using turbidity to predict total suspended solids in mined...
TRANSCRIPT
STATE OF ALASKA
DEPARTMENT OF NATURAL RESOURCES
DIVISION OF GEOLOGICAL AND GEOPHYSICAL SURVEYS
Steve Cowper, Governor
Judith M. Brady, Commissioner
Robert B. Forbes, Director and Stute Geologist
August 1988
Thie report ia a preliminary publication of DGGS.The author ia mlely reqxxuible for itr content andwill appreciate candid comments on the accuracy ofthe data aa well as euggeetionr to improve the report.
Report of Investigations 88-2USING TURBIDITY TO PREDICT TOTAL SUSPENDED
SOLIDS IN MINED STREAMS ININTERIOR ALASKA
B YStephen F. Mack
STATE OF ALASKADepartment of Natural Resources
DIVISION OF GEOLOGICAL 6 GEOPHYSICAL SURVEYS
According to Alaska Statute 41, the Alaska Division of Geological andGeophysical Surveys is charged with conducting 'geological and geophysicalsurveys to determine the potential of Alaskan land for production of metals,minerals, fuels, and geothermal resources; the locations and supplies ofground water and construction materials; the potential geologic hazards tobuildings, roads, bridges, and other installations and structures; and shallconduct such other surveys and investigations as will advance knowledge ofthe geology of Alaska.'
In addition, the Division of Geological and Geophysical Surveys shallcollect, record, evaluate, and distribute data on the quantity, quality, andlocation of underground, surface, and coastal water of the state; publish orhave published data on the water of the state and require that the resultsand findings of surveys of water quality, quantity, and location be filed;require that water-well contractors file basic water and aquifer data,including but not limited to well location, estimated elevation, well-driller's logs, pumping tests, flow measurements, and water-qualitydeterminations; accept and spend funds for the purposes of this section, AS41.08.017 and 41.08.035, and enter into agreements with individuals, publicor private agencies, communities, private industry, and state and federalagencies; collect, record, evaluate, archive, and distribute data on seismicevents and engineering geology of the state; and identify and inform publicofficials and industry about potential seismic hazards that might affectdevelopment in the state.
Administrative functions are performed under the direction of theDirector, who maintains his office in Fairbanks. The locations of DGGSoffices are listed below:
.794 University Avenue .400 Willoughby Avenue(Suite 200) (3rd floor)Fairbanks, Alaska 99709 Juneau, Alaska 99801(907)474-7147 (907) 465-2533
.3700 Airport Way .18225 Fish Hatchery RoadFairbanks, Alaska 99709 P.O. Box 772116(907)451-2760 Eagle River, Alaska 99577
(907) 696-0070
This report is for sale by DGGS for $3. DGGS publications may be %n-spected at the following locations. Mail orders should be addressed to fheFairbanks office,
.3700 Airport WayFairbanks, Alaska 99709
,400 Willoughby Avenue(3rd floor)Juneau, Alaska 99801
.U.S. Geological SurveyPublic Information Office701 C StreetAnchorage, Alaska 99513
. Information SpecialistU.S. Geological Survey4230 University Drive, Room rO1Anchorage, Alaska 99508
ii
CONTENTS
Page
Abstract ............................................................... 1Acknowledgments ........................................................ 1Introduction ........................................................... 1Background ............................................................. 2Methods ................................................................ 5
Sources of data ................................................... 5Geographical organization ......................................... 7Statistical methods ............................................... 8Model validation .................................................. 10Velocity-turbidity multiple regression model ...................... 11
Results ................................................................ 11Summary statistics ................................................ 11Regression equations .............................................. 18Analysis of covariance ............................................ 24Model validation .................................................. 26Velocity-turbidity multiple regression ............................ 28
Discussion ............................................................. 29Conclusion ............................................................. 30References cited ....................................................... 31Appendix A. Turbidity and TSS data from interior Alaska streams ....... 33Appendix B. Explanation of statistical techniques ..................... 45Appendix C. Model validation results .................................. 47
FIGURES
Figure 1.
2.3.4.5.
6.
7.8.9.10.11.
12.
13.
14.
Plot of turbidity and TSS above and below mining, EagleCreek in Birch Creek basin ..............................
Interior Alaska basins with placer mining data ...............Plot of turbidity and TSS for streams in Birch Creek basin ...Plot of turbidity and TSS for streams in Crooked Creek basin.Plot of turbidity and TSS for streams in the Chatanika
and Goldstream basins ...................................Plot of turbidity and TSS for streams in the Upper Tolovana,
Chena and Koyukuk basins ................................Plot of turbidity and TSS for sites on Crooked Creek .........Plot of turbidity and TSS for sites on Chatanika River .......Plot of turbidity and TSS for sites on Fish Creek ............Plot of turbidity and TSS for sites on Tolovana River ........Plot of turbidity-TSS regression lines for seven basins in
interior Alaska .........................................Plot of turbidity-TSS regression lines for streams in
Birch Creek basin .......................................Plot of turbidity-TSS regression lines for streams in
Crooked Creek basin .....................................Plot of turbidity-TSS regression lines for streams in
Chatanika and Goldstream basins .........................
89
1414
15
1516161717
20
20
21
21
iii
15. Plot of turbidity-TSS regression lines for streams in theUpper Tolovana and Chena basins......................... 22
16. Plot of turbidity-TSS regression lines for sites onCrooked Creek . . . . . . . . . . . . ..I............................ 22
17. Plot of turbidity-TSS regression lines for sites onChatanika River ..*...,....,............................. 23
18. Plot of turbidity-TSS regression lines for sites in theUpper Tolovana River and Chena basins................... 23
19. Plot of regression equation coefficients of determinationand standard errors of estimate......................... 25
20. Plot of turbidity and TSS by year, Chatanika River belowFaith Creek, 1983-84 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
21. Plot of turbidity and TSS by year, Crooked Creek .at Central, 1984-85 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
22. Z score distribution for 1985 Chatanika and Tolovana dataand 1983 ACFRU data . . . . . . . . . ..*......................... 28
TABLES
Table 1. Summary statistics for streams and sites with 15 or moreobservations . . . . . . ..a................................... 12
2. Summary of regression equations and covariance analysisfor basins, streams and sites in interior Alaska........ 18
3. Summary of covariance analysis by year.,......,.............. 264. Comparison of multiple and simple linear regression equations
from Crooked Creek basin ,.*.,.,.............,........o.. 29
iV
USING TURBIDITY TO PREDICT TOTAL SUSPENDED SOLIDSIN MINED STREAMS IN INTERIOR ALASKA
byStephen F. Mack'
ABSTRACT
Data from mined streams in interior Alaska were used to determine theextent to which data from different locations can be combined to predicttotal suspended solids (TSS) from turbidity measurements. Data were trans-formed into logarithms with log TSS regressed on log turbidity using linearregression. Coefficients of determination (r') for equations derived frommeasurements in seven basins, 15 streams and 18 sites ranged from 0.261 to0.996 with standard errors of estimate (SEE) ranging from +155 percent(-61 percent) to +14 (-13 percent). Covariance analysis indicated relation-ships between TSS and turbidity data collected from different basins to bestatistically different; turbidity-TSS relationships of data from differentstreams within a basin may also differ, and relationships of data fromdifferent sites within a stream may differ. Also, data collected in separateyears may have statistically different relationships. Model validation con-firmed the uncertainty of using previous years' data. At one site, multipleregression with turbidity and average velocity used as predictors for TSSimproved the r2 from 0.20 of a simple turbidity-TSS model to 0.68 and reducedSEE from +98 percent (-49 percent) to +49 percent (-33 percent).
ACKNOWLEDGMENTS
I wish to express my appreciation to the following for their assistancewith this project: Dr. Robert Carlson for his patience, encouragement, andcritical evaluation throughout the project history; Dr. Dana Thomas for hisassistance with the statistical methods; and other investigators whose data(and advice about the data) were essential to this project---in particular,Bruce Cleland, Larry Peterson, Gary Nichols, Jacqueline LaPerriere, PhyllisWeber, and Jeffrey Hock. My appreciation is also extended to JosephSt. Sauver and others at the University of Alaska computer nodes for theirassistance with the SAS computer programs, to Shirley Liss for assistancewith manuscript figures, to Roman Motyka for review of and comments on thefinal draft of the manuscript, and to my advisory committee at the Universityof Alaska-Fairbanks for their critical evaluation of the manuscript.
INTRODUCTION
This report presents the results of an investigation of the statisticalrelationship between turbidity and total suspended solids (TSS) in free-flowing, placer-mined streams in interior Alaska. Because of high levels ofsediment discharge, increasing scrutiny is being directed at the placermining industry, To determine the impact of discharged sediment, samples
1DGGS, 3700 Airport Way, Fairbanks, Alaska 99709 (current address: NevadaRegional Water Board, P.O. Box 857, Sparks, NV 89432.
-l-
from mined streams are collected and analyzed, both for turbidity and forTSS. The turbidity parameter is easier, less time consuming, and lessexpensive to measure. If a good statistical relationship between turbidityand TSS can be established, turbidity analysis would serve for most purposes.A good statistical relationship is defined as one which has an acceptablecoefficient of determination (r2) and standard error of estimate.
Several government agencies and consulting firms have collected aconsiderable amount of paired turbidity and TSS data from placer-minedstreams in interior Alaska during the past 3 yr. I have organized theseobservations on a basin-stream-site basis and applied statistical techniquesto determine the feasibility of predicting TSS from turbidity, using existingdata.
BACKGROUND
Placer mining entails locating free gold in alluvial (placer) depositsnear bedrock, uncovering the gold-bearing layer (stripping), and separatinggold from sand and gravels (sluicing). Stripping and sluicing, as practicedin Alaska, often results in the discharge of noticeable amounts of sedimentinto many bodies of water that otherwise would be virtually sediment free.This is contrary to state and federal laws by which the placer miningindustry is more and more being governed.
Two parameters by which the impact of placer mining on water bodies ismeasured are turbidity (which relates to the muddiness or cloudiness of thewater), and TSS (which describes the physical amount of sediment in the watercolumn).
Turbidity is defined by APHA (1985) as ‘the expression of the opticalproperty that causes light to be scattered and absorbed rather than trans-mitted in straight lines through the sample.’ Such scattering and absorptionis caused by particles--clays or silts, algae, organic detritus, and otherfine insoluble sediments --suspended in the water (Hach and others, 1984). InAlaska , turbidity is measured by turbidimeter, in nephelometric turbidityunits (NTU). Nephelometry is the measurement of light scattered at rightangles to the incident light beam passing through a sample (Hach and others,1984). The deleterious effects of turbidity include, but are not limited to,aesthetic and functional impairment of recreational use, impaired product-ivity and adverse impacts on the food chain because of reduced light penetra-tion, avoidance by fish populations, and impaired treatment of drinking water(Peterson and others, 1985).
Turbidity measurement requires a properly calibrated turbidimeter andappropriate glassware. Portable turbidimeters are available which canaccurately measure turbidity in the field. Nephelometric turbidimeters canmeasure values to 100 NTUs; however, the standard method requires dilutionsto below 40 NTU (APHA, 1985). Placer-mined streams are often above 100 NTUand may require several dilutions.
TSS is defined by APHA (1985) as ‘the portion of total solids retainedby a glass-fiber filter. ’ TSS is reported in concentrations (usually milli-
-2-
grams per liter) and represents the mass of non-dissolved solids contained inthe water column. TSS is not to be confused with settleable solids, which isthe volumetric quantity of solids that will settle in an Imhoff cone in 1 hr(APHA, 1985) and are reported in milliliters per liter. This project did notinvestigate any relationship between turbidity and settleable solids.
Recent research connects high TSS concentrations to biota damage,including impacts on fish at various life stages and impacts on invertebrates(Peterson and others, 1985). TSS, combined with discharge, gives an estimateof sediment load, which is the total amount of sediment carried by a stream.
TSS measurement requires ovens, analytical balances, and glassware forfiltering samples, and is not practical outside a properly equipped labora-tory. TSS analysis requires more time than turbidity measurements. Samplesmust be filtered (which can take hours with silt-laden samples) and dried inan oven. Turbidity measurements can be done in the field and require onlytime for the turbidimeter readout to stabilize and, for highly turbidsamples, time for dilutions.
Extensive literature exists on the relationship of turbidity to TSS.Measurement of turbidity was developed as an index of suspended materialconcentrations, but it has been long recognized that no single, universalrelationship is applicable (Lloyd, 1985); turbidity is an optical measurementof reflected light, whereas TSS is measured by the actual mass of particlesretained on filter paper. Investigators have found that particles with verylittle mass can cause turbidity; in fact, much of the variation in turbidityis attributed to particles 10 microns or smaller (Nichols, 1986). Sampleswith identical TSS measurements but differing particle sizes can have verydifferent turbidity measurements. Conversely, of two samples with similarturbidity measurements, the sample with coarser material can measure substan-tially higher in TSS (Nichols, 1986). Particle size may vary less in streamsaffected by placer mining because of effluent treatment, which is usually inthe form of settling ponds. Settling ponds do a poor job of removing part-icles smaller than 25 microns (Dames and Woore, 1986), and because finerparticles are also most responsible for turbidity, placer-mined streams mayexhibit less variability from differences in particle size.
