-------
In this example, temporal variation of surface erosion rate for continuous
cornland in central Indiana was calculated. Again, the erosion index dis-
tribution Curve No. 16 on Figure A-2d was used. Assumptions were conven-
tional tillage, a yield average of 40 to 59 bu of corn per acre, and corn-
stalks left on the field after harvesting. The dates, C values, and per-
cent of erosion index for five-crop stages, and RC products are:
Percent R
Crop stage, Cover
starting- factor,
ending date CS.'
Turn plowing, 5/1-5/19 0.55
Seeding, 5/20-6/19 0.70
Establishment, 6/20-7/19 0.58
Growing crop, 7/20-10/9 0.32
Harvest and 0.50
stubble, 10/10-4/30
Total
Reading^-/
13.8
19.5
36.0
57.3
91.0
Percent
in the
period
5.7
16.5
21.3
33.7
22.8
100
RC
product
3.14
11.55
12.35
10.78
11.40
49.22
a_/ Reference source: USDA-Agricultural Research Service Handbook No.
282,27 Table 2.
b/ Reading from Figure A-2d (Curve 16) for starting date.
The annual C factor is estimated at 0.49. Temporal variation of surface
erosion rate, in terms of percent of annual total, is shown in Figure 3-12.
It is seen that the maximum erosion from this continuous cornland would
occur in mid-June through mid-July, nearly identical to the period of
maximum erosion with constant soil cover (Figure 3-11). The 30-day max-
imum is approximately 3.2 times average daily, which is higher than the
previous (constant C factor) case due to the magnifying effect caused
by the overlapping of a high R period with a high C period. Figure
3-12 also shows that minimum erosion would occur during the winter sea-
son; the 30 day minimum is one-fourth of the average daily load.
3.2.7 Source Areal Data
Information and data of considerable variety are needed to assess sediment
loading by surface erosion from various sources. Pertinent source charac-
teristic data including soil erodibility, rainfall erosivity, slope length,
slope gradient, vegetative cover, conservation practices, and delivery
ratio, have been presented in the previous sections. This section presents
sources of data relevant to acreages of land use and land disturbance.
74
-------
1.0 r
— 30-Day Maximum
Cropstage
F - Turn Plowing
Cl - Seeding
C2 - Establishment
C3 - Growing Crop
C4 - Harvest and Stubble
1/1 2/1 3/1 4/1 5/1 6/1 7/1 8/1 9/1 10/1 11/1 12/1 1/1
Date Month/Day
Figure 3-12. Projected variation of soil erosion on continuous corn
lands in central Indiana^'
Source: Midwest Research Institute.
75
-------
The following are sources of areal data which are pertinent to assessing
sediment loadings from various nonpoint sources.
Land use -
"Conservation Needs Inventory" - Soil Conservation Service
"Census of Agriculture" - Bureau of Census
State cropland and livestock reports - State Agriculture Department
Forest survey reports - Forest Service
Range survey reports - Soil Conservation Service, Forest Service
Forest cutting and fire reports - Forest Service, State Foresters
"Watershed Conservation and Development Field Data" - Bureau of Land
Management
Housing construction -
Statistical Yearbook - U.S. Department of Housing and Urban
Development
County and City Data Book - U.S. Bureau of Census
"U.S. Census of Population and Housing" - U.S. Bureau of Census
"Housing Authorized by Building Permits and Public Contracts" - U.S.
Bureau of Census
"Construction Report" - U.S. Bureau of Census
Mining activities -
Mineral Yearbook - U.S. Bureau of Mines
Mining permits - State
Highways and roads -
U.S. Federal Highway Administration
State Highway Department
The following data sources are particularly pertinent to assessment of
surface erosion for large areas.
Data for agricultural lands — the Conservation Needs Inventory (CNI) - The
CNI is one of the major sources of data for agricultural land in the
United States. The first inventory was made in 1958 to 1960 and updated
in 1967. The objective of the inventory was to develop current, detailed
data on land use and conservation treatment needs on rural land and to ob-
tain data on watershed project needs on both privately and publicly owned
land in the U.S. The inventory includes all acreage except urban and
built-up areas and land owned by the federal government, other than crop-
land operated under lease or permit.
76
-------
Inventoried lands are compiled by county in terms of land use, land capa-
bility class and subclass,'1- and conservation treatment needs as shown in
Table 3-11. The seven major rural land use categories are subdivided into
18 secondary land use classifications and current (1967) conservation
treatment needs. Each group is inventoried according to land capability
classes and subclasses.
It is important to note that not all land was classified or inventoried
in the CNI. For the noninventoried land (including federal noncropland,
urban buildup, and small water bodies), there has been thus far no infor-
mation concerning use of land by capability. For most regions the propor-
tion of total land in the noninventory group is not significant. However,
in the western states the proportion of this group may be very high.
Copies of state inventories may be obtained from the State Conservation
Needs Inventory Committee, and/or University Agricultural Extension Service,
Magnetic tapes of the inventory are available from the Statistical Labora-
tory, Iowa State University, Ames, Iowa.
The U.S. Soil Conservation Service in 1972 solicited soil scientists in
the United States for the soil data relevant to surface erosion, in format
compatible with the format of CNI. Data are reported by Land Resources
(LRA) and by land capability class and subclass. For all LRAs east
* Land Capability Classification is one of a number of interpretive group-
ing of soil survey maps made primarily for agricultural purposes.
In this classification, the arable soils are grouped according
to their potentialities and limitations for sustained production of
the common cultivated crops that do not require specialized site con-
ditioning or site treatment. Nonarable soils (soils not suitable for
long-time sustained use for cultivated crops) are grouped according
to their potentialities and limitations for the production and per-
manent vegetation and according to their risks of soil damage if
mismanaged.
The capability classification provides three major categories:
(a) capability unit; (b) capability subclass; and (c) capability
class. The reader is advised to consult with State Conservation
Needs Inventory for detailed descriptions of classifications.
** Land Resource Areas (LRA), as delineated by the Soil Conservation
Service, U.S. Department of Agriculture, are broad, geographic areas
having similar patterns of soil (including slope and erosion), climate,
water resources, land use, and type of farming. Delineation and de-
scription of LRAs are available in USDA-SCS, Agriculture Handbook
No. 296, "Land Resource Areas of the United States," December 1965,
and USDA-ERA series on "The Look of Our Land--An Airphoto Atlas of
the Rural United States."
77
-------
Table 3-11.
LAND USE AND TREATMENT NEEDS CATEGORIES OF THE
CONSERVATION NEEDS INVENTORY
Primary use
classification
Secondary use
classification
Treatment classification
Cropland in tillage
rotation
Corn and sorghum
Other row crops
Close-grown crops
Summer fallow
Rotation hay and pasture
Hayland
Conservation use only
Idle
Other cropland
Orchards, vineyards and
bush fruit
Open land formerly
cropped
Pastureland
Rangeland
Treatment adequate
Treatment needed--nonirrigated
Residue and annual cover
Sod in rotation
Contouring
Strip-cropping or terracing
diversion
Permanent cover
Drainage
Treatment needed — irrigated
Cultural and management
practices
Improved system
Water management
Treatment adequate
Treatment not adequate
Treatment adequate
Treatment unfeasible
Needs change in land use
Protection only
Improvement only
Improvement and brush control
Reestablishment of vegetative
cover
Reestablishment and brush
control
Treatment adequate
Treatment unfeasible
Needs change in land use
Protection only
Improvement only
Improvement and brush control
Reestablishment of vegetative
cover
Reestablishment and brush
control
78
-------
Table 3-11.(Concluded)
Primary use
classification
Forestland
Forestland grazed
Secondary use
classification
Commercial
Other land
Noncommercial
Commercial
Noncommercial
On farms
Not on farms
Treatment classification
Treatment adequate
Noncommercial — stand establish-
ment and reinforcement
Commercial--stand establishment
and reinforcement
Commercial--timberstand improve-
ment
Treatment adequate
Forage improvement
Reduction or elimination of
grazing
Treatment adequate
Treatment not adequate
79
-------
of the continental divide, information solicited includes name of dominant
soil, dominant slope length, dominant slope percent, and K factor. These
data were reported in Data Form 1.
For LRAs west of the continental divide, where K factors had not been de-
veloped before the survey, information solicited includes dominant soil
name, dominant slope length, slope percent, and estimated soil losses
(tons/acre/year) from selected cropping systems. Data were solicited in
Form 1W.
For convenience of use, the MRI study group has combined factors in Form 1
and calculated K-LS indexes for various land capability classes and sub-
classes for LRAs in the areas east of the continental divide. Values of
the K'LS index, and questionnaire returns in Form 1W (for LRAs west of
continental divide) are presented in the Appendices D and E, respectively,
of this handbook. These data can be used together with land-use data in
the State Conservation Needs Inventory for assessing gross erosion from
agricultural lands in large areas.
Data for commercial forests - The most recent data on state and national
levels are presented in "The Outlook for Timber in the United States,"
U.S. Department of Agriculture, Forest Service, Forest Resource Report No.
20, October 1973. This is a report on the nation's timber supply and
demand situation and outlook, related primarily to the commercial timber-
lands in the U.S. that are suitable for production of timber crops. This
report provides statistical data, as of 1970, on the current area and con-
dition of the nation's forestland, inventories of standing timber, and
timber growth and removals by individual states. Information is also in-
cluded on recent trends in forestland and timber resources, trends in util-
ization of the nation's forest for timber and other purposes, and trends
in consumption of wood products. This report represents the latest in a
series of similar timber appraisals prepared by the Forest Service in the
past.
If more local detail data are needed, they likely can be provided by the
forest and range experiment stations. An important timber resources in-
ventory on a local level available from the forest and range experiment
stations is "Forest Statistics" (or "The Timber Resources"). The recent
publications present inventories of timber resources on the state and
county levels. The forest resource data and the accompanying discussions
of forest area, volume, growth, and cut are useful for planners.
Despite the availability of considerable information on the United States
timber inventory, there are important gaps in information necessary to
assess pollutant loadings from forested areas. There is far more informa-
tion available today concerning standing timber volume on forestland than
80
-------
there is concerning soil and topographic characteristics, the acreage of
forest harvested, method of harvest, mileage of roads built and maintained,
percent canopy and ground cover situation, and current soil and water con-
servation practices. One possible method of obtaining such information is
through personal contact with local knowledgeable persons. The following
are individuals who may be able to supply such needed data:
U.S. Forest Service
Resource management staff officers
District rangers
Forest supervisors
Regional foresters
U.S. Bureau of Land Management
State director
District manager
State and local agencies
State foresters
County foresters
Private forest industries
Data for mining and construction activities^ - The extent of construction
and mining activities in a given locale can be estimated directly from
sources such as building permits, construction reports, and mining permits.
Similar data also can be obtained from some other sources, such as census
data for housing units, highways, roads, utility transmission lines, etc.,
in which data are assembled periodically. Data gathered in different years
can be translated into average annual acreages of land being disturbed by
construction activities.
For example, the census in County and City Data Book, U.S. Department of
Commerce, Bureau of the Census, includes the total number of housing units
between 1967 and 1972. Also given are the number of units in single family
units and the number in multiple units. From these figures the average
annual number of new single and multiple dwelling units can be determined.
With actual data or an approximation of acreage per housing unit, one may
estimate the average annual acres of land used for new housing.
81
-------
Construction activities for a given site are generally of limited duration,
and so is sediment production. MRI economists estimated the average dura-
tion of construction to be:
6 months for residential buildings,
11 months for nonresidential buildings, and
18 months for nonbuilding construction.
3.3 SEDIMENT LOADINGS FROM OTHER SOURCES: GULLIES, STREAMBANKS, AND MASS
SOIL MOVEMENT
3.3.1 Overview
3.3.1.1 Gully erosion -
Gully erosion is caused by temporary concentration of runoff during and
immediately after rainfall. Sediment production from gullies is accom-
plished by scouring on the bottom or sides by running water, by slides of
materials into gullies from the side, and by erosion over the well-defined
headscarp.
Gully erosion is common to most regions in the United States. Expansion
of gully development is most vividly apparent in arid and semi-arid areas
such as southwestern U.S. where climatic changes are easily expressed in
network changes, and also in those areas where the influence of man has
been substantial or rapid, or both.
Gullies usually are found on slopes greater than 5 degrees. Gullies are
especially active during the rainy season, and are particularly well-
developed on the margins of uplands composed of highly friable sandstones.
Development of gullies is associated with improper land use and severe
climatic events. The effect of land use on gully development is connected
with modification of land cover and soil conditions, and subsequent changes
in runoff patterns. Gullies have developed following the removal of trees
on the lower part of the sides of glacial troughs, and following compaction
of ground, change in topsoil, and changes in infiltration characteristics.
The impact of land use on gully development is most striking when original
plant cover on steep slopes is removed and runoff occurs with little im-
pediment.
Climatic fluctuations also may cause gully development. Climatic fluctua-
tion may cause disappearance of vegetation cover, and lead to vivid gully-
ing activities.
82
-------
Sediment production from gully development has been described for some
regions in the U.S.?ff»29/ The quantity, though often large, is usually
less than that produced by surface erosion. However, economic losses from
dissection of uplands, damage to roads and drainage structures, and deposi-
tion of relatively infertile overwash on flood plains are disproportionately
large.
The prediction of gully growth has thus far received little attention, al-
though some studies have developed empirical prediction procedures for
specific localities.
3.3.1.2 Streambank erosion -
In the streambank erosion process, energy from streamflow, ice, and floating
debris, and the force of gravity are applied to the streambank and stream-
bed. If the energy is greater than the resistance of soil particles form-
ing the channel, erosion results. Brown?.' suggests that in most forest and
range country and in areas with less than 51 cm (20 in.) of precipitation
annually, channel-type erosion (including gully, streambank, etc.) generally
produces the greater part of the sediment. Where a watershed is primarily
agricultural and has more than 5.1 cm (20 in.) of precipitation, a major
part of the sediment production is generally from sheet erosion. Gottschalk—'
suggests that streambank erosion is dominant in the semiarid and arid areas
of the United States and in the mountainous areas of the Central and South
Pacific Coast regions. Andersoni!/ estimated sediment yields from the North
Coast watersheds of California, and the Williamette Basin of western Oregon,
and concluded that sediment contribution from streambank erosion in that
part of the country is greater than from other sources combined.
In 1969, the Corps of Engineers, in conjunction with Soil Conservation Ser-
vice personnel, completed the "National Assessment of Stream Bank Erosion."—'
All districts in the nation provided information on the amounts of stream-
bank erosion in their areas. Stream density by land resource area was used
to determine total stream miles and bank miles. Estimates were then made
on how many of these banks erosion was negligible, moderate, and serious.
Damages were determined at the site where erosion occurred and where the
ensuing sediment was deposited. Cost of treatment was calculated for both
moderate and serious cases.
A report on the nationwide assessment was issued by the Corps in October
1969. Regional inventory reports are available from appropriate district
offices.
83
-------
3.3.1.3 Mass soil movement -
Mass soil movement is the downs lope movement of a portion of the land sur-
face under the effect of gravity. Such movements may take the form of
landslide, mudflow, or downward creep of an entire hillside, and contribute
to sediment loadings to surface waters. In many areas this source of supply
is unimportant. However, mass soil movement may constitute the dominant
process of erosion in areas with exceptionally steep slopes, high rainfall,
or low-strength soil, such as that of mountainous areas of western North
America, as well as of southern California. In such areas, soil may ^-main
in place as the result of a delicate balance between forces tending tc
cause downslope movement and various forces tending to resist it. Any dis-
turbance may upset this delicate balance and result in initiation or accel-
eration of mass soil movement.
