NC0031879_Report_19810803

DIVISION OF .E IVI RME TAL. VACIAGEENT
Aug s t 30 1981
Mr. Howard .Zeller: . tiro Director
Enforcement Division
U. S.. Environmental P rotiect9 on Agency
Jgion IV
345 Courtlattd Street, K.E.
Atlanta, Geoneie 30355
Subject: *equine -for Modification
NHS Discharge Permit
Corpening Creek Wastreatar Treatment
Plant
City of Marion
Nevi 1 County, Norge Car of i r ,
Dear -Mr. _Wien
The City -of Marion has 'requested thst the4Mvision oi' Environeental
etenagement review the effiteet lialts •contained es a part of tiPIES Dis-
charge Pert ire. NCQU3187Ss, to determine if the limitations platted upon
$shosptspr t s :and exygen..constaingooaste,,are valid in -light of present day
standards :and allocation amtho4iogy.
Limitations for ogyeen consuming waste were the result of a modified
Level E, evaluation tanning limited field work specific to Corpening Cry
was utilized for purposes of developing constants for use in the modified
Streat r.Pheips Equation. phone limits .ware okiyeloped primarily as
a product of l 1 mi ted worst by the C. S. Env1 ron etetal Protection Agency
to the headeater lakes of the. Catawba River. The. City of Marion has ex-
pressed intermit in gaining sera relief..fwa .present -flail effluent limits
for they reason .that the City might thereby experience saving in oiler
tional merits at its Corpening'Creek .Wasteeater Treatment 'Plant,
The Division of Environmental ibutagesmnt is willing to invest tip
seeking the deterizination as to whether conditions as they exist tad th
Lake [deodiss or as they mayexist in the near future might allow difica-
tionon of the phosphorus limits for the Careening Creek -Wastewater Treatment
Plant. Additionally, the Division of Environmental 1enagament is willing
t3 invest its resources 1n the .conduct of a Level C study of Corpening •=s _z
Creek which would inc.l an intensive .survey with .slug stapling to dam'/ 89, ,
mine if the present graft limits for oxygen -aoesema1ng waste _are in
of modification. Obviously* the tneesis nt of time and elonay in
effort by the € i vi si or, of Environmental Management would not be_ i
if the U. S . Erivimegental Protection Agency.would not allow vela
certain l i rl is should the above described steer demenstrat+e that
as r",=,
fleet.9 `
Steldeeets would be protacted in Corpen1ng k by the distiitrge
taztemater° cf E lesser-q. .11ty -then prasen.tly *scribed .in the City*WPEES"

i .. Boward-.Zo11or
August
Ponta - . intviouslys .effluent limits could,be-ne less stringent that required.
• br applicable. effluent guidelines. The4tate-of.Nortit tsrolina would be ":ie-
t®a .tad in giving this rt11af to the= CI ,t -of Ration if justified so 'that
the •City ie " tun night be In a positioe-toes ri - cost savings as.. a- .
sul t of less stringent effluent requitetalmtep
would appreciate yottirgiving this r . -careful contideratits.
Should it be helpful to you in arriving at..-e -decision to *lease this pru-
,posal further9.. lone contact °. L. P. - liontone..Chief of tho tioter.QoU ty
Rio at ►
Sincerrl ►ems.
Original Sips. By
ROBERT" F. HELMS
Robert F. Helm
Director
eica 4ir. L. P. BentOa o Jr..s tlief
bIater Qoaa1-1y Section
14r. tie Earl Dastielsa Amster -
City
of -art:
Pti..:.-Neath-P.:Akibsoni, Regional Engines
Vattir Quality. Section o Asheville
t-.

INTENSIVE SURVEY
LAKE RHODHISS
CITY OF MARION - CORPENING CREEK WWTP STUDY
JULY 28 AND 29, 1982
DIVISION OF ENVIRONMENTAL MANAGEMENT
WATER QUALITY SECTION
TECHNICAL SERVICES UNIT
PHYSICAL AND CHEMICAL MONITORING

Introduction
The North Carolina Department of Natural Resources and Community
Development, Division of Environmental Management, at the request of
the City of Marion has agreed to review the effluent limits for their
Corpening Creek WWTP. Of special concern are the limits on phorphorus
EPA imposed based on a study the agency did in 1973 and 1974.
Since that study, several larger municipal WWTP's have begun
discharging into the Catawba River drainage basin which feeds Lake
Rhodhiss. Because of the possible changes in phosphorus loading rates
and the cost of phosphorus removal, the City of Marion feels the
Division of Environmental Management should re-evaluate the need for
the levels of phosphorus reduction they are required to meet.
On July 27, 28, and 29, 1982, the Technical Services Branch of
DEM made a detailed study of the Marion Corpening Creek WWTP, and
the Lake Rhodhiss drainage area, including other permitted dischargers
in question. All sampling was done according to accepted methods of
water data acquisation of the N. C. Department of NRCD, EPA, USGS, and
Standard Methods.

Description of Lake Rhodhiss Study Area
Lake Rhodhiss is located on the southeastern border of Caldwell County
and in the central portion to eastern border of Burke County. The lake is
generally of riverine type with an approximate length of 13 miles. The
width of the lake is about 1/8 mile at its headwaters and graduates to about
1/4 mile at its eastern end where it is blocked by the Lake Rhodhiss Dam.
The average depth of the lake was determined to be 20.6 feet with a surface
area of 11.1 sq.. Km. The estimated retention time of the lake is 17 to 21
days. Lake Rhodhiss is fed primarily by the Catawba river which drains sub -
basin 03-08-30 upstream and the sub -basin 03-08-31 which drains directly into
the lake and its immediate headwaters. Both drainage areas are considered
in this report although Lake James is located in sub -basin 03-08-30 and due
to its large size serves as a natural water quality treatment area. With
this fact in consideration greater emphasis was placed on sub -basin 03-08-31.
The general topography of the area is considered to be that of the
Appalachian foothills with the Hydro Environmental Area (HEA) being classified
as upper Piedmont. The area surrounding the lake is basicly forested with
small industrial towns as well as small recreational facilities.
There are 13 major tributaries entering Lake Rhodhiss. Of these 13, 12
are classified as C streams, and 1 is classified as an A -II stream. Lake
Rhodhiss, and the Catawba river from the Rhodhiss Dam upstream to Lake James
is classified as A -II.

Point Source Inputs into the Lake Rhodhiss Drainage Area
There are 72 permitted dischargers in the two sub -basins within the
Lake Rhodhiss drainage area. Forty of the dischargers are located in
sub -basin (03-08-31) which immediately surrounds the lake area, with
32 dischargers being located in the upstream sub -basin (03-08-31).
All dischargers are located on the sub -basin maps with permit numbers,
receiving stream and geographic location on the following pages. of the 40
discharges in sub -basin (03-08-31), there are 3 water treatment plants,
8 wastewater treatment plants, 14 industrial sites, 11 schools, rest
homes or correctional centers and 4 domestic permitted point sources.
0f the 32 dischargers in sub -basin (03-08-30), there are, 4 wastewater
treatment plants, 9 industrial sites, 6 schools, rest homes or
correctional centers and 13 domestic permitted dischargers. In sub -
basin (03-08-30) where the Marion-Corpening Creek WWTP is located, there
is a combined design flow of 11.41 MGD entering the area. Marion-Corpening
Creek WWTP has a design flow of 3.0 MGD which accounts for 26.0% of the
combined design flow of that sub -basin. In sub -basin (03-08-31) there
is a combined design flow of 29.97 MGD, which when added to sub -basin
(03-08-30) yeild a total design flow for the two sub -basins of 41.38.
Marion-Corpening Creek WWTP accounts for 7.25% of the combined design
flow of the 2 sub -basins which drain the Lake Rhodhiss area.

• Station Locations and Parametric Coverage
The Marion-Corpening Creek / Lake Rhodhiss Study was divided into
3 strategic sampling areas which consisted of lake stations, stream
(tributary) stations and point source discharger stations. By determining
background nutrients entering the lake via stream, permitted discharger in-
put, and existing lake nutrient status, a total nutrient budget could be
concluded.
Stream sampling was performed at 7 key locations to determine nutrient
loadings into Lake Rhodhiss by way of tributary introduction. Sampling was
performed from bridges at all stations using lab -line samplers for
obtaining grab samples and a Hydrolab 4041 for all physical parameters.
Each station was sampled 7 times between 0845 on 82/07/28 and 0710 on 7/29/82.
Grab sample parameters consisted of N and P, chlorophyll a, pheophytin and
phytoplankton. Physical parameters included dissolved oxygen, pH, temperature
and conductivity.
Seven lake stations were sampled to evaluate nutrient concentration levels
and determine potential lake status. Secchi disk depth and dissolved oxygen
were taken to determine the euphotic zone and the hypolimnion. Integrated
samples were taken from the euphotic zone for analysis of N and P series and
Chlorophyll a. A separate sample for N and P was also taken from the
hypolimnion. Conductivity, temperature and pH were also taken throughout
the water column.
Point source dischargers into the Lake Rhodhiss drainage area are a
primary concern in nutrient loadings. All dischargers with a design flow
of equal to or greater than 0.5 MGD were sampled during the study to
evaluate their nutrient input. Isco samplers were used for all time con-
troled composite samples. Marion-Corpening creek WWTP was composite sampled
at four locations: influent (before P removal), effluent (after P removal),
upstream outfall (Corpening Creek), downstream outfall (Corpening-Creek).
Morganton - Catawba River WWTP, Lenoir - Lower Creek WWTP, Great Lakes Carbon
WWTP and Valdese - McGalliard Creek WWTP were composite sampled at their
effluent outfall. Valdese - Hoyle Creek WWTP and the Drexel - Howard Creek
WWTP were grab sampled three times in a 24 hr. period.

FY:�t-o. *' {"-•!�' ° 7 :aft` . r •• .�7:.✓'- t's., j. "fi • .r• ;•a .,sa-``._ ... Yx• SL•.- .--^Vx.':-V .a•.: Kam: ..i, e vti' - 'sc.-
S+.:L3d`�'�s�:i ciR.Iys,"�v �s�i�W..-.r..=c��.�i: t....s.r_..s��j�.:?� ��i..x-:�S.r...'�t�a�. ,.:v^'�..�-i�.'d'i..'i�V�V1 �.c5�c`i�:.�1r.`.•.��4A`i�u�.f...f-•ICs'+r:� .� ' �F" . i•..y��i • :^U-� � • ma's. � .awh `s•a}`�;;'��'•.."'�T!s�:R�'-"�..:.�sfaxa ii3�.:'�iM7��`a`w •. r .1� .�y �.a: dr : •{..^-L cam'} r R.r�:. ,i:• s;� .�yr. u�. '+ �: �'i"•'Yx.;::•rur:
C[' 'rc -a5 .�°S •• 1.•i.*• i...:�.• .c.2S. � :.t a w.i+��..rw+. � .xtt�•S �'�.1i': _ � - .. ...
Map of Dischargers in the Lake Rhodhiss Drainage Area
(Sub -basin 03 08 30)

