Using a New Groundwater-RegressionModel to Forecast Nitrogen Loading

from the Delmarva Peninsula

Ward Sanford, USGS, Reston, VirginiaJason Pope, USGS, Richmond, VirginiaDavid Selnick, USGS, Reston, Virginia

Objectives:· To develop a groundwater flow model that can simulate return-times to streams

(base-flow ages) on the Delmarva Peninsula

· To explain the spatial and temporal trends in nitrate on the Delmarva Peninsula using a mass-balance regression equation that includes the base-flow age distributions obtained from the flow model

· To use the calibrated equation to forecast total nitrogen loading to the Bay from the Eastern Shore

· To forecast changes in future loadings to the bay given different loading application rates at the land surface

· To develop maps that will help resource managers target areas that will respond most efficiently to better management practices

Groundwater Model—Delmarva Peninsula

MODFLOW 2005500 ft cell resolution7 Model Layers4+ million active cells30-m DEM, LIDAR300 ft deepSteady State FlowMODPATH travel timesSanford and others (2012)

USGS Open File Report2012-1140 (in press).

1 2

34

5

6

7

Seven watersheds hadsubstantial stream nitrateData and were used:

1. Morgan Creek

2. Chesterville Branch

3. Choptank River

4. Marshyhope Creek

5. Nanticoke River

6. Pocomoke River

7. Nassawango Creek

1950 1960 1970 1980 1990 2000 2010 20200.00.20.40.60.81.01.21.41.61.82.0

Choptank River near Greensboro, MD

Annual Averages

Multi-Year Averages

YEAR

Nitra

te a

s N,

in m

g/L

1950 1960 1970 1980 1990 2000 2010 20200.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0Nanticoke River near Bridgeville, DE

Yearly Averages

Multi-Year Averages

YEAR

Nitra

te a

s N,

in m

g/L

1950 1960 1970 1980 1990 2000 2010 20200.0

0.4

0.8

1.2

1.6

2.0

2.4

2.8

3.2

3.6

4.0Morgan Creek near Kennedyville, MD

Yearly AveragesMulti-Year Averages

YEAR

Nitra

te a

s N,

in m

g/L

1 10 100 1000 100000

1

10Choptank River near Greensboro, MD

Discharge, in Cubic Feet per Second

Nitra

te a

s N,

in m

g/L

Nitrate Mass-Balance Regression Equation

Simulatedsurface waterconcentration

in stream

=

Concentrationat water table

at recharge year

SoilTerm

RiparianTermX

Denitrification

Simulatedgroundwaterconcentrationin ag field well

=

Concentrationat water table

at recharge year

XSoilTerm

Concentrationat water table

at recharge year= Fertilizer

Load [ XPoultryLoad

RechargeRate ]X X{ [] }1 - 1 -FUE PUE+

FUE & PUE =Fertilizer and PoultryUptake Efficiencies

SoilTerm = (S/5)m

Denitrification ratealong flow path

RiparianTerm = (AREA)n

S = soil drainagefactor

AREA is for the watershedIn square miles

Sn=1

No.pathlines

XX

Denitrification ratealong flow pathX

0

10

20

30

40

50

60

1940 1950 1960 1970 1980 1990 2000 2010

Year

Nitr

ogen

Loa

ding

from

Fer

tiliz

er, m

g/L

Kent

New CastleSussex

CarolineCecil

Dorchester

KentQueen Annes

Somerset

Talbot

WicomicoWorcester

Accomackaverage

1 4 7 10 13 16 19 22 25 28 310.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

Stream and Riparian Denitrification

Euivalent distance traveled along the stream, in miles

Frac

tion

Nitra

te R

emai

ning

For calibrated watershed areaexponent parameter = -0.119

Soil Factor (S) in Equation:

1. Very poorly drained2. Somewhat poorly drained3. Poorly drained4. Moderately well drained5. Well drained6. Somewhat excessively drained7. Excessively drained