Landslide is influenced by the slope of the land, composition of soil, and
oo /
water content of the soil. Dyrness—' indicated that stony soils from
basalt and andesite were 14 to 37 times more stable than those from tuffs
and breccias, which are volcanic parent materials, and normally weather
rapidly to silts and clays. Silts and clays can retain large quantities
of water. The water adds to the soil burden and reduces its strength,
thus promoting landslides. In Oregon, landslides normally occur near peak
stream flow from winter storm runoff when the water content of soil is at
the maximum.
Man's activities may play an important role in initiation and acceleration
of mass soil movements. In a review of mass erosion research in the
western United States, Swanston—' made the following statements about the
effect of disrupting activities of man on mass soil movements:
"Road building stands out at the present time as the most damaging
activity. Soil failures relating to this activity are the result
primarily of slope loading from road fill and sidecasting, inade-
quate provision for slope drainage, and of bank cutting.
Fire, natural and man-caused, is a second major contributor to
accelerated soil-mass movement in some areas. This relates largely
to the destruction of the natural mechanical support of soils, often
abetted by surface denudation of the soil mantle, opening it to the
effects of surface erosion.
Logging affects slope stability mainly through destruction of pro-
tective surface vegetation, obstruction of main drainage channels
by logging debris, and the progressive loss of mechanical support
on the slopes as anchoring root systems decay."
84
-------
Very little work has been done to establish quantitative cause and effect
relationships between mass soil movements and causative factors, including
natural characteristics and man's activities in watersheds.
3.3.2 Methods for Quantifying Sediment Loading from Gullies, Streambanks,
and Mass Soil Movement
The cause/effect interrelationships of gully erosion, streambank erosion,
and mass soil movement have yet to be put into proper perspective. Methods
are therefore not available for any given locality and any set of existing
or assumed conditions for accurately predicting contributions of sediment
loading from these sources. The discussion and general facts presented in
the preceding paragraphs will serve as guidelines for estimation of channel
erosion and mass soil movement. These guidelines generally apply to two
options, presented below, for estimating gully and streambank erosion and
mass soil movement at the local/regional level. These options may be used
separately or in combination.
3.3.2.1 Estimation from historical local data and research results -
The local history of gully erosion, streambank erosion, and mass soil move-
ment can be obtained by local interview and from existing research results.
Research results are available in engineering surveys and basin and project
reports. Public agencies which have these results include: Department of
Army--Corps of Engineers; Department of the Interior—Bureau of Land
Management, Bureau of Mines, Fish and Wildlife Service, and National Park
Service; Department of Agriculture--Forest Service and Soil Conservation
Service; state departments of water resources; public works authorities;
and planning commissions.
3.3.2.2 Estimation from historical topographic data -
Quantification of sediment production from gullies, streambanks, and mass
soil movement also can be made through use of aerial photographs. A large
area of the United States was photographed from the air about 35 years ago.
Many areas have been rephotographed periodically. These aerial photographs
provide valuable tools to determine the boundaries and lateral movement of
channels during various periods of time and are used extensively in water-
shed investigations whenever available. The following agencies and organi-
zations have aerial photographs of parts of the United States: Department
of the Interior—Geological Survey, Topographic Division; Department of
Agriculture—Agriculture Stabilization and Conservation Service, Soil Con-
servation Service, and Forest Service; Department of Commerce--Coast and
Geodetic Survey; Department of the Air Force; National Aeronautics and
Space Administration; various state agencies; and commercial aerial survey
and mapping firms.
85
-------
REFERENCES
1. Meyer, L. D., and W. H. Wischmeier, "Mathematical Simulation of the
Process of Soil Erosion by Water," paper presented at the 1968 Winter
Meeting of the American Society of Agricultural Engineers, Chicago,
Illinois, 10-13 December 1968.
2. Brown, C. B., "Effect of Land Use and Treatment on Pollution,"
Proceedings of the National Conference on Water Pollution, PHS,
Department pf Health, Education and Welfare, Washington, D.C. (1960).
3. Megahan, W. F. , "Logging, Erosion, Sedimentation—Are They Dirty Words,"
J. Forestry. 7£(7) (1972).
4. Ralston, C. W., and G. E. Hatchell, "Effects of Prescribed Burning on
Physical Properties of Soil," in Proceedings, Prescribed Burning
Symposium, pp. 68-85, 14-15 April 1971, USDA Forest Service.
5. Collier, C. R., et al., "Influence of Strip Mining on the Hydrologic
Environment of Beaver Creek Basin, Kentucky, 1955-1959," USGS Pro-
fessional Paper 427-B (1964).
6. USDA Soil Conservation Service, "Controlling Erosion on Construction
Sites," Agriculture Information Bulletin 347 (1970).
7. USDA Soil Conservation Service, National Engineering Handbook, Sec-
tion 3, "Sedimentation," Washington, D.C., April 1971.
8. USDA Agriculture Research Service, "Present and Prospective Technology
for Predicting Sediment Yield and Sources," Proceedings of the
Sediment=Yield Workshop, USDA Sedimentation Laboratory, Oxford,
Mississippi, 28-30 November 1972.
9. Wischmeier, W. H., and D. D. Smith, "Predicting Rainfall—Erosion
Losses from Cropland East of the Rocky Mountains," Agriculture
Handbook 282, U.S. Department of Agriculture, Agriculture Research
Service, May 1965.
10. Wischmeier, W. H., "Upland Erosion Control," in Environment Impact
on Rivers, p. 15-1 to 15-26, H. W. Shen (ed.), Fort Collins,
Colorado (1972).
11. U.S. Department of Agriculture, Soils Technical Note No. 3, Soil
Conservation Service, Honolulu, Hawaii, May 1974.
86
-------
12. Wischmeler, W. H. , and D. D. Smith, "Rainfall Energy and Its
Relationship to Soil Loss," Transaction, 39:285-291, American
Geophysical Union (1958) .
13. U.S. Department of Agriculture Conservation Agronomy Technical
Note No. 32, Soil Conservation Service, West Technical Service
Center, Portland, Oregon, September 1974.
14- Garstka, W. U., "Snow and Snow Survey," in Handbook of Applied Hydrology.
V. T. Chow (ed.), McGraw-Hill, Inc., New York, New York (1964).
15. Porter, G. R., and W. H. Wischmeier, "Evaluating Irregular Slopes for
Soil Loss Prediction," presented before the American Society of
Agricultural Engineers, Paper No. 73-227, St. Joseph, Michigan (1973).
16. Wischmeier, W. H., "Estimating the Cover and Management Factor for
Undisturbed Areas," presented at USDA Sediment Yield Workshop,
Oxford, Mississippi (1972).
17- Water Resources Administration, "Technical Guide to Erosion and
Sediment Control Design (Draft)," Maryland Department of Natural
Resources, Annapolis, Maryland, September 1973.
18. U.S. Environmental Protection Agency, "Effect of Hydrologic Modifi-
cations on Water Quality," report draft by the MITRE Corporation,
October 1974.
19. U.S. Department of Agriculture, Engineering Technical Note No. 16,
Soil Conservation Service, Des Moines, Iowa, 21 March 1973.
20. Smith, K. G., "Standards for Grading Texture of Erosional Topography,"
Amer. J. Sci.. 248^:655-668 (1950).
21* Schumm, S. A., "The Evolution of Drainage Systems and Slopes in Bad-
lands at Perth Amboy, New Jersey," Geo. Soc. Amer. Bull.. 67_: 597-646
(1956)
22. Strahler, A. N., "Quantitative Geomorphology of Drainage Basin and
Channel Network," in Handbook of Applied Hydrology, pp. 4-39 to
4-76, V. T. Chow (ed.), McGraw-Hill, Inc., New York, New York (1964).
9 Q
"• Strahler, A. N., "Hypsometric (Area-Altitude) Analysis of Erosional
Topography," Geo. Soc. Amer. Bull.. 63:1117-1142 (1952).
87
-------
24. Melton, M. A., "An Analysis of the Relations Among Elements of Climate,
Surface Properties and Geomorphology," Project No. NR 389-042,
Technical Report No. 11, Columbia University, Department of Geology,
New York (1957).
25. Maxwell, J. C., "Quantitative Geomorphology of the San Dimas Experi-
mental Forest, California," Project No. NR 389-042, Technical
Report No. 19, Columbia University, Department of Geology, New
York (1960).
26. Smith, K. G., "Erosional Processes and Landforms in Badlands National
Monument, South Dakota," Geo. Soc. Amer. Bull.. 69_:975-1008 (1958).
27. Carlston, C. W., and W. B. Langbein, "Rapid Approximation of Drainage
Density: Line Intersection Method," U.S. Geological Survey, Water
Resource Division, Bulletin 11 (1960).
28. Glymph, L. M., "Relation of Sedimentation to Accelerated Erosion in
the Missouri River Basin," Soil Conservation Service, Technical
Paper No. 102 (1951).
29. Leopold, L. B., W. W. Emmett, and R. M. Myrick, "Channel and Hillslope
Process in a Semiarid Area in New Mexico," UiS. Geological Survey,
Paper No. 102 (1966).
30. Gottschalk, L. C., "Effect of Watershed Protection Measures on Reduc-
tion of Erosion and Sediment Damages in the United States," Int.
Assoc. Sci. Hyd. Pub. , 59.:426-427 (1962).
31. Anderson, H. W., "Relative Contribution of Sediment from Source Areas
and Transport Processes," in Proceedings of a Symposium on Forest
Land Uses and Stream Environment, Oregon State University, pp. 55-
63, August 1972.
32. U.S. Army Corps of Engineers, "A Study of Streambank Erosion in the
United States," submitted to Committee on Public Works, House of
Representatives, October 1969 (available from the U.S. Government
Printing Office, Washington, D.C.).
33. Dyrness, C. T., "Mass Soil Movements in the H. J. Andrews Experimental
Forest," USDA Forest Service Research Paper PNW-42, Pacific Northwest
Forest and Range Experimental Station, Portland, Oregon, 12 pages
(1967).
88
-------
34. Swanston, D. N., "Principal Mass Movement Processes Influenced by
Logging, Road Building, and Fire," in Proceedings of a Symposium
Forest Land Uses and Stream Environment, Oregon State University,
Corvallis, Oregon, 19-21 October 1970.
89
-------
SECTION 4.0
NUTRIENTS AND ORGANIC MATTER
4.1 INTRODUCTION
Nitrogen and phosphorus are the primary nutrients which are important in
agricultural and silvicultural practices. The effect of these nutrients on
receiving waters is the increased potential for algal blooms — especially in
lakes and reservoirs--thus interfering with many beneficial uses of these
waters. Of the two nutrient elements, phosphorus has received greater em-
phasis because of the available technology to control phosphorus discharges
from municipal and industrial sources. Nitrogen is also important as a
rate-limiting nutrient for algal growth in some surface waters; however,
the nitrogen pathways in plant nutrition are relatively more complex than
those of phosphorus. Technology for controlling nitrogen emissions from
point sources is not sufficiently advanced to economically justify its
adaptation to nonpoint pollutant emissions.
The magnitude of losses of these two nutrient elements from different
source activities can, in principle, be calculated by making nutrient bud-
gets of all source inputs and outputs, and specifically determining out-
puts to surface waters. Methods for estimation of quantities involved in
the several parts of a nutrient budget are not well enough developed for
use in nutrient loading functions. In addition, the quantities of nu-
trients that actually reach a stream from a given source are subject to
variation depending upon the nature of the intervening terrain. The pre-
diction of nutrient losses from various land uses can in part be accomp-
lished by loading functions which describe the changes of nutrient con-
tent in the soil in response to various external variables such as cultural
practices, fertilizers, and climatic differences, and which account for
soil losses by erosion.
Organic matter from cropland and pastureland carries oxygen-consuming ma-
terials that can degrade the quality of receiving waters by stripping its
oxygen content, and carries potentially pathogenic microorganisms from
livestock wastes and other rural runoff. A loading function for organic
90
-------
matter has been developed based on the organic matter content of soil
and sediment yield.
These assumptions are more nearly correct for nitrogen when erosion is
moderate to extensive, and are less correct when erosion is slight or
when surface runoff is negligible. In the latter cases dissolved forms
of nitrogen are the principle nitrogen pollutants. These are transported
either to subsurface waters or directly to surface waters in runoff.
Functions which describe either of the latter phenomena are not yet avail-
able, and the approach to estimating dissolved forms of nitrogen accord-
ingly involves a combination of local or regional experience supplemented
by measurements of soluble nitrogen forms in runoff and baseflow.
Nutrient and organic matter loading functions presented in this section
are accordingly based on the sediment loading function developed in Sec-
tion 3.0 entitled "Sediment Loading Functions." It is assumed that the
nutrients and organic matter are carried through surface runoff and that
most of these are removed with sediment.
Because the currently available data applicable to the entire U.S. may
not reflect the local conditions, it is suggested that local data when-
ever available be used in preference to the general data presented in
this section.
4.2 NITROGEN
4.2.1 Introduction
Soil nitrogen is derived from several sources which include geologic
weathering, microbial reactions, precipitation, and chemical fixation.
Addition of chemical fertilizers and organic residues to soil constitutes
man's effort to increase or supplement nitrogen forms which can be read-
ily utilized by plants. Although the cultivated soils contain a large
reservoir of total nitrogen in the plowed layer—about 2 to 4 tons/acre--
available nitrogen is usually quite small—a few pounds per acre. The
significance of this available nitrogen to water pollution is great, how-
ever. As much as 95% of total nitrogen in the soil is organically bound
and is not readily released in solutions for plant growth. The ammonium
ion in soil which is tightly bound to clay or other anionic molecules in
soil is also not readily available for plant growth. Nitrate which is
not held by soil particles can be readily transported through the soil
profile to below the root zone in the absence of an actively growing crop
and can eventually join the groundwater pool. The time of migration of
groundwater nitrogen to surface waters can extend to several decades
91
-------
depending upon groundwater hydrology relative to surface water hydrology.
Significant nitrogen losses to the air occur through volatilization and
denitrification processes.
4.2.2 Precipitation
Precipitation contains significant quantities of numerous substances,
including nitrogen and phosphorus .ii^' That precipitation which falls
on surface waters carries with it a load which becomes a part of the
total pollutant load. The direct contribution via precipitation is neg-
ligible for surface streams, and may be substantial for lakes or still-
standing waters—as much as 5070 of the total nutrient input.—' Contri-
butions of precipitation-borne nutrients to surface waters via overland
runoff will vary in proportion to both precipitation and runoff. The
simplest approach is to assume that overland runoff carries with it,
without loss to the soil, the nitrogen and phosphorus load which it con-
tained when it reached the earth. Overland runoff is seldom very direct
except in high intensity/high quantity storm events or in certain types
of snowmelt, and rainfall entrained nutrients will in most runoff events
be exposed to mineral and organic matter in the soils. Phosphorus and
nitrogen should be somewhat attenuated by exposure to the soil.
That fraction of precipitation-borne phosphorus carried in precipitation
which does not discharge to streams via overland runoff becomes a part
of the inventory of phosphorus in the soil, and becomes relatively im-
mobile in the surface layers of soil. The surface-sorbed phosphorus be-
comes a nonpoint pollutant when it is discharged to streams on eroded
sediment.
That fraction of precipitation-borne nitrogen which is not immediately
carried off in overland runoff also enters the soil compartment where it
continues its participation in the complex nitrogen cycle: some stays
in the root zone, and may be completely utilized by plant life; some
moves below the root zone, and thus becomes involved in a very ill-
defined physical-chemical-biological-hydrologic system; some of that
which stays in the root zone is a candidate for transport, later, to
surface streams in overland runoff.
