. quality.

Figure 7. Incremental delivered nitrogen yield to Delaware Basin from SPARROW model. Figure 7. Incremental delivered nitrogen yield to Delaware Basin from SPARROW model.

Table 5. Nitrogen loads in Atlantic Coast river basins from USGS SPARROW model. Table 5. Nitrogen loads in Atlantic Coast river basins from USGS SPARROW model. Drainage Area Nitrogen Load Unit N Load River Basin (km2) (kg/Year)Nitrogen Load(kg/km2/Year)Unit N Load River Basin Drainage Area (km2) Susquehanna 71,199 66,320,320(kg/Year) 931 (kg/km2/Year) SusquehannaDelaware 71,19930,611 45,876,700 66,320,320 1499 931 DelawarePotomac 30,61137,964 40,593,956 45,876,700 1069 1499 PotomacHudson 37,96434,610 26,069,588 40,593,956 753 1069 HudsonJames 34,61026,778 15,873,656 26,069,588 593 753 JamesConnecticut 26,77829,166 15,650,288 15,873,656 537 593 Connecticut 29,166 15,650,288 537 In the Delaware Basin, our analysis using the SPARROW model indicates that almost half (46%) of theIn nitrogen the Delaware flows Basin, from our wastewater analysis using discharges the SPARROW and a third model (29%) indicates emanates that almostfrom agriculture half (46%) offertilizer/manure the nitrogen ﬂows runoff from (Figure wastewater 8). Urban/suburban discharges stormwater and a third from (29%) the emanates cities and from suburbs agriculture delivers fertilizer/manurejust over 10% of the runoff N load. (Figure The8 ).airshed Urban/suburban of the Delaware stormwater River is from 10 times the cities larger and than suburbs its watershed delivers justand overatmospheric 10% of the deposition N load. The from airshed industries of the and Delaware vehicles River contributes is 10 times 12% larger of the than nitrogen its watershed to the andestuary. atmospheric Together, deposition Pennsylvania from and industries New Jersey and vehicles discharge contributes over 90% 12% of the of the nitrogen nitrogen to tothe the Delaware estuary. Together,Basin with Pennsylvania half from wastewater and New Jerseydischarges discharge and a over quarter 90% to of a the third nitrogen from agriculture. to the Delaware New Basin York with and halfDelaware from wastewatercontribute 4% discharges and 3% of and the a nitrogen, quarter to respectively. a third from agriculture. New York and Delaware contributeFrom 4%the andUSGS 3% SPARROW of the nitrogen, model, respectively. we found that three watersheds—the Delaware River at Trenton,From Schuylkill the USGS River, SPARROW and above-Philadelphia model, we found tributaries—deliver that three watersheds—the 80% of the Delaware nitrogen River load atto Trenton,the estuary. Schuylkill Above River,Trenton, and the above-Philadelphia Lehigh River contri tributaries—deliverbutes 9% of the N 80%load ofto thethe nitrogenDelaware load River. to theBelow estuary. Philadelphia, Above Trenton, the Brandywine-Christina, the Lehigh River contributes Delaware 9%River of above the N loadWilmington, to the Delaware and Delaware River. BelowEstuary Philadelphia, at Prime Hook the Brandywine-Christina, watersheds contribute Delaware 7%, 8%, Riverand 3% above of Wilmington,the N loads, and respectively. Delaware EstuaryWastewater at Prime discharges Hook watersheds are the predominant contribute 7%,sources 8%, andof N 3% in the of the Delaware N loads, River respectively. above Philadelphia Wastewater discharges(82%), Schuylkill are the predominant(46%), and above-Wilmington sources of N in the Delaware (68%) watersheds. River above Agriculture Philadelphia is (82%), the primary Schuylkill N source in the Delaware River at the Trenton (34%), Brandywine-Christina (77%), and Delaware Bay at Prime Hook (72%) watersheds and the second-highest N source in the Schuylkill watershed (35%).

Water 2018, 10, 95 10 of 21

(46%), and above-Wilmington (68%) watersheds. Agriculture is the primary N source in the Delaware River at the Trenton (34%), Brandywine-Christina (77%), and Delaware Bay at Prime Hook (72%) watersheds and the second-highest N source in the Schuylkill watershed (35%). Water 2018, 10, x FOR PEER REVIEW 10 of 20

Nitrogen Loads Delaware River Basin

Atmospheric Deposition 6,063 tons/yr Agriculture 12% 14,625 tons/yr 29%

Wastewater Suburban/ 23,241 tons/yr Urban 45% 7,073 tons/yr 14%

Figure 8. Annual nitrogen loads in the Delaware Basin from USGS SPARROW Model. Figure 8. Annual nitrogen loads in the Delaware Basin from USGS SPARROW Model. We observed that nitrogen loads at USGS gages compare within 25–30% (±) of modeled loads from WeSPARROW observed [51]. that The nitrogen observed loads annual at N USGS load gagesat Delaware compare River withinat Trenton 25–30% is 14.2 ( ±million) of modeled kg (31.2 loads million lb) compared to 11.4 million kg (25.1 million lb) from SPARROW. Along the Schuylkill, the from SPARROW [51]. The observed annual N load at Delaware River at Trenton is 14.2 million kg observed annual N load is 9.4 million kg (20.8 million lb) versus 13.1 million kg (28.9 million lb) from (31.2 millionthe SPARROW lb) compared model. The to 11.4 variance million may kgbe due (25.1 tomillion the difference lb) from between SPARROW. loads modeled Along during the Schuylkill, the the observedSPARROW annual base Nyear load of 2002 is 9.4 and million the period kg (20.8of record millioning at lb) each versus of the 13.1 stream million monitoring kg (28.9 stations. million lb) from theAlso, SPARROW the observed model. data would The varianceinclude nitrogen may be load dues delivered to the difference by groundwater between to the loads streams modeled and the during the SPARROWSPARROW base model year may of under-report 2002 and the groundwater period of recordingsources of nitrogen. at each ofGiven the streamthe variance monitoring of 25–30% stations. Also, the observed data would include nitrogen loads delivered by groundwater to the streams and the

