Movement of fertilizer and herbicide through irrigated sands

MISSING IMAGE

Material Information

Title:
Movement of fertilizer and herbicide through irrigated sands
Series Title:
Florida Water Resources Research Center Publication Number 38
Physical Description:
Book
Creator:
Mansell, R. S.
Rhoads, F. M.
Hammond, L. C.
Selim, H. M.
Wheeler, W. B.
Zelazny, L. W.
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Notes

Abstract:
The simultaneous movement of water and selected agrichemicals (fertilizer nutrients and herbicide) through sandy soils is of particular importance to the efficient use of fertilizers and irrigation water by agricultural crops. Efficient use of fertilizers and herbicides applied to Florida's sandy soils is desirable for maintaining optimum growth of plants and for minimizing groundwater contamination. Laboratory and field experiments as well as mathematical models were used to study water and solute (potassium and phosphorus nutrients and 2,4-D herbicide) transport in two representative Florida soils: Wauchula sand and Troup sand. In an irrigated and fertilized corn experiment, grain yields and efficiency of water use were observed to be mutually related to both the irrigation and the fertilizer application treatments. Leaching of applied nutrients and·irrigation water from the soil "rooting zone" resulted in decreased water use efficiency in these soils. Mathematical models were developed and used to stimulate transport and chemical-physical reactions for potassium, phosphorus and 2 ,4-D herbicide in these soils. Reactions such as adsorption -desorption, chemical precipitation and immobilization (fixed) greatly influenced the movement and thus potential leaching of these solutes through the soil.

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the source institution and holding location.
System ID:
AA00001512:00001


This item is only available as the following downloads:


Full Text






-.


Publication No. 38


Movement of Fertilizer and Herbicide Through
Irrigated Sands


By


R.S. Mansell, F.M. Rhoads, L.C. Hammond, H.M. Selim,
W.B. Wheeler and L. W. Zelazny


Department of Soil Science, IFAS
University of Florida
Gainesville


S.--






- ..~


7 7 *- '"- : ".


W114-
















TABLE OF CONTENTS


Title . . . . . .

Table of Contents .. . . . .

Acknowledgments . . . .

Abstract . . . . .

Chapter 1: Introduction . . ..

Chapter 2: Objectives . . .

Chapter 3: Field Experiments

A. Description of Soils . . .

B. Experiments on Wauchula sand, a Spodosol .

C. Experiments on Troup sand, an Ultisol .

Chapter 4: Mathematical Models . . .

A. Water and 2,4-D Models . .......

B. Phosphorus Models . . .

C. Potassium Models . . . .

Chapter 5: Summary and Conclusions . .

Literature Cited . . . ..

Appendix: Titles and abstracts of published papers
from this research . . . .


. . . .

. . . .

. . ... .


* 0 0 0 0 0 0 0 0 0


0 0 0 0 0 0 0 0


resulting
. . . .


Page

i

ii

iii

iv

1

2



3

3

22

29

29

30

33

42

44


46


. .


a a .


. .


. . .





iii




ACKNOWLEDG4EENTS




















Contributions by the following technical staff of the Soil Science
Department provided valuable assistance in performing this research:
7i:->!-i.d McCurdy, Bill Porthier, and Ron Jessup. We also acknowledge Chris
Y and Ann Barry for typing this report as well as manuscripts for ar-
ticles published in scientific journals. Special thanks are extended to
each of these individuals.








ABSTRACT












The simultaneous movement of water and selected agrichemicals (ferti-
lizer nutrients and herbicide) through sandy soils is of particular impor-
tance to the efficient use of fertilizers and irrigation water by agricul-
tural crops. Efficient use of fertilizers and herbicides applied to
Florida's sandy soils is desirable for maintaining optimum growth of plants
and for minimizing groundwater contamination.

Laboratory and field experiments as well as mathematical models were
used to study water and solute (potassium and phosphorus nutrients and 2,4-
D herbicide) transport in two representative Florida soils: Wauchula sand
and Troup sand. In an irrigated and fertilized corn experiment, grain
yields and efficiency of water use were observed to be mutually related to
both the irrigation and the fertilizer application treatments. Leaching of
applied nutrients and irrigation water from the soil "rooting zone" re-
sulted in decreased water use efficiency in these soils. Mathematical
models were developed and used to simulate transport and chemical-physical
reactions for potassium, phosphorus and 2,4-D herbicide in these soils.
Reactions such as adsorption-desorption, chemical precipitation and immo-
bilization (fixed) greatly influenced the movement and thus potential leach-
ing of these solutes through the soil.









Si"TER 1: T :i ,' ,'ION


Fertilizers, herbicides, and irrigation water are commonly applied to
Florida s sandy soils for the --... !- .-e of maintaining high levels of food
production from soil-water-plant systems. These soils generally have
limited capacities to temporarily store applied water and solutes, and thus
S-. .: ... ,:-: of .-.- .:..- and water applications is needed to mini-
mize 1 :. losses of nutrients and herbicides from the "rooting zone" of
the soil *. File. Excessive leaching, which may occur during periods of
intense rainfall greatly increases the potential for contamination of the
underlying groundwater with ..-ichemicals. Efficient usage of fertilizers,
herbicides, and water therefore provides two very important beneficial
results: (1) optimum plant growth and yields, and (2) minimal pollution of
grotmdwater resources.

Management of crops growing in these sands to produce optimum crop
yields requires irrigation during periods of drought and frequent applica-
tions of fertilizer during the growing season. Although the average annual
rainfall for the state ".,. --..:'; from near 52 inches on the central and
northern peninsula to nearly 65 inches in the panhandle west of Tallahassee
..." (Butson and Prine, 1968), severe drought commonly occurs during the
6._;, growi .L; season followed by heavy rains during the summer when the
state receives approximately 60% (Jones, 1967) of its rainfall. The uneven
rainfall di:7 -. ..-. coupled with low retention capacities of the surface
soil of the sands for water and solutes result in relatively high leaching
losses .r.f applied agricultural chemicals, Also crop yields may be de-
creased by periods of drought as a result from soil-water deficiency or
high osmotic pressures of the soil solution due to improper timing of ferti-
lization.

Water, fertilizers, and herbicides are applied to the sands to create
a favorable -*-.t root environment with optimum supplies of available water
and nutrients. The herb: -_,-. :: are used to destroy weeds and undesirable
grasses which compete with the crop plants for water, nutrients, and light.
Failure to maintain minimum threshold levels of water and fertilizer nu-
trients in the root zone during critical periods of plant growth can result
in decreased yields. However, levels of water and nutrients in the root
zone in excess of plant requirements may result in waste of resources,
possible detrimental effects on crop growth, and potential contamination
of ground water with herbicide or fertilizer solutes.

To insure o,:iT-,- crop growth on sands, irrigation during periods of
low rainfall is needed to maintain low soil-water suction (0 to -200 cm of
water), Consequently, the profile will of necessity be high in soil-water
content and undergo some drainage or redistribution of water at all times.
Since the hydraulic conductivity of the deep sands increases greatly with
soil-water content, heavy irrigation or rainfall imposed upon relatively
moist soil will result in much of the infiltrating water being lost from the
root zone by dr_ :.: .

The simultaneous transport of water and soluble chemicals through soil
has important implications with :'- Ig to the efficient management of
fertilizer, herbicide, and irrigation water applied to agricultural crops
.-'-....-' in sandy soils. Transport of water and nutrients through sandy
soils is -...-:.cularly i':..t-.' to phenomena such as plant uptake of nu-








trients and water which tends to increase fertilizer and water use effi-
ciencies. Thus, improper water management combined with inefficient appli-
cation of herbicides and fertilizer nutrients to sands may result in leach-
ing loss of some of the chemicals from the "rooting zone" of the soil pro-
file. Thus potential contamination of groundwater may result as leached
herbicides and nutrients move deeper into the soil. Fortunately,pro-
cesses such as adsorption-desorption, ionic exchange, microbiological
transformations, chemical interactions, and uptake by plant roots tend to
decrease the leaching loss of herbicides and nutrients from soils.

Therefore, detailed knowledge of simultaneous movement of water and
solutes in Florida's sandy soils is critically needed to insure efficient
use of water, fertilizer, and herbicide resources for crop production with-
out undesirable contamination of the underlying groundwater.



CHAPTER 2: OBJECTIVES

Specific objectives for this project were as follows:

(1) To determine rates of movement of soil-applied herbicide and
fertilizer (nutrients) solutes in irrigated agricultural sands and to deter-
mine leaching losses of these chemicals from the "rooting zone" of the soil
profile;

(2) To determine the influence of limestone application to acid, sandy
soils upon movement and leaching losses of soil-applied fertilizer solutes
from the "rooting zone;"

(3) To determine the influence of adsorption and desorption processes
upon rates of movement of 2,4-D herbicide, potassium fertilizer and ortho-
phosphate fertilizer with water through irrigated sandy soils; and

(4) To utilize existing mathematical models to describe the simul-
taneous movement of water, selected herbicides, and fertilizer nutrients
in sandy soils during periods of water infiltration and redistribution.










" 3: Fi'.;1 r 7 _MINUS

A. Description of Soils

Most Florida soils can be classified into either of four orders (Fig.
1): Spodosols, Ultisols, Entisols, and Histosols, Representatives of two
of these soil orders were selected for the location of field experiments to
determine the sJfaultanreouns movement of water and agricultural chemicals in
the -t :t-- ..-- of irrigated and fertilized corn.- One experiment was
located at the University of Florida -f '- e.rch Unit near Gainesville
on a Wauchula sand which is a Spodosol (family: sandy-over-loamy, sili-
ceous and hyperthernic; !- 'roup: Ultic Haplaquods). Although the Wauchula
soil is a Spodosol it appears in an area which is broadly characterized by
the pr esence of Ultisls. Another oxpcranct was located at the Agricul----
tural '. ch and Education Center of the University of Florida near Quincy
on a Troup sand which is an Ultisol (family: loamy, siliceous, and thermic;
subgroup: Grossarenic Paleudults).

Spodosols and associated flatwood soils represent the most extensive
order (Fig. 1) of Florida soils (Zelazny and Carlisle, 1971) and account
for one fourth of the total land area. Most of these soils occur on nearly
level to gently sloping landscapes with a generally shallow ground water
table which fluctuates near the soil surface during periods of high rain-
fall (Brasfield et al. 1973). Spodosols are characterized by the presence
of a subsurface spodic horizon which is an accumulation of organic matter
with ir, amounts of aluminum and iron. Brasfield et al. (1973) state
that the : ...,c horizon has a 1 i. ionic exchange capacity, large specific
surface are..- high water -. ::;; and high exchangeable acidity. The
spodic horizon commonly occurs less than 75 cm beneath the soil surface and
is overlain by sandy A2 eluvial and Al surface horizons, These soils may
be classed as :,- -:..ly acid sands with low fertility and base saturation.
Although the .dic horizon is generally slowly permeable, the overlying
Al and A2 transmit water : y.

Ultisols are the most extensive soils (Fig. 1) of northwest Florida
(Carlisle and Zelazny, 1973). These soils are also located in northcentral
. orida. -, .- et al. (1973) describe these soils as having B horizons
that contain an .-. i.'-,l amount of translocated silicate clay but few
bases, Sandy or loamy surface horizons generally are underlain by horizons
with loamy or clayey texture. Ultisols are typically acid, relatively
infertile, and have a low base saturation (< 35%) within about 2 meters of
the soil surface.

B. Experiments on -JCHULA j IA, a Spodosol

Nutrients -..`-._d as fertilizers to crops growing on acid, sandy soils
in Florida's humid climate are susceptable to partial leaching loss from
the -- '-; : --; of the soil. r---tilizers are typically applied to the
soil surface as dry solid materials which eventually 1..:- dissolution in
infil-' ".-;,, rainwater (or :rri-:. ion water). Thus with time a portion of
the nutrients become solutes in the soil solution. As the soil solution
moves downward in the soil y--:..i-le, nutrient solutes may be removed from








solution by uptake through plant roots, sorption onto soil particles, chemi-
cal precipitation, and biological degradation (denitrification). As the
solution moves further from the soil surface, nutrient solutes are subject
to loss by drainage in tile-drained soil and by deep seepage to the ground-
water in well-drained soil. Highly mobile nutrients such as NO3 (Thomas,
1970) are particularly susceptible to leaching from the soil; whereas the
amount of a reactive solute such as P moving in the soil solution is usually
very low relative to the total quantity of P in the soil. The mobility of
potassium in soil is usually intermediate to that for NO3-N and orthophos-
phate-P.

During the spring of 1974 a field experiment was established on a sub-
surface-drained Wauchula sand (a Spodosol) to determine in situ distribu-
tions of NH4-N, NO3-N, PO4-P, and K in the solution phase-ofTtKe soil pro-
file during the growth of corn. A subsurface system of parallel (10-cm clay
tile) drains 75 cm deep and spaced 6 m apart provided drainage for the soil
profile. Nutrient distributions with soil depth were determined in soil
receiving two levels of dolomite limestone application: a low level, 567
kg/ha and a high level, 9,070 kg/ha, The experiment was located at the
University of Florida Beef Research Unit, approximately 10 miles northeast
of Gainesville, A randomized block design with four blocks and two lime-
stone application levels was used. Each block contained six replications
to give an overall replication of 24 for each lime treatment. Each plot,
3.6 m wide and 9 m long, contained four rows of corn.

A commercial fertilizer 4-7-16 (N-P205-K20) with micronutrients was
applied broadcast to the soil surface on March 26 at the rate of 2841 kg/ha
which contained 113.6, 99.2 and 377.3 kg/ha each of N, P, and K, respec-
tively. McNair 73011 hybrid seed corn was planted on April 1 (day 0) in
rows 90 cm apart and at 15-cm intervals within each row. On May 14 (day
43) 250 kg/ha of N as NH4NO3 was applied in a narrow (5 cm) band near each
corn row. The actual amount of N applied in the band and adjusted for the
band width was 4,570 kg/ha.

Soil solution samplers and soil water tensiometers were installed in
the middle of six plots for each of the two limestone applications within
one block. Solution samplers composed of 6-cm diameter porous ceramic cups
(bubbling pressure of 1 bar) and attached to the bottom of plastic pipes
were installed in a corn row adjacent to the tensiometer installations.

The porous cups were located at 30, 60, 90, 120, and 150-cm depths,
and samples of soil solution were removed periodically. Approximately 10
to 50 ml of soil solution were collected from each sampler, and samples were
analyzed tor K, NH4-N, NO3-N, and P04-P. Methods of analyses were flame
photometry for K, specific ion electrodes for NH4-N and N03-N, and the as-
corbic acid technique as described by Watanabe and Olsen (1965) for P.
Tensiometers with 2.3-cm diameter porous ceramic cups (1 bar bubbling
pressure) and mercury manometers were installed adjacent to corn rows at
depths of 15, 30, 60, 90, 120, and 150 cm. Manometer readings provided
distributions of soil water suction, h, and total hydraulic head, H. Soil
water content-suction characteristic curves for undisturbed soil cores were
determined in the laboratory by the method of Hammond, Carlisle, and Rogers
(1971). By use of these drainage characteristic curves, values of h for a









given depth were converted to volumetric water contents, 0. This conver-
sion is normally adequate for conditions of soil water desorption, but at
best .-.: ides an approximation for water sorption due to possible error
caused by soil water hysteresis.

Soil water content-suction and hydraulic conductivity-water content
curves 4Mansell, et al. 1975) for soil materials taken from several depths
showed that the hydrologic properties among the Ap (0-25 cm), A2 (25-75 cm),
(45-75 cm), B2t (75-145 cm, and C I(>145 cm) horizons were greatly
d:.' :ent. For practical purposes water movement through the soil profile
was considered to be that for a soil having a pervious zone overlying a
slowly pervious zone. Soil above 45 cm (Ap and A2) had saturated conduc-
tivities at least 10-fold greater than that for soil between 45 and 150 cm
depths fB2i, B2t and CQ

.:i-,-: of corn seedlings occurred on April 7 (day 8), tasseling
began on June 14 (day 74), and development of ears began on June 24 (day
84). Growth curves showing relative heights of corn plants versus time for
567 and 9,070 kg/ha of dolomite limestone applied to the soil are presented
in Fig. 2. Maximum plant .;hts (Lf) for both low and high lime treat-
ments were 275 cm, Growth curves were sigmoid in shape and differences
between the treatments were not appreciable. Plant heights reached 95% of
the maximni after 80 days.

Grain and stover were harvested September 5 (day 157) and their oven-
dry weights were determined. Even though appreciable differences of corn
growth curves did not occur between limestone treatments, average grain
yields for high limestone application were 25% greater than for the low
limestone application. Average yields were 6,346 and 7,907 kg/ha for low
and high limestone treatments.

-I-..g the 157 days between planting and harvest of the corn crop, the
experimental site received a total of 76.5 cm of rainfall. Since only 3.4
cm of rainfall was received during the first 30 days after planting, 3.5 cm
of water was applied to the experimental site on day 29 by sprinkler irri-
gation. During the first 30 days, depths to the water table exceeded 100
cm, but for the remainder of the growing season the water table fluctuated
about the 75 cm depth. Thus, one month following planting of the corn, con-
ditions of water saturation and poor aeration occurred in the soil beneath
90 cm. depth. For soil depths above 60 cm, water pressure heads were always
negative indicating water unsaturation and favorable aeration conditions,
Negative pressure heads at 15 cm depth fluctuated widely with time relative
tc that for the subsoil. Extremes in soil water suction head of 30 and 200
cm of water were observed in the soil at 15 cm depth and these heads cor-
respond to volumetric water contents of 28 and 8%, respectively. Relative
to underlying soil horizons, water contents in the top 15 cm of soil in-
creased rapidly during rainfall events and then subsequently decreased dur-
ing the first 3 to 5 days of the post-infiltration period.

Distributions of K, NH4-N, N03-N, and P04-P nutrients in the soil solu-
tion as a function of depth and time are presented in Figs. 3-6 for both
lime treatments. Each data point represents the mass of a specific nutrient
in the solution phase per unit volume of the bulk soil, as determined from









the nutrient concentrations (pg/cm3 of soil solution) in the soil solution
and volumetric water contents (cm3 of soil solution per cm3 of bulk soil).
Furthermore, each data point represents an average from six replicate plots.
Distributions of nutrients in the solution phase of Wauchula soil provide a
means for describing movement of the various solutes downward through the
profile. Although 80 cm of irrigation and rain water occurred during the
first 150 days after fertilization, nutrients did not move appreciably to
depths below approximately 70 cm. Nutrient contents in the soil solution
for profile depths greater than 70 cm showed only small changes with time
throughout the entire-growing season of the corn crop.

