United States
Environmental Protection
Agency
Robert S. Kerr Environmental Research
Laboratory
Ada OK 74820
EPA-600/4-79-050
August 1979
Research and Development
&EFK
Identifying
Sources of
Subsurface Nitrate
Pollution with
Stable Nitrogen
Isotopes
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/4-79-050
August 1979
IDENTIFYING SOURCES OF SUBSURFACE
NITRATE POLLUTION WITH
STABLE NITROGEN ISOTOPES
by
Timothy J. Wolterink, Hugh J. Williamson,
David C. Jones, Thomas W. Grimshaw, and
W. F. Holland
Radian Corporation
Austin, Texas 78758
Contract No. 68-03-2450
Project Officer
Jack W. Keeley
Ground Water Research Branch
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr
Environmental Research Laboratory, U.S. Environmental Protection
Agency and approved for publication. Approval does not signify
that the contents necessarily reflect the views and policies of
the U.S. Environmental Protection Agency, nor does mention of
trade names or commercial products constitute endorsement or
recommendation for use.
11
-------
FOREWORD
The Environmental Protection Agency was established to
coordinate administration of the major Federal programs designed
to protect the quality of our environment.
An important part of the Agency's effort involves the search
for information about environmental problems , management tech-
niques , and new technologies through which optimum use of the
Nation's land and water resources can be assured and the threat
pollution poses to the welfare of the American people can be
minimized.
EPA's Office of Research and Development conducts this search
through a nationwide network of research facilities .
As one of these facilities, the Robert S. Kerr Environmental
Research Laboratory is responsible for the management of programs
to: (a) investigate the nature, transport, fate, and management
of pollutants in ground water; (b) develop and demonstrate methods
for treating wastewaters with soil and other natural systems ;
(c) develop and demonstrate pollution control technologies for
irrigation return flows; (d) develop and demonstrate pollution
control technologies for animal production wastes; (e) develop
and demonstrate technologies to prevent, control or abate pollu-
tion from the petroleum refining and petrochemical industries ;
and (f) develop and demonstrate technologies to manage pollution
resulting from combinations of industrial wastewaters or indus-
trial/municipal wastewaters.
This report contributes to that knowledge which is essential
in order for EPA to establish and enforce pollution control stand-
ards which are reasonable, cost effective, and provide adequate
environmental protection for the American public.
William C. Galegar
Director
111
-------
ABSTRACT
This report describes the methods, results, conclusions,
and recommendations of an investigation of a technique to iden-
tify sources of nitrate ions in ground water. A discussion of
the theoretical basis of the technique is also provided. Over
300 soil and ground-water samples were collected for this study.
The samples are from numerous sites from within the United
States, representing a variety of environmental conditions. The
nitrate ions in 66 of these samples were separated from other
nitrogen species, converted to nitrogen (No) gas. purified, and
analyzed to determine the nitrogen-isotope ratio (15M/lttN).
These data were combined with the results of analyses performed
previously by Jones (1) and Kreitler (2). Standard statistical
techniques were used to analyze the observed variations in 615N
values, with respect to several nitrate-ion sources and various
environmental factors. It was found that nitrate ions from feed-
lots, barnyards , and septic tanks can be distinguished from nat-
ural soil nitrate ions on the basis of the 615N values. These
nitrate ions from various sources cannot, however, be distin-
guished from each other. Environmental factors contributed to
the observed variation in 615N values.
This report was submitted in fulfillment of Contract No.
68-03-2450 by Radian Corporation under the sponsorship of the
U.S. Environmental Protection Agency. This report covers the
period August 23, 1976 to March 31, 1978, and work was completed
as of November 1, 1978.
IV
-------
CONTENTS
Foreword iii
Abstract. iv
Figures vii
Tables viii
Acknowledgements x
1. Introduction 1
Goals and Approach of This Study. 1
Significance of the Nitrate Problem 2
2 . Conclusions 4
3 . Recommendations 6
4. Rationale for Tracing Nitrate Via Stable Isotope
Ratios 7
Nitrogen Isotope Abundance 7
Isotope Fractionation Mechanisms 9
Nitrogen in the Environment 13
Nitrate Source Determination 18
Other Stable Isotopes As Tracers 21
5 . Experimental Procedures 26
Selection of Sampling Sites 26
Sampling Procedures. 28
Sample Preparation and Analysis 30
Analysis of Standards 37
6. Sampling Sites 39
7 . Results and Discussion 46
Statistical Analyses of the Data. 46
Analytical Problems. 68
Summary and Conclusions. 73
v
-------
CONTENTS (Continued)
References 77
Appendices 79
A. Reagents Used in Isotopic Analyses 79
B. Description of Radian's Intensive Sampling Program,
By Sites. 81
C. Sample Numbering System 122
D. Basic Data. 125
E. Climate Classifications 134
F. Soil Classifications 136
G. Statistical Analyses of the Entire Data Base 139
H. Separate Statistical Analyses for the Different
Source Types 144
VI
-------
FIGURES
Number Page
1 Simplified Schematic of the Nitrogen Cycle 16
2 Mechanisms of Nitrate Formation 19
3 Distillation Apparatus Used for Removal of
Interfering Nitrogen Species 32
4 Nitrogen Gas Generation and Purification
Apparatus 34
5 Nitrogen Gas Generation Tube 35
6 Locations of Sampling Sites 42
7 Histograms of Del Values by Source Category 50
8 Illustration of the Meaning of Correlation
Coefficient, r 55
9 Depth Profile of Nitrate Concentration and 615N
for the Greeley, Colorado, Animal Feedlot Site,
Hole H2 71
10 Nitrate and 615N Versus Depth Beneath a Barnyard
With a Definite Animal Waste Contribution of
Nitrogen, In Runnels County, Texas 72
VII
-------
TABLES
Number Page
1 Nitrogen Isotope Ratios for Various Substances.... 8
2 Isotope Equilibrium Constants 11
3 Kinetic Fractionation Factors For Various Non-
Equilibrium Reactions 14
4 Summary of Samples Collected In Radian's
Intensive Sampling Program. 41
5 Key to Figure 6 43
6 Site Environmental Characteristics 45
7 Summary of Samples Included in the Final Data Base,
By Source Category and Site. 47
8 Means and Standard Deviations of 615N For
Different Nitrate Source Categories 48
9 Definitions of Variables Used In the Correlation
Analyses 52
10 Value Required For A Correlation Coefficient (r)
To Be Significant, As A Function of Significance
Level and Sample Size 57
11 Variables Which Have Statistically Significant
Correlations With DEL ' 58
12 Correlations Involving Variables Most Closely
Related to DEL 58
13 Variables Which Have Statistically Significant
Correlations With PPM 61
14 Variables Which Have Statistically Significant
Correlations With DEL: Septic Tank Sources 63
15 Variables Which Have Statistically Significant
Correlations With DEL: Feedlot Sources 64
viii
-------
TABLES (Continued)
Number Page
16 Variables Which Have Statistically Significant
Correlations with DEI,: Irrigation Sources 65
17 Composite Relationships Between DEL and Predictor
Variables 67
18 Variables Which Correlate Most Highly With PPM
(At the .01 Level of Significance) for the Various
Source Categories 68
IX
-------
ACKNOWLEDGEMENTS
This study was sponsored by the Environmental Protection
Agency. The Project Officer for EPA was Jack W. Keeley. The
investigation was carried out by Radian Corporation. The
Project Director for Radian was David C. Jones. Timothy J.
Wolterink directed the field investigations and sampling program,
and oversaw the analytical activities. Thomas W. Grimshaw as-
sisted in the sampling program. Hugh J. Williamson conducted
the statistical analyses and provided much valuable assistance
in interpreting their results. W.F. (Kirk) Holland provided
many helpful suggestions and assisted in interpreting the results
John T. Robinson prepared the samples for isotopic analysis.
Many people outside of Radian contributed much to the com-
pletion of this study. In particular, Dr. Charles W. Kreitler,
of the Bureau of Economic Geology, University of Texas at Austin,
performed the isotopic analyses, provided isotope standards, and
participated in many helpful discussions.
A great many people provided invaluable assistance in ob-
taining samples or background information on various sites.
Special thanks go to Kirk W. Brown (Department of Soil and Crop
Sciences, Texas A&M University); Sam B. Upchurch (Department of
Geology, University of South Florida at Tampa); Karl Ford
(Jefferson County, Colorado, Health Department); Richard Bissell
(Geological Survey Division, Michigan Department of Natural
Resources); Thomas B. Bahr and Tom Burton (Institute of Water
Research, Michigan State University); Dave Meier (Roscommon
County, Michigan, Health Department); Louise Sullenger (Houghton
Lake, Michigan); and Bill Leseman (City of Tallahassee, Florida).
x
-------
Finally, thanks go to the personnel of the Oregon Depart-
ment of Environmental Quality; the West Virginia State Health
Department; the Minnesota Department of Health; and the Rolla,
Missouri, District Office of the U.S. Geological Survey for
agreeing to provide samples via the kit sampling program.
XI
-------
SECTION 1
INTRODUCTION
GOALS AND APPROACH OF THIS STUDY
This report describes the results of an investigation of a
technique to identify sources of nitrate (NOs) in ground water.
The investigation was funded by the U.S. Environmental Protection
Agency.
The technique under investigation was developed previously
by Radian Corporation under EPA Grant 16060 HNI (1). It is based
on the premise that nitrate in soils and ground water will have
a nitrogen-isotope ratio (15N/11+N) that is determined by (a) the
series of reactions which formed the nitrate, and (b) the nitrogen-
isotope composition of its precursors. Previous work (1, 2) has
indicated that soil and ground-water nitrates that were derived
from animal wastes can be distinguished from other soil nitrates
on the basis of their nitrogen isotope ratios. A detailed descrip-
tion of the technique is presented in Section 4 of this report.
The overall objective of this study was to assess the general
feasibility of using stable nitrogen isotope ratios to identify
the source(s) of nitrate pollution in ground water. Within this
context, two key questions were addressed:
1. For a given waste source, are the isotope
ratios of nitrate-nitrogen statistically
equivalent in different geographic and climatic
-------
regions of the country? and
2, For different waste sources, is the variance
of the isotope ratio means adequate to dif-
ferentiate between these sources?
Particular emphasis was placed on evaluating the technique with
respect to its ability to differentiate between domestic septic
tank, animal feedlot, and municipal wastewater land-application
(irrigation) sources of nitrate.
In pursuing this study, soil and ground-water samples were
obtained from numerous sites around the United States. Emphasis
was placed on acquiring sufficient numbers of samples to obtain
statistically significant results. Samples were prepared and
analyzed using equipment and procedures previously developed by
Radian under EPA Grant 16060 NHI (1). These procedures are
described in Section 5 of this report. Isotopic ratio determina-
tions were performed at the mass spectrometer facility of the
Center for Research in Water Resources, University of Texas at
Austin. Standard statistical procedures were then used to evaluate
the results of the analyses.
SIGNIFICANCE OF THE NITRATE PROBLEM
The presence of excessive nitrate in drinking water sup-
plies is a well-recognized public health hazard. The best known
health problem arising from nitrate consumption is methemoglo-
binemia, a cyanosis which is brought about by the reduction of
nitrate (NOa) to nitrite (NOa) by bacteria in the digestive
tract, followed by absorption of the nitrite into the blood-
I I
stream. There, the nitrite oxidizes the ferrous ion (Fe ) in
I ||
hemoglobin to ferric ion (Fe ), thereby preventing the trans-
port of oxygen by the hemoglobin. This results in a gradual
suffocation (cyanosis). It is noteworthy that infants are most
-------
susceptible since the acidity of their stomach is considerably
less than that of adults, resulting in a more favorable environ-
ment for the nitrate-reducing bacteria. The digestive system of
animals such as cattle (ruminants) also is conducive to this
bacterial action. Therefore, when a nitrate problem occurs, it
is first reflected in health problems of cattle or human infants
It should also be noted that the whole problem of methemo-
globinemia is poorly understood. The nitrate or nitrite concen-
tration at which methemoglobinemia becomes a problem may vary
widely, and is most likely influenced by some as yet unidenti-
fied factor or factors. Because of the methemoglobinemia prob-
lem, the U.S. Environmental Protection Agency has established
a maximum contaminant level (MCL) for nitrate (as nitrogen) of
10 milligrams per liter (mg/£). This MCL was established for
drinking water supplies, under the National Interim Primary
Drinking Water Regulations (3).
Another potential health effect of nitrate relates to can-
cer. Nitrate and/or nitrite can react with secondary amines to
form nitrosamines, many of which are carcinogenic. Numerous
studies indicate a possible link between nitrate and cancer (4,
5,6).
If ground waters become contaminated with nitrate as a
result of man's activities, a public health hazard could arise.
Moreover, some ground waters naturally contain high levels of
nitrate. The technique under investigation in this study has
been proposed as a means of identifying the source(s) of nitrate
in ground water. This investigation was aimed toward identify-
ing any limitations and potential applications of the technique.
-------
SECTION 2
CONCLUSIONS
1. Nitrates from septic tanks, feedlots, and barnyards can be
distinguished from natural soil nitrates on the basis of the
isotopic compositions.
2. Nitrates from septic tanks, feedlots, and barnyards cannot
be distinguished from each other.
3. The average nitrogen isotope ratio del value (615N) of
nitrates from municipal wastewater irrigation sites falls
between the average for feedlots and septic tanks and the
average for natural soil nitrates.
4. The range of 615N values of nitrates from municipal waste-
water irrigation sites overlaps the range for feedlots and
septic tanks and the range for natural soil nitrates.
5. Therefore, nitrates from municipal wastewater irrigation
sites cannot be distinguished from nitrates derived from
septic tanks, feedlots, barnyards, or from natural soil
nitrates.
6. The isotopic composition of nitrates from municipal waste-
water irrigation sites suggests that these nitrates are a
mixture of nitrates derived from the waste and nitrates
derived from natural soil processes.
7. Environmental factors influence the 615N value of soil and
ground-water nitrates.
4
-------
8. Only about 30 percent of the observed variation in 615N can
be attributed to the environmental factors assessed in this
study.
9. Identifying causal relationships between 6 * 5N values and
specific environmental factors will require additional
research.
10. The ranges of 615N values for specific nitrate source
categories overlap.
11. Distinguishing between nitrates from animal or human wastes
and natural soil nitrates will therefore require a compari-
son of mean 6l5N values.
12. Analysis of large numbers of samples will improve the abil-
ity to distinguish between mean 615N values.
13. The isotope technique for identifying sources of nitrate
is not suitable for use as a "routine" procedure, because:
a. sophisticated analytical equipment, which is not
commonly available, is required to prepare and an-
lyze samples; and
b. a high level of familiarity with the technique,
the chemistry of nitrogen, and nitrogen isotopic
systematics is required on the part of the inves-
tigator, in order to insure the proper analysis
of samples and interpretation of results.
-------
SECTION 3
RECOMMENDATIONS
Additional information is needed to make the nitrogen isotope
ratio technique a generally applicable tool for tracing nitrate
sources. Studies should be performed to determine whether varia-
bles such as soil oxidation-reduction potential (Eh), pH, and
bacteria exert a major influence on the del value.
These studies also should include variables considered in
this program, e.g. precipitation, elevation, soils, etc. Corre-
lation studies similar to those used in this study should be
performed on the larger data base to determine whether variations
in del value can be explained.
If these expanded studies are successful in explaining the
del values, the results must be studied to develop a practical
scheme which would permit the sampling of a location to identify
the source of subsurface nitrate. This might require on-site
measurements of parameters such as soil Eh and pH.
-------
SECTION 4
RATIONALE FOR TRACING NITRATE
VIA STABLE ISOTOPE RATIOS
NITROGEN ISOTOPE ABUNDANCE
There are two naturally occurring, stable (non-radioactive)
isotopes of nitrogen, l "N and 15N. The nucleus of a * "*N atom
contains seven protons and seven neutrons, while the nucleus of
a l 5N atom contains seven protons and eight neutrons. The nitro-
gen gas in air is composed of 99 . 632 percent 1 4N and 0.368 percent
15N. Junk and Svec analyzed air samples from many localities ana
found no significant variation in the isotope composition of at-
mospheric nitrogen (7). In other nitrogen- containing compounds the
15N occurrence may be difiverent from tliat in air, however.
The 1 5N content of any particular nitrogen species is ex-
pressed in terms of the nitrogen isotope ratio del value (6) ,
in parts per thousand (ppt) , defined as follows:
i5 = 15N/^N (sample) - l 5N/ 1 "N (air)
1514
N/14N (air)
It can be seen that the nitrogen isotope- ratio in air is
used as a reference, and therefore has a 6 value of zero. Sub-
stances with a higher 1 5N content than air will have a positive
6 value, while substances with a lower 1 5N content than air will
have a negative 6 value. Del values of naturally occurring sub-
stances generally range from -20 to +20. Selected nitrogen iso-
tope ratios determined for various substances are shown in Table 1
-------
TABLE 1. NITROGEN ISOTOPE RATIOS FOR VARIOUS SUBSTANCES*
Sample (Source)
Inorganic Nitrogen
Chile nitratite (Chile)
Sal ammoniac (Mexico)
Rocks and Minerals
Pitchblende (Canada)
Granite (Maine)
Oil and Gas
Bitt No. 1 Oil (Oklahoma)
Natural Gas (Arkansas)
Plaisted No. 1 Oil (Oklahoma)
Peat and Coal
Peat (New York)
Bituminous coal (Pennsylvania)
Peat (Eire)
615N (ppt)
-2.6
+13.0
-2.3
-0.2
-11.5
-5.9
-3.5
-2.8
-0.9
+1.9
Animal Protein
Rat liver (Unknown) -3.3 to +9.3
Lamb flesh (Unknown) +5.0
Clam flesh (Atlantic Ocean) +7.3
Marine zooplankton (Unknown) +10.0
Marine fish flesh (Unknown) +15.0
Plant Protein
White Clover leaves (Local) -6.5
Dandelion leaves (Local) -2.8
Oats (Unknown) +6.2
Seaweed (Unknown) +7.0
*From Jones (1) and Kreitler (2).
8
-------
An isotope ratio mass spectrometer is used to measure 6 values.
ISOTOPE FRACTIONATION MECHANISMS
Differences in the nitrogen isotope ratios of different
species occur because of the difference in mass between l5N and
14N atoms. Since an atom of 15N contains an extra neutron, it
has a mass approximately 7 percent greater than a 14N atom. This
greater mass causes 15N atoms to behave slightly differently from
1^N atoms in physical and chemical processes. These behavioral
differences can produce variations in the 15N/1!4N ratios of
different nitrogen compounds (isotopic fractionation).
The isotopic ratio of a nitrogen compound is the result of:
1. Physical fractionation;
2. Chemical equilibrium fractionation;
3. Chemical kinetic fractionation; and
4. The original isotopic composition of the material from
which the compound was formed.
Physical fractionation, chemical equilibrium fractionation,
and chemical kinetic fractionation are all important in the iso-
topic fractionation of nitrogen in natural biogeochemical systems.
However, their relative importance is only beginning to be under-
stood.
Physical fractionation can occur through diffusion, evapora-
tion, and sublimation. In diffusion the light isotope (ll*N) will
have a higher velocity. In evaporation and sublimation the
lighter isotope will have a higher vapor pressure.
Chemical equilibrium fractionation (isotope exchange equilib-
rium) is the tendency of an isotope to concentrate in one species
of a chemical equilibrium reaction. When two chemical species
9
-------
are in equilibrium, the isotopic distribution also reaches an
equilibrium (i.e., the isotopic ratio in one species may not be
the same as the ratio in the the other species). The distribu-
tion of isotopes between the species is dependent on the differ-
ent energy levels of the molecules, and is described by the iso-
tope equilibrium constant (K or a). Isotope equilibrium con-
stants may be calculated or determined experimentally.
The calculated isotope equilibrium constant is "K", where
[Products]
K =
[Reactants ]
The experimentally determined isotope equilibrium constant
is "a". K and a are related as follows:
v-
a = K
where n is the number of equivalent exchangeable atoms in the
reaction .
Isotope equilibrium constants have been calculated from
partition functions and/or determined for several nitrogen equi-
librium reactions relevant to surface or near-surface conditions
These are listed in Table 2.