Water 2018, 10, 95 11 of 21

SPARROW model may under-report groundwater sources of nitrogen. Given the variance of 25–30% (±) between the observed and modeled nitrogen loads in the Delaware River Basin, we expect the costs of pollutant load reduction programs in the watershed will vary by a similar amount. Water 2018, 10, x FOR PEER REVIEW 11 of 20 Our synthesis of the literature [5,46–49] reveals that nitrogen reduction costs vary from $2.64–$24.20/kg(±) between the observed N ($1.20–$11.0 and modeled nitrogen $/lb N) loads for in agricultural the Delaware River conservation, Basin, we expect $18.83–$174/kg the costs N ($8.56–$79.00of pollutant $/lb load N) reduction for wastewater programs treatment, in the watershed $165–$639/kg will vary by N a ($75.00–$132.00 similar amount. $/lb N) for airborne emissions controls,Our synthesis and of $198–$1100/kg the literature [5,46–49] N ($90.00–$500.00 reveals that nitrogen $/lb N)reduction for urban/suburban costs vary from $2.64– stormwater retrofit$24.20/kg BMPs (Table N ($1.20–$11.06) . Wastewater $/lb N) for treatment agricultural N conservation, load reduction $18.83–$174/kg costs vary N from ($8.56–$79.00 $18.83–$60.83/kg $/lb N ($8.56–$27.65N) for wastewater $/lb N)treatment, in the $165–$639/kg Chesapeake N Bay ($75.00– watershed$132.00 $/lb to N) $38.06/kg for airborne N emissions ($17.30 $/lbcontrols, N) in the and $198–$1100/kg N ($90.00–$500.00 $/lb N) for urban/suburban stormwater retrofit BMPs (Table Connecticut River Basin and $139.00–$174.00/kg N ($63.00–$79.00 $/lb N) in Maine and New 6). Wastewater treatment N load reduction costs vary from $18.83–$60.83/kg N ($8.56–$27.65 $/lb N) Hampshire.in the Chesapeake Airborne depositionBay watershed nitrogen to $38.06/kg load N reduction ($17.30 $/lb costs N) in inthe the Connecticut Chesapeake River BayBasin range and from $165/kg$139.00–$174.00/kg N ($75.00/lb N) N for($63.00–$79.00 Clean Air $/lb Act N) programs in Maine to and $639/kg New Hampshire. N ($132.00/lb Airborne N) for deposition low-emission vehiclenitrogen programs. load Urbanreduction stormwater costs in the retroﬁtting Chesapeake is Bay a more range expensive from $165/kg option N ($75.00/lb with costs N) thatfor Clean range from $198–$1100/kgAir Act programs N ($90–$500/lb to $639/kg N N) ($132.00/lb reduced. N) for low-emission vehicle programs. Urban stormwater Ourretrofitting survey is of a themore literature expensive indicates option with that co agriculturalsts that range conservationfrom $198–$1100/kg practices N ($90–$500/lb reduce N N) loads by reduced. 40% for grass buffers to 90% for cover crops at unit costs that range from $2.64/kg N ($1.20/lb N) Our survey of the literature indicates that agricultural conservation practices reduce N loads by for forest40% buffersfor grass to buffers $22.24/kg to 90% Nfor ($10.11/lb cover crops N)at unit for costs cover that crops. range Agricultural from $2.64/kg nutrientN ($1.20/lb management N) for plans reduceforest buffers N by to 20% $22.24/kg at a cost N ($10.11/lb of $4.41 N) $/lb for N.cover No-till crops.cropping Agricultural can nutrient reduce management N by 55% plans at a cost of $7.04/kgreduce N ($3.20/lbN by 20% at N) a reduced.cost of $4.41 Forest $/lb N. buffers No-till cropping remove can 50% reduce of N atN by $2.64 55% to at $14.94/kga cost of $7.04/kg N ($1.20 to $6.79 $/lbN ($3.20/lb N). Grass N) reduced. buffers Forest remove buffers 40% remove of N at 50% $3.45 of toN at $14.87/kg $2.64 to $14.94/kg N ($1.57 N to ($1.20 6.76 to $/lb $6.79 N). $/lb WeN). calculated Grass buffers the remove costs to 40% reduce of N at nitrogen $3.45 to $14.87/kg loads by N 20% ($1.57 (25th to 6.76 percentile), $/lb N). 32% (median), and 48% (75th percentile)We calculated to meet the a morecosts to stringent reduce nitrogen DRBC loads dissolved by 20% oxygen (25th percentile), standard 32% by multiplying(median), and N load 48% (75th percentile) to meet a more stringent DRBC dissolved oxygen standard by multiplying N reduction rates (kg/year) by the unit cost ($/kg) (in 2010 dollars) for atmospheric NOX reduction load reduction rates (kg/year) by the unit cost ($/kg) (in 2010 dollars) for atmospheric NOX reduction $165/kg$165/kg ($75.00/lb), ($75.00/lb), wastewater wastewater treatment treatment $61.60/kg $61.60/kg ($28.00/lb), ($28.00/lb), agriculture agriculture conservation conservation $11.00/kg $11.00/kg ($5.00/lb),($5.00/lb), and urban/suburbanand urban/suburban $440/kg $440/kg ($200/lb) ($200/lb) BMPs BMPs (Figure (Figure 9). 9).

Per-Pound Costs of Reducing Nitrogen Pollution in the Chesapeake Bay Region

200 $200.00

150

100 $92.40 $75.00 ($/lb N) 50 $27.65 $21.90 $7.00 $4.70 $3.30 $3.20 $3.20 $3.10 $1.50 $1.20 0

t s ) n e n a rs ofi tio an ture e tr c Pl nt w Sit du (Me acul Buff r Re e u Ne re q te X eme t. rad g r a wa m g a Cover Crops Grass Buffers g te Forest Buffers Forest rm e NO an s Land Retirement M M Sto rn P Up t Oy Restored Wetlands o n Conservation Tillage rb WT ie Ai W tr rmwater o Nu St

Figure 9. Costs (in 2010 dollars) to reduce nitrogen pollution in the Chesapeake Bay region [5,46]. Figure 9. Costs (in 2010 dollars) to reduce nitrogen pollution in the Chesapeake Bay region [5,46]. Table 6. Nitrogen reduction costs by source [5,46–49]. Table 6. Nitrogen reduction costs by source [5,46–49]. Atmospheric Wastewater Urban/Sub. Agriculture Location Deposition Treatment Stormwater Conservation Atmospheric Deposition Wastewater Treatment Urban/Sub. Stormwater Agriculture Conservation Location ($/lb N) ($/kg N) ($/lb N) ($/kg N) ($/lb N) ($/kg N) ($/lb N) ($/kg N) Chesapeake Bay ($/lb N) 75 ($/kg N) 165 ($/lb N) 27.65 ($/kg N) 60.83 ($/lb 200–500 N) 440–1100 ($/kg N) 1.20–4.70 ($/lb N) 2.64–10.34 ($/kg N) ChesapeakeNew Hampshire Bay 75 165 27.65 63–79 60.83 139–174 200–500 440–1100 1.20–4.70 2.64–10.34 New HampshireLong Island Sound 63–79 17.30 139–174 38.06 137 301 4.93 10.85 Long Island SoundIowa 17.30 38.06 13790 198 301 2.00–11.00 4.93 4.40–24.20 10.85 IowaChesapeake Bay 8.56 18.83 90 >100 >220 198 1.57–4.41 2.00–11.00 3.45–9.70 4.40–24.20 Chesapeake Bay 8.56 18.83 >100 >220 1.57–4.41 3.45–9.70 United States 75–132 165–639 United States 75–132 165–639 MarylandMaryland 104–210104–210 229–462 229–462 1.57–10.11 1.57–10.11 3.45–22.24 3.45–22.24

Water 2018, 10, 95 12 of 21

Under Option 1, we estimate that nitrogen loads must be reduced by 14.7 million kg/year (32.3 million lb/year) to achieve 32% reductions applied equally to all sources (Table7). Under this uniform load reduction scenario, wastewater N loads are reduced by 6.8 million kg/year (14.9 million lb/year) followed by agriculture by 4.3 million kg/year (9.4 million lb/year), urban/suburban by 2.0 million kg/year (4.5 million lb/year), and atmospheric by 1.8 million kg/year (3.9 million lb/year). The cost to reduce N loads evenly by 32% for each source is $1.66 billion/year with the largest costs borne by urban/suburban stormwater retroﬁtting ($905 million/year) with the highest unit cost, followed by wastewater discharge ($416 million/year), atmospheric NOX reduction ($291 million/year), and agriculture conservation ($47 million) with the lowest unit cost.

Table 7. Least Cost (in 2010 dollars) by state to reduce nitrogen by the median 32% in Delaware Basin.