A consideration of the sources of error in turbidity and TSS measure-ments is necessary for developing a relationship between turbidity and TSS.Nichols (1986) identified four sources of error: (1) error in sample collec-tion; (2) subsample error; (3) error in turbidity analyses; and (4) error inTSS analyses.
The first source, 'error in sample collection,' refers to whether thesample collected is representative of the whole stream cross section; thiscategory is not applicable to the project reported here. Development ofregression equations require only that TSS and turbidity samples be taken atthe same time and at the same location, regardless of whether samples arerepresentative of an entire cross section.
-3-
The second source, 'subsample error,' however, is important to theproject reported here. TSS and turbidity samples are commonly collected inbottles with a capacity in excess of what is needed for analysis, and sub-samples are then taken from these bottles for the actual analysis. Thesubsample error factor becomes most critical when samples contain coarseparticles, because these start settling immediately after a thorough shaking,and the subsample may not contain a representative proportion of the coarserparticles.
The third source, 'error in turbidity analyses,' has received the mostattention. Pickering (1976) recommended that the U.S. Geological Survey stopreporting turbidity because of measurement error. Nichols (1986) extensivelystudied this type of error. In the past, turbidity was measured by variousmethods which reported in similar, but not identical, units. Nephelometry isnow the standard method and is used in Alaska for measuring turbidity inplacer-mined streams. Although nephelometry is the only method used, severalbrands --and models within brands --of nephelometric turbidimeters are used,and there is concern that these instruments do not report identical results.Nichols (1986) tested three turbidimeters on replicate samples from a placer-mined stream and found the results varied from 6 to 20 percent between instru-ments. For each set of replicates, the coefficients of variation for theinstruments ranged from 1 to 15 percent. Rounding data according to standardmethods (APHA, 1985) may help reduce error due to variation in turbidimeterbrand or model (Peterson and others, 1985).
The fourth source, 'error in TSS analyses,' appears to be attributablemainly to subsample error (Nichols 1986). Paralleling turbidity variabilitytrials cited above, Nichols also tested TSS variability of replicate samplesand found higher coefficients of variation for TSS replicates (10 to 33 per-cent) than for turbidity (2 to 10 percent) between corresponding replicatesets.
In spite of problems in relating TSS to turbidity, numerous attemptshave been and continue to be made to relate the two parameters. Lloyd (1985),Peterson and others (1985), and Nichols (1986) have summarized the attemptsof others, and Lloyd and Nichols have added their own equations. It isapparent from viewing the equations and their graphical representations thatno one equation best describes the TSS-turbidity relationship (Peterson andothers, 1985). Nichols found a statistical rationale for the common practiceof using a logarithmic transformation of the data and commented that althoughall authors report the coefficient of determination (r'), few give an esti-mate of the equation error. Both Nichols (1986) and Peterson and others(1985) caution that although turbidity-TSS equations can be useful, the errorassociated with the correlation must be known. Scatterplots of the data mustbe analyzed to determine if data are clustered into discrete groups, and therelationship should be periodically updated. The regression model must con-sider drainage, season, and discharge and is best based on data from similarsources, such as glacial streams or placer-mined streams (Peterson and others,1985).
Nichols (1986) tested these recommendations on a placer-mined streamnear Fairbanks. Collecting samples above mining, directly below sluicing,
- 4.-
and below settling ponds, he found the da,ta clustered in distinct groups.Regression equations for the clusters predicted TSS with average errors of25 to 30 percent, a result which compares well with those of other investiga-tors. The error associated with predicting individual TSS concentrationsfrom turbidity was much higher---600 to 1,700 percent.
The investigation reported here follows the work of Lloyd, Peterson andothers, and Nichols. A quantity of data exists, collected by several investi-gators from several sites in interior Alaska, and, although the experience ofother investigators indicates that equations from different areas differstatistically, it was hypothesized that because placer mining is essentiallysimilar throughout interior Alaska, equations predicting TSS from turbiditymight be similar enough to formulate one equation for the entire area or forthe area within a single basin. By organizing data on a geographic basis,using the computer to generate site, stream and basin-specific equations, andapplying appropriate statistical techniques, one might determine to whatextent historical data can be used and whether the concept of one predictiveequation has merit.
In natural streams with no large point source of sediment such as placermining, a positive relationship exists between sediment concentration anddischarge or velocity (Leopold and Haddock, 1953). In streams affected byplacer mining, the point source input from sluicing operations overwhelmsthis balance to the extent that dilution from extreme events may result in anegative relationship. However, in such streams, sediment settles from thewater column onto stream bottom during low flows and resuspends during highflows, which affects the turbidity-TSS relationship. All other things beingequal, particle size distributions in the water column will vary with flow,and coarser particles will be suspended at higher velocities. Becauseturbidity-TSS relationship is affected by changes in particle size distri-butions within the water column, variation in particle size distributionsover a wide range of flows may introduce considerable error into a simpleregression which uses turbidity as the predictor variable. To investigatethis, I constructed a multiple regression model using turbidity and velocityvariables to predict TSS.
Discharge data containing information needed to estimate velocity wereavailable for many observations from the Crooked Creek basin, but investi-gators have not routinely measured discharge during water quality sampling,so multiple regression could not be applied to the entire database. Velocitywas used as a variable in order to combine observations from different sitesand construct a basin model.
METHODS
Sources of Data
Eight data sources were used in the development of the project database:
1. Alaska Division of Geological and Geophysical Surveys (DGGS)placer mining research program (Mack and Moorman 1986);
-5-
2.
3.
4.
5.
6.
7.
8.
United States Environmental Protection Agency (EPA) STQRETdatabase (USEPA 1985) ;
Alaska Department of Environmental Conservation (DEC), Environ-mental Quality Monitoring and Laboratory Operations data from1983-85 (ADEC 1984, ADEC 1985, Hock 1986);
Alaska Department of Fish and Game (ADF&G), Habitat Divisionmiscellaneous data from 1983-5 (Weber 1985);
‘Fairbanks Area Ambient Water Quality Study, Placer RelatedB a s i n s , 1 9 8 4 , ’ ( d r a f t ) , J e r r y H i l g e r t , I n s t i t u t e o f N o r t h e r nForestry (INF), USDA;
‘Placer Mining Wastewater Settling Pond Demonstration ProjectReport, ’ R&M Consultants, Inc., 1982;
‘Placer Mining Wastewater Treatment Technology Project,’ Phase 2Report, Shannon & Wilson, Inc., 1985; and
data collected by the Alaska Cooperative Fishery Research Unit(ACFRU) investigators f o r s e v e r a l p r o j e c t s d u r i n g 1 9 8 2 - 8 3(Wagener 1984).
T h e t o t a l d a t a b a s e o f o v e r 1 , 1 0 0 o b s e r v a t i o n s d o e s n o t c o n t a i n a l lavailable data. Data collected directly below a sluice or pond outlet wasn o t i n c l u d e d , b e c a u s e p a r t i c l e s i z e d i s t r i b u t i o n s a f f e c t t h e t u r b i d i t y - T S Sr e l a t i o n s h i p a s l a r g e r p a r t i c l e s s e t t l e o u t i n s e t t l i n g p o n d s a n d i n t h estream channel. By avoiding data so directly affected by mining, the effectof particle size distributions was minimized. No data from R&M Consultants(1982) were used, and other data sources - - - p a r t i c u l a r l y Shannon and Wilson(1985) ---were scrutinized to make certain that only data from sites 500 ft orfarther from mining operation outlets were included in the database.
The EPA STORET database contains sample replication where, in someinstances, an investigator collected multiple samples within a short timespan. Because of concern that replicates might bias the results toward thereplicated samples, only data from the first sample was included when sampleswere taken less than 30 min apart by the same investigator at the same site.Even with this restriction, the database is not temporaliy homogeneous. Muchof the data came from intensive, short studies at sites where, for example,samples might be collected on a 3-hr basis for 3 days. Because of thediurnal change in turbidity and TSS below a mining operation due to startingand stopping of work, a range of values will be included; but it must beassumed that the relationship present for this short time did not varythroughout the operating season. These types of data are mixed with observa-tions taken on a daily or weekly basis, or miscellaneous samples that werenot part of a systematic monitoring program.
Paired turbidity-TSS data not determined from weighing a dried filterwere not used in development of the equations. TSS data reported by Wagener(1984) were calculated from total solids, using a conversion developed from
-6-
conductivity. Although this is a standard method, I felt that inclusion ofthese data might introduce additional error to the equations. Wagener's datawere used later to check the predictive value of the equations.
Considerable scatter can exist in the reported data at lower levels ofturbidity and TSS. Figure 1, a plot of turbidity and TSS from Eagle Creekabove and below mining, is a vivid demonstration. It shows well the cluster-ing described by Nichols (1986). When these data are combined, the samplecoefficient of determination (r') value (0.952) is high; however, a correla-tion based only on data from sites above mining operations results in a poorr2 value (0.031). A correlation analysis based on data from sites downstreamfrom mining operations results in a poorer r2 (0.837) than the combined data,but the equation is more descriptive of the turbidity-TSS relationship withinplacer-mining areas, and the equation error is less. In this instance, thestandard error of estimate (SEE) for combined data is 0.412 (+158, -61 per-cent), and for data from sites below mining activity SEE = 0.115 (+30,-23percent).
A problem arose in using data from different sources, because of differ-ing TSS reporting procedures among laboratories. Various labs reported lowTSS values to within one to three significant figures; thus, for differentlabs, 1 could be equivalent to 0.6 or 1.4, which, in turn, could be equiva-lent to 0.56 or 1.44. This was further complicated by varied lower detectionor reporting limits. Detection limits for data used in this study rangedfrom 0.01 mg/L to 4 mg/L. Because 4 mg/L is a high detection limit for clearstreams, considerable scatter can be introduced when paired with turbiditydata reported to the nearest hundredth, down to 0.01 NTU. Less variabilitywas noticed in the reporting procedures for turbidity. These reportingproblems may not greatly affect the sample coefficient of determination, butmay affect the equation error.
Because of the reporting and clustering problems with lower value ob-servations, the database used for regression analyses included only thoseobservations with turbidity greater than 5 NTUs. Although admittedlyarbitrary for the purposes of this project, 5 NTUs is a justifiable limit,because it is the background turbidity drinking water supply standard for theState of Alaska (ADEC, 1979). Deletion of observations with turbidity lessthan 5 NTUs reduced the database to 885 observations.
Geographical Organization
Investigations were conducted mainly in placer-mining areas accessibleby road, near Fairbanks and along the Steese, Elliot, and Dalton Highways.Streams in these areas eventually drain into the Yukon River via the Tananaand Koyukuk Rivers and Birch Creek. Major drainage basins used in the studyare described in the draft U.S. Geological Survey Hydrological Unit Map ofAlaska (USGS, 1985); smaller basins were delineated where data were available.Seven basins were selected: Birch Creek, Crooked Creek, Chena River, ChatanikaRiver, Goldstream Creek, Upper Tolovana River, and Koyukuk River (fig. 2).Analysis was broken down further to creeks and rivers within the basins, andto sites on those creeks.
-7-
;mE
I:li
I
il-l0)
ulull-
l-5510000~ 0
1 0 0 0 . 0
1 0 0 . 0
1 0 . 0
:
1+ * 034+
*
OQ 40 o
l- o- 0 0 0”L-EGENOm
88oo
0
0 00 oooao0 0 0 crbov=
0 0+ + * b-low
0 . 1 0 0 0 0 0 0I s I I I 8
-I-ur-bidity valu-r i n N - l - U 11 0
1 0 01 0 0 0
0 1 0 0 0 0
1 0 0 0 0 0
Figure 1. Plot of turbidity and TSS above and below mining, Eagle Creek inBirch Creek basin.
Statistical Methods
Statistical methods employed for this project included logarithmictransformation of data, simple and multiple linear regression, coefficient ofdetermination, standard error of estimate, and analysis of covariance models.Turbidity and TSS values were transformed to logarithms for regressionanalyses. The wide range of values displayed well on a logarithmic scale,and an initial plot of the data on linear scale showed a power curve thatappeared straight on a logarithmic scale. Nichols (1986) investigated therationale behind logarithmic transformation of data in the development ofturbidity-TSS relationships, and his residual analysis indicated that alogarithmic transformation of both turbidity and TSS best fit the data.
Linear regression uses the relation between two or more variables topredict one from the other(s) (Neter, Wasserman, and Kutner, 1985). A simplelinear regression model is expressed in the equation y = a + b(x), where 2 isthe predictor variable (in this case, turbidity), y is the response variable(TSS), b is the slope of the line, and 2 is the y-axis intercept. Becausethe analyses were performed on log transformed data, the regression equations
-8-
6--K
I-
Figure 2. Interior Alaska Basins with placer mining data.
can be expressed as power functions in the form of 11. = ax($), where theterms are defined as above.