Initially, downward movement of the zone of maximum concentration
occurred from the soil surface with time to day 43. Following day 43, this
zone showed slow penetration of these solutes beyond 50-70 cm depths. This
apparent slow transport of solutes was due to the presence of a slowly-
permeable layer (B 2t horizon) at a relatively shallow depth (75 cm). Under
such conditions, flow of water as well as nutrients in the B2h horizon
above the impermeable B 2t layer occurred predominantly in the lateral direc-
tion rather than vertically downward. Obviously much deeper penetration of
the solutes would be expected in deep, uniform, well-drained soil profiles,
reasulL., in mAle rapid leaching loss of nutrients than observed for the
Spodosol reported here. Such nutrient losses are minimized when a shallow
water table is present; however, the solute concentrations become more di-
lute depending upon depth to an impermeable layer. Also, one would expect
nutrient losses in Spodosols to be closely related to the frequency of high
intensity rainfall. Such rainfall patterns would cause appreciable lateral
flow of water andsolutes to drain tiles. Therefore, we conclude that the
presence of a shallow, slowly permeable layer such as the B 2t horizon in
Wauchula sand is beneficial in minimizing rapid leaching losses of nutrients
from a tile-drained soil.

During 1975 a second corn experiment was performed at the Beef Re-
search Unit, but three irrigation treatments--no irrigation, daily irriga-
tion (0.64 cm/day), and controlled irrigation-were imposed. Controlled
irrigation was maintained by irrigating with 1.3 cm of water when the soil
water suction at 15 cm depth exceeded 100 cm of water. A total of 2.6 cm
of water was applied during the entire season. Line sources (plastic tub-
ing) with discrete emitter holes were placed adjacent to corn rows to pro-
vide irrigation by the trickle method. For the daily irrigated plots,
irrigation was provided between April 2 (day 8) and May 30 (day 70). Water
was pumped at a constant hydraulic head through the emitters in the plastic
tubing to provide a constant volume discharge of water with time for unit
length of the irrigation tube. The trickle concept offers advantages of
increased water conservation and improved control of soil water matric
potential in the root zone. Approximately 2000 to 3000 acres in Florida
are currently irrigated (private communication with Dalton Harrison, Ex-
tension Irrigation Specialist, Department of Agricultural Engineering,
University of Florida) by trickle irrigation. A disadvantage of the con-
cept from a technological standpoint is that the irrigation water must under-
go intensive filtering of Fe, S, particulate matter, and algae to prevent
premature clogging of trickle nozzles. Theoretical and experimental analy-
sis of transient infiltration from a trickle source irrigation system have
been presented by Brandt et al (1971) and Bresler et al. (1971).









v.D fertilizer .treatments--conventional and programmed
,.olication--were also established in the 1975 experiment. Prior to plant-
ing of corn, 4,235 kg/--i of a 5-10-15 (percentages for WI P205-K20) fertili-
zer was applied broadcast to the soil surface for the conventional applica-
tion. The corn was planted on March 26 (day 0), and 672 kg/ha of 'Hi-l N3
fertilizer was : ':7-d to the soil on May 28 (day 68). The programmed
-rtilizer treatment was established by applying 5% of the total N, P anrd. K
fertilizer (4, :. kg/. of 5-10-15 fertilizer .._-., 622 kg/ik of NH4NO3) in
a hand near each corn row at the time of seedling emergence, another 10%
two weeks later, 10% at 4 weeks, 15% at 6 weeks, and 20% each at o,-jt, ten,
and 12 weeks. The tasseling stage for the corn began on May 30 (day 70) and
the corn was harvested on August 11 (day 152).

1.5 ,-,*.i of daily irrigation upon the soil water curtion head at 15
on beneath the soil s-Y Pf- can be seen in Fig. 7 for the p:-.m.. ,."d fer-
tilization. During the period when irrigation was applied daily (day 8 to
70) the soil water suction head in the irrigated plot was maintained be-
tween 40 (15% water content) and 60 (10% water content) cm of water; where-
as the suction head in the plot that received no irrigation fluctuated from
as low as 45 (14% water content) to as high as 380 (less than 6% water
content) cm of water. During days when rainfall occurred (Table 1) soil
water suction for the unirrigated plot was greatly decreased but increased
shar-.: during periods of minimal rainfall. After day 70 soil water suc-
tion in both 7ts tended to decrease during periods of rainfall and to
increase during the days after a rainfall. The increases in soil water
suction in both plots resulted from soil water being removed by evaporation
at the soil surface, water uptake by plant roots (transpiration) and down-
ward water movement by soil water redistribution (gravitational drainage).

As euxp -t-1 the effect of trickle irrigation upon soil water suction
was less evident for soil depths of 30 (Fig, 8) and 60 (Fig. 9) ncm than at
the shallower 15 cm (Fig. 7) depth. This effect can also be seen in Figs.
10 and 11 where distributions of soil water pressure head with soil depth are
presented for the unirrigated and irrigated plots during a 26-day period of
minimal rainfall. In the plot receiving no irrigation, soil at the 15 cm
< .. underwent the greatest drying in the soil profile. Soil at the 30 cm
depth also exhibited drying but to a lesser extent than at the 15 cm


At day 20 the distributions of soil water pressure head with soil
depth were similar for both the irrigated and unirrigated plots. Water
tables ( -__ "e head = 0) were located at 70 and 65 cm depths, respectively,
for the irr<.:-t-3 and unir-1i: t soil profiles, indicating conditions for
water saturation for soil depths greater than these depths and unsaturation
at shallower d;] 2.... Vertical gradients for hydraulic head, H, with depth
were approximately -0.011 in both irrigated and unirrigated soil for all
depths greater than 60 cm. This small negative gradient of hydraulic
head indicates that water is slowly draining from the soil profile. As
time ".,.. :. from day 20 to day 46 the gradients of H in the top 30 cm of the
uni- \:ated soil reflected a net upward water movement due to evaporation and
water ., A..-: by plant roots. For example at day 30 the gradient of H between
15 and 30 cm depths was +0.933; whereas for the irrigated soil that value









was -1.067 indicating a net downward flow of water due to the daily addition
of 0.64 cm of water. Thus the distributions of soil water pressure head
shown in Figs. 10 and 11 show that the daily application of 0.64 cm of
water by trickle irrigation was more than sufficient to compensate for water
loss in the top 60 cm of soil due to evaporation, transpiration, and drain-
age. The amount of 0.64 cm of water per day was selected because it re-
presents the maximum potential removal of water from the soil by evapotran-
spiration for the summer months.

Average grain yields for-corn grown on Wauchula-sand during 1975 under
three irrigation and two fertilizer management programs are presented in
Table 2. For either the conventional or programmed fertilization, daily
irrigation resulted in higher yields of grain, but the percentage yield
increases due to daily irrigation were relatively small (21.7% for conven-
tional fertilization and 20.5% for programmed fertilization). For conven-
tional fertilization, grain yields for controlled irrigation were only
slightly higher than for corn irrigated daily. The grain yield reported
for corn receiving programmed fertilization and controlled irrigation was
unexplainably lower than other treatments.

Irrigation use efficiencies (calculated by subtracting grain yields
for unirrigated corn from yields for corn receiving daily irrigation and
controlled irrigation and dividing the results by 39.7 cm of water for the
daily irrigation treatment and 2.6 cm for the controlled irrigation) were
31.1 and 510.4 kg/ha/cm for conventionally fertilized corn receiving daily
and controlled irrigation. These values in conjunction with soil water
suction and grain yield data indicate that the controlled irrigation pro-
vided much more efficient use of the water applied and that the daily irri-
gation resulted in water application to the soil in excess of that actually
needed to produce economically profitable grain yields in Wauchula sand.
Even the grain yields for unirrigated corn could be considered to be econo-
mically profitable for the Wauchula sand. The relatively large yields for
unirrigated corn was partially attributable to the presence of a water-
saturated zone in the lower portion of the soil profile during the growing
season for the corn. For the daily irrigated corn, irrigation use effi-
ciencies were not greatly influenced by the method of fertilizer applica-
tion (31.1 and 33.0 kg/ha/cm for conventional and programmed fertilization,
respectively). That result is also supported by the observation from the
1974 experiment which showed that leaching fertilizer nutrients in Wauchula
sand was primarily limited to the top 50-70 cm of soil.





Entisi
H n i s I''s;,N


H is o i "


F'.
,- sPooosa:,

S SPODOSL: ',
,,i




ii -


4






U
~.t. -2,0'







N"
C,



'a


7~,~~
'ii
4r, 2
2,~ I' -.
- II I,


4'









H. -1-1s
0


S0-j

w I-
(D O
(ID z
r+e
< 75-


O z


00
7q 0



(. Dol omite


25 3.49cm567 kg/ha
0L I rr igatio x 9,070
g -- |Seedlfng 4/
SEmergence



0 25 50 75 100
0 25 50 75 100


TIME (days )





-11-


\ -- "*---v---3

\---43
----- ---\ 116














'67 kg/ha Lime 9070 kg/ha Lime
0 [-h L Lime





















Figure 3: Distributions of potassium with depth in Wauchula sand treated with
567 and 9,070 kg/ha of limestone during 1974.




-12-


25


NH4-N ( yg/cm3soil)
59 0 25


567 kg/ha Lime


9C


29 days
---------30
-- 43
116


)70 kg/ha


Lime


Distributions of NH4-N with depth in Wauchula sand treated with 567 and
9,070 kg/ha of limestone during 1974.


I


'C
u





0
o

C
Ln


if


Figure 4:


0U .*~i -ji. TuTg -i .rm n.. 11 ii~ g i 1 f i i in im m i W u ~ ^ii i* m i^ ~iILii I*


0






-13-


Nitrate-N
f020


E
U


I I.,




N;


567 ha Lime


/cm3soi I


-- 29 days
------ 30
- 43
......... 116


9070 kg/ha Lime


Distributions of N03-N with depth in Wauchula sand treated with 567 and
9,070 kg/ha of limestone during 1974.


Figure 5:


>




-14-


0.2


0.4


P04


9 0 0
I-
I ~





* /
* /
/
/


567 kg/ha Lime


-P (pg
C)


---


/cm3soil)
Q2


0
/


29 days
---- 30
--43
.......... ... 116


9070 kg/ha


Lime


Distributions of PO4-P with depth in Wauchula sand treated with 567 ani
9,070 kg/ha of limestone during 1974.


E8
U
-0
-f


0

'O
o )


Figure 6:


04




-15-


11/H


r


A
- 8


'I


Irr iot.d


U i r gatd


. Period .-f Irri action


0
9' ~


Days


Figure 7- Soil water suction head at 15 cm soil depth in Wauchula sand during the
1975 growing season for daily irrigated and unirrigated corn.


io<
r
[


r


0

U-




-16-


L- 3
10

Soil Depth: 30 cm
'4-
0

E Irrigated
S Unirrigated










U) -Period of Irrigation


10
O 40 80 120
Days

Figure 8: Soil water suction head at 30 cm soil depth in Wauchula sand during the
1975 growing season for daily irrigated and unirrigated corn.




-i/-


1 p4
p4


% I
\ e0
% S


Irrigated

Un irrigated


Period ,of


Irrigation


*4i 0


Days


Soil water suction head at 60 cm soil depth in Wauchula sand during the
1975 growing season for daily irrigated and unirrigated corn.


t t
t :h
/ "


8 0


Figure 9:


0 L


12 '
b^^l.






Soil
-300


Water
-200


Pressure(cm of water)


30


Day


Day 46


Unirrigated


Figure 10:


Distributions of soil water pressure head with soil depth in Wauchula
sand during a selected 26-day period of the growing season for un-
irrigated corn in 1975.


+100


-CD

-J
1)
a)


20


















n oq o

o 0
-j (D H-j

(D~


0 03

CDo

(CD


H 0D

tH.
CA
(D


w


cr

0..
U)
Qm


0
00





-20-


Table 1: Rainfall at the Beef Research Unit near Gainesville
during the spring and summer of 1975.


Day Rainfall Day Rainfall Day Rainfall
(cm) (cm) .(cm)

*0-8 0.0 72 1.8 96 1.0

9 0.3 73 0.1 97-99 0.0

10-14 0.0 74-77 0.0 100 0.3

15-26 10.7 78 0.4 101-105 0.0

27-40 0.0 79 0.1 106-108 0.8

41 0.1 80 1.4 109 0.6

42-43 0.0 81 0.0 110 0.2

44 0.2 82 0.2 111 0.0

45-49 0.0 83 0.1 112 2.4

50 0.8 84 0.0 113 2.0

51-53 0.0 85 0.2 114 0.6

54-60 2.7 86 0.0 115 0.2

61-64 0.0 87 0.3 116-118 0.0

65 0.4 88-91 0.0 119 2.7

66 1.3 92 0o9 120 0.2

67 0.0 93 0.1 121-123 0.0

68 6ol 94 1.1 124-131 3.6

69 0.7 95 0.0 132 0.6

70-71 0.0 133 2.4

Total 47.o


* Day 0 was March 26





-21-


Table 2: Corn grain yields on Wauchula sand under three
irrigation and two fertilizer management programs,
Gainesville, 1975,


Grain Yield, kg/ha1
Fertilizer No Daily Controlled
management irrigation irrigation2 irrigation3

Conventional 5708 ac 6949 ab 7035 ab

Prograamed 6366 abc 7672 b 5206 c

1LSD 05 1612 kg/ha. Grain yields followed by the same
alphabetical letter (a,b or c) or combination of letters
did not differ from each other. LSDO.05 is the least signi-
ficant difference at a 5% probability level(Steel and
Torrie 1960).
2Application of 0.64 cm/day from day 8 to day 70; total
application, 39.7 cm.
3Irrigation with 1.3 cm of water when soil water suction at
15 cm depth exceeded 120 cm of water; total application,
2.6 cm.




-22-


C. EXPERIMENTS ON TROUP SAND, AN ULTISOL

During 1975 a field experiment of irrigated corn was established on
Troup sand (an Ultisol) at the University of Florida Agricultural Research
and Education Center near Quincy. The purpose of the experiment was to
determine distributions of soil water pressure head with depth during the
growth of irrigated and unirrigated corn and to determine grain yields for
corn grown under 3 irrigation and 2 fertilizer management programs. The
soil profile had a compacted plow layer in the top 15 cm but was essential-
ly a uniform sand between 15 to 150 cm depths with a sandy clay material
located at 150 cm. This soil was originally classified as Lakeland sand,
but the profile characteristics more closely resemble Troup sand. The names
Lakeland sand and Troup sand are used to represent the same soil in this
report. Graphs of hydraulic conductivity versus soil water content for
selected horizons of this soil are presented elsewhere (Elzeftawy and
Mansell, 1975).

Coker 77 hybrid corn was planted on March 24 (day 0) in rows 90 cm
apart and with 15 cm spacing between seed in each row. Three soil-water
management treatments were maintained: (a) no irrigation, (b) daily irri-
gation of 0.64 cm of water applied through plastic trickle irrigation tub-
ing placed near each corn row, and (c) 1.30 cm of water applied with
trickle tubing when the soil water suction head at 15 cm depth exceeded
100 cm of water. Two fertilizer treatments were also used: (a) conven-
tional fertilization where 2240 kg/ha of a 5-4.4-12.4 (percentages of N-
P-K) fertilizer was applied prior to planting of the corn and 670 kg/ha of
NH4NO3 was later applied in two equal applications (May 14 and 17) and
(b) programmed fertilization of 2240 kg/ha of 5-4.4-12.4 fertilizer and
670 kg/ha NH4NO3 applied in several split applications. In the programmed
fertilization treatment 5% of the N, P and K was applied at the time of
seedling emergence, 5% two weeks later, 10% at four weeks, and 20% appli-
cations at 6, 8, 10 and 12 weeks.

Tensiometers were installed at 15, 30, 60, 90, 120, and 150 cm depths
beneath the soil surface and adjacent to the corn rows. Soil water pressure
head measurements were obtained from the tensiometers at selected times
during the growth of the corn. The grain was harvested on August 8 (day
141).

Major fluctuations in soil-water suction head with time at the 15 and
30 cm soil depths were observed in the unirrigated plots (Fig. 12). At the
15 cm depth suction heads ranged from as low as 100 cm of water immediately
following a rain to as high as 800 cm of water several days following rain.
Major periods during which the surface soil underwent extensive drying were
observed during 30-40, 55-59, 70-80, and 86-100 days after planting of the
corn, Only minor fluctuations in suction head were observed deeper in the
soil profile. For example,at the 90 cm soil depth the suction head ranged
from near 0 to slightly above 100 cm of water between 25 and 100 days after
planting of the corn. Between days 25 and 37, positive soil water pressure
heads were temporarily observed above the slowly permeable sandy clay iayer
at 150 cm depth. Heavy rains in the early part of the season were responsi-
ble for the positive soil water pressure at 150 cm depth.






23-




Daily Virrigation of 0.64 cm of water resulted in soil water suction
head;ls less than 100 cm at ill depths (Fig. 13) in the profile during the
rowingg season or thie corn.i A total of 56 an of water was applied as
I IV tion during the growing season for the corn. Major fluctuations in
sit ion haead as seen in the unirrigated plot were not observed at the 15 cm
depth in the daily irrigated treatment. Thus daily irrigation of 0.64 cm
of water was :.. -.-i'::ly more than sufficient to compensate for water uptake
by plant roots. At the 60 oa soil depth the suction head ranged from 55 to
80 cm of water.

Soil water pressure head for 15, 60 and 150 cm soil depths are shown
in Fig. 14 for the treatment which was irrigated with 1.30 cm of water when
the suction head at 15 cm depth exceeded 100 cm of water. A total of 31 cm
of water was applied by irrigation during the growth season for the corn.
The irrigation minimized the magnitude of major fluctuations in the soil
suction head at the 15 cm depth. The suction head ranged from a low of 40
cm of water to a high of 300 cm of water in the surface soil.

Yields of corn grain harvested from plots receiving the three soil-
water management and the two fertilizer management treatments are presented
in Table 3. The highest yields and the highest use efficiency for the
irrigation water (Table 4) were obtained for corn grown with programmed
fertilization and controlled irrigation (1.30 cm of water applied when the
soil water suction head at 15 cm depth exceeded 100 cm of water). In
u.,i i. i.ated plots the fertilizer treatment did not affect the grain yields;
however, in irrigated plots programmed fertilization resulted in higher
grain yields than when the fertilizer was applied in a conventional manner.
Programmed fertilization resulted in more efficient use of irrigation water
in the in ii; --. corn plots. In plots receiving daily irrigation, water
use efficiency for programmed fertilization was approximately 4 times larger
than for conventional fertilization, and in plots receiving controlled irri-
gation, water use efficiency for programmed fertilization was approximately
twice that for conventional fertilization. For plots receiving either con-
ventional or programmed fertilization, grain yields as well as irrigation
use efficiencies were higher for controlled irrigation where tensiometers
were used to schedule water applications than for daily applications of
Irrigation water,

One of the major conclusions that can be obtained from this experiment
is that fertilizer and water management for corn growing in sandy soil
should be treated as mutually dependent phenomena. For example, excessive
and indescriminate use of irrigation may decrease use efficiency of ferti-
lizer and irrigation water by excessive leaching from the "rooting zone" of
the soil profile and can result in nutrient contamination of ground-water.
Also programmed fertilization tended to improve the use efficiency of
irrigation water applied to corn.