Nitrogen Agriculture Atmospheric Wastewater Urban/Suburban Drainage Area Reduction Conservation State Deposition (5%) Discharge (10%) (5%) (km2) (32%) (90%) (M kg/Year) (M kg/Year) (M kg/Year) (M kg/Year) (M kg/Year) Pennsylvania 15,506 10.91 0.18 1.55 0.23 8.95 New Jersey 6374 2.73 0.05 0.50 0.05 2.14 New York 6358 0.45 0.05 0.01 0.01 0.36 Delaware 2352 0.55 0.00 0.05 0.01 0.45 Maryland 21 0.02 0.00 0.00 0.00 0.00 Del. Basin 30,611 14.68 0.27 2.09 0.32 12.00 Atmospheric Wastewater Agriculture Nitrogen Urban/Suburban Drainage Area Deposition Discharge Conservation State Reduction Cost ($440/kg N) (km2) ($165/kg N) ($62/kg N) ($11/kg N) ($M/Year) ($M/Year) ($M/Year) ($M/Year) ($M/Year) Pennsylvania 15,506 322 27 94 102 99 New Jersey 6374 87 8 31 25 23 New York 6358 19 8 0.6 6 4 Delaware 2352 16 1 3 6 5 Maryland 21 0.3 0.02 0 0.08 0.2 Del. Basin 30,611 449 45 130 142 132

According to Option 1, to reduce N equally by 32% from all sources, we estimate that annual loads must be reduced by 10.6 million kg in Pennsylvania for $1.2 billion, 3.0 million kg in New Jersey for $317 million, 600,000 kg in New York for $95 million, 500,000 kg in Delaware for $60 million, and 11,000 kg in Maryland for $700,000. Our review of the TMDL models suggest that nitrogen loads should be reduced by a median 32% within a range of 20% (25th percentile) to 48% (75th percentile) to increase dissolved oxygen levels from the current DRBC criterion (3.5 mg/L) to meet a future standard (5.0 mg/L) in the Delaware River. By maximizing least cost agricultural and wastewater BMP options and minimizing higher cost airborne emissions and urban stormwater BMPs (Figure 10), annual costs to reduce N loads by 32% in the Delaware Basin are cut from $1.66 billion for Option 1 (reduce loads evenly for all sources) to $845 million for Option 2 (reduce Ag N by 32%), $652 million for Option 3 (reduce Ag N by 60%), $552 million for Option 4 (reduce Ag N by 75%), and $449 million for Option 5 (reduce Ag N by 90%). We found that the least cost (Option 5) would reduce N loads by the median 32% or 14.5 million kg/year (32 million lb/year) by reducing atmospheric NOX by 5%, wastewater N by 10%, urban/suburban N by 5%, and agricultural N by 90%. Annual costs range from $334, $449, and $904 million to reduce N loads by 20% (25th percentile), 32% (median), and 48% (75th percentile), respectively. Water 2018, 10, 95 13 of 21 Water 2018, 10, x FOR PEER REVIEW 13 of 20

Water 2018, 10, x FOR PEER REVIEW 13 of 20 Water 2018, 10, x FOR PEER REVIEW 13 of 20

Figure 10. Costs (in 2010 dollars) to reduce nitrogen loads by 32% in the Delaware Basin. Figure 10. Costs (in 2010 dollars) to reduce nitrogen loads by 32% in the Delaware Basin.

AccordingFigure to 10. Option Costs (in 5, 2010we concludedollars) to thatreduce the nitrogen annual loads least bycost 32% to in reduce the Delaware N loads Basin. by 32% in the Figure 10. Costs (in 2010 dollars) to reduce nitrogen loads by 32% in the Delaware Basin. DelawareAccording Basin to is Option $449 million, 5, we concludeincluding that$141 the million annual for leasturban/suburban cost to reduce retrofitting, N loads $132 by 32% million in the According to Option 5, we conclude that the annual least cost to reduce N loads by 32% in the Delawarefor agriculture Basinis conservation, $449 million, $130 including million $141for wastewater million for treatment, urban/suburban and $45 million retroﬁtting, for atmospheric $132 million DelawareAccording Basin tois $449Option million, 5, we includingconclude that$141 the million annu foral least urban/suburban cost to reduce retrofitting, N loads by $132 32% million in the forNOX agriculture reduction conservation, (Figure 11). $130Covering million half for of wastewaterthe Basin, Pe treatment,nnsylvania’s and annual $45 million share is for $322 atmospheric million forDelaware agriculture Basin conservation, is $449 million, $130 including million for$141 wastewater million for treatment, urban/suburban and $45 retrofitting, million for atmospheric$132 million NOXor 72% reduction of the N (Figure load reduction 11). Covering cost (Figure half of 12). the Ne Basin,w Jersey Pennsylvania’s would bear $87 annual million share or is19% $322 of the million cost. or NOXfor agriculture reduction conservation, (Figure 11). Covering $130 million half for of wastewaterthe Basin, Pe treatment,nnsylvania’s and annual $45 million share for is atmospheric$322 million 72%New of theYork N State load would reduction contribute cost (Figure $19 million 12). Newor 4% Jerseyof the N would reduction bear cost. $87 millionDelaware or would 19% of assume the cost. orNOX 72% reduction of the N load(Figure reduction 11). Covering cost (Figure half of12). the Ne Basin,w Jersey Pe nnsylvania’swould bear $87 annual million share or 19%is $322 of the million cost. $16 million or 4% of the cost. Maryland’s share would be $337,000. NewNewor 72% York York of State the State N would loadwould reduction contribute contribute cost $19 $19 (Figure million million 12). oror Ne 4%w of Jersey the N Nwould reduction reduction bear cost.$87 cost. million Delaware Delaware or 19% would would of the assume assumecost. $16$16New million million York or State or 4% 4% would of of the the cost.contribute cost. Maryland’s Maryland’s $19 million shareshare or would 4% of thebe $337,000. $337,000.N reduction cost. Delaware would assume $16 million or 4% of the cost. Maryland’s share would be $337,000.

Figure 11. Least cost (in 2010 dollars) by source to reduce nitrogen loads by 32% in Delaware Basin.

Figure 11. Least cost (in 2010 dollars) by source to reduce nitrogen loads by 32% in Delaware Basin. Figure 11. Least cost (in 2010 dollars) by source to reduce nitrogen loads by 32% in Delaware Basin. Figure 11. Least cost (in 2010 dollars) by source to reduce nitrogen loads by 32% in Delaware Basin.

Figure 12. Least cost (in 2010 dollars) by state to reduce nitrogen loads by 32% in Delaware Basin.

OnFigure a watershed 12. Least cost basis, (in 2010we estimate dollars) by that state the to Delawarereduce nitrogen River loads at Trenton by 32% incontributes Delaware Basin. 25% of the Figure 12. Least cost (in 2010 dollars) by state to reduce nitrogen loads by 32% in Delaware Basin. nitrogenFigure load, 12. Leastpredominately cost (in 2010 from dollars) agricultural by state source to reduces, for nitrogen $132 million loads or by 30% 32% of in the Delaware total cost Basin. (Figure On a watershed basis, we estimate that the Delaware River at Trenton contributes 25% of the nitrogen On aload, watershed predominately basis, we from estimate agricultural that the source Delawares, for $132 River million at Trenton or 30% contributesof the total cost25% (Figure of the nitrogen load, predominately from agricultural sources, for $132 million or 30% of the total cost (Figure