The meaning of the isotope equilibrium constant may be
illustrated using the second reaction in Table 2:
1 'tMtr i 1 5|arj _ 1 SVTTT i ItMTj
+ NH3(g) - NH4(aq) + NH3(g)
This is the nitrogen isotope exchange equilibrium reaction when
ammonium is in chemical equilibrium with gaseous ammonia (ammonia
volatilization) . From the equation for the calculated equilibrium
10
-------
TABLE 2. ISOTOPE EQUILIBRIUM CONSTANTS*
Reaction
T°K
Isotope Equilibrium Constant
Calculated (K) Experimental (a)
5N2) = !
298.1
1.023
+15NH3 =15NH1| +1J+NH3
(g) (aq) (g)
298.1
1.035
1.034 ± 0.002
NH1|+1 5NO=:
298.1
1.038
298.1
1.012
1UNH3 +15NH3 =15NH3 +llfNH3
(aq) (g) (aq) (g)
298.1
1.005
15
NH
+14NH4 =11*NH3 +15NH!4
(aq) (aq) (aq) (aq)
298.1
1.029
*From Kreitler (2).
-------
constant, the second reaction in the table is written:
K - 1.035 =
(15NH3) (15NH3/ll>NH3)
(aq) (g) (g)
This equation states that there should be a 35 parts per thou-
sand (ppt) enrichment of 1 5N in aqueous ammonium compared to
ammonia gas .
In the equations in Table 2 there is only one exchange atom
(i.e., n = 1) , and the relation between K and a is therefore:
a = K.
For the second reaction in Table 2,
a = 1.034 =
1 5M/ 1 if
N/14N (NH3)
This equation states that a 34 ppt enrichment of l 5N in ammonium
compared to ammonia, was determined.
Chemical kinetic fractionation describes the tendency of
the two isotopes to react at different rates when undergoing
the same reaction. If all the reactants are not converted into
products (because of side reactions or incomplete reactions)
the nitrogen isotope ratio will differ between the reactants and
products. The particular isotope ratio for the products will
depend on (a) the initial isotope ratio of the reactants, and
(b) the particular reaction which occurred.
12
-------
It should be noted that environmental parameters can
affect isotope fractionation. For example, high temperatures
(hundreds of degrees) tend to lessen the tendency of the two iso
topes to react at different rates. Thus, less isotope fraction-
ation would be expected at high temperatures This is because
the slight differences in energy levels (due to the slightly
different masses of the nitrogen atoms) become less important
as the total energy of the system is increased by the added heat
(molecular motion) . Bacterially-aided reactions are another
example of environmental influence on isotope fractionation.
Such reactions often exhibit significant fractionation.
The kinetic fractionation factor B is used to describe
isotopic fractionations in nonequilibrium reactions. Table 3
lists the values for B that have been determined for several
nitrogen reactions, as summarized by Kreitler (2). Since no
single convention for the usage of B has been established, the
equation defining B is given for each of the reactions listed in
Table 3. For example, the equation for B for the first step of
the nitrification process (oxidation of ammonia to nitrite;
reaction #2 in Table 3) is written:
= ^N/^N (NH3) = ,
*N (N02)
The value given for B, 1.026, indicates that there is a 26 ppt
enrichment of l 5N in ammonia relative to nitrite.
NITROGEN IN THE ENVIRONMENT
The Nitrogen Cycle
The earth nay be viewed as having a natural "metabolism",
in which materials are continually circulated through the atmo
sphere; to the oceans via streams; in and out of the biota via
13
-------
TABLE 3. KINETIC FRACTIONATION FACTORS FOR
VARIOUS NONEQUILIBRIUM REACTIONS*
Reaction
Equation for B
Measured
Values of B
1. Nonsymbiotic bacterial
nitrogen fixation.
B =
i S
15
(atmospheric N2)
(fixed nitrogen)
1.0022-1.004
2. Nitrification: two-step
bacterial oxidation of
ammonia to nitrate.
Only the first step has
been measured.
1 5,
B =
1 5 X
(NH3)
(N02)
1.026
3. Denitrification: bacterial
reduction of nitrate to
nitrogen gas.
B =
(NO 3)
(N2)
1.0173-1.02
4. Cation exchange of
ammonium with exchange
resins or kaolinitic
clay.
B =
(solution NHQ
15N/1!tN (adsorbed
0.99926-0.99922
5. Anion exchange of
nitrate with exchange
resin.
B =
15N/1[tN (solution N03)
15N/lltN (adsorbed N03)
1.0021
*Reported by Kreitler (2).
-------
photosynthesis, growth, and decay; and through the lithosphere
via sedimentation, weathering, and erosion. Each element follows
a particular path or set of pathways through the natural system.
The paths are determined by the biogeochemical and physical prop-
erties of the elements. Each has its own natural biogeochemical
cycle.
The global nitrogen cycle is extremely complex, and many
of its aspects are poorly understood. A simplified schematic
of the nitrogen cycle is shown in Figure 1. The figure shows
the various nitrogen reservoirs (e.g., the atmosphere, the bio-
sphere, sedimentary rocks, etc.) and the gross aspects of nitro-
gen transfer between these reservoirs.
The major nitrogen reservoir is the atmosphere. In this
respect, nitrogen is somewhat unique. Other elements for which
global cycles have been investigated (e.g., carbon, sulfur, and
phosphorus) have their major inventories in the sedimentary rock
reservoir (8). Several nitrogen species are significant in the
atmosphere. Of these, however, nitrogen gas is by far the most
abundant, accounting for more than 99.999 percent of all atmos-
pheric nitrogen. Ammonia, ammonium, and oxides of nitrogen are
also present, but together they account for less than 0.001 per-
cent of the atmospheric nitrogen (8).
The transfer of nitrogen from one reservoir to another is
indicated in Figure 1 by the various arrows. The mechanisms by
which nitrogen is transferred are very complex, involving a
large number of pathways and many nitrogen species. Many of
the mechanisms are only beginning to be understood. The situa-
tion is complicated by such factors as the recycling of nitrogen
in the biosphere, the complex photochemical reactions involving
nitrogen oxides in the atmosphere, and the influence of man's
15
-------
Uplift,
Weathering
Deposition
Igneous
Rocks
Sedimentary
Rocks
Figure 1. Simplified Schematic of the Nitrogen Cycle,
16
-------
activities (e.g., fixation of atmospheric nitrogen gas for
fertilizer manufacture).
Nitrate In the Terrestrial Environment
In the context of this study, the significant portions
of the nitrogen cycle are those that lead to the presence of
nitrate (N03) in soils and ground waters.
The major input of nitrogen to the soil-water-biomass
subsystem of the nitrogen cycle is the fixation of atmospheric
N2. This is accomplished primarily through natural bacterial
fixation. Within the last 100 years, however, the demand for
fixed nitrogen compounds as fertilizers has grown to the point
where the. amount of industrially fixed atmospheric N2 nearly
equals the amount fixed naturally (8).
Other somewhat less important inputs to the soil-water-
biomass system also exist. Ammonium and nitrate are directly
deposited on the land surface by rainout of both natural and
pollutionally derived ammonium nitrate and ammonium sulfate.
Estimates of the rate of addition of nitrogen by this mechanism
are very uncertain, but some estimates suggest that it may be
significant.
Nitrogen may also enter the soil-water-biomass system via
the weathering of nitrogen-bearing rocks. This appears to be
a very minor source, however. The rate of input, estimated on
the basis of global weathering rates and the average nitrogen
contents of rocks, is only about 2 percent of the rate of input
from atmospheric sources (8) .
Assuming the atmosphere to be the important source of
nitrogen, several reaction paths for the production of nitrate
can be visualized. The most reasonable paths are shown in Figure
17
-------
2. These include (a) nitrate from fertilizer, (b) nitrate due to
fixation by free-living bacteria, (c) nitrate from decayed plants,
(d) nitrate from legumes, (e) nitrate from animal wastes, and (f)
nitrate from human wastes. Not all possible paths are shown,
e.g., rainout of atmospheric ammonium nitrate.
An examination of Figure 2 shows that the overall reaction
schemes can be very complex due to possible recycling of the
nitrate. This is only true at depths within the root zone of
plants, however. Once the nitrate is flushed beyond the root
zone it should be unavailable for further recycling. It should
be noted that many of the reactions shown are caused or aided
by bacteria.
It can be seen in Figure 2 that the ammonia and nitrate
from human and animal wastes are not assumed to participate in
recycling. This is because these wastes are concentrated at
particular sites (e.g., barnyards or septic tanks) where plants
either do not grow or are not numerous enough to consume signif-
icant amounts of the nitrate.
NITRATE SOURCE DETERMINATION
The technique under investigation in this study relies
on the dependence of the nitrogen isotope ratio of a particular
compound (specifically, nitrate) on the type of reaction or series
of reactions which formed that compound. A complete knowledge
of all possible nitrate formation pathways, and of the nitrogen
isotopic fractionations that occur along those pathways, would
enable one to predict the isotopic composition of any particular
nitrate sample. Such detailed information on the nitrogen iso-
tope behavior of nitrate and its precursors does not exist,
however.
18
-------
Industrial fixation
V
Legumes plus
rhizodium bacteria
Bacteria:
Azobacter,
Clostridium
V
Commercial
Fertilizer
v/
Animals
Animal
Wastes
V
Human
Wastes
Oxidation
Oxidation
\K
X*
NO 3
Oxidation
C D E
Figure 2. Mechanisms of Nitrate Formation.
-------
The technique for identifying sources of nitrate in ground
water is based on more of a "black box" approach. It is sufficient
to know (a) that isotopic fractionations exist for various reac-
tions (such as those presented in Tables 2 and 3), and (b) that
nitrate may be formed by different reaction pathways (such as
those shown in Figure 2) that depend, at least in part, on the
source of the nitrogen. As an example, nitrate formed during
the manufacture of fertilizer should go through a different
series of reactions than nitrate formed from animal wastes, and
should therefore exhibit a different nitrogen isotope ratio.
Jones (1) and Kreitler (2) tested this hypothesis in Runnels
County, Texas. They found that nitrates derived from septic
tanks and animal wastes had a different nitrogen isotope ratio
from nitrates derived from "natural" processes (i.e., bacterial
fixation and plant uptake, growth, and decay). This was demonstrated
by obtaining samples of soil nitrate from sites where the source
of the nitrate could be reasonably predicted. Nitrate derived
from the decomposition of animal wastes (including human wastes)
was found to have a characteristic isotopic composition (615N) of
from +10 ppt to +22 ppt, while nitrate derived from natural
soil-bacteria-vegetation systems had an isotopic composition of
from +2 ppt to +8 ppt.
It is possible that these results are applicable only to
the particular systems studied in Runnels County, Texas. Environ-
mental factors, such as climate and soil type, could affect the
isotopic composition of a soil nitrate sample by affecting the
types of bacteria present and the reaction pathways leading to
nitrate formation.
One purpose of this study is to determine whether environ-
mental factors influence the isotopic composition of nitrates
derived from a particular waste source category (e.g., septic
tanks). This is accomplished by analyzing soil and ground-water
20
-------
samples obtained from various sites across the country, represent-
ing a range of environmental conditions.
The other major goal of this study is to determine the ex-
tent to which nitrates from various sources can be differentiated
on the basis of their isotopic composition. The work of Kreitler
and Jones indicated that for one geographic area nitrate from
animal and human wastes can be distinguished from nitrate derived
from "natural" soils or fertilizers. This study addresses the
question of whether or not nitrate from various sources can be
distinguished on the basis of nitrogen isotope ratio for the
country as a whole.
OTHER STABLE ISOTOPES AS TRACERS
It may be possible to identify sources of ground-water
contamination using stable isotopes of elements other than
nitrogen. In the case of nitrate (N03), there is only one other
element (oxygen) to consider. Other contaminants (e.g., phos-
phates, sulfates and sulfides, detergents, and other organic
compounds) contain elements for which isotopic studies may
prove useful, however. These elements include hydrogen, oxygen,
carbon, sulfur, chlorine, and some trace metals.
The usefulness of these elements as waste-source indicators
is difficult to assess. It should be remembered that for stable
isotope identification of waste sources to be successful, it is
necessary that relatively unique isotope fractionations exist
for given waste types or waste sources. This has not been
established for the elements mentioned above.
A complete theoretical evaluation of the potential for
using stable isotopes of elements other than nitrogen is beyond
the scope of this study. Rather, brief discussions are provided
21
-------
for several elements that are considered to be "possible candidates"
for this use.
Hydrogen
There are two stable isotopes of hydrogen, JH and 2H (2H
is commonly called deuterium, and is designated by the symbol
"D"). The nucleus of a 1E atom contains a single proton, while
the nucleus of a D atom contains a proton and a neutron.
The relative mass difference between :H and D is larger
than that of any other pair of stable isotopes of other elements:
the mass of the JH nucleus is essentially half that of the D
nucleus. Because of this, hydrogen exhibits the largest fraction-
ations of any stable isotopes. Fractionations of several hundred
parts per thousand have been observed for various earth materials
(9).
The average natural abundances of the stable hydrogen iso-
topes are as follows:
1H: 99.9844%
D: 0.0156 %.
Hydrogen is an extremely abundant element in natural bio-
geochemical systems. It occurs in water, clay minerals, dis-
solved inorganic compounds and organics. Appreciable fractiona-
tions have been observed in isotope exchange equilibrium reactions,
and significant kinetic fractionations seem possible in biochem-
ical processes (e.g., photosynthesis and bacterial production of
H2 and CHO(9) .
Carbon
Carbon has two stable isotopes: 12C and 13C. Carbon-12
accounts for 98.89 percent of natural carbon. Natural frac-
22
-------
tionations of greater than 10 percent have been observed (9).
Extensive fractionations take place in the biosphere, during
photosynthesis and metabolism. Environmental factors (e.g.,
temperature and pH) have also been observed to influence carbon
isotope fractionations (9).
Carbon occurs extensively in the environment as organic
compounds, carbon dioxide (C02)gas, and bicarbonate (HCOa) and
carbonate (C03) ions. Kinetic fractionations are significant in
the system: atmospheric C02 - dissolved HCO~ ions - Calcium car-
bonate (CaC03) (9).
Oxygen
Oxygen has three stable isotopes. Their natural abundances
are as follows:
160: 99.7637.
170: 0.0375%
180: 0.1995%
In oxygen stable isotope studies the 180/160 ratio is usually
measured, due to the higher abundance and greater mass difference
(9).
Oxygen plays a critical role in natural biogeochemical pro-
cesses. It is the most abundant element on earth and occurs in
gaseous, liquid, and solid compounds. Natural variations in the
180/160 ratio of up to about 10 percent have been measured.
Isotope equilibrium excnange fractionations are significant
in inorganic geochemical systems. The 180/160 ratios of mineral
pairs are temperature sensitive, and oxygen isotope studies have
been used extensively to estimate the "temperature of formation"
of various mineral samples. Chemical kinetic fractionations of
23
-------
oxygen isotopes occur during biochemical processes (e.g., photo-
synthesis and respiration)(9).
Sulfur
There are four stable isotopes of sulfur: 32S, 33S, 3"S,
and 36S. 32S and 3"*S are most abundant:
32S = 95.02%
3*S = 4.21%
These are the isotopes that are generally of interest in sulfur
isotope studies.
Sulfur is present in nearly all biogeochemical environments,
in a number of forms including sulfides, organic substances,
elemental sulfur, and sulfates. Natural variations in 31|S/32S
ratios on the order of 15 percent have been measured. Kinetic
effects and isotope exchange reactions both produce isotope
ratio variations. Kinetic effects are particularly noticeable
in biochemical reactions, such as the bacterial reduction of sul-
fate to hydrogen sulfide (H2S). Isotope exchange between sulfate
and sulfide and between different sulfide species is significant
in geochemical systems (9).
Chlorine
The natural abrndances of the two stable chlorine isotopes
are:
35C1: 75.53%
37C1: 24.47%
The chloride ion is a major anion in natural waters, and is
present in relatively high concentrations in waste materials
24
-------
(see, for example, the analysis of septic tank effluent in Table
B-l, Appendix B). Theoretically one might expect to see small
variations in the isotopic composition of chlorine. No signifi-
cant fractionations have been observed, however (9).
Trace Elements
Many of the more environmentally significant trace elements
have more than one stable isotope: chromium (4 isotopes), iron
(4 isotopes), nickel (5 isotopes), copper (2 isotopes), zinc
(5 isotopes), selenium (6 isotopes), silver (2 isotopes), cadmium
(7 isotopes), tin (10 isotopes), mercury (7 isotopes), and lead
C3 isotopes). Some of these may be useful in tracing pollution
sources. Little, if any, previous work has been done in this
area with these elements, however.
Summary
From a theoretical standpoint, stable nitrogen isotopes ap-
pear to be most useful for identifying pollutant sources that
contribute nitrate to subsurface water-soil systems. The re-
stricted sources of nitrogen and the better defined pathways
reduce the uncertainty inherent in this approach. Especially
helpful in this regard is the fact that most of the nitrogen
in natural soils is ultimately derived from a single source
(the atmosphere) which has a uniform nitrogen isotopic composi-
tion.
25
-------
SECTION 5
EXPERIMENTAL PROCEDURES
The procedures followed in selecting sampling sites and in
collecting, preserving, transporting, storing, and analyzing
samples are discussed below.
SELECTION OF SAMPLING SITES
Samples of soil and/or ground water were obtained from nu-
merous sites across the United States (see Section 6). The pro-
cess of selecting these sites was based primarily on a review of
published literature relating to nitrate contamination of ground
waters. Over 300 references were assembled in the course of this
review. To the extent possible, these references were subdivided
as to suspected, presumed, or known sources of the nitrate and as
to their geographic distribution within the source categories.
The objective of the geographic classification was not to select
sampling sites in all areas of the country, but to identify sites
that were representative of a wide range of environmental con-
ditions (e.g.,, climate, soil type, latitude, elevation). This
classification enabled numerous candidate sites to be identified.
Final site selection was based on an assessment of how well
the candidate sites met the set of selection criteria described
below. In some cases, inquiries were directed toward cognizant
local personnel to better define the suitability of a site for
sampling.
26
-------
In order for a candidate site to be selected for sampling,
the site must first have been characterized as having a single,
identified source of environmental nitrate. Three sources of
nitrate were of primary interest: domestic septic tanks, animal
feedlots, and municipal wastewater land application operations.
Particular emphasis was placed on finding sites where domestic
septic tanks were considered to be the known or probable source
of any environmental nitrate.
Additional criteria were that the final choice of sites
should provide a wide geographic representation and that the
geologic and soil conditions at the sites be well-characterized.
Finally, the sites had to be accessible for sampling.
Fourteen sites were chosen for intensive sampling. Samples
of soil and ground water were collected from these sites by
Radian personnel according to the procedures discussed below. An
additional five sites were selected, to be sampled by local per-
sonnel of the United States Geological Survey, various state
agencies, or universities. These sites were selected primarily
to provide additional geographic coverage, and were sampled less
intensively than the others.
Data on five other sites were obtained from Kreitler (per-
sonal communication, 2). These data provided additional geo-
graphic coverage and enlarged the final data base, particularly
for the animal waste source category. They also provided cover-
age of "natural" sources, i.e., sites where the nitrates were
derived from the natural growth and decay of plants.
The various sites and the samples obtained from them are
discussed in Section 6. More detailed descriptions of the 14
intensively sampled sites are given in Appendix B.
27
-------
SAMPLING PROCEDURES
Ground-Water Samples
Ground-water samples were collected from existing wells in
the vicinity of the sites whenever wells were available for sam-
pling. In a few cases, water samples were also obtained from
augered soil sampling holes. At two sites, well points were
driven to collect ground-water samples.
The procedure for collecting samples from wells was straight-
forward. Samples were collected from points as near to the
wellhead as possible. The presence of any treatment or storage
systems (e.g., water softeners or pressure tanks) between
the wellhead and the sampling point was noted, along with any
other information about the well that was available (such as
well type, use, date installed, diameter, and depth). Prior to
collecting a sample, the well was pumped for a period of at least
five minutes to assure the collection of a fresh water sample.
Information on well construction is needed because the man-
ner of well completion can affect the observed 615N value of
nitrate in the well water. For example, a well completed in
two distinct aquifers may yield water containing nitrate from
two distinct sources.
In many cases, however, details of well construction were
lacking. All wells sampled were assumed to tap a single aquifer,
based on observations of depth, well diameter, use, and pump
size.