The coefficient of determination (r2) and standard error of estimate(SEE) indicate how well the regression equation fits; r2 can be interpretedas the proportionate reduction of variation in the response variable assoc-iated with the predictor variable. It always lies between 0 and 1; thecloser to 1, the greater the linear association between the two variables(Neter, Wasserman, and Kutner, 1985).
The r2 indicates how well two variables are linearly associated but doesnot show how much error would be involved if the model were used for predic-tive purposes. Since the predictive value of the turbidity-TSS relationshipis of primary importance to this project, error analysis is crucial.Standard error of estimate (SEE) is one way of reporting error. SEE, thepositive square root of the regression model mean square error, is an esti-mator of regression model standard deviation (Neter, Wasserman, and Kutner,1985). For this project, SEE, reported in percent, was used as an estimatorof standard deviation for the predicted TSS for any turbidity value.Appendix B describes calculation method and contains sample calculations.
In order to determine to what extent data from different areas can becombined to develop useful predictive equations, it was first necessary todetermine whether the predictive regression equations for different groups ofdata (for example, data from different basins) were similar at a specifiedconfidence level. To determine the similarity of data from different groups,a covariance model was developed by adding qualitative indicator variablesfor each data group, and tested to determine if indicator variables improvethe model. The assumption was that if indicator variables do not improve themodel, they are not needed, and the data can be combined to develop oneequation. Covariance analysis assumes (1) independence of observations,(2) normality of residuals, and (3) common variability of the points aroundthe individual regression lines. Data used for this project were independentobservations. The latter two assumptions were not studied but were assumedto hold. Appendix B contains a more detailed description of covarianceanalysis.
The calculations were performed on the University of Alaska-FairbanksVAX computer using the GLM (general linear model) procedure of the SASstatistical package (SAS, 1985a,b). Both turbidity and TSS were transformedinto base-10 logarithms, and all analyses were performed on transformed data,All pairs had site, stream, basin, collection date, and source descriptors toenable analysis on any of these. Geographical descriptors were based on theUSGS hydrologic unit map and hierarchical in nature, which allowed analysisof subbasins or streams within larger basins.
Model Validation
Following the statistical practice of Neter, Wasserman, and Kutner(1985) to measure the predictive value of a model with data not used in themodel development, paired data from placer-mined streams in interior Alaskawhich had not been included in the principal database were used to measure
- 10 -
the predictive ability of the equations. DEC fiscal year 1986 placer-miningdata from the 1985 summer (DEC, 1986) and Alaska Cooperative Fishery ResearchUnit data from the 1983 summer (Wagener, 1984) were used. TSS was estimatedfrom turbidity values reported by those researchers, by using the mostappropriate regression equation indicated from analysis of covariance. Re-sults were compared with reported TSS, and a Z score was calculated by divid-ing the difference between the reported and predicted TSS by the regressionequation SEE. The Z score gives a relative measure of how close, inmultiples of SEE, the predicted value is to the reported value. A negative Zscore means the model overpredicted.
Velocity-turbidity Multiple Regression Model
Velocity estimates were available for 76 paired turbidity-TSS observa-tions from the Crooked Creek basin, including 16 observations on CrookedCreek at Central. These estimates were developed from staff gage readings byusing velocity rating curves. Multiple regression models and accompanyingstatistics were developed using the GLM procedure of the SAS statisticalpackage (SAS, 1985b).
RESULTS
Summary Statistics
The complete database used for this project contains 1,100 observationsfrom approximately 140 sites in seven basins: Birch Creek (excluding CrookedCreek), Crooked Creek, Chena River, Chatanika River, Goldstream Creek, UpperTolovana River, and Koyukuk River (app. A).
Regression equations used only those observations where turbidity wasgreater than 5 NTU. Of these 885 observations, 552 observations (62 percent)came from 18 individual sites which had 15 or more observations, and 766observations (87 percent) came from 15 streams with 15 or more observations.Summary statistics for these sites and streams are presented in table 1. On7 of the 15 streeams (Eagle, Gold Dust, Deadwcod, Ketchem, Mammoth, Gilmore,and Goldstream Creeks), 70 percent of the observations came from one of the18 individual sites (above), and on 4 (Crooked and Fish Creeks, and Chatanikaand Tolovana Rivers) over 70 percent came from 2 or 3 sites with 15 or moreobservations. Even though the observations came from a large geographicarea, most data came from relatively few sites on a few streams. Investiga-tors from other agencies and consulting firms also use these road-accessiblesites.
The Koyukuk River basin was an exception ---probably because of its dis-tance from Fairbanks. No stream in this basin had even 10 observations.Existing data were mainly from sites along the Dalton Highway.
Figures 3 through 10 present plots of paired observations groupedaccording to stream or site location. None of the stream data exhibit thedefinite cluster pattern demonstrated by figure 1, but the site data do showa more clustered pattern. Figure 9 points up the problem with using datafrom different sources. The data from Fish Creek below Lucky 7 were
- 11 -
Table 1. Summary statistics for streams and sites with 15 or moreobservations.
Turbidity Total suspended solids(in NT&) (w/L)
Location Na- Mean SDb Max Min Mean SDb Max Min-I_
A. Birch Creek Basin
1. Lower Birch Cra. Birch ab
Crooked Cr
44 39.2 46.6 240 6.4 75.1 138 770 12.716 15.08 9.48 32 6.4 71.6 187 770 14.8
2. Eagle Cra. Eagle b GHD
3. Gold Dust Cra. Gold Dust
b GDM
4746
1818
17701654
15901590
1150860
12201220
70003500
50005000
130130
100100
1450 14401312 695
1180 9471180 947
10000 853190 85
3040 523040 52
4. Upper Birch Cr 16 739 542 2100 270 872 688 2640 244
B. Crooked Creek Basin
1. Crooked Cra. Crooked Cr
at Centralb. Crooked Cr
ab mouth
9638
19
459 412663 482
33 392 36133 564 417
134 68.1
19001900
310 60 110 55.9
1530 371532 37
250 55.2
2. Deadwood Cra. Deadwood Cr
at CHSR
36 875 991 3500 45 1540 1540 5980 2332 866 995 3500 45 1559 1569 5980 23
3. Ketchem Cra. Ketchem Cr
at CHSR
22 1640 1700 5100 110 2600 3200 9300 97.620 1737 1750 5100 210 2800 3290 9300 97.6
4. Mammoth Cra. Mammoth Cr
at Steese
32 383 324 1300 16 493 457 1810 8827 380 286 1200 50 496 459 1810 88
5. Porcupine Cr 34 167 162 750 23 186 270 1470 16.5
baNumber of observations.Standard deviation.
- 12 -
Location
C. Chena River Basin
1. Fish Cra. Fish Cr
b Gold Dredgeb. Fish Cr
b Lucky 7
D. Chatanika River Basin
1. Chatanika Ra. Chatanika R
at 39 mileb. Chatanika R
at Long Crc. Chatanika R
b Faith Cr
2. Faith Cra. Faith Cr
at Steese
E. Goldstream Creek Basin
1. Goldstream Cra. Goldstream Cr
b Fox
2. Gilmore Cra. Gilmore Cr
b BD Mining
F. Tolovana River Basin
1. Tolovana Ra. Tolovana R
at TAPSb. Tolovana R
ab West Fork
Table 1. Continued.
Turbidity Total suspended solids(in NTUs) &g/L)
Na Mean SDb Max Min Mean SDb Max Min- - P--P-- -
67 21422 16.5
43 623
225 1100 6.9 192 225 950 157.18 36 6.9 51 78.4 396 15
212 1100 45 271 242 950 20
151 40.2 51 310 5.1 52.2 82.2 500 215 12.7 14.9 65 5.1 10.5 10.2 32 2
53 21.4 20 95 6.2 20 22.8 100 3
56 74.6 68.1 310 6.2 102 113 500 6
27 215 498 2600 6.7 233 375 1890 1417 75.1 43.1 140 14 120 112 416 14
50 269 123 800 30 323 241 1400 3036 284 105 800 65 335 239 1400 140
50 1650 1100 5300 60 479 271 1300 2044 1810 1070 5300 280 506 273 1300 20
76 20.830 18.1
36 18
23.8 180 5.4 61.6 176 1400 7.210.1 40 6.1 39.1 43.6 238 11
9.16 38 5.4 33.9 19.2 83 13
collected by a consulting firm (R&M) for a summer-long project and reflect avariety of seasonal conditions. The data from Fish Creek below Gold Dredgewere collected by EPA researchers during a 3-day span and have a much tightercluster pattern.
- 13 -
1-SS10000I
Figure 3. P'lot of turbidity and TSS for streams in Birch Creek Basin.
TSS1 0 0 0 0
11 0 0 0
1 0 0
1 0
:
0 0 0 0* * + E3ONANZA0 0 0 C R O O K E D Cc 0 . DEAOWOOOA A A KETCHEM C. n . M A M M O T H+ + + P O R C U P I N E
l -, 1 I
.
I3 g+* rt* 0
0 0* *+ ++; **+@*P f l3
++e+ $ ,+*o *o *.D O 0 +
5 P 0
0 0 0* + * E A G L E C R0 Cl 0 COLD OUST5 0 0 L O W E R BZRC+ + + U P P E R EIXRC
S O SOD so00-T-l-II-bid1et-y vr=lllLrrr i n N - I - U
AA b
C. * +’ o 0A AC6 A
s S O 5 0 0 so00
l-L-l~-bidit~ "calL-J~- in N - l - U
Figure 4. Plot of turbidity and TSS for streams in Crooked Creek basin.
- 14 -
1 - S S1 0 0 0 0
1
1
;mE
1 0 0 0
1 0 0
1 0
0 + *0
* *+*
*O wq
+*c **,‘* **+
k**t/ *CL
** *a’ t* **+* * **
.
*
.A
0l05
I + Al
0 0l * Chatan i ka
0 0 Faith t-. D Flumr ct-A A ti lmcr-r or. . Goldrt~meam
11I I 1
s S O so0 so00
tur--bidity i n N- f - l - l
Figure 5. Plot of turbidity and TSS for streams in the Chatanika andGoldstream basins.
-I-ssoooo-
1 ooo-
O00 O
0
Figure 6. Plot of turbidity and TSS for streams in the Upper Tolovana, Chenaand Koyukuk basins.
- 15 -
rl
\mE
Ii
\
mE
-I-SS1 0 0 0 0
1 0 0 0
1 0 0
1 0
1
t*
0 El.;%3# cl+
*
;
0 0 0 C R O O K E O A sol-0* * + C R O O K E O h C E N0 0 0 C R O O K E D A M l - l - i
* I I IEi S O 500 5 0 0 0
t-r-bldity iI-! N - I - U
Figure 7. Plot of turbidity and TSS for sites on Crooked Creek.
-I-SS1 ooo-
1 OO-
lo-
l-
0
0
0
Oo
* 0 0
* *
AD** +
o* #O*
0 +O
0 0e 0 CHATANI'KA A 38l * * Cl-IA-l-AN I KA ALON0 0 0 CHATANIKA S ,=A
s S O so0
tul-bidity i n N - T - U
Figure 8. Plot of turbidity and TSS for sites on Chatanika River.
- 16 -
TSS1 oooo-
1 ooo-
1 oo-
;
r
1 o-
l-
0 0 0 FISH 0 GOLO ORt l 1 FISl4 0 L U C K Y 7
-Is SO 5 0 0 3 0 0 0
t-t-bidity in- N - t - U
Figure 9. Plot of turbidity and TSS for sites on Fish Creek.
4*000
0
0 0 0 TOLOVANA A TAP+ * + TOLOVANA A W,=
11 1 I
5 S O so0
turbidity in N - I - U
Figure 10. Plot of turbidity and TSS for sites on Tolovana River.
- 17 -
Regression Equations
Table 2 presents regression equation coefficients with descriptiveparameters for all sites and streams with 15 or more observations, for theseven basins, and for the combined interior Alaska database along with theresults of the analysis of covariance.
Table 2. Summary of regression equations and covariance analysis for basins,streams, and sites in interior Alaska.
