western portion of the county and alongmajor transportation corridors. Agricul-tural land use, which now accounts foronly about 1 percent of the county, peakedaround 1920. However, stream channelsand floodplains may still have large sur-pluses of sediment resulting from pooragricultural soil conservation practicesof that period (Trimble, 1969).

The major hydrologic feature of GwinnettCounty is the Eastern Continental Divide.

Streams southeast of the divide flow tothe Atlantic Ocean, draining 74 percentof the county. Streams northwest of thedivide flow to the Gulf of Mexico,draining 26 percent of the county. TheChattahoochee River is the principalwater supply for Metropolitan Atlantaand forms the 25-mile northwesternboundary of the county. Buford Damimpounds Lake Sidney Lanier at theextreme northern boundary of the county.

Cha

ttah

oochee

Riv

er

Lake Sidney Lanier

Chattahoochee–Flint River Watershed

Ocmulgee–Oconee–Altamaha River Watershed

Georgia

Atl

antic

Oce

an

Buford Dam

Yellow RiverWatershed

SuwaneeCreek Watershed

AlcovyRiver

Watershed

Brushy ForkCreekWatershed

BigHaynesCreekWater-shed

CrookedCreek

Watershed84°15'

84°7'30"

84°

83°52'30"

33°52'30"

34°

34°7'30"

GWINNETT COUNTY

Snellville

Lilburn

BufordSugar Hill

Suwanee

Lawrenceville

Duluth

Norcross

About15 milesto Atlanta

Gwinnett County land use in 1998(Gwinnett County Department of Public Utilities, 2000).

Park3 percent

Transportation9 percent

Agriculture1 percent

Water1 percent

High- densityresidential4 percent

Commercial, industrial, and utilities 10 percent

Numbers have been rounded

Vacant(undeveloped)

25 percent

Low-to-mediumdensity

residential28 percent

Estateresidential19 percent

Stream quality is a barometer of urbanland-use pressures on a watershed.Managing land use in a watershed is vitalto protecting drinking-water supplies,recreational opportunities, and streamecosystem health. However, the effectof land use and management practiceson streams is difficult to assess and oftenunmeasured. Pollution due to land use,or nonpoint-source pollution, is complexin its origin, transport, impacts, andresponse to management practices. Howdoes urbanization affect sediment ornutrient loads in streams? How effectiveis a detention pond or a stream-bufferrequirement? In six watersheds of Gwin-nett County, Georgia, monitoring resultsof an ongoing study show the effects ofland use on streamflow and on loads ofsuspended solids, metals, and nutrients.

Gwinnett County background

Gwinnett County lies in northeasternMetropolitan Atlanta, Georgia, and isone of the most rapidly developing coun-ties in the United States. The populationof Gwinnett County grew by more than250 percent from 1980 to 2000, and the2001 population is estimated at 621,500(U.S. Census Bureau, 2002).

Land use in the county is broadly mixed,with residential being the largest. Urbandevelopment is most dense in the south-

0 4 6 MILES2

0 6 KILOMETERS2 4

Population

0

200,000

400,000

600,000

1900

1920

1940

1960

1980

2000

EXPLANATION

Eastern Continental Divide

Elevation, in feetHigh: 1,288

Low: 720

Monitoring station

Monitored watershed

Does land use affect our streams?A watershed example from Gwinnett County, Georgia, 1998– 2001

U.S. Department of the InteriorU.S. Geological Survey

U.S. Geological Survey Water-Resources Investigations Report 02-4281December 2002

designed and installed during 1996 asa cooperative investigation of the U.S.Geological Survey and the GwinnettCounty Department of Public Utilities.Stream gages in six watersheds continu-ously monitor water level, streamflow,and precipitation, and periodically collectwater-quality samples. At each streamgage, three storm-composite samples andthree baseflow samples are collectedevery 6 months. In baseflow conditions,streamflow is from ground water thatseeps through the channel bed and banks.Discharge measurements also are routinelymade to define stage-discharge relations.Aquatic invertebrate sampling and habitatassessments were conducted in four water-sheds. This report includes data collectedthrough September 2001. During 2002,monitoring began at six additionalwatersheds in Gwinnett County.