Permitted Dischargers in Sub -basin 03-08-30
Map Design
# Flow Discharger
1 4.200
2 .200
3 .016
4 .005
5 .050
6 .005
Chalet Mtr Lodge
NC0030996
American Thread
NC0004243
Travenol Labs
'NC0006554
Glenwood Elem. School
NC0033499
22 .070 Crossmore WWTP NC00266654
23 .018 Div. cf Forest Resources
NC0040339
24 .050 Linville Land Harbor
NC00227 36
25 .001 Mill :'l.mber Nursery
NC0046221
Old Fort Finishing
NC0005002
Old Fort WWTP NC0021229
C.E. Air Preheater Co.
NC0039934
Pleasant Garden Elem. Sch
NC003464
McDowell Co. Sr. High Sch
NC003481
Marion Plaza Shoppping
Ctr. NC0030911
7 .001 Burger Boy Drive -In
NC0044415
8 .001 Broyhill Ind. NC0006726
9 .008 West Marion Elem Sch
NC003456
10 .005 (Artrip Petroleum Inc)
Collins Truck Stop
NC0029831
11 :008 Westwood Chateau Apts.
NC0039322
12 .016 NC DOT I-40 Rest Area
NC0024864
13 .005 Pinnacle Rest Home
NC0035157
14 .300 Marion WWTP #2
NC0027413
15 3.00 Marion WWTP (Corpening)
NC0031879
16 .020 Lemon Tree Inn
NC0040291
17 .008 Nebo School NC0033472
18 .010
19 2.000
20 1.200
21 .009
Receiving Str-am
Latitude
Catawba River (McDowell) 35°37'53"
Curtis Creek Dowel l )
UT to Catawba fiver
(McDowell)
Ut to Buck Cr. McDowell)
Catawba River 9c Dowe l l )
Garden Cr. (Mc.; well)
Garden Cr (McD ell)
Town Creek (Mc
UT to N. Muddy ree,
(McDowell)
UT to N. Muddy r
(McDowell)
Garden Creek (M! Dowell )
North Muddy Cre(<
(McDowell)
S. Muddy Cr (Mc.I )well )
UT to Catawba Ri ter
(McDowell)
Youngs Fork (McL)well)
Hicks Branch (Mc.)owell)
UT to Shadrick C-
(McDowell)
UT to Buchanan C.
.
(McDowell)
N. Fork Catawba ?dyer
(McDowell)
N. Fork Catawba '?fiver
(McDowell)
Goose Creek (McDwell)
:,a 11 Timer Co. (,very)
Linville River (: very )
Linville River (Avery)
Mill T� tuber Cr (Avery
35°38'43"
35°40'59"
35°41'27"
35°41'52"
35°41'44.5"
35°41'29"
35°40'48"
35°38'14.5"
Longitude
82°10'18"
82°09'32.5
82°03'55.5
82°04'02"
82°02'50.5
82°01'31.5
82'01'19.5'
82'01'06"
82°01'43"
35°38'52.5" 82°02'07"
35°41'41"
35°38'19.5"
35°39'00.5"
35°42'16.5"
35°39'O1"
35°38'25"
35°42'56"
35°50'52"
35°47'34"
35'47'33"
35'36'38"
36°01'03"
35°00'41.5"
82°01'29.5'
82°01'15"
81°52'35.5'
81 °59' 55"
81°57'27.5'
81°59'25.5"
81 °55' 30"
82°05'17.5"
82`01'09"
82°0;'08"
81 °58' 38"
81`55'21"
81'55'54"
36°02'13.5" 81 °53' 33"
36'01'30" gl°53'41"

Map Design
# Flow Discharger Receiving Stream Latitude Longitude
26 .008 Norris Ind. Stacey Creek (Avery) 36°00'31.5" 81°53'41"
NC0045543
27 .070 G.F. Co. Develop Linville River (Avery) 36°05'15.5" 81°51'55.!
NC0023124
28 .050 Linville Resorts Linville River (Avery) 36°04'17.5" 81°52'22"
NC0039446
29 .001 Eckenrod Apts. NC004771 UT to Garden Cr (McDowell) 35°42'15" 82°O1'32"
30 .024 Columbia Carolina Corp. Catawba River (McDowell) 35°39'38" 82'05'17"
NC0048542
31 .002 Quick As A Wink Car Wash Garden Cr (McDowell) 35°42'00.5" 82°01'33"
NC0047660
32 .050 Duke Power -Bridgewater Catawba River (Burke) 35°44'31.5" 81°50'11"
NC0004936

4
"w` ir,•_c.....-,.:...^* i :f �� `'-S}:�xY��'ti'-srr:.:::..: Wit,.• -• - ..-.- .....w..wr.-� ...xs+ ��. ...�
_. '�;: Nam': °c/�iR�`�':I�iL!. -� k•.4-a.. 1+5:: �.vv,?.,gym':';tc: '«•v,T:'+.r:•�e-F-,cM�...: � :+�:.ra:'�:�- • .-n,.M-�..:.w:��:��w.rx •r-r� r+..Jrr ^.a..,.,.... e:.�x_,,:+>+r•a ,.. ....�..-P:s �.•aR..�••r...�-.... -.. _. _�..

Map
Discharger
1 Valdese WTP
NCOo44369
2 Valdese - Hoyle Cr. WWTP
NC0021199
Receiving Stream
Latitude Longitude
Lake Rhodhiss 35°46'36" 81°33'34"
(Burke)
Hoyle Creek
(Burke)
3 Valdese - Lake Rhodhiss WWTP Lake Rhodhiss
NC0041696 (Burke)
4 Valdese - McGalliard Creek WWTP McGalliard Creek
NC0020753 (Burke)
5 Drexel WTP
NC0043613
Howard Creek
(Burke)
6 Drexel WWTP Howard Creek
NC0021296 (Burke)
35°45'37" 81°32'58"
35°43'36.5" 81°32'35.
35045' 26.5" 81°34' 34"
35046105.5" 81°36'37"
35046'25" 81034158.
7 Baton Elementary UT to Stafford Creek 35°48'29" 81°32'17"
NC0030783 (Caldwell)
8 Gamewell Middle & Elementary
NC0041173 (Caldwell)
9 Warrior Fork WTP
NC00144148
10 Monte Carlo MHP
NC0037346
Warrior Fork
(Burke)
Lower Creek
(Burke)
11 Morganton WWTP Hunting Creek
NC0021431 (Burke)
12 Synthron Inc. Hunting Creek
NC0033120. (Burke)
13 Essex Group Inc.
NC0006645
14 Burke Co. Youth Center
Little Silver Creek
(Burke)
UT to Lower Creek 35°52'10.5" 81°35'55"
35°46'50" 81°42'07"
35049141.5" 81°38'06.
35045 '50"
35°45'36"
35°43'39"
81°39'�43.
81°39'19"
81°44'34.
UT to Fiddlers Run 35°42'19.5" 81°39'43.!
NC0027618 (Burke)
15 Catawba (Morganton) WTP
NC0044156 (Burke)
Catawba River 35°414'20.5" 81°43'48.
16 Mull School UT to Fiddlers Run 35°42'13" 81°39'38"
NC0040193 (Burke)
17 Western Correctional Center Hunting Creek 35°40'42.5" 81°41' 15.`
NC00=7669 (Burke)

Map
Discharger
18 Knob Creek Of Morganton WWTP
NC0034517 ,
19 Broyhill (Lenoir Furniture)
NC0046566 - 005
- 006
20 Broyhill (National Venere)
NC0046558 - 007
- 008
- 009
- 010
21 Broyhill (Occasional #1)
NC0046574 - 011
012
- 013
22 Broyhill (Harper Plant)
NC0006505 - 001
- 002
-004
23 Broyhill - Lenoir Chair # 4
NC0006602
24 Southeastern Adhesives
N00047431
25 District 9 School
NC0044687
26 Lenoir - Lower Cr. - WWTP
N00023981
27 Rospatch Tape
NC0026930
28 Shades Pavilion Rest Home
NC0047147
29 Green Mountain Park:
NC0040274
30 Crossroads Oil Co.
NCOO40835
31 Southern Devices
NC0004278
32 Great Lakes Carbon
NC0005258
Receiving Stream
McGalliard Creek
(Burke)
Town Creek
(Caldwell)
Town Creek
(Caldwell)
Town Creek
(Caldwell)
Town Creek
(Caldwell)
Lower Creek
(Caldwell)
Lower Creek
(Caldwell)
UT to Lower Creek
(Caldwell)
Lower Creek
(Caldwell)
UT to Spainhour Creek
(Caldwell)
UT to Spainhour Creek
(Caldwell)
UT to Zacks Fork
(Caldwell)
Blair Fork
(Caldwell)
Little Silver Cr
(Burke)
Silver Creek
(Burke)
Latitude
35°43'47"
35°54'19"
35°54'16"
35°54'12"
35° 54' 12"
35°54'09.5"
35°54'05"
35°54'03.5
35°53'58.5"
35°53'54"
35°53'55.5"
35054'54"
35°54'50"
35°54'46.5"
35°53'50"
35°53'50"
35°52'32.5"
35°52'36"
35°55'09.5"
35°55'56.5"
Longitude
81°37'27"
81°33'O1"
81°33'02"
81°33'07"
81°33'07„
81°33'07"
81°33'07„
81°33'07„
81°33'10"
81°33'10„
81°33'10"
81°32'41"
81°32'42"
81°32'41"
81°53'17.
81°33'17.
81°34'15.
81°34'47.
81°33'31.
81°33'24"
35°57'25.5" 81°31'O1"
35°56'06"
81°32'42"
35043'33.5" 81044'57"
35°43'41.5" 81°43'37"

Map
fi Discharger
33 Oak Hill School
NC0033243 (Burke)
Receiving Stream
UT to Canoe Creek
Latitude Longitude
35°46►30" 81°45'26,
34 Morganton - Catawba River WWTP Catawba River 35°46'42.5" 80°39'48'
NC0026573 (Burke)
35 Burke Long Term Care Facility UT to Canoe Cr 35°45'11" 81°44'09'
NC0046027 (Burke)
36 Rutherford College Elem. UT to Island Cr. 35°45'12" 81°31'21'
NC0040207 (Burke)
37 Medusa Aggregates Lower Creek 35°55'59.5" 81°39'06'
NC0047139 (Caldwell)
38 Cedar Rock Country Club UT to Lower Creek 35°56'32.5" 81°27'40'
NC0043231 (Caldwell)
39 Lenoir WTP Rhodhiss Lake 35°47'04.5" 81°28'51'
NC0044164. (Caldwell)
40 Hibriten High School UT to Lower Creek 35°55'28.5" 81°29'53'
NC0041149 (Caldwell)
41 Sawmill Elem. School Hayes Mill Run 35°49'09" 81°28'36
NC0041165 (Caldwell)
42 Hammary Furniture Plant #4
(Dellwood Furniture) Hayes Mill Run 35°49'11" 81°28'36
NC0034851 (Caldwel10
43 Duke Power - Rhodhiss Catawba River 35°46'26" 81°26'16
NC0004529 (Caldwell)
44 Glen Alpine School UT to Little Silver 35°43'22.5" 81°46'48'
NC0040541 (Burke)
45 Inmont UT to Little Silver CR 35°43'42" 81°45'21'
NC0041203 (Burke)
46 Collettsville Elem. School UT to Johns River 35°55'37.5" 81°40'26'
NC0041203 (Burke)
47 Richard D. Swanson UT to Lower Creek 35°56'48.5" 81°28'33'
NC0049638 (Caldwell)
48 Monte Carlo MHP
NC0048755
49 Outward Bound School
NC0040754
Lower Creek
(Burke)
•
Roses Creek
(Burke)
35°49'50" 81°37'49'
35053115.5" 81°52'30'

Map
Discharger
50 Cellu Products -- Lenoir
NC0047627
51 Thomasville Furniture
I C0005592
52 Hosiery Mfg. Co
NC0024341
Receiving Stream
Blair Fork
(Caldwell)
•
Blair Fork
(Caldwell)
Latitude Longitude
35°56'39" 81°34'09,
35'55'56.5" 81°32'35'
UT to Hunting Cr_ 35°45'12" 81°40'47.
(Burke)

Stream Stations - Lake Rhodhiss Study
Station # Stream Location
S-1 Muddy Creek Hwy. 70
S-2 Silver Creek (CTB029) Hwy 70W
S-3 Warrior Fork SR 1440
S-4 Johns River (CTB031-D1) * SR 1438
S-5 Lower Creek (CTB034-1A) * SR 1501
S-6 Catawba River (CTB028A) * SR 1147
S-7 Catawba River NC 18
* Historical Ambient monitoring data available
Parametric Coverage at Stream Stations
Sampling was preformed from bridges at all stations. Each station
was sampled 7 times between 0845 on 82/07/28 and 0710 on 7/29/82. Grab
samples were taken for N an P, chlorophyll a, pheophytin, and phyto-
plankton. Also taken at stream stations were dissolved oxygen, pH,
temperature and conductivity.

r ,
Lake Stations - Lake Rhodhiss Study
Station # Location
L1 at SR 1501
L2 near mouth of Smoky Creek
L3 by island just below mouth of Howard Cr.
L4 at SR 1001
L5 by boat ramp just beyond mouth of Freeman Cr.
L6 where lake turns left 5000 ft. upstream from Dam
L7 just above Rhodhiss Dam
Parametric Coverage at Lake Stations
Sampling was preformed at the center of the channel. Secchi
disk depth and dissolved oxygen were taken at each station in order
to determine the euphotic zone and the hypolimnion. An integrated sample
was taken from the euphotic zone for analysis of N and P series and
chlorophyll a. A sample for N and P series was also taken from the
hypolimnion. Conductivity temperature and pH were also taken throughout
the water column.