[Equations in the,form y = a*(~~), where y = TSS, 2 = turbidity,a = v axis intercept, and b = slone. N = number of observations.]- L
Location N-
885
133
441647461818
-
a-
2.317
2.630
3.5402.1581.4162.0461.2591.259
.
b r2-
0.813
0.899
0.4680.3720.8470.8370.6710.671
+SEE(%) -SEE(%) F*<F?l
Interior Alaska
Birch Cr Basin
1. Lower Birch CrBirch Cr ab CC
2. Eagle CrEagle Cr b GHD
3. Gold Dust CrGold Dust Crb GDM
4. Upper Birch Cr
Crooked Cr Basin
1. Crooked CrCrooked Cr
ab BoulderCrooked Cr
at CentralCrooked Cr
ab mouth2. Deadwood Cr
Deadwood Crat CHSR
3. Ketchem CrKetchem Cr
at CHSR4. Mammoth Cr
Mammoth Crat Steese
5. Porcupine Cr
1
0.851
0.840
0.7311.0140.9240.8710.9110.911
112
75
1041193330102102
53 no
43 yes
515425235151
16 1.249 0.989 0.944 17 15
239 2.000 0.900 0.730 103 51 no
96 3.589 0.748 0.553 73 42 Yes9 0.032 1.504 0.549 23 19
38 14.655 0.535 0.261 123 55
19 2.178 0.821 0.256 97 49
36 5.012 0.859 0.767 82 4532 4.656 0.863 0.769 86 46
22 1.982 1.028 0.839 82 4520 1.406 0.999 0.863 74 43
32 10.328 0.638 0.711 52 3427 1.858 0.928 0.808 40 28
34 0.713 1.044 0.696 81 45
'A 'no' in this column indicates that the equations which, when combined,would make up this geographical unit are statistically different at the95 percent confidence level. For example, the 'no' for the interior Alaskaequation indicates that the basin equations within interior Alaska arestatistically different from each other. A 'yes' indicates the equationsare statistically similar.
- 18 -
Location
Chena River Basin
1. Fish CrFish Cr
b Gold DredgeFish Cr
b Lucky 72. Little Chena
Chatanika R Basin
1. Chatanika RChatanika R
at 39mChatanika R
at LongChatanika R
b Faith2. Faith Cr
Faith Crat Steese
Goldstream Cr Basin
1. Goldstream CrGoldstream Cr
b Fox2. Gilmore Cr
Gilmore Crb BD Mining
Upper TolovanaRiver Basin
1. Tolovana RTolovana R
at TAPSTolovana R
ab West Fork2. Livengood Cr
Koyukuk R Basin
N
96
6722
43
14
186
15115
53
56
2717
a b
3.311 0.771
5.598 0.6301.153 1.261
1.315 0.879
0.124 2.108
0.932 1.034
0.729 1.0980.771 0.965
0.473 1.179
2.280 0.844
1.770 0.9300.611 1.186
r2-
0.648
0.6290.627
0.370
155
10755
124
95
90
88115
47
85
5657
61
5235
55
0.782
0.789
0.7430.418
0.803
0.610
0.8810.787
49
47
4754
32
46
3636
112 5.808 0.651 0.602 97 49
50 5.781 0.694 0.320 76 4336 1.274 0.967 0.385 52 34
50 4.560 0.627 0.657 51 3444 0.848 0.852 0.719 44 31
88 1.500 1.083 0.841 35
76 1.233 1.157 0.77830 1.419 1.088 0.673
36 3.126 0.814 0.722
12 1.871 1.015 0.882
31 5.768 0.867 0.635
53
5053
34
74
140
3335
25
43
58
Table 2. Continued.
+SEE(%) -SEE(%) F*<F?l
no
no
yesno
no
yes
yes
Figures 11 through 18 show regression lines plotted by basin and streamlocation. The regression which included all 885 observations had a coeffic-ient of determination of 0.813 but a standard error of estimate of+112 percent (-53 percent). Coefficients of determination for the basinequations ranged from 0.602 (Goldstream Creek basin) to 0.899 (Birch Creekbasin). Four of seven equations had standard errors of estimate less than+lOO percent.
- 19 -
1 0
1
1
0 0 0
1 0 0
t E1RCt-l A L L1 0 ; CHATANIKA R I V E R
-8 n 6- CHENA R I V E R- C R O O K E D C R4 COI-OS-I-REAM C R: KOYUKUK R I V E R
TOLOVANA R I V E R
1-iI 1 1 1
s S O so0 5 0 0 0
turbidity i n N T U
Figure 11. Plot of turbidity-TSS regression lines for seven basins ininterior Alaska.
I
1 0
1t-4\InE
- L O W E R SIRCt-42 U P P E R l3IRCW
11,
I 1 I
9 S O 5 0 0 5 0 0 0
t u r b i d i t y i n N - I - U
Figure 12. Plot of turbidity-TSS regression lines for streams in Birch Creekbasin.
- 20 -
ki
\ulE
T-651 oooo-
1 ooo-
loo-
lo-
l-
A- KETCHLM C R: M A M M O T H C R
P O R C U P I N E C R
5 30 m o o moo0
~ur-bidity im N-I-U
Figure 13. Plot of turbidity-TSS regression lines for streams in CrookedCreek basin.
-I-SS1 0 0 0 0
1 0 0 0
1 0 0
1 0
1
Figure
a = - OILMORE CR5 GOLOSTREAM C R
s s o so0 5 0 0 0 .
turbadity %&-I N-I-U
4. Plot of turbidity-TSS regression lines for streams in Chatanikaand Goldstream basins.
- 21 -
l o -
-+ : '; F1St-i CR2 LITTLE CHENA R I V3 I-IVENGOOO CR: TOl-OVANA R I V E R
11I 1 8 I
1
5 S O SO0 5 0 0 0
fur-bidity *r, N-l-U
Figure 15. Plot of turbidity-T%? regression lines for streams in the UpperTolovana and Chena basins.
lo-
BOUI-DE+- C R O O K E D A CENTRA
C R O O K E D A MOUTII
l-I I 1 !5 5 0 5 0 0 so00
tul-bic9ie.y i.t-l N - l - U
Figure 16. Plot of turbidity-TSS regression lines for sites on Crooked Creek.
- 22 -
A
\IJlE
: C H A T A N I K A A JWMC C H A T A N I K A AC C H A T A N I K A A LONGLONG
-8-8 - C H A T A N I K A w- C H A T A N I K A w F A I TF A I Tl-l-
II II ,, r-r-!?5 S O 500 5 0 0 0
turbidity lr, N - l - U
Figure 17. Plot of turbidity-TSS regression lines for sites on ChatanikaRiver.
l -
+ F1St-i W GOLO O R E O+ $ 4- FISH W L U C K Y 7
B- TOI-OVANA A T A P S- TOLOVANA A Wf=
s 5550 500 z5000
turtai.dity II-I N - I - U
Figure 18. Plot of turbidity-TSS regression lines for sites in the UpperTolovana River and Chena basins.
- 23 -
For the stream data equations, the equation coefficients and regressionparameters ---coefficient of determination (r2) and standard error of estimate(SEE) varied considerably; r2 ranged from 0.320 (Goldstream Creek) to 0.996(Upper Birch Creek), and r2 in 13 of 15 equations was over 0.50. SEE variedfrom +107 percent (-52 percent) with Fish Creek data to +17 percent(-15 percent) for Upper Birch Creek, and +SEE value was less than 100 percentin 12 of 15 equations.
The variation of equation descriptors (r2, SEE) for site equations wassimilar to that of stream equations; r2 ranged from 0.262 at Crooked Creek atthe bridge to 0.863 at Ketchem Creek at the Circle Hot Springs Road. In13 of 18 equations, r2 was over 0.50. Other sites with relatively poor r2values were Birch above Crooked Creek (0.372), Fish Creek below Lucky 7(0.370); Chatanika at 39 mile Steese (0.418), and Goldstream below Fox(0.389).
SEE for site equations ranged from +30 percent (-23 percent) to+124 percent (-55 percent) and 13 of 18 were less than +lOO percent. Aninverse relationship generally existed between SEE and r2 for the site equa-tions; that is, equations with the lowest r2 had the highest SEE. Figure 19,a plot of coefficients of determination and corresponding standard errors ofestimate for site and stream equations, demonstrates the scatter that oc-curred with these equations. No general conclusion can be drawn aboutwhether combination of data into stream equations improved, reduced, oraveraged the regression parameters.
Analysis of Covariance
For streams with two or more sites, for basins with two or more streams,and for all interior Alaska data, analysis of covariance was performed. Theresults of this work are presented in column 8 (F*cF?) of table 2. A ‘yes’in this column indicates that the equations describing the data groups in-cluded in the covariance analysis were statistically similar at the 95 per-cent level, and that the equation describing the combined data would there-fore be the most appropriate.
The analysis of covariance results were mixed; the seven basin equationsfor interior Alaska were statistically different , which indicated that thesedata should not be combined to develop one equation. At the basin level, thefour streams in Birch Creek, the two streams in the Chatanika River basin,and the two streams in the Upper Tolovana River basin had statistically simi-lar equations for each basin. The six streams in the Crooked Creek basin,the two streams in the Chena River basin, the two streams in the GoldstreamCreek basin, were statistically different for each basin. At the streamlevel, the F value comparison indicated that the three sites on Crooked Creekand the two sites on the Upper Tolovana River had statistically similar re-gression equations. The three sites on the Chatanika River and the two siteson Fish Creek were statistically different.
Of note is the reversal in the Chatanika River basin, One might expectsites on one stream to have similar regression equations if the total streamequation were similar to that of a tributary stream. That was not the case
- 24 -
+,n0
uF c0 0
U
L L0 0L QL0 c
-.‘DLuT)c0v
n
160 -I
1 3 0 -
1 0 0 -
7 0 -
I LOCA-I- I O N
*0
0 *
.O * 0* * *
*
0
* o *
40 0 *
*O l *0
04 0 - 0
00 *
0*
-m-b- -___
Figure 19. Plot of regression equation coefficients of determination andstandard errors of est imate.
with the Chatanika River. Covariance analysis indicated that the regressionequations for the sites on the Chatanika River were different, yet the equa-tion for the combined data from the Chatanika River was not significantlydifferent from the equation for Faith Creek. When only 1984 data were used,regressions for the Chatanika River sites were statistically similar, butwhen the 1983 data were included the difference occurred.
Whether regression equations are similar between sites, streams andbasins was a central question for this project. Also of interest was whetherregression equations are similar between years. Does the equation developedfrom data collected in 1983 and 1984 accurately predict in 1985? Covarianceanalysis was used to investigate whether the equations for the combined data-base and equations for site data from Crooked Creek at Central, ChatanikaRiver below Faith Creek, and Chatanika River above Long Creek differed be-tween years. The results, presented in table 3, show that regression equa-tions can differ statistically from year to year. When all data were com-bined, the regression equations for each year (1983-85) were different. How-ever, earlier analysis demonstrated that one should not combine data fromdifferent basins. To rule out the possibility that the difference by year ofthe combined data might be a function of basin differences, three individualsites--Crooked Creek at Central, Chatanika below Faith Creek, and Chatanikaat Long Creek---were investigated. Covariance analysis based on year showed
- 25 -
Table 3. Summary of covariance analysis by year.
[Equations in the form 41 = g*(sb), where y = TSS, x =-a = y axis intercept, and b = slope.]-
Location (yr) N a b r2- +SEE (X) -SEE(X) F*<F?!
Int. Alaska (all) 885 2.317 0.851 0.81Int. Alaska (83) 1 5 8 0.689 1.082 0.92Int. Alaska (84) 543 3.236 0.799 0.80Int. Alaska (85) 1 8 4 2.871 0.820 0.74
1 1 2 53 no56 36
1 1 9 5 4101 50
Crooked Cen (all) 38 14.655 0.535 0.26 123 55 n oCrooked Cen (84) 19 234.423 0.156 0.04 1 2 1 55Crooked Cen (85) 1 9 2.009 0.831 0.41 8 7 47
Chat b Faith (all) 56 2.280 0.844 0.61 8 5 48Chat b Faith (83) 3 2 1.611 0.894 0.77 3 4 2 5Chat b Faith (84) 2 4 2.553 0.865 0.62 1 3 7 58
y e s
Chat a Long (all) 5 3 0.473 1.179 0.80 4 7Chat a Long (83) 2 8 0.514 1.092 0.55 3 3Chat a Long (84) 2 5 0.813 1.055 0.82 5 2
32 no2534
turbidity,
that the regression equations for Chatanika at Long Creek and Crooked Creekat Central were different, whereas regression equations for Chatanika belowFaith were similar (figs. 20 and 21).
Model Validation
Model validation was done with 1985 data from the Chatanika and Tolov~naRivers and Goldstream Creek (DEC, 1986) and 1983 data from Upper Birch Creek,Crooked Creek, and Chatanika River (Wagener, 1984). Appendix C presents theresults of these comparisons. Figure 22 is a histogram of Z scores for 1989Chatanika and Tolovana DEC data and 1983 data reported by Wagener (1984).The Chatanika data had an average Z score of -1.07; 55 percent of theobservations were within one standard error of estimate and 98 percent werekwithin two standard errors of estimate of the reported values. The Tolovanadata had an average Z score of -0.20, with 89 percent within one standarderror of estimate and 95 percent within two standard errors of estimate ofreported values. The 1983 data had an average Z score of 0.56, with 58 perr-cent within one standard error of estimate and 88 percent within two standarderrors of estimate of the reported values.
The disparity between the 1985 Tolovana and Chatanika results was note-worthy. These data were collected by the same people using the same methodsduring a 2-wk period, Results from the 1983 data were underpredicted, onaverage, and distribution was spread out more than in the other two groups ofdata.