Baseflow samples are collected usingdepth and width integrating techniques.Quality-assurance/quality-control mea-sures ensure that sampled data are notcontaminated by outside sources and arerepresentative of the entire stream crosssection. All stormflow and baseflowwater-quality samples are processedand analyzed at USGS laboratories.

Properties analyzed

During sample collection, standard fieldproperties are measured including pH,specific conductance, water temperature,and dissolved oxygen. Water samples areanalyzed for the following constituents:biological oxygen demand, chemicaloxygen demand, total suspended solids,turbidity, total dissolved solids, totalphosphorus, dissolved phosphorus,several nitrogen species, and hardness.Trace metals analyzed include cadmium,copper, lead, zinc, chromium,and magnesium.

Stream monitoring

Monitoring streams is critical to water-shed assessment and informed manage-

ment. Long-term monitor-ing usingconsistentmethods isessential tocompareconditionsbetweenwatershedsand toidentifytrends. Thewatershed-monitoringnetwork forGwinnettCounty was

Stream-corridor land use

Streams are particularly sensitive to landuse in and near their floodplains. Whenprotected from development, these streamcorridors may filter out and substantiallyreduce pollutants entering streams. TheGwinnett County Stormwater DesignManual (1999) provides incentives topreserve additional stream-buffer areasbeyond the 25-foot requirement ofGeorgia law.

Stream-corridor land use is comparedwith watershedwide land use. A corridorwidth of 30 meters (98 feet) was chosenbased on the limited resolution ofavailable land-use data. Land is muchless developed in the stream corridorthan watershedwide. In the more de-veloped watersheds, however, a greaterpercentage of the stream corridor isdeveloped. The bar chart on the rightillustrates this for transportation land use,which has a high impact on water quality.

0

4

8

12

16

TR

AN

SP

OR

TA

TIO

N L

AN

DU

SE

, IN

PE

RC

EN

T

WATERSHED

Watershed-wide

Stream corridorof 30 meters

Lessdeveloped

Moredeveloped

BrushyFork

Creek

BigHaynesCreek

AlcovyRiver

SuwaneeCreek

YellowRiver

CrookedCreek

02207120 GWLT-02 Yellow River at SR 124 near Lithonia, Ga. 16202207385 GWLT-04 Big Haynes Creek at Lenora Road near Snellville, Ga. 17.302207400 GWLT-05 Brushy Fork Creek at Beaver Road near Loganville, Ga. 8.1502208150 GWLT-07 Alcovy River at New Hope Road near Grayson, Ga. 30.802334885 GWLT-12 Suwanee Creek at Buford Highway near Suwanee, Ga. 4702335350 GWLT-01 Crooked Creek at Spalding Drive near Norcross, Ga. 8.89

USGSstationnumber

GwinnettCounty

site numberMonitoring-station name

Area, insquaremiles

Assessment of nonpoint-source pollutionrequires water-quality sampling duringrainfall runoff. Stormwater samples arecollected with automatic samplers toobtain a discharge-weighted compositesample of the storm runoff.

0

40

80

120

160

200

JANUARY 2001

ST

RE

AM

FLO

W, I

N C

UB

ICF

EE

T P

ER

SE

CO

ND

19 20

Sample taken

Big Haynes Creek

Stream-monitoring stationat Crooked Creek.

Crooked Creek during baseflow. Crooked Creek during stormflow.

Monitoring stations for the six monitored watersheds

The seasonal nature of streamflow is evident in the monthly mean water yield. Water yield is the total volumeof streamflow during a month or year divided by watershed area. The average yield for the wettest month(March) is more than three times greater than that of the driest month (October). Annual variability isevident in the annual mean flow and yield. The 1998 average flow was at least twice the flow of the droughtyears 1999–2001. For the period 1998–2001 in the Yellow River watershed, the average total water yieldwas 23.7 inches, which is 52 percent of the average precipitation.