1 2 3 4 50.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

Soil Factor, S

Frac

tion

Nitra

te R

emai

ning

For calibrated soil drainageexponent parameter = 0.337

Very poorlydrained

Welldrained

Fertiliz

er in Rec

harge

After C

rop Uptake

After S

oil Den

itrific

ation

After n

on-ag di

lution

After p

re-ag

dilutio

n

After r

iparian

denitri

ficati

on

observe

d0

5

10

15

20

25

30

35

40

45

50

Morgan Creek

Pocomoke River

Process Leading to Nitrate Loss or Dilution

Con

cent

ratio

n on

Nitr

ate

as N

, in

mg/

L

Regression Number

Fertilizer

Uptake

Efficiency

Manure Uptake Efficiency

Percent Increase in

Uptake Efficiencies

Riparian and Stream

Denitrificait

on loss exponent

Soil Deintirficaiton loss

exponent

Number of

Parameters

estimated

Sum of Squared

Weighted Residual

s

Standard Error of

Regressi

on

1 83% 83% 0 0 0 2 6141 89

2 82% 70% 0 0 0.670 3 5721 83

3 72% 63% 0 -0.152 0 3 2394 35

4 72% 46% 0 -0.146 0.425 4 2139 31

5 67% 47% 24% -0.119 0.377 5 1651 25

10% upper parameter value

limit70% 62% 40% -0.120 0.77 5 2056 40

10% lower parameter value

limit64% 32% 12% -0.155 0.1 5 2056 40

10% limits as percent of value 4% 32% 38% 13% 82% 5 2056 40

1950 1960 1970 1980 1990 2000 20100.00.20.40.60.81.01.21.41.61.82.0

Choptank River near Greensboro, MD

Yearly AveragesMulti-Year AveragesRegression Model

YEAR

Nitra

te a

s N,

in m

g/L

Best Fit for Four Parameterswith constantFertilizer and ManureUptake Efficiencesthrough Time

1950 1960 1970 1980 1990 2000 20100.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0Nanticoke River near Bridgeville, DE

Yearly AveragesMulti-Year AveragesRegression Model

YEAR

Nitra

te a

s N,

in m

g/L

1950 1960 1970 1980 1990 2000 2010 20200.0

0.4

0.8

1.2

1.6

2.0

2.4

2.8

3.2

3.6Morgan Creek near Kennedyville, MD

Yearly AveragesMulti-Year AveragesRegression Model

YEAR

Nitra

te a

s N,

in m

g/L

1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 20500

4

8

12

16

20

24 Forecast of Nitrogen Loading

Total Flow N Loading

Had there been no BMPs

High Flow Component

YEAR

Nitr

ogen

Loa

d to

the

Bay

, in

Mill

ions

of P

ound

s pe

r Yea

r

1980 1990 2000 2010 2020 2030 2040 20500

2

4

6

8

10

12

14

16

18

20

Forecast of Nitrogen Loading

13% load reduction

40% load reduction

100% load reduction

0% load reduction

YEAR

Nitr

ogen

Loa

d to

the

Bay

, in

Mill

ions

of P

ound

s pe

r Yea

r

EPA TARGET

Targeting Matrix

Targeting that Includes Response Timeand Nitrogen Delivered to the Bay

Groundwater Return Time

< 7 yrs 7 - 20 yrs > 20 yrs

NitrateConcentration

< 4 mg/L

5 - 7 mg/L

>7 mg/L

Targeting Matrix

Targeting that Includes Response Timeand Nitrogen Delivered to the Bay

Groundwater Return Time

< 7 yrs 7 - 20 yrs > 20 yrs

NitrateConcentration

< 4 mg/L

5 - 7 mg/L

>7 mg/L

Summary and ConclusionsResults from a groundwater flow model were coupled to a nitrate-mass-

balance regression model and calibrated against stream nitrate data.

The calibrated model suggests that nitrogen uptake efficiencies on the Eastern Shore may be improving over time.

Response time of nitrogen delivery to the Bay on the Eastern Shoreis on the order of several decades

EPA targets are for reduced loading of ~20% (3 million lbs/yr) on the Eastern Shore. This cannot be accomplished by reducing landsurface applications by 20%, as loads will continue to rise 13%.

The new model can help target areas where reduced nitrogen loadings would be the most beneficial at reducing total loadings to the Bay.