Water samples from augered soil sampling holes were obtained
with a small peristaltic pump or by bailing with a small bucket.
The holes were bailed out several times before a sample was col-
lected. Water samples obtained with well points were collected
28
-------
with a small peristaltic pump. The pump was run for several
minutes before a sample was collected to flush the pump and hoses,
All water samples were collected into one-liter polyethylene
bottles.
Soil Samples
Most soil samples were collected using a 4-inch diameter
soil auger. Some near-surface samples were obtained using a
shovel. The depth intervals from which the samples were col-
lected were recorded. Soil samples were placed in plastic bags
for shipping and storage.
At one site, a set of 1.25-inch diameter cores was obtained.
These samples were obtained by driving a core-barrel fitted with
a tapered cutting bit. The core barrel was lined with 1.25-inch
thin-wall PVC pipe. When the maximum depth had been reached,
the PVC pipe, with the core-sample inside, was extracted. The
core was cut into convenient lengths, and the ends of the PVC
tube were sealed.
Sampling Kits
Sampling kits were sent to local personnel at the five sites
not visited by Radian personnel (see Section 6). These kits were
designed to supply all of the materials required for collecting
a sample. Each kit consisted of a styrofoam-lined shipping con-
tainer, water sample bottles, labels, data sheets, vials of sam-
ple preservative, plastic soil sample bags, and an instruction
sheet.
Sample Preservation, Transport, and Storage
Water samples were preserved by adding 10 milliliters (ml) of
1 normal (N) sulfuric acid (t^SO^) to each liter of sample. The
29
-------
samples were refrigerated, frozen, or placed on ice as soon after
collection as possible. Deviations from this procedure are noted
in the descriptions of the sampling sites given in Appendix B.
No preserving agents were applied to the soil samples. These
samples were refrigerated and/or frozen as soon after collection
as possible. Usually this was within an hour, and in no case
were samples left unrefrigerated for more than eight hours.
Coolers and dry ice were taken to the sites when cold storage
facilities were unavailable.
The samples were shipped , by either air or surface carrier,
packed in dry ice. Upon receipt in Austin some frozen samples
were found to have thawed in transit, but no samples had reached
ambient temperature.
All samples were kept frozen after they reached Austin. Samples
were not thawed until immediately prior to analysis. Frozen samples
were stored for up to nine months before being analyzed.
SAMPLE PREPARATION AND ANALYSIS
Isotope analysis is most conveniently performed on nitrogen
gas (N2). This means that all nitrate samples must be reduced
to nitrogen gas, and this conversion must be performed without
introducing any isotope fractionation during the sample prepara-
tion. Once the nitrate is converted to nitrogen gas, the gas
must be purified to remove species such as oxygen which might
damage the filaments in the mass spectrometer. The techniques
which were developed for sample preparation were as follows.
Soil Sample Leaching
Prior to analysis, the soluble nitrogen species were ex-
tracted from the soil samples by leaching. A one-kilogram
30
-------
aliquot of soil was placed in a 4.2 liter polyethylene wide
mouth bottle. One liter of deionized water was added and the
bottle was placed on an orbital shaker table for four hours.
After four hours of shaking, the sample was removed and
allowed to settle for 15 to 20 minutes. A 50 ml aliquot of
the supernatant liquid was analyzed for nitrate with a specific-
ion electrode. Approximately 65 milligrams (mg) of silver sulfate
were added to minimize chloride interference during the nitrate
concentration determinations.
Those samples that yielded a leachate nitrate concentration
of less than 5 mg/£ were not analyzed further. A concentration
of at least 5 mg/Jl was found to be necessary in order to obtain
a sufficiently large nitrogen gas sample for isotopic analysis.
Those soil samples yielding a leachate nitrate concentration
greater than 5 mg/£ were filtered using an 11-centimeter Buchner
funnel and a 1,000-ml vacuum flask attached to an aspirator. Di-
atomaceous earth was added as a filter aid. The subsequent sample
preparation procedures were identical for soil sample leachates
and ground-water samples.
Sample Preparation
The procedures followed in preparing samples for nitrogen
isotopic analysis were essentially the same as those described
by Krietler (2). Since only the isotopic ratio of the nitrate
was of interest, other species of nitrogen had to be removed.
These included ammonia and organic nitrogen. Removal of these
species and extraction of nitrate was accomplished using a two-
stage distillation process. The apparatus used for this process
is illustrated in Figure 3.
31
-------
DISTILLATION
FLASK
HOT PLATE
SUPPORT
PLATE
MAGNETIC
STIRRER
BLOCK
SUPPORT
DISTILLATION
COLUMN
Figure 3. Distillation Apparatus Used for Removal
of Interfering Nitrogen Species.
32
-------
The pH of the sample (soil leachate or ground water) was
first raised to between 9.5 and 10.0 by addition of powdered
magnesium oxide (MgO). The solution was agitated to suspend
the MgO (see Appendix A), transferred to a two-port round bottom
distillation flask, and distilled until 50 ml of distillate
were collected. This caused the conversion of NH4 to NH3 and
the breakdown of any labile organic compounds to ammonia. At
this stage, interfering forms of nitrogen were removed from
the sample, having been collected in the distillate.
The hot plate was then replaced with the magnetic stirrer,
and the exit port of the water-jacketed distillation column was
immersed in a beaker containing 50 ml of 0.1 normal hydrochloric
acid (HC1) (see Appendix A). Five grams of finely ground De-
varda's Alloy (see Appendix A) were added through the side port
of the flask. The flask was immediately restoppered and the
solution magnetically stirred for 10 minutes without heating.
This step reduced the nitrate in the sample to ammonia. Finally,
the solution was redistilled until 75 ml of distillate were col-
lected in the 0.1 HC1 solution. Relatively concentrated HC1
was used to insure that all ammonia was absorbed, producing an
acidic ammonium chloride solution. The volume of this solution
was reduced to 5 to 10 ml by heating on a hot plate at 70°C to
80°C.
Nitrogen Gas Generation
A special nitrogen gas generation and purification apparatus
was required to further prepare a sample for isotopic analysis.
A schematic of the apparatus is shown in Figure 4. The appara-
tus consists of the following: a nitrogen gas generation tube,
a freeze-thaw manifold, two furnaces (containing copper and
copper oxide, respectively), a Toeppler pump, a gas collection
flask, two mechanical vacuum pumps, and a mercury diffusion pump.
33
-------
VACUUM
THERMOCOUPLE
FREEZE-THAW MANIFOLD
GAS
COLLECTION
FLASK
MERCURY
DIFFUSION
PUMP
NITROGEN
GAS
GENERATION
TUBE
Figure 4. Nitrogen Gas Generation
and Purification Apparatus
MECHANICAL VACUUM PUMPS
-------
After the ammonium chloride solution was concentrated to
an appropriate volume (5 to 10 ml), as described above, it was
transferred to the large volume side of a nitrogen gas generating
tube (see Figure 5). Approximately 5 ml of sodium hypobromite
(NaOBr) solution (see Appendix A) were added to the small side-
arm. This assembly was attached to the freeze-thaw manifold of
the vacuum preparation apparatus.
NH/.CL
NAOBR
Figure 5. Nitrogen Gas Generation Tube
The two separate solutions were frozen by immersing the
generating tube in an equilibrated dry ice-acetone mixture.
Once frozen, the stopcock on the reaction tube was opened to
the vacuum system allowing the atmosphere above the frozen
solutions to be withdrawn. The stopcock was reclosed, and the
35
-------
solutions were allowed to thaw. This sequence was carried out
three times to completely degas the solutions and thereby elim-
inate atmospheric nitrogen contamination.
The nitrogen gas generation tube was transferred to the
sample injection port below stopcock D (see Figure 4). With the
gas generation tube isolated from the rest of the system (i.e.,
stopcock D closed) , the side-arm was rotated 180° to drain the
sodium hypobromite solution into the ammonium chloride solution.
The resulting reaction liberates nitrogen gas:
2NH3 + 3NaOBr - SNaBr + 3H20 + N2
N2 Gas Purification
When the N2 gas generation reaction was complete, stopcocks
D and F were opened and the Toeppler pump was activated (stop-
cocks E, G, and J closed). The nitrogen gas was pumped from
the generation tube, through the furnaces, and into the tubing
segment bounded by stopcocks G, I, and J. When a pressure of
3 to 5 mm Hg was accumulated (as indicated by vacuum thermocouple
#2) , stopcock F was closed and stopcock G was opened. A closed
loop was thus created through the furnaces and the Toeppler pump.
The nitrogen gas was circulated through the hot copper
furnace (T = 400°C), the hot copper oxide furnace (T = 800°C),
and a liquid nitrogen cold trap for one hour. This eliminated
any oxygen, carbon monoxide, and gaseous oxides of nitrogen that
may have been present.
To collect the purified nitrogen sample the three-way stop-
cock I was turned, which redirected the purified sample into the
gas collection flask. The stopcock on the collection flask as
well as the stopcock on the collection port (0) were then closed
and the gas collection flask removed.
36
-------
Isotopic Analysis
Purified nitrogen gas samples were analyzed for their nitro-
gen isotope compositions by personnel of the Center for Research
in Water Resources of the University of Texas at Austin. The
isotope ratio was measured by comparing mass 28 (14N14N) to mass
29 (15N14N) in a double-collector, gas-source, mass spectrometer,
using atmospheric nitrogen as a standard.
ANALYSIS OF STANDARDS
The analytical technique used in this study was perfected
by Kreitler (2), who found that the experimental error in the
technique was about ±1 ppt. Careful adherence to the details
of the procedure is necessary to achieve this level of precision,
however. Deviations from the procedure can introduce isotopic
fractionations in the sample.
An additional check on the reliability of the analytical
procedures was made in this study by running standards of known
isotopic composition. A set of potassium nitrate (KN03) standards
was supplied by Kreitler. Four of these were run through a dis-
tillation, preparation, and purification system. The resulting
nitrogen gas samples were returned to Kreitler for isotopic anal-
ysis. Kreitler compared the 615N values of the investigator's
standards with the results of a set of analyses that were per-
formed on the same material in Kreitler's laboratory. Kreitler
concluded that the 615N values were in agreement, within the
limits of the mass spectrometer, and that the investigator's
preparation techniques were acceptable.
Additional standards were periodically run during the course
of the program. These standards were prepared from a stock solu-
tion of KNOs provided by Kreitler. A sample of atmospheric
nitrogen was also run.
37
-------
All of these standards, except one, yielded 615N values
within ±0.5 ppt of the known 615N values. The one exception was
a KN03 standard (615N = -0.7) which yielded an apparent 615N value
of -10.7 ppt. Calculations involving the mass of nitrogen in the
standard, the mass of Devarda's Alloy in the nitrate-to-ammonium
reaction, and the mass of NaOBr in the ammonium-to-nitrogen gas
reaction indicated that incomplete conversion reactions were not
the cause of this discrepancy. This data point is viewed as an
analytical "outlier," since none of the known opportunities for
isotopic fractionation in the analytical procedures could be
blamed for the discrepancy.
As an additional check on the procedure, the measured
<51 5N values of the samples prepared by Radian were examined as
a function of the time during the program in which the samples
were analyzed. Most of the analytical difficulties experienced
(see Section 7) occurred in the early stages of the program.
The examination of 615N versus time was to determine whether
there was any consistent and/or progressive error in the proce-
dures as the program progressed. No relationship was found
between 6l5N and the time a sample was analyzed.
38
-------
SECTION 6
SAMPLING SITES
The intensive field sampling program carried out by Radian
personnel resulted in the acquisition of 324 discrete soil and
water samples, plus a total of about 45 feet of 1%-inch core.
Two hundred forty-three of the discrete samples were soils;
eighty-one were water samples.
Fourteen sites were visited in the intensive program:
! Texas A&M University (TAMU) at College Station, Texas;
2- Rockdale , Texas ;
3* Tampa, Florida (2 sites);
4- Indian Hills, Colorado;
5- Pueblo, Colorado;
6- Houghton Lake, Michigan (2 sites);
7* Tallahassee, Florida;
8- Abilene, Texas;
9- Michigan State University (MSU) at East Lansing, Michigan;
10- San Angelo, Texas;
39
-------
!! Greeley, Colorado; and
12* Dimmitt, Texas
These sites are associated with three nitrate source categories:
domestic septic tank systems (7 sites), municipal wastewater
land application systems (5 sites) , and cattle feedlots (2 sites)
A breakdown of the samples collected from each of these sites
is given in Table 4.
Sampling kits were sent to an additional 5 sites, all of
which are associated with septic tank nitrate sources. Provi-
sion was made in each kit for the collection of 3 samples.
In addition to these 19 sites, data were obtained from Dr.
C.W. Kreitler, of the Texas Bureau of Economic Geology, for
samples that he obtained from 5 other sites. The Queens County,
New York, and Grand Cayman Island sites are believed to represent
septic tank sources of nitrate. The Macon County, Missouri, and
Lockhart-Taylor, Texas,sites are believed to represent animal-
waste nitrate. In Runnels County, Texas, data for septic tank,
animal waste (barnyards), and natural nitrate sources were ob-
tained. The geographic locations of the 24 sites are shown in
Figure 6. A key identifying each site shown in the figure is
given in Table 5.
Environmental factors were assessed for each of the 24 sites
The environmental variables considered were related to climate
and soil type:
mean annual precipitation,
mean annual air temperature,
40
-------
TABLE 4. SUMMARY OF SAMPLES COLLECTED IN
RADIAN'S INTENSIVE SAMPLING PROGRAM.
Location
TAMU
Rockdale
Tampa
Indian Hills
Houghton Lake
(A)
Houghton Lake
(B)
Pueblo
Tampa
Tallahassee
Abilene
San Angelo
MSU
Greeley
Dimmitt
N03 Source
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Wastewater
Wastewater
Wastewater
Wastewater
Wastewater
Feedlot
Feedlot
No. of Soil
Samples
8
23
64
5
8
11
10
21*
19
16
15
16
13
14_
243
Total Combined
Depth
Interval
2.0-4.5 ft.
1.0-5.0 ft.
0.0-11.5 ft.
0.5-5.1 ft.
0.0-4.9 ft.
0.0-5.7 ft.
0.1-10.9 ft
0.0-20.0 ft.
0.0-9.5 ft.
0.0-7.2 ft.
0.5-6.4 ft.
0.0-6.2 ft.
0.0-6.0 ft.
0.0-10.3 ft.
No. of Water
Samples
12
0
8
6
0
0
1
29
6
6
4
7
1
1
81
*Plus four 1 1/4-inch cores from the surface to a maximum depth of 20 feet.
41
-------
Figure 6. Locations of sampling
sites (see key in Table 5)
-------
TABLE 5. KEY TO FIGURE 6.
Map
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Location
Texas A&M University
Rockdale, Texas
Tampa, Florida
Indian Hills, Colorado
Houghton Lake, Michigan, Site A
Houghton Lake, Michigan, Site B
Pueblo, Colorado
Tampa, Florida
Tallahassee, Florida
Abilene, Texas
San Angelo, Texas
Michigan State University
Greeley, Colorado
Dimmitt, Texas
Wahkon, Minnesota
College Station, Texas
Jefferson County, West Virginia
Joplin Area, Missouri
Portland Area, Oregon
Lockhart-Taylor Area, Texas
Grand Cayman Island
Macon County, Missouri
Queens County, New York
Runnels County, Texas
N03 Source
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Wastewater
Wastewater
Wastewater
Wastewater
Wastewater
Feedlot
Feedlot
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Septic Tank
Barnyard
Septic Tank
Barnyard
Septic Tank
Septic Tank,
Barnyard , and
"natural"
Sampled
By*
A
A
A
A
A
A
A
A
A
A
A
A
A
A
B
B
B
B
B
C
C
D
D
D
^Samples by: A = Radian Intensive Sampling Program; B = Radian Kit Sampling
Program; C = Kreitler (unpublished data); D = Kreitler (2).
43
-------
Koppen Climate Classification,
regional soil type,
elevation, and
latitude.
The values of these parameters are listed in Table 6 for each
of the sites. The Koppen Climate Classification system and Soil
Classification system are described in Appendices E and F,
respectively.
44
-------
TABLE 6. SITE ENVIRONMENTAL CHARACTERISTICS
Site Location
Texas A&M University
Rockdale, Texas
Tampa, Florida (2 sites)
Indian Hills, Colorado
Houghton Lake, Michigan, Site A
Houghton Lake, Michigan, Site B
Pueblo, Colorado
Tallahassee, Florida
Abilene, Texas
San Angelo, Texas
Michigan State University
Greeley, Colorado
Dimnitt, Texas
Wahkon, Minnesota
College Station, Texas
Jefferson Co. , West Virginia
Joplin Area, Missouri
Portland Area, Oregon
Lockhart-Taylor Area, Texas
Grand Cayman Island
Macon Co., Missouri
Queens Co. , New York
Runnels Co., Texas
Sampled
ByL
A
A
A
A
A
A
A
A
A
A
A
A
A
B
B
B
B
B
C
C
D
D
D
Average
Annual
Temperature
68
67
72
42
43
43
51
58
64
65
47
46
56
39
68
52
57
54
68
82 1!
52
50
66
Average
Annual Koppen
Precipitation Climate
(Inches) Classification*
36
34
52
24
30
30
12
56
24
20
31
12
18
28
36
40
40
48
32
80//
36
44
24
Caf
Caf
Caf
H
Dbf
Dbf
BSk
Caf
BSk
BSh
Daf
BSk
BSk
Dbf
Caf
Daf
Caf
Cb
Caf
Au
Daf
Daf
BSh
Elevation
(Feet)
200
490
50
7,200
1,150
1,270
3.500//
250)?
1,700
1,800
870
4,630
3,900
1,500*
200
500*
1,500*
250*
500
50*
750l?
1000
1,600
Latitude
(Degrees)
30.50
30.75
28.00
39.75
44.25
44.25
38.25
30.25
32.50
31.50
42.75
40.50
34.50
46.00
30.50
39.50
37.00
45.50
30.25
19.00
39.75
40.75
31.75
Soil
Type4
5
5
3
]0
1
1
6
3
6
6
2
6
6
1
5
2
3
10
5
4
5
2
6
Sampled by: A = Radian Intensive Sampling Program; B = Radian Kit Sampling Program;
C = Kreltler (unpublished data); D = Kreitler (2).
*See Appendix E for definitions of Climate Classes.
+See Appendix F for definitions of Soil Classes.
^Estimated value.
-------
SECTION 7
RESULTS AND DISCUSSION
STATISTICAL ANALYSES OF THE DATA
The analytical data used in this study are tabulated in
Appendix D. This data base includes the results of the isotopic
analyses performed under this study, and also includes both pub-
lished (2) and unpublished data provided by Dr. C.W. Kreitler
of the Texas Bureau of Economic Geology. The distribution of
the data by nitrate source category and site is presented in
Table 7, It should be noted that not all of the samples collected
in Radian's sampling program were analyzed, and that some sites
are not represented in the data base. The reasons for this are
discussed in the following subsection on Analytical Problems.
This data base was analyzed using several statistical
techniques. These included an analysis of means and variances
of 615N for the various waste source categories, preparation of
histograms, and correlation and regression analysis.
Correlation, multiple correlation, and partial correlation
analyses were performed for the entire data base and for each
separate nitrate source category. The analysis of the entire
data base examines the "behavior" of
-------
TABLE 7. SUMMARY OF SAMPLES INCLUDED IN THE FINAL
DATA BASE, BY SOURCE CATEGORY AND SITE.
NOa Source Category
Septic Tanks
Animal Wastes (Feedlots
and Barnyards)
Municipal Wastewater
Irrigation
Natural Soil-Plant Systems
Site
Houghton Lake, Michigan, Site A
Houghton Lake, Michigan, Site B
Puehlo, Colorado
Tampa, Florida
Indian Hills, Colorado
Rockdale, Texas
Texas A&M University
Grand Cayman Island
Macon County, Missouri'
Queens County, New York
Runnels County, Texas
Subtotal
DJmmltt, Texas
Greeley, Colorado
Lockhart-Taylor Area, Texas
Macon County, Missouri
Runnels County, Texas
Subtotal
Abilene, Texas
San Angelo, Texas
Michigan State University
Subtotal
Runnels County, Texas
TOTAL, ALL SOURCE CATEGORIES AND SITES
Number of
Analyses
6
4
6
13
5
4
3
4
1
2
4
52
8
7
11
10
16
52
2
5
_3
10
J*
178
Analyzed
By*
A
A
A
A
A
A
A
B
C
C
C
A
A
B
C
C
A
A
A
C
*Analyzed by: A = Radian; B = Kreitler, unpublished data; C = Kreitler (2).