Annual mean flow and water yield

0

100

200

300

400

500

1998 1999 2000 2001 1998–2001

FLO

W, I

N C

UB

ICF

EE

T P

ER

SE

CO

ND

FLO

W, I

N C

UB

ICF

EE

T P

ER

SE

CO

ND

Monthly mean flow and water yield, 1998–2001

0

1

2

3

4

0

114

228

342

456

570

Oct Nov Dec Jan Feb Mar Apr May June July Aug Sept

YIE

LD, I

N IN

CH

ESStormflow

Baseflow

0

7

14

21

28

35

42

YIE

LD, I

N IN

CH

ES

Yellow RiverWatershed

Jackson Creek

Creek

Beaver Ruin Creek

Yello

wJa

cks

Riv

er

Yello

w

River

Sw

eetwater

Creek

84°7'30"

84°

33°57'45"

33°50'42"

0 1 4 5 KILOMETERS2 3

0 1 2 3 4 5 MILES

Transportation12 percent

Park3 percent

High-density residential2 percent

Numbers have been rounded. Areas less than

1 percent not shown

Low-to-mediumdensity residential

42 percent

Commercial, industrial, and utilities 14 percent

Vacant(undeveloped)

17 percent

Estateresidential10 percent

A closer look at land use and hydrology,Yellow River watershed

The Yellow River watershed is the largest(162 square miles) in the county and liesin the Ocmulgee River watershed. It issituated in the more densely developed,southwestern part of the county. Land use(in 1998) is shown for the portion (97percent) of the watershed located withinGwinnett County. The predominate landuse (42 percent) is low-to-medium densityresidential. Commercial and industrial de-velopment are concentrated along the majortransportation corridors. Four wastewater-treatment facilities discharge into the water-shed upstream from the monitoring station.

Continuously monitored streamflow can bedivided into components of stormflow andbaseflow using hydrograph separation tech-niques (Sloto and Crouse, 1996). Stormflowis the rainfall runoff that enters the drainagenetwork from overland flow. As a percentof total flow, stormflow is greatest duringwet months and years. Baseflow enters thedrainage network through the streambedand banks and is more stable.

Yellow River watershed land use in 1998.

Land use fromGwinnett CountyDepartment of PubicUtilities, 2000

Yellow RiverWatershed

Gwinnett County

EXPLANATION

Monitoring station

Wastewater-treatment facility

Streamflow concentrations and watershedloads and yields, Yellow River watershed

Streamflow samples are analyzed to deter-mine constituent concentrations. These datadefine whether the stream quality is adequatefor uses such as recreation, for drinking-water supply, for fishing, and as a habitatfor aquatic organisms. However, one mustknow more than the condition of a stream inorder to mitigate pollution and manage theresource. Is the pollution from point sourcesor nonpoint sources? When and how are con-stituents transported? An intensive monitor-ing program, such as the one for GwinnettCounty, can address these questions.

Constituent load, shown to the right inunits of tons per year, is the product ofconcentration and discharge. Constituentyield, shown in units of pounds per acreper year, is the load divided by the water-shed area. Loads are computed using arelatively simple rating-curve model(Crawford, 1996). The model is cali-brated and run using a stormflow load-to-discharge curve, and a baseflow load-to-discharge curve, and applied to the dailytime-step hydrograph for the Yellow River.The results, summarized in the graphs tothe right, illustrate several important points.

• Water quality typically is worse duringstormflow than during baseflow.

• Concentration-to-discharge curvesare dissimilar for stormflow andbaseflow for suspended solids and forconstituents having an affinity for solidphase particulates (such as metals);whereas the curves are similar fortotal dissolved solids.

• The absence (lead) or presence (totalnitrogen) of constituents from pointsources is evident in the baseflowconcentration-to-discharge curves.Concentrations of constituents frompoint sources are diluted and decreasewith increasing baseflow.

• Nonpoint-source constituent concen-trations increase with increasing

streamflow because larger storms typi-cally wash off more constituent massfrom a larger area of the watershed.