Major * Dischargers in Lake Rhodhiss Drainage Area
Map 11 Discharger Design Flow
1 Morganton - Catawba River WWTP NC0026573 8.0 MGD
2 Valdese - Lake Rhodhiss WWTP NC00111696 7.5 MGD
3 Lenoir - Lower Creek WWTP NC0023981 6.0 MGD
4 Valdese -McGalliard Creek WWTP NC0020753 3.2 MGD
5 Marion - Corpening Creek WWTP NC0031879 3.0 MGD
6 Great Lakes Carbon - Dilver Creek NC0005258 2.6 MGD
7 Valdese - Hoyle Creek WWTP NC0021199 1.0 MGD
•8 Drexel - Howard Creek WWTP NC0021296 0.5 MGD
* This list includes all dischargers with a design
flow equal to or greater than 0.5 MGD.
NOTE: Valdese - Lake Rhodhiss WWTP is not in
operation and is presently under construction.
Parametric Coverage at Major Dischargers
Composit and grab samples for N and P were taken from the effluents
of these WWTPs. Marion - Corpening Creek WWTP was composit sampled at
four sites: Influent (before P removal), Effluent (after P removal),
Upstream outfall (Corpening Creek), Downstream outfall (Corpening Creek).
Morganton - Catawba River WWTP, Lenoir - Lower Creek WWTP, Great Lakes
Carbon WWTP and Valdese - McGalliard Creek WWTP were composit sampled.
Valdese - Hoyle Creek WWTP and the Drexel - Howard Creek WWTP were grab
sampled three times in a 24 hrs period.

N
Area Schematic

Algal Growth Potential Test (AGPT)
The Environmental Protection Agency, Region IV, Surveillance
Analysis Division, Athens, Georgia, is and
g� � presently in the process of running
an Algal Growth Potential Test (AGPT) on the waters of Lake R
nutrients and sunlight are abundant, algae are capableRhodhiss. When
multiplication often resulting in serious water of rapid growth and
quality problems. By
introduction of a controlled laboratory stock algae (Selenastrum actual lake water in question, the capricornatum)
q potential of waters, sewage and in-
dustrial effluents to accelerate or inhibit algal growth can
As indicated in a letter from Mr. Roanld Raschke be determined.
Branch, Athens, Ga.)early (EPA, Ecological support
in July an AGPT could be easily integrated into
the Lake Rhodhiss study. Mr. Raschke stated that an AGPT on
waters could provide the Division of EnvironmentalLake Rhodhiss
minimum, limiting nutrient status of the Management with, at
lake. Additionally, he stated that
the AGPT results could assist in refining the predictive capabilities
model by providing information on available phosphorusof the
meaningful than total phosphorus estimates p p orus, which could be more
of sampled waters with P-media and only. By incorporating dilutions
conducting a limiting nutrient assessment,
EPA will provide information useful to DEM in these ways:
1) predicting in -lake concentration reductions of
whether it is the limiting nutrient or not phosphorus
2) determining available phosphorus loading, '
3) determining phtoplankton nuisance
p potential for Lake Rhodhiss,
4) determing WWTP influence upon stream and lake nutrient
5) determining available phosphorus loads with regimes,
and without
phosphorus removal at the Marion - Corpening Creek WWTP.
Composite samples were taken as requested by Mr. Don Schultz
Support Branch, Athens, Ga.) (EPA,
Ecological
who worked on -site in conjuration with
the 'technical Services field personnel on July 28 and 29
from the Rhodhiss Lake stations, stream stations, �� Samples were taken
WWTP influent and effluent, as well as upstreamMarlo.. - Corpening Creek
and downstream locations.
Because AGPT tests take several
weeks to run and the EPA Athens lab has a heavy
work load, no date for the completion of the report has been
will n set. However, the
AGPT report from EPA
1 be attached to this report as 2r addendum upon
completion.

tit
N
co
rn
r-i
4.3
ro
L-1. Comp 7-28
bot.
Nitrogen, Phosphorus, Solids, and Chlorophyll Concentrations
at Lake Rhodhiss Stations
r--1tto r-1
E
1300 .15 .4
.05
r-f
E E
•
Z O
.a
m4-4
Q..,
O
as
N �t
z
.36 .08
.22 .08
L-2 comp 7-28 1237 .12 .6
bot NS NS
L-3 Comp 7-28 1230 .09 .6
bot .06 .4
Comp 7-28 1032 .06 .6
bot .07 1.1
1,-5 Comp 7-28 1007 .08
hot .08
L_6 romp 7-28 0920 .13
bot .20
L-7 Comp
bot
7-28
0900 .13 .6
.17 .06
NS NS
.13 .06
.22 .08
.10 .05
.23 .68
.10 .04
.20 .04
1.3 1.6
.18 .04
.19 .09
.18 .04
.01 .08
a)
r-I
4-) 'Ly
o •,-I
CL
r--I
E
E bO
a
0)
U
3 11 3
24
tto
Pheophytin a
21
40
35
9 35 29
.27 25
25 21 5
<1 24 21 4
Avg = 22 ug/1

0
cd
0
O
N
CO
rn
,-i
Hwy70 S-1
7-28 1110
Table
Temperature
22
Stream Station Data Lake Rhodhiss Study
E U
r--1 �� n
O CV
E
z M e
•
0
O
U
7.8 6.7 110
7-28 1320
7-28 1545
7-28 1800
7-28 2312
7-29 0100
23
23
23
7-29 0710
ativer C Hwy 70
S-2
M
r-1
E E
•
0
r n cL
0
z
ON
E0•
.01 .2 .31 .06
0
.02
8.1 7.3 118 .01 .3 .27 .04 .01
7.8 7.1 115
7.6 6.8 110 .02 .3 .28 .08 .03
.03 .3 .32 .07 K.01
7-28 0830 21 7.8 6.6 55
.04 .4 .30 .09 .02
.35 .13 .02
.27 .08 .01
.49 .05 .01
7-28 1140 22 7.8 7.0 50 .01 . 28 .01
7-28 1400 23.5 7.7 7.1 55 .01 .2
.30 .04 .4 . of
7-28 1615 23.5 7.5 6.9 53 .01 2 .26 .01 .04
7-28 2111
7-29 0230
7-29 0600
2
.06 1.0 .30 .06 .01
.01 .2 .33 .04 „...01
.01 .2 .33 . .05 _ .01
r--t
bO
Pheophytin a
1

.^-.. R....:.:.� 'J -,ti1n: ^s;C +� �• .:y.: �'. S .e:a.ne: .. ` "�• .� rv. �:-a.w.r, 'y}g.,;'
•s'q' ¢-,n; :._, r�-� :i+r� �' .:I -•.a c .�. r : rs rr • �4 art?%.?''•"t�:^ ` 'a` R 'K.,°y":+S
Table Cont. Stream Station Data Lake Rhodhiss Study
E
L
4.3
�
Location1
Station No.
CVe--1
CO
.0
nM
E
E.
O
1
d)
i,
Z
M
E
H
r-1
-Co
E
0
ca
a.
umhos/cm
Spec. Cond. at 25°C
b0
E
M
z
,-I
DO
E
Z
H
r--1
E
`�
1 M
CV.�
2
f-1
E
0
QQ.r
H
r-1
d)
.0
a
L.
rl
`
L
F
c�
r-1
r-1
`
S..
L
0
C..)1)
n1
r-i
r-i
m
C
•1-1
a
N
wThrrior For. k
MO
1 S-3
7-28
1015
►22.5
7.6
6.7
30
.01
.1
.18
,03
<.01.
2
1
< 1
7-28
1245
23
7.9
6.8
30
<,01
.2
.16
<.01
<.03
7-28
1515
23
7.8
7.0
30
.01
<.1
.16
.02
<.01
1
<.1
< 1
7-28
1730 ,24
7.6
6.8
30
.01
.1
.16
.02
<.01
_
7-28
2229
.02
.2
.15
.03
<.01
_
7-29
0140
.03
.2
.19
.03
<.01
7-29
0645
i
1
.01
.2
.17
.03
<.01
iohn R _ ;
1438
S-4
7-28
1000
23.5
7.6
6.7
30
.01
.2
.07
.03
<.01
2
2
.41
7-28
1230
24
8.0
6.9
30
<.01
.2
.07
.03
<.01
7-28
1500
24.5
8.2
7.1
30
.01
.2
.08
.02
<.01
7-28
1720
24.5
8.5
6.9
30
.01
.2
.07
.02
(.01
7-28
2215
.02
.1
.08
.03
<.01
•
7-29
0150
,
.01
.1
.07
.02
Z.01
7-29
0630
.01
.2
.08
..03
.01
1
_
_

MILOW
451)111WriallaStin
Table Cont. Stream Station Data Lake Rhodhiss Study
Mo
o
S.
rn
L
.,�
N
0
.1
.Station No.
N
Co
Cr,
n)
4-)
C)
a)
E
E-.
Temperature °C
.--1
E
0
Q
O.
Spec. Cond, umoh/om
at 25 C
r-i
I'D
z
1
M
z
.--I
z
1
Z
E-1
r-i
-.
E
z
orn
7
+
ON z
Total Phos. mg/1
r-I
E
00
-
a.
0
.s~
O
r-1
E
•.
E-+
M
.--1
U
r-I
E
it
0
U
M
r-I
U
r
r4
-
W
cis
1
.1.)
a
o
W
a
•
natacaba R .
SR
1147
S-6
7-28
1045
16
8.4
6.3
50
< .01
.1
.27
.02
< .01
1
1
.. 1
7-28
1300
20
8.6
7.5
70
.01
.2
.28
.02
•
7-28
1530
17
7.0
6.8
50
4.01
.2
.29
.02
.(.01
3
3
< 1
7-28
1745
16
6.5
6.8
48
.01
.2
.28
.01
s
7-28
2252
.02
.2
..30
.02
•
7-29
0120
.02
.2
.31
.01
.01
7-29
f0700
.01
.1
.26
.02
. .01
.
atawba R
NC18
S-7
7-28
0915
19
7.7
6.7
50
.01
.2
.29
.02
<.01
2
1
1
7-28
1150
20
7.8
7.0
50
..c.01
.3
.27
.02
.01
7-28
1415
22
8.0
7.2
48
.01
.2
.20
.03
<.01
1
.< 1
<-1
7-28
1635
19
8.6
7.0
50
.01
.2
.27
.03
<.01
7-28
2131
.28
.6
.52
.03
<.01
L .01
7-29
0215
.01
.2
.08
.03
7-29
0610
.01
.2
.24
.02
4 .01
i
�
-r