Water 2018, 10, 95 14 of 21

On a watershed basis, we estimate that the Delaware River at Trenton contributes 25% of the Water 2018, 10, x FOR PEER REVIEW 14 of 20 nitrogen load, predominately from agricultural sources, for $132 million or 30% of the total cost (Figure13 and Table 13 and 8). TableThe Schuylkill8). The Schuylkillcontributes contributes 30% of the N 30% load, of mostly the N from load, wastewater mostly from and wastewater agricultural andsources, agricultural at a costsources, of $124 atmillion a cost or of 28% $124 of million the total or cost. 28% ofThe the tributaries total cost. between The tributaries Philadelphia between and PhiladelphiaTrenton contribute and Trenton 29% of the contribute N load, 29%mostly of from the N wastewater, load, mostly with from a cost wastewater, of $104 million with or a cost24% ofof $104the total. million The orBrandywine-Christina 24% of the total. The watershed, Brandywine-Christina where over three-quarters watershed, whereof the overnitrogen three-quarters flows from ofagriculture, the nitrogen would ﬂows cost from $37 agriculture, million (8% would of the cost N load $37 millionreduction (8% cost). of the The N loadwatersheds reduction between cost). TheWilmington watersheds and between Philadelphia Wilmington would andrequire Philadelphia $32 million would or 7% require of the $32 total million cost orto 7%reduce of the mostly total costwastewater to reduce N mostlyloads. The wastewater Delaware N Bay loads. watershed The Delaware between Bay Prime watershed Hook and between Wilmington Prime would Hook andcost Wilmington$13 million to would reduce cost the $13 mostly million agricultural to reduce N the loads mostly from agricultural the Coastal N Plain loads streams from the on Coastaleither side Plain of streamsthe bay. on either side of the bay.

Figure 13. Least cost by watershed to reduce nitrogen loads by 32% in the Delaware Basin. Figure 13. Least cost by watershed to reduce nitrogen loads by 32% in the Delaware Basin.

Table 8. Least cost by watershed to reduce nitrogen by the median 32% in the Delaware Basin. Table 8. Least cost by watershed to reduce nitrogen by the median 32% in the Delaware Basin. Nitrogen Atmospheric Wastewater Agriculture Urban/Suburban Drainage ReductionNitrogen Deposition Discharge ConservationAgriculture Watershed Atmospheric Wastewater Urban/Suburban(5%) Drainage 2 Reduction Conservation Watershed Area (mi ) (32%) Deposition(5%) (5%) Discharge(10%) (10%) (5%) (90%) Area (mi2) (32%) (M kg/Year) (90%) (M kg/Year) (M(M kg/Year) kg/Year) (M(M kg/Year) kg/Year) (M kg/Year) (M kg/Year) (M kg/Year) (M kg/Year) Del. R. at Trenton 17,731 3.91 0.18 0.15 0.14 3.45 Del. R. at Trenton 17,731 3.91 0.18 0.15 0.14 3.45 Del.Del. R. R. above Phila.Phila. 3227 3227 1.82 0.02 0.02 1.09 1.09 0.05 0.05 0.64 0.64 SchuylkillSchuylkill R. R. 4905 4905 4.91 0.05 0.05 0.59 0.59 0.09 0.09 4.18 4.18 Brandywine-ChristinaBrandywine-Christina 14531453 2.36 0.01 0.01 0.01 0.01 0.02 0.02 2.32 2.32 Del. R. above Wilmington 1264 0.68 0.01 0.23 0.03 0.41 Del.Del. R. Bayabove at Prime Wilmington Hook 1896 1264 0.820.68 0.01 0.01 0.23 0.01 0.03 0.00 0.82 0.41 Del. DelawareBay at Prime Basin Hook 30,477 1896 14.55 0.82 0.01 0.27 0.01 2.09 0.00 0.32 11.86 0.82 Delaware Basin 30,477 Nitrogen14.55 Atmospheric0.27 Wastewater2.09 Urban/Sub.0.32 Agriculture11.86 Drainage Watershed ReductionNitrogen Atmospheric($165/kg N) Wastewater($62/kg N) Urban/Sub.($440/kg N) Agriculture($11/kg N) ReaDrainage (mi2) Watershed Reduction($M/Year) ($165/kg($M/Year) N) ($62/kg($M/Year) N) ($440/kg($M/Year) N) ($M/Year)($11/kg N) Rea (mi2) Del. R. at Trenton 17,731($M/Year) 132 ($M/Year) 28 ($M/Year) 9 ($M/Year) 57 ($M/Year) 38 Del.Del. R.R. above. at Trenton Phila. 322717,731 104132 28 3 9 67 57 27 7 38 Schuylkill R. 4905 124 8 37 33 46 Brandywine-ChristinaDel. R. above. Phila. 1453 3227 104 37 3 2 67 1 27 8 26 7 Del. R.Schuylkill above Wilmington R. 1264 4905 124 32 8 1 37 15 33 11 5 46 Brandywine-ChristinaDel. Bay at Prime Hook 18961453 1337 2 1 1 1 8 2 9 26 Delaware Basin 30,477 442 44 130 138 130 Del. R. above Wilmington 1264 32 1 15 11 5 Del. Bay at Prime Hook 1896 13 1 1 2 9 WeDelaware constructed Basin marginal30,477 abatement442 cost curves44 (Figures130 14 and 15) that 138 illustrate the 130 least costs to reduce nitrogen loads by the median 32% within a range of 20% (25th percentile) to 48% (75thWe percentile). constructed The marginal nitrogen abatement MAC curve cost indicates curves (Figures that, by 14 prioritizing and 15) that investments illustrate the in least agriculture costs to conservationreduce nitrogen and someloads wastewaterby the median treatment 32% projectswithin ﬁrst,a range 90% of of the20% nitrogen (25th percentile) (13.6 million to kg48% N/year) (75th (30percentile). million lbThe N/year) nitrogen can MAC be removed curve indicates for just 35%that, ($160 by prioritizing million) of theinvestments $449 million in agriculture total cost. Theconservation remaining and 0.9 some million wastewater kg N/year treatment (2 million projects lb N/year) first, 90% or 10% of th ofe thenitrogen N load (13.6 reduction million will kg N/year) require 65%(30 million ($290 million/year) lb N/year) can of thebe removed total cost tofor implement just 35% ($160 the remaining million) of wastewater the $449 million treatment total followed cost. The by moreremaining expensive 0.9 million atmospheric kg N/year deposition, (2 million and lb N/year) urban/suburban or 10% of projects.the N load The reduction marginal will abatement require cost65% ($290 million/year) of the total cost to implement the remaining wastewater treatment followed by more expensive atmospheric deposition, and urban/suburban projects. The marginal abatement cost (MAC) curves define the most cost-effective combination of nitrogen reduction strategies to improve dissolved oxygen to a future DRBC standard to provide year-round propagation of anadromous fish. The least

Water 2018, 10, 95 15 of 21

(MAC) curves define the most cost-effective combination of nitrogen reduction strategies to improve Water 2018, 10, x FOR PEER REVIEW 15 of 20 dissolved oxygen to a future DRBC standard to provide year-round propagation of anadromous fish. Water 2018, 10, x FOR PEER REVIEW 15 of 20 costThe leastagriculture cost agriculture and wastewater and wastewater treatment treatment reduct reductionsions would would be beimplemented implemented first first in in priority priority watersheds,cost agriculture followed followed and by wastewater by higher higher cost cost treatmentatmospheric atmospheric reduct dep depositionositionions would and andurban be urban suburbanimplemented suburban runoff first runoff controls. in controls.priority After theAfterwatersheds, less the costly less followed agricultural costly agricultural by higher and wastewater cost and atmospheric wastewater BMPs ardep BMPse implemented,osition are and implemented, urban nitrogen suburban reduction nitrogen runoff reductionin controls. the Delaware in After the BasinDelawarethe less becomes costly Basin agricultural incrementally becomes incrementally and less wastewater cost-effective less BMPs cost-effective afte arer implemented, 30% after N reduction 30% Nnitrogen reduction as the reduction slope as the of slopein the the cost ofDelaware the curve cost flattens.curveBasin becomes flattens. Increasingly Increasinglyincrementally higher invest higher less mentscost-effective investments in costly afte in atmospheric costlyr 30% atmosphericN reduction and urban/suburban andas the urban/suburban slope controls of the costprovide controls curve a diminishingprovideflattens. aIncreasingly diminishing rate of return higher rate inof investterms returnments of inpo termsllutant in costly of removal pollutant atmospheric efficie removalncy and per ur efficiencyban/suburban dollar spent. per dollar controls spent. provide a diminishing rate of return in terms of pollutant removal efficiency per dollar spent.