• In constituent loads and yields, the effectof dissimilar baseflow and stormflowconcentration characteristics is multi-plied. Stormflow loads of nonpoint-source constituents vary exponentiallywith stream discharge because concen-tration increases with discharge, and the

load is a product of concentrationand discharge.

• Seasonal and annual loads vary withprecipitation and total streamflow. Base-flow loads are relatively constant fromyear to year.

These findings are relevant to analyses oftotal maximum daily load, watershed assess-ments, and watershed protection strategies.

1

10

100

1,000

10

100

1,000

10,000

0

40

80

120

140

0

1

2

3

4

DISCHARGE, IN CUBIC FEET PER SECOND

0

10

20

30

40

0100200300400500600700

Annual yield and loadConcentration and discharge

0

1,000

2,000

3,000

4,000

5,000 260,000

195,000

130,000

6,500

0

36,000

27,000

18,000

9,000

0

1,300

975

650

325

0

8

6

4

2

00

4

8

12

16

1998

1,000100

1999 2000 2001 1998–2001

0

5

10

15

20

25

CO

NC

EN

TR

ATIO

N, I

N M

ILLI

GR

AM

S P

ER

LIT

ER

CO

NS

TIT

UE

NT

YIE

LD,

IN P

OU

ND

S P

ER

AC

RE

PE

R Y

EA

R

CO

NS

TIT

UE

NT

LO

AD

, IN

TO

NS

PE

R Y

EA

R

CO

NC

EN

TR

ATIO

N,

IN M

ICR

O-

GR

AM

S P

ER

LIT

ER

CO

NC

EN

TR

ATIO

N,

IN C

OLO

NIE

S P

ER

100

MIL

LILI

TE

RS

Stormflow

Stormflow

Baseflow

Baseflow

A. Total suspended solids A. Total suspended solids

B. Total dissolved solids B. Total dissolved solids

C. Total nitrogen C. Total nitrogen

D. Total lead

E. Fecal-coliform bacteria

D. Total lead

Collecting a stream sample at baseflow.

A. Excessive total suspended solids are a major cause of habitat degradation innorth Georgia streams. Stormflow and baseflow have dissimilar total-suspended-solidcharacteristics; 99 percent of the annual load is carried by stormflow. B. Total dissolvedsolids are naturally higher in baseflow, which has filtered through the ground. The totaldissolved solids-to-discharge curve is similar for baseflow and stormflow. C. The dilutionof point-source nitrogen is evident in baseflow. Nonpoint-source nitrogen increases withdischarge. These nitrogen levels do not indicate an impaired status for the Yellow River.D. Lead concentrations are typically below detection limits during baseflow, indicatingno lead point sources. Stormflow carries 97 percent of the annual lead load. E. Fecal-coliform bacteria serve as an indicator for possible pathogenic bacteria in streams.Fecal-coliform data are highly variable, but stormflow concentrations typically aremuch greater than baseflow concentrations.

Monitoring data from six GwinnettCounty watersheds show strongrelations between transportationland use in the stream corridor andwatershed yield of total suspendedsolids and lead.

Effects of land use on stream hydrology

Bare or impervious areas and a moredeveloped stormwater drainage systemproduce greater volumes of high-energystormflow and reduce baseflow in astream. A measure of this effect is theaverage stormflow of a watershed, takenas a percentage of total watershed flow.The bar chart below shows that with

Nitrogen yield can be estimated from water-shed land use and point-source indices.

Effects of land use on stream loadsand yields

Increased stormwater runoff accelerateserosion of the land surface, particularlyfor areas under construction, and causesscouring of stream channels resulting inmuch higher stream-sediment loads. Totalsuspended solids are used as a measureof overall development impacts in theGwinnett County Watershed ProtectionPlan (2000). The usefulness of thatmanagement strategy is illustrated herein the strong relation shown betweentotal suspended solids and transportationland use in the stream corridor.

Lead levels in most streams have de-creased since the early 1970s as a resultof compliance with environmental laws.Lead yields in the six monitored water-sheds are closely related to transportationland use in the stream corridor. Thismay be due to increased lead depositionin paved areas by transportation andrelated land uses; and because of thegreater efficiency of wash off fromthese surfaces, accompanied by reducedfiltration of runoff before it reaches thedrainage network.