Since only a small fraction of precipitation incident on land enters
surface waters by overland runoff, the great majority of precipitation-
borne phosphorus and nitrogen is deposited on the land and becomes a part
of its continually changing inventory of nutrients. The present discus-
sion is concerned with estimation of the fractions of the precipitation-
borne nutrients transported directly, via overland runoff, to surface
waters. An analysis of "national average" data is instructive.
92
-------
Annual average precipitation is 76 cm (30 in.)- Annual average runoff
via all processes is 25 cm (10 in.)- The fraction of runoff occurring
by the overland varies widely; for purposes of discussion 20% of total
runoff will suffice. Average annual overland runoff is thus 5 cm (2
in.), or about 7% of precipitation.
Reported deposition rates of nitrogen and phosphorus in rainfall range
from about 5 to 10 kg/ha/year (4.4 to 8.9 Ib/acre/year) for nitrogen,
and reportedly average 0.05 to 0.06 kg/ha/year (0.045 to 0.055 lb/acre/
year) for phosphorus.J^r'
Seven percent of the precipitation-borne phosphorus and nitrogen might
thus be carried directly to surface waters, if no absorption on soil is
assumed. Nonattenuated yield rates, for stream deposition, national
average basis, would accordingly be 0.35 to 0.7 kg/ha (0.31 to 0.62-
lb/acre) of nitrogen, and 0.0035 to 0.004 kg/ha (0.0031 to 0.0036 Ib/
acre) of phosphorus. If one assumes that phosphorus is 5070 attenuated
and nitrogen 25% attenuated, the net yields become 0.28 to 0.53 kg/ha
(0.25 to 0.47 lb/acre) of nitrogen, and 0.0018 to 0.002 kg/ha (0.0016
to 0.0018 lb/acre) of phosphorus.
If one translates the above data into in-stream concentrations (assum-
ing no in-stream transformations), the results are 0.11 to 0.21 ppm
nitrogen, and 0.7 to 0.8 ppb of phosphorus. Comparison of these con-
centrations with the national benchmark station data summarized in
Figures 12-3 and 12-4 reveals the perhaps fortuituous comparison that
nitrogen concentrations estimated from precipitation are the same as
what appears to be an average for nationally observed concentrations
in locations relatively unaffected by man. The above estimated concen-
trations for phosphorus are lower than benchmark station concentrations
(0.7 to 0.8 ppb vs 10 to 200 ppb of total phosphorus). This compari-
son indicates that the load of precipitation-borne phosphorus is a small
fraction of the phosphorus nonpoint contribution to surface streams, but
that nitrogen contributions are a significant part of the in-stream bur-
den of available forms of nitrogen (particularly nitrate).
A comparison of nutrient contribution from precipitation with that from
croplands reveals that, on a national basis, the eroded soil from crop-
lands yields about 20 kg/ha/year (18 Ib/acre/year) of total nitrogen..1'
Assuming a 77» value for the available fraction in total nitrogen, the
load of available nitrogen from cropland becomes 1.42 kg/ha/year (1.26
Ib/acre/year). This value compares with 0.28 to 0.53 kg/ha/year (0.25
to 0.47 Ib/acre/year) of available nitrogen in precipitation. Since the
93
-------
cropland nitrogen loading function does not account for precipitation
loads, the total contribution to a stream should include both these
sources. The total load of "available" nitrogen thus is about 1.8 kg/
ha/year (1.6 Ib/acre/year), on a national average basis, from cropland.
Although available nitrogen is extremely significant in the enrichment
of stream nutrition, the role of the remainder of the total nitrogen
carried on eroded sediment is also substantial. Since streams are dy-
namic in nature, there is a continuous mineralization of soil nitrogen
by the microorganisms in the bottom sediment which is supplied with
oxygen from both stream reaeration processes and photosynthetic pro-
cesses. Thus, the delayed release of available nitrogen to the aquatic
systems can be as significant as the available nitrogen in precipi-
tation and eroded soil. For example, in-stream nitrogen burdens averaged
over the Missouri River basin translate to an average yield of about 3
lb/acre/year_' of nitrate-nitrogen, which is two to three times the de-
livered rate from nonpoint sources and precipitation.
Nitrogen loading from precipitation should be added to that from surface
erosion processes to obtain the total load. Since the load for phosphorus
from precipitation is small, the phosphorus loading function does not in-
clude the contribution from precipitation.
4.2-3 Nitrogen Loading Function
While the complex interactions in soil, air, water, and plants are rea-
sonably well understood, methods for quantifying movements within the
system are still in the research stage. Methods which are suitable for
general use oversimplify the problem, must be used with discretion, and
may be quite inadequate in certain cases. In particular, it is not
presently possible to describe leaching processes for soluble forms of
nitrogen. The nitrogen loading function is made up of two sources: (a)
erosion; and (b) precipitation. Total nitrogen loading is obtained by
adding the yields from both sources. The loading functions exclude
leaching losses, and predict the amount of total nitrogen that is re-
leased to surface waters by runoff and erosion. The nitrogen in precip-
itation is mostly in available form.
Nitrogen loading function for erosion loss is:
Y(NT)E = a'Y(S)E-Cs(NT)-rN (4-1)
94
-------
where Y(NT)g = total nitrogen loading from erosion, kg/year (Ib/year)
a = dimensional constant (10 metric, 20 English)
CS(NT) = total nitrogen concentration in soil, g/100 g
Y(S)E = sediment loading from surface erosion, MT/year (tons/
year)
r^j = nitrogen enrichment ratio
Available nitrogen can be obtained by using a fraction f^ which is the
ratio of available N to total N in sediment. Thus, the available
nitrogen in sediment is
Y(NA)E = Y(NT)E'fN- (4-2)
Nitrogen loading function for precipitation is
Q(OR)
Y(N)p = AO - .Npr-b (4-3)
^ Q(Pr)
where Y(N)pr = stream nitrogen load from precipitation, kg/year
(Ib/year)
A = area, ha (acres)
Q(OR) = overland flow from precipitation, cm/year (in/year)
Q(Pr) = total amount of precipitation, cm/year (in/year)
Npr = nitrogen load in precipitation, kg/ha/year (lb/acre/
year)
b = attenuation factor
Almost all of Y(N)pr will be in the available form so that the total
available nitrogen from both erosion and precipitation may be obtained
by adding Eqs. (4-2) and (4-3). Thus,
Y(NA) = Y(NT)E'fN + Y(N)pr (4-4)
95
-------
4.2.4 Evaluation of Parameters in the Nitrogen Loading Function
The value of Y(S)g can be evaluated from the sediment loading function
presented in Section 3.0 "Sediment Loading Functions." The value of the
enrichment ratio r^ is variable according to the soil texture and cul-
tural treatment. VietsJ/ presented the values of r^ using data from
small experimental plots (see Table 4-1). Hagin and Amberger,.?-' as well
as Stoltenberg and White ,Z' have proposed an rN value of 2.0. Massey
et a.1.§.' estimated the value of r^ as 2.7. Because of wide variations
in the properties of erodible soil, a single value of rN is not prob-
able; the values reported range from 2.0 to 4.0, and a value in this
range should be selected for a specific location unless local data are
available.
Table 4-1. NUTRIENT AND SEDIMENT LOSSES^
5/
Source
Total loss (kg/ha)
Soil N
Enrichment
ratio, r
N P
Check
Rye winter cover crop
Manure (45 MT/ha)
Rye and manure (45 MT/ha)
29,100
13 , 160
18,390
8,130
74.5
38.9
52.8
21.5
75.8
37.7
44.3
19.6
3.88
4.08
4.28
3.35
1.59
1.56
1.47
1.47
Nutrient losses from forest soils are typically very low. Kilmer^/
cited several authors to show that nutrient losses from forestlands are
insignificant. Clear-cutting and burning of forest areas appear to be
the most important practice involved in accelerated release of nutrients.
Table 4-2 shows that clear-cutting and nitrogen fertilization accelerated
nitrogen and phosphorus losses to a slight extent.
96
-------
Table 4-2. EFFECT OF CLEAR-CUTTING AND FERTILIZATION ON NUTRIENT
OUTPUT IN DOUGLAS FIR FORESTS^/
N P
Treatment (kg/ha) (kg/ha)
Control 0.21 0.01
Clear-cut 0.39 0.05
Fertilized (200 Ib/acre) urea 0.28 0.03
Ammonium sulfate 0.43 0.07
a/ Source: Cole and Gessel (1965) cited by Kilmer.2'
Nitrogen losses by leaching are also negligible from actively growing
grassland. However, losses from legume grass mixtures can be high.
Lysimeter studies by Low and Armitage (page 7 of Ref. 9) showed that
clover produced about 10 times as much N loss in drainage as that in
actively growing grass; however, the loss was 100 times as much when
the clover crop died.
Runoff losses of nitrogen from grass sod plots ranged from 27° of applied
nitrogen when soil moisture was 12.5%, to 14% at 25.8% moisture.—/
Timmons et al.ii' determined N and P losses in runoff solution and sed-
iment in Minnesota. Their results indicate that leaching losses from
a hay rotation could contribute to substantial N and P losses in solu-
tion.
The value of CS(NT) in the plowed layer of soil is variable from location
to location and from time to time. Estimates of native soil nitrogen
in the U.S. indicate a range between 0.02 and 0,^%.—' Parker et al.
published a map in 1946 showing the nitrogen content in the top 1-ft
layer in the U.S.M' (see Figure 4-1). Since 1946 the nitrogen content
of the cropland soil has most probably decreased due to cultivation.
However, fertilizer inputs to cropland have offset part of the depleted
nitrogen.
Precipitation also contributes to the soil nitrogen. Atmospheric nitro-
gen extracted by soil microbes becomes incorporated into soil organic
matter; animal manures, crop residues, and other wastes contribute sig-
nificant amounts of nitrogen to the soil. Jenny—' expressed the nitro-
gen content of the soil in terms of temperature, T, and a humidity fac-
tor, H. Jenny's equation is:
97
-------
z
LLJ
o
o
cu p
w
t-l 4J
CU C
> cu
•H -H
Q 0
|
* £
3 •£
^O
0
C)
k.
Ov
o
o
1
o
0
1
o
CM
O
1
>o o
o
CM
O
O
CO
o
CO
4J
o
o
(U
o
dJ
3
to
C
•H
W)
O
M
•U
•H
C
4J
C
(U
O
•<)•
cu
^
60
98
-------
CS(NT) = 0.55 e-0-08T (1 _ e-0.005H} (4_5)
H = "(4-6)
where P = precipitation, mm/year
CS(NT) = concentration of soil nitrogen, g/100 g
T = annual average temperature , °C
RH = relative humidity, %
SVPt = saturated vapor pressure at given temperature, mm
of Hg
Equation (4-7) shows the relation between SVPt and T.*
- 2360/(273+T)]
The solution of Eq. (4-5) is shown graphically in Figure 4-2. The value
of humidity factor, H, can be determined from Eqs. (4-6) and (4-7). A
nomograph solution of H is shown in Figure 4-3. For given values of
precipitation, relative humidity and temperature, the value of H can be
quickly and accurately established from Figure 4-3. For example, given
P-L = 500 mm/year (19.7 in/year), RH^ = 60%, and TI = 5°C (41°F), the
value of H factor can be determined as follows: using a straight-edge
ruler, align P^ and RH^ to intersect on the index line at "a" as shown
on the inset of Figure 4-3. Align "a" with T^ on the temperature scale
to intersect the H scale. The result on the H scale is 194.
Data in Figure 4-1 may be used as a check on current data. Equations
(4-6) and (4-7) may be used to calculate nitrogen content of soil more
precisely if necessary data are available for using these equations. If
local data are considered to be more reliable than those presented herein,
local data may be preferentially used.
The fraction of available nitrogen to total nitrogen in soil, f>r is
variable, depending upon many factors such as soil characteristics, degree
of mineralization, and organic matter content. The most important forms
of available nitrogen are NH/+, N0o~, and certain simple organic compounds
containing free amide or amino groups. Nitrate is only a minor source of
available nitrogen in soil.
Modified from Gladstone, S., Elements of Physical Chemistry, D. Van
Nostrand Company, Inc., New York, New York (1946).
99
-------
0.01
100
200
300 400
H, HUMIDITY FACTOR
500
600
700
Figure 4-2. Soil nitrogen vs humidity factor and temperature
100
-------
1 I.I
1 '
m
i
0
CO
i . . .1 u i . i .
M 1 f I 1
O
O
,,,!,.
1
in
o
.1.. ,,!,,,
1 ' M '
O
« —
O
•o
L, 1.11,1,., ,
M 1 f 1 1 1
in
o
i. . . .I. .
1 i ' •
o
CM
.,1...
If,
in
CN
o
00
.1 ..
i i i
* 1 '
o
CO
o
o*
.I...,)
"1 1 1 1
in
CO
IE
u
t_
i — i «>
|_ ^O.
Q_
1— sQ + "\ *
> CN {^J <- E • —
1/1 CM £ _ ._
o , 8 g |
2 o- "- 'z x
^ £ .2-J a>
°^ fK ~^ *^ .—
i i — . 'E '5 "5
- o = £ ~Z
•—• ' •— JC Q_ a:
M II II II II
(U
x £ S "i
> -g
LO £
C
6
'o
I
V)
1/1
ro
X
w-
O
E
V
3
'o
k_
a.
V
t—
f ^
o
W
U.
D
S
1 X~
i—
t>
D)
S o
CO
1
U
•»-
._
c
. "5
/ ^™ * o
/ s «
!/
4) *- "5>
> y _si 4>
"^ f **• -1
" >— • I?
H- a-
M
O
u
cd
•r-l
e
JO4ODJ X4ipiiun(-| '
oo o o
oo o o
oo o o
o
o
o
CM »—
I O O O
) O O O
I •«*• CO CXJ
O
O
oo o o
CO CN
CN
o
M-l
60
O
o
o
CO
o
co
o
in
o
CO
^ I
o o
CN •—
aui-| xapu|
'uoj4D4|dpajj '
O
CN
| ,
1
00 Is
o
o
in
in 1 1
1 1 '
3 0 C
s -O U
o
o o o
o o o
*— 00 hx
i i 1 , ii i , i
, , ,1,1.
DO O
"} '"^ CO
o o o o o c
O O O O O u'
•o in -«t co O
1 O O O O 0 O
f— co hx ^o in •<•
, , . , I . I , l I l . 1 , , i,,,,
1 j ' r I i'1'1 '"' i
3 in •* co CN m
^-'
o o in c
CO CM ~- -
...,|,.,. ..,.1..,, 1 , , , ,
'Ml 'd , \ , j ,
o o oor^ -o in -^
— ooo o o c
en
i
60
•r-l
101
-------
The available nitrogen in soil rarely exceeds 10% of total nitrogen.
Data from Lopez and Galvezlz' suggest that about 8% of total nitrogen
in soil is available in mineralized form for plant growth.
The values of Q(OR) and Q(Pr) may be obtained from local data sources.
The value of Q(Pr) (annual average precipitation) is usually obtained
from the weather bureau statistics for the area. The value of overland
runoff can be roughly estimated from stream flow data. A user unfamiliar
with hydrology should consult with qualified personnel in state conser-
vation services, agricultural extension service, the Corps of Engineers,
or the Agricultural Research Service for assistance in interpretation of
stream flows. These resources will also have historical information on
overland runoff in relation to precipitation.
Values of Np are usually available from measurements made in the
local research stations. In the absence of actual data, data in Figure
4-4 may be used.