tiggfirewlag ru •ems-•. r.wu;• . ,i,_ "+= erg:.(• „jy.ai ^i -. �,y..,, s5 -;►. . �:.
�,r
i. �.�ra.�Crae. r;.s•. ,, �:' �. i .1v?r'..,La� +t}; wr,n►
'+ Sod: - • ,vg
Stream Station Data Lake Rhodhiss Study
E
(0w
i.
10
Location1
Station No.
(NI
co
ON
w
.0
11
w
E
'r'1
F.
Temperature oC
r--i
E
O
•
Q
."L'.
0.
Spec. Cond, umoh/cm
at 25 C
r-^I
hO
E
z
on
.�I".
Z
r-i
@
z
ae
E--r
r-1
dO
E
=
M
O
+
N
O
Z
r-1
h0
E
.
0
c
m
o
H
r-1
-
bO
E
0
ai
.c
4
�..
0
r--1
Gp
•�
L.
..
N
r--1
.z
0
r--I
4O
L.
0
N
r-1
.S�
U
.-I
ha
N
•
.0
0
W
.>r
a
Lower Cr.
1501
S-5
7-28
0930
22
7.0
6.8
80
.05
.5
.90
.13
.05
2
< 1
2
7-28
1215
24
6.8
6.9
80
.04
.5
.87
.13
.05
7-28
1430
26
6.6
6.9
80
.05
.5
.90
.13
.04
2
2
<1
7-28
1650
26
6.3
6.9
80
.06
.6
.51
.14
.06
7-28
2145
.11
.8
1.1
.18
.08
7-29
0200
.05
.7
1.1
.06
.06
7-29
0620
.07
.8
1.1
.16
.07
-
f-
_
z

-ti.• . ..i`ra-. •i. M'_'„ j; • s..-cS:: ";:-7 _v. r, r •i. 4 i i'7rr i�i-.`:�•a�rt�� i'^h=v 'Lr F.r tiG �a r.:. y•(.1; • x
'�' szF�i"i'.. rf4S •-'may `'i:= �+_.•�^'�.i � •,, . � �'.. ieii r!'ir=ty✓i"':S.'=�T.;..i..
Nitrogen and Phosphorus Values in Influent and Effluent of
the City of Marion Corpening Creek WWTP
C C
•r-f .,-1
GO bO 'C3
W O C
m aq w
a) a) a)
4.3 E 4-)
aS •i-i N
0 (-' Q
7-27 1635 7-27 2000
7-27 2100 7-27 2400
7-28 0100 7-28 0400
7-28 0800 7-28 1800
7-28 1900 7-28 2300
7-28 2400 7-29 0400
7-29 0500 7-29 0900
7-28 0900 7-28 1800
7-28 1900 7-28 2300
7-28 2400 7-29 0400
7-29
<.05
.05
.06
9.2 -c.05 9.6 6.6
14. <.05 28. 17.
1.3 4.9
1.3 4.2
1.3 4.3
1.2 3.6
.75 .65
.73 .63

Upstream
,. ...,..�r......�..._.�o.,. ...�,.....�.��...-�., a.�aw -�.v. u.e ...,.sue .,,,rm �rs�.r�er�e.+-.r+�tixmwirrwea��� +f-ei. nna'a rEP- rariaa cem- Anff *yry r...GR[,,MITOMMAIK. .RFetiIIV
Nitrogen and Phosphorus Concentrations Upstream and Downstream of
the City of Marion Corpening Creek WWTP 0utfall
1500 7-27
2100 7-28
0300 7-28
0900 7-28
7-28 1900 7-28 24400
Downstream 7-27 1605
7-27 L2010
7-28
7-29
7-29
0010
0510 7-28
0900 7-28
1030
1035
E
.06
4.05
4.05
1
E
O
L
O
.8 1.6 .40 .21
1.3 .22 .17
.05 3.5 1.1 .22 .15
.9 .6 .30 .25

• r:. '.^'_'c`_ .:.. .may. �a ... r-sti rC�Ys b.;�.`;. wF•.[:.. _:::
•
Nitrogen and Phosphorus Concentrations in Effluents from
Wastewater Treatment Plants of Morganton, Lenoir, and Valdese
L.
o
•4
(0
,a
Date Begin
c
,1
b
a)
m
w
E
•,-i
Date End
Time End
E
Z
1
rl
=
e
Z
1
Z
ae
E
Z
1
M
O
2
N
O
Total Phosphorus mg/1
Ortho Phosphorus mg/1
Morganton
7-28
1230
7-28
1730
3.2
13.
G•05
5.9
3.5
cn Lriwhn R .
7-28
1830
7-28
2330
7.8
14.
.05
8.3
4.9
...
dTP
7-29
0030
7-29
0530
6.8
22.
{.05
8.8
5.2
-�-
Fffluent
7-29
0630
7-29
1130
6.7
20.
<.05
8.2
4.9
Lenoir
7-28
1345
7-28
1845
7.4
14.
6.3
3.2
2.7
i.ewer Cr.
7-28
1945
7-29
0045
7.7
12.
6.8
3.2
2.5
WWTP
7-29
0145
7-29
0645
7.5
14.
6.7
3.4
2.6
Effluent ;
7-29
0745
7-29
1145
6.5
12.
6.5
3.4
2.7
Valdese
7-28
1100
7-29
1300
4.1
12.
C.05
15.
12.
r•icGa 1 ]. iard
7-28
2200
3.6
12.
.07
17.
14.
:r . WWTP
7-29
1400
3.3
7.6
..05
11.
8.5
.
!affluent
7-28
1515
4.0
11.
< .05
32.
15.
Valdese
7-28
1530
1.6
4.8
.06
7.8
5.9
Hoyle Cr.
7-28
2230
1.2
4.2
.07
6.2
4.8-
____________
.1WTP ____
7-29
1330
2.9
18.
.07
9.8
4.7_�-
---
-

Nitrogen and Phosphorus Concentrations in Effluents from
Drexel WWTP and Great Lakes Carbon
c:
o
• ,+
y_)
(13
1-)
(f)
c
•I
60
01)
cr_t
2
M
a)
C
-+
W
a)
in
f E
•r{
H
Date End
Time End
Z
'M
`...T.'
Z
r1
Z
z
M
F.
r1
z
1
cn
0
Z
N
0
Z
,-1
to
0
.0
a.
r-I
•0
0
H
.-1
fA
0
.0
a.
O
L.
0
Drexel
7-28
1430
17.
20.
1.4
5.7
4.4
Howard rd c r .
7-28
2215
15.
28.
1.2
6.0
4.4
WWII
7
7-29
1320
21.
35.
1.8
5.5
4.1
Effluent
treat Lake
7-28
1130
7-28
1630
<.05
.1
.31
4.05
c.05
Carbon
7-28
1730
7-28
2330
<.05
.1
.29
.05
.05
1•:ffluent
7-29
0030
7-29
0530
4.05
<.1
.30
<.05
<.05
;.
7-29
0630
7-29
1030
4.05
.c.1
.31
<.05
c.05
y
l
i
I

,
T .
LAKE STATIONS
SHEET
FIELD DATA
Station Number
Date
Time
DM
Depth
300
D. 0.
10
Temp
400
pH
94
Conduct.
77
Secchi
CTB-040A
7128182
1032
'
5
11.6
27.4
8.4
77
44"
L-4
7128,82
1
11.3
27.3
8.5
77
7128182
1 25
2t; 9
_
7128182
1.5
26.E
7128i 82
2
8.6
25.5
7.2
69
7128182
3
5.0
24.2
6.5
69
7128182
4
4.1
23.4
_
6.2
69
7128182
5
4.4
22.7
_
1
614
7 121; 182
_
6
5.5
22.2
6.0
61
7 128 182
7
6.1
21.9
'
• 6.0
'
61
7 128 182
8
5.2
21.6
6.0
61
7 128 ,82
9
4.8
21.1
5.9
61
7 128182
9.5
:
4_.5.
20.4
_
5.9 ,
62
,
,
_
CTB-0361A
7 128182
1230
5
'
11.5
28.5
8.5
73
36"
L-3
1 A
1
27.5
,
8.5
70
1 1
4
1.25
,10.9
26.3
1 1
1.50
25.5
1 i
r
i I
2
7.7
24.9
,
7.0
69
t 1
3
6.9
24.4
6.9
65
t 1
4
5.5
6.3
67
1 t
5
_23.0
5.7
22.5
6.0
65
1 1
6
5.5
22.4
6 0
65
I
6.5
4.9
22.0
5.5
65
t 1
-
CTB-034A
7 ,28 182
1300
5
-
7.3
24.1
6.3
73
20"
c-1
1 1
1
6.9
23.0
6.1
74
1 1
1.25
6.5
22.5
6.1
74
t
CTB-040B
7128'82
0900
5
11.7
28.7
8.4
71
41"
L-7
t 1
1
11.3
28.9
8.4
72
•
1 1
2
10.2
28.5
8.4
69
t 1
2.25
27.1
1 I
2 5
25.2
I•' 1
•
3
3.0
23.5
6.5
60
1 1
t1
3 4.
22 6
6 2.
60
1 1
5
3.5
21.7
5.9
66
1 I
6
3.2
21.3
6.0
67
1 A
7
3.4
21.0
6.0
69
t i
,
R
q 9
20.9
y
6.0
65
_
0
3 3
20..
6.0
67
1 t
10
3.";20.5
6.0
64
115
0
17.3
D.2
95

7• '
•
Lake Stations (Continued)
Station Number
Date
Time
DM
Depth
300
D. 0.
10
Temp
400
pH
94
Conduct.
77
Secchi
L-2
7 t28t82
1237
5
11.0
28.5
8.4
73
32"
0.25
27_5
t t
.5
_
27.2
1 t
1.
9.1
2.6 . 0
7 7
fi 9
1.25
24.3
t
1.5
24.4
•t
2
6.5
23.0
6.4
69
t t
,
3
6.3
_
23.0
6.1
69
3.5
6.1
22.7 ,
6,0
69
t t
1250
Bottom
_
L-6
7 28 t82
0920
5
11.5
28.7
8.5
,
72
43"
I !
1
11.1
28.7
8.5
72
! t
1.5
28.5
t t
1.75
28.3
i 1
2
10.9
27.3 .
8.2
68
t
3
4.3
24.3
6.6
_
64
t t
4
3.3
22.2
5.8
64
1 t
5
3.7
21.6
5.8
66
6
3.8
21.4
5.8
66
t
7
4.0
21.2
5.8
64
t t
8
4.6
21.0
5.8
62
9
3.8
20.9
5.8
,
66
!
10
3.9
,
20.7
5.8
63
t t
11.25
, 3.8
20.2
5.8
,
64
0955
Bottom
t,
L-7
7 28[ 82
1007
5
11.6
28.0
8,5
74
, 41"
1
11.3
27.9
8.6
75
` 2
9.9
27.4
8.4
71
`
I !
2.5
26.4
! 1
_
3
5.9
24.7
6.8
64
t
4
3.6
23.0 ,
6.8
66
5
3.8
22.3
6.0
65
1 t
6
4.3
21.9
6.0
64
7
5.2
21.7
6.0
60
j
8
5.1
21.5
6.0
60
t_
9
,
4.9,
21.1
5.9 .
, 60
t t
'10
4.6
20.7
5.8
62
t t
_
10.5
4.8
20.4
5.9
61
1 I
1020
Bottom
•
1 I

.ti
Month
Marion - Corpening Creek Wastewater Treatment Plant
Effluent
Total P mg/1
1981 Self Monitoring Data
NH3-N mg/1
Upstream Downstream
Total P mg/1 Total P mg/1
January 3.65 1.0 - -
February 2.5 .94 .087 .261
March 1.0 .35 .033 .369
April 1.6 2.8 .041 .256
May 2.3 2.9 .095 .571
June 1.3 .46 .100 .200
July 1.22 .20 .143 .291
August 6.9 .47 .099 .523
September 3.09 .70 .060 .920
* October 7.60 .20 .040 2.0
* November 6.87 .39 .041 1.7
* December 9.50 .46 .091 1.1
Average 3.96
Maximum 9.50
Minimum 1.00
.91
2.90
.20
.075 .74
.143 2.0
.033 .20
* Note: During October, November, and December Marion's Lime feed equipment
was out of order resulting in higher than normal phosphorus readings.