-24-


45 65


105 125


Days


Figure 12:


Soil water pressure head at 15, 30, 90, and 150 an soil depths in
Troup sand during the 1975 growing season for unirrigated corn.


0
4

0
0
E



01
-!


L
a,
0




-25-


45 65


85


105 125


Days


figure 13:


Soil water pressure head at 15, 60, and 150 cm soil depths in Troup
sand during the 1975 growing season for daily irrigated corn,


>/


-10w


C:




rL




, i'
<~*I


[^


Irrigated Daily
(0.5 4 c m/d a y)
/


Soil 15 cm
.-l15cm


-15 0


I I I p I I I


-40


0
+20
+40


5










ai)
0



E
U
c)
L
en
.)

L_
0L
L
U



0
if)


25


105 125


Days


Figure 14:


Soil water pressure head at 15, 60, and 150 cm soil depths in Troup
sand during the 1975 growing season for corn irrigated when the pr
sure head became less than -100 cm of water at 15 cm depth.


-26-


L


-300


-200


-100


0
+20
+40


Controlled Irrigation




Soil Depth
-15 cm
I
| I I

fi I e




- ---
p p p p V 6


45


65


85





























Table 3: Grain yields for corn grown under three irrigation and two fer-
tilizer management programs on Troup sand, Quincy, 1975.


Fertilizer Treatment


Conventional

Programmed


Grain Yield, kg/ha'
-S Da-lay rrT-gaTlon Controlled
Irrigation (0.64 cm/day) Irrigation2


2,860 a

2,710 a


3,380 a

4,950 b


4,760 b

6,600 c


LSDO.05 (a,b,c) ; 1190

LSD0o05 is the least significant difference at a 5% probability level
(Steel and Torrie 1960).

Grain yields followed by the same alphabetical letters (a,b or c) or
combination of letters did not differ from each other.

1o30 cm of irrigation water applied when the soil water suction head at
15 cm depth exceeded 100 cm of water.




-28-


Table 4: Irrigation use efficiency* for corn grown under two irri-
gation and two fertilizer management programs.


Irrigation Treatments
Fertilizer Program Daily irrigation Controlled Irrigation
(kg of grain/ha/cm of irrigation water)

Conventional 9.3 61.3

Programmed 37.3 120.6

* Irrigation use efficiency was calculated by subtracting grain yield
for the unirrigated plots from those for the irrigated plots and divid-
ing the result by 56 cm for the daily irrigated treatment and by 31 cm
for the controlled irrigation treatment.






-29-


CHAPTER 4: MATHEMATICAL MODELS

During the course of this study several mathematical models have been
developed for the purpose of describing soil water movement during infil-
tration and redistribution and the transport and interactions of applied
2,4-D1, orthophosphate, and potassium in soils. Initially, existing models
were utilized to describe the fate of these solutes in soils. However, it
was found that new models were needed in order to account for the various
mechanisms governing solute interactions which influence the movement of
these solutes in the soil-water system. The capability of such models to
predict the simultaneous movement of water and applied solutes is clearly
dependent upon their validity under controlled laboratory conditions and
more crucially under field conditions. In this study, the water movement
model as well as the 2,4-D and orthophosphate models have been tested
against field and laboratory results.

A. Water and 2,4-D Nbdels:

The water and 2,4-D models are based on the numerical solution of the
water flow equation and the solute transport equation for transient and
unsaturated water flow conditions. The transient one-dimensional water
flow equation for unsaturated soils is


Dt Dzz z

where t denotes the time (hr); z the vertical distance in the soil (cm)
positive downward; 6 the volumetric soil water content (cm3/cm3); D(9) the
soil water diffusivity (cm'/hr); and K(9) the hydraulic conductivity (cm/hr).
The corresponding equation governing transient one-dimensional solute trans-
port through soils may be written as

D(S + 6C) (D() C (vc) (2)


where C is the solute concentration in soil solution (mg/cm3); S the amount
of solute sorbed by the soil matrix per unit volume of soil (mg/cm3); v the
Darcy water flux (cm/hr); Ds(v) the solute dispersion coefficient (cm2/hr);
and v (= v/e) the pore water velocity (cm/hr). The dispersion coefficient
Ds(9) of equation (2) is assumed to be a function of pore water velocity,
V. Adsorption isotherms which express relationships between quantities of
solute adsorbed by the soil, S, and corresponding concentration in soil
solution, C, were used to evaluate the rate-of-sorption term DS/8to Equa-
tions (1) and (2) subject to appropriate boundary and initial conditions
were solved numerically using explicit-implicit finite difference approxi-
mation techniques (see Selim and Mansell, 1974; Selim, Mansell, and
Elzeftawy, 1976). Simulated results from the water and 2,4-D models were
compared with field measurements during continuous water infiltration in
a Lakeland sandy soil (Ultisol) near Quincy, Florida. It was found that
increasing the water flux from 2.08 to 4.10 cm/hr during irrigation caused
the maximum soil water content, 6, at the surface to increase (Fig. S1 and
16) from 0.26 to 0.29 cm'/cm3. The higher flux resulted in more rapid pene-
tration of the wetting front. Predicted profiles of soil water as obtained
using the numerical solution of the water flow equation (1) agreed well with
water contents measured in the field; differences were within 0.02 to 0.05
cm'/cm3. Thus predicted water content profiles enabled evaluation of water
penetration with time into the soil for the infiltration period studied.




-30-


Figures 17 and 18 show measured and calculated 2,4-D distributions
during water infiltration with flux of 2.08 and 4.10 cm/hr. These field
measurements indicate that the calculated 2,4-D distributions lagged behind
the measured ones for all times during the infiltration process. The lag
between calculated and measured distributions were between 2 and 5 cm.
Retardation of calculated distributions with respect to measured values
did not appear extensive; thus adequate agreement may be concluded. How-
ever, such retardation of calculated solute distributions could be attri-
buted to overestimation of 2,4-D adsorption as determined by equilibrium
isotherms,

Results of 2,4-D and water movement in laboratory soil columns of
Lakeland fine sand (Elzeftawy, Mansell, and Selim, 1976) showed essentially
the same trends as that from the field studies. Moreover, from the field
and laboratory results it was clear that following water application of
14.3 cm,.penetration of 2,4-D did not exceed 30 cm below the soil surface.
Therefore, we may conclude that 2,4-D movement in the soil is extremely
slow which minimizes potential contamination of groundwater resources when
small applications of 2,4-D herbicide are used.

Bo Phosphorus Models:

It was necessary to develop several models in order to describe the
fate of phosphorus in a Spodosol. These models (see Mansell et al., 1976),
are based upon solving the solute transport equation with several phos-
phorus interaction mechanisms.

The transport equation governing one-dimensional phosphorus movement
through soils may be written as

S C 2C C '(3)
P'+ e -= 0 D T-2 q -

where C is the concentration of the solute in soil water (pg/cm3), p the
soil bulk density (g/cm3), 9 the soil water content (cmD/cms), Ds the hy-
drodynamic dispersion coefficient (cm2/hr), q is the soil water flux (cm/
hr), S the amount of solute sorbed by the soil particles (pg/gm), Q the
rate of solute removal per unit volume of soil (pg/cm3/hr), t is time (hr),
and z is distance in the soil (cm). The term 3S/t represents the rate for
reversible adsorption (or exchange) of solute by soil particle surfaces,
and the sink term, Q, is the rate for irreversible removal (i.e., precipi-
tation and immobilization) of solute from the soil.

In this study, three phosphorus sorption processes were considered
for the reversible term,3./ t,in equation (3). First, adsorption was
assumed to be linear and instantaneous

S = KC (4)

where K is the solute distribution coefficient (cm3/g) as determined from
adsorption isotherms. Secondly, a nonlinear instantaneous adsorption
mechanism was considered,

S = K CN (5)





-31-


where N < 1. Thirdly, a kinetic reaction between S and C such that


t _\ I k K: ) (6)

was used. This equation describes a finite adsorption rate proportional to
the difference between the equilibrium adsorption and the amount which is
already adsorbed. The parameter ka is the forward (adsorption) and kd is
the backward (desorption) reaction rate coefficients (hr-1), respectively.
S the value of N ci. :'s unity, equation (6) describes a reversible,
first-order kinetic reaction. For many reactive solutes such as herbicides
and phosphorus, the value of N has been found to be less than unity.

Irreversible sorption processes assigned to the Q term of equation
(1) were considered in two ways. First, the sink term was given the form

Q k (6C) (7)

where kc is the rate coefficient (hr-1) for precipitation and/or chemi-
sorption of solute from the soil solution. Secondly, the sink term Q is
expressed as,

Q = ks(pS) (8)

where ks is considered the rate coefficient (hr-1) for chemical immobili-
zation of physically adsorbed phosphorus.

The phosphorus models were used to describe experimental data for
miscible displacement of phosphorus solution in Oldsmar fine sand (Spodo-
sol) columns, The soil was obtained from an uncultivated and unfertilized
area at the University of Florida Agricultural Research Center located
near Fort Pierce, Florida. Duplicate samples from undisturbed soil cores
(5.4 an inside diameter and 10 cm in length) were collected from surface
41, and subsurface, A2, horizons.

Phosphorus breakthrough curve from the A2 subsurface soil is shown
in hig. 19 where the relative concentration in the effluent (C/Co) is
plotted against the number of pore volumes (V/Vo). Calculated concentra-
tion distributions in the effluent using equilibrium adsorption isotherm
(S = 0.23 C) overestimated the measured retardation of phosphorus transport
in the A2 soil. Only when the distribution coefficient, K, was arbitrarily
decreased tenfold (S = 0,023 C) did the simulated curve fit experimental
results., This tenfold decrease in K assumes that only 10% of the adsorp-
tion sites were active in phosphorus removal. Therefore, we conclude that
the equilibrium adsorption isotherm did not adequately describe phosphorus
transport for the relatively high soil water pore velocities used in the
subsurface A2 soil. However, using the kinetic equation (6) to represent
phosphorus adsorpt- calculated breakthrough curves provided an ade-
quate description of phosphorus transport in the A2 soil. Therefore, we
conclude that phosphorus transport in the A2 soil appeared to occur under
fully reversible conditions. In addition, the adsorption-desorption
mechanism is of the first order kinetic type.




-32-


The phosphorus breakthrough curve from surface Al soil core (Fig. 20)
is distinctly different from that for the A2 soil (Fig. 19). The break-
through curve shows a large retardation, assymmetry, and extensive tailing
during desorption (right-hand side). Calculated results from the equili-
brium adsorption model (curve not shown) did not describe the shape or
position of the experimental data. However, the calculated phosphorus
breakthrough curve based upon the kinetic adsorption model (the NO SINK
curve in Fig. 20) adequately described the left-hand side of the experi-
mental data. Deviation between calculated and experimental results as well
as incomplete recovery of applied phosphorus (80% recovery after 32 pore
volumes) suggests that phosphorus retention in the A1 soil may involve
irreversible as well as reversible processes.

Since complete recovery of phosphorus in the effluent was not attained
in the Al soil, an irreversible sink term Q was included in the kinetic
transport model for the purpose of providing better description of the
experimental data. Equation (7) was first used to represent irreversible
precipitation or chemisorption, and equation (8) to describe chemical
immobilization of physically adsorbed solute. Calculated breakthrough
curves (Fig. 20 and 21) clearly depend upon the magnitude of these rate
coefficients; however, optimum values for kc and ks were 0.1 hr-1. Thus,
phosphorus sorption during transport through the A1 soil apparently
occurred under conditions greatly removed from equilibrium with sorption-
desorption mechanisms occurring primarily as reversible processes but also
accompanied by irreversible processes such as chemical immobilization or
precipitation.

Although the sink models for irreversible chemical immobilization and
precipitation of phosphorus improved the agreement with experimental data,
such agreements were not considered satisfactory. Furthermore, the exact
mechanisms of irreversible adsorption were not identified. Therefore, a
new phosphorus model (see Mansell, Selim, and Fiskell, 1976) was developed
to describe various transformations which might occur for different soil
and under different conditions. The model is a multistep one where phos-
phorus is considered to be present in four phases: A) solution, B) adsorb-
ed, C) immobilized, and D) precipitated. The four phases of phosphorus
are interrelated to each other by the rates of individual transformations
which are determined by both the rate coefficients and the amounts of
phosphorus in each phase. Under steady flow conditions, the transport and
transformations of these four phases of phosphorus may be expressed as

A= D v [kAN + ksA] + [k2B + k6D] (9)


(B) = k,6AN [k2 + k3]pB + k4pC (10)


S(pC) = k pB k pC (11)

S(pD) k OAN k pD (12)
at 5 6











0 = volumetric soil water content (cm3/cm3),


p = soil bulk density (g/cm'),

t = time (hr),

z = transport distance in soil (cm)

v = average pore water velocity (cm/hr),

Ds = dispersion coefficient (cm2/hr),

B, C and D = amounts of phosphorus per gram of soil for adsorbed,
immobilized, and precipitated phases, respectively (pg/g),

A = phosphorus concentration in soil solution phase (pg/cm3)

N = constant representing the order of the adsorption pro-
cess,

and kj, k2, k3, k4, k5, k6 = rate coefficients for adsorption, desorption,
immobilization, mobilization, precipitation,
and dissolution, respectively (hr-1).

At any given time, the sum of the amounts of phosphorus per gram of soil in
the adsorbed (B), immobilized (C), and precipitated (D) phases will be
referred to collectively as the total amount of phosphorus sorbed per gram
of soil, S. The rate coefficients are assumed to depend upon characteris-
tics of the soil environment such as level of acidity (pH), concentrations
of reactive Fe and Al present and the redox potential of the soil. Ad-
sorption was assumed to be Nti order where N = 0.35 (Mansell et al. 1976),
whereas desorption, immobilization, mobilization, precipitation, and
dissolution reactions were assumed to follow first order kinetic reactions.
The model, however, is versatile and other transformation processes can be
included as deemed necessary. Furthermore, validation of this model is
needed in order to fully predict the fate of applied phosphorus in soils.

C. Potassium Model:

A mathematical model was developed (Selim, Mansell, and Zelazny, 1976)
for the purpose of describing potassium transport and transformations in a
soil profile. Potassium was considered to be present in the soil in four
phases; solution, exchangeable, nonexchangeable (complex secondary miner-
als), and primary mineral. Simulated results were presented for assumed
conditions of first-order kinetic reactions between the exchangeable, non-
exchangeable, and primary mineral phases. However, themodel is flexible
and can be adapted to incorporate other transformation processes as well
as variable rate coefficients with soil depth and time.

Potassium distribution results (see Selim, Mansell, and Zelazny, 1976)
illustrate the dependence of the leaching loss and transformation among
the various potassium phases upon the rate coefficients governing the


whio re




-34-




transformation mechanism. Since water flow through the soil was maintained
constant at all times, i.e., continuous water infiltration, the results
presented here provide optimum conditions for transport and leaching losses
of potassium. Transformations and transport of potassium under the more
realistic conditions of water infiltration and redistribution should result
in less movement of applied potassium. Such infiltration and redistribution
conditions have been considered previously and can be adapted for the trans-
port and transformations of potassium.

Verification of the model presented here is needed in order to fully
describe the chemical fate of applied potassium in soil. Experimental
evaluation of our model for various potassium transformation mechanisms
is currently being investigated for several soils.






-35-


Water Content


3 3
cm /cm


.0 .05 .10 .15 .20 .25


N30

-CF


540


0
n 50



60


initial


Calculated


Experimental


Figure 15. Experimental and calculated soil water profiles in Lakeland fine sand
near Quincy, Florida during water infiltration with 2.08 cm/hr flux.




-36-


WATER CONTENT 0 cm3/cm3


Figure 16.


Experimental and calculated soil water profiles in Lakeland fine sand
near Quincy, Florida during water infiltration with 4.10 cm/hr flux.





-37-


Co. .entration


30


C,
40


A


ppm
50


1.5 hr

2.5 hr


3.5 hr


.5 hr
Experimental x 1.5
A 2.5
o 3.5
Calculated


Figure 17. Distributions of 2,4-D in the solution phase of the soil profile of
Lakeland fine sand near Quincy, Florida ,during water infiltration with
2.08 cm/hr flux.


20


2,4-D
10


0.
L 15


o
if 20


25







0


5


10
E
U

N15
-C

S20


025


30


35


40


45


2,4-D
10


-38-
Concentration


20


30


C 4
40


ppm
50


Figure 18o Distributions of 2,4-D in the solution phase of the soil profile of
Lakeland fine sand near Quincy, Florida during water infiltration with
4.10 cm/hr flux.






1 0


0
(3


0.4


0.2


0-
0


Figure 19.


Experimental and predicted phosphorus breakthrough curves for a
(A2) Oldsmar soil.


core of subsurface


O6























4 8 12 16 20 24 28 32
V/Vo


Figure 20.


Predicted phosphorus breakthrough curves with and without a sink term for irreversible immobilization
of adsorbed P during transport through a wa' -unsaturated core of surface (CA1 Oldsmar soil.


1.0


0.8


0

U


0.6


0.4


0.2-


0-






1.0













0.2


Figure 21.


Predicted phosphorus breakthrough curves with and without a sink term for irreversible precipitation
of soluble P during transport through a water-unsaturated core of surface (Al) Oldsmar soil.




-42-


CHAPTER 5: Summary and Conclusions

Sandy agricultural soils of Florida are typically limited in their
capacities to retain irrigation water and applied chemicals such as fer-
tilizers and herbicides. Therefore in order to economically grow optimum
yields of crops on these soils under the humid climate of Florida, frequent
applications of fertilizer is often required to compensate for leaching
losses of nutrients from the "rooting zone" of the soil profile. Since
these soils also have limited capacity to store infiltrated water during
periods of redistribution, irrigation is also required during extended
periods of infrequent rainfall. The simultaneous transport of nutrients
and water through these sandy soils is therefore of utmost importance to
the efficient use of both fertilizers and irrigation water for agricultural
production of food. For example, leaching of nutrients from the soil tends
to decrease the efficient use of fertilizers, whereas uptake of nutrients
by plant roots tends to increase the use efficiency. Not only is the leach-
ing of nutrients detrimental to food production but leached nutrients that
move with the soil solution into underlying groundwater also pose a poten-
tial problem for environmental pollution. Results from this research indi-
cate that irrigation and fertilizer management of corn growing on sandy
soils in a humid climate mutually effect the individual use efficiencies for
applied water and nutrients.

Field experiments of irrigated and fertilized corn were established
on two representative Florida soils: Wauchula sand, a Spodosol, and Troup
sand, an Ultisol. The Wauchula sand was characterized by relatively slow
rates of natural internal drainage due to the presence of subsurface hori-
zons of slowly-permeable material in the soil profile and due to relatively
flat terrain. By contrast, the Troup sand was characterized by a rela-
tively uniform profile of sand which transmits soil water rapidly downward.
Beneath 150 cm depth, however, the Troup sand was underlain by a slowly-
permeable sandy clay material. Water tables were commonly observed in the
Wauchula soil profile between depths ranging from 60 to 100 cm during the
growing season for corn; whereas, water tables rarely existed in the Troup
soil profile. Although the Troup soil profile was primarily composed of
sand material to a 150 cm depth, a compacted plow layer occurred in the top
30 cm of soil which was observed to limit much of the plant root growth to
the top 15 cm of the entire 150 cm profile. Soil water movement was not
observed to be decreased by the presence of the plow layer.