4.3 PHOSPHORUS
4.3.1 Introduction
Phosphorus occurs naturally in soil from weathering of primary phosphorus-
bearing minerals in the parent material. Additions of plant residues
and fertilizers by man enhances the phosphorus content of the surface
soil layer.
Phosphorus in soils occurs either as organic or inorganic phosphorus.
The relative proportion of the phosphorus in these two categories varies
widely. Organic phosphorus is generally high in surface soils where or-
ganic matter tends to accumulate. Inorganic forms are prevalent in sub-
soils. Soil phosphorus is readily immobilized due to its affinity to
certain minerals. In strongly acid soils the formation of iron and
aluminum phosphates, and in alkaline soils, the formation of tricalcium
phosphate reduces the availability of soil phosphorus. Once it is lost
to a stream, the nature of phosphorus existing in sediment or in solu-
tion becomes significant in the nutrition of aquatic microorganisms.
Phosphorus transport from a given site to stream can occur either by ero-
sion or by leaching. The predominant mode of transport is via soil ero-
sion. Soil solution usually contains less than 0.1 ug of phosphorus per
milliliter; the leaching losses are thus extremely low even in well-
drained soils. Exceptions are sands and peats which have little tendency
to react with phosphorus.
102
-------
oo
^
w
p*
>^ (U
rt C
« O
5-1 O
CJ W
•H
• 5
•H
a,
•r-l
o
0
•I-l
103
-------
Phosphorus losses from well managed pastures and forested soils are
usually low. For example, unfertilized pastures lost about 0.03 kg/ha
of P during a 6-month period, while addition of 45 kg of P per hectare
resulted in an escape of only 0.04 kg/ha during a similiar period of
time.2/
4.3.2 Phosphorus Loading Function
The loading function for phosphorus is based on the soil erosion mecha-
nism. The loading function is:
Y(PT) = a.Y(s)E-Cs(PT)-rp (4-8)
where Y(PT) = total phosphorus loading, kg/year (Ib/year)
a = a dimensional constant (10 metric, 20 English)
Y(S)g = sediment loading, MT/year (tons/year)
C0(PT) - total phosphorus concentration in soil, g/100 g
o
rp - phosphorus enrichment ratio
Available phosphorus may be computed from Eq. (4-9):
Y(PA) = Y(PT)-fp (4-9)
where Y(PA) = yield of available phosphorus, kg/year (Ib/year)
fp = ratio of available phosphorus to total phosphorus
4.3.3 Evaluation of Parameters in Phosphorus Loading Function
Sediment loading, Y(S)g , may be obtained from procedures outlined in
Section 3.0 "Sediment Loading Functions."
The value of Cg(PT) , the total phosphorus content of the soil, is
variable. For any given location, current and local data are preferred
to generalized values given in this report. No central repository of
current nationwide data exists. Parker et al.J^' published data on the
phosphorus content of soil in the top 30 cm (1 ft) for the 48 states, as
shown in Figure 4-5. Parker's data, although obtained 30 years ago, will
serve as a check on current data. Soil surveys periodically made by the
104
-------
105
-------
Soil Conservation Service contain more recent information on soil phos-
phorus content. State agricultural extension service personnel can also
provide reasonable estimates of soil phosphorus content in a given area.
These sources should be given priority in determining the phosphorus
content of the soil.
The enrichment ratio, rp , has been the least researched parameter in
the loading function. As reported in Table 4-1, the reported rp values
Q /
average about 1.5. Massey et al.—' obtained an rp value of 3.4, and
Stoltenberg and WhiteZ' reported a value of 2.0. Hagin and Amberger^.'
have used a value of rp of 2.5 in their simulation model for nutrient
losses from agricultural sources. Massey et al.®.' have developed an
empirical equation to determine rp :
log rp = 0.319 + 0.25 (-log X) + 0.098 (-log Y) (4-10)
where X = sediment loss, tons/acre-in of runoff
Y = sediment loss, tons/acre
The determination of available phosphorus in the soil is difficult. Most
reported data fail to distinguish between soluble phosphorus, adsorbed
or particulate phosphorus, and organic phosphorus in sediment runoff.
Total phosphorus is a somewhat meaningless parameter, since only the
soluble orthophosphate form is readily available for uptake by aquatic
organisms. Other forms of phosphorus in sediment can, however, act as
a source or sink for subsequent release in available form.
Schuman et al. have reported an empirical relation between sediment phos-
phorus (concentration in ppm, Cs(PT) ) and soluble phosphorus (concen-
tration in ppm, Cq(P) ) for Iowa soils. The relation may be stated as:
CQ(P) = a + b-Cs(PT) (4-11)
where a and b are regression coefficients. The reported values of
a and b are 0.018 and 0.047, respectively.!^/ Equation (4-11) shows
that the ratio of solution phosphorus to sediment phosphorus is just
under 1:20.
Taylor—' suggested that about 107o of the total phosphorus in eroded
soil would be available for aquatic plant growth.
106
-------
4.4 ORGANIC MATTER
4.4.1 Organic Matter Loading Function
The loading function is:
Y(OM)E = a-Cs(OM)-Y(S)E-rOM (4-12)
where Y(OM)E = organic loading, kg/year (Ib/year)
a = a dimensional constant (10 metric, 20 English)
CC(OM) = organic matter concentration of soil, g/100 g
D
Y(S)E = sediment loading, MT/year (tons/year)
rOM = enrichment ratio for organic matter in eroded soil
4.4.2 Evaluation of Parameters in the Organic Matter Loading Function
The value of Y(S)E can be obtained from procedures discussed in
Section 3.0. The value of CS(OM) should be obtained preferably from
current or historical data for a given area, e.g., from the extension
service. For approximate values, Cg(OM) may be taken as equal to
20 x Cs(NT) , where Cs(NT) is the total nitrogen concentration in
the soil.lZ/
The value of TQ^ , the enrichment ratio, is more difficult to assess
due to lack of research data. Values of rg^ are in the range of 1 to
5. The enrichment ratio for sandy soils will be high. Conversely, the
enrichment ratio will be low when the mineral fraction of the soil is
finely divided and highly erodible. The user should consult with local
soil experts and should use local data when available.
4.5 ACCURACY OF LOADING FUNCTIONS
The accuracy of predicting loads using the loading functions presented
in the preceding sections depends, to a large extent, on the availability
of reasonably accurate data for evaluating the various parameters in the
functions. For example, the nitrogen loading function is composed of
several parameters each of which is in turn a function of several other
variables. In addition, several options are available to the user to
develop the parameter values from his own sources of information which
may alter the prediction accuracy. However, if the used values reflect
the long-term average rather than a specific year, and if reasonably
large areas are used such as large watersheds (> 100 sq miles) rather
107
-------
than individual plots or small watersheds, the expected accuracy can be
reasonably estimated. Using the reasoning that the error in individual
parameters will tend to cumulate to a larger error, the expected ranges
of predicted values for given "true" or estimated values of load are
presented in Table 4-3.
Table 4-3. PROBABLE RANGE OF LOADING VALUES FOR
NUTRIENTS AND ORGANIC MATTER
Estimated value Probable range
Loading function (kg/ha/year) (kg/ha/year)
Total N sediment*/ 1 0.1-10
Total N sediment 10 5-20
Total N sediment 50 30-75
Total N precipitation*!/ 0.3 0.1-0.6
Total P.£/ 1.0 0.5-3.0
Total P 5.0 2-10
Total P 10.0 5-20
Organic matter 10.0 5-20
Organic matter 100 50-200
a./ Available N in sediment will range from 3 to 87<, of total N.
b/ Available N is equal to total N in precipitation.
£/ Available P in sediment will range from 5 to 10% of total P.
4.6 EXAMPLE OF LOADING COMPUTATION
The watershed given in Section 3.0, entitled "Sediment Loading Func-
tions," for Parke County in Indiana will be used to illustrate the metho-
dology presented in this section for computing the loads. It is required
to compute available nitrogen, available phosphorus, and organic matter
loading for the given area for the following conditions:
Average, daily loading;
Maximum daily loading during a 30 consecutive day period; and
Minimum daily loading during a 30 consecutive day period.
108
-------
The following data, plus soil loss data, are required:
Soil nitrogen content.
Soil phosphorus content.
In the absence of reliable soil nitrogen data, soil nitrogen con-
tent may be calculated from average annual temperature, average
annual precipitation, average annual relative humidity, and
saturated vapor pressure at given temperature.
4.6.1 Nitrogen Loading
Using the following data, soil nitrogen content is calculated:
Average annual temperature = 10°C
Average annual precipitation = 96.5 cm
Average annual relative humidity = 70%
Using the nomograph given in Figure 4-3, the value of H factor was de-
termined to be 350. From Figure 4-2, and using H = 350 and T = 10°C,
the value of Cs(NT) , the soil nitrogen content was estimated to be
0.204% or 0.204 g/100 g. Using Eq. (4-1).
Y(NT)E = 20-Y(S)E-0.204-2.0 (4-13)
= 8.16-Y(S)E
Assuming that 6% of total nitrogen is available, Y(NA)g = 0.49'Y(S)£.
The values of areal sediment yield as given in the example in Section
3.0, entitled "Sediment Loading Functions," are shown below in Table 4-4.
Table 4-4. SEDIMENT YIELD IN EXAMPLE
Sediment yield (tons/day)
Land use
Cropland
Pasture
Woodland
Daily average
2.88
0.33
0.39
Maximum 30 days
9.36
0.84
0.95
Minimum 30 days
0.72
0.09
0.09
Total 3.60 11.15 0.90
109
-------
The nitrogen loadings are shown in Table 4-5 using the data in Table
4-4 and Eq. (4-13).
Table 4-5. AVAILABLE NITROGEN LOADING, Y(NA)E , IN EXAMPLE
Land use
Cropland
Pasture
Woodland
Daily average
1.41
0.16
0.19
Nitrogen loading (Ib/day)
Maximum 30 days
4.59
0.41
0.47
Minimum 30 days
0.35
0.04
0.04
Total 1.76 5.47 0.43
4.6.2 Phosphorus Loading
Assume CS(PT) = 0.255g/lOOg for the area, 10% of CS(PT) is available
phosphorus, Cg(PA) ; and rp is 1.5, and using Eq. (4-8);
Y(PA)E = 20-Y(S)E-0.255-1.5-0.10 (4-14)
= 0.765 Y(S)E
Phosphorus loadings computed from Table 4-2 and Eq. (4-8) are shown in
Table 4-6.
Table 4-6. AVAILABLE PHOSPHORUS LOADING, Y(PA)s , IN EXAMPLE
Land use
Cropland
Pasture
Woodland
Daily average
2.20
0.25
0.30
Phosphorus loading
Maximum 30 days
7.16
0.64
0.73
(Ib/day)
Minimum 30 days
0.55
0.07
0.07
Total 2.75 8.53 0.69
4.6.3 Organic Matter Loading
Isomg Eq. (4-12), data for Cg(OM) , Y(S)g , rQM are needed.
110
-------
Assume that the value of CS(OM)/CS(NT) equals 20 and rQM =2.5,
Y(OM)E = 20-2.5-Y(S)E'20-Cs(NT)
= 1000-CS(NT).Y(S)E
Using CS(NT) = 0.2%,
Y(OM)E = 200-Y(S)£ (4-15)
The values of organic loading are computed from Eq. (4-12) and presented
in Table 4-7.
Table 4-7. ORGANIC MATTER LOADINGS IN EXAMPLE
Organic matter loading (Ib/day)
Land use Daily average Maximum 30 days Minimum 30 days
Cropland 576 1,872 144
Pasture 66 168 18
Woodland 78 190 18
Total 720 2,230 180
111
-------
REFERENCES
1. Carroll, D., "Rainwater as a Chemical Agent of Geological Processes -
A Review," GS WS Paper 1535-G, U.S. Geologic Survey (1962).
2. Loehr, R. C., "Characteristics and Comparative Magnitude of Nonpoint
Sources," JWPCF, 46(8):1849 (August 1974).
3. Phase II Report, EPA Contract No. 68-01-2293 (Draft submitted in
November 1975).
4. McElroy, A. D., S. Y. Chiu, and A. Aleti, "Analysis of Nonpoint
Source Pollutants in the Missouri Basin Region," Office of Re-
search and Development, Environmental Protection Agency, Report
No. EPA-600/5-75-004 (March 1975).
5. Viets, F. G., Jr., "Fertilizer Use in Relation to Surface and
Groundwater Pollution," In: Fertilizer Technology and Use (2nd
ed.), p. 517, Soil Science Society of America, Madison, Wisconsin
(1971).
6. Hagin, J., and A. Amberger, "Contribution of Fertilizers and Manures
to the Nitrogen and Phosphorus Load of Waters. A Computer Simula-
tion," Technion-Israel Institute of Technology, Haifa, Israel
(1974).
7. Stoltenberg, N. L., and J. L. White, "Selective Loss of Plant Nu-
trients by Erosion," Soil Science Society of America, Proceedings,
17:406-410 (1953).
8. Massey, H. F., M. L. Jackson, and 0. E. Hays, "Fertility Erosion on
Two Wisconsin Soils," Agron. J., 45:543-547 (1953).
9. Kilmer, V. J., "Nutrient Losses Through Leaching and Runoff,"
Tennessee Valley Authority, Muscle Shoals, Alabama (undated manu-
script).
10. Moe, P. G., J. V. Mannering, and C. G. Johnson, "Loss of Fertilizer
Nitrogen in Surface Runoff Water," Soil Sci. , 104j389-394 (1967).
11. Timmons, D. R., R. F. Holt, and J. J. Latterell, "Leaching of Crop
Residues as a Source of Nutrients in Surface Runoff Water," Water
Resources Research, 6:1367-1375 (1970).
112
-------
12. Jenny, H. , "A Study on the Influence of Climate Upon the Nitrogen
and Organic Matter Content of the Soil," Missouri Agr. Exp. Sta.
Res. Bui. 152 (1930).
13. Parker, C. A. et al., "Fertilizers and Lime in the United States,"
USDA Misc. Pub. No. 586 (1946).
14. Lopez, A. B., and N. L. Galvez, "The Mineralization of the Organic
Matter of Some Philippine Soils Under Submerged Conditions,"
Philippine Agr. , 4-2:281-291 (1958), cited in Ref. 17.
15. Schuman, G. E., R. G. Spomer, and R. F. Piest, "Phosphorus Losses
from Four Agricultural Watersheds on Missouri Valley Loess,"
Soil Science Society of America, Proceedings, 3_7_(2):424 (1970).
16. Taylor, A. W., "Phosphorus and Water Pollution," J. Soil and Water
Conserv. , £2:228-231 (1967).
17. Buckman, H. 0., and N. C. Brady, The Nature and Properties of Soil,
7th ed., The MacMillan Company, New York (1969).
113
-------
SECTION 5.0
PESTICIDES
5.1 INTRODUCTION
Pesticides dissipate by several mechanisms: chemical degradation (hy-
drolysis; oxidation); biochemical degradation by soil organisms and
enzymatic systems; volatilization; absorption in plant or animal tissue,
with or without decomposition; leaching into subsurface soils, possibly
into subsurface aquifers; and overland transport in surface runoff and
eroded sediment. Losses by leaching processes and by overland transport
mechanisms are relevant to contamination of water. Pesticide loading
functions must relate mechanisms for these processes to quantities de-
posited in surface waters. The total load of pesticide deposited in
surface waters equals the sum of (a) pesticide transported overland,
and (b) pesticide transported by subsurface processes (leaching, soil
moisture movement, drainage water movement, groundwater discharge to
surface). Soluble pesticides are subject to leaching into subsurface
soils and waters, solubilize in overland runoff water, and are also
transported overland as sediment-bound material. Insoluble pesticides
are transported to surface waters primarily by being carried on eroded
sediment.
