Irrigation efficiency and controlled root-zone wetting in deep sands

MISSING IMAGE

Material Information

Title:
Irrigation efficiency and controlled root-zone wetting in deep sands
Series Title:
Florida Water Resources Research Center Publication Number 52
Physical Description:
Book
Creator:
Hammond, L. C.
Mansell, R. S.
Robertson, W. K.
Johnson, J. T.
Selim, H. M.
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Notes

Abstract:
Ten field water management experiments were conducted with corn, soybeans and peanuts during 1977, 1978, and 1979 on well-drained sandy soils at Gainesville and Live Oak. Irrigation scheduling treatments varied from one to eight per experiment. A simple computer simulation model provided an estimate of seasonal evapotranspiration (ET), drainage, and change in soil storage. These data and calculated water use efficiencies served as a test of the effectiveness of various irrigation scheduling strategies, and provided information on crop response to water stress. Yields increased linearly with estimated ET. Regression of corn yields on estimated ET gave the following 3-year average results for Gainesville and Live Oak, respectively: Y = 661X - 22,895 (R2 0.90) Y = 471X - 12,964 (R2 = 0.73) where Y is grain yield, kg/ha-cm, and X is ET, cm. These findings were interpreted to mean that the most efficient crop production use of water is obtained when water is supplied to meet the full easonal ET needs imposed by the atmosphere. In addition a strategy of light, frequent irrigation of only the top 30 cm of Bandy soils will produce high yields with minimum deep seepage loss of water and nutrients.

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:
AA00001528:00001


This item is only available as the following downloads:


Full Text














IRRIGATION EFFICIENCY AND CONTROLLED
ROOT-ZONE WETTING IN DEEP SANDS



By



I.. C. Hammond, R. S. Mansell, W. K. Robertson,
J. T. Johnson, and H. M. Selim


PUBLICATION NO. 52















IRRIGATION EFFICIENCY AND CONTROLLED
ROOT-ZONE WETTING IN DEEP SANDS



By



L. C. Hammond, R. S. Mansell, W. K. Robertson,
J. T. Johnson, and H. M. Selim


PUBLICATION NO. 52



FLORIDA WATER RESOURCES RESEARCH CENTER



RESEARCH PROJECT TECHNICAL COMPLETION REPORT



OWRT Project Number A-034-FLA



Annual Allotment Agreement Numbers

14-34-0001-7019
14-34-0001-7020
14-34-0001-8010
14-34-0001-9010



Report Submitted January, 1981



The work upon which this report is based was supported in part
by funds provided by the United States Department of the
Interior, Office of Water Research and Technology
as Authorized under the Water Resources
Research Act of 1964 as amended.









-ii-


TABLE OF CONTENTS


Tit I . . . . . .


T'r le of Contents . . . .


Acknowledgements . . . .


Abstract . . .


Chapter I. Introduction .


Chapter II. Field Experiments,

A. Corn, 1977, Experiment

B. Corn, 1977, Experiment

C. Peanuts, 1977 .

1). Corn, 1978 . .

I. IP.e.muts, 1978 .

F. Soybc;ins, 1978 .

C. Corn, 1979 . .


a n

I .

II


Chapter

A.

B.

C.


Chapter

A.

B.

C.


TI1. Field Experiments, Live

Corn Experiment, 1977 .

Corn Experiment, 1978 .

Corn Experiment, 1979 .


IV. Summary Discussion .

Water-Use Efficiency .

Irrigation Scheduling .

Water Policy for Agriculture


Oak

. .


in Florida


Literature Cited . . .


Abstracts of Published Papers .


Appendix Table . . .


. . . i


. . . ii


G i ill


Ies V e.... .

. .

. .

. .

. .

. .

. .

. .D








-iii-




ACKNOWLEDGEMENTS

We are grateful for contributions from other professional members
of associated research teams: J. M. Bennett, K. J. Boote, J. A. Cornell,
J. M. Davidson, J. W. Jones, P. S. C. Rao, and A. G. Smajstrala. A host
of technical staff in the Departments of Agricultural Engineering,
Agronomy, and Soil Science and at the Agricultural Research Center, Live
Oak, participated in this research activity.