Corn was grown under three irrigation treatments--no irrigation, daily
irrigation (0.64 cm/day), and controlled irrigation (1.3 cm of water was
applied vhen the soil water suction at 15 cm soil depth exceeded 100 cm of
water) and two fertilizer application treatments--conventional (bulk of
fertilizer applied prior to planting of corn and nitrogen fertilizer
applied later as a sidedressing and programmed (fertilizer applied in
several split applications). Without irrigation, grain yields of corn were
approximately twice as great for either fertilizer treatment in the WauchuLa
sand as compared to those in the Troup sand. The larger yields were attri-
buted to the larger water contents and less leaching of fertilizer nutrients
that occurred in the Wauchula soil profile as compared to the Troup soil dur-
ing the growing season. For both soils with corn grown at either irrigation






-43-


treatment, grain yields for corn fertilized in a programmed manner were
higher than yields where Fertilizer was applied conventionally. Water use
ciFicLlewv (not grain yield divided by amount of irrigation water applied)
was only sligh i tly ;altered by changing the method of fertilization for corn
growing iIn the "!' i, .lai soil; however, for corn growing in Troup sand irri-
gaition water was used much more efficiently by applying the fertilizer in a
progemunmed manner. The differences in water use efficiencies for corn on
the two soils were attributed to the greater magnitude of nutrient leaching
observed in the .. : sand relative to that in the Troup sand. Water use
efficiency for corn growing in both soils was greater for controlled irriga-
tion than for daily irrigation. In the Troup soil highest grain yields and
water use efficiencies were observed for corn fertilized in a programmed
manner and irrigated in a controlled fashion. Daily irrigation apparently
resulted in excessive leaching of both nutrients and irrigation water in the
Troup sand. Thus corn growing in both soils used irrigation water and
fertilizer nutrients more efficiently for programmed versus conventional
fertilization and more efficiently for controlled versus daily irrigation.

Several mathematical models were developed and used to describe the
simultaneous transport of water and applied agrichemicals (fertilizer nu-
trients and herbicides) to sandy soils. A transport model for water alone
was successfully verified with experimental results from laboratory and
field experiments for periods of infiltration and redistribution. This model
is flexible and can be adjusted to incorporate a variety of initial and bound-
ary conditions for a range of soil types. Nutrient (phosphorus and potassium)
and herbicide (2,4-D) models were also developed and used to describe chemi-
cal-physical interactions and movement of these reactive solutes through
soil. Simulated and limited experimental results associated with these
models indicate that mechanisms such as physical adsorption-desorption,
chemical precipitation, and immobilization or fixation have a large influence
upon the transport of potassium and phosphorus, even in sandy soils. Fur-
ther experimental validation of possible interactions of these agricultural
chemicals during movement with water through soil is needed in order to
precisely predict the leaching loss of agrichemicals applied to sandy soils
of humid climates.




-44-


LITERATURE CITED

1. Brandt, A., E. Bresler, N. Diner, J. Ben-Asher, J. Heller, and D.
Goldberg. 1971. Infiltration from a trickle source: I. Mathemati-
cal models. Soil Sci. Soc. Amer. Proc. 35:675-682.

2. Brasfield, J. F., V. W. Carlisle, and R. W. Johnson. 1973. Spodosols--
soils with a spodic horizon. Pages 57-60 of Soils of the Southern
States and Puerto Rico, Bulletin No. 174, Agricultural Experiment
Stations of the Southern- States and Puerto Rico and the Soil Conser-
vation Service of the United States Department of Agriculture.

3. Bresler, E., J. Heller, N. Diner, I. Ben-Asher, A. Brandt, and D.
Goldberg. 1971. Infiltration from a trickle source: II. Experi-
mental data and theoretical predictions. Soil Sci. Soc. Amer. Proc.
35:683-689.

4. Butson, K. D., and G. M. Prine. 1968. Weekly rainfall frequencies in
Florida. Circular S-197, Agricultural Experiment Stations, University
of Florida, Gainesville.

5. Carlisle, V. W. and L. W. Zelazny. 1973. Mineralogy of selected
Florida Paleudults. Soil and Crop Sci. Soc. Florida Proc. 33:136-139.

6. Elzeftawy, Atef and R. S. Mansell. 1975. Hydraulic conductivity cal-
culations for unsaturated steady-state and transient-state flow in
sand. Soil Sci. Soc. Amer. Proc. 39:599-603.

7. Elzeftawy, Atef, R. S. Mansell, and H. M. Selim. 1976. Distributions
of water and herbicide in Lakeland sand during initial stages of in-
filtration. Soil Sci. (accepted for publication).

8. Hammond, L. C., V. W. Carlisle, and J. S. Rogers. 1971. Physical and
mineralogical characteristics of soil in SWAP experimental site at
Fort Pierce, Florida. Soil and Crop Sci. Soc. Florida Proc. 31:210-214.

9. Jones, G. C. 1967. Some observations on Florida's water research
needs, an economic appraisal. Publication No. 3, Institute of Food and
Agricultural Sciences, University of Florida, Gainesville. Pages 105-
114.

10. Mansell, R. S., H. M. Selim, and J. G. A. Fiskell. 1976. Simulated
transformations and transport of phosphorus in soil. Soil Sci.
(accepted for publication).

11. Mansell, R. S., H. M. Selim, P. Kanchanasut, J. M. Davidson, and J. G.
A. Fiskell. 1976. Experimental and simulated transport of phosphorus
through sandy soils. Water Resources Research (accepted for publica-
tion).

12. Mansell, R. S., L. W. Zelazny, L. C. Hammond, and H. M. Selim. 1975.
Nutrient distributions in a Spodosol during corn growth. Soil and
Crop Sci. Soc. Florida Proc. 34:24-29.











13. Perkins, HL F., H. J, Byrd, and F. F. Ritchie, Jr. 1973. Ultisols--
light-colored soils of the warm temperate forest lands. Pages 73-86 of
Soils o(f the Southern States and Puerto Rico, Bulletin No. 174, Agri-
i T irTl expriment S-tatonsof T ckeouthee States and Puerto Rico
aidl the Soil Conservation Service of the United States Department of
Agriculture.

14. Selim, IH. M. and R. S. Mansell. 1976. Analytical solution of the
equation for transport of reactive solutes through soils. Water Re-
sources Research 12:528-532.

15. Selim, H. M., R. S. Mansell, and Atef Elzeftawy. 1976. Distributions
of 2,4-D and water in sniiLdurDg infiltration and redistribution.
Soil Sci. 121:176-183.

16., Selim, H. M., R. S. Mansell, and L. W. Zelazny. 1976. Modeling re-
actions and transport of potassium in soils. Soil Sci. 22:77-84.

17. Steel, R. G. D. and J. H. Torrie. 1960. Principles and Procedures of
Statistics. McGraw-Hill Book Co., New York, N. Y., pages 106, 107
and 114.

18. Thomas, G. W. 1970. Soil and climatic factors which affect nutrient
mobility. In O.P. Engelstad (ed.) Nutrient Mobility in Soils: Assi-
mulation and Losses, Special Publication No. 4, Soil Sci. Soc. America
Inc., Madison, Wisconsin, p. 1-20.

19. Watanabe, F. W., and S. R. Olsen. 1965. Test of an ascorbic acid
method for determining phosphorus in water and NaHCO3 extracts from
soil. Soil Sci. Soc. Amer. Proc. 29:677-678.

20. Zelazny, L. W. and V. W. Carlisle. 1971. Mineralogy of Florida Aeric
Haplaquods. Soil and Crop Sci. Soc. Florida Proc. 31:161-165.









APPENDIX: Titles and abstracts of published papers resulting from this
research.

1. Saxena, G. K., R. S. Mansell, and C. C. Hortenstine. 1975. Drainage of
vertical columns of Lakeland sand. Soil Sci. 120:1-12.

A drainage experiment was conducted to determine the time-depen-
dence of soil-water retention during drainage of a quasi-uniform
column of Lakeland sand and similar columns to which phosphatic clay was
added as an amendment or-a-layer. Phosphatic-clay was observed to in-
crease and prolong water retention in the soil. Saturated hydraulic
conductivity of the column with soil amended with 5 percent phosphatic
clay in the surface 30 cm was 13 cm/hr as compared to 25 cm/hr for the
quasi-uniform column of Lakeland sand. Drainage from the column with
a 1-cm-thick layer of phosphatic clay and the column with a 2-cm-thick
layer of phosphatic clay aggregates was greatly restricted. The non-
aggregated phosphatic clay layer provided nearly constant impedence to
flow across the layer, whereas successive water desaturation of the
larger pores in the layer aggregates resulted in a time-dependent im-
pedence to water flow. At saturation the hydraulic conductivities of the
aggregated and nonaggregated clay layers were calculated to be 5.0 and
0.8 cm/hr, respectively. Maximum hydraulic head gradients across the
aggregated and nonaggregated clay layers were 4 and 13, respectively.
Values of unsaturated hydraulic conductivity calculated from soil-
water characteristic curves for the Lakeland sand and measured water-
saturated hydraulic conductivity were in good agreement with values
measured experimentally from the soil columns.

2. Elzeftawy, Atef and R. S. Mansell. 1975. Hydraulic conductivity cal-
culations for unsaturated steady-state and transient-state flow in sand.
Soil Sci. Soc. Amer. Proc. 39:599-603.

Using a method employed by Green and Corey (1971), hydraulic con-
ductivity was calculated as a function of water content for Lakeland
fine sand. A gamma ray transmission method for measuring soil water
content and a tensiometer-pressure transducer arrangement for measuring
soil water suction were also used to experimentally determine values of
hydraulic conductivity for a similar range of soil water contents in
undisturbed soil cores and hand-packed soil columns. Measured and cal-
culated values were in good agreement for steady flow.

During transient flow soil water content was observed to be a non-
unique function of suction for water desorption, but depended upon the
state of flow. Higher water contents were found at a given pressure
head during unsteady flow than during steady flow or static equilibrium
(zero flow). Graphs of water content versus soil water suction were
similar for cases of steady and no-flow conditions. For transient flow,
the soil water pressure depended upon the soil-water content and rate of
change of pressure head with time.

3. Mansell, R. S., L. W. Zelazny, L. C. Hammond, and H. M. Selim. 1975.
Nutrient distributions in a Spodosol during corn growth. Soil and
Crop Sci. Soc. Florida Proc. 34:24-29.

Distributions of K, P, N114-N, and N03-N in a tile-drained Wauchula






-47-


sand (Spodosol) were determined under corn cultivation in the spring of
1974. Grain yields of 6,346 and 7,907 kg/ha were obtained from plots
receiving 567 and 9,070 kg/ha of limestone, respectively. Nutrient
distributions in the soil solution phase were used to describe nutrient
movement with time in the soil profile. Movement of NO3-N, NH4-N, K,
and Cl proceeded downward in the profile during the first 43 days.
However, for times greater than 43 days downward movement of these nu-
trients did not exceed 50-70 cm depth. Restricted downward transport of
nutrients was attributed to the presence of a slowly-permeable B2t
horizon located 75 cm from the soil surface.

4. Selim, H. M. 1975. Water flow through a multilayer stratified hill-
side. Water -Rsures-esearch ll:949-957o

The objective of this study is to present a mathematical analysis
for steady state saturated flow through multilayer stratified hillsides
of semi-infinite depth. Two soil surface shapes were considered: a
constant soil surface slope and a surface of arbitrary shape. Potential
and stream functions were obtained for one-, two-, and three-layered
hillsides. The method of solution was based on the Gram-Schmidt
orthonormalization method. For two-layered hillsides the hydraulic
conductivities were KI:K2 = 1:10 and 10:1. For three-layered hill-
sides the hydraulic conductivities were Kl:K2:K3 = 1:10:1 and 10:1:10.
Flow nets, seepage velocities, and flow rates are presented. These
results are useful particularly with regard to subsurface flow, runoff,
erosion, and solute movement through sloping soils.

5. Selim, H. M. and R. S. Mansell. 1976. Analytical solution of the equa-
tion for transport of reactive solutes through soils. Water Resources
Research 12:528-532.

Mathematical solutions of the differential equation governing
reactive solute transport in a finite soil column were developed for
two specific cases: continuous solute input and pulse-type solute in-
put at the soil surface. These solutions incorporate reversible linear
adsorption as well as irreversible solute adsorption. The irreversible
adsorption was expressed by a sink/source term which either may be a
constant or may have a concentration-dependent form. The boundary con-
dition used across the surface (X = 0) was that of the third type,
which accounts for advection as well as dispersion. To illustrate the
significance of using the proper boundary conditions, comparisons were
made with two other mathematical solutions, one by Cleary and Adrian
(1973) and another by Lindstrom et al. (1967). We conclude that the
solution presented here is highly recommended for low flow velocities
or specifically for v0L/Ds values less than 20. For large pore velo-
cities, or specifically for voL/Ds* >20, all three solutions are in
agreement o

6. Selim, H, M., R. S. Mansell, and Atef Elzeftawy. 1976. Distributions
of 2,4-D and water in soil during infiltration and redistribution.
Soil Sci. 121:176-183.


A numerical model was developed to predict transient water and




-48-


reactive solute transport in water-unsaturated soil. An implicit-
explicit method of finite difference approximation was used to simul-
taneously solve the water and solute flow equations. The model was
used to calculate transport of water and 2,4-D (2,4-Dichlorophenoxy-
acetic acid) in a field profile of Lakeland fine sand during infiltra-
tion at constant intensities (2.08 and 4.10 cm/hr). Calculated values
of the soil-water distributions compared well to experimental deter-
minations, but agreement obtained between 2,4-D distributions was only
adequate. Simulated results were also obtained to study the effect of
irrigation intensity on water and herbicide transport during redis-
tribution following irrigation. It was found that for advanced stages
of redistribution solute movement became negligible and the maximum
herbicide concentration was located at the same depth in the soil pro-
file regardless of irrigation intensity.

7. Selim, H. M., R. S. Mansell, and L. W. Zelazny. 1976. Modeling reac-
tions and transport of potassium in soils. Soil Sci. 22:77-84.

A mathematical model was developed to describe potassium reac-
tions and transport in soils. Kinetic reactions were assumed to gov-
ern the transformation between solution, exchangeable, nonexchange-
able (secondary minerals), and primary mineral phases of potassium.
Simulated results were presented for two soils, a weakly sorbing soil
and a strongly sorbing soil. The effect of kinetic rate coefficients
upon transport and transformation of applied potassium was also inves-
tigated. The model is flexible and can be adapted to incorporate vari-
ous transformation mechanisms between the different phases of potassium.
Model validation with the aid of experimental data is needed to further
describe the fate of potassium in soil.

8. Elzeftawy, Atef, R. S. Mansell, and H. M. Selim. 1976. Distributions
of water and herbicide in Lakeland sand during initial stages of in-
filtration. Soil Sci. (accepted for publication).

Penetration depths and concentration distributions were determined
in columns of Lakeland sand for surface-applied 2,4-D herbicide during
initial stages of steady water infiltration. Water was applied with
application rates, R, of 2 and 4 cm/hr to soil with initial water con-
tents, 6i, of 0.20, 0.11 and 0.01 cm3/cm3. Increasing R resulted in
faster rates of advance for both the water wetting fronts and the
depths of 2,4-D peaks. Larger values of 0-, however, increased the
rate of advance for the wetting front but Aid not affect the location
for 2,4-D peaks. Penetration of 2,4-D peaks per unit of water applied
to the soil was observed to be independent of both R and 6i during ini-
tial stages of water infiltration. Distributions of 2,4-D concentra-
tion as calculated from water and solute transport equations were ob-
served to lag slightly behind experimentally measured distributions.

9. Mansell, R. S., H. M. Selim, and J. G. A. Fiskell. 1976. -Simulated
transformations and transport of phosphorus in soil. Soil Sci. (accepted
for publication).






-49-


A mechanistic, multistep model was developed using chemical kin-
etics and mass transport theory to describe transformations and move-
ment of orthophosphate in soil. Soil phosphorus was assumed to occur
simultaneously in any of four primary phases: water-soluble, physi-
cally adsorbed, immobilized, and precipitated. Kinetic reactions which
control the transformation of phosphorus between any two of the four
phases were considered to be reversible and of Nth order. A range of
values for the reaction rate coefficients were used inthe model to
describe the transport of applied phosphorus in the solution phase of
the soil profile.

10. Mansell, R. S., H, M. Selim, P. Kanchanasut, J. M. Davidson, and J. G.
A. Fiskell. 1976. Experimental and simulated transport of phosphorus
through sandy soils. Water Resources Research (accepted for publica-
tion).

Reversible equilibrium adsorption-desorption relationships were
inadequate for describing the transport of orthophosphate through water-
saturated and unsaturated cores from surface (AI) and subsurface (A2)
horizons of Oldsmar fine sand (a Spodosol). Using a kinetic model
with nonlinear reversible adsorption-desorption improved descriptions
of phosphorus transport through these soils. Phosphorus effluent con-
centrations were described best using an irreversible sink for chemical
immobilization or precipitation with a nonlinear reversible, kinetic
adsorption-desorption equation.

11 Selim, H. M. and R. S. Mansell. 1974. Transient one-dimensional and
simultaneous solute and water flow in soils. Program No. 360 D-17.4.003,
SIl'.]' Program Library Agency, Triangle Universities Computation Center,
Research Triangle Park, North Carolina.

A computer program has been developed for the problem of solute
and water movement in unsaturated soils or porous media under transient
flow conditions, The two nonlinear partial differential equations
governing the solute and water flow are solved simultaneously for the
water content and solute concentration at any specified time and lo-
cation as desired. The initial conditions used are uniform salt and
water content distributions at time t=0. The boundary conditions at
the soil surface are water flux and constant salt concentration condi-
tions. The method of solution is a numerical one which utilizes the
explicit-implicit finite difference technique.

The computer program is written in FORTRAN language and consists
of a source program, eleven subprograms, and an input data section.
An important feature of the program is that incremental distance and
time steps are adjusted automatically to satisfy stability and conver-
gence criteria for the water and solute finite difference criteria. A
second feature is that the number of nodal points are automatically
calculated from the length of the flow region. A third feature of the
program is that output data of water content, water flux, solute con-
centration, and solute flux in the flow region are provided at speci-
fied times as desired.




-50-


12. Elzeftawy, Atef A. 1974. Water and solute transport in Lakeland fine
sand. Ph.D. Dissertation, Soil Science Department, University of
Florida, Gainesville.