-------
separate correlation analyses examine the extent to which this
observed variation in 615N for specific nitrate sources can be
explained in terms of the various environmental factors.
Descriptive Statistics for j>
15N
The means and standard deviations of the isotope ratios
of the samples were computed for each of the nitrate source
categories. These statistics are presented in Table 8.
The means and variances of 615N for the different source
groups were compared using a one-way analysis of variance (F-
test) and other statistical tests (t-tests) . This analysis in-
dicated that the mean 615N values for the septic tank (+10.9),
animal waste (+12.4), and municipal wastewater irrigation (+9.5)
source categories do not differ significantly (i.e., these sources
of nitrate cannot be distinguished on the basis of 615N values).
Moreover, the mean 615N values for the irrigation source (+9.5)
and "natural" nitrate sources (+7.3) do not differ significantly.
TABLE 8. MEANS AND STANDARD DEVIATIONS OF 615N
FOR DIFFERENT NITRATE SOURCE CATEGORIES
Mean Standard Number of
N03 Source (PPt) Deviation Data Points
Septic Tanks +10.9 9.8 52
Animal Wastes +12.4 9.4 52
(Feedlots and Barnyards)
Municipal Waste +9.5 7.2 10
Water Irrigation
Other* +7.3 5.4 64
^"Natural" soil systems: primarily unfertilized cultivated
fields.
48
-------
The analysis indicated, however, that the mean 6l 5N values
of both the septic tank (+10.9) and the animal waste (+12.4)
source categories differ from the mean 615N value for "natural"
soil nitrates (+7.3) at the 0.05 level of significance. (The
"level of significance" is discussed below in the section on
correlation analysis.) This supports Kreitler's conclusion (2)
that nitrates derived from animal wastes (including human wastes)
can be distinguished from natural soil nitrates on the basis of
their nitrogen-isotope ratios.
The distribution of 6l 5N values for the various N0$ source
categories are plotted as histograms in Figure 7. These histo-
grams (and the standard deviations given in Table 8) show that
the isotopic composition for each source category varies over a
wide range of values. These ranges overlap considerably (in
contrast to Kreitler's (2) findings).
The variation in del within the source categories was ex-
amined using correlation analysis. This analysis was carried
out to determine whether environmental factors associated with
the sampling locations were the cause of the variations observed.
Correlation Analysis
Definition of Variables--
The environmental variables associated with the sampling
sites were described in Section 6 (see Table 6). They cover
such environmental factors as latitude, climate class, soil
type, and elevation. Some of the variables are "continuous."
For example, the mean annual precipitation at a site may logi-
cally assume any real number value, even though the actual
values used were integers ranging from 12 to 80. Other variables
are "binary." These variables are assigned a value of either
zero or one. A value of "1" means that a sample is character-
ized by that variable. For example, climate zone Caf (see
49
-------
0.3- .
0.2-
u. 0. 1
SEPTIC TANK
0.0-
nn
-20 -10
10 20 30 40
Del value (ppt)
0.3- -
0. 1-
0.0.
IRRIGATION
-20 -10
Del value (ppt)
10 20 30 40
0.3-
0.2- '
i o.i-
0.0
0.3--
0.2- -
i o.i--
o.o-
OTHER
-20 -10
i n p
10 20 30 40
Del value (ppt)
Del value (ppt)
Figure 7. Histograms of Del Values by Source Category.
-------
Appendix E) is designated by the variable name "C CAF." Thus, a
sample acquired from a site located in climate zone Caf will have
a value of C CAF = 1, while a sample acquired from a site located
in climate zone Aw (see Appendix E) will have a value of C CAF = 0
The definitions of the variables used in the correlation analyses
are given in Table 9. These definitions are provided because
coded variable names (e.g., C CAF) are used in the following
discussion of the results of the correlation analyses.
Meaning of the Correlation Coefficient--
The correlation coefficient is a very useful measure of the
extent to which two variables are related. The correlation
coefficient r is defined as follows:
N
r =
z
i = 1
(X± - X)
(Y. - Y)
N
r
i = 1
av\ 2
A )
1
N
Z (Y. - Y)2
j = 1
where
N = number of data points,
X = average of the X values, and
Y = average of the Y values.
51
-------
TABLE 9. DEFINITIONS OF VARIABLES USED IN THE
CORRELATION ANALYSES
A. Variables with Continuous Values
1 5
DEL 6 N value, in parts per thousand.
PPM N03 concentration, in parts per million (mg/&).
D*SOIL Depth from which a particular soil sample was ob-
tained, in meters.
TEMP Mean annual air temperature at a site, in de-
grees F.
PRECIP Mean annual precipitation at a site, in inches.
ELEV Elevation of a site, in feet above mean sea
level.
LAX Latitude of a site, in decimal degrees.
SEQNCE Sequence number of samples analyzed in this
study.
B. Variables with Zero-One Values'5'
1. Miscellaneous Variables
RADSAM The sample was prepared by Radian under this study.
SOIL The sample was a soil sample.
2. Nitrate Source Variablest
SEPTIC The sample was from a septic tank site.
FEEDL The sample was from an animal waste (feedlot or
barnyard) site.
IRRIG The sample was from a municipal wastewater irri-
gation site.
3. Climate VariablesT
C AW The sample was from a site in climate zone Aw.
C BSK The sample was from a site in climate zone BSk.
C CAF The sample was from a site in climate zone Caf.
C DAF The sample was from a site in climate zone Daf.
C DBF The sample was from a site in climate zone Dbf.
C H The sample was from a site in climate zone H.
(Continued)
52
-------
TABLE 9. DEFINITIONS OF VARIABLES USED IN THE
CORRELATION ANALYSES (Continued)
4. Soil Type Variables§
SOIL 1 The sample was from a site with type 1 soils.
SOIL 2 The sample was from a site with type 2 soils.
SOIL 3 The sample was from a site with type 3 soils.
SOIL 4 The sample was from a site with type 4 soils.
SOIL 5 The sample was from a site with type 5 soils.
SOIL 6 The sample was from a site with type 6 soils.
SOIL 10 The sample was from a site with type 10 soils.
5. Geographic Variables
HLMA The sample was from Houghton Lake, Michigan, site A.
HLMB The sample was from Houghton Lake, Michigan, site B.
PU The sample was from Pueblo, Colorado.
TAM The sample was from Tampa, Florida.
IH The sample was from Indian Hills, Colorado.
RT The sample was from Rockdale, Texas.
TAMU The sample was from Texas A&M University.
ABL The sample was from Abilene, Texas.
SNGLO The sample was from San Angelo, Texas.
MSU The sample was from Michigan State University.
DI The sample was from Dimmitt, Texas.
GR The sample was from Greeley, Colorado.
LOCTAY The sample was from the Lockhart-Taylor area, Texas.
GNDCMN The sample was from Grand Cayman Island in the
Caribbean.
MACON The sample was from Macon County, Missouri.
QUEENS The sample was from Queens County, New York.
RUNLS The sample was from Runnels County, Texas.
*A value of "1" indicates that sample is characterized by a particular
variable; a value of "0" indicates that it is not.
tSamples of "natural" soil nitrate were indicated by a value of "0" for each
of the three nitrate source variables.
tClimate classifications are defined in Appendix E.
§Soil classifications are defined in Appendix F.
53
-------
The correlation coefficient is a number between -1 and +1,
and r2 can be interpreted as the proportion of the variation in
Y that can be "explained" or "predicted" in terms of X. For ex-
ample, the correlation coefficient between DEL (Y) and ELEV (X)
is (from Table 11)
r = .31.
Since
r2 = (-.31)2 = .096,
it follows that 9.6 percent of the total variation in DEL can
be. explained in terms of ELEV.
Figure 8 illustrates the meaning of a non-negative corre-
lation coefficient. A negative correlation between two variables
occurs if one increases as the other decreases.
The multiple correlation coefficient is a measure of the
strength of the composite relationship between a single variable
Y (e.g., DEL) and a set of predictor variables. The square of
the multiple correlation coefficient is the proportion of the
total variation in Y that can be explained in terms of the set
of predictor variables. For example, consider the multiple
correlation analysis of the data for the septic tank source
category. The correlation coefficient between DEL and the set
of predictor variables (D*SOIL, ELEV, and SOIL) is (from Table 17)
r = .56.
Since
r2 = (.56)2 = .314,
54
-------
y-
6-
4-
2-
x and y are
perfectly related
-2-
4-
-6-
-8-1
r = 0
-2
-1
No relationship exists
between x and y
y ~i
6-
= 7
-6-
x and y are
imperfectly related
-8-1
Figure 8. Illustration of the
Meaning of Correlation
Coefficient, r.
55
-------
it follows that 3.14 percent of the variation in DEL can be
explained in terms of the composite influence of D*SOIL, ELEV,
and SOIL.
The partial correlation coefficient is a measure of the
strength of the relationship between two variables given that
one or more other variables are fixed.
Criterion for Choosing Variables Which Are Correlated--
If two variables had no true relationship, their correla-
tion computed from a finite data sample would still almost
certainly be non-zero. The question, then, is how large should
the correlation coefficient be to indicate a true or repeatable
relationship between the two variables?
There is a small chance (one in ten) that an r value as
large as .123 in magnitude would be obtained if the two variables
being considered were unrelated. (The threshold correlation
.123 was computed as a function of the sample size, 178). Thus,
if two variables have a correlation coefficient larger than
.123 or less than -.123, there is a high probability that a
true relationship exists between them. In this case, the cor-
relation is said to be "statistically significant."
The probability (.10 in the example above) of erroneously
concluding that a true correlation exists is called the level
of significance, or "confidence" level. Values of the sample
correlation coefficient (r) required to achieve statistical
significance for different levels of significance and sample
sizes used in this study are given in Table 10.
56
-------
Correlation, multiple correlation, and partial correla-
tion analyses were performed for the entire data base and for
each separate nitrate source category. The analysis of the
entire data base examines the "behavior" of DEL (and PPM) as a
function of all other independent variables, including SEPTIC,
FEEDL, and IRRIG (see Table 9).
TABLE 10. VALUE REQUIRED FOR A CORRELATION
COEFFICIENT (r) TO BE SIGNIFICANT,
AS A FUNCTION OF SIGNIFICANCE LEVEL
AND SAMPLE SIZE.
Sample Size
(N)
10
52
64
178
Value of r at
.10
.539
.231
.207
.123
Significance
.05
.619
.273
.246
.146
Level
.01
.746
.354
.320
.191
Separate correlation analyses were performed within each
of the four nitrate source categories. The values of 615N (DEL)
were shown above to vary over fairly wide ranges (see Table 8
and the histograms in Figure 7). The separate analyses examine
the extent to which this observed variation in DEL for specific
nitrate sources can be explained in terms of the various environ-
mental factors.
Correlations in the Entire Data Base
The most significant results of the correlation analysis
of the entire data base are presented in Tables 11 and 12. Ad-
ditional details of the statistical analyses are presented in
Appendix G.
57
-------
TABLE 11. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL
TEMP
PRECIP
ELEV
FEEDL
C AW
C BSK
C DAF
C H
SOIL 4
SOIL 5
SOIL 6
SOIL 10
HLMB
IH
DI
LOCTAY
GNDCMN
MACON
RUNLS
Correlation
with DEL
.13
.29
-.31
.19
.15
-.13
.18
-.18
.15
.27
-.26
-.18
-.13
-.18
-.19
.22
.15
.15
-.15
Level of
Significance
.10
.01
.01
.05
.05
.10
.05
.05
.05
.01
.01
.05
.10
.05
.05
.01
.05
.05
.05
TABLE 12. CORRELATIONS INVOLVING VARIABLES
MOST CLOSELY RELATED TO DEL
DEL
PRECIP
ELEV
SOIL 5
SOIL 6
LOCTAY
DEL
1.
.29
-.39
.27
-.26
.22
PRECIP
.29
1.
-.61
.21
-.69
.08
ELEV
-.31
-.61
1.
-.35
.35
-.21
SOIL 5
.27
.21
-.35
1.
-.58
.58
SOIL 6
-.26
-.69
.35
-.58
1.
-.33
LOCTAY
.22
.08
-.21
.58
-.33
1.
58
-------
Table 11 lists the significant correlations with DEL and
the levels of significance of those correlations. None of the
other variables listed in Table 9 are significantly correlated
with DEL in the entire data base. Table 12 presents the complete
correlation matrix for DEL and the five variables which corre-
late most highly (at the .01 level) with DEL. Table 12 indicates
the extent to which these independent variables are correlated
with each other. Note especially the relatively high correla-
tion between PRECIP and ELEV (-.61); these are the two variables
with the highest correlations with DEL. These correlations
suggest that, for the sites and samples considered in this study,
615N values tend to be higher in areas of high precipitation
and/or low elevation.
It is impossible to determine, however, whether a true
relationship exists between DEL and PRECIP, between DEL and
ELEV, or between DEL and the combination (PRECIP + ELEV), solely
on the basis of the correlations presented in Table 12. The
fact that PRECIP and ELEV are correlated with each other may
be causing the observed correlation between DEL and PRECIP or
between DEL and ELEV.
For example, it is possible that a true relationship exists
between DEL and PRECIP (i.e., that the mean annual precipitation
influences the biogeochemical environment of soil/ground-water
systems, and thereby influences the nitrogen isotope ratio).
This would show up as a significant correlation. The relation-
ship between DEL and ELEV, however, may be caused by the cor-
relation of PRECIP and ELEV. This correlation may be real,
or may be the result of a sampling bias (e.g., no samples were
obtained from low elevation deserts). In either case, ELEV
could be linked to DEL through PRECIP. The fact that sites
with high precipitation are at relatively low elevations (in
this study) would yield a correlation between DEL and ELEV,
59
-------
even if the elevation, in and of itself, has no influence on
615N. Additional research will be required to clarify the
influence of these environmental variables on 615N in nitrates.
The most significant correlation in Table 11 is the posi-
tive (although small) correlation between FEEDL and DEL. The
DEL value is higher in the feedlot (and barnyard) data than
in the data for septic tank, irrigation, and natural ("other")
nitrate sources as a whole. This difference is primarily due
to the difference between feedlot DEL values and natural ("other")
DEL values (see the earlier discussion of mean 615N values for
the various source categories).
The correlation between FEEDL and DEL cannot be "explained"
by other environmental factors. An analysis of partial correla-
tions (see Appendix G) indicates that the correlation between
DEL and FEEDL changes very little as the influence of other
variables (including ELEV) is removed. Thus, a true relation
between animal waste nitrate sources (feedlots and barnyards)
and higher DEL values appears to exist.
The correlations between nitrate concentration (PPM) and
various other variables were also examined. The variables
which have significant correlations with PPM are listed in
Table 13. Six variables are correlated with PPM at the .01
level: RUNLS (.36), RADSAM (-.36), FEEDL (.29), C CAF (.25),
SEPTIC (-.25), and SOIL 6 (.23).
Thus, for the data examined in this study, higher nitrate
concentrations were found in Runnels County, Texas; in feedlots
and barnyards; in climate zone CAF; and in areas with type 6
soils (Runnels County soils are classified as type 6).
Samples collected for this study (RADSAM = 1) generally
had lower nitrate concentrations than samples analyzed previously
60
-------
TABLE 13. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH PPM
Variable
TEMP
PRECIP
LAT
RADSAM
SOIL
SEPTIC
FEEDL
IRRIG
C BSK
C CAF
SOIL 3
SOIL 6
TAM
RUNLS
Correlation
with PPM
.17
-.14
-.13
-.36
-.16
-.25
.29
-.13
-.15
.25
-.14
.23
.14
.36
Level of
Significance
.05
.10
.10
.01
.05
.01
.01
.10
.05
.01
.10
.01
.10
.01
-------
by Kreitler (2) . This situation had been observed prior to the
statistical analyses: few sites had samples with nitrate con-
centrations as high as those found in Runnels County, Texas,
by Jones (1) and Kreitler (2).
Finally, samples obtained from around septic tank laterals
tended to have relatively low nitrate concentrations (compared
primarily to samples from feedlots and barnyards).
Correlations Within Source Categories
Correlation analyses were also performed to determine
whether any of the observed variation in DEL within a particular
source category (see histograms in Figure 9) can be "explained"
by environmental factors. Only the more significant results of
these analyses are presented here. Additional details of the
source-by-source analyses are presented in Appendix H.
Correlations With Del: Septic Tank Sources--
The significant correlations with DEL in the septic tank
source category are listed in Table 14. The two variables that
correlate most highly (at the .01 level) with DEL are TEMP (.35)
and ELEV (-.39). These variables are also correlated with each
other, and with some of the other variables that are less highly
correlated with DEL (PRECIP and LAT). Thus, it appears that the
615N values of nitrates from septic tanks are higher in warm,
moist, low elevation and low latitude areas than in cool, dry,
high elevation, and/or more northerly areas. The specific
causal relationship between DEL and these four factors cannot
be identified. A true relationship may exist between DEL and
any one of these variables, between DEL and a combination of
these variables, or between DEL and a fifth unidentified variable
which is also correlated with TEMP, PRECIP, ELEV, or LAT.
62
-------
TABLE 14. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL:
SEPTIC TANK SOURCES
Variable
TEMP
PRECIP
ELEV
LAT
RADSAM
C H
SOIL 10
HLMB
IH
Correlation
with DEL
.35
.33
-.39
-.31
-.27
-.32
-.32
-.25
-.32
Level of
Significance
.01
.05
.01
.05
.10
.05
.05
.10
.05
NOTE: C H, SOIL 10, and IH occur simultaneously in this data subset
and, therefore, are indistinguishable for the purposes of
statistics.
It should be noted that the variables C H, SOIL 10, and
IH occur simultaneously in the septic tank source category data.
That is, the correlation between each of these three variables
is r = 1.0. They are therefore indistinguishable in this analy-
sis .
The variable RADSAM is negatively correlated with DEL
(-.27) in this data subset. This means that the <515N values
of samples collected by Radian in this study were generally
lower than the 515N values for septic tank samples provided by
Kreitler (2).
Correlation With DEL: Feedlot Sources--
The significant correlations with DEL in the feedlot (and
barnyard) source category are listed in Table 15. The variables
which correlate most highly with DEL are DI, RADSAM, C BSK, and
ELEV. All of these correlations are negative (i.e., these
variables are inversely related to DEL). These correlations
are strongly influenced by the fact that six of the eleven
63
-------
negative 615N values that were determined in this study are in-
cluded in the feedlot data subset. Five of these six are as-
sociated with the DI (Dimmitt, Texas) site variable. It is
interesting to note that nitrates from feedlots had the highest
average 615N values of any source category (see the discussions
of means and correlations in the entire data base above) in
spite of this relatively high prevalence of negative <515N values
TABLE 15. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL:
FEEDLOT SOURCES
Variable
PRECIP
ELEV
SEQNCE
RADSAM
C BSK
C CAF
SOIL 5
SOIL 6
DI
LOCTAY
Correlation
with DEL
.32
-.37
-.30
-.38
-.38
.25
.32
-.32
-.44
.27
Level of
Significance
.05
.01
.05
.01
.01
.10
.05
.05
.01
.10
The negative correlation of climate class BSk with 615N
can largely be attributed to the DI samples. The Dimmitt site
is located in climate zone BSk.
The negative correlation with RADSAM indicates that the
del values of samples collected and analyzed for this study are
lower than the del values reported by Kreitler for barnyards in
Runnels County, Texas, the Lockhart-Taylor Area, Texas, and
Macon County, Missouri.
Correlations With DEL: Irrigation Sources--
Only one variable (SOIL) was correlated with DEL in the
waste water irrigation data subset (see Table 16). This is a
64
-------
fairly "weak" negative correlation (i.e., the level of signifi-
cance is .10). It indicates the two water samples had higher
615N values than the eight soil samples that were analyzed.