Data requirements for a precise pesticide loading function are as fol-
lows:
1. Quantity of pesticide in the source, expressed as some suitable
function of the area, volume, or mass of the source, e.g., concentration
in erodible soil layer; concentration and concentration distribution in
leachable soil profile. The quantity information should be time spe-
cific, i.e., detail source quantities/concentrations as a function of
time elapsed since application, season, etc. Since most pesticides de-
grade, rates of degradation are needed to enable calculation of source
quantities as a function of time.
114
-------
2. Quantitative data on overland runoff, by month, season, and year.
3. Quantitative data on sediment transported from the source and deliv-
ered to surface streams.
4. Quantitative data on percolation; seepage; drainage water inventories
and movement; and groundwater inventories and movement.
5. Accurate coefficients, rate constants, etc., for desorption--
solubilization--leach transport of pesticides through soil columns, of
numerous possible soil types.
6. Information on miscellaneous modes of pesticide removal from the
source, such as by volatilization or by removal in harvested vegetative
matter.
Some of the required data is not available or is unknown, and other data
are known or available in varying accuracy and degree of coverage of
source situations.
The approach to estimation of contamination of water by pesticides will
therefore vary in response to a combination of three factors: (a) degree
of required accuracy; (b) availability of data; and (c) capabilities of
predictive functions. The greatest impediment is lack of data. Loading
functions and approaches to estimation of pesticide pollution are pre-
sented, in succeeding sections, for three source conditions. These are:
Case 1 - Water Insoluble Pesticides: Average concentrations of pesticide
in soils known. Pesticide load is calculated as a function of sediment
loads. Approach most applicable to large areas. Use limited to annual
average loads.
Case 2 - Water Insoluble Pesticides: Pesticide use history accurately
known, soil analytical data current and extensive, pesticide properties
(especially rates of disappearance) well known. Calculate load as fun-
ction of sediment loss; useful for annual average, 30-day maximum, 30-day
minimum.
Case 3 - Water Soluble and Water Insoluble Pesticides: Concentrations
in runoff waters known, runoff water flows known (stream source approach).
Calculate loads at watershed discharge points, distribute load over water-
shed land uses in proportion to known or probable pesticide use.
These approaches or options do not treat pesticides discharged to ground-
water aquifers and subsurface drainage. The latter can be treated if
115
-------
drainage discharge flows are known, together with concentrations of pes-
ticides in the drainage. Pesticide contamination in groundwater aquifers
is presently a research area.
These approaches do not preclude the use, in special or highly documented
situations, of research models or approaches which are being locally de-
veloped by research scientists.
5.2 PESTICIDE LOADING FUNCTIONS
5.2.1 Case 1; Insoluble Pesticides, Average Soil Concentrations L jwn
The loading function is:
Y(HIF)* = Y(S)E-C(HIF)-1(T6-A (5-1)
where Y(HIF) = pesticide yield for source, kg/day (Ib/day)
Y(S)E = sediment yield, kg/ha/day (Ib/acre/day)
C(HIF) = concentration of pesticide in soil (ppm)
A = size of source, ha (acres)
Sediment yields, Y(S)E , for the source are calculated by methods pre-
sented in Section 3.0.
Pesticide concentrations in soils throughout the United States are being
monitored by the Environmental Protection Agency, Office of Pesticide
Programs, in the National Pesticide Monitoring Program (NPMP). Data re-
positories for this monitoring program are a source of average soil con-
centration data. Results for 35 pesticides are summarized, for FY 1969
in Pesticides Monitoring Journal, ^(3):194-228, 1972. (This article is
reproduced in Appendix F.) The FY 1969 data cover cropland soils in 43
states and noncropland soils in 11 states.
The NPMP FY 1969 study is a source of soil concentration data which may
be used as C(HIF) values in Eq. (5-1). The "range of detected residues"
will serve as input for calculation, with Eq. (5-1), of the range of pes-
ticide loads which may be expected in the area of interest. Similarly,
the "percent positive sites" indicate whether a particular pesticide is
* KEF denotes Herbicide, Insecticide, and fungicide.
116
-------
distributed over much of the area or has limited distribution. The NPMP
information thus tends to be useful chiefly for estimating possible ex-
tremes in pesticide loads and for estimating total loads from a large
area such as a minor river basin.
It is imperative that the user of this function obtains up-to-date site
or area-specific data on soil concentrations. Current NPMP data should
be consulted, as should local sources of data, notably universities,
state and local health departments, and environmental agencies.
5.2.2 Case 2; Water Insoluble Pesticides, Current Area-Specific Data
Available
Case 2 covers the source with well-documented concentration data obtained
by analysis of samples taken from the source, in combination with pesti-
cide use data and knowledge of the persistence of the pesticide. If the
source is sampled frequently at well-distributed sampling sites, other
information may be unnecessary. If the sampling is less complete, in-
formation on application rates and persistency will help deduce concen-
trations. The basic loading function is the same as for Case 1, i.e.,
Eq. (5-1). The values used for C(HIF) are determined from different
sources than the sources for Case 1. Guidelines for determining C(HIF)
follows:
1. Document beginning of the season residual concentrations, if any, of
pesticides of interest.
2. Obtain data on application rates and schedules. Calculate concentra-
tion in surface soils (3 to 5 cm (1 to 2 in,)) of applied pesticide, tak-
ing into account the fraction of the pesticide which reaches the soil
surface, and the depth the pesticide is mixed into the soil.
3. Add values from Steps 1 and 2 to obtain an initial concentration.
4. From information on pesticide persistency, estimate fraction of pes-
ticide which remains after appropriate intervals of time: days for short-
lived pesticides; months for pesticides with growing season persistency;
and years for long-lived pesticides.
5. If pesticide is applied more than once per season, repeat Steps 1,
2, 3, and 4 for each application and estimate concentration throughout
growing season and up to the start of the next growing season.
6. Calculate sediment loads, Y(S)g , from sources by procedures pre-
sented in Section 3.0.
117
-------
Calculate annual average Y(S)_ if pesticides are relatively persistent
Ei
and a reasonable yearly average value can be deduced. Calculate Y(HIF)
from Eq. (5-1):
Y(HIF) average = *(S)E average-C(HIF) average-1Q-6
Calculate Y(S),., by months if pesticide concentrations vary widely through-
out the year. Calculate Y(HIF) annual average, 30-day maximum and 30-day
minimum by calculating monthly loads.
Y(HIF) monthly = Y(S)E monthlyC(HIF)•10"6
Sum for a year to obtain annual average. Select 30-day maximum and 30-
day minimum from computed monthly loads.
5.2.3 Case 3; Water Soluble and Water Insoluble Pesticides, Stream to
Source Approach
Water soluble pesticides are in part transported overland in surface
runoff and absorbed on sediments; they are also susceptible to migration
downward in the soil column, where they are not subject to overland trans-
port mechanisms. For lack of a procedure for predicting the ultimate fate
of the fraction which moves downward from the surface, it has been by-
passed in loading function development. That fraction transported over-
land may be estimated if runoff is measured and analyzed for pesticides.
Specifically, watershed hydrographs for storm events must be determined
by measurement, or calculated from predictive models,!—-' and concentra-
tions of pesticides determined for water samples collected at various
stages of the hydrograph(s)." The data so obtained convert to pesticide
loads by multiplying increments of flow by the respective concentration
values:
Y(HIF) storm event = SQ^i' a (5-2)
where Qj[ = increment of flow
Ci = C(HIF) of the ith increment of flow
a = 10 if dimensions of Q and C are liters and ppm
f\ ^
a = 62 x 10 if dimensions of Q and C are feet and ppm
Units of Y(HIF): kilograms (Ib)
* Base flow (nonstorm event) stream data on flows and concentrations
will not suffice. Many pesticides decompose in water and may be-
come trapped in bottom sediments. Concentrations under base flow
conditions do not accurately reflect storm event loads.
118
-------
The storm event load can be distributed back to the land by several op-
tions; for example:
Uniformly over the watershed.
Nonuniformly to broad categories of sources, e.g., row crops.
Specifically to identified or suspect sources, in proportion to
source size.
It will be necessary to sum storm events for the season, perhaps for the
year, to obtain annual loads. The 30 day maximum loadings fall naturally
out of cumulative storm event loads.
This procedure has several limitations and disadvantages. Extensive use
will be costly, and limited use will not suffice to adequately describe
large areas. An appropriate use is as follows: with selective runoff
measurement and analysis it will be possible to develop the relatively
modest inventory of data and experience needed to estimate pesticide
loads for sensitive areas, e.g., an intensive agricultural area which
depends heavily on herbicides and insecticides, and has a relatively
stable and predictable pattern of use. Combination of accumulated in-
formation on pesticide use patterns with representative measured con-
centrations and loads of pesticides will more than adequately serve as
a predictive "loading function." Since many of the persistent pesti-
cides are being phased out, the peak loads which occur in storms which
follow pesticide application are increasingly important. This basic
approach will, if properly used, deal with this problem adequately.
5.3 GENERAL INFORMATION
5.3.1 Pesticide Solubility
The dividing line between solubility and insolubility is diffuse and is
affected by factors such as the presence of other constituents in the
solution phase, pH, soil acidity, and organic matter in soil. Solubility
denotes, for purposes of the handbook, relatively little to moderate re-
sistance to leaching, and insolubility denotes moderate to high resistance
to leaching. Limited solubility data and indices of leachability are
presented in Appendix G, Table G-2. A pesticide with a leaching index
of one or two is treated as "insoluble." An index of three or four is
treated as "soluble."
5.3.2 Pesticide Persistence
General information on persistence is presented in Appendix G. Particu-
larly relevant to load calculation are the data which, though only semi-
quantitative, permit estimation of rates of disappearance in soils. Resi-
dues, concentrations and percent losses of selected pesticides are compiled
from recent literature and presented in Table G-3.
119
-------
5.4 LOAD CALCULATION: EXAMPLES
Case 1 Method
Conditions: Refer to Section 3.0, entitled "Sediment Loading Function."
Dieldrin
Continuous corn
A = 73 ha
Y(S)g (30-day maximum) = 117 kg/ha/day
Y(S) (annual average) = 36 kg/ha/day
C(I) range, 0.01 to 0.58 ppm from Appendix F, Table F-3
Probable minimum load, annual average
Y(I) = 36-0.01-73 x 10-6 = 26 x 10-6 kg/day
Probable maximum load, annual average
Y(I) = 36-0.58-73 x 10-6 = 1,524 x 10~6 kg/day
Case 2 Method
Conditions: Refer to above example.
2,4-D
Application rate: 5 kg/ha
Application date: 15 June
Persistence: 4 weeks (Appendix G)
Residue zero at season start
Y(S)E = 117 kg/ha/day, for 1-month period, 15 June to 15 July
Calculations
Initial concentration in erodible soil layer (5 cm), about 5 ppm
Average concentration, estimated from persistence information
equals 2 to 3 ppm for 15 June to 15 July period
Y(H) = 117 x 2.5 x 73 x 10'6 = 0.0214 kg/day
Y(H) (30-day maximum) = 0.0214 kg/day
5.5 LIMITATIONS IN USE
As stated earlier, pesticide behavior in the environment is both complex
and variable, and the accuracy of estimation reflects these complexities.
120
-------
The National Pesticide Monitoring Program, which serves as the basis for
Case I, contains data which generally indicate levels of pesticides in
soils throughout the country, and the frequency at which pesticides are
observed is an indication of the intensity of the use pattern. The data
and the Case I method should, however, be used only to derive an estimate
of loads over very large areas, and the results should be presented with
two qualifications: (a) that peak loads for nonpersistent pesticides are
apt to be overlooked by the method; and (b) that pesticides which leach
readily into the soil (and thus may contaminate subsurface waters) will
not be accounted for. Examination of the range of values reported in the
NPMP system reveals the fact that loads calculated from that data base
may differ substantially from actual loads, especially if one wishes to
apply calculated loads to a specific small area.
The Case II method depends upon area-specific and pesticide-specific data,
and thus will calculate loads considerably closer to actual values than
Case I. Since the data requirement is fairly extensive, its use is prob-
ably restricted to a small region--several counties perhaps—in which
pesticide use is uniform and other parameters are also relatively uniform.
The Case II method will, with care in use, be somewhat sensitive to peak
loads, i.e., when it rains soon after pesticide application.
The Case III method can be accurate with care in use. As indicated in
Section 5.2, the approach is probably best used to develop data and ex-
perience at local or regional levels, so that pesticide loads can be
estimated with confidence but not necessarily with a high degree of ac-
curacy.
Estimates of accuracy expected for Cases I through III are presented in
Table 5-1.
121
-------
Table 5-1. ESTIMATES OF ACCURACY FOR PESTICIDES
Annual average
(g/ha/year)
Probable
Estimated range
Storm event
(g/ha/day)
Probable
Estimated range
Case 1 method
(insoluble pesticide)
1-10
0.001-100
Not applicable
Case 2 method 1
(insoluble pesticide) 20
Case 3 method 1
(soluble and insoluble 20
pesticides)
0.01-10
5-50
0.1-5
10-50
1
20
1
20
0.1-10
5-50
0.1-5
10-50
122
-------
References
1. Holton, H. N., and N. C. Yokes, "USDA HL-73 Revised Model of Water-
shed Hydrology," Plant Physiology Report No. 1, ARS-USDA (1973).
2. Crawford, N. H., and R. K. Linsley, "Stanford Watershed Model IV,"
Stanford University, Stanford, California, Technical Report No.
39 (1966).
3. Crawford, N. H., and A. S. Donigian, Jr., "Pesticide Transport and
Runoff Model for Agricultural Lands," Office of Research and
Development, U.S. Environmental Protection Agency, EPA-660/2-74-
013 (December 1973).
4. Frere, M. H., C. A. Onstad, and H. N. Holtan, "ACTMO, an Agricultural
Chemical Transport Model," ARS-H-3, ARS-USDA, June 1975.
123
-------
Bibliography
Bailey, G. W., and J. L. White, "Review of Adsorption and Desorption of
Organic Pesticides by Soil Colloids with Implications Concerning Pes-
ticide Bioactivity," J. Agr. Food Chem., 12_ (1964).
Edwards, C. A., "insecticide Residues in Soils," Res. Reviews, JL3 (1966).
Frissel, M. J8, "The Adsorption of Organic Compounds, Especially Herbi-
cides on Clay Minerals," Verslag, Landbouwk, 7j5_ (1961) „
Getzin, L. W., "The Effect of Soils Upon the Efficiency of Systemic In-
secticides with Special Reference to Thimet," Dissertation Abstract,
19. (1958).
Hamaker, J« W., Mathematicl Prediction of Cumulative Levels of Pesticides
in Soils, Advances in Chemistry 60-Organic Pesticides in the Environ-
ment, American Chemical Society, Washington, D.C«
Kiigemagi, U», "Biological and Chemical Studies on the Decline of Soil
Insecticides," J. EC on. Entomol., 51^ (1958).
Lichtenstein, Es P., and K. R. Schulz, "Insecticide Residues Colorimetric
Determinations of Heptachlor in Soils and Some Crops," J. Agr. Food
Chem., 12. (1964).
Nauman, K., Einfluss von Pflanzenschutzmittel auf die Bodenmikroflora,
Hit. Biol. Bund., anst. Berlin, 9^ (1959).
"Production, Distribution, Use and Environmental Impact Potential of
Selected Pesticides," Final Report by Midwest Research Institute,
Kansas City, Missouri and RvR Consultants, Shawnee Mission, Kansas,
March 1974.