The objective of this study was to investigate effects of three
water supply rates--2, 4, and 8 cm/hr -- and three initial soil water
contents -- 1.2, 10.9, and 20.2% by volume -- upon the simultaneous
transport of water and solutes -- 2,4-D herbicide and chloride -- in
vertical columns of Lakeland fine sand. Columns were prepared by
packing air-dry soil into cylinders 7.6-cm diameter and 107 cm long.
A specific volume of aqueous solution containing 57.9 ppm chloride and
5 ppm 2,4-D was introduced and displaced through each column,. Gammuia-
ray attenuation and pressure-transducer-tensiometers were used to
precisely monitor soil-water content and pressure distributions with
time. Soil solution was extracted at selected depth intervals along
the soil columns and extracted samples were then analyzed for 2,4-D
and chloride content.

Depths to which chloride and 2,4-D moved for a given quantity of
water infiltrated into the surface of the soil was found to depend
upon the surface water flux. Increasing water application rates re-
sulted in an increased water content in surface soil and in shallower
displacement of chloride and 2,4-D for equal quantities of accumula-
tive infiltration. For a given quantity of water infiltrated, initial
soil-water content did not influence depths of chloride or 2,4-D trans-
port. Adsorption, caused 2,4-D distributions to lag behind those for
chloride for all experiments.




Full Text

PAGE 1

r ( WATER IiRESOURCES researc center Publication No. 38 Movement of Fertilizer and Herbicide Through Irrigated Sands By R.S. Mansell, F.M. Rhoads, L.C. Hammond, H.M. Selim, W.B. Wheeler and L. W; Zelazny Department of Soil Science, IF AS University of Florida Gainesville UNIVERSITY OF FLORIDA

PAGE 3

TABLE OF CONTENTS Page Title i Table of Contents ii Acknowledgments .. iii Abstract iv Chapter 1: Introduction 1 Chapter 2: Objectives 2 Chapter 3: Field Experiments A. Description of Soils 3 B. Experiments on Wauchula sand, a Spodoso1 3 C. Experiments on Troup sand, an U1tiso1 22 Chapter 4: l\fathematica1 l\bde1s 29 A. Water and 2,4-D MOdels 29 B. Phosphorus l\bde1s 30 C. Potassium Mbde1s 33 Chapter 5: Summary and Conclusions 42 Literature Cited 44 Appendix: Titles and abstracts of published papers resulting from this research 46

PAGE 5

iii ACKNOWLEffiEMENTS Contributions by the following technical staff of the Soil Science Department provided valuable assistance in performing this research: Richard McCurdy, Bill Porthier, and Ron Jessup. We also acknowledge Chris Young and Ann Barry for typing this report as well as manuscripts for articles published in scientific journals. Special thanks are extended to each of these individuals.

PAGE 6

iv ABSTRACT The simultaneous movement of water and selected agrichemicals (fertilizer nutrients and herbicide) through sandy soils is of particular impor tance to the efficient use of fertilizers and irrigation water by agricultural crops. Efficient use of fertilizers and herbicides applied to Florida's sandy soils is desirable for maiptaining optimum growth of plants and for minimizing groundwater contamination. Laboratory and field experiments as well as mathematical models were used to study water and solute (potassium and phosphorus nutrients and 2,4-D herbicide) transport in two representative Florida soils: Wauchula sand and Troup sand. In an irrigated and fertilized corn experiment, grain yields and efficiency of water use were observed to be mutually related to both the irrigation and the fertilizer application treatments. Leaching of applied nutrients andirrigation water from the soil "rooting zone" resulted in decreased water use efficiency in these soils. Mathematical models were developed and used to sinrulate transport and chemical-physical reactions for potassium, phosphorus and 2 herbicide in these soils. Reactions such as adsorption -desorption, chemical precipitation and immobilization (fixed) greatly influenced the movement and thus potential leach ing of these solutes through the soil.

PAGE 7

CHAPTER 1: INTRODUCTION Fertilizers, herbicides, and irrigation water are commonly applied to Florida! 5 sandy soils for the purpose of maintaining high levels of food production from soil-water-plant systems. These soils generally have capacities to temporarily store applied water and solutes, and thus proper management of agrichemical and water applications is needed to mini mize leaching losses of nutrients and herbicides from the "rooting zone" of the soil profile. Excessive leaching, which may occur during periods of intense rainfall greatly increases the potential for contamination of the underlying groundwater with agrichemicals. Efficient usage of fertilizers, herbicides, and water therefore provides two very important beneficial results: (1) optimum plant growth and yields, and (2) minimal pollution of groundwater resources. Management of crops growing in these sands to produce optimum crop yields requires irrigation during periods of drought and frequent applica tions of fertilizer during the growing season. Although the average annual rainfall for the state ranges from near 52 inches on the central and northern peninsula to nearly 65 inches in the panhandle west of Tallahassee (Butson and Prine, 1968), severe drought commonly occurs during the spring growing season followed by heavy rains during the summer when the state receives approximately 60% (Jones, 1967) of its rainfall. The uneven rainfall distribution coupled with low retention capacities of the surface soil of the sands for water and solutes result in relatively high leaching losses of applied agricultural chemicals. Also crop yields may be de creased by periods of drought as a result from soil-water deficiency or high osmotic pressures of the soil solution due to improper timing of ferti lization. Water, fertilizers, and herbicides are applied to the sands to create a favorable plant root environment with optimum supplies of available water and nutrients. The herbicides are used to destroy weeds and undesirable grasses which compete with the crop plants for water, nutrients, and light. Failure to maintain minimum threshold levels of water and fertilizer nu trients in the root zone during critical periods of plant growth can result in decreased yields. However, levels of water and nutrients in the root zone in excess of plant requirements may result in waste of resources, possible detrimental effects on crop growth, and potential contamination of ground water with herbicide or fertilizer solutes. To insure optimum crop growth on sands, irrigation during periods of low rainfall is needed to maintain low soil-water suction (0 to -200 em of water). Consequently, the profile will of necessity be high in soil-water content and undergo some drainage or redistribution of water at all times. Since the hydraulic conductivity of the deep sands increases greatly with soil-water content, heavy irrigation or rainfall imposed upon relatively moist soil will result in much of the infiltrating water being lost from the root zone by drainage. The simultaneous transport of water and soluble chemicals through soil has imp0Ytant implications with regards to the efficient management of fertilizer, herbicide, and irrigation water applied to agricultural crops growing in sandy soils. Transport of water and nutrients through sandy soils is particularly important to phenomena such as plant uptake of nu-

PAGE 8

-2-trients and water which tends to increase fertilizer and water use efficiencies. Thus, improper water management combined with inefficientapplication of herbicides and fertilizer nutrients to sands may result in leaching loss of some of the chemicals from the "rooting zone" of the soil profile. Thus potential contamination of groundwater may result as leached herbicides and nutrients move deeper into the soil. Fortunately,processes such as adsorption-desorption, ionic exchange, microbiological transformations, chemical interactions, and uptake by plant roots tend to decrease the leaching loss of herbicides and nutrients from soils. Therefore ,detailed-knowledge ef simul taneeus movement of water and solutes in Florida's sandy soils is critically needed to insure efficient use of water, fertilizer, and herbicide resources for crop production without unqesirable contamination of the underlying groundwater. CHAPTER 2: OBJECTIVES Specific objectives for this project were as follows: (1) To determine rates of movement of soil-applied herbicide and fertilizer solutes in irrigated agricultural sands and to determine leaching losses of these chemicals from the "rooting zone" of the soil profile; (2) To determine the influence of limestone application to acid, sandy soils upon movement anq leaching losses of soil-applied fertilizer solutes from the "rooting zone;" (3) To determine the of adsorption and desorption processes upon rates of movement of 2,4-D herbicide, potassium fertilizer and orthophosphate fertilizer with water through irrigated sandy soils; and (4) To utilize existing mathematical models to describe the simul taneous movement of water, selected herbicides, and fertilizer nutrients in sandy soils during periods of water infiltration and redistribution.

PAGE 9

-3-CHAPTER 3: FIELD EXPERIMENTS A. of Soils Most Florida soils can be classified into either of four orders (Fig. 1): Spodosols, Ultisols, Entisols, and Histosols. Representatives of two of these soil orders were selected for the location of field experiments to detennine the s imul taneous movement of water and agricultural chemicals in the "rooting zone" of irrigated and fertilized corn. One experiment was located at the University of Florida Beef Research Unit near Gainesville on a Wauchula sand which is a Spodosol (family: sandy-over-loamy, siliceous and hyperthermic; subgroup: UI tic Hap I aquods ) AI though the Wauchula soil is a Spodosol it appears in an area which is broadly characterized by the presence of Ultisols. l\nother experiment located at trre Aglicul-tural Research and Education Center of the University of Florida near Quincy on a Troup sand which is an Ultisol (family: loamy, siliceous, and thennic; subgroup: Grossarenic Paleudults). Spodosols and associated flatwood soils represent the most extensive order (Fig. 1) of Florida soils (Zelazny and Carlisle, 1971) and account for one fourth of the total land area. Most of these soils occur on nearly level to gently sloping landscapes with a generally shallow ground water table which fluctuates near the soil surface during periods of high rainfall (Brasfield et ale 1973). Spodosols are characterized by the presence ofa subsurface spodic horizon which is an accumulation of organic matter with varying amounts of all@inum and iron. Brasfield et ale (1973) state that the spodic horizon has a high ionic exchange capacity, large specific surface area, high water retention, and high exchangeable acidity. The spodic horizon commonly occurs less than 75 em beneath the soil surface and is overlain by sandy A2 eluvial and Al surface horizons. These soils may be classed as strongly acid sands with low fertility and base saturation. Although the spodic horizon is generally slowly permeable, the overlying Al and A2 trrulsmit water rapidly. Ultiso1s are the most extensive soils (Fig. 1) of northwest Florida (Carlisle llild Zelazny, 1973). These soils are also located in northcentral Florida. Perkins et al. (1973) describe these soils as having B horizons that contain rul appreciable amount of trllilslocated silicate clay but few bases. Sandy or loamy surface horizons generally are underlain by horizons with loamy or clayey texture. Ultisols are typically acid, relatively infertile, ruld have a low base saturation 35%) within about 2 meters of the soil surface. -B. Experiments on WAUCHULA SAND, a Spodosol Nutrients applied as fertilizers to crops growing on acid, sllildy soils in Florida's humid climate are susceptable to partial leaching loss from the "rooting zone" of the soil. Fertilizers are typically applied to the soil surface as dry solid materials which eventually undergo dissolution in infiltrating rainwater (or irrigation water). Thus with time a portion of the nutrients become solutes in the soil solution. As the soil solution moves downward in the soil profile, nutrient solutes may be removed from

PAGE 10

-4-solution by uptake through plant roots, sorption onto soil particles, chemical precipitation, and biological degradation (denitrification). As the solution moves further from the soil surface, nutrient solutes are subject to loss by drainage in tile-drained soil and by deep seepage to the groundwater in well-drained soil. Highly mobile nutrients such as N03 (Thomas, 1970) are particularly susceptible to leaching from the soil; whereas the amount of a reactive solute such as P moving in the soil solution is usually very low relative to the total quantity of P in the soil. The mobility of potassium in soil is usually intermediate to that for N03-N and orthophos_ ___ _________ ____ During the spring of 1974 a field experiment was established on a subsurface-drained Wauchula sand (a Spodosol) to determine in situ distributions of NH4-N, N03 -N, P04-P, and K in the solution phaseoTThe soil profile during the growth of corn. A subsurface system of parallel (lO-em clay tile) drains 75 em deep and spaced 6 ill apart provided drainage for the soil profile. Nutrient distributions with soil depth were determined in soil receiving two levels of dolomite limestone application: a low level, 567 kg/ha and a high level. 9.070 kg/ha. The experiment was located at the University of Florida Beef Research Unit, approximately 10 miles northeast of Gainesville. A randomized block design with four blocks and two limestone application levels was used. Each block contained six replications to give an overall replication of 24 for each lime treatment. Each plot, 3.6 m wide and 9 m long, contained four rows of corn. A commercial fertilizer 4-7-16 (N-P20S-KZO) with micronutrients was applied broadcast to the soil surface on March 26 at the rate of 2841 kg/ha which contained 113.6, 99.2 and 377.3 kg/ha each of N, P, and K, respectively. MCNair 73011 hybrid seed corn was planted on April 1 (day 0) in rows 90 em apart and at IS-em intervals within each row. On May 14 (day 43) 250 kg/ha of N as NH4N03 was applied in a narrow (5 em) band near each corn row. The actual amount of N applied in the band and adjusted for the band width was 4,570 kg/ha. Soil solution samplers and soil water tensiometers were installed in the middle of six plots for each of the two limestone applications within one block. Solution samplers composed of 6-cm diameter porous ceramic cups (bubbling pressure of 1 bar) and attached to the bottom of plastic pipes were installed in a corn row adjacent to the tensiometer installations. The porous cups were located at 30, 60, 90, 120. and 150-em depths, and samples of soil solution were removed periodically. Approximately 10 to 50 m] of solution were collected from each sampler, and samples were analyzed tor K, NH4-N, and P04-P, Methods of analyses were flame photometry for K, specific ion electrodes for NH4-N and N03-N, and tile ascorbic acid technique as described by Watanabe and Olsen (1965) for P. Tensiometers with 2.3-cm diameter porous ceramic cups (1 bar bubbling pressure) and mercury manometers were installed adjacent to corn rows at depths of 15, 30, 60, 90, 120, and 150 cm. Manometer readings provided distributions of soil water suction, h, and total hydraulic head, H. Soil water content-suction characteristic curves for undisturbed soil cores were determined in the laboratory by the method of Hammond, Carlisle, and Rogers (1971). By use of these drainage characteristic curves, values of h for a

PAGE 11

given depth were converted to volumetric water contents, 8. This conver sion is nonnally adequate for conditions of soil water desorption, but at best provides an approximation for water sorption due to possible error caused by soil water hysteresis. Soil water content-suction and hydraulic conductivity-water content curves (Mansell, et aL 1975) for soil materials taken from several depths showed that the hydrologic properties among the Ap (0-25 em), AZ (25-75 an), B:/t (45-75 em) 7 BZt (75-145 em, and C T:> 145 on) horizons were greatly For practical purposes water movement through the soil profile was considered to be that for a soil having a pervious zone overlying a slowly pervious zone. Soil above 45 on CAp and A2) had saturated conductivities at least lO-fold greater than that for soil between 45 and 150 em depths (BZb. B2t and C). Emergence of corn seedlings occurred on April 7 (day 8), tasseling began on Jillle 14 (day 74), and development of ears began on JlUle 24 (day 84). Growth curves showing relative heights of corn plants versus time for 567 and 9,070 kg/ha of dolomite limestone applied to the soil are presented in Fig, 2. MaximuJn plant heights (Lf) for both low and high lime treatments were 275 em. Growth curves were sigmoid in shape and differences between the treatments were not appreciable. Plant heights reached 95% of the mCL'Cimum after 80 days. Grain and stover were harvested September 5 (day 157) and their ovendry weights were determined. Even though appreciable differences of corn growth curves did not occur between limestone treatments, average grain yields for high 1 lines tone application were 25% greater than for the low limestone application. Average yields were 6,346 and 7,907 kg/ha for low and high limestone treatments. During the 157 days between planting and harvest of the corn crop, the experimental site received a total of 76,5 em of rainfall. Since only 3.4 enl of rainfall was received during the first 30 days after planting, 3.5 em of\vater was applied to the experimental site on day 29 by sprinkler irrigation. During the first 30 days, depths to the water table exceeded 100 em, but for the remainder of the growing season the water table fluctuated about the 75 em depth. Thus, one month following planting of the corn, conditions of water saturation and poor aeration occurred in the soil beneath 90 em depth. For soil depths above 60 em, water pressure heads were always negative indicating water lli1saturation and favorable aeration conditions .. Negative pressure heads at 15 cm depth fluctuated widely with time relative to that for the subsoil. Extremes in soil water suction head of 30 and 200 cm of water were observed in the soil at 15 em depth and these heads correspond to volumetric water contents of 28 and 8%, respectively. Relative to underlying soil horizons, water contents in the top 15 em of soil increased rapidly during rainfall events and then subsequently decreased during the first 3 to 5 days of the post-infiltration period. Distributions of K, NH4-N, N03-N, and P04-P nutrients in the soil solution as a function of depth and time are presented in Figs. 3-6 for both lime treatments. Each data point represents the mass of a specific nutrient in the solution phase per unit volume of the bulk soil, as determined from

PAGE 12

-6-the nutrient concentrations of soil solution) in the soil solution and volumetric water contents (em3 of soil solution per em3 of bulk soil). Furthermore, each data point represents an average from six replicate plots. Distributions of nutrients in the solution phase of Wauchula soil provide a means for describing movement of the various solutes downward through the profile. Although 80 em of irrigation and rain water occurred during the first 150 days after fertilization, nutrients did not move appreciably to depths below approximately 70 cm. Nutrient contents in the soil solution for profile depths greater than 70 cm showed only smull changes with time -throughout the entire growing-season of-the com crDp. Initially, downward movement of the zone of maximum concentration occurred from the soil surface with time to day 43. Following day 43, this zone showed slow penetration of these solutes beyond 50-70 em depths. This apparent slow transport of solutes was due to the presence of a slowly permeable layer (B 2t horizon) at a relatively shallow depth (75 cm). Under such conditions, flow of water as well as nutrients in the B2h horizon above the impermeable B 2t layer occurred predominantly in the lateral direction rather than vertically downward. Obviously much deeper penetration of the solutes would be expected in deep, lU1iform, well-drained soil profiles, resulting in lTeJ10 rapid leaching loss of nutrients than observed for the Spodoso1 reported here. Sud1 nutrient losses are minimized when a shallow water table is present; however, the solute concentrations become more dilute depending upon depth to an impermeable layer. Also, one would expect nutrient losses in Spodosols to be closely related to the frequency of high intensity rainfall. Such rainfall patterns would cause appreciable lateral flow of water and solutes to drain tiles. Therefore, we conclude that the presence of a shallow, slowly permeable layer such as the B 2t horizon in Wauchula sand is beneficial in minimizing rapid leaching losses of nutrients from a tile-drained soil. During 1975 a second corn experiment was performed at the Beef Research Unit, but three irrigation treatments--no irrigation, daily irrigation (0.64 em/day), and controlled irrigation-were imposed. Controlled irrigation was maintained by irrigating with 1.3 em of water when the soil water suction at 15 em depth exceeded 100 em of water. A total of 2.6 em of water was applied during the entire season. Line sources (plastic tubing) with discrete emitter holes were placed adjacent to corn rows to provide irrigation by the trickle method. For the daily irrigated plots, irrigation was provided between April 2 (day 8) and 30 (day 70). Water was pumped at a constant hydraulic head through the emitters in the plastic tubing to prov:ide a constant volume discharge of water with time for lli1i t length o the irrigation tube. The trickle concept offers advantages of increased water conservation and improved control of soil water matric potential in the root zone. Approximately 2000 to 3000 acres in Florida are currently irrigated (private communication with Dalton Harrison, Ex tension Irrigation Specialist, Department of Agricultural Engineering, University of Florida) by trickle irrigation. A disadvantage of the concept from a technological standpoint is that the irrigation water must illlder go intensive filtering of Fe, S, particulate matter, and algae to prevent premature clogging of trickle nozzles. Theoretical and experimental analysis of transient infiltration from a trickle source irrigation system have been presented by Brandt et al (1971) and Bresler et al. (1971).