Correlations With DEL: Other Sources--
There were no meaningful correlations of variables with
DEL in the "other" nitrate source category. It is important
to note that all of the points in this data subset are from
Runnels County, Texas. Thus, the only meaningful variables
that could be correlated with DEL are PPM, D*SOIL, and SOIL.
The fact that no significant correlations were found indicates
simply that the influence of environmental factors (e.g., PRECIP,
TEMP, etc.) was not assessed for this source category in this
study
TABLE 16. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL:
IRRIGATION SOURCES
Correlation Level of
Variable with DEL Significance
Soil -.55 .10
NOTE: This data set included only 10 points, so a correlation of at
least .54 was required for significance at the .10 level.
Regression Models for DEL--
Multiple regression models for DEL were constructed for
each of the nitrate source categories. The regression models
were developed by using stepwise regression analysis. This
method adds variables to and deletes variables from a model one
at a time until no further significant improvement in the pre-
dictive accuracy can be made.
The primary purpose of this analysis is not to develop equa-
tions to be used to estimate DEL, but to investigate the strength
65
-------
of the relationship between DEL and the prediction variables.
The set of candidate predictor variables included all except
the site variables (HLMA, HLMB, etc.), RADSAM, and SEQNCE.
The variables that were included in the final regression
model equations are not necessarily the ones that are most
highly correlated with DEL in Tables 14, 15, and 16. If two
predictor variables are highly correlated with each other, it
is unlikely that both variables will be selected for inclusion
in the equation. This is due to the fact that if one of the two
variables is already in the equation, the other variable will
make very little additional contribution to the predictive
accuracy.
When there are many candidate predictor variables which
are correlated with each other (as is the case in this analysis)
the set of variables which is included in the equation is a
complicated function of the nature of the correlations. The
variables included in an equation for DEL and the list of vari-
ables which have the highest correlations with DEL may therefore
be very different.
The results of the multiple regression analyses are sum-
marized in Table 17. Additional details of the analyses, in-
cluding the forms of all four equations for DEL and the analysis
of variance tables, are given in Appendix H.
It is seen from Table 17 that the multiple correlations
for septic, feedlot, and irrigation sources are all between .54
and .56; thus, only about
C.55)2 = .30
or 30 percent of the total variation in DEL can be explained in
terms of the predictor variables used in this study. In the
66
-------
TABLE 17. COMPOSITE RELATIONSHIPS BETWEEN
DEL AND PREDICTOR VARIABLES
Source
Septic
Feedlot
Irrigation
Other
Terms
in Multiple
Modelt Correlation
D*SOIL . 56
ELEV
SOIL
ELEV .54
SOIL
C BSK
SOIL 5
SOIL .55
PPM .23
SOIL
Significant Standard
at Level Error
.01 8.3
.01 8.3
.10* 6.4
.25 5.4
t
These variables were selected for inclusion in the equations by using
the stepwise regression procedure.
-'Borderline. The small sample size (10) accounts for the lower level
of significance in this case.
-------
case of other sources, only about 5 percent can be explained.
It should be remembered that the only meaningful predictor
variables in the "other" data subset are PPM, D*SOIL, and SOIL.
Correlations With PPM By Source--
The correlations between nitrate concentration (PPM) and
the various environmental factors were also examined. Complete
tables of the significant correlations with PPM are included in
Appendix H. A summary of the variables which correlate most
highly with PPM is presented in Table 18.
TABLE 18. VARIABLES WHICH CORRELATE MOST HIGHLY WITH
PPM (AT THE .01 LEVEL OF SIGNIFICANCE) FOR
THE VARIOUS SOURCE CATEGORIES
Source
Septic Tank
Feedlot
Irrigation
Other
Variable
RADSAM
SOIL 6
RUNLS
TEMP
LAT
SEQNCE
RADSAM
C BSK
C CAP
RUNLS
SNGLO
D*SOIL
SOIL
Correlation
with PPM
-.41
.39
.64
.46
-.40
-.38
-.40
-.40
-.52
.63
.77t
-.34
-.33
This data set included only 10 points, so a correlation of at least
.75 was required for significnace at the .01 level (see Table 10),
ANALYTICAL PROBLEMS
When sample analysis activities began, several problems
occurred with the preparation and purification apparatus (the
68
-------
"prep-line" is described in Section 5. The problems, which were
related to equipment failure, included a ruptured Toeppler pump,
a cracked mercury diffusion pump, and numerous leaks. Leaking
S tnnrnr'ks TA7PTP> rort^a t-^rll \T t--rnnl-> 1 o a/-vmo
a. \_o_ a.v_jxcu uicj-L;U..L y U.J.JL±. us JLUII puny-i, ana
stopcocks were repeatedly troublesome.
These various problems led, in some cases, to the loss of
samples. Cumulatively, these problems resulted in considerable
lost time. Because of this, although over 320 discrete soil and
ground-water samples were obtained in our intensive sampling pro-
gram, only about 140 samples were leached and had their nitrate
concentrations determined. Only 87 of these (62 percent) contained
sufficient nitrate to warrant further processing and isotopic
analysis (see below). Of these 87, only 66 were successfully
analyzed for their isotopic compositions. The remainder were
either contaminated with interferring species such as carbon
monoxide (CO) or atmospheric nitrogen (N2) gas, or produced too
little N2 gas for a successful isotopic analysis.
Another problem was that none of the samples obtained via
sampling kits were received in time to be analyzed in this study.
The exclusion of these samples from the final data base reduced
the geographic coverage and the size of the "septic tank" source
data subset.
Another problem was related to the nitrate concentration
of the samples. The configuration of the preparation apparatus,
and the size of the samples, were such that a nitrate concentra-
tion of at least 5 mg/£ was required in the soil leachates and
ground-water samples in order to generate sufficient N2 gas for
an isotopic analysis. Because of this constraint, none of the
samples obtained from the Tallahassee, Florida, or Tampa, Florida
spray irrigation sites were analyzed. This limited both the
size of the data base and the geographic coverage for the
"Irrigation" waste source category.
69
-------
Table 7 presented a summary of the numbers of samples suc-
cessfully analyzed for each site, according to nitrate source
category. Also included in the table are the numbers of samples
for which data was provided by Kreitler.
A final problem is a limitation of the method of study
rather than an analytical problem. It is possible that the total
history of a sample is unknown. Thus, the nitrate in a sample
could conceivably have been derived from a variety of sources.
During the sampling, every effort was made to minimize this
possibility by sampling as close to the waste source as possible
(e.g., within an active pen of a cattle feedlot). The results
of the nitrate concentration and 6l 5N analyses are tabulated in
Appendix D.
Profiles of nitrate concentration and 615N versus depth were
examined for the various sampling holes. An example of such a
profile is presented in Figure 9. As can be seen in the figure,
both nitrate concentration and del value vary over a fairly wide
range with depth. This degree of variation was fairly typical
for the profiles developed in this study. In the previous work
by Kreitler (2), however, the profiles of del value were much
more constant with depth. A representative profile prepared by
Kreitler is reproduced in Figure 10 for comparison.
This difference in the behavior of <515N between different
sites with similar waste sources can be accounted for by two
possible explanations. The first explanation*is that some un-
identified deviation from the standard analytical procedure
produced a variation in 615N values. Due to the time losses
discussed above, duplicate analyses of samples were not per-
formed. Standards of known isotopic composition were run,
however. The results of these analyses did not indicate any
systematic analytical errors (see Section 5).
70
-------
4-1
0)
0)
MH
CX
0)
P
20 25 615N (PPt)
105 120 ppm NOj
__.-.--I
Nitrate Concentration
' 615N
Figure 9. Depth Profile of Nitrate Concentration
and 615N for the Greeley, Colorado,
Animal Feedlot Site, Hole H2.
-------
2-
3-
)M
4-
200 400
NO (mq/kq)
-5 0 5 10
6N'5(%0)
20
Figure 10.
Nitrate and 615N Versus Depth
Beneath a Barnyard With a Definite
Animal Waste Contribution of
Nitrogen, In Runnels County, Texas
[From Kreitler (2)].
The second explanation is that the differences are real.
Differences in the biogeochemical environments of the sites
could be causing the observed variations in isotopic behavior.
Several depth-related factors could be involved. In par-
ticular, variations in the Eh-pH conditions of the soil column
could have a strong influence on the depth distribution of
both nitrate concentration and 615N value. The variations
could also be related to changes or differences in soil bac-
teria populations, clay content or mineralogy, chemical composi
tion or soil texture.
72
-------
It is impossible to explain the differences in isotopic be-
havior illustrated in Figures 9 and 10 with the data generated
in this study. Detailed, site-specific studies, conducted under
a variety of biogeochemical conditions, will be required to fully
explain the N-isotope fractionations which occur in soil columns.
SUMMARY AND CONCLUSIONS
The objectives of this study were to:
1. determine whether 615N values of nitrates
derived from a given waste source are con-
sistent in different geographic regions of
the country and to
2. determine whether 615N values of nitrates
from different waste sources are sufficiently
different to discriminate between the sources.
In order to meet these objectives, samples of soil- and
ground water-nitrate were obtained from different regions of
the country, from sites where the source of the nitrate could
be reasonably predicted. Four nitrate source categories were
examined:
1. nitrates from domestic septic tanks;
2. nitrates from f eedlots and barnyards ;
3, nitrates from municipal wastewater irrigation
sites; and
4 "natural" nitrates from unfertilized soil-
bacteria-vegetation systems (primarily cul-
tivated fields).
73
-------
The nitrate contents and the N-isotope compositions of
these samples were determined. The results were examined using
standard statistical procedures.
The means and standard deviations of the 615N values were
computed for each of the nitrate source categories. These statis-
tics were compared using a one-way analysis of variance (F-test)
and other tests. This analysis indicated that nitrates from sep-
tic tanks, feedlots, and barnyards can be differentiated from
natural soil nitrates on the basis of their isotopic compositions.
The mean 615N values for septic tank nitrates and for feedlot/
barnyard nitrates were +10.9 and +12.4, respectively, while the
mean 6l5N value for natural soil nitrates was +7.3.
The analysis of variance of the means also indicated that
nitrates from septic tanks and from feedlots and barnyards can-
not be differentiated from each other. Finally, nitrates from
municipal wastewater irrigation sites cannot be distinguished
from nitrates from any of the other sources. The mean 615N
value for irrigation nitrates was +9.5.
It should be remembered when considering these results that
the data for natural soil nitrates are all from Runnels County,
Texas, samples. Additional research is required to verify
that the mean 6:5N value for natural soil nitrates (+7.3) is
representative for other regions of the country.
The finding that nitrates from wastewater irrigation sites
cannot be differentiated from nitrates from natural soils, septic
tanks, or feedlots and barnyards may be attributable to the
particular set of conditions that occur at wastewater irrigation
sites. Wastewater irrigation operations are oriented toward
removal of nutrients from the wastewater, via uptake by plants.
The water generally receives some treatment before irrigation,
and is then sprayed or flooded over the land surface. Two of
74
-------
the three sites included in the irrigation data base were actively
cultivated. The one site which was not had the highest 615N
values. These factors tend to suggest that the nitrates at
wastewater irrigation sites represent a mixture of nitrates
derived from the waste and nitrates derived from natural soil-
plant interactions.
Correlation analyses of the entire data base were carried
out. The results of these analyses confirmed the ability to
distinguish nitrates from feedlots and barnyards from natural
soil nitrates (the variable FEEDL was positively correlated with
DEL). This analysis also indicated that some of the variation
in 615N values can result from environmental influences. The
6l5N value was correlated with mean annual precipitation (high
515N values occurred in areas of high precipitation) and negatively
correlated with elevation (high 615N values occurred at low eleva-
tions). Identifying specific causal relationships between 615N
and environmental variables will require additional study.
Correlation analyses were also carried out within each
nitrate source category. These analyses explored the variation
in 615N values determined for each nitrate source. The source-
by-source analysis also included multiple regression analyses.
These analyses indicated that about 30 percent of the variation
in 615N for septic tank, feedlot and barnyard, and irrigation
sources can be "explained" in terms of the environmental factors
considered in this study.
As regards natural soil nitrates, only about 5 percent of
the variation in 615N can be explained. The only meaningful
variables in this source category are nitrate concentration (PPM),
the depth from which soil samples were obtained (D*SOIL), and
whether the sample was a soil sample or a ground-water sample
(SOIL).
75
-------
It is fairly clear that environmental factors influence
615N values. Additional research will be required to precisely
identify the nature of this influence. It is also apparent,
however, that this influence is not so great as to completely
prevent the identification of nitrate sources by nitrogen iso-
tope studies.
It is important to note that the mean 615N values of the
nitrate from septic tanks and from feedlots and barnyards were
statistically different from the mean value for natural soil
nitrates. The ranges of S15N values overlap considerably. Thus,
it will be impossible to identify the source of the nitrate in
any single sample unless the 615N value is greater than about
+24 ppt. This is the mean for natural soil nitrate, plus three
standard deviations: 7.3 + 3(5.4) = 23.5. The more samples
that are analyzed, the more confident one would be of whether
or not the samples represent contamination by a particular
waste source or mixture of sources.
76
-------
REFERENCES
1. Jones, B.C. An Investigation of the Nitrate Problem in
Runnels County, Texas. EPA-R2-73-267, U.S. Environmental
Protection Agency, Washington, D.C., 1973. 214 pp.
2. Kreitler, C.W. Determining the Source of Nitrate in
Ground Water by Nitrogen Isotope Studies. Report of
Investigations No. 83, Bureau of Economic Geology, Austin,
Texas, 1975. 57 pp.
3. U.S. Environmental Protection Agency. National Interim
Primary Drinking Water Regulations. Title 40, Code of
Federal Regulations, Part 141. Promulgated December 24,
1975.
4. Gelperin, A. The Health Effects of Nitrate in Water.
Proc. 12th Sanitary Engineering Conference. University
of Illinois, 1970. pp. 51-52.
5. Asahina, S., et al. Acute Synergistic Toxicity and Hepatic
Neurosis Following Oral Administration of Sodium Nitrite
and Secondary Amines to Mice. Cancer Research. 31, p.
1201, 1971.
6. Alam, B., et al. Synthesis of Nitrosopiperidine From
Nitrate and Piperidine in the Gastro-Intestinal Tract of
the Rat. Nature, 232, p. 199, 1971.
77
-------
7. Junk, H. and G. Svec. The Absolute Abundance of the
Nitrogen Isotopes in the Atmosphere and Compressed Gas
From Various Sources. Geochimica and Cosmochimica Acta,
14. p. 234, 1958.
8. Garrels, R.M., F.T. Mackenzie, and C.H. Hunt. Chemical
Cycles and the Global Environment: Assessing Human
Influences. William Kaufman, Inc., Los Altos, California,
1975. 206 pp.
9. Hoefs, J. Stable Isotope Geochemistry. Springer-Verlag,
New York, 1973. 140 pp.
10. Brown, K.W., J. Frank Slowey, and H.W. Wolf. Accumulation
and Passage of Pollutants In Domestic Septic Tank Disposal
Fields, 2nd Annual Report. College Station, Texas, Texas A&M
University, Civil Engineering Department and Soil and Crop
Sciences Division, 1976.
11. Trewartha, G.T., A.H. Robinson, and E.H. Hammond. Funda-
mentals of Physical Geography. McGraw-Hill, New York, 1961.
409 pp.
78
-------
APPENDIX A
REAGENTS USED IN ISOTOPIC ANALYSES
1. Magnesium Oxide - Finely ground "Baker Analyzed" reagent
grade magnesium oxide was used directly from the bottle.
2. Devarda's Alloy - Reagent grade alloy from Baker composed
of 50 percent copper, 44.5 percent aluminum, and 5.5 percent
zinc was ground in a rotary ball mill using carborundum
balls until it would pass through a 100-mesh screen and
75 percent would pass through a 200-mesh screen. In Radian's
Norton rotary ball mill, two pounds could be processed to this
specification after 48 hours of operation. It is absolutely
necessary that the Devarda's Alloy be ground to a fine pow-
der. If it is not, reduction of N03 may be incomplete and
isotopic fractionation may occur.
3. 0.1 Normal Hydrochloric Acid - This solution was prepared
from a 0.1 normal hydrochloric acid Baker "Dilut-It" con-
centrate.
4. Silver Sulfate - Approximately 65 milligrams was added to
minimize chloride interference during nitrate (NO^) analyses
with the specific ion electrode.
5. Sodium Hypobromite-Iodide Solution - 200 grams of sodium
hydroxide were dissolved in 300 milliliters of deionized
water and cooled to 0°C in ice-sodium chloride bath. Half
of the cooled solution was transferred to a 500 milliliter
Erlenmeyer flask immersed in crushed ice. Sixty milliliters
79
-------
of bromine (Br2) were added over a period of 30 minutes. The
solution was stirred vigorously during the addition, and the
rate of addition regulated so that the temperature of the
solution did not exceed 5°C. When the addition of bromine
was complete, the remainder of the sodium hydroxide solution
was added and stirred. The solution was then stored in a
refrigerator for six days. During the period of cold storage
a precipitate of sodium bromide formed. The solution was
then filtered through a glass-fibre filter and an equal
volume of potassium iodide solution (2 grams KI in one liter
of water) was added. The sodium hypo-bromite-iodide solution
was stored in a tightly stoppered amber bottle at 4°C. One
milliliter of this solution will oxidize five to six milli-
grams of NHij to N2 .
-------
APPENDIX B
DESCRIPTION OF RADIAN'S INTENSIVE
SAMPLING PROGRAM, BY SITES
TEXAS A&M UNIVERSITY EXPERIMENTAL SEPTIC TANK
The experimental septic tank system at Texas A&M was lo-
cated on the TAMU Research Farm, about 12 miles west of the
university campus. A plan of the site is shown in Figure B-l.
Operation of the site ceased in 1977.
The site consisted of three tile fields and a lysimeter
complex. These four basic elements were tied into a central
effluent distribution system. The effluent was derived from a
septic tank serving several houses located about a mile from
the site. It was delivered to the site via pipeline. The
average composition of the effluent is presented in Table B-l.
The lysimeter complex consisted of three lysimeters, each
measuring about 7 feet long, 6 feet across, and 6 feet deep.
Each lysimeter contained a different soil type: Lakeland
sand; Norwood sandy loam; and Miller clay. Effluent was ap-
plied at a constant rate by peristaltic pumps.
There were three tile fields at the site: the dump field;
the Norwood field; and the Miller field. The dump field con-
sisted of four parallel laterals, each of which was 50 feet
long. They were spaced about 10 feet apart. The dump field
was in Norwood sandy loam soil and was used to dispose of ex-
cess effluent. No record was kept of the effluent application
rate to this field.
81
-------
TO SEPTIC TANK*
5000'
DUMP FIELD
DESIGN
RATE
NORWOOD
J FIELD PLOT
LYSIMETERS UNDER
RAIN-SHELTERS
DEMAND.
RATE
100
-DESIGN
RATE
Figure B-l.
Plan of TAMU Experimental
Septic Tank System
[From Brown (10) ]
82
-------
TABLE B-l. TAMU EXPERIMENTAL SEPTIC TANK EFFLUENT QUALITY
FOR 14 MONTH PERIOD
pH
T alkalinity
Conductivity
Hardness
COD
T-PO^-P
0-PO^-P
Organic N
NH3-N
N03-N
N02-N
Cl
SOtt
TOC
IOC
Fecal Coliform
Total Coliform
TC
TN
Mean
7.5
952
2,346
42
292
8.3
6.6
9.2
23.4
1.55
<0.55
166
19
128
125
1.8 x 106
1.8 x 106
220
30.4
Standard
Deviation
± 0.3
± 246 mg/1
± 438 mhos/cm
± 30 mg/1
± 212 mg/1
± 4.5 mg/1
± 3.0 mg/1
± 9.5 mg/1
± 9.4 mg/1
+ 0.44 mg/1
± 64 mg/1
± 12 mg/1
± 81 mg/1
± 39 mg/1
± 1.03 x 106 /100 ml
± 1.03 x 106 /100 ml
± 9.6 mg/1
± 13.6 mg/1
Source: Brown, et al. (10).
-------
The Norwood field consisted of two parallel, 50-foot long
laterals in Norwood sandy loam. One lateral was a design rate
line which received effluent at a standard 40 gallons per day.