Nitrogen loads in the watershed have bothnonpoint and point sources. Nonpointsources include natural organic detritusand applied nitrogen, such as fromfertilizer. Nitrogen yield was estimatedbased on an index of whether or not the

increasing watershed development,stormflow increases as a percentage oftotal watershed discharge.

Transportation land use can be anindicator of the total impervious areathat is connected to a drainage network.Stormflow was correlated with transpor-tation land use in the stream corridor and

watershed contains wastewater-treatmentfacilities, and based on land-use percent-ages in the following categories: low-to-medium density residential, high-densityresidential, commercial/utilities,and transportation.

The annual yield of a constituent instreamflow is affected by an array ofland uses, point sources, hydrologicfactors, chemical processes, and otherfactors. Effective water-resource manage-ment is difficult because of these manyfactors. However, stream monitoring canprovide information and tools that arevaluable for watershed assessments,land-use planning, and informed water-resource management.

watershedwide. The correlation is slightlybetter for stream-corridor land use. Addi-tional stream-corridor analyses are neededusing higher resolution land-use data.

4 6 8 10 12 14

MEASURED TOTAL NITROGEN YIELD,IN POUNDS PER ACRE PER YEAR

4

6

8

10

12

14

ES

TIM

ATE

D T

OTA

L N

ITR

OG

EN

YIE

LD,

IN P

OU

ND

S P

ER

AC

RE

PE

R Y

EA

R

BR

SU

BH

CR

YL

AL

0 5 10 15TRANSPORTATION LAND USE, IN PERCENT

30

25

35

40

45

50

55

60

STO

RM

FLO

W (

1998

–20

01)

AS

PE

RC

EN

TAG

E O

F T

OTA

L F

LOW

Watershedwide

Streamcorridor

BR

SU

BH

CR

YL

AL

BR

SU

BH

CR

YL

AL

0

10

20

30

YIE

LD, I

N IN

CH

ES

StormflowBaseflow

Mean annual water yield, 1998–2001

BrushyFork

Creek(BR)

BigHaynesCreek(BH)

AlcovyRiver(AL)

SuwaneeCreek(SU)

YellowRiver(YL)

CrookedCreek(CR)

WATERSHEDLessdeveloped

Moredeveloped

20 4 6 8 100.02

0.04

0.06

0.08

BR

SU

BH

CR

YL

AL

TRANSPORTATION LAND USEIN THE STREAM CORRIDOR,

IN PERCENT

TOTA

L LE

AD

YIE

LD, I

N P

OU

ND

SP

ER

AC

RE

PE

R Y

EA

R

500

1,000

1,500

2,000

2,500

TOTA

L S

US

PE

ND

ED

SO

LID

S Y

IELD

,IN

PO

UN

DS

PE

R A

CR

E P

ER

YE

AR

BR

SU

BH

CR

YL

AL

20 4 6 8 10

Sediment from stormflow into theYellow River.

Stormflow, as a percent of total flow,increases with increasing transportationland use.

Effects of land use on stream ecosystems

Just as stream-water quality is affectedby watershed land use, so are the aquaticorganisms. Stream communities in Gwin-nett County are composed of a relativelydiverse group of organisms rarely seenby most people. These stream inhabitantsspend all or parts of their life cycles inaquatic environments and include insectsas well as crayfishes, snails, mussels, andaquatic worms. Since these organismsgenerally live in the same section of streamduring their entire lives, the invertebratecommunity that develops is correlated withwater-quality and habitat conditions of thatstream section.

To determine the Georgia EcologicalCondition Scores (Georgia Departmentof Natural Resources, 1997), invertebratescollected during site visits are identifiedand counted by a trained taxonomist. Simi-lar species are grouped according to howtolerant they are to pollution or how theyfeed in the aquatic environment. Using

these data, indices that have been shownto be related to environmental degradationare calculated.