- 26 -
TSS1 0 0 0 0 . o-
1 0 0 0 . o-
* l*
8loo- o- * o* $ll
** l
O *p
* l o*f8
t1 0 . o-
l **
YeEAR1 . o-
0 0 0 8!33
* * 8 04
0 . 11 I I 1 I I
-I-SS10000~ o-
1000~ o-
lOO.O-
1 0 . o -
*0 Mgg l
l *�,�c Ot
** d
*
0
-1
1
b
11 0
1 0 01 0 0 00 0 0 0
- I0 0 0 0
turbidiey in N - I - U
l- o-
0 . l -
1 k1 0 0
1 0 0 00 1 0 0 0 0
-1 b 0 b b b
tur-taidiizy i n N - i - U
Figure 20. Plot of turbidity and TSS by year, Chatanika River below FaithCreek, 1983-84.
Figure 21. Plot of turbidity and TSS by year, Crooked Creek at Central,1984-85,
- 27 -
-------___-_-
OA
0.0
QA
1;
53 0 . 3
s
‘1
Ol-
O.l-
o.o-
Figure 22. Z score distributions for 1985 Chatanika and Tolovana and 1983ACFRLJ data.
The Chatanika data came mostly from two sites---Chatanika below FaithCreek and Chatanika at Long Creek ---which had different Z score distributions.At the site on Chatanika below Faith Creek, 92 percent of the Z scores (22 of24) fell within the greater thansite on Chatanika at Long Creek,
-1.0 and less than -0.5 interval and at the
-1.0. In particular,81 percent of the Z scores were less than
the predicting equation for the site on Chatanika belowFaith Creek may not be accurate for this set of data, but theprecision--- 92 percent within one Z score interval---was good.
Velocity-Turbidity Multiple Regression
Velocity estimates were available for 76 observations within the CrookedCreek basin.produced an r2
Simple regression of the log transformed turbidity and TSS dataof 0.82 with an SEE of 0.296 (+98,-49 percent). Velocity by
itself does not have significant relationship with total suspended solids.The multiple regression model with log velocity as the second predictor vari-able produced an r2 of 0.85 and an SEE of 0.271 (+87,-46 percent).not substantial improvements,
These arebut the added velocity variable is statisti-
cally significant at the 95-percent confidence level.
- 28 -
When only data from Crooked Creek at Central were considered, there wasmarked improvement. Multiple regression (log turbidity and log velocity)improved the simple regression (log turbidity) r2 of 0.207 to 0.686 andreduced the SEE from +98 (-49 percent) to +56 (-36 percent). Table 4 pre-sents a comparison of the multiple regression analyses.
DISCUSSION
The underlying premise of this project was that, because placer miningmethods are similar throughout interior Alaska, the turbidity-TSS relation-
ship in placer-mined streams in interior Alaska also may be similar and mayallow the use of one equation to define that relationship. This was notborne out by the analysis. Regression equations for the seven basins werestatistically different. Of six basins that had two or more streams with 15or more observations, only three produced statistically similar regressionequations and, in one of those, equations for the individual sites are notsimilar. Of four streams that had two or more sites with 15 or more observa-tions, two had statistically different regressions.
Covariance analysis also indicated that one should be careful usingequations developed from data of previous years to predict TSS. The equa-tions using all data from interior Alaska were different for 1983, 1984, and1985. Covariance analysis of three sites indicated that at two of thosesites the equations differed between years. Model validation supported thisuncertainty. Estimates from 1985 Chatanika River site data averaged morethan one standard error of estimate from reported TSS.
Error as indicated by the standard error of estimate is reasonable formost equations. Considerable variation may occur among individual observa-tions. Inspection of the data from the site equations with the worst errorterms showed that these sites were close to sluice operations or included
Table 4. Comparison of multiple and simple linear regression equations fromCrooked Creek basin.
[Equations in the form 2 = 2 * (~~bl)a~d(12~:~,c~~~~~c~~n~st~rbidity.
52= velocity, and El, -2, ._
Location N abl b2 r2 +SEE -SEE
- - - ~ - -
Crooked Creek BasinSimple regression (turb) 72 1.211 0.985 0.788 91 48Simple regression (vel) 72 134.896 0.165 0.005 305 75Multiple regression 72 0.851 1.016 0.456 0.828 79 44
Crooked Creek at CentralSimple regression (turb) 16 7.447 0.622 0.207 98 49Simple regression (vel) 16 210.863 0.073 0.002 114 53Multiple regression 16 0.001 1.919 2.127 0.686 56 36
- 29 -
data from a variety of flow conditions. It is important to note that theseequations cannot be used to estimate TSS outside the range of values in thedata sets used to develop the equations , particularly turbidity values lessthan 5 NTU. Also, these equations cannot be used to predict TSS in non-placer-mined streams.
Stream flow levels ---discharge or velocity---can affect theturbidity-TSS relationship over a wide range of flows. When velocity wasadded to the poor relationship at Crooked Creek at Central, the r2 improvedremarkably and the error was reduced equally well. Inspection of the datashowed much different turbidity-TSS relationships at high flows.Observations in May and early June showed TSS values much higher than theaccompanying turbidity values. Observations from Crooked Creek basin in lateJune and mid-August, 1985---times of high flows---revealed similarrelationships. Low flows in early August may partly explain the poorprediction performance of the Chatanika site equations on 1985 DEC data.Lack of measured or estimated discharge and velocity data limits a morethorough exploration of this. Addition of a discharge or velocity variableis essential for adequate prediction of TSS from turbidity over a wide rangeof flows, although a simple regression may be acceptable for average-levelsummer flows.
The research conducted here indicated that the most appropriate use ofregression models to predict TSS from turbidity in mined streams is on asingle site basis. Analyses indicated that regression equations should beused with care if developed for more that one site, if used on sites that didnot contribute data to the model development, or if used for years that didnot contribute data to the model development. A simple regression equationdeveloped with data collected during normal flows will underestimate TSS athigh flows and overestimate TSS at low flows. Analysis of covarianceindicated that the relationship may stay the same between years, sites, orstreams, but this constancy of relationship requires verification and cannotbe assumed,
A strong, if not perfect, relationship exists between TSS and turbidity;turbidity, as well as being much less expensive to collect, has a more en-forceable standard. Excess amounts of sediment which cause ecological andaesthetic damage can be accurately monitored or estimated by either para-meter, and this report has demonstrated a way to estimate sediment loads witha minimum amount of TSS analysis.
As state and federal funding declines and interest remains constant insolutions to the issue of water quality in placer-mining areas, funds to doall desired analyses may not be available. If water-quality monitoring inplacer-mined streams requires both turbidity and suspended sedimentinformation, then the turbidity-TSS models recommended here can help stretchthe analysis dollar.
CONCLUSION
The results of the analyses conducted in this report support the conclu-sion that equations are most useful in predicting TSS values from turbiditymeasurements when developed on a site basis. Combining all data from
- 30 -
interior Alaska into one equation is not supported by the analysis, nor iscombining data within a basin or stream. The turbidity-TSS relationship maychange from one year to the next. Multiple regression models using turbidityand velocity to predict TSS give improved results over a wide range of flows.
REFERENCES CITED
Alaska Department of Environmental Conservation, 1979, Water qualitystandards, Alaska Water Quality Pollution Control Program, Juneau,Alaska, 34 p.
Alaska Department of Environmental Conservation, 1984, Suspended solids inmainstem drainages downstream from placer mines, Fairbanks and vicinity,Alaska, August 3-17, 1983, Juneau, Alaska, 27 p.
1985, Suspended solids and turbidity in drainages subjected toimpacts from placer mines, Interior Alaska, August 7-16, 1984: Workingpaper for environmental quality monitoring and laboratory operations,Juneau, Alaska, 56 p.
APHA, 1985, Standard methods for the examination of water and wastewater:Sixteenth edition, Washington, D.C., 1268 p.
Dames and Moore, 1986, A water use assessment of selected Alaska streambasins affected by placer gold mining (draft): Prepared for the AlaskaDepartment of Environmental Conservation by Dames and Moore, ArcticHydrologic Consultants, Stephen R. Braund Associates, L.A. Peterson 6Associates, and Hellenthal & Associates, Anchorage, Alaska, 159 p.
Hach, Clifford C., and others, 1984, Understanding turbidity measurement:Technical Information Series - Booklet no. 11, Hach Company, Loveland,Colorado, 10 p.
Hilgert, Jerry W., 1985, Fairbanks area ambient water Quality Study (draft):Institute of Northern Forestry, USDA, Forest Service, Fairbanks, Alaska,27 p.
Hock, Jeffrey, 1986, Unpublished data from 1985 field season: Alaska Depart-ment of Environmental Conservation, Juneau, Alaska.
Leopold, Luna B., and Thomas Maddock, Jr., 1953, The hydraulic geometry ofstream channels and some physiographic implications: Geological SurveyProfessional Paper 252, United States Government Printing Office,Washington, D.C., 57 p.
Lloyd, Denby S., 1985, Turbidity in freshwater habitats of Alaska - A reviewof published and unpublished literature relevant to the use of turbidityas a water quality standard: Habitat Division Report 85-1, AlaskaDepartment of Fish and Game, Juneau, Alaska, 101 p.
Mack, Stephen F. and Mary Moorman, 1986, Hydrologic and water qualityinvestigations related to the occurrence of placer mining, summers,1984-85 (draft): Alaska Division of Geological and Geophysical Surveys,Fairbanks, Alaska, 29 p.
Neter, John, William Wasserman, and Michael H. Kutner, 1985, Applied linearstatistical models: Second edition, Richard D. Irwin, Inc., Homewood,Illinois, 1127 p.
Nichols, Gary E., 1986, Sediment and turbidity in the stream environment:Master of Science Thesis, University of Alaska, Fairbanks, Alaska,106 p.
- 31 -
Peterson, Laurence A., and others, 1985, Alaska particulates criteria review:Prepared for Alaska Department of Environmental Conservation by L.A.Peterson and Associates, Inc., Fairbanks, Alaska, 133 p.
Pickering, R.J., 1976, Measurement of "turbidity" and related characteristicsof natural waters: U.S. Geological Survey Open-file Report 76-153,Reston, Virginia, 7 p.
R&M Consultants, Inc., 1982, Placer mining wastewater settling pond demonstra-tion project: Prepared for the Alaska Department of Environmental Con-servation by R&M Consultants, Inc., Fairbanks, Alaska, 60 p.
Shannon and Wilson, Inc., 1985, Placer mining wastewater treatment technologyproject, Phase 2 Report: Laboratory study, pilot study, field study,prepared for the Alaska Department of Environmental Conservation byShannon and Wilson, Inc., Fairbanks, Alaska, 241 p.
SAS Institute Inc., 1985a, SAS User's Guide: Basics, 1985 Edition: SASInstitute, Cary, North Carolina, 921 p.
1985b, SAS User's Guide: Statistics, 1985 Edition: SAS Institute,Cary, North Carolina, 584 p.
U.S. Environmental Protection Agency, 1985, STORET data for placer minedstreams in Alaska, 1984-85:Region X, Seattle, Washington.
U.S. Geological Survey, 1985, Alaska hydrologic unit map (draft): U.S.Geological Survey, Fairbanks, Alaska.
Wagener, Stephen Mitchell, 1984, Effects of gold placer mining on streammacroinvertebrates of Interior Alaska: Master of Science Thesis,University of Alaska, Fairbanks, Alaska, 98 p,
Weber, Phyllis, 1985, Miscellaneous turbidity and TSS measurements collectedby Alaska Department of Fish and Game, analyzed by Northern TestingLaboratories, Inc.: Alaska Department of Fish and Game, Fairbanks,Alaska.