The other was a demand rate line. They were located about 50
feet apart at a depth of 45 cm. A stand pipe was connected
to the demand field via a 2 cm diameter tube which was posi-
tioned below the level of the line so that the system remained
sealed from the air. A water stage recorder was positioned on
top of the stand pipe to record the changes in water level within
the field. Three stainless steel rods were used in the pipe as
electrodes. These activated the pump when the water level
dropped to 5 cm from the bottom of the field and turned it off
when the level rose to within 5 cm of the top of the field.
This demand field began operation near the end of April, 1976.
The Miller field was set up the same as the Norwood field.
It was located in Miller clay. The Miller clay is less cap-
able of accepting effluent than is Norwood sandy loam. The
Miller design field was therefore set at an application rate
of 20 gpd.
Sampling of the various fields was accomplished in a vari-
ety of ways. The lysimeters contained tubes terminating in
porous ceramic suction cups through which soil water samples
were obtained. The cups were located in an array at different
depths and distances from the line of inflow. Effluent was
applied through a horizontal perforated pipe enclosed in a
gravel pack. The pipe was located at a depth of 45 cm along
one wall of the lysimeter. Soil water samples were obtained
by applying a vacuum to the suction cup tubes.
Twelve water samples were obtained from lysimeters at TAMU.
Three samples were obtained from a lysimeter containing Lakeland
84
-------
sand, 4 were obtained from a lysimeter containing Norwood sandy
loam, and 5 samples were obtained from a lysimeter containing
Miller clay.
Sampling along a dump field lateral was achieved through
hand augering. Eight soil samples were obtained from the dump
field. The positions of the sampling holes are shown in Figure
B-2. The depth intervals sampled from each hole are shown in
Figure B.-3. No samples were obtained from the Norwood field or
the Miller field, because sampling these fields would have inter-
fered with the TAMU experimental program.
ROCKDALE, TEXAS, SPETIC TANK SITE
A septic tank system located approximately 5 miles east
of Rockdale, Texas, was sampled. The system has an overall
length of 100 feet. The. current effective length of the lateral
is only about 30 feet, however.
The system was installed in April, 1966, and has been in
continuous operation since then. The lateral extends to the
southeast from the south side of a single residence. It is
buried at a depth of about 18 inches and is surrounded by a
gravel pack one foot thick.
At a distance of about 30 feet from the tank, there is
a 3 foot break in the lateral. The tiles were removed from
this section of the lateral in October of 1976, when they be-
came clogged with roots. No effluent passes this break at
present.
Twenty-three soil samples were collected from around the
30 feet section of lateral still in operation. They were col-
lected using a 4-inch, hand operated soil auger. They were placed
in plastic bags and frozen upon return to Austin (within 8 hours).
-------
INFLOW
2.0'
/
1.5'
X
2.0'
2.0'
"°
-10
-20
DISTANCE
FROM
INFLOW
(FEET)
30
40
THESE THREE LINES
NOT IN USE
Figure B-2. Locations of Soil Sampling Holes H1-H4,
TAMU Experimental Septic Tank Dump Field.
86
-------
GROUND
GRAVEL
ENVELOPE
SEPTIC
LINE
GROUND
H2
SURFACE
GRAVEL
ENVELOPE [ Ii5,
SEPTIC
LINE
SAMPLED ,
INTERVALS^
^
-T 0
- I
- 2
- 3
- 4
-1- 5
GROUND
H4
GROUND
H3
GRAVEL
ENVELOPE
SEPTIC
LINE
SURFACE
^-
\
_r\
\J
- 2.0'
/
/
SAMPLED /
INTERVALS\
1
TT1
^
1
GRAVEL
ENVELOPE
SEPTIC
LINE
SURFACE
H
\
r^
\J
-2.0'
SAMPLED .
INTERVALS\
\
H
I
I
-r 0
- I
- Z
- 3
- 4
-1- 5
Figure B-3. Depth Intervals Sampled in Holes H1-H4,
TAMU Experimental Septic Tank Dump Field
87
-------
The positions of the sampling holes and cross sections
showing the depth intervals sampled are shown in Figure B-4
and B-5, respectively.
TAMPA, FLORIDA, SEPTIC TANK SITE
A plan of the Tampa septic tank site showing the locations
of the sampling holes is presented in Figure B-6. The site is
owned by a private homeowner. The septic tank handles the
wastewater load of about seven people. Disposal of the tank
effluent is to the. ground via a recharge pit. This occupies
the area marked "Input" on Figure B-6.
Soil samples were collected at discrete depth intervals
from auger holes at the six locations indicated in Figure B-6.
The maximum depth sampled was 11.5 feet. A total of 64 soil
samples were acquired.
Water samples were collected from auger holes H3, H4, H5,
and H6. When these holes encountered the water table, the sam-
ples were collected using a small peristaltic pump. In addition,
samples of lake water and water from a well located about 275
feet up-gradient (southeast) from the. recharge pit were collected
The ground-water gradient is to the northwest, toward the lake.
INDIAN HILLS, COLORADO, SEPTIC TANK SITE
The Indian Hills site was visited twice, once for water
samples and a second time for a suite of soil samples. Indian
Hills is an old mountain community (development) about 20 miles
southwest of Denver that has been occupied since early in the
century. Although a public water supply system has been in-
stituted in recent years, most residences are on individual
water wells and septic tank systems. The aquifer supplying the
88
-------
5H
0)
01
10 H
IP 20-
OO
t-H
Q
15 H
25 H
3(H
HI
H2
rf
H3
H4
rf-
H5 H6 H7
+
H8
Figure B-4.
Locations of Soil Sampling Holes,
Rockdale, Texas, Septic Tank Site,
89
-------
HI
H2
H3
GROUND SURFACE
GROUND SURFACE
GROUND SURFACE
GROUND SURFACE.
GRAVEL
ENVELOPE
SEPTIC-""
LINE
K-l
r^\
^
o1-
--)
\v
\\^
\\v
GRAVEL
ENVELOPE 1
A
SEPTIC--"
B LINE
C
K-i-
r^
^
-------
LAKE
H6,
H5'
H3
0
_SCALE
10 20
30 FEET
H2
HI
Figure B-6.
Plan of the Tampa, Florida,
Septic Tank Site Showing
Locations of Sampling Holes
91
-------
domestic water supplies is a typical aquifer in crystalline
rock terrain. The porosity is fracture porosity, and the water
is at shallow depths. The community is now apparently recycling
part of the water inadvertently through the septic tanks and
wells, with resulting high levels of nitrate. The Jefferson
County Health Department is aware of the situation and is moni-
toring the water quality in wells and regulating septic tank
design. An intensive study has been conducted in cooperation
with the U.S.G.S.
During the first visit, a total of six water samples were
collected from water wells having the highest nitrate concentra-
tions. Most of these were collected from outside lawn spigots
after letting them run for 5 to 15 minutes.
During the second trip to Indian Hills, a likely candidate
septic field was selected for soil sampling. A plan of the site
is shown in Figure B-7. The log of the hole showing the sampled
intervals is shown in Figure B-8. No previous work has been
done specifically at this site. A more ideal sample point would
have been within the septic field, but the Jefferson County
Health Department advised against this, due to the fact that the
investigation might "put someone in violation of county regula-
tions" by penetrating an engineered septic field structure. Five
soil samples were collected.
PUEBLO, COLORADO, SEPTIC TANK
This site is located in a rural area about 10 miles north of
Pueblo, Colorado. No known previous studies have been made at
the site, which is located on the alluvial bottomlands of
Fountain Creek.
92
-------
GENERAL DIRECTION
OF SLOPE
Figure B-7. Plan of the Indian Hills, Colorado,
Septic Tank Site (not to scale).
93
-------
GROUND
SURFACE
LU
LU
UL
LlJ
Q
SAMPLED
INTERVAL
Figure B-8.
Depth Intervals Sampled at the Indian Hills,
Colorado, Septic Tank Site.
94
-------
A plan of the sample site is shown in Figure B-9. The
sewage disposal system at the site consists of three parts:
a "grease trap", a septic tank, and a cesspool with laterals.
The samples were collected near one of the laterals leading
away from the cesspool. A simplified sample log showing the
depths at which samples were collected is shown in Figure B-10.
The soil was moist and appeared to have been exposed to effluent,
although no odor was detected.
A total of 10 soil samples were collected from the sample
hole. In addition, a water sample was collected from the water
well, which is located about 50 feet from the septic tank. This
well is a shallow well (about 30 feet deep) in the Fountain
Creek alluvium.
HOUGHTON LAKE, MICHIGAN.. SEPTIC TANK SITES
Houghton Lake is located in the north-central part of
Michigan's Lower Penninsula. The area surrounding the lake is
a popular resort community. Nearly all of the 27 mile shore-
line is developed. Permanent residences are interspersed with
summer homes and cottages. Until recently, all of the resi-
dences were serviced by individual water wells and septic
tank systems.
In 1971, Michigan State University and the State of
Michigan Department of Natural Resources began an intensive
study of the deteriorating water quality in Houghton Lake. This
study identified septic tank disposal systems as a primary source
of excess nutrient loading in the lake. Construction of a cen-
tralized sewage collection and treatment facility has been under-
way for several years. At present about 75 percent of the im-
mediate shoreline community is connected to this system.
95
-------
N
GARAGE
\
WATER
WELL
CESSPOOL
SAMPLE
POINT
Figure B-9. Plan of the Pueblo, Colorado,
Septic Tank Site (not to scale)
96
-------
GROUND
LU
LU
X
I
Q_
6j
7
8
9-1
10 H
SURFACE
SAMPLED
INTERVAL
Figure B-10.
Depth Intervals Sampled at the Pueblo,
Colorado, Septic Tank Site.
97
-------
Soil samples were collected adjacent to two active septic
tank/tile field systems in the Houghton Lake area. A combined
total of 19 soil samples were collected from these two sites.
Site A is located about one-fourth mile from the lake in
the community of Houghton Lake Heights. A plan of the site is
shown in Figure B-ll. The system serves a single family resi-
dence and has been in operation for about six years. Soil sam-
ples were collected at several depth intervals from a hole at
the low end of the drain field. The depth intervals sampled
are indicated in Figure B-12. In addition, a single near-surface
sample was obtained right at the edge of the drain field (see
Figure B-ll).
Site B is located about one-half mile from the lake on a
glacial moraine. The site is about 135 feet above the level
of the lake. The water table is fairly deep (>20') and the.
soils are composed of sand and gravel. The septic tank system
at Site B serves a nursing home with about 25 to 30 occupants.
The system consists of three tanks in series. The tank effluent
is discharged to two drain fields. The system is overloaded
and effluent was observed seeping from the ground at one of
the tile fields. A plan of the site is shown in Figure B-13.
One of the tile fields at this site is functioning pro-
perly. A suite of soil samples was obtained from a point
immediately downslope from this field. This sample point is
designated B-l in Figure B-13. The depth intervals sampled
at this point are shown in Figure B-12. Sampling was only car-
ried out to a depth of 3.2 feet because a dense gravel layer
was encountered.
98
-------
ROAD
1
f
1
I
EXTENT OF
GRAVEL FIELD
LATERALS «=
SEPTIC
TANK
/
/
1
1
/
J
;
*
~^.
\ «
HOUSE
SURFACE
SAMPLE
) ©
V
SAMPLE
HOLE
o
a
Figure B-ll. Plan of Houghton Lake, Michigan,
Septic Tank Site A (not to scale)
99
-------
O
O
SITE A
GROUND SURFACE
1 -
4-
SITE B-l
GROUND SURFACE
1-
O)
Ol
2-
a.
LU
Q
- SAMPLED INTERVAL
SITE B-2
GROUND SURFACE
1-
OJ
O)
Q.
LU
O
3-
4 -
5-
Figure B-12. Depth Intervals Sampled at the Houghton Lake,
Michigan Septic Tank Sites.
-------
SAMPLE
POINT B-2
LU
s
o
x
CD
CO
en.
SURFACE
SAMPLE
GENERAL DIRECTION
OF SLOPE
N
TILE
FIELD
SAMPLE POINT
B- 1
Figure B-13.
Plan of Houghton Lake, Michigan,
Septic Tank Site B (not to scale)
101
-------
The other tile field at the nursing home site is over-
loaded. The exact location of the tile field was not delineated
because of the heavily overgrown, spongy nature of the soil.
Effluent was observed seeping from the soil surface and flowing
for a distance of about 10 feet down the side of the moraine.
Samples were acquired from a point immediately downs lope from
the point where the effluent reentered the soil. This sample
point is designated B-2 in Figure B-13. The sampled intervals
for Point B-2 are also indicated in Figure B-12.
It was obvious that during periods of heavy use and/or
rainfall, the area of seepage to the surface is much greater
than that observed on the sampling data. A near-surface (0.2-
0.6 feet) sample was, therefore, taken at a point about five
feet downslope from Point B-2. This sample point is also in-
dicated in Figure B-13.
CITY OF TAMPA, FLORIDA, SPRAY IRRIGATION SITE
A plan of the Tampa Fowler Avenue Spray Irrigation Site,
showing the locations of the monitoring wells that were sampled,
is presented in Figure B-14. Duplicate samples were obtained
from each of the 11 wells. The wells range from 38 to 103 feet
in depth. The wells were pumped for about five minutes prior
to collecting the water samples. Ten ml of IN HaSOit were added
to each sample as a preservative. The samples were stored at
4°C prior to shipment (14 days). They were frozen upon receipt
in Austin.
Soil samples were obtained at three locations in Spray Lot
7, adjacent to wells IB and 1C (see Figure B-14). Spray Lots
6 and 7 have been in operation since September of 1973.
102
-------
LT
UNIVERSITY OF SOUTH FLORIDA
WELL 6C
EAST FOWLER AVE.
WELL 10
LOT 3
RAILROAD R/H
0.1 MILE
LOT 4
WELL 15
LOT 5
"RAILROAD T7₯
LOT 11
WELL 13
LOT 12
WELL 14
LOT 6
WELL 1B,,.WELL1C
LOT 7
T7ULROAD R/U
UTILITY EASEMENT
FOWLfUjAVENU
WASTEWATER
TREATMENT PLANT
LOT 19
WELL 3B
LOT 20
Figure B-14.
Plan of Tampa, Florida, Wastewater
Spray Irrigation Site.
-------
The sample collected at Location A consists of a single 1.25
inch diameter core. The core sample extends from the surface
to a depth of 10 feet. Those collected at Location B consist
of both core and augered samples. The core extends to a maximum
depth of 20 feet. Augered samples were collected from discrete
intervals to a depth of about 7 feet.
Samples collected at Location C consist of two cores and a
suite of auger samples. One core extends to a depth of 20
feet, and the other to a depth of 1.9 feet. The augered samples
were collected from discrete intervals to a depth of 8 feet.
A total of 21 discrete augered samples were collected from the
three locations.
TALLAHASSEE, FLORIDA, SPRAY IRRIGATION SITE
The Tallahassee spray irrigation site is located adjacent
to the Thomas P. Smith Wastewater Renovation Plant. Figure B-15
is a plan of the plant and spray fields, showing the locations
of the three soil sample holes and four of the six wells that
were sampled. These four wells are maintained by the City of
Tallahassee. The average depth is about 40 feet. The two other
wells that were sampled are maintained by the U.S. Geological
Survey. They are located about 0.1 and 0.5 miles south (down-
gradient) of the irrigation fields. Well No. 17 is 152 feet
deep and Well No. 21 is 247 feet deep. Water levels vary over
the region and during the year from about 15 feet to 35 feet
below land surface.
The four city wells are completed in sand overlying the St.
Marks Limestone (Floridan Aquifer). The two U.S.G.S. wells are
cased through the surficial sand deposits and are open to the
limestone of the Floridan Aquifer. The sand and the underlying
Floridan Aquifer are hydraulically connected.
104
-------
o
Ol
SINGLE-GUN
AREA
TALLAHASSEE CITY LIMIT
\ \
\ \
\ \
CAPITAL CIRCLE SR 261
\
O O O On
I
PONDS
IRRIG
FIE
112
HI
CB8D
e-3
AT I ON
LDS
WELL
m AUGER HOLE
A SPRAY NOZZLE
Figure B-15.
500 FEET
0
150 METERS
Plan of Tallahassee, Florida, Spray Irrigation Site
Showing Locations of Sampled Wells and Auger Holes.
-------
Samples from the six wells were obtained by pumping the
wells for about five minutes, or until the water level dropped
below the pump. Two one-quart samples were then collected. Ten
ml of IN H2SOIt were added to the samples collected from Wells
4-2, 6-3, and 17. No preservative was added to the samples col-
lected from Wells CB8D, 4-1, and 21. These samples were placed
in a freezer within one hour of the time they were collected,
and had frozen within five hours.
The soil samples were collected using a 4 inch diameter
soil auger. The samples were taken from measured depth incre-
ments and were sealed in plastic bags. They were put on dry
ice within three hours of the time they were collected. The
locations of the three soil sampling holes are shown in Figure
B-15. The depth intervals sampled in each hole are shown in
Figure B-16. A total of 19 soil samples were collected.
ABILENE, TEXAS, LAND APPLICATION SITE
The Abilene wastewater irrigation farm is located in Jones
and Shackleford Counties, Texas, about eight miles northeast of
the city off State Route 351. A map of the site is presented
in Figure B-17. The treatment plant, immediately adjacent to
the farm, handles an average wastewater flow of slightly more
than ten million gallons per day. After receiving secondary
treatment, the water is stored in several large holding ponds.
The water is delivered to the fields via ditches. Irrigation
is by flooding (overland flow and/or ridge and furrow). When
no irrigation is required, the water is discharged from the
ponds to the Deadman Creek drainage system, a part of the Brazos
River Basin. The irrigation schedule is determined by the farm-
ers who lease the land from the city.
A total of 16 soil samples were acquired from two locations
at the farm. The samples are from specific depth intervals from
the surface to a maximum depth of about seven feet. The sampling
106
-------
TAL - HI
TAL - H3
TAL - H2
GROUND SURFACE
S
GROUND SURFACE
GROUND SURFACE
D.
U-l
Figure E-16.
Depth Intervals Sampled at the Tallahassee, Florida,
Wastewater Spray Irrigation Site.
-------
EXPLANATION
HOLDING PONDS
TREATMENT PLANT
FARM BOUNDARY
Figure B-17. Plan of the Abilene Municipal Wastewater
Land Application Site, Showing Sampling
Locations.
108
-------
locations, designated Hi and H2, are indicated in Figure B-17.
The depth intervals sampled from each site are shown in
Figure B-18.
Six water samples were obtained at the Abilene site. Two
of the samples are from wells located within the irrigated area.
The wells are shallow (about 25 and 50 feet). One is not in
use. The other is used for watering a lawn and small vegetable
garden. Both are completed in a sand and gravel bed.
Two of the water samples are of the irrigation water
(treated effluent). These samples were taken from ditches
delivering water to the fields. The last two water samples
were obtained from soil sampling hole HI. Water was encountered
in HI at a depth of 5.2 feet. The hole was extended to 6.3 feet
and a sample of water was collected. The hole was then bailed
out four times, over a period of about 30 minutes, and another
sample was acquired. The locations from which the water samples
were obtained are shown in Figure B-17.
SAN ANGELO, TEXAS, LAND APPLICATION SITE
The San Angelo wastewater irrigation farm is located about
five miles east of the city off Texas Farm/Market Road 380. A
map of the site is presented in Figure B-19. At present, waste-
water receives primary treatment only. The plant is currently
being expanded to provide secondary treatment, -pj^g method of
irrigation is flooding.
Soil samples were collected at four locations on the farm.
These are indicated in Figure B-19. At two of the sampling
locations (HI and H3) surface samples were collected. Samples
were collected from discrete depth intervals at the other two
locations (H2 and H4) to a maximum depth of 6.4 feet. The depth
intervals sampled at these two sites are shown in Figure B-20.
The total number of samples collected was 15.
109
-------
ABL-H1
ABL-H2
GROUND SURFACE
1 -
2 -
4 -
6-
M.T.
SAMPLED
INTERVAL
Figure B-18.
GROUND SURFACE
3 -
O-
LoJ
7-
\\
Depth Intervals Sampled At the
Abilene Land Application Site.
-------
EXPLANATION
HOLDING PONDS
FARM BOUNDARY
0-"> EFF 1
FARM
BUILDINGS
SECONDARY TREATMENT PLANT
(UNDER CONSTRUCTION)
Figure B-19.