Part of the Georgia Ecological ConditionScore is composed of a separate indexknown as the Georgia Visual AssessmentHabitat Score (Georgia Department ofNatural Resources, 1997). This indexprovides managers with an assessmentof the physical condition of the streamhabitats present at a site and allows com-parison among streams included in thestudy as well as to regional referencesites. Habitat data allow more accurateinterpretation of the biological dataespecially when viewed in conjunctionwith water-quality data.

Invertebrate samples were collectednear the stream gages of four of the sixmonitored watersheds. A reference site,located on the Alcovy river in the upperAlcovy watershed, was selected andsampled to represent relatively unimpairedecological conditions in Gwinnett County.Georgia Ecological Condition Scores areexpressed as a percentage of the score ofthe reference site and rated from verygood to very poor, in comparison withconditions across northern Georgia,including Gwinnett County.

The most densely urbanized watershed,Crooked Creek, has the lowest EcologicalCondition Score. However, for the threeother watersheds, the comparative eco-logical condition does not agree with thecomparative land-use development. Thismay be largely due to the localized effectsof past channelization on aquatic habitat in

For more information contact:

State RepresentativeU.S. Geological Survey3039 Amwiler Road, Suite 130Atlanta, GA 30360-2824770-903-9100http://ga.water.usgs.gov

Written byMark N. LandersPaul D. AnkcornKeith W. McFaddenM. Brian Gregory

Design and layout byCaryl J. Wipperfurth

U.S. DEPARTMENT OF THE INTERIORGALE A. NORTON, SecretaryU.S. Geological SurveyCharles G. Groat, Director

Big Haynes Creek and Brushy Fork Creek.The effects of channelization also areapparent in the habitat scores.

Streams are strongly affected by watershedland use. Land-use effects are evident inthe amount of streamflow during storms,in the amount of constituents in a stream,and in the stream ecosystem. Monitoringand understanding of these effects provideinformation for the wise management ofland use and water resources.

Georgia Ecological Condition Scores

0

20

40

60

80

100Verygood

Good

Fair

Poor

Verypoor

PE

RC

EN

T O

F R

EF

ER

EN

CE

CO

ND

ITIO

N

Georgia Visual Assessment Habitat Scores

0

30

60

90

120

150

TO

TA

L H

AB

ITA

T S

CO

RE

Alcovy River

BigHaynesCreek

BrushyFork

Creek

CrookedCreek

Reference site score

hydrograph separation and analysis:U.S. Geological Survey Water-ResourcesInvestigations Report 96-4040, 46 p.

Trimble, S.W., 1969, Culturally accelerated sedi-mentation on the middle Georgia Piedmont:Athens, Georgia, The University of Georgia,unpublished masters thesis, 110 p.

U.S. Census Bureau, State and County Quick-Facts, accessed on September 1, 2002, athttp://quickfacts.census.gov.

Base maps from Gwinnett County Board of Com-missioners, Department of Support Services.

Gwinnett County Department of Public Utilities,2000, Land-use classification for 1998.

Gwinnett County Stormwater Design Manual,1999, revised October 2001: GwinnettCounty Georgia Department of PublicUtilities, Stormwater ManagementDivision, 277 p.

Gwinnett County Watershed Protection Plan,2000: Gwinnett County Department ofPublic Utilities, 54 p.

Sloto, R.A., and Crouse, M.Y., 1996, HYSEP:A computer program for streamflow

References citedCrawford, C.G., 1996, Estimating mean con-

stituent loads in rivers by the rating-curveand flow-duration rating-curve methods:Bloomington, Indiana, Indiana University,unpublished Ph.D. dissertation, 245 p.

Georgia Department of Natural Resources,1997, Draft standard operating procedures—freshwater macroinvertebrate biologicalassessment: Atlanta, Ga., EnvironmentalProtection Division, Water ProtectionBranch, 139 p.

U.S. Geological Survey Water-Resources Investigations Report 02-4281December 2002

Collecting an invertebrate samplefrom riffle habitat in a stream inMetropolitan Atlanta.

Georgia’s Ecological Condition andVisual Assessment Habitat Scores forselected watersheds in Gwinnett County.