PAGE 13

-7-lltJO fertilizer application treatments--conventional and progrannned application--were also established in the 1975 experiment. Prior to planting of COI11, 4,235 kg/ha of a 5-10-15 (percentages for N-P20S-K20) fertilizer was applied broadcast to the soil surface for the conventional applica tiol1. The com was planted on March 26 (day 0), and 672 kg/ha of NH4N03 was to the soil on l'v1ay 28 (day 68). The programmed treatment was established by applying 5% of the total P and. K fertilizer kg/1m of 5,10-15 fertilizer plus 622 kg/ha of NH4N03) in a h;cmd near each corn row at the time of seedling emergence another 10% tlvO weeks later, 10% at 4 weeks, 15% at 6 weeks, and 20% each at eight, ten, and weeks. The tasseling stage for the CODI began on l'v1ay 30 (day 70) and the corn was harvested on August 11 (day lS2). The effect of daily irrigation upon the soil water slJction head at 15 on beneath the soil surface can be seen in Fig. 7 for the programmed fertilization. During the period when irrigation was applied daily (day 8 to 70) the soil water suction head in the irrigated plot was maintained between (15% water content) and 60 (10% water content) em of water; whereas the suction head the plot that received no irrigation fluctuated from as low as 45 (14% water content) to as high as 380 (less than 6% water content) cm of water. During days when rainfall occurred (Table 1) soil water suction for the unirrigated plot was greatly decreased but increased sharply during periods of minimal rainfall. After day 70 soil water suction in both plots tended to decrease during periods of rainfall and to increase during the days after a rainfall. The increases in soil water suction in both plots resulted from soil water being removed by evaporation at the soil surface, water uptake by plant roots (transpiration) and down ward water movement by soil water redistribution (gravitational drainage). As the effect of trickle irrigation upon soil water suction was less evident for soil depths of 30 (Fig. 8) and 60 (Fig. 9) an than at the shallower 15 on (Fig. 7) depth. This effect can also be seen in Figs. 10 and 11 where distributions of soil water pressure head with soil depth are presented for the unirrigated and irrigated plots during a 26-day period of lilinimal rainfall. In the plot receiv:lng no irrigation, soil at the 15 an depth underwent the greatest drying in the soil profile. Soil at the 30 em depth also exhibited drying but to a lesser extent than at the 15 an depth. At day 20 the distributions of soil water pressure head with soil depth were similar for both the irrigated and lll1irrigated plots. Water tables (pressure head ::= 0) were located at 70 and 65 cm depths, respectively, irrigated and lll1irrigated soH profiles, indicating conditions for water saturation for soil depths greater than these depths and unsaturation at shallowe:r depths@ Vertical gradients for hydraUlic head, with depth were approximately -00011 in both irrigated and unirrigated soil for all depths greater than 60 on. This small negative gradient of hydraulic head indicates that water is slowly draining from the soil profile. As time passed from day 20 to day 46 the gradients of H in the top 30 an of the unirrigated soil reflected a net upward water movement due to evaporation and water uptake by plant roots. For example at day 30 the gradient of H between IS and 30 C.TIl depths was +0.933; whereas for the irrigated soil that value

PAGE 14

-8-was -1.067 indicating a net downward flow of water due to the daily addition of 0.64 em of water. Thus the distributions of soil water pressure head shown in Figs. 10 and 11 show that the daily application of 0.64 em of water by trickle irrigation was more than sufficient to compensate for water loss in the top 60 cm of soil due to evaporation, transpiration, and drainage. The amount of 0.64 em of water per day was selected because it represents the maximum potential removal of water from the soil by evapotran-spiration for the summer months. Ave-rage.--gra-iny-ields---for-co-m--grown three irrigation and two fertilizer management programs are presented-in Table 2. For either the conventional or programmed fertilization, daily irrigation resulted in higher yields of grain, but the percentage yield increases due to daily irrigation were relatively small (21.7% for conventional fertilization and 20.5% for progrannned fertilization). For conventional fertilization, grain yields for controlled irrigation were only slightly higher than for corn irrigated daily. The grain yield reported for corn receiving programmed fertilization and controlled irrigation was lIDexplainably lower than other treatments IrTigation use efficiencies (calculated by subtracting grain yields for lIDirrigated corn from yields for corn receiving daily irrigation and controlled irrigation and dividing the results by 39.7 ern of water for the daily irrigation treatment and 2.6 cm for the controlled irrigation) were 31.1 and 510.4 kg/ha/cffi for conventionally fertilized corn receiving daily and controlled irrigation. These values in conjunction with soil water suction and grain yield data indicate that the controlled irrigation provided much more efficient use of the water applied and that the daily irrigation resulted in water application to the soil in excess of that actually needed to produce economically profitable grain yields in Wauchula sand. Even the grain yields for unirrigated corn could be considered to be economically profitable for the Wauchula sand. The relatively large yields for lIDirrigated corn was partially attributable to the presence of a watersaturated zone in the lower portion of the soil profile during the growing season for the corn. For the daily irrigated corn, irrigation use efficiencies were not greatly influenced by the method of fertilizer application (31.1 and 33.0 kg/ha/ern for conventional and programmed fertilization, respectively). That result is also supported by the observation from the 1974 experiment which showed that leaching fertilizer nutrients in Wauchula sand was primarily limited to the top 50-70 ern of soil.

PAGE 15

-9(\) 0 > (/) 0::: (\) 0 C .-J (f) cu LL (/) 0 (/) (9 0 (/) (/) 0 LL (/) 0 OUl 0 >. -0 (/) ._ +' o ......, (/) U (/) a. c .c (f)::)WI .-J 0 0 (f) 1: Map showing location of major soils of Florida

PAGE 16

-10d (l) ..c -+0) E ..Y:. 0 f' 0 to 1"--0 l() 0 0 Of x E 0 u ...... +' 0> ('1 ..... L Q) L U 0) C C Q) \J Q) Q)-?
PAGE 17

-11Fi,!"TUre 3: Distributions of potassium with depth in Wauchula sand treated with 567 and 9,070 kg/ha of limestone during 1974.

PAGE 18

o I.f) -12-N H4 -N (}Jg/cm3 soi I) 567 kg/ha Lime 1" : ,,, '.'. : :' / ./ / ,// 29 days 30 43 116 9070 kg/ha Lime Figure 4: Distributions of NH4-N with depth in Wauchula sand treated with 567 and 9,070 kg/ha of limestone during 1974.

PAGE 19

-----------------,--.-----------13-Nitrate-N CjJg/cm3so i I) __ -=2rO __ E u ........ ..c -t-J a. Q) o o If) "l'"o L() r567 kg/ha Lime I ....... .... .. .. // /" /. ... I ,/. ", .... : I : if' : : I I \ : 29 days 30 43 1 \ ,/ '\ 116 9070 kg/ha Lime FihTUre 5: Distributions of N03-N with depth in Wauchula sand treated with 567 and 9,070 kg/ha of limestone during 1974.

PAGE 20

-14P04 PC jJg/cr-r-r3soi I) __ __ ____ 567 kg/ha Lime I I r \ I V I \ \ \ '. I r I \. / / --29 days ------30 --43 -_ ....... __ ..... 116 9070 kg/ha Lime Figure 6: Distributions of P04 -p with depth in Wauchula sand treated with 567 awl 9,070 kg/ha of limestone during 1974.

PAGE 21

-15 -t--J o S "t-o C o +-' U 2 Soil Depth: 15cm ... _-" '----", I rr i gated Unirrigated Period of Irrigation o (j) o 40 80 120 Days Figure 7: Soil water suction head at 15 on soil depth in Wauchula sand during the 1975 growing season for daily irrigated and unirrigated corn.

PAGE 22

-16......, o 5---1------'+o E c o -t-J g102 (/) -L (l) +-' o '-o (j) 10 o Figure 8: Period of Irrigation t --.-. I 40 80 120 Days Soil water suction head at 30 on soil depth in Wauchula sand during the 1975 growing season for daily irrigated and lIDirrigated corn.

PAGE 23

-1./.OJ j-J o E c o Soi I I \. \I I I ,." ,-'" Depth: 60cm ,4 I I \ , , f-J 1 II \, o tf) Irrigated Un i rr i gated Period of Irrigation I I o 40 80 120 Days Figure 9: Soil water suction head at 60 em soil depth in Wauchula sand during the 1975 growing season for daily irrigated and unirrigated corn.

PAGE 24

! Soi I Water Pressu re(cm of wa ter) o -300 -200 -100 0 : \ \ 26' I ',\, \c \ I ", '\ '"'' I I 4 I 3 30 : Day 20 I I =0 -(0 J ..J ::l. 1>0 )(J) ) Day 46 Unirrigated () l Figure 10: Distributions of soil water pressure head with soil depth in Wauchula o sand during a selected 26-day period of the growing season for unN irrigated corn in 1975. I I I I I +100 I I-' 00 I

PAGE 25

0 0 't""'+ ........ -1> ..J J ) :J, 1> :JO J) r ..... L L ())O r-JO oc\( > > o f) o (Y) 06 0 (Y) 0 N 0 0 o (") -19<0 "\t 0 0 0 2E au .) L LO '-'" ......... o 0 ) (J) (W:)Lf+daO o C\J 't""'Figure 11: Distributions of soil water pressure head with soil depth in Wauchula sand during a selected 26-day period of the growing season for daily irrigated corn in 1975. o LO

PAGE 26

-20Table 1: Rainfall at the Beef Research Unit near Gainesville during the spring and summer of 1975. Day Rainfall Day Rainfall Day Rainfall (em) (em) (ern) *0-8 0.0 72 1.8 96 1.0 9 0.3 73 0.1 97-99 0.0 10-14 0.0 74-77 0.0 100 0.3 15-26 10.7 78 0.4 101-105 0.0 27-40 0.0 79 0.1 106-108 0.8 41 0.1 80 1.4 109 0.6 42-43 0.0 81 0.0 110 0.2 44 0.2 82 0.2 111 0.0 45-49 0.0 83 0.1 112 2.4 50 0.8 84 0.0 113 2.0 51-53 0.0 85 0.2 114 0.6 54-60 2.7 86 0.0 115 0.2 61-64 0.0 87 0.3 116-118 0.0 65 0.4 88-91 0.0 119 2.7 66 1.3 92 0 0 9 120 0.2 67 0.0 93 0.1 121-123 0.0 68 601 94 1.1 124-131 3.6 69 0.7 95 0.0 132 0.6 70-71 0.0 133 2.4 Total 47.b Day 0 was March 26

PAGE 27

-21Table 2: Corn grain yields on Wauchula sand under three irrigation and two fertilizer management programs, Gainesville, 1975. Grain Yield, kg/hal Fertilizer No Dally management irrigation irrigation2 IrrIgatIon Conventional 5708 ac 6949 ah 7035 ab Programmed 6366 abc 7672 b 5206 c lLSDO 05 = 1612 kg/ha. Grain yields followed by the same alphabetical letter (a,b or c) or combination of letters did not differ from each other. LSDO.05 is the least significant difference at a 5% probability level (Steel and Torrie 1960). 2Application of 0.64 em/day from day 8 to day 70; total application, 39.7 cm. 3Irrigation with 1.3 em of water when soil water suction at 15 cm depth exceeded 120 em of water; total application, 2.6 em.

PAGE 28

-22-C. EXPERIMENfS ON TROUP SAND, AN UL TI SOL During 1975 a field experiment of irrigated corn was established on Troup sand (an Ultisol) at the University of Florida Agricultural Research and Education Center near Quincy. The purpose of the experiment was to determine distributions of soil water pressure head with depth during the growth of irrigated and unirrigated corn and to determine grain yields for corn grown under 3 irrigation and 2 fertilizer management programs. The :tgp lS })utJia se11J:ial-__ ly a Wliform sand between 15 to 150 an depths with a sandy clay material located at 150 em. This soil was originally classified. as Lakeland sand, but the profile characteristics more closely resemble Troup sand. The names Lakeland sand and TrO'l;lp sand are used to represent the same soil in this report. Graphs of hydraulic conductivity versus soil water content for selected horizons of this soil are presented elsewhere (Elzeftawy and Mansell, 1975). Coker 77 hybrid corn was planted on March 24 (day 0) in rows 90 on apart and with 15 an spacing between seed in each row. Three soil-water management treatments were maintained: (a) no irrigation, (b) daily irrigation of 0.64 cm of water applied through plastic trickle irrigation tubing placed near each corn row, and (c) 1.30 an of water applied with trickle tubing when the soil water suction head at 15 an depth exceeded 100 em of water. Two fertilizer treatments were also used: (a) conventional fertilization where 2240 kg/ha of a (percentages of N P-K) fertilizer was applied prior to planting of the corn and 670 kg/ha of NH4N03 was later applied in two equal applications (May 14 and 17) and (b) programmed fertilization of 2240 kg/ha of 5-4.4-12.4 fertilizer and 670 kg/ha NH4N03 applied in several split applications. In the programmed fertilization treatment 5% of the N, P and K was applied at the time of seedling emergence, 5% two weeks later, 10% at four weeks, and 20% applications at 6, 8, 10 and 12 weeks. Tensiometers were installed at 15, 30, 60, 90, 120, and 150 em depths beneath the soil surface and adjacent to the cornrrows. Soil water pressure head measurements were obtained from the tensiometers at selected times during the growth of the corn. The grain was harvested on August 8 (day 141). Major fluctuations in soil-water suction head with time at the 15 and 30 an soil depths were observed in the unirrigated plots (Fig. 12). At the 15 cm depth suction heads ranged from as low as 100 em of water immediately following a rain to as high as 800 em of water several days following rain. Major periods during which the surface soil underwent extensive drying were observed during 30-40, 55-59, 70-80, and 86-100 days after planting of the corn. Only minor fluctuations in suction head were observed deeper in the soil profile. For example,at the 90 an soil depth the suction head ranged from near 0 to slightly above 100 on of water between 25 and 100 days after planting of the corn. Between days 25 and 37, positive soil water pressure heads were temporarily observed above the slowly penncable sandy clay 1 ;lye1' at l5b em depth. Heavy rains in the early part of the sC,lson were rcspons i -ble for the positive soil water pressure at 150 on depth.

PAGE 29

-23])aily irrigation of 0.64 on of water resulted in soil water suction he;lds I ess 1 no Ull at all depths (Fi g. 13) in the profile during the ,\',nl\viilg S(';ISOI1 Cor tilt:.' corn. A total of S6 on of water was applied as i during the grow.ing season 1'01' the com. Major fluctuations in stlet ion head as seen in the lmirrigated plot were not observed at the 15 an depth in the daily irrigated treatment. 'Thus, daily irrigation of 0.64 an of water was apparently more than sufficient compensate for water uptake by plant roots. At the 60 on soil depth the suction head ranged from 55 to 80 cm of water. Soil water pressure head for 15, 60 and 150 cm soil depths are shown in Fig. 14 for the treatment which was irrigated with 1.30 on of water when the Sllctj on head at 15 an depth exceeded 100 on of water. A total of 31 an of water was applied by irrigation during the growth season for the corn. The irrigation minimized the magnitude of major fluctuations in the soil suction head at the 15 an depth. The suction head ranged from a low of 40 on of water to a high of 300 an of water in the surface soil. Yields of corn grain harvested from plots receiving the three soilwater management and the two fertilizer management treatments are presented in Table 3. 1ne highest yields and the highest use efficiency for the irrigation water (Table 4) were obtained for corn grown with progrannned fertilization and controlled irrigation (1.30 on of water applied when the soil water suction head at 15 em depth exceeded 100 on of water). In lmirrigated plots the fertilizer treatment did not affect the grain yields; 110wever, in irrigated plots programmed fertilization resulted in higher grain yields than when the fertilizer was applied in a conventional manner. Programmed fertilization resulted in more efficient use of irrigation water in the irrigated corn plots. In plots receiving daily irrigation, water use efficiency for programmed fertilization was approximately 4 times larger than for conventional fertilization, and in plots receiving controlled irrigation, water use efficiency for progrannned fertilization was approximately twice that for conventional fertilization. For plots receiving either conventional or programmed fertilization, grain yields as well as irrigation use efficiencies were higher for controlled irrigation where tensiometers were used to schedule water applications than for daily applications of irrigation water. One of the major conclusions that can be obtained from this experiment is that fertilizer and water management for corn growing in sandy soil should be treated as mutually dependent phenomena. For example, excessive and indescriminate use of irrigation may decrease use efficiency of fertilizer and irrigation water by excessive leaching from the "rooting zone" of the soil profile and can result in nutrient contamination of ground-water. Also programmed fertilization tended to improve the use efficiency of irrigation water applied to corn.

PAGE 30

-----_. '0 E Q) L-700 ::J ; (/) (/) 0-. -300,' -100 o -24,'. I I I I r I I :1' I II I I I : , I I Soil Depth ...---15 em , .----30 o (j) + 1 00 25' 4565 85 105 125 Days Figure 12: Soil water pressure head at 15, 30, 90, and 150 an soil depths in Troup sand during the 1975 growing season for unirrigated corn.

PAGE 31

o (}) L :J -100 (f) (f) -80 (1) L -60 (L L -40 ()) -20 +-J 0 $ 0 0 +20 +40 (/) 25 -25-Irrigated Daily CO.64 em/day) 45 65 Soil Depth ,..--15cm \ '. I .--60 t ...-150 85 105 125 Day's 13: Soil water pressure head at 15, 60, and 150 em soil depths in Troup sand during the 1975 growing season for daily irrigated corn.

PAGE 32

L (l) +..J o 13 E ...SJ (l) S-300 U1 U) DL (l) -100 o (f) -26Controlled Irrigation Soil Depth ...---15 em I' .. t \ ,: ,. ,J ........... ... ..-150 45 65 85 105 125 Days Figure 14: Soil water pressure head at 15, 60, and 150 em soil depths in Troup sand during the 1975 growing season for corn irrigated when the pr sure head became less than -100 on of water at 15 em depth.

PAGE 33

-27.Tahle 3: Grain yields for corn grown under three irrigation and two fertilizer management programs on Troup sand, Quincy, 1975. Fertilizer Treatment Conventional Progrannned LSDO.05 (a,b,c) 1190 Grain Yield, kg/hal Irrigation (0.64 em/day) Irrigation2 2,860 a 2,7l0 a 3,380 a 4,950 b 4,760 b 6,600 c \SDO.05 is the least difference at a 5% probability level (Steel and Torrie 1960). ) l;rain yie Ids followed by the same alphabetical letters (a, b or c) or combination of letters did not differ from each other. em of irrigation water applied when the soil water suction head at 15 cm depth exceeded 100 cm of water.