U.S. Department of Agriculture, "Quantities of Pesticides Used by
Farmers in 1971," Economic Research Service, Washington, D.C., in
press (1974).
124
-------
SECTION 6.0
SALINITY IN IRRIGATION RETURN FLOW
6.1 INTRODUCTION
The accurate prediction of salinity emissions in irrigation return flows
requires detailed knowledge of the particular system being studied. Prac-
tice has shown that salinity in irrigation return flows varies widely in
differing regions of the country because of the specific natures of the
soils, underlying geological formations, regional topography, and irriga-
tion practices. As a result, a simple "loading function" applicable to
all irrigation cases has no validity under present state of the art. A
discussion of the data needs for irrigation return flow salinity models
pointing out this fact has been prepared by the Environmental Protection
Agency.i'
For purposes of making assessments of salinity from irrigation return
flow, three optional methods are suggested in this section. The user is
cautioned, however, that the methods are not universally applicable and
hence may yield estimates that are not accurate. The most accurate pre-
diction method remains long-term monitoring of the particular irrigation
area to quantify actual salinity outputs in irrigation return flow.
The three procedures presented here for estimating salinity in irrigation
return flow are:
Option I - Source to Stream Approach: The first option involves the es-
timation of irrigation water percolating into groundwater. The irriga-
tion water acts as a hydraulic "head" on the groundwater, which pushes
the groundwater into surface waters as subsurface return. This approach
is valid for only a few areas of the country when valid relationships
between applied water and return flows exist. Furthermore, this option
should not be used in cases of spray irrigation where evaporative losses
associated with the applied water are significant. This option is most
valid in those cases where the total dissolved solids in groundwater
125
-------
contributing to return flow are very high (ca. 10 times) compared to total
dissolved solids concentrations in applied water.
Option II - Stream to Source Approach: The second option involves a back-
estimation procedure for salinity discharges in irrigation return flow.
Salinity measurements taken at sampling points above and below irrigated
areas will establish the amount of salt discharged in the area drained
by the stream between the two points. This salt load, however, includes
that discharged from background, salt springs, and point sources, as well
as that discharged from irrigation return flow. This method requires a
good definition of salinity sources other than irrigation return flow,
particularly that of background. This method is the one which has been
most widely used by others, especially where the total salinity loads are
measured at the discharge points of drainage basins.
Option III - Loading Values for Salinity in Irrigation Return Flow: A
third method for estimating salinity loads in irrigation return flows is
the use of loading values established for given areas through reduction
of stream monitoring data. A list of such loading values for areas in
the Colorado River basin are presented. These values are applicable only
to the particular region and should not be used except where indicated.
6.2 OPTION I: SOURCE TO STREAM APPROACH
6.2.1 Load Estimation Equation and Information Needs
An equation to estimate salinity in irrigation return flow has been form-
ulated based upon data reported by Skogerboe et al._' The equation is:
Y(TDS)IRF = a-A-C(TDS)GW-[lRR + Pr - Cu] (6-1)
where Y(TDS)jTvp = salinity load in irrigation return flow, kg/day (lb/
day)
A - area under irrigation, ha (acre)
IRR = volume of water added to crop root zone annually for
irrigation, cm (in.)
Pr = annual precipitation, cm (in.)
CU = annual consumptive use of water in growing crops, cm
(in.)
126
-------
= average concentration of total dissolved solids in
groundwater contributing to subsurface return, ppm
a = conversion factor to obtain proper units of load. If
Y is in kg/day, a = 2.7 x 10~4; if Y is in Ib/day,
a = 6.2 x 1CT4.
The volume of water applied to the crop root zone, IRR , can be deter-
mined by subtracting the volume of tailwater from the total water de-
livered to the irrigation site. This information would be available
from local irrigation districts.
Annual precipitation, Pr , is available from local weather data. Aver-
age annual precipitation can be used for purposes of estimating gross
salinity loads.
The CU factor, consumptive use, can be estimated by standard formulae
such as Jensen-Haise Method or the Blaney-Criddle Method. The Jensen-
Haise Method for estimating consumptive use is described in detail in
p /
the Skogerboe et al. report on irrigation scheduling.—' Information
needed for the Blaney-Criddle consumptive use formula can be found in
Todd's Water Encyclopedia.3/
The key data needed in the irrigation return flow loading function are
the groundwater total dissolved solids concentrations, C(TDS)Q^J. These
values must represent groundwater which maintains perennial strearaflow.
In general, the quality of water in perched water tables is the proper
information. For large irrigation areas, one should use an average
groundwater TDS value obtained from several observation wells.
The user is cautioned to avoid using the Option I method for cases involv-
ing sprinkler irrigation methods. This method does not account for
evaporation losses during application. If valid information is available
concerning evaporation losses, it should be incorporated into the esti-
mation procedure. Evaporation basically will cause an increase in the
TDS of applied water which will show up as increased TDS in the ground-
water contributing to return flow.
6.2.2 Load Calculation - Irrigation Return Flow
Load calculation involves three basic steps:
1. Obtain necessary information for Eqs. (6-1), (6-2), or (6-3) from
sources identified above.
127
-------
2. Substitute values into appropriate equations (Eqs. (6-1), (6-2), or
(6-3)).
3. Compute loads.
The Option I loading value equation (Eq. (6-1)) has been used to esti-
mate loads which can be compared directly to those reported by Skogerboe
I/
Skogerboe.
et al.—' Data used as inputs to the equation were those measured by
Data inputs for Eq. (6-1) are tabulated in Table 6-1, to-
gether with calculated loads. These are compared with reported loads.
Table 6-1. COMPARISON OF SALINITY LOADS OBTAINED WITH OPTION I LOAD
ESTIMATION EQUATION WITH REPORTED SALINITY LOADS2-/ IN THE
GRAND VALLEY, COLORADO
(Essential information: a = 6.2 x 10"^; C(TDS)GW = 6,700 ppm)
Equation
factors Plot No. 1 Plot No. 2 Plot No. 3 Plot No. 4 Plot No. 5
A (acre)
IRR (in.)
Pr (in.)
CU (in.)
Calculated
load
(Ib/day)
Reported
load
(Ib/day)
8.5
31.4
1.0
26.9
194
379
8.
23,
4.1
19.
293
344
25.7
42.1
1.2
33.5
1,046
15.0
29.1
2.7
20.7
692
1,291
521
10
24
3.3
17.
484
545
As can be seen from the comparison, the calculated loads compare reason-
ably with reported loads in four out of the five cases. One reason for
discrepancies between the calculated and reported values may be that the
equation disregards changes in soil moisture storage during the year.
In general, the changes in soil moisture storage which occur during and
between irrigation events should add to zero over an annual period, and
hence would have little effect on annual irrigation return flow volume.
Some irrigation water applied to the crop root zone is retained as soil
moisture, and hence does not show up as either consumptive use or irriga-
tion return flow. Soil moisture storage is an information input which is
not readily accessible.
128
-------
The Option I loading value equation should be considered only as a first
approximation method for estimating salinity in irrigation return flow.
Its usefulness will depend primarily upon three factors: (1) the concen-
tration of total dissolved solids in shallow groundwater which is trans-
mitted to surface waters as subsurface return; (2) reliable estimates of
the volume of applied water, tailwater, and return flows; and (3) good
information pertaining to consumptive losses in the complete irrigation
system. If these data are deemed insufficient, one should estimate sa-
linity in irrigation return by other procedures.
6.3 OPTION II: STREAM TO SOURCE APPROACH
6.3.1 Loading Equation and Information Needs
A second method for estimating salinity loads from irrigation return
flow involves the stream to source approach. In this option, salinity
loads in streams are determined above and below areas of irrigation.
Differences in salinity loads represent total salt being discharged by
the area by background and point sources, as well as irrigation return
flow. Therefore, salt loadings from irrigation return flow are deter-
mined by subtracting out contributions from background and from point
sources.
The Option II loading value equation is:
Y(TDS)1RF = a-[Q(str)B-C(TDS)B - Q(str)A«C(TDS)A] - Y(TDS)BG - Y(TDS)pT (6-4)
where Y(TDS)jgj. = yield of salinity in irrigation return flow, kg/day
(Ib/day)
Y(TDS)BG = salinity load contribution of background, kg/day (lb/
day)
Y(TDS)p,j, = salinity load contribution of point sources, kg/day
(Ib/day)
Q(str)g = streamflow of surface water be low irrigated areas,
liters/sec (cfs)
Q(str)^ = streamflow of surface waters above irrigated areas,
liters/sec (cfs)
C(TDS)g = concentration of total dissolved solids in stream
below irrigated area, ppm
129
-------
C(TDS) = concentration of total dissolved solids in stream
above irrigated areas, ppm
a = conversion constant needed to obtain proper units of
load. If flow units are liters/sec, a = 0.0864
(metric system, kg/day). If flow units are cfs,
a = 5.39 (English system, Ib/day).
Flow and concentration data obtained above and below irrigated areas can
be obtained from U.S. Geological Survey records of the region, or in some
cases from local water quality monitoring data. The use of these data in
the loading value equation will indicate total salt added to surface waters
between two points.
The salt load from point sources in the area under consideration can be
determined using information supplied by persons responsible for the point
sources. Point source contributions may be estimated from data contained
in discharge permit applications available from state and local pollution
control agencies, and from regional Environmental Protection Agency offices.
The total dissolved solids from the individual point sources in the area
are summed to yield total point source contributions.
The most difficult piece of information to be obtained is quantities of
salt discharged from background. In many cases, particularly in the arid
and semiarid regions where irrigation is intensive, this estimation can
only be accomplished by knowledge of the characteristics of the particular
area.
This estimation relies upon the judicious use of information concerning
background in a particular region. The use of broad general definitions
of background such as those presented in Section 12.0 of this handbook is
not recommended for the Option II method for salinity in irrigation return
flow. An estimation of background TDS levels may be made using the U.S.
Geological Survey's Hydrologic Investigations Atlas, HA-61, Plate l.^y
This plate contains information concerning dissolved solids concentration
for surface waters throughout the conterminous United States. It does
not differentiate between point and nonpoint contributions to salinity,
nor does it account for cumulative effects of runoff from a wide variety
of sources into stream water. The use of this map is recommended as a
first approximation of background.
The equation needed to define background total dissolved solids load can
be formulated in two ways, depending upon the units of flow. If flow is
measured as annual average runoff, the equation is:
130
-------
Y(TDS)BG = a-A-Q(R)-C(TDS)BG (6-5)
where Y(TDS)BQ = salinity load from background, kg/day (Ib/day)
A = area under consideration, ha (acre)
Q(R) = flow, as annual average runoff, cm (in.)
C(TDS)BG = concentration of background total dissolved solids as
determined by local information, ppm
a = conversion constant to obtain proper units of load.
If load is kg/day, a = 2.7 x 10"^; if load is Ib/day,
a = 6.2 x 10-4.
If flow is measured as actual flow in liters per sec (cfs), the equation
for estimating salinity loads in background becomes:
Y(TDS)BG = a-C(TDS)BG-[Q(str)B - Q(str)A] (6-6)
where Q(str)g and Q(str)^ are the flows be low and above the irrigated
areas, respectively. If the load is kg/day, a = 0.0864; if the load is
Ib/day, a = 5.39. The concentration of total dissolved solids in back-
ground, C(TDS)gQ , is the same as defined previously.
After proper information has been obtained, it is substituted into the
correct background total dissolved solids equation (Eqs. (6-5) or (6-6)),
and background total dissolved solids load computed.
6.3.2 Option II Load Calculation
The step-by-step procedure presented below is used for Option II stream
to source load calculations.
1. Obtain needed flow and concentration information for points above and
below irrigated areas. In many cases, information obtained at the mouth
of a drainage basin containing irrigated agriculture is sufficient, thus
obviating the need for above stream data.
2. Estimate total salinity loads above and below irrigated areas using
proper flow and concentration data. The total salinity load from irri-
gated areas, including its nonirrigated land uses, is determined by sub-
tracting upstream load from downstream load, via Eq. (6-4).
131
-------
3. Obtain data pertaining to point source contribution and sum indi-
vidual point sources to obtain total point load.
4. Determine background total dissolved solids load using Eqs. (6-5)
or (6-6) and procedures outlined previously in this section.
5. Estimate salinity load from irrigation return flow by subtracting
values obtained in Steps 3 and 4 from the value obtained in Step 2.
Y(TDS)IRF = Step 2 - Step 3 - Step 4
= Y(total) - Y(background) - Y(point)
The Option II stream to source approach for estimating salinity loads
in irrigation return flow has been applied to several subbasins of the
Colorado River. Values generated by the Option II load estimation equa-
tion have been compared with values reported by the Environmental Pro-
tection Agency in Appendix A to their report concerning the "Mineral
Quality Problem in the Colorado River Basin." Results of the compari-
son are presented in Table 6-2.
Table 6-2. COMPARISON OF SALINITY LOADS ESTIMATED BY OPTION II
METHODS WITH THOSE REPORTED BY EPA§/
Flow at
basin
mouth
Basin (cfs)
C(TDS) at
basin
mouth
(ppm)
C(TDS)BG
estimate
(ppm)
Calculated
load using
Option II
(tons/day)
Reported
load
(tons/day)
Black Forkil/
Gunnison£'
Big Sandy
Whit&§/
663
3,100
140
901
495
558
2,190
472
200
200
1,300
300
527
2,990
336
217
481
3,100
200
20
a/ U.S. Environmental Protection Agency, Regions VIII and IX, "Natural
and Man-Made Conditions Affecting Mineral Quality," Appendix A of
EPA Report, The Mineral Quality Problem in the Colorado River
Basin (1971).
b_/ Reference a, Figure 20.
£/ Reference a, Figure 34.
d/ Reference a, Figure 18.
e/ Reference a, Figure 25.
132
-------
From the data in Table 6-2, it is seen that Option II tends to overpre-
dict salinity in irrigation return flow. The overprediction may be due
to conservative estimates of background contributions, or to emissions
from unknown natural point sources such as salt springs. The data in
Table 6-2 clearly point out the fact that background, particularly that
in arid or semiarid areas, needs to be carefully considered. For example,
the high background level in the Big Sandy Creek area is due to water
seepage from saline lake beds in the area. Such characteristics must be
known if the Option II approach is to yield valid results.
6.4 OPTION III: LOADING VALUES FOR SALINITY LOADS IN IRRIGATION RETURN
FLOW
Perhaps the most useful method of estimating salinity loads is through
loading values determined for particular regions. Lists of such values
are presented in Tables 6-3 through 6-7 for subbasins in the Colorado
River basin, and for irrigated regions in California.
Studies in the Twin Falls area and the Colorado River basin indicate that
the range of values for salt pickup from irrigated lands is roughly 1.3
to 22 MT/ha/year (0.5 to 8 tons/acre/year)..£' An average salt pickup rate
might be 5 MT/ha/year (2 tons/acre/year). On a per day basis, the range
becomes 3 to 50 kg/ha/day (3 to 44 Ib/acre/day), and the average becomes
12 kg/ha/day (11 Ib/acre/day).
6.5 ESTIMATED RANGE OF ACCURACY
The accuracy of the three optional procedures for estimating salinity
loads from irrigation return flow will be no better than the accuracy of
the input data. For this particular system, the quality of the input
data is likely to be quite variable. More often than not, the quality
of input data will be less than that desired by the user. In addition,
the estimation procedure mechanisms tend to compound errors inherent in
the input data.
With these factors taken into account, ranges of error for Options I and
II have been estimated. The Option III method--loading values--is the
most accurate method if proper input data are available. However, its
use requires loading values generated from on-site data, and such data
are most often not available.