.+.- — ....y.' _. 's. a`.•.7 is , . :f,s _, a.�;'.. .�'a _ . •[:. ..:r; ° ....'� :Y ;,�._ r ,'r: �'_ �:r:l .nr • .y.} a 1: .,a Fes':' ��r.�aiti :ti,`: ..
•
Nutrient data from 24 hour composite sample at selected dischargers into
Lake Rhodhiss drainage area. Samples taken by regional field office.
Discharger
Valdese - McGalliard Creek WWTP
NC0020753
Drexel - Howard Creek WWTP
NC0021296
Date Time NH3 TKN NO2-NO3 PO4 Total P
81/12/10 1100 .43 9.3 1.2 15 15
81/12/11 1100
81/1/10 1000 17 26 1.3 4.8 5.0
81/1/11 1000
Valdese - Hoyle Creek WWTP 81/12/10 1200 2.5 8.3 .19 5.1 5.1
NC0021199 81/12/10 1200
Western Correctional Center 82/02/09 1200 1.5 5.1 6.1 2.4 2.6
N00027669103 WWTP 82/02/10 1100
Great Lakes Carbon - Silver Creek 82/02/08 1100 < .05 .1 .42 .07 .78
NC0005258 WWTP 82/02/09 1100
Morganton - Catawba River WWTP
NC0026573
Marion - corpening Creek WWTP
NC0031879
82/02/08 1030 13
82/02/09 1030
19 .47 2.3 3.4
82/02/01 0800
82/02/02 0800 .49 2.7 7.7 .36 .54

Morganton
Catawba
Valdese
McGalliard
Lenoir
Lower Cr.
Valdese
r _ _ — —
Moyle
Drexel
Howard
Marion
Corpening
;rent Lake._,
Carbon
p stream
Corpening
Downstream
Corpening
'Marion
.:/o lime
±.reatment
.5
3.44
6
7.9
.06
.03
Phosphorus and Nitrogen Loadings in Lake Rhodhiss Drainage
g Basin
4.48 .30 � .22 8.2
1.6 15 15 200 200

v•
17. Limekiln Creek at SR-1556
McDowell CTB-0I1 Secondary
18. Limekiln Creek at SR-1559
McDowell CTB-012 Secondary
19. North Fork Catawba River near SR-1559 McDowell CTB-013 North Fork Catawba River at SR-1552 3 Secondary
McDowell CTB-013A Secondar
21. Lake James 1 mile below N.F. CatawbaY
McDowell CTB-013B Secondar
22. Forsyth Creek at US-70 Y
McDowell CTB-0I4 Secondary
23. Forsyth Creek at End of SR-1525
McDowell CTB-015 Secondary
24. Lake James near Nebo
McDowell CTB-015A Secondary
11. Catawba River at SR-150I
McDowell CTB-008A Secondary
12. Catawba River at Clinchfield RR
McDowell CTB-008B Secondary
13. North Fork Catawba River at SR-1560
McDowell CTB-009 Secondary
14• Armstrong Creek at NC226A
McDowell CTB-009A Secondary
7. Jake Creek at SR-1214
McDowell CTB-006 Secondary
8. Jake Creek near SR-1214
McDowell CTB-007 Secondary
3. Catawba River near SR-1234
McDowell CTB-002A Secondary
4. Curtis Creek at US-70
McDowell CTB-003 Secondary
Historical Water Quality Station Locations in the Lake Rhodhiss Drainage Ares
Station Location
County Station # Station Type
1. Catawba River at SR-1103
McDowell 02136500 Primary
CTB-001
Secondary
2. Ut to Catawba River at SR-1235
McDowell CTB-002 Secondary
5. Curtis Creek at I-40
McDowell CTB-004 Secondary
6. Catawba River at I-40
McDowell CTB-005 Secondary
9. Catawba River at SR-1221
McDowell CTB-008 Secondary
10. Catawba River at NC-226
McDowell 02138000 USGS & Primary
15. Armstrong Creek at NC-226
McDowell CTB-009B Secondary
16. Armstrong Creek at SR-1556
McDowell CTB-010 Secondary
25. Lake James at NC-126
Burke CTB-015B Secondary
26. Lake James 1.8 miles below NC-126
Burke CTB-015C Secondary
27. Linville River at NC-105
Avery 02138288 Primary
CTB-016 Secondary
28. Linville River at SR-1349
Avery CTB-017 Secondary
29. Linville River at US-221
Avery CTB-018 Secondary
3.0. Stacey Creek at NC-181
Avery CTB-019 Secondary
31. Stacey Creek below Hanes Knitting
Avery CTB-020 Secondary
34. Linville River at SR-15^r
3 every
' 'C - B-023 Secondary
32. Mil' Timber Creek at SR-1524
Avery CTB-021 Secondary
33. Mill Timber Creek at Mouth •. Secondary
Avery
CTB-022 Secondary

Station Location County Station # Station Type
3,5. sake James .5 miles up Linville ARm Burke CTB-023B Secondary
36. North Muddy Creek at SR-1796 McDowell CTB-024 Secondary
37. Corpening Creek at SR-1819 McDowell CTB-025 Secondary
38. Corpening Creek at SR-1794 McDowell CTB-026 Secondary
39. North Muddy Creek at NC-226 McDowell CTB-027 Secondary
40. North Muddy Creek at SR-1747 McDowell CTB-028 Secondary
41. Catawba River at SR-1147 Burke 021390036 Primary
42. Silver Creek at US-70 Burke CTB-029 Secondary
43. Catawba River .5 miles above Warrior Cr. Burke CTB-029A Secondary
44. Catawba River at NC-18 Burke CTB-029B Secondary
45. Catawba River at NC-181 Burke 02139282 Primary
46. Hunting Creek at SR-1512 Burke CTB-030 Secondary
47. Hunting Creek off SR -1512 Burke CTB-031 Secondary
48. Wilson Creek at SR-1514 Avery 02140304 Primary
49. Johns River near SR-1356 Caldwell CTB-031A Secondary
50. Johns River at SR-1356 Caldwell CTB-031B Secondary
51. Mulberry Creek at SR-1310 Caldwell CTB-031C Secondary
52. Mulberry Creek at SR-1338 Caldwell CTB-031D Secondry
53. Zacks Fork Creek at US-321A Caldwell CTB-031E Secondary
54. Lower Creek at US-321A Caldwell 02141148 Primary
CTB-032 Secondary
55. Blair Fork at US-321A Caldwell CTB-032A Secondary
56. Blair Fork at Mouth Caldwell CTB-032B Secondary
57. Spainhour Creek at NC-18 Byp. Caldwell CTB-032C Secondary
58. Lower Creek at SR-1188 Caldwell CTB-033 Secondary
59. Lower Creek at SR-1143 Caldwell CTB-034 Secondary
60. Rhodhiss Lake at SR-1501 Burke CTB-034A Secondary
61. ?ropst Creek at SR-1536 Burke CTB-035 Secondary
62. Prcpst Creek at SR-1512 Burke CTB-036 Secondary
63. ::::edh:ss Lake off SR-154/2 Burke CTB-036A Secondary
64. m.c aI'1 _are .'reek at US-64 i 70
Burke CTB-037 Secondary
65. X:....- _ : arc' '2reer: at S RR-15 38 Burke CTB-038 Secondary
66. ''c• ••::: a rd Cove cif Rhodhiss Lake Burke CTB-038A Secondary
b7. ii; .:, 'reek ,.1; rtrea:7: S ;-, 546
Burke CTB-039 Secondary
Burke CTB-040 Secondary
69. : :r_., .-'; `'i Burke CTB-040A Secondary
Burke C,B-040B Secondary

Depth and Channel Mapping
Lake Rhodhiss
A reliable estimate of average lake depth is essential for use
in the loading models assessing the effects of the Marion - Corpening
Creek WWTP on the water quality of Lake Rhodhiss.
On July 28, 1982, the Technical Services branch of the Division of
Environmental Management conducted a series of depth measurements at
predetermined stations on Lake Rhodhiss. A president Model 410 recording
depth meter was used to obtain the readings with accuracy being confirmed
by manual sounding to + or - 6 inches. A complete depth profile was
taken along the center channel of the lake from the headwaters at Morganton
WWTP to the Rhodhiss Dam. By using preset anchored buoys as well as existing
lake stations marked by land refrence points, accurate distances between
stations were assured. Time, depth scale, and station # were recorded
directly onto the strip chart. Con tin ous depth measurements were also
taken at twelve transects located at various points along the length of
the lake. While taking all measurements a constant boat speed was maintained
by keeping a steady tachometer reading. By assuming the boat was traveling
at a constant speed and by knowing the distance between stations, the average
depths along the profiel and `ransects could be determined. The average
depth of the lake was calculated by distance weighting the average depths
between stations. The average depth of the lake (from station B to the
Rhodhiss Dam)was determined to be 20.6 feet.

Avg. Depth of Depth Center
Distance Center Channel Avg. Depth Channel of Average Depth
Station Between Stations Between Stations of Transect Transect Between Stations
(Ft) (Ft) (Ft) (Ft) (Ft)
A - -
9605 4.2
A-1 • 6.4 7.3
2802 5.2 4.6
R 7.9 9.0
4002 8.2 7.1
B 5.2 6.1
2802 7.4
C - - 7.1
3002 8.1
L-1 6.6 6.9
3002 7.3 6.5
Marker 5 7.6 9.2
2601 7.3
E 7.1
2001 14.1
L-2 5.8 10.6
5803 13.1 •
F - - 8.6
1901 17.2
Marker 3 1 4 . 8 21.9
5003 15.6
- - 12.9
2802 14.1
Marker 2 21.1 20.3
4202. 26.8 22.7
Marker 1 27.4 32.4
4002 31.9
L-4 - - 28.2
6003 33.9
Marker 6 31.7 36.9
5003 37.5 31
Marker 7 29.3 36.9
7204 35.1 30.7
Marker 8 32. 46
5203 44.1 No depth transects (est..)
r
a: ter marker

Avg. Depth of Depth Center
Distance Center Channel Avg. Depth Channel of Average Depth
Station Between Stations Between Stations of Transect Transect Between Statios
(Ft) (Ft) (Ft) (Ft) (Ft)
Marker 9
3002 32 (est.)
Marker 10 4+6.8
2401 From Marker 8 32 (est.)
to the dam