- 32 -
Appendix A. Turbidity and TSS data from interior Alaska streams
UXATXON HYUNXT SOURCE DATE TIME TURB TSS
FmRHmuAsTEFTARNIGANArnTWELmAwTTWELVMXLEAHT'fWELVE?4IUAMT'IUELVMXLE A KTTUEWPIILE BNFE;IiEWgkRC
CLUMSAMTHCLW A VOLCANOCROOKED A HARNGHARRXNGTON A MTEAGLEAGHDEAGLEAGHDEAGLEAGHDEAGLEACHDE'AGLEAGHDEAGLEAGHDEAGLE AGHDEAGLEAGHDEAGLEAGHDEAGtEAGHDEAGLE A GHDEAGLE A GHDEAGLEAGHDEXXEAGHDEAGLEAGHDEAGLE A GHDEAGLE A GHDEAGLEAGHDEAGLEAGHDJZAGLEAGHDJZAGLE A GHDEAGLEAGHDEAGLEAGHDEAGUAGHDEAGLEAGHDEAGLEAGHDEAGLEAGHDEAGLEAGHDEAGLEAGHDEAGLEAGHDEAGLE A GHDEAGLEAGHDEAGLE A GHDEAGLE A GHDEAGLEAGHDEAGLE A GHDEAGLEAGHDEAGLEAGHDEAGLEAGHDEAGLEAGHD
4040203404020340402034040203404020340402034040203
%i:::40402044040204404020440402044040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205404020540402054040205
i3
i
:44
::4
;222222222
s2
I22
2'2
I222
:22
s2
z22222
84-08-2284-08-30
84-08-21
85-06-10
iEt-284x21
84-06-22
2-EkO6-124
E-z-1:84:0711784-W-1884-m-18
84-07-21
84-08-10
Kz-1:8410811184-08-12
180012201916193812531035
ii::1315
1%150610481545a55
114015102115830
2::2030910
14401840850
1130135511001520193093012051640820
12001455835
11151600700
10301300120016301945830
12001425125515552020925
4.601.000.400.200.300.300.300.550.350.401.200.650.600.481.301.901.902.001.201.001.001.500.501.101.201.001.300.800.290.190.220.400.270.450.18
0":;0.190.420.230.23
00:2n30.590.210.750.280.420.820.450.470.320.38
2.80
7::4:oo4.000.402.00
i:G1.70
:*:i1:201.600.400.800.201.000.400.400.100.200.050.050.400.050.600.400.401.700.801.101.901.600.050.100.300.050.300.500.050.050.601.100.701.900.100.300.300.050.100.500.10
33 -
Appendix A. (Continued)
OBS LOCATION HYUNIT SOURCE DATE TIM TM3 ‘fSS
1 6 01611 6 2
l21651661671 6 81691 7 01711721731 7 41 7 51 7 61 7 71781791 8 0181
!E
z51871 8 81891901911921931941951961971981 9 9
2':202203
%:
2':2082092102 1 1212
GouJDuSTBmGOUDUSBQ)IGOLDWSTBQ)IGOLDDUSTBCZM
~~~;z
ii!iEE~~GOLDWSTBCWGOLD DUST B c;DHGOLDDUSTBCZMGOLDDUSTBGDHGQLDDUSTBQ)HGOLDDUSTBcalGOLDDlJSrBmCXLDDUSTBQ)MGOLD DUST B GMHARRISON A BIRCHARRISON A BIRCHARRISON A BIRCHARRISON A MHHARRXSONAMTHHARRISON A SQUAHARRIWN B SQUASq0A.U A HARRIS0BIRCH A 12 MILEBIRCH A 12 MILEBIRW A 12 MILEBIRCH A 12HILEBIRCH A 12tlILEBIRCH A BUTTE CBIRCHAGOLDDSBIRQi AB NF CONBIRCH B 12 MILEBIRCH B 12MILEBIRCH B 1WILEBIROI B 1WILEBIRCH B BEAR CBIRCH B NF CONBIRCH B S’TARHIGBIRCHBWILLIMCROOKED A ALBERCROOKED A BLDRICROOKED A BOLDRCROOKED A BOLDRCROOKED A BOLDRCROOKFJI A BOLDRCROOKED A BOLDRCROOKED A BOLDRCROOKED A BOLDRCROOKED A BOLDRCROOKED A BOLDRCROOKED A BOLDR
40402064040206404O2064040206
E;:4040206
E%4040206
ZE:40402064O4O20640402064040206404020640402O7.40402W4040207404020740402U?4O4O2O74O402U740402O74040208404020840402084040208404020840402084040208404020840402084O4020840402084040208404020840402084040208404020840402104O402104040210404021040402104040210404021040402104040210404021040402104040210
84-06-2484-08-20
EE84:oa2184-08-22
84-08-2483-08-0884-08-0884-08-08
84-08-09
iifz.c'$t
1715130016001720134018351240141018302015
:ii?:1300174018151115134015001500
1:ll;930
1635162516251438122016151258104512151648184514221915193013011300185011501345155013251743103016121 1 4714
1314
190::114714
2000.0380.0
1200.01700.0
ZE
;E:o"1200:o
'6E1OOo:o1800.01800.0loo.0100.0500.0240.0240.0190.0
456"* ii4oo:o420.0220.0
1000.0400.0450.0400.0270.01800.0650.0320.0
iii*:5oo:o320.0950.0280.0
2100.01100.0460.0380.01100.01400.01300,o360.0330.0340.0370.0370.0500.0450.0
125
12802180820
1670
3%
'%86 5
24402000
52100408
zi320745
2::11001200970420603
2%2640694360
$0"720410960244
23801150410
:,g:12001400
269241236248161398327
3 4 -
Appendix A. (Continued)
OBS UXATION HYUNIT -. -'SOURCE DATE TIX
266m268
271z:;:$22 8 1
sg28420526628728828929029’2922932942952%297298299
%302303304
:z
;:
3103113123133143 1 53 1 63 1 7318
CRooKEDAm 4040210CROCXEDAHl?l 40110210CROOKEDAUl% 4040210CRUXEDAHTH 4o40210CROOKEDAHTH 4040210CRWKEDAI(TH 4040210CROOKEDAMH 4040210CROOKEDAHTH 4040210CROOKEDAMTH 4O40210CROOKED A MH 4040210CROOKEDAMTM 4040210CROOKEDAMH 4040210CROOIEDAMl'HI 4040210CROOKEDAbt'IWI 4040210CROOKED A MTH I 4040210CROOKED A UBALB 4040210CROOKEDAWBALB 4040210CROOKED B ALBER 4040210CROOKED B BEDRK 4040210CROOKED B B&DRK 4040210CROOKED B BEDRK 4040210CROOKEDB BEDRK 4040210CROOKED B BEDRK 4040210CROOKED B BEDRK 4040210CROOKED B DEADU 4040210CROOKED B DEADW 4040210CROOKED B DEADW 4040210CROOKED B EBALB 4040210CROOKED B KIRC 4040210CROOKED B WBAIE 4040210CROOKED N KETCH 40#2lOALBERT A BRDG 4040211ALBERT A BRlX 4O40211ALBERTAMH 4040211ALBEKT A STEESE 404o211ALBERT EB A CC 4040211ALBERT EB A CRK 4040211ALBERT UB A CC 4040211ALBERT MB A CRK 4040211BEDRo(;x A STEES 4040211BEDROCKASTEIES 4040211BEDROCK A STEES 4040211BEDROCK A STEES 4040211BEDROCK A STEES 4040211BOULDER A CC 4040211BOULDER A CC 4040211BOULDER A CC 4040211BCWLDER A CC 4040211BCULDER A GRNHR 4040211BOULDER A GRNHR 4040211BOULDER A GRNHR 4040211BOULDER A STEES 4o40211BOULDER A STEES 4040211
es-ul-2585-07-25as-07-26
84-08-0884.08-4984.08-09f%J?-23&d-2485-U7-25
84-oa-08
a4-08.og84-08-0984-08-0885-06-16a5-06-17
iz-284x2484-q-26
16302230430
‘9301030163022304301338151513001700la3012151218lo5012201545‘930‘9301030163022304301244
ii:1545loo61230100016101008952
16301540154012351540122014501130lo551440131419141014lo-141220170017381220lo22
TURB
110.00
7%95:oo
$00*~814o:oo75.00
220.00230.00140.00loo.0050.00
130.00loo.00160.00290.00270.00120.00120.00lOo.00110.00220.00220.00750.00700.00550.00270.00340.00250.00500.0015.00
:38* io"11:oo10.00
;*:oolo:oo1.004.500.290.41
36"0:37
1.001.40
TSS
101.0080.3066.40
EE64:8095.9055.20
122.00170.00137.00124.00132.00101.00103.00130.00240.00310.0095.10
103.0074.0094.80
162.00166.00550.00
;5:Ki31o:oo410.001go.00330.0064.00
293.00105.0019.0010.006.006.0010.004.00
;*:a"2:67
27.901.701.917.671.29
26.00101.00
4.602.001.80
35 -
Appendix A. (Continued)
;;:376377378
%381382383
2386387388389
;;395
%
;xs”4004014024034044054064074084094104 1 1412413414
:;z4174184194204 2 1422423424
DCATION HYUNIT
iEKm4ACHSRKETCHEHACHSRKETCHMACHSRKETCHD! A CHSRKETCHEM A CHSRKETCHEM A QiSRKETCHEl4 A CHSRKETCHEM A CHSRKETCHEM A CHSR
!Ez ii ii;:KETCHR4 A QiSRKETCHEMA CHSRKEXCHEHACHSRKETCHEM A CHSRKETCHEM A CHSRKETCHEM A HININKETCHEM A CHSRKETCHEM A CHSRKETCHEMNCCKETCHEM N CCMAMHDTHAMntFtAMKlTH A STEESMAMMDTH A STEESl’UMKl2H A STEESMAMMDZH A STEFSt4AMFQTH A STEESlWUWIY A STEESMAM14TH A STEESMAMKITH A STEESl”liW4lTI.l A STEESMAM?Q’IH A STEESMAMK)Ttl A STEEStdAMMl?H A STEESWTH A STEESt4AMlQTl.l A STEESMAHH)TM A STEESMAHMDTHASTEESHAHMDTNASTEESMAHM3TH A SEESlW%VfH A SEESItAMKlTHASTEESlfAl4KITj.l A STEESkI.AMM3TH A STEESt4AMMXM A SEESMAMKITH A STEES!+lAMQ’I’H A STEESMAMMITH A STEESMAMMDTIi A STEESMASTODON A MINEMASTODON A MIHMASTODQN B WILKMILLER A MINING
4040214404I214404021440402144040214404021440402144040214404021440402144040214404021440402144040214404021440402144040214404021440402144040214404021440402154040215404021540402154040215404021540402154040215404021540402154040215404021540402154040215404021540402154040215404021540402154040215404021540402154040215404021540402154040215404021540402154040215404021540402154040215
DATE TIE TURB TSS
84-08-08
84-08-09
~-EzLO849
i%:8410~1084-08-1084-08-2785-06-16
N-08-2384-08-0884-08-0984-08-0984-08-0184-08-0184-08-0884-08-0884-O&0884-oao884-08-0984-08-0984.oa0984-08-0984-08-09
E-:Eg84:08-0984.oat084-08-1084-08-1084-08-1084-08-21as-a-1785-06-2085-07-23B-07-24
84-08-0284-08-0184-m-31
36 -
15502210
:t
;z1545184220481 2 46569531124122016389151030
1%
1z1004122016201150150517522115
63;900
1157150418002057
5$0
1z60
1155144015151515615
12151815-s
1 51100122013001000
4600.002500.00710.00160.00
E%3600:00650.001400.005100.00
390.00240.00450.00
3300.00400.00
1300.000.75
2000.003400.001100.00340.00280.00
1200.001000.00300.00340.00600.00170.00500.00370.00300.0050.00
210.00130.00120.00220.00600.00340.00400.00370.00110.00270.00
1000.00180.00250.00230.00150.00450.00400.00
0.50370.00
1300.001.10
EfoO160:o350.0E*i
gO3:;2700:o7100.0
%%31o:o
7600.0
:"s*:.
19lE261o:o1000.0130.0350.0
1812.01810.0270.0480.0990.0240.0660.0370.0420.0173.0160.0210.0250.0280.0770.0360.0560.0400.088.0
358.01205.0199.0239.0199.0146.0394.0349.0
43E1340.0
0.8
Appendix A. (Continued)
O B S fAWS’ION HYUNIT *y SOURCE DATE TIS TURB TSS
478
tg481
::
ig
ig4884894904914924934944954964974984995005 0 1502503504
s":
5'z509510511512513514
5175185195205 2 1522523524
;2
$3529530
CHE?UANORDALE
izz f: EEECHENA A SM TRACCHENA A WENDELLCHENA A WENDELLCHENA A WENDELLCHENA A WENDELLCHENA A WENDELLCHENA A WENDELLCXENA MF A MINECHENA MF B FONDCHENA MF B FONDCHENA MF B FONDCHENA MF B PONDCHENA NF A EFCHENA NR 2 RICHENA NR 2 RICHENA NR TYO RICHENA NR TWO RICHENA NR TWO RICHENAJF AB MTHCHENA,EF AB KI’HCHENA,EF AB MTHCHENA,NF AB EFCHENA,NF AB EFCHENA,NF AB EFCRIPPLE A CHENACRIPPLE A CHENACRIPPLE A CHENAFAIRBANKS A KMFAIRBANKS A MTHFAIRBANKS A MTHFAIRBANKS A KI’HFAIRBANKS A PAXFAIRBANKS A SATFAIRBANKS A SATFAIRBANKS A SATFAIRBANKS A SATFISH AT GOLD DRFISH AT GOLD DRFISH B GOLD DRGFISH B COLD DRGFISH B GOLD DRGFISH B GOLD DRGFISH B GOLD DRGFISH B GOLD DRGFISH B GOLD DRGFISH B GOLD DRGFISH B GOLD DRGFISH B GOLD DRGFISH B GOLD DRGFISH B GOLD DRG
40506014050601405060140506014050601
:E%405060140506014050601405060140506014050601
:g%14050601405060140506014050601405060140506014050601
:::%i405060140506014050601405060240506024050602405060340506034050603405wo3405060340506034050603
:gz;4050604405060440506044050604
::;%i40506044050604405060440506044050604
44gE4050604
i44
1'11
i
z
i
:
z1
i
i1
i14441
;1
;1
:11111111111111
84-08-1384-08-1384-08-13
83-08-05
83-08-0583-08-1083-08-108wE4984-05-1585-05-1584-08-1084-08-1384-08-1684-08-2004-08-1684-08-1084-08-1384-08-1684-08-2084-08-10&I-08-2084-08-13
Ki-it8410811484-08-1484-08-1484-oa-1484-08-1484-08-15
i%;:84:08:15
1837214012001200121017551250210011202058115012101211121213001405130517451800134514301620
1:::1725
1;;o"12001200120019102030192518152100202016452040
E183020002300200500800
110014001700200500
11001400
2 . 10.7
15.013.02.73.05.23.0
25'
30::3.84.93.50.2
25"3 . 41.32.22.5
;:;
i* :0:4
45.0250.026.00.8
kk;0 . 5
120.060.0
360.01800.0
2::19.0
;I*;;:;
.