Figure 14. Nitrogen marginal abatement cost curves in the Delaware Basin. Figure 14. Nitrogen marginal abatement cost curves in the Delaware Basin. Figure 14. Nitrogen marginal abatement cost curves in the Delaware Basin.

Figure 15. Nitrogen reduction cost curve for the Delaware Basin (in 2010 dollars). Figure 15. Nitrogen reduction cost curve for the Delaware Basin (in 2010 dollars). Figure 15. Nitrogen reduction cost curve for the Delaware Basin (in 2010 dollars). We compared the 1966 Delaware River economic valuation study to our 2017 modern study. AdjustingWe compared to 2010 dollars the 1966 and Delaware starting fromRiver a economicbase dissolved valuation oxygen study level to ofour 3 mg/L,2017 modern our review study. of We compared the 1966 Delaware River economic valuation study to our 2017 modern study. theAdjusting 1966 Delaware to 2010 dollarsEstuary and economic starting study from indicate a base dissolveds that the annualoxygen costslevel toof improve3 mg/L, ourwater review quality of Adjusting to 2010 dollars and starting from a base dissolved oxygen level of 3 mg/L, our review of rangedthe 1966 from Delaware $58–$87 Estuary million economic to achieve study summer indicate disss olvedthat the oxygen annual of costs4.0 mg/L to improve to $180–$209 water millionquality the 1966 Delaware Estuary economic study indicates that the annual costs to improve water quality toranged reach from 4.5 mg/L $58–$87 (Table million 9). These to achieve estimates summer from dissthe olved50-year oxygen old economic of 4.0 mg/L study to compare$180–$209 with million our ranged from $58–$87 million to achieve summer dissolved oxygen of 4.0 mg/L to $180–$209 million 21stto reach century 4.5 mg/Lleast cost(Table (Option 9). These 5), whichestimates indicates from thethat 50-year annual old costs economic of $50 million study comparewould need with to our be to reach 4.5 mg/L (Table9). These estimates from the 50-year old economic study compare with our invested21st century to reduce least cost pollutant (Option loads 5), which to improve indicates dissolved that annual oxygen costs levels of $50 to million 4.0 mg/L, would $150 need million to be 21st century least cost (Option 5), which indicates that annual costs of $50 million would need to be wouldinvested be toneeded reduce to pollutantreach a dissolved loads to oxygenimprove leve dissolvedl of 4.5 mg/L, oxygen and levels $449 tomillion 4.0 mg/L, would $150 be neededmillion invested to reduce pollutant loads to improve dissolved oxygen levels to 4.0 mg/L, $150 million would towould reach be a dissolvedneeded to oxygenreach a leveldissolved of 5.0 oxygenmg/L in leve the lDelaware of 4.5 mg/L, River. and $449 million would be needed be needed to reach a dissolved oxygen level of 4.5 mg/L, and $449 million would be needed to reach a to reachWhile a dissolved the costs ofoxygen the 1966 level study of 5.0 and mg/L our in 2017 the economic Delaware study River. compare favorably, the methods dissolved oxygen level of 5.0 mg/L in the Delaware River. and approachesWhile the costs differ. of the The 1966 1966 study economic and our study 2017 economicexplored studyonly comparethe costs favorably, of point thesource methods load While the costs of the 1966 study and our 2017 economic study compare favorably, the methods reductionsand approaches (wastewater differ. treatmentThe 1966 planteconomic upgrades) study of explored nitrogen onlywhereas the ourcosts 2017 of analysispoint source estimated load and approaches differ. The 1966 economic study explored only the costs of point source load reductions thereductions costs of (wastewater point sources treatment such as plant wastewater upgrades) treatment of nitrogen plants whereas and nonpointour 2017 analysissources estimated(airborne (wastewater treatment plant upgrades) of nitrogen whereas our 2017 analysis estimated the costs deposition,the costs of stormwater, point sources and such agriculture) as wastewater sources oftreatment nitrogen. plants Also, theand 1966 nonpoint study estimatedsources (airborne costs of of point sources such as wastewater treatment plants and nonpoint sources (airborne deposition, nitrogendeposition, load stormwater, reduction startingand agriculture) from a baseline sources of of very nitrogen. poor waterAlso, thequality 1966 (dissolved study estimated oxygen costs levels of atnitrogen or near load zero) reduction and the 2017 starting study from estimates a baseline costs of of very load poor reduction water starting quality at(dissolved a baseline oxygen of moderate levels waterat or near quality zero) (dissolved and the 2017 oxygen study level estimates at 3.5 mg/L) costs of in load the Delaware reduction River. starting at a baseline of moderate water quality (dissolved oxygen level at 3.5 mg/L) in the Delaware River.

Water 2018, 10, x FOR PEER REVIEW 16 of 20 Water 2018, 10, 95 16 of 21

Table 9. Comparison of costs to meet water quality criteria along the Delaware River. stormwater, and agriculture) sources of nitrogen. Also, the 1966 study estimated costs of nitrogen load Dissolved Oxygen Costs (in 2010 Dollars) ($M/Year) reduction startingObjective from a baseline of very poor water quality (dissolved oxygen levels at or near zero) (mg/L) FWPCA1 [32] (1966) Modern Study (2017) and the 2017 study estimates costs of load reduction starting at a baseline of moderate water quality 5.0 449 (dissolved oxygen level at 3.5 mg/L) in the Delaware River. I 4.5 180–209 150 TableII 9. Comparison of4.0 costs to meet water quality 58–87 criteria along the Delaware 50 River. III 3.0 0 0 Note: 1. Adjusted from 1964 dollars to 2010 dollars by 3% annuallyCosts based (in 2010 on change Dollars) in ($M/Year) Consumer Price Index. Objective Dissolved Oxygen (mg/L) We point out an important consideration concerningFWPCA the1 [32 inverse] (1966) relationshipModern between Study (2017) dissolved oxygen saturation and water temperature5.0 in the Delaware River (Figure 16). The costs 449 of achieving improved Iwater quality in the Delaware 4.5 River to meet more 180–209 stringent dissolved oxygen 150 criteria are based on conditionsII where water 4.0 temperatures peak near 58–87 30 °C (86 °F), usually in July 50 and August. III 3.0 0 0 At 30 °C, freshwater dissolved oxygen 100% saturation is 7.54 mg/L and dissolved oxygen is 46% Note: 1. Adjusted from 1964 dollars to 2010 dollars by 3% annually based on change in Consumer Price Index. saturated at 3.5 mg/L, 53% saturated at 4.0 mg/L, 66% saturated at 5.0 mg/L, and 80% saturated at 6.0 mg/L. Should water temperatures in the tidal Delaware River increase by 2 °C to peak summer levels of 30We °C, point based out on an saturation, important dissolved consideration oxygen concerning levels will the decline inverse by relationship about 0.2 betweenmg/L without dissolved any oxygendecrease saturation in nutrient and loading. water temperatureMore research in theis needed Delaware utilizing River a (Figure new hydrodynamic 16). The costs ofmodel achieving to be improveddeveloped waterby the quality DRBC in that the Delawarewould explore River the to meet influence more stringentof water dissolvedtemperature oxygen and criteriasalinity areon based on conditions where water temperatures peak near 30 ◦C (86 ◦F), usually in July and August. dissolved◦ oxygen in the Delaware Estuary. At 30BasedC, freshwater on the delivery dissolved fraction oxygen of nitrogen 100% saturation (i.e., fraction is 7.54 of nitrogen mg/L and load dissolved delivered oxygen to the outlet), is 46% saturatedimplementation at 3.5 mg/L,of BMPs 53% in saturated watersheds at 4.0 closest mg/L, to 66%the saturatedDelaware atEstuary 5.0 mg/L, would and provide 80% saturated the most at ◦ 6.0immediate mg/L. Shouldimprovements water temperatures in water quality. in the The tidal SPARROW Delaware model River increaseindicates by that 2 theC to delivered peak summer yield ◦ levelsof nitrogen of 30 fromC, based watershe on saturation,ds far from dissolved the estuary, oxygen such levels as head willwaters decline of bythe aboutDelaware 0.2 mg/L River withoutin New anyYork decrease State and in upper nutrient Lehigh loading. and Schuylkill More research basins, is are needed less likely utilizing to influence a new hydrodynamic dissolved oxygen model levels to bein developedthe Delaware by theEstuary. DRBC Nitrogen that would reduction explore theprac inﬂuencetices should of water be temperaturecost-effectively and invested salinity onin dissolvedwatersheds oxygen that deliver in the the Delaware highest Estuary. yield of nitrogen and are close to the estuary.