PAGE 34

-28-Table 4: Irrigation use efficiency* for corn grown under two irrigation and two fertilizer management programs. Fertilizer Program Conventional Programmed Irrigation Treatments Daily Irrigation Controlled Irrigation (kg of grain/ha/em of lrrigation water) 9.3 37.3 61.3 120.6 Irrigation use efficiency was calculated by subtracting grain yield for theunirrigated plots from those for the irrigated plots and dividing the result by 56 em for the daily irrigated treatment and hy 31 em for the controlled irrigation treatment.

PAGE 35

-29CHAPTER 4: MA'IHEMATICAL MODELS During the course of this study several mathematical models have been devclopeJ for the purpose of describing soil water movement during infil tration ,ulll redistribution and the transport and interactions of applied 2,4-D, orthophosphate, and potassium in soils. Initially, existing models were utilized to describe the fate of these solutes in soils. However, it was found that new models were needed in order to account for the various mechanisms governing solute interactions which influence the movement of .these solutes in the soil-water system. The capability of such models to predict the simultaneous movement of water and applied solutes is clearly dependent upon their validity under controlled laboratory conditions and more crucially under field conditions. In this sDldy, the water movement model as well as the 2,4-D and orthophosphate models have been tested against field and laboratory results. A. Water and 2,4-D Models: The water and 2,4-P models are based on the numerical solution of the water -flow equation and the solute transport equation for transient and unsaturated water flow conditions. The transient one-dimensional water flow equation for unsaturated soils is a8 = D(8) aK(8) dt az az az (1) where t denotes the time (hr); z the vertical distance in the soil (em) positive downward; 8 the volumetric soil water content (em3/em3); the soil water diffusivity (em3/hr); and K@) the hydraulic conductivity (crn/hr). The corresponding equation governing transient one-dimensional solute transport through soils may be written as a(s + 8C) = (8D (v) ac) a(vC) at az s az az (2) where C is the solute concentration in soil solution (mg/em3); S the amount of solute sorbed by the soil mattix per unit volume of soil (mg/em3); v the Darcy water flux (em/hr); Ds(v) the solute dispersion coefficient (em2/hr); and v (= v/8) the pore water velocity (em/hr). The dispersion coefficient of equation (2) is assumed to be a function of pore water velocity, v. Adsorption isotherms which express relationships between quantities of solute aQsorbed by the soil, S, and corresponding concentration in soil solution, C, were used to evaluate the rate-of-sorption term as/ato Equations (1) and (2) subject to appropriate boundary and initial conditions were solved numerically using explicit-implicit finite difference approxi mation tecJmiques (see Selim and Mansell, 1974; Selim, Mansell, and Elzeftawy, 1976). SiTIRllated results from the water and 2,4-D models were compared with field measurements during continuous water infiltration in a Lakelmld sandy soil (Ultisol) near Quincy, Florida. It was found that increasing the water flux from 2.08 to 4010 em/hr during irrigation caused the maximum soil water content, 8, at the surface to increase (Fig. 15 and 16) from 0.26 to 0.29 em3/em3 The higher flux resulted in more rapid penetration of the wetting front. Predicted profiles of soil water as obtained using the numerical solution of the water flow equation (1) agreed well with water contents measured in the field; differences were within 0.02 to 0.05 crn3/em3 Thus predicted water content profiles enabled evaluation of water penetration with time into the soil for the infiltration period studied.

PAGE 36

-30Figures 17 and 18 show measured and calculated 2,4-D distributions during water infiltration with flux of 2.08 and 4.10 ern/hr. These field measurements indicate that the calculated 2,4-D distributions lagged behind the measured ones for all times during the infiltration process. The lag between calculated and measured distributions were between 2 and 5 ern. Retardation of calculated distributions with respect to measured values did not appear extensive; thus adequate agreement may be concluded. However, such retardation of calculated solute distributions could be attriof 2,4-D adsQJlltion_as determined isothenns. Results of 2,4-D and water movement in laboratory soil columns of Lakeland fine sand (Elzeftawy, Mansell, and Selim, 1976) showed essentially the same trends as that from the field studies. Moreover, from the field and laboratory results it was clear that following water application of 14.3 dm,penetration of 2,4-D did not exceed 30 ern below the soil surface. Therefore, we may conclude that 2,4-D movement in the soil is extremely slow which minirnizespotential contamination of groundwater resources small applications of 2,4-D herbicide are used. Bo Phosphorus Models: It was necessary to develop several models in order to describe ,the fate of phosphorus in a Spodosol. These models (see Mansell et al., 1976), are based upon solving the solute transport equation with several phos phorus interaction mechanisms. The transport equation governing one-dimensional phosphorus movement through soils may be written as + e = e -q -Q, (3) where C is the concentration of the solute in soil water p the soil bulk density ,(g/ern 3), e the soil water content (cm3 /cm 3), Ds the hy drodynamic dispersion coefficient (ern2/hr), q is the soil water flux (ern/ hr), S the amount of solute sorbed by the soil particles Q the rate of solute removal per unit volume of soil t is time (hr), and z is distance in the soil (ern). The term as/at represents the rate for reversible adsorption (or exchange) of solute by soil particle surfaces, and the sink term, Q, is the rate for irreversible removal (i.e., precipitation and immobilization) of solute from the soil. In this study, three phosphorus sorption processes were considered for the reversible tenn, a ,I at, in equation (3). First, adsorption was assumed to be linear and instantaneous S = KC (4) where K is the solute distribution coefficient (cm3/g) as deteI111ined from adsorption isothenns. Secondly, a nonlinear instantaneous adsorption mechanism was considered, S = K CN (5)

PAGE 37

-31where N < 1. TIlird1y, a kinetic reaction between S and C such that as 0 _. = -k eN k,t (KeN -S) 3t pac (6) was used. This equation describes a finite adsorption rate proportional to the difference between the equilibrium adsorption and the amount which is already adsorbed. TIle parameter ka is the forward (adsorption) and kd is the backward (desorption) reaction rate coefficients (hr-I), respectively. lNhen the value of N equals unity, equation (6) describes a reversible, first-order kinetic reaction. For many reactive solutes such as herbicides and phosphorus, the value of N has been found to be less than unity. Irreverslble sorption processes assigned to the Q term of equation ll) were considered in two ways. First, the sink term was given the form where kc is the rate coefficient (hr-l ) for precipitation and/or chemisorption of solute from the soil solution. Secondly, the sink term Q is expressed as, Q = ks(pS) (7) (8) where ks is considered the rate coefficient (hr-l ) for chemical immobilization of physically adsorbed phosphorus. 'TIle phosphorus models were used to describe experimental data for miscible displacement of phosphorus solution in Oldsmar fine sand (Spodo sol) collmms. The soil was obtained from an uncultivated and unfertilized area at the University of Florida Agricultural Research Center located Ilear Fort Pierce, Florida. Duplicate samples from undisturbed soil cores (5.4 on ins ide diameter and lOon in 1 ength) were collected from surface i\ 1, and subsurface, A2, horizons. Phosphorus breakthrough curve from the A2 subsurface soil is shown in Fig. 19 where the relative concentration in the effluent (C/Co ) is plotted against the number of pore volumes (V/Vo). Calculated concentration distributions in the effluent using equilibrium adsorption isotherm (S = 0.23 C) overestimated the measured retardation of phosphorus transport in the A2 soil. Only "vhen the distribution coefficient, K, was arbitrarily decreased tenfold (S = 0.023 C) did the simulated curve fit results. TIlis tenfold decrease in K assumes that only 10% of the adsorption sites were active in phosphorus removal. Therefore, we conclude that the equilibrium adsorption isotherm did not adequately describe phosphorus transport for the relatively high soil water pore velocities used in the subsurface A2 soil. However, using the kinetic equation (6) to represent phosphorus adsorption, calculated breakthrougll curves provided an adequate description of phosphorus transport in the A2 soil. Therefore, we conclude that phosphorus transport in the A2 soil appeared to occur under fully reversible conditions. In addition, the adsorption-desorption mechanism is of the first order kinetic type.

PAGE 38

-32-The phosphorus breakthrough curve from surface Al soil core (Fig. 20) is distinctly different from that for the A2 soil (Fig. 19). The breakthroug}l curve shows a large retardation, assymmetry, and extensive tailing during desorption (right-hand side). Calculated results from the equilibrium adsorption model (curve not shown) did not describe the shape or position of the experimental data. However, the calculated phosphorus breakthrough curve based upon the kinetic adsorption model (the NO SINK curve in Fig. 20) adequately described the left-hand side of the experi mentaLdata. De)[iat_ion between calculated and experimental results as well as incomplete recovery of applied phosphorus (80% recovery after 32 pore volumes) suggests that phosphorus retention in the Al soil may involve irreversible as well as reversible processes. Since complete recovery of phosphorus in the effluent was not attained in the Al soil, an irreversible sink term Q was included in the kinetic transport model for the purpose of providing better description of the experimental data. Equation (7) was first used to represent irreversible precipitation or chemisorption, and equation (8) to describe chemical irrnnobilization of physically adsorbed solute. Calculated breakthrough curves (Fig. 20 and 21) clearly depend upon the magnitude of these rate coefficients; however, values for kc and ks were 0.1 hr-l. Thus, phosphorl5 sorption during transport through the Al soil apparently occurred under conditions greatly removed from equllibrium with sorptiondesorption mechanisms occurring primarily as reversible processes but also accompanied by irreversible processes such as chemical or precipitation. Although the sink models for irreversible chemical irrnnobilizationand precipitation of phosphorus improved the agreement with experimental data, such agreements were not considered satisfactory. Furthermore, the exact mechanisms of irreversible adsorption were not identified. 'Therefore, a new phosphorus model (see Mansell, Selim, and Fiskell, 1976) was developed to describe various transformations which might occur for different soil and under different conditions. The model is a multistep one where phos phorus is considered to be present in four phases: A) solution, B) adsorbed, C) irrnnobilized, and D) precipitated. The four phases of phosphorus are interrelated to each other by the rates of individual transfonnations which are determined by both the rate coefficients and the amowlts of phosphorus in each phase. Under steady flow conditions, the transport and transformations of these four phases of phosphorus may be expressed as 8A 8 2 A 8A N + [k2B + k6D] (9) -= D v -[k1A + ksA] 8t s 3Z2 8Z a (pB) N [k2 + k3]pB + k4PC (10) 8t kI8A 8 (pC) k pB k4PC (n) 8t 3 a (pD) k 8AN -k pD (12) at 5 6

PAGE 39

-33where 0 volumetric soil water content (cm3 /cm 3 ) p soil bulk uensity (g/cm3 ) t time (hr) z = transport distance in soil (ern) v = average pore water velocity (ern/hr), Ds dispersion coefficient (cm2/hr). B, C and D amOlmts of phosphorus per gram of soil for adsorbed, immobilized, and precipitated phases, respectively A = phosphorus concentration in soil solution phase N = constant representing the order of the adsorption process, rate coefficients for adsorption, desorption, immobilization, mobilization, precipitation, and dissolution, respectively (hr-l). At any given time, the sum of the amounts of phosphorus per gram of soil in the adsorbed (B), immobilized (C), and precipitated (D) phases will be referred to collectively as the total amount of phosphorus sorbed per gram of soil, S. The rate coefficients are assumed to depend upon characteristics of the soil environment such as level of acidity (pH), concentrations of reactive Fe and AI presentl and the redox potential of the soil. Adsorption was assumed to be Ntn order where N = 0.35 (Mansell et al. 1976), whereas desorption, immobilization, mobilization, precipitation, and dissolution reactions were assumed to follow first order kinetic reactions. 1110 model, however, is versatile and other transformation processes can be included as deemed necessary. Furthermore, validation of this model is needed in order to fully predict the fate of applied phosphorus in soils. C. PotassiLUll Model: A mathematical model was developed (Selim, l\1a.nsell, and Zelazny, 1976) for the purpose of describing potassium transport and transfonnations in a soil profile. Potassium was considered to be present in the soil in four phases; solution, exchangeable, nonexchangeable (complex secondary minerals), and primary mineral. Sinrulated results were presented for assumed conditions of first-order kinetic reactions between the exchangeable, non exchangeable, and primary mineral phases. However, the.model is flexible and can be adapted to incorporate other transformation processes as well as variable rate coefficients with soil depth and time. Potassium distribution results (see Selim, Mansell, and Zelazny, 1976) illustrate the dependence of the leaching loss and transformation among the various potassium phases upon the rate coefficients governing the

PAGE 40

-34-transformation mechanism. Since water flow through the soil was maintained constant at all times, i.e., continuous water infiltration, the results presented here provide optimum conditions for transport and leaching losses of potassium. Transformations and transport of potassium under the more realistic conditions of water infiltration and redistribution should result in less movement of applied potassium. Such infiltration and redistribution conditions have been considered previously and can be adapted for the transport and transfonnations of potassium. Verification of the model-presented here is needed in order to fully describe the chemical fate of applied potassium in soil. Experimental evaluation of our model for various potassium transformation mechanisms is currently being investigated for several soils.

PAGE 41

-35-Water Content 3 3 e I cm /cm o .05 .10 .15 .20 .25 10 N 30 ..c 0. 40 o If) 50 60 70 Calculated x Experimental o FLUX = 2.08 cm/hr Figure 15. and calculated soil water profiles in Lakeland fine sand near Quincy, Florida during water infiltration with 2.08 on/hr flux.

PAGE 42

-36-WATER CONTENT e cm3/cm3 .05 .10 .20 .25 .30 0 10 20 30 E u 40 N 50 I r-60 a.. w 0 initial 70 0 80 If) 90 0 4 hI" 100 -Calculated Experimental X 110 A 0 FLUX = 4.1 cm/hr 120 Figure 16. Experimental and calculated soil water profiles in Lakc1 and rille silnd near Quincy, Florida during water infiltration with 4.111 em/Ill' fIlLX.

PAGE 43

-372)4-0 Concentration CJ ppm o 10 20 30 40 50 o 5 E u .. 10 N ..c +-' 2-15 o 20 25 ---------5 h --_ r X -.', ",,/ 1.5 hr 2.5 hr 3.5 hr 5 hr Experimental x 1.5 !!1 2.5 30 t-______ -=C:....:;:a:..:.lculated o 3.5 FLUX=2.08 cm/hr Figure 17. Distributions of 2,4-D in the solution phase of the soil profile of Lakeland fine sand near Quincy, Florida ,during water infiltration with 2.08 crn/hr flux.

PAGE 44

-382J4-D Concentration C J ppm o 10 20, 30 40 50 5 10 ---1.5 E --u -. ...... ."" -N 15 ----X ..c I +-' Q. \ OJ 20 \ 0 2.5 hr \ \ 025 \ If) ....... ""-, ..... ........ ...... 30 ........ ........ ,0 ....... 3.5 hr ... .,. ---------35 Calculated 40 5 hr Experimental x 1.5 6 2.5 45 FLUX =4.1 cm/hr 3.5 Figure 180 Distributions of 2,4-D in the solution phase of the soil profile of Lakeland fine sand near Quincy, Florida during water infiltration with 4.10 cm/hr flux.

PAGE 45

1.0 I / 0.8 I: I : I I JO.6 I I f I u 0.4 .. I O. 2 I!. f 2 4 \. \ \ \ I A2 soil \ \ V=54.5 cm/hr \ \ 1t = 0. -SJ \ --, \ '.----\ \ t. \ 6 8 vjvo =0.23C I =0.023C 10 Figure 19. Experimental and predicted phosphorus breakthrough curves for a water-unsaturated core of subsurface CA2) Oldsmar soil. I w '" I

PAGE 46

o u 1.0. r--'" -------------------, 0.8 0.6 0.4 0.2 I 00 No Sink4 ........ -"\ .. -,,' I / I ., /' I" !Ie / .. II l .. :/. !, ., 4 8 8 \ \ 12 l AI SOilJ V = 9.27 cm/hr Sink Term: Q = ks S ----ks = 0.05 hr-1 -1 ks 0.10 hr ks = 0 .. 25 h r-1 16 20 24 28 32 Figure 20. Predicted phosphorus breakthrough curves with and lI'ithout a sink tenn for irreversible immobilization of adsorbed P during transport through a IvaJ '-unsaturated core of surface CAl) Oldsmar soil. I +: o I

PAGE 47

1.0rl 0.8L f/ ( I A, soj I /'/ t L I, ... 9.27 cm/hr V= U 0.6f iZ -1f iI /,., Sink Q = kc C / .. / I I_ 'U' 0041 -1 I I ------k = 0.025 hr I .... c -1 k =0.100 hr c kc = 0.150 hr-1 O.2r ffi --.-..4 8 12 16 20 24 28 32 ViVo Figure 21. Predicted phosphorus breakthrough curves with and without a sink term fori irreversible precipitation of soluble P during transport through a water-unsaturated core of surface (AI) Oldsmar soil. I .j::--I-' I

PAGE 48

-42CHAPTER 5: Summary and Conclusions Sandy agricultural soils of Florida are typically limited in their capacities to retain irrigation water and applied chemicals such as fertilizers and herbicides. Therefore in order to economically grow optimum yields of crops on these soils under the humid climate of Florida, frequent applications of fertilizer is often required to compensate for leaching losses of nutrients from the "rooting zone" of the soil profile. Since these soils also have limited capacity to store infiltrated water during periods of redistribution," irrig(itio:i1 is also required during extended periods of infrequent rainfall. The simultaneous transport of nutrients and water through these sandy soils is therefore of utmost importance to the efficient use of both fertilizers and irrigation water for agricultural production of food. For example, leaching of nutrients from the soil tends to decrease the efficient use of fertilizers, whereas uptake of nutrients by plant roots tends to increase the use efficiency. Not only is the leaching of nutrients detrimental to food production but leached nutrients that move with the soil solution into underlying groundwater also pose a potential problem for environmental pollution. Results from this research indicate that irrigation and fertilizer management of com growing on sandy soils in a climate mutually effect the individual use efficiencies for applied water and nutrients. Field experiments of irrigated and fertilized com were established on two representative Florida soils: Wauchula sand, a Spodosol, and Troup sand, an Ultisol. The Wauchula sand was characterized by relatively slow rates of natural internal drainage due to the presence of subsurface hori zons of slowly-penneable material in the soil profile and due to relati vel\' flat terrain. By contrast, the Troup sand was characterized hy a rela tively unifonn profile of sand which transmits soil water rapidly downward. Beneath 150 em depth, however, the Troup sand was underlain by a slowlypenneable sandy clay material. Water tables were commonly observed in the Wauchula soil profile between depths ranging from 60 to 100 em during the growing season for corn; whereas, water tables rarely existed .in the Tronp soil profile. Although the Troup soil profile was primarily composed of sand material to a 150 em depth, a compacted plow layer occurred in the top 30 em of soil which was observed to limit much of the plant root growth to the top 15 cm of the entire 150 em profile. Soil water movement was not observed to be decreased by the presence of the plow layer. Com was grown under three irrigation treatments--no irrigation, daily irrigation (0.64 em/day), and controlled irrigation (1.3 cm of water was applied the soil water suction at 15 em soil depth exceeded 100 em of water) and two fertilizer application treatments--conventional (bulk of fertilizer applied prior to planting of corn and nitrogen fertilizer applied later as a sidedressing) and progranuned (fertilizer applied in several split applications). Without irrigation, grajn yields of corn \vere approximately twice as great for either fertilizer treatment in the WauchuLl sand as compared to those in the Troup sand. The larger yields were attri buted to the larger water contents and less leaching of fertilizer nutrients that occurred in the Wauchula soil profile as compared to the Troup soil dur ing the growing season. For both soils with com grown at either irrigation

PAGE 49

-43treatment, grain yields for corn fertilized in a progranrrned manner were higher than yields where rertilizer was applied conventionally. Water use {'1-riciC'lll'\, (net grain yield tlividecl by amount of irrigation water applied) I";IS only sl ightly liltered hy changing the method of fertilization for corn r I"l)\v i ng ill the Wauchub soi 1; however, for corn growing in Troup sand irri gilt ion WlltCI" was used much more efficiently by applying the fertilizer in a maImer. 'TI1e tlifferences in water use efficiencies for corn on the two soils were attributed to the greater magnitude of nutrient leaching observed in the Wauchula sand relative to that in the Troup sand. Water use efficiency for corn growing in both soils was greater for controlled irrigation than for daily irrigation. In the Troup soil highest grain yields and water use efficiencies were observed for corn fertilized in a programmed manner and irrigated in a controlled fashion. Daily irrigation apparently resulted in excessive leaching of both nutrients and irrigation water in the Troup sand. Thus corn growing in both soils used irrigation water and fertilizer nutrients more efficiently for programmed versus conventional fertilization and more efficiently for controlled versus daily irrigation. Several mathematical models were developed and used to describe the simultaneous transport of water and applied agrichemica1s (fertilizer nutrients and herbicides) to sandy soils. A transport model for water alone was successfully verified with experimental results from laboratory and field experbnents for periods of infiltration and redistribution. This model is flexible and can be adjusted to incorporate a variety of initial and boundary conditions for a range of soil types. Nutrient (phosphorus and potassium) and herbicide (2,4-D) models were also developed and used to describe chemical-'physical interactions and movement of these reactive solutes through soil. Simulated and limited experimental results associated with these TIlodels indicate that mechanisms such as physical adsorption-desorption, chemical precipitation, and immobilization or fixation have a large influence upon the transport of potassium and phosphorus, even in sandy soils. Further experimental validation of possible interactions of these agricultural chemicals during movement with water through soil is needed in order to precisely predict the leaching loss of agrichemica1s applied to sandy soils of humid climates.