.. - .. .S.4* 77 -.. "L-�'...,:a • r.J,.�?E: ''.h`. 9:+I'i'C+rosL.r.:+°I=t*..'i..L•.1,e,5r.:"a:r•:Y1+c.*.•r;.:_r•,1'
Al 1(.' • R
Ic •i/
,
I
1-
• 1•
••••••••),„••••••, • • , . \
•� 1,
-1
(h
IV
)
rAI1_11c F
•
31
L'71ij11 ` \ I.
DEPTH AND CHANNEL MAPPING STATIONS
LAKE r'! !n^'.r-
MAPPING STATION ® MAPPING STATION WITH TRANSECT �$
U
( i j T'� 1
vAa
•/ • 1.,
.934 • V
r: ::= r. r •<
1 a•.) 1
�a 1 •../ir{
I', hVv
r t r) y r LS r r t,
r rl7�)
76
' l ♦
•
1911 /
j 71
1911
UN T Y
f•
•�. L4
LAKE
•' RUTIIERVORO \,
COUrcr
trig ,/
•
•
' •
•S
ifcroon
:`
�Jt
9
111N Miff \
FIEY. 1000\�4
/ A' I
r� ` AI•er'1 Chapel Ch 4,40
Above
r9ElK UM I791-1':`,$.`
1i
1 4,1 I
)
• •r1A
1t'•
Ine�
"" ram,
•
•z, 1.. /
• ,
I••4r 1•
r.h•11,n1
rro••
1,
011OI1HIr,S
POI 77!
alf,wlll +27
IU9U ln:
10/1
'y
.\a
74 f 114111(.rf
r,1r 1,1140
11/4111(4C) •9 Y
t :o,
40)
1
7.
'14 1$ •'
• Ifirl"•: 1111 POP),
A1•n:•.ipu1 r.11' 71• 'S •
ION.: VIEW
►U1 : 14 •
(CAtAwDA l 111 1
6VRKf 11^
N

ENVIRONMENTAL PROTECTION AGENCY
REGION IV
SURVEILLANCE AND ANALYSIS DIVISION
ATHENS, GEORGIA 3011oIIx 30613
OCT 2 2 1982
REF: 4ES-EC
Mr. Steve Tedder
North Carolina Department of Natural Resources
and Community Development
Division of Environmental Management
P. 0. Box 27687
Raleigh, North Carolina 27611
Dear Steve:
I discussed the Rhodhiss Lake problem with Russ Todd, and because
of the urgency of the situation, he agreed that I should forward
the AGPT data with our preliminary assessment to you immediately.
We do have a draft report prepared, and it will soon follow after
undergoing peer review. Our assessment is as follows:
o The waters of Muddy Creek, Catawba River and Lake Rhodhiss
are sensitive to additions of bioavailable phosphorus
(Table 3).
o Potential mean maximum standing crop (MSC) ranged from
0.57 mg dry weight per liter at the Johns River Station
S-4 to 41.61 mg dry weight per liter at the Lower Creek
Station S-5 (Table 4). In my opinion, potential dry
weights of 5 mg/L or less are not unreasonable. This
concentration of 5 mg/L translates into 11.6 ug of bio-
available phosphorus per liter (Table 4).
o The Marion-Corpening WWTP effluent seems to have only a
moderate enriching effect upon its receiving stream, Muddy
Creek, and the Catawba River as evidenced by the low algal
growth potentials of 3.41 mg dry weight per liter at Station
S-1 and 0.68 mg dry weight per liter at Station S-6 further
downstream.
o Except for the lack of phosphate removal capability, the
Great Lakes Corbon-Silver-Creek (GLCSC) WWTP has a sewage
treatment load similar to the Marion-Corpening Creek WWTP.
The GLCSC effluent enters the Catawba River downstream
from the Marion Corpening Creek WWTP, yet the GLCSC ef-
fluent had little effect upon potential growth as evi-
denced.by the very low MSC of 0.58 mg dry weight per liter
at Station S-7.

-2-
o The enrichment impact of the Marion Corpening WWTP and
the Great Lakes Corbon-Silver-Creek WWTP appears to be
minimal in the Catawba River as bioavailable phosphorus
concentration decreases from 7.9 ug/L to 1.3 ug/L, a range
of concentration within acceptable limits for potential
algal production.
o Higher MSC's ranging from 7.50 mg dry weight per liter to
13.57 mg dry weight per liter at the headwaters of Lake
Rhodhiss seem to reflect nearby greater WWTP loadings to
the lake (Figure 1, Tables 4 and 5).
o Further down -lake, algal growth potentials decrease to
acceptable levels ranging in MSC concentration from 0.68
mg dry weight per liter to 3.91 mg dry weight per liter
(Table 4); lake station L-7, near the dam, had a greater
MSC than two successive up -lake stations, probably because
increased discharge velocity at the dam distributed nutrient
rich hypoliminetic waters up into the epilimnion at the
time of sampling.
Sincerely yours,
Ronald L. Raschke
Ecological Support Branch

TABLE 1. STATION LOCATIONS.
S-1 South Muddy Creek downstream fro the Marion-Corpening STP.
S-2 Silver Creek downstream Great Lakas STP.
S-3 Warrior Creek (not downstream from any STP).
S-4 Johns River (not downstream from any STP).
S-5 Lower Creek downstream from Lenoir STP.
S-6 Catawba River downstream of South Muddy Creek confluence.
S-7 Catawba River downstream of Silver Creek confluence.
L-1 Lake Rhodhiss at county road AR1301.
L-2 Lake Rhodhiss near. mouth of Smoky Creek.
L-3 Lake Rhodhiss by island just below mouth of Howard Creek.
L-4 Lake Rhodh i s s at county road SR1001.
L-5 Lake Rhodhi©© by boat ramp just beyond mouth of Freemason Creek.
L-6 Lake Rhodhiss where lake turns sharply to right below mouth
of Hayes Mill Creek.
L-7 Lake Rhodhiss just above dam.

t
` TABLE 2. TREATMENT SCHEME FOR LAKE RHODHISS.
Treatments
1. Untreated water sample (control).
2. Control plus 1.0 mg nitrogen per liter.
3. Control plus 0.05 mg phosphorus per liter.

•
TABLE 3. MEAN MAXIMUM STANDING CROP OF
SELENASTRUM CAPRICORNUTUM (mg/L),
LAKE RHODHISS, NORTH CAROLINA, JULY 1982.
Treatment
Control
4N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Control
+N
+P
Source
Muddy Creek
Silver Creek
Warrior Creek
Johns River
Lower Creek
Catawba River
Catawba River
Lake Rhodhiss
Headwaters
Lake Rhodhiss
Lake Rhodhiss
Lake Rhodhiss
Lake Rhodhiss
Lake Rhodhiss
Lake Rhodhiss
near dam
Station
S.D. - Standard Deviation
Replicates
1 • 2 3
S-1 .2.67 3.88
11.57 10.39
19.26 17.73
S-2 2.41 2.27
/.22 2.01
20.23 17.26
8-3 0.81 0.84
0.55 0.54
9.30 9.85
S-4 0.55 0.59
0.43 0.61
4.09 4.81
S-5 39.99 41.27
55.33 56.79
45.02 48.86
S-6 0.67 0.67
0.48 0.53
19.63 21.51
S-7 0.55 0.65
0.40 0.42
22.72 15.08
L-1 13.83 12.28
21.68 21.92
18.70 21.40
L-2 7.47 7.81
12.55 11.90
25.94 25.72
L-3 6.98 9.74
13.06 9.06
20.97 21.50
L-4 5.36 3.54
9.74 9.50
18.95 15.97
L-5 0.68 0.67
0.67 0.60
18.06 21.93
L-6 0.73 0.74
0.52 0.56
10.88 14.69
L-7 3.84 2.57
2.43 3.22
3.76 3.91
3.67
10.63
20.07
1.74
1.44
20.42
0.81
0.56
9.13
0.59
0.41
4.27
43.58
47.01
50.49
0.69
0.50
19.49
0.54
0.40
23.14
14.61
14.80
21.27
7.23
7.99
26.06
8.68
12.42
22.55
2.84
8.38
17.41
0.71
0.65
16.23
0.58
0.44
13.58
2.83
1.98
4.50
X
3.41
10.86
18.85
2.14
1.89
19.30
O .82
0.55
9.43
O .57
0.48
4.39
41.61
53.04
48.12
0.68
0.50
2041
0.58
0.41
20.31
13.57
19.47
20.46
7.50
10.81
25.91
8.47
11.51
21.67
3.91
9.21
17.44
0.69
0.64
18.74
0.68
0.51
13.05
3.10
2.54
4.06
O .D.
0.65
0.62
1.46
0.35
O .40
1.77
0.02
O .01
O .38
O .02
O .11
O .37
1.82
5.28
2.81
0.01
0.03
1.13
O .06
0.01
4.54
1.19
4.04
1.52
0.29
2.47
0.17
1.39
2.15
O .80
1.30
0.73
1.49
O .02
O .04
2.91
O .09
0.06
1.96
0.70
0.63
0.39

hi : .
K'
1�3r +�-sse�'.:�e'.a *_* ,..e_di �4F!S s� ',:� ` u of .:Sv'r'Ni� i'• ., •;: 3 .. <: r3 `i
x
•c.
rti
M1,
.. "::'77'. .tk=�•r ::�' u�::ti ' •,?.7�= �. �•r•`' .;rit c'S:��'P..; •,'j,=: -•.; ;•c.>. •-
.. .. .. a ..'>�'•. ... w'�c'-'�, tip. � .� ..._ ..T:, _ $,� �•�-.. -.... -- x4]� • - .. .
w1•-.
TABLE 4. CONTROL ALGAL GROWTH POTENTIAL WITH ASSOCIATED BIOAVAILABLE PHOSPHORUS,
LAKE RHODHISS, NORTH CAROLINA, JULY 1982.
Source
AGP Mean
Maximum
Standing Crop
Station (mg/L)
AGPa Bioavailable
Phosphorus (ug/L)
Muddy Creek • S-1 3.41 7.9
Silver Creek S-2 2.14 5.0.
Warrior Creek S-3 0.82 1.9
Johns River S-4 0.57 1.3
Lower Creek S-5 41.61 96.8
Catawba River S-6 0.68 1.6
Catawba River S-7 0.58 1.3
Lake Rhodhiss headwaters L-1 13.57 31.6
Lake Rhodhiss L-2 7.50 17.4
Lake Rhodhiss L-3 8.47 19.7
Lake Rhodhiss L-4 3.91 9.1
Lake Rhodhiss L-5 0.69 1.6
Lake Rhodhiss L-6 0.68 1.6
Lake Rhodhiss near dam L-7 3.10 7.2
aBioavailable PhosphorusMSC
-
in ug,/L• 430

TABLE 5. MAJOR* SEWAGE TREATMEN..PLANT (STP) DISCHARGES IN
LAKE RHODHISS DRAINE AREA.(MGD)
- , .. .
STP Design Flow (MGD)
1 t-r Morganton -Catawba River 8.0
2. Valdese -Lake Rhoades 7.5
3. Lenoir -Lower Creek 6.0•
4. Valdese-McGalliard Creek 3.2
5. Marion-Corpening Creek 3.0
6. Great Lakes Carbon -Silver Creek 2.6
7. Valdese -Hoyle Creek 1.0
8. Drexel -Howard Creek 0.5
Total Discharge 31.8
*Only those STP with a design flow equal to or greater than 0.5 MGD.

.ti
M �
A METHOD FOR THE REDUCTION OF LAKE MODEL PREDICTION ERROR
by
Kenneth H. Reckhow
School of Forestry and Environmental Studies
Duke University
Durham, North Carolina 27706
USA
Short Title: Lake Model Error Reduction
May, 1982

A METHOD FOR THE REDUCTION OF LAKE MODEL PREDICTION ERROR
ABSTRACT
Recent work on methods of error analysis to accompany the application of
lake models has not enjoyed great acceptance in part because of the magnitude of
the error term. For the models that have undergone a rigorous error analysis
(generally the single equation cross -sectional regression models), prediction
errors for various water quality variables are often + 30% or more. A procedure
is proposed herein for the reduction of the error associated with the prediction
of lake phosphorus concentration from land use and hydrologic data. Existing
lake quality data are used in the prediction, and the model is employed only to
project changes from the present state. This obviates the need to project all
land use impacts with the model; only those proposed to change are projected.
The result is a substantial reduction in prediction error for many planning
scenarios.