Table 6-8 presents the estimated range of error for the Option I (Source
to Stream) procedure. The error is estimated for several ranges of areas
which emit an average annual load of either 1 or 10 MT/year.
133
-------
Table 6-3. SALT YIELDS FROM IRRIGATION IN GREEN RIVER SUBBASIN3./
Average salt yield
Area (tons/acre/yr) (kg/ha/day) (Ib/acre/day)
Green River above New Fork River 0.1 0.6 0.5
Big Sandy Creek 5.6 34.3 30.7
Blacks Fork in Lyman area 2.4 14.7 13.2
Hams Fork 0.3 1.8 1.6
Henry's Fork 4.9 30.1 26.9
Yampa River above Steamboat Springs 0.2 1.2 1.1
Yampa River, Steamboat Springs to Craig 0.4 2.5 2.2
Milk Creek 1.0 6.1 5.4
Williams Fork River 0.3 1.8 1.6
Little Sanke above Dixon 0.3 1.8 1.6
Little Sanke, Dixon to Baggs 0.5 3.1 2.7
Ashley Creek 4.2 25.8 23.0
Duchesne River 3.0 18.4 16.4
White River below Meeker 2.0 12.3 11.0
Price River 8.5 52.2 46.6
San Rafael River 2.9 17.8 15.9
aj U.S. Environmental Protection Agency, Regions VIII and IX, "Natural and Man-
Made Conditions Affecting Mineral Quality," Appendix A of EPA Report, The
Mineral Quality Problem in the Colorado River Basin (1971).
Table 6-4. SALT YIELDS FROM IRRIGATION IN UPPER COLORADO MAIN STEM SUBBASIN§/
Average salt yield
Area (tons/acre/yr) (kg/ha/day) (Ib/acre/day)
Main stem above Hot Sulphur Springs 0.3 1.8 1.6
Main stem, Hot Sulphur Springs to 0.9 5.5 4.9
Kremmling
Muddy Creek Drainage Area 2.4 14.7 13.2
Brush Creek 0.7 4.3 3.8
Roaring Fork River 3.5 21.5 19.2
Colorado River Valley, Glenwood Springs 2.3 14.1 12.6
to Silt
Colorado River, Silt to Cameo 3.5 21.5 19.2
Grand Valley 8.0 49.1 43.8
Plateau Creek 0.9 5.5 4.9
Gunnison River above Gunnison 0.3 1.8 1.6
Tomichi Creek above Parlin 0.3 1.8 1.6
Tomichi Creek, Parlin to mouth 0.3 1.8 1.6
Uncompahgre above Dallas Creek 4.5 27.6 24.7
Lower Gunnison 6.7 41.1 36.7
Naturita Creek near Norwood 2.8 17.2 15.3
aj U.S. Environmental Protection Agency, Regions VIII and IX, "Natural and Man-
Made Conditions Affecting Mineral Quality," Appendix A of EPA Report, The
Mineral Quality Problem in the Colorado River Basin (1971).
134
-------
Table 6-5. SALT YIELDS FROM IRRIGATION IN SAN JUAN RIVER SUBBASIN^/
Average salt yield
Area
Fremont River above Torrey, Utah
Fremont River, Torrey to
Hanksville, Utah
Muddy Creek above Hanksville, Utah
San Juan above Carracas
Florida, Los Pinos, Animas drainage
Lower Animas Basin
LaPlata River in Colorado
LaPlate River in New Mexico
(tons/acre/yr)
0.4
5.8
3.1
2.7
0.2
3.5
1.4
0.3
(kg/ha/day)
2.5
35.6
19.0
16.6
1.2
21.5
8.6
1.8
(Ib/acre/day)
2.2
31.8
17.0
14.8
1.1
19.2
7.7
1.6
£/ U.S. Environmental Protection Agency, Regions VIII and IX, "Natural and Man-
Made Conditions Affecting Mineral Quality," Appendix A of EPA Report, The
Mineral Quality Problem in the Colorado River Basin (1971).
Table 6-6. SALT YIELDS FROM IRRIGATION IN LOWER COLORADO RIVER BASIN^-/
Average salt yield
Area
Virgin River
Colorado River Indian Reservation
Palo Verde Irrigation District
Below Imperial Dam
(Gila and Yuma projects)
(tons/acre/yr)
2.3
0.5
2.1
variable
(kg/ha/day)
14.1
3.1
12.9
-
(Ib/acre/day)^
12.6
2.7
11.5
_
a/ U.S. Environmental Protection Agency, Regions VIII and IX, "Natural and Man-
Made Conditions Affecting Mineral Quality," Appendix A of EPA Report, The
Mineral Quality Problem in the Colorado River Basin (1971).
135
-------
Table 6-7. SALT YIELDS FROM IRRIGATION FOR SELECTED
AREAS IN CALIFORNIA^/
Average salt yield
Area
North coastal
Central coastal
Sacramento
Delta-Central Sierra
San Joaquin
Tulare
Colorado Desert
(tons/acre/year)
0.353
0.808
0.707
0.974
0.827
0.768
10.9
(kg/ha/day)
2.2
5.0
4.3
6.0
5.1
4.7
67
(lb/ acre /day)
1.9
4.4
3.9
5.3
4.5
4.2
60
sj California Regional Framework Study Committee for Pacific Southwest
Inter-Agency Committee, Water Resources Council, "Comprehensive
Framework Study, California Region, Appendix XV, Water Quality,
Pollution, and Health Factors," June 1971.
Table 6-8. ESTIMATED RANGE OF ACCURACY FOR OPTION I (SOURCE TO STREAM)
PROCEDURE FOR ESTIMATING SALINITY FROM IRRIGATION RETURN FLOW
Area
considered
(ha)
< 100
100 - 1,000
1,000 - 10,000
> 100,000
Calculated
load
(MT/ha/year)
1
10
1
10
1
10
1
10
Probable range
of loads
(MT/ha/year)
0.7 - 1.5
8-13
0.5 - 3
6-15
0.3 - 5
4-20
0.1 - 10
2-25
136
-------
As can be seen from the table, the Option I procedure is deemed most ac-
curate when used for small areas, and when larger loads are calculated.
This aspect of accuracy arises because the Option I function is totally
dependent upon local conditions such as total dissolved solids in ground-
water, water consumptive use variations from crop to crop, irrigation
water supplied to specific fields, and variation of deep percolation
losses. If any of these data are extrapolated to larger areas, the vari-
ations in input data become wider, and hence the procedure becomes less
accurate for large areas.
In principle, error in the Option I method can be minimized for large
areas by summing up the values obtained for small areas. However, it
is questionable whether such a summation would yield calculated values
with any higher accuracy than those obtained using the Option II method,,
Estimated ranges of error for the Option II (Stream to Source) procedure
are given in Table 6-9. When Option II is used, the most accurate loads
will be calculated when large areas are considered. The ranges shown in
Table 6-9 assume that background salinity loads have been carefully con-
sidered. Since these background loads are the most uncertain component
of the procedure, the breadth of the error range is determined by this
uncertainty.
Table 6-9. ESTIMATED RANGE OF ACCURACY FOR OPTION II (STREAM TO SOURCE)
PROCEDURE FOR ESTIMATING SALINITY FROM IRRIGATION RETURN FLOW
Area
considered
(ha)
< 1,000
1,000 - 10,000
> 100,000
Calculated
load
(kg/ha/day)
1
10
1
10
1
10
Probable range
of loads
(kg/ha/day)
0.2 - 5
4-30
0.5 - 3
6-20
0.8 - 1.5
8-13
137
-------
The Option II method is deemed to be less accurate when small areas are
considered. The decrease in accuracy for small areas is inherent due to
uncertainty in flow measurements as well as uncertainty in background.
In general, small areas are associated with small streams draining the
area. The amount of variation for small streams is usually quite high
(and more unpredictable) than that of large streams.
No estimate of error has been given for the Option III procedure for
estimating salinity loads from irrigation return flow. The accuracy
of this option—use of salinity loading values—depends chiefly on the
trouble taken by the user to characterize his region and develop site-
specific information on his loadings. This option can be the most ac-
curate of the three discussed, provided that the values used are accu-
rate.
The availability of accurate loading values for the Option III approach
is quite limited. Accurate values can be obtained through long term
monitoring and analysis of irrigated areas, an expensive and time con-
suming operation. However, various mathematical methods for predicting
salinity in irrigation return flow are being developed. These models
will tend to describe the complicated relationships between the water
used for irrigation and the land being irrigated which result in salin-
ity emissions. It may be that at some future time, these models will
be sufficiently validated so that their outputs can produce loading
values for use in the Option III procedure. The user of this handbook
is encouraged to keep abreast of these modeling projects so that their
output can be used to obtain accurate estimates of salinity from irri-
gation return flow.
138
-------
REFERENCES
1. Hornsby, A. G. , "Prediction Modeling for Salinity Control in Irriga-
tion Return Flow," U.S. Environmental Protection Agency, Report
No. EPA-R2-73-168, March 1973.
2. Skogerboe, G. V., W. R. Walker, J. H. Taylor, and R. S. Bennett,
"Evaluation of Irrigation Scheduling for Salinity Control in Grand
Valley," Grant No. S-800278, U.S. Environmental Protection Agency,
Report No. EPA-660/7-74-052, June 1974.
3. Todd, D. K., The Water Encyclopedia, pp. 101-108, Water Information
Center, Port Washington, New York (1970).
4. Rainwater, F. H., "Stream Composition of the Conterminous United
States," U.S. Geological Survey, Hydrologic Investigations Atlas,
HA-61, Washington, D.C. (1962).
5. Skogerboe, G. V., and J. P. Law, Jr., "Research Needs for Irrigation
Return Flow Quality Control," U.S. Environmental Protection Agency,
Report No. 13030-11/71, November 1971.
139
-------
SECTION 7.0
ACID MINE DRAINAGE
7.1 INTRODUCTION
The emission of acid mine drainage arises from land disturbances created
by coal and metals mining activities. The mine drainage arises because
of atmospheric and hydrologic actions on pyritic materials associated
with the mined materials. The pyritic materials may be in residues left
behind at the mined-out site, or in residues produced by coal processing
or mineral beneficiation. If pyrites (or other sulfurous materials) are
not associated with a particular mined product, e.g., quarrying, sand and
gravel operations, etc., then acid mine drainage will not occur. Thus, the
presence or absence of pyritic materials is the determining factor for
nonpoint emissions of mine drainage.
Mine drainage can arise from active and inactive mines and from under-
ground and surface activities. In addition, mine drainage can arise from
processing wastes, e.g., tailings piles and gob piles. In considering
nonpoint emissions from these latter sources, processing wastes disposed
of on the land surface are considered as surface mines.
Basically, regional mine drainage problems arise because of an assemblage
of individual sources in an area. A procedure for estimating mine drain-
age loads based upon the statistical distribution of individual sources
is presented here as Option I. The procedure was developed using data
gathered by Environmental Quality Systems, Inc., in a study dealing with
estimation of mine drainage emissions in the Monongahela River Basin,—'
and from data obtained for the Appalachian Regional Commission^' for their
140
-------
report concerning mine drainage in Appalachia.—' This procedure is funda-
mentally a source to stream loading function. On the other hand, sulfate
analysis of surface waters are key indicators of nonpoint emissions of
mine drainage, since sulfate is the end product of atmospheric/hydrologic
reactions with pyrite. Thus, an Option II estimation procedure is pre-
sented which uses the stream to source approach and is based on sulfate
concentrations in surface waters. A brief description of these two op-
tions follow.
Option I - Source to Stream Approach: A loading estimation procedure is
presented which relates the number of total sources in an area, the dis-
tribution of these sources among four categories (active underground,
active surface, inactive underground, and inactive surface), and neutral-
ization of acidic products of pyrite weathering with background alkalinity.
This approach is particularly useful for heavily mined areas of the coun-
try, such as the coal mining regions of Appalachia. In other areas where
mining is less concentrated, this statistical approach may not be adequate.
Option II - Stream to Source Approach: The second option involves compar-
ing sulfate loadings found in surface waters with sulfate loadings ex-
pected from natural background. Increases in sulfate loading as surface
waters move through an area over the background contribution can be at-
tributed to nonpoint emissions of mine drainage in the area. This second
approach should be considered when detailed information about the number
of sources is unknown, where mining density is low, or when streamflow
data are' deemed more appropriate to use.
7.2 OPTION I: SOURCE TO STREAM APPROACH
7.2.1 Loading Function and Information Needs
The loading function for the Option I approach contains three fundamental
elements: the number of potential sources of mine drainage; the amount
of raw acidity formed from the sources; and the neutralization capacity
of the background. The second element—amount of raw drainage formed--
involves the statistical distribution term to account for the widely var-
iable source-to-source loads arising from the individual sources. The
loading function is:
Y(AMD) = N[Ka-(IAU + IAS + Ij-u + IIS) - Kb'Q(R)-C(Alk)BG] (7-1)
where N = total number of sources which are potential emitters of
acid mine drainage
141
-------
Statistical Distribution Term
Ka = constant representing the raw acid load generated by
the "typical" site. A range of values for Ka is
presented in Table 7-1, and discussed in Section
7.2.2.
IA.U> ''-AS' -""IIP
I-rc = load index values for the number of Active Underground
sources, Active Surface sources, Inactive Underground
sources, and Inactive Surface sources. The load in-
dex values are presented in Table 7-2, and discussed
in Section 7.2.3.
Background Neutralization Term
Kjj = constant representing the neutralization capacity of
background alkalinity for raw acid produced at the
"typical" site. A range of values for K, is pre-
sented in Table 7-1, and discussed in Section 7.2.2.
Q(R) = flow as annual average runoff in the area, cm/year
(in/year)
C(Alk)BG = concentration of background alkalinity in the area,
ppm as CaC03. C(Alk)RG can be determined through
use of an isoalkalinity map (see Figure 7-1, Section
7.2.4).
7.2.2 Constants Ka and Kb in Option I Loading Function
Description of the acid mine drainage discharge from the "typical" source
was determined by subjecting a number of mine drainage data obtained in
the Monongahela River Basin!' to regression analysis. These data repre-
sented the acid load discharged at specific sites from about 7,000 poten-
tial sites. The regression analysis indicated that the distribution of
mine drainage quantities from the 7,000 sources could be well fit (index
of determination = 0.998) to a hyperbolic function dependent upon (a) the
number of sources, (b) the quantity of mine drainage from the largest
source, and (c) the cumulative amount of mine drainage emitted from all
sources. The regression equation has the form:
lim A'N = 1 (7-2)
N—-»» "
.f^B. + A'N
1=1 i
142
-------
Table 7-1. VALUES OF Ka AND Kb FOR ACID MINE
DRAINAGE OPTION I LOADING FUNCTION
Units of
load
Metric kg/day
English Ib/day
Value of
K
130
280
Range of
a
110-150
250-320
Value of
Kb
0.15
0.62
Range of
Kb
0.10-0.20
0.35-0.75
Table 7-2. LOAD INDEX VALUES FOR ACTIVE AND INACTIVE
SURFACE AND UNDERGROUND MINES
Fraction
of mines
in category
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
Active
underground
0.00
0.33
0.50
0.60
0.67
0.71
0.75
0.78
0.80
0.82
0.83
0.85
0.86
0.87
0.88
0.88
0.89
0.89
0.90
0.90
0.91
Load
Active
surface
0.00
0.08
0.16
0.22
0.27
0.32
0.36
0.39
0.43
0.45
0.48
0.50
0.53
0.55
0.56
0.58
0.60
0.61
0.63
0.64
0.65
index
Inactive
underground
0.00
0.13
0.23
0.31
0.37
0.42
0.47
0.51
0.54
0.57
0.60
0.62
0.64
0.66
0.67
0.69
0.70
0.71
0.72
0.74
0.75
Inactive
surface
0.00
0.03
0.06
0.08
0.11
0.13
0.15
0.17
0.19
0.21
0.23
0.24
0.26
0.28
0.29
0.31
0.32
0.33
0.35
0.36
0.37
143
-------
where A = the quantity of mine drainage from the largest source
N
EB- = the cumulative amount of mine drainage from all sources
i=lx
N = the number of potential mine drainage sources in the
area
N
The ratio of £ Bj_ to N thus determines the acid load from the "typi-
1=1
cal" site. Furthermore, the equation implies that the load will be more
accurate when the number of sources considered becomes very large.