Figure 16. Dissolved oxygen/water temperature along the Delaware River at Ben Franklin Bridge. Figure 16. Dissolved oxygen/water temperature along the Delaware River at Ben Franklin Bridge.

Groundwater can contribute significant nitrogen and phosphorus loads to surface waters. For instance,Based half on of the the delivery nonpoint fraction source of N nitrogen load to the (i.e., Chesapeake fraction of nitrogenBay flows load through delivered groundwater to the outlet), and implementationthe other half flows of BMPsto the inbay watersheds via surface closest runoff to[52]. the Depending Delaware on Estuary soil permeability, would provide it could the most take immediateyears for nutrients improvements such as inN waterand P quality.in groundwater The SPARROW to reach model the Delaware indicates Estuary that the from delivered source yield waters of nitrogen[53]. The from implementation watersheds far of from nitrogen the estuary, source such controls as headwaters for airborne of the Delawareemissions River and inwastewater New York Statetreatment and upperwould Lehigh have andan immediate Schuylkill basins,effect in are improving less likely towater inﬂuence quality dissolved in the Delaware oxygen levels River, in whereas urban, suburban, and agriculture BMPS could take months to years to make an impact on

Water 2018, 10, 95 17 of 21 the Delaware Estuary. Nitrogen reduction practices should be cost-effectively invested in watersheds that deliver the highest yield of nitrogen and are close to the estuary. Groundwater can contribute significant nitrogen and phosphorus loads to surface waters. For instance, half of the nonpoint source N load to the Chesapeake Bay flows through groundwater and the other half flows to the bay via surface runoff [52]. Depending on soil permeability, it could take years for nutrients such as N and P in groundwater to reach the Delaware Estuary from source waters [53]. The implementation of nitrogen source controls for airborne emissions and wastewater treatment would have an immediate effect in improving water quality in the Delaware River, whereas urban, suburban, and agriculture BMPS could take months to years to make an impact on water quality, depending on soils and the physiographic province. The SPARROW model does not account for direct contributions of nitrogen from groundwater; therefore, it is likely that nitrogen loads to the tidal Delaware River are underestimated in this analysis. Additional modeling—particularly geographically resolved hydrodynamic modeling, with explicit inclusion of groundwater transport—is needed to address this quantitatively.

6. Conclusions The Delaware River and its tributaries have made a notable recovery in the half-century since JFK signed the DRBC Compact in 1961, Richard Nixon formed the EPA in 1970, and Congress passed the Federal Clean Water Act Amendments during the 1970s. A first-of-its-kind 1966 benefit–cost analysis conducted by the Federal Water Pollution Control Administration (FWPCA) concluded that it would be cost-effective for the DRBC to fund a multi-million-dollar per year waste load abatement program to raise dissolved oxygen levels to boatable and fishable standards that would, in turn, generate economic activity. In 1967, the DRBC used this benefit–cost analysis to set dissolved oxygen criterion at 3.5 mg/L along the river from Philadelphia to Wilmington, where this water quality standard has stood for over four decades. The FWPCA and DRBC were indeed prescient, as multi-billion-dollar investments in Delaware River water pollution control programs have boosted the water quality, as measured by dissolved oxygen—from anoxia during the 1960s, to levels that meet the DRBC criterion of 3.5 mg/L most of the year (except during the increasingly hot summers). While the water quality in the Delaware River has recovered over the last half-century, the Partnership for the Delaware Estuary Science and Technical Advisory Committee (PDE STAC) has called for raising the dissolved oxygen standard from 3.5 mg/L, which has stood since the 1960s, to a higher level of protection. A more rigorous standard of at least 5 mg/L or 6 mg/L would provide for more year-round protection of anadromous fish, such as the recovering American shad and the nearly extirpated Atlantic sturgeon (just placed on the Federal Endangered Species List). A more rigorous dissolved oxygen standard would also provide protection against atmospheric warming that is projected to increase water temperatures, raise sea levels, and elevate chloride levels, all of which act in combination to reduce dissolved oxygen saturation. By maximizing least cost agricultural and wastewater BMPs and minimizing higher cost airborne emissions and urban stormwater retrofitting BMPs, we estimate that the annual costs to reduce nitrogen loads by 32% in the Delaware Basin are reduced by more than 300%, from $1.66 billion for Option 1 (reduce N equally from all sources by 32%) to $449 million for Option 5 (reduce Agriculture N by 90%). Our interpretation of the nitrogen marginal abatement cost (MAC) curves shows it to be more cost-effective to prioritize upstream investments in agricultural conservation and wastewater treatment, as these controls have lower unit nitrogen reduction costs that are up to an order of magnitude less than the more expensive airborne emission source control and urban/suburban best management practices. This research concludes, through a marginal abatement cost (MAC) analysis, that it would be cost-effective to raise dissolved oxygen levels to meet a more stringent standard by prioritizing agricultural conservation and some wastewater treatment investments in the Delaware River watershed to remove 90% of the nitrogen load by 13.6 million kg N/year (30 million lb N/year) for just 35% Water 2018, 10, 95 18 of 21