INTRODUCTION
As resources become scarce, we must become increasingly demanding of the
justification for policy and planning changes. This justification often takes
the form of analytical planning methods, as for example, issues of a quantita-
tive nature are appropriately analyzed using mathematical methods. These
methods, however, must be scrutinized both in terms of their applicability and
their ability to describe system response reliably. Without accurate and
precise mathematical methods, there is no assurance that projected impacts from
policy and planning changes are likely to occur.
In the water quality arena, investigators (e.g., Beck and van Straten,
1982), have recognized the importance of uncertainty and risk. This informa-
tion, in turn, is gradually being incorporated into water quality management
planning (e.g., Burges and Lettenmaier, 1975; Reckhow, et al., 1980). These
approaches have been slower to infiltrate the lake eutrophication modeling field
in part because of the scarcity of data and in part because of the perceived
complexity of the problem. Nonetheless, error analysis methodologies have
recently been proposed (Reckhow, 1979a; Chapra and Reckhow, 1979; Scavia, 1980)
for several lake models.
A discouraging feature of these efforts to estimate prediction uncertainty
associated with water quality modeling has been the magnitude of the error term.
For example, when a.complete error analysis has been undertaken, the prediction
error for most lake trophic state variables (as a function of watershed charac-
teristics and hydrologic variables) is at least + 30% and sometimes spans an
order of magnitude on a logarithmic scale. With errors this large, it is
perhaps understandable that error analysis is not widely used in water quality
management planning.
-1-

This unfortunate state of affairs does not lessen the need for measures of
reliability in our planning procedures. Rather, it makes this need more acute,
if only to alert model users to the sizeable error terms associated with the
methods being employed. Further, it represents .a challenge to researchers to
develop and propose ways to reduce this prediction error. Contributions of that
nature are likely to be among the best measures to increase the effective use of
mathematical models in water, and specifically lake, quality management.
Uncertainty
Reliability, or its converse, uncertainty, serves three vital functions in
planning studies:
1. Reliability represents an estimate of the value of information. If
the reliability associated with a prediction is low, the prediction
is uncertain and imprecise, and the predictive information is not
particularly valuable. Alternatively, if the reliability of a
prediction is high, the prediction is precise, and the predictive
information can be valuable.
2. Important factors that are poorly characterized (i.e., have high
uncertainty) may be identified when reliability is assessed. The
analysis of uncertainty, or error, helps model developers and model
users in a sensitivity analysis exercise. Specifically, estimation
of errors allows the analyst to identify those characteristics that
have significant error and have a significant effect on the prediction.
The analyst then realizes that in order to obtain a precise prediction,
these uncertain, prediction -sensitive terms must be better defined.
3. Discrimination among control strategies may be explicitly evaluated
with an assessment of reliability. Without uncertainty analysis,
-2-

one is given the impression that prediction differences of one micro-
gram per liter or less are significant and indicate a well-defined
ordering of quality states. With uncertainty analysis, the prediction
error identifies a region in which land use strategies may be
predictively indistinguishable. The error analysis allows the planner
to determine when land use strategy impacts can be predictively
distinguished, given the error associated with model applications.
Uncertainty Analysis: Shortcomings in Lake Modeling Applications
Several phosphorus lake models have been proposed recently that incorporate
a procedure for estimating prediction uncertainty (Chapra and Reckhow, 1979;
Reckhow, 1979a,b; Reckhow and Simpson, 1980; Reckhow and Chapra, 1982; Reckhow,
et al., 1980). These approaches represent an improvement over the purely
deterministic analyses presented in previous papers, since the error estimate is
a measure of prediction information value.
However, there are two problems or shortcomings with the existing error
analysis methods. Errors arise in model applications because of error in the
model, the model parameters, and the model variables. In a more fundamental
sense, one may also say that the errors are caused by natural variability,
inadequate sampling design, measurement error and bias, and model specification
error. When the data set variables used to construct a model contain error,
then this error is transmitted to the model error term for the fitted model.
Since virtually all limnological statistics contain error as a result of the
aforementioned causes, then lake models developed from these data contain error
associated with the data error. This means that the model standard error term
includes an error component associated with errors in the model variables. This
-3-

is an unwanted component, yet it is unavoidably there given present knowledge
and data.
Models can be employed in a descriptive or a predictive mode. When used
descriptively, models may be used to relate observed inputs (the independent
variables) to observed outputs (the dependent variable). In a descriptive
application, when all variables are directly measured in the same manner as the
variables in the model development data set were measured, then there is no need
to add additional application lake variable error. This is because the appro-
priate variable error is already contained in the model error term. However,
when the model is used in a predictive mode, the dependent variables generally
cannot be measured (because the predictive nature implies conditions not yet
physically realized). Predictive applications of a model require that the
analyst extrapolate variable values from other points in time and/or space.
This extrapolation process introduces error beyond
model error term. Thus, predictive use of a
error analysis that includes variable error.
must be undertaken thoughtfully, however, to
model
This
avoid
that already contained in the
should be accompanied by an
errors -in -variables analysis
"double counting" errors (due
to the errors -in -variables term already contained in the model standard error
term). The first problem of existing error analysis methods, therefore, is that
their application may lead to error double counting.
The second shortcoming associated with existing methods is of greater
importance, given the fact that with care, double counting can probably be kept
at an acceptably low level. The second problem relates to the magnitude of the
error term. The input-output empirical phosphorus lake models of concern here
are developed from cross -sectional analyses. The models are simple, our
knowledge of limnology is limited, and all phosphorus -settling processes are
aggregated into one empirically -determined model parameter. As a result, the
-4-

model error term is large, since it represents cross -sectional variability,
measurement and sampling error, and model specification error. With a decrease
in cross -sectional heterogeneity (culminating in the single lake over time), the
model error term may correspondingly be reduced. However, errors in the model
variables, particularly in phosphorus loading, can be substantial for certain
applications. The combined effect of these error terms is a large prediction
error using existing error analysis methods. The magnitude of this error term,
and the associated prediction intervals, is such that the analyst is often
unable to find "statistically significant differences" among competing lake
management options.
A PROPOSED ERROR ANALYSIS METHODOLOGY
An alternative error analysis methodology proposed herein will substantial-
ly reduce prediction error over existing error analysis techniques for most
applications. This procedure exploits two features that are common to many lake
quality management planning situations:
1. For most projected planning scenarios, the land use area expected to
change is small in comparison to the land use area that is expected
to remain constant during the planning period.
the impact of the change is generally small in
impact of the existing, unchanging land uses.
2. Existing phosphorus lake data reflect the impact of present
Stated another way,
comparison
to the
land use
conditions. Furthermore, the variability in these data represent the
variability in impact response. These data could already be in
existence or they could be acquired upon initiation of this modeling
program.

I
Here, and below, land use is intended to refer to any activity that results in
materials (in this case, phosphorus) discharge to the water body of concern.
This includes point sources of discharge such as wastewater treatment
plants.
To see how these features can lead to prediction error reduction, consider
a planning scenario in which no change is projected so that existing land use
and future land use are equivalent. In that case, two methods may be used to
predict future lake phosphorus concentration (ignoring temporal variability for
the moment):
1. A phosphorus lake model may be applied to relate land use to phosphorus
concentration through literature export coefficients. This standard
procedure is accompanied by a high prediction error.
2. Existing phosphorus lake data may be used to describe future lake
quality under unchanging watershed conditions. Here the error term
is a function of the standard error of the estimate for the data and
of the representativeness of the data.
In virtually all cases, with even a modest amount of phosphorus lake data,
the error for the second method will be considerably smaller than the error for
the first method, given the size of the phosphorus loading and model error
terms.
If this scenario is modified slightly to a situation common in lake quality
management planning, the new error analysis methodology may be outlined.
Consider a planning scenario in which a relatively small land use change is
projected. The new modeling and error analysis methodology stipulates that the
analyst use:
-6-

1. Existing lake data to evaluate the impact of unchanging land uses; and,
2. The model to evaluate the impact of the land use change and the impact
of hydrologic variability.
Existing error analysis methods do not permit the analyst to distinguish
between land uses that are projected to change and land uses that are to remain
constant. This means that the impacts of all watershed land uses on lake
phosphorus concentration are evaluated through the model. Since the impact of
unchanging land uses is manifested in recent lake phosphorus concentration data,
information (the lake phosphorus concentration data) is wasted and high
prediction errors result.
To indicate the magnitude of the error reduction associated with the
procedure outlined herein, consider the following model (Reckhow, I979b):
where:
L
P 11.6 + 1.2gs (1)
P = lake phosphorus concentration (mg/1)
L = annual areal phosphorus loading (g/m2-yr)
qs = annual areal water loading (m/yr)
The model standard error is 0.128 in logarithmically -transformed concentration
units. This translates to about a + 30% prediction error when the antilog is
determined for a particular concentration. The difference between the existing
and proposed error analysis methodologies may best be stated through a
hypothetical comparison.
With the existing error analysis procedures, the model is used to predict
the impact (on lake phosphorus concentration) from all land uses. The model
error alone (to which errors in variables must eventually be added) is
approximately + 30%. For oligotrophic lakes, this model error is relatively
-7-

small. However, planning frequently occurs on lakes with phosphorus
concentrations in the range of 0.020 mg/1 to 0.060 mg/I.
For example, to demonstrate this procedure, assume that a lake with the
following characteristics is to be modeled using Equation 1:
where:
P = 0.040 mg/1
qs = 10.0 m/yr
L = 1.00 g/m2-yr
(observed concentration)
This yields a predicted phosphorus concentration of:
L 1.00
11.6 + 1.2 qs 11.6 + 1.2 (10)
P = 0.042 mg/1
Assume that a land use change is projected leading to the conversion of a
forested portion of the watershed to agriculture. Using phosphorus export
coefficients, the phosphorus loading to the lake from the forested land (to be
eliminated) is estimated as 0.02 g/m2-yr, and the phosphorus loading to the
lake expected from the new agricultural activity is estimated as 0.10 g/m2-yr.
Note that these values have been converted to square meters of lake surface area
so that they may be directly combined with the phosphorus loading term in the
model above. This results in a projected phosphorus loading of:
P = 1.00 - 0.02 + 0.10 = 1.08 g/m2-yr
The projected lake phosphorus concentration associated with this land use change
is:
-8-

r
L 1.08
11.6 + 1.2 qs 11.6 + 1.2 (10)
P = 0.0458 mg/1
Using the model standard error of 0.128 (in logarithmically -transformed
concentration units), the prediction error may be estimated as:
or:
Prediction interval bounds = 101og (0.0458) + 0.128
= 0.0341 mg/1, 0.0615 mg/1
0.0341 mg/1 < P < 0.0615 mg/1
As stated above, this is, on average, a prediction interval of about + 30%,
which for convenience may be approximately expressed:
P = 0.0458 mg/1 + .0135 mg/1
This is a substantial error term, and it may both discourage the planner from
using error analysis and obscure the differences among management strategy
impacts.
With the error analysis procedure proposed herein, the model is used to
predict the impact (on lake phosphorus concentration) for the changing land uses
only. As described below, the analyst must use the model to evaluate the impact
for both the old and the new land use. Most projected changes in land use have
a relatively minor impact on lake phosphorus concentration in comparison to the
impact from all watershed land uses.
For the example presented above, the impact of the forested land (to be
eliminated) is:
-9-