Plan of the San Angelo, Texas, Municipal
Wastewater Land Application Site, Showing
Sampling Locations.
Ill
-------
SNGLO-H2
SNGLO-H4
GROUND SURFACE
1 -
2 -
a.
LU
a
4 -
6 -
SAMPLE
INTERVAL
GROUND SURFACE
1-
a.
LU
Q
3-
4 -
6-
Figure B-20.
Depth Intervals Sampled at the
San Angelo, Texas, Land Applica-
tion Site.
-------
Four water samples were collected in San Angelo. Three
are well-water samples from wells in the irrigated areas. The
fourth is a sample of the irrigation water.
Well No. 1 serves the farm building complex. The well is
78 feet deep. The water level at the time of sampling was
about 35 feet below land surface. It is operated with an elec-
tric submersible pump. The sample was collected at the well
head after pumping for a period of 20 minutes. This well may
penetrate more than one water-bearing zone.
Well No. 2 is located adjacent to a small corral and is
used to supply a watering trough. It is 48 feet deep. The
water level was at 22 feet. The sample was collected in the
same manner as at Well No. 1.
Well No. 3 is a shallow well (exact depth unknown) used to
maintain a stock watering pond. It is operated by a windmill
and pumps nearly continuously at about 4 gpm. The sample was
collected from the pipe where the well discharges to the pond.
The sample of irrigation water was collected as the water
was being applied to a field. The delivery system is made up
of buried 12-inch pipe along the boundaries of the fields.
Water is applied directly from the holding ponds. The water
sampling locations are indicated in Figure B-19.
MICHIGAN STATE UNIVERSITY WASTEWATER LAND APPLICATION SITE
Michigan State University began its program to investigate
the effects of land application of municipal wastewater in 1973.
A plan of the irrigation site is shown in Figure B-21. The site
was once farmed but has been lying idle for about 15 years. The
vegetation cover is mainly old-field vegetation. About one-
fifth of the site is comprised of a Beech-Maple woodlot.
113
-------
FELTON DRAIN
Typical arrangement
of surface aluminum
irrigation laterals
UNDERGROUND
PIPELINE
t
REMOTELY
CONTROLLED
VALVES
IRRIGATION
BOUNDARY
BUFFER ZONE
100m
Figure B-21. Plan of the MSU Experimental Wastewater Application
Site, Showing Sampling Locations (A,B,C).
-------
Samples were collected from three locations at the MSU
spray site (see Figure B-21). Sample Point A is located in an
old-field area receiving 4 inches of effluent per week. Sample
Point B is located in a control area which receives no effluent.
The third soil sampling, Point C, is located in the Beech-Maple
woodlot. This area receives 4 inches of effluent per week. The
depth intervals sampled at each point are indicated in Figure
B-22. Sixteen soil samples were collected in all.
In addition to these soil samples, seven water samples were
collected. Five of these are of soil water obtained from lysi-
meters in the old-field and woodlot areas. The sixth is from
a pond located in the woodlot, and the seventh is a sample of
ground water. This was obtained from the top of the water table
(depth = 5 feet) at sample Point C.
GREELEY, COLORADO, CATTLE FEEDLOT SITE
Soil samples were collected from a feedlot near Greeley,
Colorado. No known previous work has been done for this site,
with the exception of soil mapping by the Soil Conservation
Service. The feedlot is an operation consisting of about 40,000
cattle. Part of the feedlot is on the bottomland of the Cache
la Poudre River (a tributary of the South Platte), and part is
on an adjoining terrace. Runoff from the feedlot is apparently
impounded to prevent it from entering the Cache la Poudre River.
Two sites were selected for sampling at this feedlot. Both
are in cattle pens. The first site (A) was chosen in the older
part of the feedlot, which has been operating since the 1930's.
This site is on a gravel terrace. The log for the sample hole
is shown in Figure B-23. A total of 5 soil samples were re-
tained for transport back to Austin. Gravel that could not be
penetrated with the soil auger was encountered at a depth of
115
-------
POINT A
POINT B
POINT C
3ROUND
SURFACE
1 -
2-
i
UJ
UJ
3-
in
(
Q_
UJ
Q
4-
5 -
6-
i
^^
is
\\x
b
Vv\
S^ SAMPLED
\
-------
only two feet, so the sample point was considered only a partial
success. A second sample site (B) was therefore chosen on the
bottomlands, as a backup. This site has been used as a feedlot
for about 12 years. The log of this sample hole is also given
in Figure B-23. Seven samples were collected from this hole to
a depth of 6 feet.
In addition to these soils samples, one water sample was
collected from a shallow well near the feedlot. No data are
available for this well, but it is apparently a shallow well
in the Cache la Poudre River alluvium. The water has a foul
odor and appears to be contaminated, but part of the contamina-
tion may be from surface runoff.
DIMMITT, TEXAS, CATTLE FEEDLOT SITE
The Dimmitt Feedlot was selected as the representative
feedlot for the North Texas High Plains. Little or no previous
work has been done at the site except for soil mapping by the
Soil Conservation Service. The site is located in the Ogallalla
Formation. The feedlot is built concentrically around one of
the "playas" that are typical of the Llano Estacado. Drainage
from the feedlot is inward into the central playa lake. The
feedlot has been in operation about eleven years. A plan of
the site is shown in Figure B-24.
Two sites were selected for soil sampling at the Dimmitt
Feed Yard. The first site (A) was selected in a cattle pen in
the oldest part of the feedlot. A log of the sample hole is
shown in Figure B-25. The hole was augered to a depth of only
48 inches because a hard, impenetrable caliche layer was en-
countered at that depth.
117
-------
POINT A
POINT B
CO
GROUND SURFACE
SAMPLED
INTERVAL
LU -,
UJ i
a_
LU
Q
2 H
GROUND SURFACE
Figure B-23. Depth Intervals Sampled at
the Greeley, Colorado Feedlot
Site.
1 -
o
i
LU
UJ
Q-
UJ
Q
5 ~'
-------
N
TEXAS
SAMPLE
POINT A
WATER WELL
SAMPLE POINT
FEEDLOT
BOUNDARY
Figure B-24.
Plan of the Dimmitt, Texas,
Feedlot Site, Showing Sampling
Locations.
119
-------
POINT A
GROUND SURFACE
SAMPLED
INTERVAL
1-
Q.
UJ
Q
3-
Figure B-25
POINT B
GROUND SURFACE
Depth Intervals Sampled
at the Dimmitt, Texas
Feedlot Site
1-
2-
3-
4-
5-
o_
UJ
6-
7-
8-
9-
10
>0\\
-------
Experience at the first sample hole and at the site at
Greeley indicated that little precipitation probably infiltrates
within the cattle pens. The tightly packed manure mat in the
pens appears to act as a virtual pavement, and most precipita-
tion probably goes to surface runoff. Consequently, it was
decided to collect a suite of samples at the edge of the playa
lake at the Dimmitt site. The log of this sample hole (B) is
also shown in Figure B-25. The upper 8 inches at this site is
black, spongy, odoriferous washed-in manure. Fourteen samples
were collected from these two holes.
A single water well sample was collected from an irrigation
well near the site (Figure B-24). No data were obtained for
this well, but the pump size indicated the well is several hun-
dred feet deep.
121
-------
APPENDIX C
SAMPLE NUMBERING SYSTEM
All samples used in the statistical analysis were assigned
a sample code number for identification. The sample numbers
consist of several parts separated by hyphens.
The first part of the sample number identifies the location
from which the sample was obtained. Location codes are defined
in Table C-l.
The second part of the sample number identifies the sampl-
ing hole, in the case of soil samples, or the well number, in
the case of ground-water samples. For samples collected by
Radian in the intensive sampling program, these are related to
the holes identified in the various sections of Appendix B.
Soil samples have a third part in their sample identifica-
tion numbers. This part is a number indicating the level within
a particular sampling hole from which the sample was obtained.
This number is roughly related to depth: deeper samples have
higher numbers. An example of the numbering system follows:
a sample is numbered:
HLMA-H2-3
HLMA indicates that the sample is from Houghton Lake, Michigan,
site A. H2 indicates that the sample is from the second sampling
122
-------
TABLE C-l,
LOCATION CODES USED IN SAMPLE
IDENTIFICATION NUMBERS
ID Code
Sampling Location
TAMU
RT
TAM
IH
PU
HLMA
HLMB
TAL
ABL
SNGLO
MSU
GR
DI
LT
GC
MCM
QNY
RU
Texas A&M University; College Station,
Texas
Rockdale, Texas
Tampa, Florida
Indian Hills, Colorado
Pueblo, Colorado
Houghton Lake, Michigan, Site A
Houghton Lake, Michigan, Site B
Tallahassee, Florida
Abilene, Texas
San Angelo, Texas
Michigan State University; East
Lansing, Michigan
Greeley, Colorado
Dimmitt, Texas
Lockhart-Taylor Area, Texas
Grand Caymen Island (Caribbean)
Macon County, Missouri
Queens County, New York
Runnels County, Texas
123
-------
hole completed at that site, and the "3" indicates that the
sample is the third sample collected from the sampling hole
H2.
These sample code numbers may be used to relate the basic
data given in Appendix D to the sites identified in Section 6
and Appendix B.
124
-------
APPENDIX D
BASIC DATA
The analytical data upon which this report is based are
presented in Tables D-l and D-2. Table D-l contains soil
sample data, and Table D-2 contains ground-water sample data
125
-------
TABLE D-l. SOIL SAMPLE DATA
Sample^
Number
HLMA-H1-1
HLMA-H2-2
HLMA-H2-3
HLMA-H2-5
HLMA-H2-6
HLMA-H2-7
HLMB-H1-1
HLMB-H1-3
HLMB-H1-4
HLMB-H2-1
PU-H1-1
PU-H1-4
PU-H1-7
PU-H1-8
PU-H1-9
PU-H1-10
IH-Hl-4
RT-H5-1
RT-H5-2
RT-H8-2
RT-H8-3
TAMU-H1-2
TAMU-H2-1
TAMU-H4-1
TAMU-H2-3
Source*
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
Depth
(meters)
0.2
0.4
0.7
1.2
1.4
1.5
0.1
0.7
0.9
0.1
0.1
1.0
2.0
2.4
2.8
3.3
1.2
0.5
0.8
0.7
1.0
1.1
0.7
0.8
0.6
NO 3
Concentration 615N ,
(ppm) (ppt) Reference
22.1
26.1
46.8
30.8
52.3
25.8
71.5
18.1
9.4
8.1
106.4
34.9
30.0
21.0
48.2
18.9
8.3
5.0
1.8
20.1
8.7
10.9
10.3
5.7
15.0
+10.0
+20.5
+12.3
+ 7.4
+13.9
+10.5
-12.8
+ 7.6
+ 7.8
+ 8.0
-12.7
+12.5
+ 7.3
+ 6.8
+ 6.7
+23.8
-20.0
+17.0
+19.0
+ 7.3
+10.0
+13.0
+ 7.2
+11.0
+ 8.0
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
(continued)
126
-------
TABLE D-l (continued)
Sample^
Number
TAM-H4-1
TAM-H4-3
TAM-H4-4
TAM-H4-5
TAM-H4-6
TAM-H4-7
TAM-H4-8
TAM-H4-9
TAM-H5-2
TAM-H5-4
TAM-H5-11
TAM-H4A- 1
DI-H1-1
DI-H1-3
DI-H2-4
DI-H2-5
DI-H2-6
DI-H2-7
DI-H2-8
DI-H2-9
GR-H2-2
GR-H2-4
GR-H2-6
GR-H2-8
GR-H2-10
GR-Hl-4
GR-H1-7
Source*
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
ST
AW
AW
AW
AW
AW
AW
AW
AW
AW
AW
AW
AW
AW
AW
AW
Depth
(meters)
0.1
0.4
0.8
0.9
1.1
1.2
1.5
1.9
0.4
0.7
3.0
3.0
0.1
0.9
0.7
1.1
1.5
1.9
2.2
2.6
0.2
0.4
0.5
0.8
1.1
0.3
0.6
NO;
Concentration 615N
(ppm) (ppt) Reference
18.0
56.0
23.2
14.2
28.8
10.4
19.3
31.7
34.2
23.8
8.0
102.6
36.2
16.8
40.4
46.8
142.9
48.8
91.9
100.7
30.1
44.0
31.5
47.0
82.0
43.8
30.5
+ 3.9
- 6.3
+ 5.8
+18.8
- 0.4
+21.0
+26.7
+ 3.8
+24.6
+25.9
+30.3
+ 8.6
+ 8.7
- 0.2
- 5.0
+10.5
- 3.9
- 5.7
+23.4
- 5.2
+27.7
+26.4
+13.4
+18.6
+14.2
+ 3.5
-22.0
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
(continued)
127
-------
TABLE D-l (continued)
Sample@
Number
MSU-H1-1
ABL-H1-4
ABL-H1-7
SNGLO-H2-1
SNGLO-H2-2
SNGLO-H2-3
SNGLO-H2-4
SNGLO-H2-6
LT-H1-1
LT-H2-1
LT-H3-1
RU-HA-1
RU-HA-2
RU-HA-3
RU-HA-4
RU-HA-5
RU-HA-6
RU-HA- 7
RU-HB-1
RU-HB-2
RU-HB-3
RU-HB-4
RU-HB-5
RU-HC-1
RU-HC-2
RU-HC-3
RU-HC-4
Source*
MWW
MWW
MWW
MWW
MWW
MWW
MWW
MWW
AW
AW
AW
0
0
0
0
0
0
0
AW
AW
AW
AW
AW
0
0
0
0
Depth
(meters)
0.1
0.5
1.2
0.2
0.5
0.9
1.2
1.9
0.3
0.3
0.3
0.6
0.9
1.2
1.5
2.1
2.4
2.7
0.3
0.9
1.2
1.8
2.3
0.3
0.9
1.5
2.1
NO;
Concentration <515N ,
(ppm) (PPt) Reference
10.4
12.6
6.3
17.1
16.5
22.8
30.8
17.0
1116.0
37.0
25.0
348.0
458.0
381.0
237.0
105.0
104.0
44.0
42.5
25.0
19.0
43.0
32.0
43.0
65.0
75.0
104.0
+ 6.9
+16.7
+ 3.5
+ 6.4
+ 8.5
+ 0.2
+10.4
+ 8.3
+17.3
+37.6
+17.6
+ 4.4
+ 3.8
+ 4.5
+ 3.6
+ 2.7
+ 2.0
+ 2.5
+15.3
+17.9
+11.8
+11.6
+ 7.0
+ 3.3
+ 3.9
+ 5.0
+ 3.8
a
a
a
a
a
a
a
a
b
b
b
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
(continued)
-------
TABLE D-l (continued)
S amp 1 e(§
Number
RU-HD-1
RU-HD-2
RU-HD-3
RU-HD-4
RU-HD-5
RU-HD-6
RU-HE-1
RU-HE-2
RU-HE-3
RU-HG-1
RU-HH-1
RU-HH-2
RU-HJ-2
RU-HK-1
RU-HK-2
RU-HK-3
RU-HK-4
RU-HK-5
RU-HL-1
RU-HL-2
RU-HL-3
RU-HM-1
RU-HN- 1
RU-HN-2
RU-HN- 3
RU-HO-1
RU-HO-2
Source*
AW
AW
AW
AW
AW
AW
0
0
0
AW
0
0
AW
0
0
0
0
0
0
0
0
AW
0
0
0
0
0
Depth
Cmeters)
0.3
0.9
1.2
1.5
1.8
3.0
0.6
1.2
2.1
0.6
0.8
1.4
0.9
0.5
1.1
1.8
2.1
2.4
0.3
0.9
1.5
0.9
0.3
0.9
1.5
0.6
0.9
NO 3
Concentration 6 * 5N ,
(ppm) (ppt) Reference
180.0
1682.0
1300.0
429.0
706.0
700.0
30.0
126.0
36.0
135.0
10.0
25.0
805.0
13.0
36.0
58.0
49.0
45.0
78.0
36.0
32.0
1032.0
37.0
32.0
19.0
225.0
450.0
+13.6
+16.5
+15.2
+13.7
+14.1
+14.0
+24.6
+31.0
+19.2
+21.9
+12.4
+ 7.8
+ 9.7
+ 5.8
+ 6.5
+ 5.6
+ 4.6
+ 4.9
+ 4.4
+ 5.7
+ 6.2
+14.0
+ 3.3
+ 8.6
+ 6.2
+ 6.7
+ 5.7
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
(continued)
129
-------
CABLE D-l (continued)
Sample(§
Number
RU-HO-3
RU-HO-4
RU-HP-1
RU-HP-2
RU-HQ-1
RU-HR-1
RU-HS-1
RU-HT-1
RU-HU- 1
RU-HU-2
Source*
0
0
ST
ST
AW
0
0
AW
ST
ST
Depth
(meters)
2.1
2.7
0.6
0.9
1.5
2.0
0.0
0.9
0.6
0.9
NO 3
Concentration 615N
(ppm) (ppt) Reference
65.0
70.0
33.0
33.0
17.0
10.0
631.0
306.0
711.0
523.0
+ 6.4
+ 6.2
+17.8
+13.0
+15.9
+10.4
+ 9.8
+14.6
+10.3
+12.4
c
c
c
c
c
c
c
c
c
c
-Sample numbers are explained in Appendix C.
k
Sources: ST = Septic Tank; AW = Animal Waste; WWW = Municipal
Wastewater; 0 = Other.
t
References: a = This work; b = Kreitler, unpublished data-
c = Kreitler (2) .
130
-------
TABLE D-2. GROUND-WATER SAMPLE DATA
Sample^
Number
IH-W1
IH-W2
IH-W4
IF-W5
MSU-W4
MSU-W5
LT-W1
LT-W2
LT-W3
LT-W4
LT-W5
LT-W6
LT-W7
LT-W8
GC-W1
GC-W2
GC-W3
GC-W4
MCM-W1
MCM-W2
MCM-W3
MCM-W4
MCM-W5
MCM-W6
MCM-W7
MCM-W8
MCM-W9
Source*
ST
ST
ST
ST
MWW
MWW
AW
AW
AW
AW
AW
AW
AW
AW
ST
ST
ST
ST
AW
Aw
AW
ST
AW
AW
AW
AW
AW
NO;
Concentration <515N ,
(ppm) (ppt) Reference
108.0
66.0
90.0
92.0
14.0
8.0
16.0
341.0
313.0
200.0
147.0
226.0
271.0
12.0
27.3
29.0
12.3
11.9
75.0
62.0
105.0
260.0
330.0
60.0
95.0
145.0
215.0
+ 6.0
+ 7.5
+10.0
+ 3.3
+ 8.0
+26.0
+18.2
+17.2
+17.8
+ 7.3
+18.4
+11.8
+ 9.8
+16.6
+21. 7
+14.2
+19.0
+18.0
+15.9
+10.8
+18.8
+14.5
+13.7
+16.6
+13.5
+16.2
+14.8
a
a
a
a
a
a
b
b
b
b
b
b
b
b
b
b
b
b
c
c
c
c
c
c
c
c
c
(continued)
131
-------
TABLE D-2 (continued)
Sample
Number
MCM-W10
MCM-W11
QNY-W1
QNY-W2
RU-W15
RU-W16
RU-W18
RU-W67
RU-W105
RU-W211
RU-W165
RU-W-201
RU-W233
RU-W234
RU-W309
RU-W366
RU-W369A
RU-W386
RU-W388
RU-W419
RU-W421
RU-W506
RU-W551
RU-W551A
RU-W552
RU-W728
RU-W865A
RU-W865B
Source*
AW
AW
ST
ST
AW
0
AW
0
0
0
0
0
0
0
0
0
0
0
AW
0
0
0
0
0
0
0
0
0
NO 3
Concentration S 1 5N ,
(ppm) (ppt) Reference
180.0
145.0
20.0
30.0
200.0
243.0
288.0
154.0
21.0
211.0
78.0
96.0
416.0
290.0
390.0
32.0
257.0
366.0
978.0
186.0
105.0
174.0
184.0
470.0
94.0
61.0
1360.0
1260.0
+12.2
+15.2
+21.3
+12.1
+ 3.3
+ 2.0
+ 6.4
+ 7.5
+ 7.4
+ 7.5
- 1.1
+ 2.2
+ 7.0
+ 3.3
+ 4.5
+ 2.5
+ 7.0
+ 5.8
+10.3
+ 5.3
+ 8.6
+ 5.2
+ 6.6
+ 8.3
+ 1.2
+ 2.7
+14.1
+13.0
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
(continued)
132
-------
TABLE D-2 (continued)
Sample^-
Number
RU-W867A
RU-W868
RU-W1002
RU-W1003
RU-W1034
RU-W1004
RU-W1005
Source*
AW
0
0
0
AW
0
0
NO;
Concentration & l 5N
(ppm) (ppt) Reference
579.0
21.0
223.0
225.0
1238.0
250.0
75.0
+10.4
+13.1
+ 5.0
+ 6.3
+12.0
+10.0
+ 6.8
c
c
c
c
c
c
c
^-Sample numbers are explained in Appendix C.