($160 million) of the $449 million total cost. We estimate that the annual costs to reduce nitrogen loads and increase dissolved oxygen to meet a more stringent standard in the Delaware River include $45 million for atmospheric NOX reduction, $130 million for wastewater treatment, $132 million for agriculture conservation, and $141 million for urban stormwater retroﬁtting. In 2010 dollars, the annual costs from the 1966 Delaware Estuary economic study range from $58–$87 million to achieve summer dissolved oxygen of 4.0 mg/L to $180–$209 million to reach dissolved oxygen of 4.5 mg/L. Estimates from this 50-year-old economic study compare with our 21st century least cost analysis (Option 5), which estimates that an annual investment of $50 million is needed to reduce nutrient loads and increase dissolved oxygen levels to 4.0 mg/L, $150 million is needed for dissolved oxygen to reach 4.5 mg/L, and $449 million is needed for dissolved oxygen to reach 5.0 mg/L in the Delaware River. This economic cost analysis of the pollutant load reductions needed to improve water quality in the Delaware River is inﬂuenced by (1) the inverse relationship between water temperature and dissolved oxygen; (2) the delivery fraction of nitrogen; and (3) the groundwater time of travel. Since dissolved oxygen (DO) saturation and water temperature is inversely related, rising water temperatures in the Delaware River to near 30 ◦C (86 ◦F) during summer may depress dissolved oxygen levels without any decrease in nutrient loading, thus offsetting future pollution load reduction efforts. Based on the delivery fraction of nitrogen (i.e., fraction of nitrogen load delivered to the outlet), nitrogen reduction practices should be cost-effectively invested in watersheds close to the estuary that deliver the highest yield of nitrogen. Depending on soil permeability, long groundwater time of travel may occur; therefore, the effects of urban/suburban stormwater and agriculture nitrogen load reduction BMPs in source water many miles from the Delaware River may take months to years to have a positive impact on water quality.

Acknowledgments: We are grateful to the Delaware River Basin Commision and Partnership for the Delaware Estuary in the support of this economic research. Conﬂicts of Interest: The authors declare there are no conﬂicts of interest.

References

1. Gilbert, M.; Madden, C.J.; Boynton, W.; Flemer, D.; Heil, C.; Sharp, D. Nutrients in Estuaries: Summary Report of National Estuarine Experts Workgroup 2005–2007; EPA Contract No. 68-C-02-91 and EP-6-07-025; United States Environmental Protection Agency: Washington, DC, USA, 2010; p. 188. 2. Stoner, N.K. Memorandum to regional administrators Regions 1–10. In Working in Partnership with States to Address Phosphorus and Nitrogen Pollution through Use of a Framework for State Nutrient Reductions; United States Environmental Protection Agency: Washington, DC, USA, 2011. 3. Schleich, J.; White, D.; Stephenson, K. Cost implications in achieving alternative water quality targets. Water Resour. Res. 1996, 32, 2879–2884. [CrossRef] 4. Evans, B.M. An evaluation of potential nitrogen load reductions to Long Island Sound from the Connecticut River Basin. In A Report to the New England Interstate Water Pollution Control Commission, Lowell, MA; Penn State Institutes of Energy and the Environment: University Park, PA, USA, 2008; p. 66. 5. Chesapeake Bay Program. Chesapeake Bay Watershed Bmp Potential Load Reductions and Cost-Effectiveness Study; Chesapeake Bay Program: Annapolis, MD, USA, 2004. 6. Rabotyagov, S.; Campbell, T.; Jha, M.; Gassman, P.W.; Arnold, J.; Kurkalova, L.; Secchi, S.; Feng, H.; Kling, C.L. Least-cost control of agricultural nutrient contributions to the Gulf of Mexico hypoxic zone. Ecol. Appl. 2010, 20, 1542–1555. [CrossRef][PubMed] 7. Lyon, R.; Farrow, S. An economic analysis of Clean Water Act issues. Water Resour. Res. 1995, 31, 213–223. [CrossRef] 8. Interstate Commission on the Delaware River Basin. The Delaware River Basin Physical Facts; Interstate Commission on the Delaware River Basin: Philadelphia, PA, USA, 1940. 9. Sharp, J.H.; Church, T.M. Biochemical modeling in coastal waters of the Middle Atlantic States. Limnol. Oceanogr. 1981, 26, 843–854. [CrossRef] Water 2018, 10, 95 19 of 21

10. Albert, R.C. The historical context of water quality management for the Delaware Estuary. Estuaries 1988, 11, 99–107. [CrossRef] 11. Mandarano, L.A.; Featherstone, J.P.; Paulsen, K. Institutions for interstate water resources management. J. Am. Water Resour. Assoc. 2008, 44, 136–147. [CrossRef] 12. Sharp, J.H.; Culberson, C.H.; Church, T.M. The chemistry of the Delaware Estuary. General considerations. Limnol. Oceanogr. 1982, 27, 1019–1028. [CrossRef] 13. Sharp, J.H.; Pennock, J.R.; Church, T.M.; Tramontano, J.M.; Cifuetes, L.A. The Estuarine Interaction of Nutrients, Organics, and Metals: A Case Study in the Delaware Estuary; The Estuary as a Filter; Academic Press Inc.: Cambridge, MA, USA, 1984. 14. Scudlark, J.R.; Church, T.M. Atmospheric input of inorganic nitrogen to Delaware Bay. Estuaries Coasts 1993, 16, 747–759. [CrossRef] 15. Sharp, J.H. How the Delaware Estuary works. In Proceedings of the Prepared for Processes Workgroup Meeting, Tiburon, CA, USA, 20 June 2006. 16. Church, T.M.; Sommerfield, C.K.; Velinsky, D.J.; Point, D.; Benoit, C.; Amouroux, D.; Plaa, D.; Donard, F.X. Marsh sediments as records of sedimentation, eutrophication and metal pollution in the urban Delaware Estuary. Mar. Chem. 2006, 102, 72–95. [CrossRef] 17. Bricker, S.; Longstaff, B.; Dennison, W.; Jones, W.; Boicourt, K.; Wicks, C.; Woerner, J. Effects of nutrient enrichment in the Nation’s estuaries: A decade of change. In NOAA Coastal Ocean Program Decision Analysis Series No. 26; National Center for Coastal Ocean Science: Silver Spring, MD, USA, 2007; p. 328. 18. Bain, M.; Walter, M.T.; Steenhuis, T.; Brutsaert, W.; Gaetano, A. Delaware River and Catskill Region hydrologic observatory. In Prospectus by the Cornell University Hydrologic Sciences Working Group 2010; Cornell University: Ithaca, NY, USA; p. 10. 19. Sharp, J.H.; Yoshiyama, K.; Parker, A.E.; Schwartz, M.C.; Curless, S.E.; Beauregard, A.Y.; Ossolinski, J.E.; Davis, A.R. A biogeochemical view of estuarine eutrophication: Seasonal and spatial trends and correlations in the Delaware Estuary. Estuaries Coasts 2009, 32, 1023–1043. [CrossRef] 20. Sharp, J.H. Estuarine oxygen dynamics: What can we learn about hypoxia from long-time records in the Delaware Estuary? Limonol. Oceanogr. 2010, 55, 535–548. [CrossRef] 21. Schneider, J. Development of numeric nutrient criteria for waters of the State of Delaware and Delaware Bay/Estuary nutrient dissolved oxygen concerns. In Proceedings of the DRBC Joint Monitoring and Water Quality Advisory Committee Meeting, West Trenton, NJ, USA, 20 February 2007. 22. Sildorff, E.; Fikslin, T.J. Continuing restoration of dissolved oxygen in the Delaware Estuary: Historical data and current efforts. In Proceedings of the 2010 AWRA Conference, Philadelphia, PA, USA, 1–4 November 2010. 23. Kauffman, G.J.; Homsey, A.R.; Belden, A.C.; Sanchez, J.R. Water quality trends in the Delaware River Basin (USA) from 1980 to 2005. Environ. Monit. Assess. 2010, 177, 193–225. [CrossRef][PubMed] 24. Sharp, J.H.; Cifuentes, L.A.; Coffin, R.B.; Pennock, J.R. The influence of river variability on the circulation, chemistry, and microbiology of the Delaware Estuary. Estuaries 1986, 9, 261–269. [CrossRef] 25. Roman, C.T.; Jaworski, N.; Short, F.T.; Findlay, S.; Warren, R.S. Estuaries of the northeastern United States: Habitat and land use signatures. Estuaries 2000, 23, 743–764. [CrossRef] 26. Delaware River Basin Commission. Administrative Manual Part III Water Quality Regulations with Amendments through July 16, 2008; Delaware River Basin Commission: Ewing Township, NJ, USA, 2008; p. 131. 27. Delaware River Basin Commission. Delaware River and Bay Integrated List Water Quality Assessment; Delaware River Basin Commission: Ewing Township, NJ, USA, 2010; p. 46. 28. Dorfman, R.; Jacoby, H.D.; Thomas, H.A. Models for Managing Regional Water Quality; Harvard University Press: Cambridge, MA, USA, 1972. 29. Maass, A.; Huffschmidt, M.; Dorfman, R.; Thomas, H.; Marglin, S.; Fair, G. Design of Water Resources Systems; Harvard University Press: Cambridge, MA, USA, 1962. 30. Reuss, M. Is it time to resurrect the Harvard Water Program? J. Water Resour. Plan. Manag. Am. Soc. Civ. Eng. 2003, 129, 357–360. [CrossRef] 31. Schaumburg, G.W. Water pollution control in the Delaware Estuary. In Harvard Water Program Discussion Paper No. 67-2; Harvard University: Cambridge, MA, USA, 1967; p. 150. 32. Federal Water Pollution Control Administration. Delaware Estuary Comprehensive Study, Preliminary Report and Findings; Federal Water Pollution Control Administration: Washington, DC, USA, 1966; p. 110. Water 2018, 10, 95 20 of 21