1;:'o12.014.017.014.018.018.0
12.005.60
86.0047.005.007.004.004.004.0011.001.00
18.0022.0026.0017.004.004.00
13.004.001.301.008.005.005.001.002.002.00
235.002c60.00226.00
0.050.200.800.80
118.0040.00
3368.00
7:iEo"62:0028.0038.0016.00
1E23:0018.0024.0018.0030.0030.0048.0046.00
37 -
Appendix A. (Continued)
OBS UXATION HYUNIT
%
:t
::;
55:
g
:z597598599
%
SE
ig
%Z6086096106 1 16126136146156166176186196206 2 1622623624625
:g628629630631632633
2;616
ki& : iiiLalMAAamsLCliENAACHRSLCHENA A CHRSLCHENA A CHSRLCHENA A MSRLCXENA A CHSRLCHENA A CHSRLCZiENA A MSRLCMENA A CHSRLCIENA A CHSRLCXENA A CHSRLCIENA A CHSRLCHENA A CHSRLCHENA A CHSRLCHENA A CHSRLCHEINA A NORDALLCHENA A NORDALLCHENA A NORDALCHATANIKA A 39MCHATANIKA A 39MWTANIKA A 39MCHATANIKA A 39MCHATANIICA A 39MCIiATANIKA A 39MCHATANIKA A 39MCXATANIKA A 39MCHATANIKA A 39MCMTAJUKA A 39MCHATANIKA A39MCliATANIKA A 39MGIATANIKA A39MCHATANIKA A39MCHATANIKA A 39MCHATANIKA A 39MCHATANIKA A 39MCHATANIKA A39MCHATANIKAA 39blCHATANIKAA 59HCHATANIKAADOTCHATANIKA A ELLCHATANIKA A ELLCHATANIKA A ELLCHATANIKA AELLCHATANIKAAELLCHATANIKA A ELLCHATANIKA AELLCHATANIKA AELLCHATANIKA A ELLClUTANIKA AELLCXATANIKA A ELLCMTANIKA AELL
4050605
:gi%g405060540506054050605
44:::4050605
:g%g40506054050605
:~~~40506054050605405cm54050605405060540509014050901405090140509014050901405090140509014050901405090140509014050901405090140509014050901405090140509014050901405090140509014050901405090140509ol4050901405090140509014050901405090140509014050901
%z:40509014050901
SOURCE
:
z:1111111
111
1:4111
;
111
:
i
:
i
t4
:1111144
:
t
DATE
84-08-1184-08-11
83-08-06
84.08-a784-08-1084-08-1484-08-1484-08-1584-08-1584-08-21
83-08-07
TIE
19051915181922002100202022003007001200130013401900130517101335
172z12001200134520001030115021401250143022551315143512551256172315301540
1;:74516301510160511102010111521451130210
'120012001122211814101200
TURB
5.8
;::2.2
::i6.19 . 28.8
:*:5:7
E9:3
7"*:12:o
4.8
ii*%12:o10.06.9
16.065.05.1
22::14.0
3 . 913.012.0
:*2"8:04.5
t:;4.6
13.016.011.0
:::29.0
10.012.015.0
65::
;:i
ix6:0
2:6:010.012.06.0
ai*:164:0258.0
:::2.05.03.01 . 0
::i18.015.028.0
3;::4.04.04.4
24.06.4
63:;9.02.04.0
2:;2.03.0
!E8:04.0
10.0227.0
- 38 -
Appendix A. (Continued)
OBS
690691
%!F6%6976986997007 0 1
:il?;z7097107 1 17127137147157167177187197207 2 1722723724
;2
7737297307 3 1
739740741742
iiEEi zEiEE tizCHATANIKA ALONGCHATANIKA ALONGQUTANIKA ALONGQiATANIKA ALONGCHATAJUKA ALONGCHATANIKA ALONGWTANIKA ALONGCHATANIKA ALONGWATANIKA ALONGCHATANIKA ALONGCHATANIKA ALONGCHATANIKA ALONGCXATANIKA ALONGWATANIKA ALONGCHATANIKA ALONGCHATANIKA B FAICHATANIKA B FAICHATANIKA B FAICXATANIKA B FAICHATANIKA B FAICHATANIKA B FAICHATANIKA B FAICHATANIKA B FAXCUTANIKA B FAXCHATANIKA B FAICHATANIKA B FAXCHATANIKA B FAICHATANIKA B FAIUiATANIKA B FAICYATANIKA B FAICHATANIKA B FAICHATANIKA B FAXQMTANIKA B FAICHATANIKA B FAICIiATANIKA B FAICHATANIKA B FAICHATANIKA B FAICHATANIICA B FAIU-IATANIKA B FAIQiATANIKA B FAICHATANIKA B FAICHATANIKA B FAICHATANIKA B FAICHATANIKA B FAICHATANIKA B FAICHATANIKA B FAICHATAJUKA B FAX
HYUNIT
40509014050901
~~:;
:gg
::;g40509014050901405090140509014OmOl4050901
::ggiz40509014050901.405090140509014050901405090140509014050901405O901405090140509014050901405090140509014O509Ol40509014OmOl405090140509014O5ogOl4050901405090140509014050901405090140509014m901405090140509014w90140509014050901405090140509014050901405090~4m901
SOURCE
333
z
:3
:
z
;
z33
i41111111111111
;1111111111
;1
111
DATE
84-08-0984-08-1084-08-1084-08-10a4-08-1084-08-10
KE:84109-2383-08-0683-08-0983-OS-O9
83-08-1083-08-10
83-08-l 183-O8-1183-08-1183-08-1183-08-1183-08-11
83-08-16
220
iii:11201420172020202320220520725820
112014201620170615541600
E16101725151017302030
2:;:530830
1130143017302030
2::oo
F$?l113014301730185520302330230
ii::1130143017301930201019302025
3.97.8 E
70.0 71:o95.0 100.0x
8:2
22.0 43.0
4.1 1::;3.6
30.0 29”::3.7
60.0 6;::85.0
2:: :;*: 24:015.0 19.09.4
;:::*:lo:o
1:::, 3.6 5.045.0 120.0;I*: 38.0
29:o Z:i38.0 g.:70.0 .75.0 90.070.0 76.0
50:o ;5s*:62.046.0 48.0
34.0 40.060.0 68.0go.0 110.0140.0 144.0110.0 118.085.0 88.060.0 42.0
70.0 72.0140.0 132.0150.0 152.0ljO.0 134.0120.0 100.085.0 64.055.0 44.0
20:o %E
42.0
48.0 22.032.0 37.0
- 39 -
Appendix A, (Continued)
OBS LDCATIOW
FAITMA-E
i%iEFAITH A STEESEFAITli A STEESEFAITH A STEESEFAITH A STEESEFAI’IW A STEESEFAITH A STEESEFAITH A STEWFAITH A STEESEFAITH A STEESEFAI’TY A STEESEFAI’Ili A STEESEFAITH A STEESEFAITH A STEESEFAITH A STEESEFAITHABMCUAIFAITH B MCINTSHFAITH B MCINTSHFAITH B MCINTSHFAITH B MCINTSHFAITH B MCINTSHFAITH B MCINTSHFAITH B HINEMC?WUS A FAIlIMC#UNUS A FAITHIWANUS A FAI’INMCMANUS A FAITHMCMANUS A FAITHMCHANUS A FAITHMCXANUS A FAITHMCXANUS A FAITHMCMANUS A FAITHMMANUS A FAI‘lliMWUS A FAITHMCMANUS A FAITHTATALINA A BRIETATALINA A BRDGTATALINA A BRDGTATALINA A BRDGTATALINA A CHTOOLDSTREAM A FXOOLDSTREAM A FXGOLDSTREAM A FXGOLDSTREAM A LRCOLDSTREAM A MTGOLDSTREAM ALOGGOLDSTREAM B FXCOLDSTREAM B FXCOLDSTREAM B FXOOLIH’REAM B FXOOLDSTREAM B FX
HYUNIT
40509044050904
~~:4050904
:gg:405090440509044O50904
:~~:4050904
:z%:4O509O440509044050904405090440509044050904405090440509044050904405090440509054050905405090540509054050905405090540509054050905405090540509054050905405090540509064O5090640509064050906405090640509104050910405091040509104050910405091040509104050910405091040509104050910
SOURCE DATE TIE TURB TSS83-08-0683-08-0683.OaOg83-084
g-E2283IOs.1683-08-1684-08-0784-084
LizE~84:08--15;-oa;:
84:oa2184-09-2384-08-2184-08-01
E-E:84:oa 1784-08-2984-08-3085-06-0983-08-0683.oaog83.oa12
2-E8Loam84-08-1084-08-1484.oa15M-08-1584-08-2184-09-23
!EELOc16
itEE;84:05-0984-05-15a5-05-1584.0115
!Eza3:oaoa
gEz83108-1483-08-14
164517401500
i7&20001910
?7?17221516
'E1825900
22301445
1g
g;13101 555450653655
z-i
g:;16521515
'E1830910
14151200120015301200132612001200120012001240120Q12251050113014551540
45.0120.0120.0
E70:o31.055.0140.0120.039.0140.014.0i20.017.065.040.012.0
2600.0190.0550.0600.0280.0130.0130.0
O*30.30.20 . 40.20.10.31 . 00.4
0":;
i:&.
A:;2.3
40.0180.075.0190.03 0 . 0
190.0
Ei3oo:o260.0no.0
44.0182.0
Eg:;
39:o78.0
260.0
%i17o:o21.0
416.014.0
148.029.019.0
1890.0339.0465.0767.0
::EnS:o
1 . 01 . 01 . 0
2-i4:o
E
E
i::16.070.0
5;::.