*
Sources: ST = Septic Tank; AW= Aniipal Waste; MWW = Municipal
Wastewater; 0 = Other.
References: a = This work: b = Kreitler, unpublished data;
c - Kreitler (2).
133
-------
APPENDIX E
CLIMATE CLASSIFICATIONS
The sites in this study were classified with respect to
climate using the Koppen system, as described by Trewartha and
others (11) . Five great climatic groups are defined under this
system as follows:
A -- Tropical Humid Climates: temperature of coolest
month over 64.4° (18°C) .
B -- Dry Climates: potential evaporation exceeds
precipitation.
C -- Humid Mesothermal Climates: temperature of
coldest month between 64.4° (18°C) and 26.6°C
D -- Humid Microthermal Climates: temperature of
coldest month under 26.6° (-3°C) ; warmest month
over 50° (10°C) .
E -- Polar Climates: temperature of warmest month
under 50° (10°C) .
In addition, a sixth climate group (H) is used to refer to
undifferentiated highland climates. The great complexity of
mountain climates depends on contrasts in altitude and in ex-
posure to sun and winds. Highland areas are therefore grouped
together under a single designation.
134
-------
The five main climate groups are subdivided into a number
of finer classifications by adding a one or two letter qualify-
ing code to the letter designation of a climate group (e.g.,
Caf). The definitions of the various qualifying codes are given
in Table E-l.
TABLE E-l. QUALIFYING CODES OF THE KOPPEN
CLIMATE CLASSIFICATION SYSTEM*
Code Definition
S Steppe
a Temperature of the warmest month over
71.6° (22°C).
b Temperature of the warmest month below
71.6° (22°C).
h Hot: average annual temperature over
64.4° (18°C).
k Cold: average annual temperature below
64.4° (18°C).
f Moist throughout the year: no month
with less than 2.4 inches of rain.
w Dry season in winter, or low sun period:
at least one month with less than 2.4
inches of rain.
*0nly those codes used in this study are given.
135
-------
APPENDIX F
SOIL CLASSIFICATIONS
The soils at the various sites were classified according
to the regional scheme presented by Trewartha and others (11).
Ten basic soil categories are used in this scheme, as shown in
Table F-l.
Podzol and tundra soils (Classes 1 and 9) are the typical
zonal soils of humid, high-latitude regions. Podzols occur in
forested areas, while tundra soils lie in the treeless areas
closer to the pole. These soil classes are largely confined
to the northern hemisphere. The surface layer of a podzol soil
typically is rich in partially decomposed organic material. Be-
low this layer, the A horizon is strongly leached. Material
leached from the A horizon accumulates in the B horizon. The
C horizon is often composed of sandy glacial drift.
Podzolic soils (Class 2) tend to be similar to the true
podzols, but their A horizons are not so heavily leached. Podzolic
soils are generally more fertile than the true podzols.
Podzolic-latoso.lic soils (Class 3) occur in the southeastern
United States. The upper horizons are weathered and clay rich.
They contain a leached zone as a result of podzolic processes.
Latosolic soils (Class 4) occur in areas of high temperature
and precipitation. They are commonly very deeply weathered, and
are rich in clay and oxides of iron and aluminum. These com-
monly occur in a layer of hardpan (laterite).
136
-------
TABLE F-l. MAJOR CLASSES OF SOILS
Soil Class Identifying Number
Podzol Soils 1
0) en
^| Podzolic Soils 2
>W h4
°T) Podzolic-latosolic Soils 3
to -H
3I Latosolic Soils 4
O X
Tundra Soils 9
w
II T)
£. C
u,3 Chernozemic Soils 5
M-l
°!2 Chernozemic-desertic Soils 6
to S
^j: Desertic Soils 7
O ,n
w 3
CO
Alluvial Soils 8
Complex Soils of Areas of
High Local Relief 10
137
-------
Chernozemic soils (Class 5) develop in dryer areas than
those above. They typically contain a deep surface layer which
is rich in organic materials and nutrients. They are not
severely leached.
Chernozemic-desertic soils (Class 6) are transitional be-
tween the chernozemic soils and the true desertic soils. They
may develop a hardpan, commonly of lime (caliche) due to the
low precipitation and high evaporation.
The true Desertic soils (Class 7) are low in organic con-
tent and typically contain decomposing rock fragments. Due to
the very low rainfall, these soils are frequently immature and
horizon differentiation is poor.
Two basically non-zonal categories of soil are the alluvial
soils (Class 8) and the soils of areas of high local relief
(Class 10).
The characteristics of alluvial soils are highly dependent
on the characteristics of the parent alluvium. Little charac-
terization of soils in the areas of high local relief can be
made because of the complex influences of slope and local cli-
mate.
138
-------
APPENDIX G
STATISTICAL ANALYSES OF THE ENTIRE DATA BASE
In this and the following appendix, statistical tables which
supplement those given in Section 7 are presented. These two
appendices are presented for completeness of documentation and
for the benefit of those who are interested in further details
of the statistical analyses. The discussions in these two appen-
dices are brief, however, since the results are discussed in Sec-
tion 7 .
In this appendix, statistical analyses of the entire data
base, including septic tank, feedlot, irrigation, and other
sources, are presented. Separate analyses of the four source
classes are discussed in the following appendix.
The primary objective is to analyze the relationship between
the dependent variable DEL and the various independent variables.
Various approaches to multivariate problems of this type exist,
such as the following: (1) correlation, partial correlation,
and regression, (2) automatic interaction detection, (3) factor
analysis, and (4) path analysis.
While all of these approaches have merits, the first was
selected for several reasons. This approach requires relatively
few a priori assumptions and allows one to explore the data
freely. Correlations and regression, moreover, allow the analy-
sis of continuous variables as continuous variables, whereas
automatic interaction detection involves discrete subdivisions
of a data set. Also, the output from the selected method is
139
-------
relatively easy to interpret, especially compared to results from
a factor analysis.
Various data plots, especially plots of regression residuals,
were produced to determine whether a nonlinear form of any of the
variables should be analyzed. No evidence of nonlinear trends
was seen, however.
Some of the tables presented in this and in the following
appendix are also presented and discussed in the text. The com-
plete set of statistical correlation tables is included here for
the convenience of the reader.
Table G-l indicates the variables which correlate signifi-
cantly with DEL, and Table G-2 presents the complete correlation
matrix for DEL and the variables which correlate most highly
with DEL. Table G-2 indicates the extent to which these inde-
pendent variables are correlated with each other. Note especial-
ly the relatively high correlation between PRECIP and ELEV; these
are the two variables with the highest correlations with DEL.
Table G-3 presents the significant correlations of PPM with
other variables.
Tables G-4 and G-5 present the correlations and partial
correlations of the variables SEPTIC, FEEDL, and IRRIG with DEL.
The partial correlation is the correlation of one variable with
another given that one or more other variables are fixed. The
most important conclusions from these tables are as follows:
(1) The DEL value is higher in the feedlot data than in
data from septic tank, irrigation, and other sources
as a whole. This is indicated by the positive (al-
though small) correlation between FEEDL and DEL.
140
-------
(2) The correlation between FEEDL and DEL cannot be
"explained" by the influence of other factors;
as other variables are fixed, the correlation
between FEEDL and DEL changes very little.
(3) The correlations for the septic and irrigation
variables also change by a fairly small amount
as other variables are fixed.
It should be recalled that the mean values for DEL for the
four sources are analyzed in Section 7. The difference between
the DEL value for feedlots and for the three other sources as
a whole is due primarily to the difference between the DEL values
for feedlots as opposed to those in the fourth, or "other" cate-
gory.
TABLE G-l. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL
TEMP
PRECIP
ELEV
FEEDL
C AW
C BSK
C DAF
C H
SOIL 4
SOIL 5
SOIL 6
SOIL 10
HLM B
IH
DI
LOCTAY
GNDCMN
MACON
RUNLS
Correlation
with DEL
.13
.29
-.31
.19
.15
-.13
.18
-.18
.15
.27
-.26
-.18
-.13
-.18
-.19
.22
.15
.15
-.15
Level of
Significance
.10
.01
.01
.05
.05
.10
.05
.05
.05
.01
.01
.05
.10
.05
.05
.01
.05
.05
.05
141
-------
TABLE G-2. CORRELATIONS INVOLVING VARIABLES
MOST CLOSELY RELATED TO DEL
DEL
PRECIP
ELEV
SOIL 5
SOIL 6
LOCTAY
DEL
1.
.29
-.31
.27
-.26
.22
PRECIP
.29
1.
-.61
.21
-.69
.08
ELEV
-.31
-.61
1.
-.35
.35
-.21
SOIL 5
.27
.21
-.35
1.
-.58
.58
SOIL 6
-.26
-.69
.35
-.58
1.
-.33
LOCTAY
.22
.08
-.21
.58
-.33
1.
TABLE G-3. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH PPM
Variable
TEMP
PRECIP
LAT
RAD SAM
SOIL
SEPTIC
FEEDL
IRRIG
C BSK
C CAP
SOIL 3
SOIL 6
TAM
RUNLS
Correlation
with PPM
.17
-.14
-.13
-.36
-.16
-.25
.29
-.13
-.15
.25
-.14
.23
.14
.36
Level of
Significance
.05
.10
.10
.01
.05
.01
.01
.10
.05
.01
.10
.01
.10
.01
142
-------
TABLE G-4. CORRELATION OF SEPTIC, FEEDL, IRRIG WITH DEL
SEPTIC
FEEDL
IRRIG
Correlation
.07
-.01
Partial
Correlation*
-.03
.15**
-.02
*The partial correlations were computed with the following variables fixed:
PPM, ELEV, SOIL, C AW, SOIL 2, SOIL 3, SOIL 5. These were selected by
stepwise regression analysis.
**Statistically significant at the .05 level.
TABLE G-5. CORRELATION OF SEPTIC, FEEDL, IRRIG WITH DEL
WITH VARIOUS INDEPENDENT VARIABLES FIXED
Variables Fixed*
SEPTIC
FEEDL
IRRG
NONE
.07
.19
-.01
ELEV
.05
.24
-.02
SOIL 5
ELEV
.06
.18
-.00
C AW
SOIL 5
ELEV
.03
.18
.01
SOIL 2
C AW
SOIL 5
ELEV
.03
.19
-.03
SOIL 3
SOIL 2
C AW
SOIL 5
ELEV
-.02
.19
-.03
PPM
SOIL 3
SOIL 2
C AW
SOIL 5
ELEV
-.01
.16
-.01
SOIL
PPM
SOIL
SOIL
C AW
SOIL
ELEV
-.03
.15
-.02
3
2
5
*The variables ELEV, SOIL 5, C AW, SOIL 2, SOIL 3, PPM, and SOIL were selected
in that order by stepwise regression analysis for inclusion in a model to
predict DEL.
143
-------
APPENDIX H
SEPARATE STATISTICAL ANALYSES
FOR THE DIFFERENT SOURCE TYPES
In this and the preceding appendix, statistical tables which
supplement those given in Section 7 are presented. In this ap-
pendix, the results done separately for the four different source
types are given. Some of the correlation tables presented here
are also discussed in the text. The complete set is given here,
however, for the convenience of the reader.
Tables H-l through H-4 present the significant correlations
of the other variables with DEL. Tables H-5 through H-8 present
similar results for the variable PPM.
Tables H-9 through H-13 present the results of multiple re-
gression analyses to develop models to predict DEL. Table H-9
summarizes the results, and Tables H-10 through H-13 present fur-
ther details, including the forms of all four equations and the
analysis of variance tables, not presented in the text.
The primary purpose of this analysis is not to develop equa-
tions to be used to estimate DEL, but to investigate the strength
of the relationship between DEL and the prediction variables.
This point is discussed further below.
The regression models were developed by using stepwise re-
gression analysis. This method adds variables to and deletes
variables from a model one at a time until no further significant
improvement in the predictive accuracy can be made. The set of
144
-------
candidate predictor variables included all except the site var-
iables (HLM A, HLM B, etc.), RADSAM, and SEQNCE. Thus, the mul-
tiple correlations presented can be thought of as indicating the
strength of the relationship between DEL and the set of candidate
physical predictor variables.
It is seen (from Table H-9) that the multiple correlations
for septic, feedlot, and irrigation sources are all between .54
and .56; thus, only about
(.55)2 = .30
or 30% of the total variation in DEL can be explained in terms
of the predictor variables used in this study. In the case of
other sources, only about 5% can be explained.
TABLE H-l. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL:
SEPTIC TANK SOURCES
Variable
TEMP
PRECIP
ELEV
LAT
RADSAM
C H
SOIL 10
HLM B
IH
Correlation
with DEL
.35
.33
-.39
-.31
-.27
-.32
-.32
-.25
-.32
Level of
Significance
.01
.05
.01
.05
.10
.05
.05
.10
.05
NOTE: C H, SOIL 10, AND IH occur simultaneously in this data subset and,
therefore, are indistinguishable for the purposes of statistics.
145
-------
TABLE H-2. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL:
FEEDLOT SOURCES
Variable
PRECIP
ELEV
SEQNCE
RADSAM
C BSK
C CAF
SOIL 5
SOIL 6
DI
LOCTAY
Correlation
with DEL
.32
-.37
-.30
-.38
-.38
.25
.32
-.32
-.44
.27
-
Level of
Significance
.05
.01
.05
.01
.01
.10
.05
.05
.01
.10
TABLE H-3. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL:
IRRIGATION SOURCES
Correlation Level of
Variable with DEL Significance
SOIL -.55 .10
NOTE: This data set included only 10 points, so a correlation of at least
.54 was required for significance at the .10 level.
TABLE H-4. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH DEL:
OTHER SOURCES
None at .10 level of significance
NOTE: TEMP, PRECIP, ELEV, LAT are all constants within this data set. Only
one climate classification (C CAF), one soil type (SOIL 6), and one
geographical location (RUNLS) were represented.
146
-------
TABLE H-5. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH PPM:
SEPTIC TANK SOURCES
Variable
SEQNCE
RADSAM
SOIL 6
MACON
RUNLS
Correlation
with PPM
-.26
-.41
.39
.23
.64
Level of
Significance
.10
.01
.01
.10
.01
TABLE H-6. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH PPM:
FEEDLOT SOURCES
Variable
TEMP
ELEV
LAT
SEQNCE
RADSAM
C BSK
C CAF
DI
GR
RUNLS
Correlation
with PPM
.46
-.25
-.40
-.38
-.40
-.40
-.52
-.26
-.26
.62
Level of
Significance
.01
.10
.01
.01
.01
.01
.01
.10
.10
.01
TABLE H-7. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH PPM:
IRRIGATION SOURCES
Correlation Level of
Variable with PPM Significance
PRECIP --63 .05
SNGLO -11 .01
147
-------
TABLE H-8. VARIABLES WHICH HAVE STATISTICALLY
SIGNIFICANT CORRELATIONS WITH PPM:
OTHER SOURCES
Correlation Level of
Variable with PPM Significance
D*SOIL -.34 .01
SOIL -.33 .01
TABLE H-9. COMPOSITE RELATIONSHIPS BETWEEN
DEL AND PREDICTOR VARIABLES
Source
Septic
Feedlot
Irrigation
Other
Terms
in Multiple
Modelt Correlation
D*SOIL .56
ELEV
SOIL
ELEV .54
SOIL
C BSK
SOIL 5
SOIL .55
PPM .23
SOIL
Significant Standard
at Level Error
.01 8.3
.01 8.3
.10* 6.4
.25 5.4
These variables were selected for inclusion in the equations by using
the stepwise regression procedure.
'^Borderline. The small sample size (10) accounts for the lower level of
significance in this case.
148
-------
TABLE H-10. REGRESSION MODEL FOR DEL: SEPTIC TANK SOURCES
Multiple Correlation: .56 Standard Error: 8.3
Regression Model: DEL = 19.8 + 4.2 (D*SOIL) - .0024 (ELEV) - 11.3 (SOIL)
Source
Regression
Residual
ANALYSIS OF VARIANCE
Degrees of Sum of
Freedom Squares
3 1542
48 3308
Mean
Square F
514 7.46*
69
^Easily significant at the .01 level.
TABLE H-ll. REGRESSION MODEL FOR DEL: FEEDLOT SOURCES
Multiple Correlation: .55 Standard Error: 6.4
Regression Model: DEL = 17.0 - 9.4 (SOIL)
Source
Regression
Residual
ANALYSIS OF VARIANCE
Degrees of Sum of
Freedom Squares
4 1327
47 3152
Mean
Square F
332 4.95*
67
^Significant at the .01 level.
149
-------
TABLE H-12. REGRESSION MODEL FOR DEL: IRRIGATION SOURCES
Multiple Correlation: .55 Standard Error: 6.4
Regression Model: DEL = 17.0 - 9.4 (SOIL)
ANALYSIS OF_ VARIANCE
Degress of Sum of Mean
Source Fredom Squares Square F _
Regression 1 141 141 3.45*
Residual 8 327 41
*Significant at .10 level (borderline).
TABLE H-13. REGRESSION MODEL FOR DEL: OTHER SOURCES
Multiple Correlation: .23 Standard Error: 5.4
Regression Model: DEL = 5.2 + 0.0036 (PPM) +2.4 (SOIL)
ANALYSIS OF VARIANCE
Source
Regression
Residual
Degrees of
Freedom
2
61
Sum of
Squares
95
1750
Mean
Square
47
29
F
1.65*
*Significant at .25 level.
150
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
REPORT NO.
EPA-600/4-79-050
3. RECIPIENT'S ACCESSION-NO.
flTLE AND SUBTITLE
IDENTIFYING SOURCES OF SUBSURFACE NITRATE
POLLUTION WITH STABLE NITROGEN ISOTOPES
5. REPORT DATE
August 1979
issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
T. J. Wolterink, H. J. Williamson, D. C.
Jones, T. W. Grimshaw, and W. F. Holland
8. PERFORMING ORGANIZATION REPORT NO
DCN 78-200-155-15
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Radian Corporation
8500 Shoal Creek Boulevard
Austin, Texas 78758
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-03-2450
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Laboratoi
Office of Research & Development
U.S. Environmental Protection Agency
Ada, Oklahoma 74820
13. TYPE OF REPORT AND PERIOD COVERED
y Final (8/76 - 3/78)
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report describes the methods, results, conclusions, and
recommendations of an investigation of a technique to identify
sources of nitrate in ground water. A discussion of the theoretical
basis of the technique is also provided. Over 300 soil and ground
water samples were collected for this study. The samples are from
numerous sites around the United States, representing a variety of
environmental conditions. The nitrate in 66 of these samples was
separated from other nitrogen species converted to N2 gas, purified,
and analyzed to determine the ratio N/1I+N. These data were com-
bined with the results of analyses performed previously by Jones (1)
and Kreitler (2). Standard statistical techniques were used to
analyze the observed variations in 61 5N values, with respect to sev-
eral nitrate sources and various environmental factors. It was found
that nitrates from feedlots, barnyards and septic tanks can be dis-
tinguished from natural soil nitrates on the basis of their 6l 5N
values. They cannot, however, be distinguished from each other.
Environmental factors contributed to the observed variation in 6 * 5N
values.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSAT1 Field/Group
Ground Water
Waste Disposal
Nitrates
Septic Tanks
Feedlots
Land Application
48 F
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
Unclassified.
21. NO. OF PAGES
163
20. SECURITY CLASS (This page)
Unclassified.
22. PRICE
EPA Form 2220-1 (9-73)
151
V US GOVERNMENT PRINTING OFFICE 1979 -657-060/54^0
------- |