A part of the raw mine drainage generated within the mining area will
have been neutralized by background alkalinity before it is discharged
to surface waters. From the consideration of the background neutraliz-
ing capacity in the Monongahela River basin, it has been possible to
establish values of Ka and K^ for the loading function (7-1) based
upon the regression analysis represented by Eq. (7-2). These values are
presented in Table 7-1.
The value Ka represents the raw acid generated at the "typical" mine
site as determined by Eq. (7-2). The value K^ represents the neutral-
ization of part of the raw acid by background alkalinity in the area
directly affected by the "typical" site.
7.2.3 Load Index Factors for Option I Loading Function
The Ka values presented in Table 7-1 have been established based on
data pertaining to the Monongahela River basin. In order to use them
in other regions of the country, the Ka term must be corrected to re-
flect the distribution of potential mine drainage sources. This correc-
tion is accomplished through the use of "load index factor" determined
in the following manner:
The total number of sources are separated into four components: number
of active underground (AU), active surface (AS), inactive underground
(IU) and inactive surface (IS). The fraction of each source is deter-
mined for each category by dividing the number of sources in a certain
category by the total number of sources.
After the category fractions have been determined, a load index value is
found in Table 7-2 for each category. The first column of Table 7-1 in-
dicates the fraction of mine in each category; subsequent columns contain
the load index value for each category. This procedure is exemplified in
Table 7-3, using a hypothetical situation involving 1,800 mines.
144
-------
Table 7-3. EXAMPLE OF DETERMINATION OF LOAD INDEX VALUES
Number of Fraction of
sources sources Load index
Active underground 180 0.10 0.50
Active surface 450 0.25 0.32
Inactive underground 630 0.35 0.51
Inactive surface 540 0.30 0.15
Total 1,800 1.00 1.48
After fractions of mines have been determined in each category, the ap-
propriate load index value is established for each category by referring
to the appropriate column in Table 7-2. After the individual load index
values have been determined, they are added together to yield a total load
index value required for the loading function. The total is the factor
(IATJ + IAC + ITTJ + ^TC) ^n tne loading function.
The load index values have been established by proportionating the total
load and total number of sources (as determined by the regression analy-
sis results of Eq. (7-2)) into contributions from active underground,
active surface, inactive underground, and inactive surface sources in the
Monongahela River basin. The bases for the proportionment were obtained
from data in the 1969 Appalachian Regional Commission report concerning
mine drainage..?/ This exercise yielded a series of four equations defin-
ing load index values for each of the four types of mine drainage sources.
The equations from which the load index values in Table 7-2 were derived
are:
(7-3)
AU 0.10 + nAU
0-54
HAC
(7-4)
0.34 + nIO -
IlS * 1.7o" .IS (7-6)
where n^y , n^s , n.^ , and n-j-g are the fractions of mines in each of
the four categories.
145
-------
nAU + nAS + nIU + nIS = 1'° (7'>
The total number of sources in an area is determined by study of state
and local historical records. Basically, what is needed is the number
of active and inactive underground and surface sources. The total num-
ber of sources need not be an exact count; a reliable estimate is quite
satisfactory for use in the loading function.
Information concerning active sources can be found in the annual Minerals
Yearbook, published by the U.S. Bureau of Mines. An alternate source of
information about active sources will be state and local permit programs.
Uncontrolled waste piles associated with active mines should be counted
as active surface mines.
Information about inactive mines may be more difficult to obtain. Prob-
ably the best source of information on inactive mines will come through
analysis of historical records of the area. These records should be
available in state archives.
7.2.4 Background Alkalinity Term for Option I Loading Function
The Kb values presented in Table 7-1 have also been established from
Monongahela River basin data. These too must be corrected in order to
reflect changes in the neutralizing capacity of background. The correc-
tion factors involve alkalinity concentrations in background and average
annual runoff.
Background alkalinity concentrations are determined by locating mining
areas on the iso-alkalinity map (Figure 7-1), estimating concentration,
and using this value in the alkalinity term. If other values of back-
ground alkalinity concentrations are deemed more appropriate than those
shown on the map, then they should be used instead. In areas afflicted
with acid mine drainage emissions, one should be cautious about using
"unaffected" stream values of alkalinity. Although data may have been
generated in areas unaffected by mining activity, unknown sources of
mine drainage may be present which would lower background alkalinity es-
timates.
Average annual runoff can be estimated from standard runoff maps such as
that in the U.S. Geological Survey's National Atlas, Plates 118 and 119.
7.2.5 Procedure for Using Option 1 Loading Function
The procedure for putting together components of the source to stream
loading function to estimate levels is outlined below.
146
-------
o
o
tfl
u
6
0,
a.
w
c
o
•rl
c
QJ
O
C
o
o
4J
•i-l
C
•H
3
O
M
60
^J
O
rt
M
3
147
-------
1. Estimate total number of potential mine drainage sources through re-
view of state and local records, permits, etc., as indicated in Section
7.2.3.
2. Establish load index values for the following categories: active
underground, active surface, inactive underground, and inactive surface,
by procedures indicated in Section 7.2.3.
3. Sum up load index values established in Step 2.
4. Determine constant Ka from Table 7-1.
5. Multiply results generated in Steps 1, 3, and 4 to obtain value of
statistical distribution term of the loading function.
6. Determine average annual runoff in area from standard runoff maps,
e.g., U.S. Geological Survey's National Atlas, Plates 118 and 119.
7. Determine background alkalinity from iso-alkalinity map (Figure 7-1)
or from other data deemed to reflect background alkalinity concentrations
more adequately.
8. Determine proper constant K^ from Table 7-1.
9. Multiply values yielded by Steps 6, 7, and 8 to obtain background
alkalinity term.
10. Subtract value obtained in Step 9 from that obtained in Step 5.
11. Multiply value obtained in Step 10 by the total number of mine drain-
age sources established in Step 1. This final step will yield the load
of acid mine drainage being emitted from the mining region under consid-
eration.
7.2.6 Examples of Option I Loading Function Utilization
The mine drainage loading function has been used to estimate loads emitted
from two basins in Appalachia--West Branch Susquehanna, and Allegheny.
These examples are presented to indicate how the mine drainage loading
function can be used.
7.2.6.1 Case I: West Branch Susquehanna
Data Source - Federal Water Pollution Control Administration, Ohio Basin
Region, U.S. Department of the Interior, "Stream Pollution by Coal Mine
148
-------
Drainage in Appalachia,11 Attachment A to Appendix C of the Appalachian
Regional Commission Report, Acid Mine Drainage in Appalachia, Washington,
D.C. (1969).
Step 1. Number of mine sources N: 4,400
Number of draining sources: 967
Step 2. Load index values (Table 7-2):
Active underground: 19; 27» = 0.02
Active surface: 17; 2% =0.02
Inactive underground: 630; 65% =0.65
Inactive surface: 301; 31% = 0.31
Total: 967 100% = 1.00
Load indexes: I^j = 0.07
IAS = 0.02
Ijy = 0.66
IIS = 0.15
Step 3. Load index summation total: 0.90
Step 4. Constant K from Table 7-1: 280
SL
Step 5. Calculation of statistical distribution term: 280 x 0.9 = 252
Step 6. Average annual runoff Q(R): 20 in.
Step 7. Background alkalinity C(Alk)gQ (from Figure 7-1): 10 ppm
Step 8. Constant Kb (from Table 7-1): 0.62
Step 9. Calculation of background alkalinity term: 0.62 x 20 x 10 = 124
Step 10. Subtract alkalinity term from statistical distribution term:
252 - 124 = 128
Step 11. Compute acid mine drainage load: 4,400 x 128 = 560,000 Ib/day
Mine drainage (calculated) = 560,000 Ib/day
Mine drainage (reported) = 500,000 Ib/day
149
-------
7.2.6.2 Case II: Allegheny River Basin (1966)
Data Sources - Appalachian Regional Commission Report, Acid Mine Drainage
in Appalachia (1969); Tybout, R. A., "A Cost Benefit Analysis of Mine
Drainage," paper presented before 2nd Symposium on Coal Mine Drainage
Research, Pittsburgh, Pennsylvania, 14-15 May 1968; and U.S. Bureau of
Mines, Minerals Yearbook, 1966. Washington, D.C. (1967).
Step 1. Number of mine sources N (Tybout): 6,626
Step 2. Load index values (Table 7-2):
Active underground: 228; 3% =0.03
Active surface: 310; 5% = 0.05
Inactive underground: 2,350; 36% =0.36
Inactive surface: 3,738; 56% = 0.56
Total: 6,626 100% =1.00
Load indexes: IAU = 0.10
IAS = 0.08
IlU = 0.51
IIS = 0.24
Step 3. Load index summation total: 0.93
Step 4. Constant Ka from Table 7-1: 280
Step 5. Calculation of statistical distribution term: 280 x 0.93 = 260
Step 6. Average annual runoff Q: 20 in.
Step 7. Background alkalinity C(Alk)BG (from Figure 7-1): 10 ppm
Step 8. Constant Kb (from Table 7-3): 0.62
Step 9. Calculation of background alkalinity term: 0.62 x 20 x 10 = 124
Step 10. Subtract alkalinity term from statistical distribution term:
260 - 124 = 136
Step 11. Compute acid mine drainage load: 6,626 x 136 = 900,000 Ib/day
Mine drainage (calculated) = 900,000 Ib/day
Mine drainage (reported) = 866,000 Ib/day
150
-------
7.3 OPTION II: STREAM TO SOURCE APPROACH
7.3.1 Loading Function and Information Needs
Since acid mine drainage is basically a discharge of sulfuric acid (and
its reaction products), the presence of sulfate in stream water analyses
is often a good indicator of nonpoint emissions from mine drainage
sources. Thus, a comparison of sulfate levels detected in streams with
that expected from natural background will yield an estimate of nonpoint
emissions of mine drainage. The loading function can be expressed in
two ways, depending upon the units of flow.
Y(AMD) = a-A-Q(R)-[C(S04) - C(S04)BQ - C(S04)pT] (7-8)
Y(AMD) = a-Q(str)-[C(S04) - C(S04)BQ - C(S04)PT] (7-9)
where Y(AMD) = yield of acid mine drainage, kg CaC03/day (Ib CaCO«/day)
A = area containing mine drainage sources, ha (acre)
Q(R)j Q(str) = flow; Eq. (7-8) requires flow units Q(R) as annual
average runoff, in cm/year (in/year). Equation (7-9)
requires flow units Q(str) as streamflow in liters/
sec (cfs).
C(SO^) = concentration of sulfate in surface waters, ppm
= concentration of sulfate in surface waters attributable
to background, ppm
= concentration of sulfate in sources attributable to
point sources, ppm
a = conversion constant for obtaining proper load
The two key elements of this loading function are in the conversion fac-
tor a and in the concentration of background sulfate C(S04)BG. The
value of a to be used in the loading function is determined by the
units of flow. A table of a values is presented in Table 7-4. The
values take into account the conversion of sulfate concentrations (in
ppm) to their calcium carbonate equivalents (ppm as CaC03). This con-
version is necessary in order to obtain load units of kilograms CaCOo
per day (Ib CaC03/day) .
151
-------
Table 7-4. CONVERSION FACTORS a TO BE USED FOR
OPTION II MINE DRAINAGE LOADING FUNCTION
Sulfate
concentration
units
ppm
ppm
ppm
ppm
Units of
flow
Q
cm/year
in/year
liters/sec
cfs
Units of
area
A
ha
acre
Value of
a
2.8 x 1CT4
6.4 x 10"4
0.090
5.61
Units of
Y(AMD)
kg CaC03/day
Ib CaC03/day
kg CaC03/day
Ib CaC03/day
The background levels of sulfate can be estimated through the use of an
iso-sulfate background map presented in Figure 7-2. The region of in-
terest is identified on the map, and sulfate levels estimated through the
contours. If more specific data are available which are believed to de-
scribe background sulfate levels more adequately, then these data would
be preferred to the use of Figure 7-2.
Other components of the loading function are obtained through standard
sources. Sulfate concentration in streams, C(S04) , and streamflow,
Q(str) , can be obtained from U.S. Geological Survey studies and from
local water quality records. Annual average runoff can be estimated with
the U.S. Geological Survey Surface Runoff Map, Plates 118 and 119, in the
National Atlas. Sulfate contributions from point sources, C(SO^)p-T. ,
can be estimated from data contained in permit applications for point
source discharges.
7.3.2 Procedure for Using Option II Mine Drainage Loading Function
The step-by-step procedure for using the Option II loading function is
outlined below.
1. Obtain necessary water quality data, streamflow data, and areal data
from U.S. Geological Survey records, local records, or other similar
sources.
2. From these data establish appropriate values for A , Q(R) , and
C(S04) .
3. Determine value for background sulfate, C(SO^)BG , using Figure 7-2,
or from local water quality information thought to be more appropriate.
152
-------
0
CX
o,
w
d
o
•i-l
4J
n!
n
.u
C
o;
o
c
o
o
r-l
3
w
T)
d
3
o
M
bO
3
bO
•i-l
153
-------
4. Determine value of conversion constant a by means of Table 7-4.
5. Insert a , A , Q(R) , 6(804) and C(S04)BG values established
in the above steps into proper form of Option II loading function.
6. Compute mine drainage loads.
7.3.3 Example of Option II Loading Function for Mine Drainage
An example of how the Option II loading function can be used is summar-
ized in Table 7-5. This table contains results for the Tioga and Juniata
river basins in Appalachia obtained by the Option II loading function.
The Option II estimates are within 80 to 110% of the reported loads.
The loading function works out well in these cases because mine drainages
(and background) are the principal sources of sulfate in the area. If
the loading function is applied to more highly industralized areas, e.g.,
the Anthracite Region of Appalachia, it tends to overpredict the nonpoint
loads of mine drainage. In the industrial areas, point source discharges
of sulfate report as nonpoint discharges within the context of the Option
II method. Therefore, the Option II approach should be used mainly in
rural areas. If amounts of the point source contributions of sulfate are
known, however, they can be subtracted from estimates yielded by the Op-
tion II loading function. This procedure would ameliorate some of the
deficiencies of using the stream to source approach in populated or in-
dustrialized areas.
7.4 ESTIMATED RANGE OF ACCURACY
A series of estimated value ranges for several acid mine drainage loads
calculated using the Option I procedure are presented in Table 7-6. Two
ranges are presented—one for Appalachia, and one for regions other than
Appalachia. As can be seen by the ranges, the loading function is more
accurate when applied to coal mining in Appalachia than it is when used
in other parts of the country.
One major source of error in the Option I loading function lies in the
number of mine drainage sources in the area being considered with the
loading function. If not enough sources are available in an area, it
is likely that their distribution of loads will not meet that of the
"typical" mine from which the loading function was developed. This prob-
lem will most often be encountered in regions outside of Appalachia where
mining activity density (number of mines in the area being considered)
is small.
154