P
L 0.02
11.6 + 1.2 qs 11.6 + 1.2 (10)
P = -0.000847 mg/1
and the impact of the projected new agricultural activity is:
L 0.10
P = + 11.6 + 1.2 qs + 11.6 + 1.2 (10)
P = +0.004237 mg/1
This results in a net projected phosphorus concentration change of:
P = -0.000847 + .004237 = 0.00339 mg/1
Since the errors may be additive, the total prediction error reflects the sum of
the projected changes (0.000847 + 0.004237 = 0.00508 mg/1), yielding an
approximate projected change prediction interval of:
Prediction interval bounds = 101og (0.00508) + 0.128
= 0.00378 mg/I, 0.00682 mg/1
As in the above example, this is, on average, a prediction interval of about +
30%, which for convenience may be approximately expressed:
Projected P change = 0.00339 mg/1 + 0.00152 mg/1
-10-

Note that the prediction interval is based on the sum of the phosphorus
concentration changes, and thus will be greater than + 30% of the net
concentration change.
If a reasonable amount of lake sampling for phosphorus concentration has
occurred (under a good sampling design), then the impact of unchanging land uses
may be evaluated objectively. Even with a modest amount of data, the standard
error of the mean phosphorus concentration will usually be small. For example,
Reckhow (1979c) evaluated phosphorus data variability in a cross -sectional study
of lakes and estimated a mean coefficient of variation of 0.6 for total
phosphorus concentration measurements in these lakes. This translates to a
standard deviation that is 0.6 times the observed mean, which, for the example
lake is:
standard deviation = (0.6) (0.040 mg/1) = 0.024 mg/1
With a relatively small data set of perhaps 25 phosphorus concentration
measurements, the standard error of the mean (1/'i times the standard deviation)
is:
standard error = (1/ 25) (0.024 mg/1) = 0.0048 mg/1)
Combining this phosphorus concentration measurement error term with the
previously calculated model'prediction error term for changing land uses (square
the error terms, add, and calculate the square root), the total prediction error
is:
prediction error = + [(0.0048 mg/1)2
= + 0.0050 mg/1
+ (0.00152 mg/1)2]1/2
-11-

The result is a prediction for phosphorus concentration in the lake after the
projec.* c land use change:
where:
P = (observed phosphorus concentration + predicted net
phosphorus concentration change) + prediction error
P = (0.040 mg/1 + 0.00339 mg/1) + 0.0050 mg/1
P = 0.0434 mg/1 + 0.0050 mg/1
This compares to the prediction and prediction error generated from the first
method:
P = 0.0458 mg/1 + 0.0135 mg/1
The comparison presented above does not include all error terms for either
methodology, but the major error terms are calculated. Note that the proposed
methodology leads to a reduction in prediction error of about 60% for the
hypothetical example. The analyst must realize, however, that this error
reduction associated with the methodology is contingent on the magnitude of the
land use change that must be evaluated using the model. Obviously, as the
magnitude of the projected impact increases, the comparative advantage of the
proposed error analysis procedure diminishes. Alternatively, additional
measurements of the lake phosphorus concentration prior to land use change will
increase the sample size and reduce the error for the proposed methodology.
This should be considered when allocating funds for lake water quality
management planning.
The hypothetical example comparing error analysis procedures includes three
of the four basic terms for the proposed methodology. The four error terms, and

their interpretations in modeling applications, are:
1. Uncertainty in the assessment of current lake phosphorus concentration.
If adequate data exist, this error term may be represented by the
standard error in the central tendency statistic (e.g., the mean).
Data "adequacy" should be determined by the representativeness of the
existing data on a spatial and temporal (within and across years)
basis. In situations with inadequate data, this error term may be
estimated through regression analysis with more comprehensive data sets
on correlated variables (e.g., Secchi disk transparency) or through
subjective determination.
2. Uncertainty in the hydrology variable, qs. Cross -sectional error
in qs already exists in the model standard error for the reason
identified earlier. This qs-component of model error has unknown
magnitude and may be sufficient for lakes with low inflow -outflow
variability. Further, when several years of in -lake phosphorus con-
centration data exist (described in error component number 1 above),
these data already exhibit the effect of qs-variability, making qs-
error analysis unnecessary. Therefore, this additional error term may
be considered optional. In cases with substantial variability in year-
to-year values of the hydrology variable (qs), and limited in -lake
phosphorus data, a qs-error term should be included, propagated
through the model (using first order analysis; see Reckhow, et al.,
1980) . This error term should represent the year-to-year variability
and the estimation or measurement error associated with the determina-
tion of qs.

3. Uncertainty in the prediction of the impact of the projected new land
use on lake water quality. This term is estimated using the phosphorus
lake model and a procedure like that presented in Reckhow, et al.,
(1980). This error component includes model error and error in the
estimate of phosphorus loading for the projected land use change.
Note that for minor land use changes (relative to the entire watershed)
the impact of this error term is small (despite the inclusion of model
error) because the fractional phosphorus loading addition is small.
4. Uncertainty in the prediction of the impact of the existing land use
in the area projected to undergo change. To properly assess the
anticipated change, the analyst must determine the impact of both the
old and the new land uses using the modeling/error analysis procedure.
These calculations are undertaken in the same manner as are the
calculations for the error term described in number 3 above. Note that
here, too, the error is small when the fractional phosphorus loading
"subtraction" is small.
In summary, the two error analysis methodologies may be compared with the
aid of Figure 1. At the top of the figure, the traditional method is undertaken
by estimating the phosphorus loading, and the loading estimation error, for all
land uses in the lake watershed. The phosphorus loading and the ,loading error
are propagated through the model for the calculation of the predicted lake
phosphorus concentration. Prediction error for this procedure is determined
primarily by the loading error and the model error. Since the model error term
is proportional to the phosphorus loading magnitude propagated through the
model, and since all phosphorus loading is propagated through the model under
the traditional procedure, the model error term is large. As a result, the
-14-

total prediction error for the traditional procedure is often + 30 (or more) of
the predicted concentration.
The new procedure often leads to a prediction error reduction because it is
not required that the model (and model error) be used to predict all land use
impacts. The watershed may be divided into land uses that are expected to
remain constant over a planning period and land uses that are expected to
change. Similarly, the average phosphorus concentration in a lake may be
divided into a fraction contributed by unchanging land use and a fraction
contributed by land use that is expected to undergo change. For the new error
analysis procedure, existing phosphorus lake data represent the impact of all
existing land uses. The variability in these data reflect estimation
uncertainty. Since no predictive model was required to assess this impact, the
uncertainty term is often small. Added to this uncertainty is the prediction
error associated with the determination of the impact of all changing land uses
calculated using the model. However, since a fraction (often a sizeable
fraction) of the land use impacts is assessed without the model, total
prediction error for the new procedure is generally much lower than it is for
the old procedure.
ISSUES FOR CONSIDERATION
An effort has been made in this paper to stress a conceptual discussion of
error analysis, and to demonstrate the attributes of a.proposed procedure
through a,simple example. For more detailed applications, there are some issues
that, at best, were alluded to above, yet warrant further consideration. These
issues are identified below in the hope that additional analysis of this topic
will be stimulated.
-15-

1. What is, or should be, the meaning of the lake phosphorus measurements
error term? It is intended to represent the impact of unchanging land
use on lake water quality.
a. Can we determine whether the lake is in steady state relative
to watershed land uses?
b. It has been indicated above that the lake phosphorus measurements
should represent spatial and temporal variability. How does the
need for temporal variability representation conflict with the
need for "steady state" and the likelihood that most lakes undergo
continuous small land use changes?
c. Can an "adequate" sampling design be defined in a general and
objective sense?
2. To what extent is time series variability represented in the lake
phosphorus measurements and to what extent must it be included in the
qs-error term?
3. Time series data for qs may not exist for an application lake. It
seems reasonable that qs-variability could be extrapolated from other
similar watersheds or perhaps from precipitation data. In those situa-
tions, an additional error term should be included, representing
possibled bias associated with the use of extrapolated data.
4. The analyst should be aware of the difference between the standard
deviation and the standard error of the estimate. The standard devia-
tion is a measure of the inherent variability in a set of data or in a
population. The standard error of the estimate, which may often be
calculated from the standard deviation (by dividing by 1in), reflects
the error in a statistic. The error analysis procedures yield a
standard error of the estimate that represents the error in the

prediction; this error does not directly reflect the variability to be
expected for (in this case) lake phosphorus concentration.
5. The error propagation equation (Benjamin and Cornell, 1970) is to be
used to calculate the impact of errors in the variables and errors
in the parameters on the total prediction uncertainty. One term in
the error propagation equation represents the error contribution
associated with variable (or parameter) correlation. Should this term
be computed for the correlation between the old, changing land use
phosphorus loading and: a) the new land use phosphorus loading, and/or
b) qs? This question is largely of a conceptual nature, since the
impact on total prediction uncertainty in either case is undoubtedly
quite small. Nevertheless, it illustrates the type of conceptual
problem that must be considered as error analysis methodologies are
proposed and refined.

4
ti
REFERENCES
Beck, M.B. and G. van Straten (1982) Uncertainty and Forecasting of Water
Quality. Pergamon Press, New York (in press).
Benjamin, J.R., and C.A
for Civil Engineers.
Burges, S.J., and D.P.
quality management.
. Cornell (1970) Probability, Statistics, and Decision
McGraw-Hill, New York, 684 pp.
Lettenmaier (1975) Probabilistic methods in stream
Water Res. Bull. 11(1):115-130.
Chapra, S.C., and K.H. Reckhow (1979) Expressing the phosphorus loading concept
in probabilistic terms. J. Fisheries Res. Board. Can., 36(2):225-229.
Reckhow, K.H. (1979a) Empirical lake models for phosphorus: development,
applications, limitations, and uncertainty, IN: Perspectives on Lake
Ecosystem Modeling. p. 193-221, D. Scavia and A. Robertson (eds.). Ann
Arbor Science Publishers, Ann Arbor, Mich.
Reckhow, K.H. (1979b) Quantitative techniques for the assessment of lake
quality. EPA-440/5-79-015. U.S. Environ. Prot. Agency, Washington, D.C.
146 pp.
Reckhow, K.H. (1979c) Lake data analysis and phosphorus variability. Paper
presented at the North American Lake Management Conference, Michigan State
University, East Lansing, Mich.
Reckhow, K.H., and J.T. Simpson (1980) A procedure using modeling and error
analysis for the prediction of lake phosphorus concentration from land
use information. Can. J. Fish. Aq. Sci., 37(9):1439-1448.
Reckhow, K.H., M.N. Beaulac, and J.T. Simpson (1980) Modeling phosphorus
loading and lake response under uncertainty: A manual and compilation
of export coefficients. EPA 440/5-80-011. U.S. Environ. Prot. Agency,
Washington, D.C. 214 pp.
Reckhow, K.H., and S.C. Chapra (1982) Engineering Approaches for Lake Manage-
ment: Volume I - Data Analysis and Empirical Modeling, Volume II -
Mechanistic Modeling. Ann Arbor Science Publishers, Inc., Ann Arbor, MI
(in press).
Scavia, D. (1980) Uncertainty analysis of a lake eutrophication model, Ph.D.
dissertation, 159 pp., Univ. of Michigan, Ann Arbor, MI.

FIGURE CAPTION
1. A Comparison of Error Propagation Methodologies.

•!-'"r'c'�°•_ �.,a :gip s •r ; ; o.., x s -.-.. - �� ti :x... .. ... .
15Ea1"-=.�3.: .... �t:ft •.itli � ... -... ��•� �'s�"Y.iis�,.f�r.'=a��iYr•••�i�:,F-.,—'•'.'•`•:•y •.. '=T:.rm+}'�. ��'E•i 7i.��: -: L' .���'.::�=?'x!,+�5r
•
Old
Procedure
New
Procedure
Phosphorus Loading
for Entire
Watershed
Phosphorus Loading
for Land Use
Changes
Measured Phosphorus
Concentration
in Lake
Phosphorus
Lake
Model
a
Prediction Error
Prediction Error