33. Thomann, R.V. River Ecology and Man. In The Delaware River—A Study in Water Quality Management; Oglesby, R.T., Carlson, C.A., McCann, J.A., Eds.; Academic Press Inc.: New York, NY, USA, 1972; pp. 99–132. 34. Kneese, A.V.; Bower, B.T. Managing Water Quality: Economics, Technology, Institutions; Resources for the Future: Washington, DC, USA, 1984; p. 328. 35. Johnson, E.L. A study in the economics of water quality management. Water Resour. Res. 1967, 3, 291–305. [CrossRef] 36. Ad-Hoc Task Force to Evaluate Dissolved Oxygen Requirements of Indigenous Estuary Fish. Dissolved Oxygen Requirements of a “Fishable” Delaware River Estuary; Report to the Delaware River Basin Commission; Delaware River Basin Commission: Trenton, NJ, USA, 1979. 37. Delaware River Fish and Wildlife Management Cooperative. A Fishery Management Plan for the American Shad (Alosa Sapidissimia) in the Delaware River Basin; Delaware River Fish and Wildlife Management Cooperative: New York, NY, USA, 1982; p. 26. 38. Secor, D.; Gunderson, T.E. Effects of hypoxia and temperature on survival, growth, and respiration of juvenile Atlantic sturgeon, Acipencer oxyrincus. Fish. Bull. 1998, 96, 603–613. 39. Campbell, J.G.; Goodman, L.R. Acute sensitivity of juvenile shortnose sturgeon to low dissolved oxygen concentrations. Trans. Am. Fish. Soc. 2004, 133, 772–776. [CrossRef] 40. Moore, R.B.; Johnston, C.M.; Smith, R.A.; Milstead, B. Source and delivery of nutrients to receiving waters in the northeastern and mid-Atlantic regions of the United States. J. Am. Water Resour. Assoc. 2011, 47, 965–990. [CrossRef][PubMed] 41. Alam, M.J.; Goodall, J.L. Toward disentangling the effect of hydrologic and nitrogen source changes from 1992 to 2001 on incremental nitrogen yield in the contiguous United States. Water Resour. Res. 2012, 48, 1–16. [CrossRef] 42. Preston, S.D.; Alexander, R.B.; Schwarz, G.E.; Crawford, C.G. Factors affecting stream nutrient loads: A synthesis of regional SPARROW model results for the continental United States. J. Am. Water Resour. Assoc. 2011, 47, 891–915. [CrossRef][PubMed] 43. Environmental Protection Agency. Progress in Water Quality: An Evaluation of the National Investment in Municipal Wastewater Treatment, Chapter 7: Delaware Estuary Case Study; Environmental Protection Agency: Washington, DC, USA, 2000; pp. 7.1–7.26. 44. Scatena, F.N.; Curley, D.; Laskowski, S.; Abbott, K.; Bardin, H.; Shieh, W.; Johnson, J. Water quality trading in the lower Delaware River Basin: A resource for practitioners. In Report to the William Penn Foundation by the Institute for Environmental Studies; University of Pennsylvania: Pennsylvania, PA, USA, 2006; p. 86. 45. Environmental Protection Agency. Revisions to Total Maximum Daily Loads for Nutrient and Low Dissolved Oxygen under High Flow Conditions Christina River Basin, Pennsylvania, Delaware and Maryland; Environmental Protection Agency: Washington, DC, USA, 2006. 46. Jones, C.; Branosky, E.; Selman, M.; Perez, M. How Nutrient Trading Could Help Restore the Chesapeake Bay, WRI Working Paper; World Resources Institute: Washington, DC, USA, 2010; p. 13. 47. Trowbridge, P. Analysis of Nitrogen Loading Reductions for Wastewater Treatment Facilities and Non-Point Sources in the Great Bay Estuary Watershed; New Hampshire Department of Environmental Services: Concord, NH, USA, 2010; p. 27. 48. Wieland, R.; Parker, D.; Gans, W.; Martin, A. Costs and Cost Efficiencies for Some Nutrient Reduction Practices in Maryland; NOAA Chesapeake Bay Office: Annapolis, MD, USA, 2009; p. 58. 49. Environmental Protection Agency. Atmospheric Nitrogen Deposition Loadings to the Chesapeake Bay, an Initial Analysis of the Cost-Effectiveness of Control Options; Environmental Protection Agency: Washington, DC, USA, 1996; p. 29. 50. Van Soesbergen, A.; Brouwer, R.; Baan, P.; Hellegers, P.; Polman, N. Assessing the cost-effectiveness of pollution abatement measures in agriculture, industry and the wastewater treatment sector. In WEMPA Report-07 2007; Water Economic Modelling and Policy Analysis, Institute for Environmental Studies: Amsterdam, The Netherlands, 2007; p. 31. 51. Trench, E.C.T.; Moore, R.B.; Ahearn, E.A.; Mullaney, J.R.; Hickman, R.E.; Schwarz, G.E. Nutrient Concentrations and Loads in the Northeastern United States-Status and Trends, 1975–2003: USGS Scientific Investigations Report 2011–5114; United States Geological Survey: Reston, VA, USA, 2012; p. 169. Water 2018, 10, 95 21 of 21

52. Phillips, S.W.; Lindsey, B.D. The Inﬂuence of Ground Water on Nitrogen Delivery to the Chesapeake Bay; USGS Fact Sheet FS-091-03; United States Geological Survey: Reston, VA, USA, 2003; p. 6. 53. Claessens, L.; Tague, C.L.; Band, L.E.; Groffman, P.M.; Kenworthy, S.T. Hydro-ecological linkages in urbanizing watersheds: An empirical assessment of in-stream nitrate loss and evidence of saturation kinetics. J. Geophys. Res. Biogeosci. 2009, 114, G04016. [CrossRef]

© 2018 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).