PAGE 50

-44LITERATURE CITED 1. Brandt, A., E. Bresler, N. Diner, J. Ben-Asher, J. Heller, and D. Goldberg. 1971. Infiltration from a trickle source: I. Mathematical models. Soil Sci. Soc. Arner. Proc. 35:675-682. 2. Brasfield, J. F., V. W. Carlisle, and R. W. Johnson. 1973. Spodosolssoils with a spodic horizon. Pages 57-60 of Soils of the Southern States and Puerto Rico, Bulletin No. 174, Agricultural Experlment StatiGns of-theSouthern -States and-Puerto -Rico and theuSoil Conser vation Service of the United States Department of Agriculture. 3. Bresler, E., J. Heller, N. Diner, I. Ben-Asher, A. Brandt, and D. Goldberg. 1971. Infiltration from a trickle source: II. Experi mental data and theoretical predictions. Soil Sci. Soc. Arner. Proc. 35:683-689. 4. Butson, K. D., and G. M. Prine. 1968. Weekly rainfall frequencies in Florida. Circular S-197, Agricultural Experiment Stations, University of Florida, Gainesville. 5. Carlisle, V. W. and 1. W. Zelazny. 1973. Mineralogy of se1ecteu Florida Paleudults. Soil and Crop Sci. Soc. Florida Proc. 33:136-1:)9. 6. Elzeftawy, Atef and R. S. Mansell. 1975. Hydraulic conductivity calculations for unsaturated steady-state and transient-state flo\v in sand. Soil Sci. Soc. Arner. Proc. 39:599-603. 7. Elzeftawy, Atef, R. S. Mansell, and H. M. Selima 1976. Distributions of water and herbicide in Lakeland sand during initial stages of infiltration. Soil Sci. (accepted for publication). 8. Hammond, L. C., V. W. Carlisle, and J. S. Rogers. 1971. Physical and mil1eralogical characteristics of soil in SWAP experimental site at Fort Pierce, Florida. Soil and Crop Sci. Soc. Florida Proc. 31:210-214. 9. Jones, G. C. 1967. Some observations on Florida's water research needs, an economic appraisal. Publication No.3, Institute of Food and Agricultural Sciences, University of Florida, Gainesville. Pages 105114. 10. Mansell, R. S., H. M. Selim, and J. G. A. Fiskell. trlllsformations and transport of phosphorus in soil. (accepted for publication). 1976. Simulated Soil Sci. 11. Mansell, R. S., H. M. Selim, P. Kanchanasut, J. M. Davidson, and J. G. A. Fiskell. 1976. Experimental and simulated transport of phosphorus through sandy soils. Water Resources Research (accepted for publication) 12. Mansell, R. S., 1. W. Zelazny, L. C. Hammond, and H. M. Selim. 1975. Nutrient distributions in a Spodosol during corn growth. Soil and Crop Sci. Soc. Florida Proc.34:24-29.

PAGE 51

-45-13. Perkins, II. F., H. J. Byrd, and F. F. Ritchie, Jr. 1973. Ultisols-1 i ght-colored soils of the wann temperate forest lands. Pages 73-86 of Soi Is or the Southern States and Puerto Rico, Bulletin No. 174, Agri Stat10ns of the Southern States and Puerto Rico ;llld thl' So it Conservation Service of the United States Department of /\g" icul ture. 14. Selim, H. M. and R. S. tv1ansel1. 1976. Analytical solution of the equation for transport of reactive solutes through soils. Water Resources Research 12:528-532. 15. Selim, H. M., R. S. Mansell, and Atef Elzeftawy. 1976. Distributions of 2,4-D and water in soi 1 during infiltration and redistribution. Soil Sci. 121:176-183. 16. Selim, H. M., R. S. Mansell, and 1. W. Zelazny. 1976. Modeling reactions and transport of potassium in soils. Soil Sci. 22:77-84. 17. Steel, R. G. D. and J. H. Torrie. 1960. Principles and Procedures of Statistics. McGraw-Hill Book Co., New York, N. Y., pages 106, 107 and 114. 18. Thomas, G. W. 1970. Soil and climatic factors which affect nutrient mobility. In O.P. Engelstad (ed.) Nutrient Mobility in Soils: Assimulation and Losses, Special Publication No.4, Soil Sci. Soc. America Madison, Wisconsin, p. 1-20. 19. Watanabe, F. W., and S. R. Olsen. 1965. Test of an ascorbic acid method for detennining phosphorus in water and NaHC03 extracts from soil. Soil Sci. Soc. Amer. Proc. 29:677-678. 20. Zelazny, L. W. and V. W. Carlisle. 1971. Mineralogy of Florida Aeric Hap1aquods. Soil and Crop Sci. Soc. Florida Proc. 31:161-165.

PAGE 52

-46-APPENDIX: Titles and abstracts of published papers resulting from this research. 1. Saxena, G. K., R. S. Mansell, and C. C. Hortenstine. 1975. Drainageof vertical columns of Lakeland sand. Soil Sci. 120:1-12. A drainage experiment was conducted to determine the time-depen dence of soil-water retention during drainage of a quasi-uniform column of Lakeland sand and similar columns to phosphatic clay was -added--as--cm-amendment-ubserve-d to-in-;;;crease and-prolong water retention in the soil. -Saturatedhydl'alTlic conductivity of the column with soil amended with 5 percent phosphatic clay in the surface 30 em was 13 em/hr as compared to 25 em/hr for the quasi-uniform column of Lakeland sand. Drainage from the column with a I-em-thick layer of phosphatic clay and the column with a 2-em-thick layer of phosphatic clay aggregates was greatly restricted. The non aggregated phosphatic clay layer provided nearly constant impedence to flow across the layer, whereas successive water desaturation of the larger pores in the layer aggregates resulted in a time-dependent im-pedence to water flow. At saturation the hydraulic conductivities of the aggregated and nonaggregated clay layers were calculated to be 5.0 and 0.8 em/hr, respectively. Maximum hydraulic head gradients across the aggregated and nonaggregated clay layers were 4 and 13, respectively. Values of unsaturated hydraulic conductivity calculated from soil-water characteristic curves for the Lakeland sand and measured saturated hydraulic conductivity were in good agreement with values measured experimentally from the soil columns. 2. Elzeftawy, Atef and R. S. Mansell. 1975. Hydraulic conductivity calculations for unsaturated steady-state and transient-state flow in sand. Soil Sci. Soc. Arner. Proc. 39:599-603. Using a method employed by Green and Corey (1971), hydraulic conductivity was calculated as a function of water content for Lakeland fine sand. A gamma ray transmission method for measuring soil water content and a tensiometer-pressure transducer arrangement for measuring soil water suction were also used to experimentally determine values of hydraulic conductivity for a similar range of soil water contents in lIDdisturbed soil cores and hand-packed soil columns. Measured and calculated.values were in good agreement for steady flow. During transient flow soil water content was observed to be a non function of suction for water desorption, but depended upon the state of flow. Higher water contents were found at a given pressure head during unsteady flow than during steady flow or static equilibrium (zero flow). Graphs of water content versus soil water suction were similar for cases of steady and no-flow conditions. For transient flow, the soil water pressure depended upon the soil-water content Dnd rnte of change of pressure head with time. 3. Mansell, R. S., L. W. Zelazny, 1. C. Hammond, and H. M. Selim. 19750 Nutrient distributions in a Spodosolduring corn growth. Soil and Crop Sci. Soc. Florida Proc. 34:24-29. Distributions of K, P, NH4-N,and N03-N in a tile-drained Wauchula

PAGE 53

-47-sand (Spodosol) were determined under corn cultivation in the spring of 1974. Grain yields of 6,346 and 7,907 kg/ha were obtained from plots receiving 567 and 9,070 kg/ha of limestone, respectively. Nutrient distributions in the soil solution phase were used to describe nutrient movement with time in the soil profile. Movement of N03-N, NH4-N, K, and Cl proceeded downward in the profile during the first 43 days. However, for times greater than 43 days dovmward movement of these nutrients did not exceed 50-70 em depth. Restricted downward transport of nutrients was attributed to the presence of a slowly-permeable B2t horizon located 75 cm from the soil surface. 4. Selim, H. M. 1975. Water flow through a multilayer stratified hill side. Water Resources Research 11:949-957. The objective of this study is to present a mathematical alulysis for steady state saturated flow through multilayer stratified hillsides of semi-infinite depth. Two soil surface shapes were considered: a constant soil surface slope and a surface of arbitrary shape. Potential and stream functions were obtained for one-, two-, and three-layered hillsides. The method of solution was based on the Gram-Schmidt orthonorrnalization method. For two-layered hillsides the hydraulic conductivities were Kl:K2 = 1:10 and 10:1. For three-layered hillsides the hydraulic conductivities were Kl:K2:K3 = 1:10:1 and 10:1:10. Flow nets, seepage velocities, and flow rates are presented. These results are useful particularly with regard to subsurface flow, runoff, erosion, and solute movement through sloping soils. 5. Selim, H. M. and R. S. I\1ansell. 1976. Analytical solution of the equation for transport of reactive solutes through soils. Water Resources Research 12:528-532. I\1athernatical solutions of the differential equation governing reactive solute transport in a finite soil column were developed for two specific cases: continuous solute input and pulse-type solute input at the soil surface. These solutions incorporate reversible linear adsorption as well as irreversible solute adsorption. The irreversible adsorption was expressed by a sink/source term which either may be a constant or may have a concentration-dependent form. The boundary condition used across the surface (X = 0) was that of the third type, wllich accounts for advection as well as dispersion. To illustrate the significance of using the proper boundary conditions, comparisons were made with two other mathematical solutions, one by Cleary and Adrian (1973) and another by Lindstrom et ale (1967). We conclude that the solution presented here is highly recommended for low flow velocities or specifically for voL/Ds values less than 20. For large pore velocities, or specifically for voL/Ds' >20, all three solutions are in agreemento 6. Selim, rio M., R. S. Mansell, and Atef Elzeftawy. 1976. Distributions of 2,4-D and water in soil during infiltration and redistribution. Soil Sci. 121:176-183. A numerical model was developed to predict transient water and

PAGE 54

-48reactive solute transport in water-unsaturated soil. An implicitexplicit method of finite difference approximation was used to simultaneously solve the water and solute flow equations. The model was used to calculate transport of water and 2,4-D (2,4-Dichlorophenoxyacetic acid) in a field profile of Lakeland fine sand during infiltration at constant intensities (2.08 and 4.10 cm/hr). Calculated values of the soil-water distributions compared well to experimental determinations, but agreement obtained between 2,4-D distributions was only were _also to S1:llclY the effect of irrigation intensity on water and herbicide transport during redistribution following irrigation. It was found that for advanced stages of redistribution solute movement became negligible and the maximum herbicide concentration was located at the same depth in the soil profile regardless of irrigation intensity. 7. Selim, H. M., R. S. Mansell, and L. W. Zelazny. 1976. Modeling reactions and transport of potassium in soils. Soil Sci. 22:77-84. A mathematical model was developed to describe potassium reactions and transport in soils. Kinetic reactions were assumed to govern the transformation between solution, exchangeable, nonexchangeable (secondary minerals), and primary mineral phases of potassium. Simulated results were presented for two soils, a weakly sorbing soil and a strongly sorbing soil. The effect of kinetic rate coefficients upon transport and transformation of applied potassium was also investigated. The model is flexible and can be adapted to incorporate various transformation mechanisms between the different phases of potassium. rvbdel validation with the aid of experimental data is needed to further describe the fate of potassium in soil. 8. Elzeftawy, Atef, R. S. Mansell, and H. M. Selim. 1976. Distributions of water and herbicide in Lakeland sand during initial stages of infiltration. Soil Sci. (accepted for publication). Penetration depths and concentration distributions were determined in columns of Lakeland sand for surface-applied 2,4-D herbicide during initial stages of steady water infiltration. Water was applied with application rates, R, of 2 and 4 cm/hr to soil with initial water contents, 8 i of 0.20, 0.11 and 0.01 cm3/cm3 Increasing R resulted in faster rates of advance for both the water wetting fronts and the depths of 2,4-D peaks. Larger values of 8, however, increased the rate of advance for the wetting front but aid not affect the location for 2,4-D peaks. Penetration of 2,4-D peaks per unit of water applied to the soil was observed to be independent of both R and 8i during initial stages of water infiltration. Distributions of 2,4-D concentration as calculated from water and solute transport equations were observed to lag Slightly behind experimentally measured distributions. 9. Mansell, R. S., H. M. Se1im, and J. G. A. Fiskell. 1976. transformations and transport of phosphorus in soil. Soil Sci. (accepted for puLlication).

PAGE 55

-49-A mechanistic, multistep model was developed using chemical kinetics and mass transport theory to describe transformations and move ment of orthophosphate in soil. Soil phosphorus was assumed to occur simultaneously in any of four primary phases: water-soluble, physically adsorbed, nmnobilized, and precipitated. Kinetic reactions which control the transformation of phosphorus between any two of the four phases were considered to be reversible and of Nth order. A range of values for the reaction rate coefficients were used inthemodel to describe the transport of applied phosphorus in the solution phase of the soil profile. 10. Mansell, R. S., H. M. Selim, P. Kanchanasut, J. M. Davidson, and J. G. A. Fiskell. 1976. Experimental and simulated transport of phosphorus through sandy soils. Water Resources Research (accepted for publication). Reversible equilibrium adsorption-desorption relationships were inadequate for describing the transport of orthophosphate through watersaturated and unsaturated cores from surface (AI) and subsurface (AZ) horizons of Oldsmar fine sand (a Spodosol). USlng a kinetic model with nonlinear reversible adsorption-desorption improved descriptions of phosphorus transport through these soils. Phosphorus effluent concentrations were described best using an irreversible sink for chemical immobilization or precipitation with a nonlinear reversible, kinetic adsorption-desorption equation. 11. Selirn, H. M. and R. S. Mansell. 1974. Transient one-dimensional and simultaneous solute and water flow in soils. Program No. 360 D-17.4.003, SHARE Program Library Agency, Triangle Universities Computation Center, Research Triangle Park, North Carolina. A computer program has been developed for the problem of solute and water movement in unsaturated soils or porous media under transient flow conditions. The two nonlinear partial differential equations governing the solute and water flow are solved simultaneously for the water content and solute concentration at any specified time and location as desired. The initial conditions used are uniform salt and water content distributions at time t=O. The boundary conditions at the soil surface are water flux and constant salt concentration conditions. The method of solution is a numerical one which utilizes the explicit-implicit finite difference technique. The computer program is written in FORTRAN language and consists of a source program, eleven subprograms, and an input data section. An important feature of the program is that incremental distance and time steps are adjusted automatically to satisfy stability and conver gence criteria for the water" and solute finite difference criteria. A second feature is that the number of nodal points are automatically calculated from the length of the flow region. A third feature of the program is that output data of water content, water flUX, solute concentration, and solute flux in the flow region are provided at specified times as desired.

PAGE 56

-5012. Elzeftawy, Atef A. 1974. Water and solute transport in Lakeland fine sand. Ph.D. Dissertation, Soil Science Department, University of Florida, Gainesville. The objective of this study was to investigate effects of three water supply rates--2, 4, and 8 cm/hr --and three initial soil water contents --1.2, 10.9, and 20.2% by volume -upon the simultaneous transport of water and solutes --2,4-D herbicide and chloride --in vertical columns of Lakeland fine sand. Columns were prepared by packing air-dry-soil intocylinders-7-. 6 cITl--diameter -and 107 cm long. A specific volume of aqueous solution containing 57.9 ppm chloridc'and 5 ppm 2,4-D was introduced and displaced through each colunll1. ray attenuation and pressure-transducer-tensiometers were used to precisely monitor soil-water content and pressure distributions with Soil solution was extracted at selected depth intervals along the soil columns and extracted samples were then analyzed for 2,4-D and chloride content. Depths to which chloride and 2,4-D moved for a given quantity of water infiltrated into the surface of the soil was found to depend upon the surface water flux. Increasing water application rates resulted in an increased water content in surface soil and in shallower displacement of chloride and 2,4-D for equal quantities of accwnulative infiltration. For a given quantity of water infiltrated, initial soil-water content did not influence depths of chloride or 2,4-D transport. Adsorption, caused 2,4-D distributions to lag behind those for chloride for all experiments.