6:::726.0128.060.0128.0556.0292.0272.0250.0282.0
40 -
OBS UKATION
902903904
E9m
E9"910911912913914915916917918919
E9229239249259269279289299309 3 1932
;:i935936937938
;t:941
;t:944
;2947948949
iFi)952
E34
GILMEBHUINGIUORE B BDMNGILHIRE B HMINGILME B BKNGILlORE B BJMNGIWRE B BIXNGILWREBBMINGILHIRE B BI3U.NGItEIoRE B BWINGIIJ4X.E B WINGIlloRE B BcE(INGILHORE B BDHINGILMIRE B BDMINGUMRE B BMINGILME B BIMNGIIJQRE B BrXMINGILKNE B BCMINGlLlgRE B BDMINGILNXE B BLMINGlIKIRE B BIMNGILJQRE B BIMNGILWE B BIWNGILJQRE B BDHINGILMXE B BCHNGILHIRE B BMINGILMORE B BCHINGILMXE B BDMINGILMIRE B BDMINGILMME B BDMINGIMRE B HWNGILJQRE B BIMNGILMORE B MINGILKIRE B BMWGILXJRE B BDMINGILMORE B BDMINGILWRE B BDMNGUtIRE B BIXINGILMIRE B mINGILMIRE B BIMNGwlIDRE B BEKNGILHDRE B BlMINGILKIRE B BDHINGILKME B BIMNGIWRE B BIXINPEDRO A MMPEDRO A MTHPEDRO A KIHPEDRO A MlWTOLOVANA A BRDCTOLOVANA A BRDGTOLOVANA A BRDCTOLOVANA A BRDGTOLOVANA A BRCC
Appendix A. (Continued)
HYUNIT - WJRCE -DATE TIM? TURB
40509124030912405091240509124050912
:gg122
:~~i340509124050912405091240509124050912405091240509124050912405091240509124050912405091240509124050912405091240509124050912405091240509124050912405091240509124050912405091240509124050912405091240509124050912405091240509124050912405091240509124050912405091340509134050913405091340509204050920405092040509204050920
22
3
;2
9
22222
:22222
2'2
z2
5222
f
52
5222221
;34
I4
84-06-1684-06-1684-06-1784-06-1784-06-17
84-08-2684-08-2684-08-27
84-08-2884-08-2884-08-2984-08-29
84-08-1084-08-1384-05-0984-05-1584-08-1285-05-1585-084
11551825940
14301830115513551510
1E1430112012501515141516101700130016101715940
13101525100012301540930
124515101210151517551135144016301130152517251130144016251135163018001445123518121900.12001200125512001440
280.001100.00500.00
EE.500:00
E*8i7oo:oo
iE:750:oo700.00900.00
5300.003400.002900.003000.003200.003100.00
%E:2600:003400.003100.00
:G%1200:oo1600.001400.001600.002200.001700.002200.002800.002100.001900.001800.002100.001600.001600.001600.001600.001700.00
70.0055.00go.0030.002.403.801.604.201.02
TSS
20.0
go"
:$:324:O
g-:19o:o195.03c5.0235.0374.0315.0
1300.0600.0620.0660.0620.0620.0670.0730.01050.0960.0820.0430.0365.0300.0
E*:445:o
iii?:720:0
1240.0560.0620.0460.0580.0460.0420.0500.0420.0440.034.070.093.034.010.030.01.0
24.01 . 2
41 -
Appendix A. (Continued)
OBS LOCATDN HYUNIT SOURCE DATE TIM TURB TSS
1008100910101011101210131014
lx1017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059lc%O
TllwvANAAuFiEE:zTaOVlVlAAwFTQLOVANAAWF
EE;:Z
ii!ikKE i zTOLOVANAAUF'IWOVANA A WFTOLOVANA A WFTOLOVANA A WFTOLOVANAAWFTOLOVANA A WFTOLOVANA A WFTOLOVANA A WFToLovAMAwFTOLOVANA A WFTOLOVANA AWFTaOVANA A UFTOLOVANA A WFTOLOVANA A WF?OLOVANA AWFTOLOVANA A WILBTOLOVANA AWILBTOLOVANABWFTOLOVANABWFTOLOVANABUFTOLOVANA B WFTJXOVANABWFTOLOVANABWFTOLOVANA B WILBTOLOVAJlABWILB'I'QLOVANAB WILBTOLOVANA WFTaovANA WFTOLOVANA WFTOLOVANAWFToLovANAwFAccTOLOVANA WF ACGlDLOVANAWFACGLIVENGOODABRDLIVENGtMDABRDLIVENGOODABRDLNENGOODABRDKVENGOODABRDLIVENGOODABRDLIVENGOODABRDLIVENGOODABRGLIVENGCNDABRGLIVENGOODABRGLIVEKOODAELL
4050920
ZzE405092040509204050920405092040509204050920
EE840509204050920405092040509204050920
::;g;:4050920405092040509204050920405092040509204050920
:::g3:4050920405092040509204050920405092040509204050920405092040509204050920405092040509204050920
:::ig:40509214050921405092140509214050921405092140509214050921405092140509214050921
3
;3;3;33:!3i3
3::1111
;11
;
a1
:443
;
:1
:4
84-08-14
Ezi-144841oc1484-08-15
EE:84Ioa-1684-08-1684-08-1684-08-16
84-08-1284-08-1684-08-1284-08-1284-08-1684.oal284.oal2
~-05-0984-05-1584-08-1284-08-1284-oa1684-08-16
83-08-1185-08-W
102013201620192022201 2 04207201020
:iz192022201 2 042072010201320162019202220120
141514151900190714301520
12:
:;2201221251910164217471602160213351420185012001200
350
1;::12001740130019251425
12.08 . 65.4
24.0
E1o:o
:Ki;;:;
a8:;.
2:;14.09.2
14.016.016.034.032.028.017.0
:::
88':14:o12.016.015.06.8
120.0180.0
1.6
0":;
::i
1.3918o:o220.0190.0260.017.012.0
230.010.0
'ZKi24:0
27.015.013.052.030.023.017.023.0
E83:o55.020.014.017.019.015.024.021.039.060.0
:*:43:o5.04.012.0
369::30.035.023.021.0
710.01400.0
4.014.04.0
25"1 . 0
%!;:;
284:o230.025.024.0
757.013.0
234.030.0105.0
42 -
OBS
1114
:1;:111711181119112011211122112311241125
KxAmN
KoTuMLsFADtKOTUKDK SF A DAKDYUKUKSFADAKOYUKUKSFAHUYPROSPECT A DALTPROSPECT A MINEPR(36pEff A MIUGPROSPECT A PIPEPROSPECT B MINGPROSPECT B MING
Appendix A. (Continued)
HYUNIT
406010140601024m201406020140602014060201
:%E4060201
zzoo14060201
S O U R C E DATE T I E TURB ISS
84484885-06-l 184-08-0384-08-0985-00-16
Ez
173522001615150011502100155012001030105019401 4 0
E5:200.700.300.902.900.342.80
55.00
2g*::.
14.0
67:;3.20.8
i::0.36 . 3
294.0194.048.073.0
180.011% PROSPECT B bUJG 4060201 85-08-15 2240 6.001127 PROSPECT B MING 4060201 4”1128
740 50.00PROSPEa B PIPE 4060201 4 1010 25.00 9 . 4
- 43 -
- 44 -
Appendix B. Further explanation of statistical techniques
A. Standard Error of Emtimata.
&caumo tha linear rograrmion u8m logarithnic transformation of
the data, the calculated l tandard l ror of l etiutm i8 a logarithm. In
thir report it ir roportad a8 a parcmtaga which im calculatad by
adding (and l ubatraoting) the SEE to the logarithm of a baas linear
value, back tramforming the ro8ult to a linear value, subtracting the
ba8o linaar value from thim fosult and dividing by the barm linear
value. Below is a l ampla calculation:
Tha standard l rror of l stimato for the log-log equation for the
combined data from Birch Creak bamin im 0.243. Asmum a linear value
of 2.00 rilligramm par liter.
+SEE(%)-[ (10 (10g(200)+*243))-2001/,,, 1.75 or 75 percent
1.43 or 43 parcant
- 45 -
Appendix B. (Continued)
different badna, mtrrrrr or ritar are rimilar, indicator
variabloti for tha different location8 arm added to the basic
turbidity-TSS modal. An F tast is performid to 1.0 if tha rlopo and y
intercept coofficionts of the full model (with indicator variables)
are statistically different from those of a roducad model (without
indicator variables) at a specified confidence level. ,?LO equation for
this relatfonahip is:
wham:
SSEp is the error sum of squares for the full model,
SSER ir the error 6um of squares for tha reduced model,
dfp i8 the degrees of freedom for the full modal, and
dfR is the degrees of freedom for the reduced model.
If the calculated P* i8 1e8s than F at a specified confidence
love1 (P values ara from an F value tablr), the inference is that the
two groups of data are not statistically different at that level
(Neter, Wa8sorman, and Rutner 1985). Thi8 typ8 of analysis can also be
used to moo if data from different years or sources can be combined.
- 46 -
Appendix C. Model validation results
S A M P L E L O C A T I O N DATB c z TURB TSS TSS Diff 2 VALUETI)CIC field raprtd calcltd Rpt-Calc
(NW wm owl 1 owl) (R-P)/SEE
A. Data collrctad by DEC at variour location in interior Alaskain 1985.
CHATANIKA A POX 85081605II a a 85081676
CHATANIKII, A 39 M 85081606CXATANIKA A L&NG 85081607
I I II II 85081650I I II II 85081651II * m 85081652II n a 85081653n II a 85081654I I II w 85081655II n n 85081656I I II w 85081657II n II 85081658I I a la 85081659I I II w 85081661I I II a 85081662n 0 w 85081663II a * 85081664II II w 85081665n w w 85081666I I w n 85081667I I II II 85081668I I II II 85081669I I II II 85081670II I) II 85081671I I w II 85081672I I m II 85081673I I II a 8508167411 n w 85081675
FAITH ABOVE MCXAN85081609 290.0 76.0 345.1 -269.1 -1.39I# w " 85081623 93.0 31.0 119.9 -88.9 -1.32
CHATANIKA B F&MII a III I II II
I I II II
I I II II
I I II I
I1 I8 n
II I II
85081625 62.0 11.0 87.6 -76.6 -1.0385081626 164.0 36.0 206.9 -170.9 -0.9785081627 104.0 28.0 138.4 -110.4 -0.9485081629 264.0 97.0 315.2 -218.2 -0.8285081630 310.0 110.0 363.3 -253.3 -0.0285081631 240.0 80.0 289.8 -209.8 -0.8585081632 276.0 100.0 327.9 -227.9 -0.8285081633 200.0 54.0 .- 246.6 -192.6 -0.92
4.26.54.4
115.030.012.07.06.58.5
18.028.044.037.034.022.020.033.022.038.030.033.024.022.021.032.044.039.038.040.0
1.0 3.58.01.0 X:i
15.0 127.27.0 26.1
12.0 8.96.0 4.73.0 4.32.0 5.94.0 14.3
21.0 24.014.0 41.012.0 33.41.0 30.24.0 18.15.0 16.24.0 29.2
11.0 18.112.0 34.58.0 26.18.0 29.2
;:::20.118.1
5.0 17.15.0 28.1
13.0 41.010.0 35.54.0 34.58.0 36.6
Avarag8 for Chatanika at
-2.5 -0.792.3 0.45
-2.0 -0.58-112.2 -1.88-19.1 -1.56
3.1 0.761.3 0.59
-1.3 -0.64-3.9 -1.41
-10.3 -1.53-3.0 -0.27
-27.0 -1.40-21.4 -1.36-29.2 -2.06-14.1 -1.66-11.2 -1.47-25.2 -1.84-7.1 -0.83
-22.5 -1.39-18.1 -1.48-21.2 -1.54-14.1 -1.49-13.1 -1.54-12.1 -1.51-23.1 -1.75-28.0 -1.45-25.5 -1.53-30.5 -1.88-28.6 -1.66
Long W-1.30
- 47 -
Appendix C. (Continued)
EAST F TOlamNA 85081954 5.50n n n 85081955 5.90II n 0 85081956 5.50n n 0 85081957 4.90n n a 85081958 4.70H n n 85081959 6.80n n u 85081960 6.50It n n 85081961 12.00H n n 85081962 6.80n n n 85081963 5.50n n 0 85081964 4.90I) n n 85081965 5.00H n n 85081966 5.50I I n a 85081967 5.60n * n 85081968 6.200 n H 85081969 6.10n n m 85081970 6.40n n n 85081971 5.10II n n 85081972 4.70
TOLOVANA A TAPS 85081973 5.80 4.0 10.1 -6.1 -1.13n It n 85081974 5.70 5.0 9.9 -4.9 -0.93n n n 85081975 5.40 5.0 9.3 -4.3 -0.87n n n 85081976 5.60 6.0 9.7 -3.7 -0.72n a I( 85081977 6.20 8.0 10.8 -2.8 -0.49n n n 85081978 6.40 10.0 11.2 -1.2 -0.20n n w 85081979 5.00 11.0 8.6 2.4 0.53 (n n n 85081980 5.50 6.0 9.5 -3.5 -0.69n n a 85081981 5.70 11.0 9.9 1,l 0.21n n n 85081982 5.40 9.0 9.3 -0.3 -0.06n n n 85081983 5.70 9.0 9.9 -0.9 -0.17W n m 85081984 6.00 7.0 10.4 -3.4 -0.62n 0 n 85081985 6.20 9.0 10.8 -1.8 -0.32n n I( 85081986 5.70 5.0 9.9 -4.9 -0.93I n 0 85081987 5.60 9.0 9.7 -0.7 -0.13H n n 85081988 5.30 6.0 9.1 -3.1 -0.65It m n 85081989 5.10 3.0 0.0 -5.8 -1.24* n 0 85081990 5.10 6.0 0.8 -2.8 -0.59It n n 85081991 4.90 5.0 a.4 -3.4 -0.76tt n n 85081992 5.00 6.0 8.6 -2.6 -0.56n n n 85081993 4.70 6.0 8.0 -2.0 -0.47n * n 85081994 4.80 9.0 8.2 0.8 0.18It n n 85081995 4.40 10.0 7.5 2.5 0.64* n n 85081996 5.30 6.0 -9.1 -3.1 -0.65
TURB TURB TSS TSS Diff Z VALUElab f iald roprtd calcltd Rpt-Calcwm mm owl 1 (w/l) (R-PI/SEE
6.0 9.513.0 10.33.0 9.56.0 8.4
12::8.0
12.010.0 11.458.0 22.112.0 12.06.0 9.56.0 8.46.0 8.66.0 9.57.0 9.78.0 10.88.0 10.67.0 11.26.0 8.05.0 8.0
Average for Tolovana ab
-3.5 -0.692.7 0.50
-6.5 -1.29-2.4 -0.54-1.0 -0.240.0 0.01
-1.4 -0.2335.9 3.050.0 0.01
-3.5 -0.69-2.4 -0.54-2.6 -0.56-3.5 -0.69-2.7 -0.52-2.8 -0.49-2.6 -0.47-4.2 -0.71-2.8 -0.59-3.0 -0.71
W88t F o r - O . 2 5
- 48 -