We are thankful for the administrative support of W. H. Morgan, J.
P. Heaney, and Mary Robinson of the Florida Water Resources Center.

In addition, this is to express appreciation to Ron Jessup for the
computer simulations, to Jennifer Johnson for the drafting, and to
Sheila Whitlock and Barbara Stokes for typing the manuscript.






-iv-


ABSTRACT

Ten field water management experiments were conducted with corn,
soybeans and peanuts during 1977, 1978, and 1979 on well-drained sandy
soils at Gainesville and Live Oak. Irrigation scheduling treatments
varied from one to eight per experiment.

A simple computer simulation model provided an estimate of seasonal
evapotranspiration (ET), drainage, and change in soil storage. These
data and calculated water use efficiencies served as a test of the
effectiveness of various irrigation scheduling strategies, and provided
information on crop response to water stress.

Yields increased linearly with estimated ET. Regression of corn
yields on estimated ET gave the following 3-year average results for
Gainesville and Live Oak, respectively:

Y = 661X 22,895 (R2 = 0.90)

Y = 471X 12,964 (R2 = 0.73)

where Y is grain yield, kg/ha-cm, and X is ET, cm.

These findings were interpreted to mean that the most efficient
crop production use of water is obtained when water is supplied to meet
the full seasonal ET needs imposed by the atmosphere. In addition, a
strategy of light, frequent irrigation of only the top 30 cm of sandy
soils will produce high yields with minimum deep seepage loss of water
and nutrients.















CHAPTER I. INTRODUCTION

Although the Florida climate is characterized by relatively high
rainfall, two factors create a need for irrigation of crops for effi-
cient production: (1) inconsistent and often unfavorable rainfall
distribution patterns, and (2) the predominance of sandy soils with low
water and plant nutrient storage capacities. The annual rainfall
distribution is highly nonuniform with approximately 80-90% of the total
occurring during the summer. Even during years with average rainfall
distribution, irrigation is commonly required during spring and late
autumn. Reasonably successful irrigation and other management practices
have evolved over the years from farmer experience and scientific find-
ings; nevertheless, recent population growth has accelerated the demand
on our large but finite water resource, and have made it necessary that
all users--agriculture, municipalities, and industry--develop more
efficient water-use systems. Favorable factors in the Florida climate
such as long growing season, warm temperatures, and high total rainfall
justify the continued development of more efficient irrigation practices
through application of present knowledge and the discovery of new
knowledge through research.

In fact, the need for new information exists for most of the humid
southeastern United States where there has been a major increase in
irrigated agriculture during the past five years (Bruce et al., 1980).
In particular, it has been found that crop production practices asso-
ciated with irrigated agriculture in arid regions are not applicable to
humid regions without considerable modification and adaptation. Humid
region agriculture depends mostly upon rainfall for crop production, and
irrigation is needed during relatively short but numerous droughts.
Consequently, when uneven rainfall distribution patterns are coupled
with soils which have characteristically restricted root zones and thus
limited water storage capacities, there is created a major problem in
the scheduling of irrigation. Timing, application intensity, method of
application, and amounts of water applied affect the fraction of added
water used by the plant, the leaching losses of pesticides and fertil-
izers, and in many cases the aeration condition in the root zone. These
problems have not yet received research attention commensurate with the
high return potential for solution to them. Of considerable current
interest and research attention, is a water management system based on
high-frequency irrigation with shallow wetting of the root zone (Rawlings
and Raats, 1979). In this system, small quantities of water are applied
frequently to meet crop needs in a manner such that a closely following
rainfall replenishes water depleted from deeper portions of the root
zone without excessive loss to deep seepage.








-2-


In Florida, most well-drained sandy soils temporarily store less
than one inch of water per foot of soil depth, so that the growing crop
can develop water stress within 3 to 7 days (depending on rooting
depth) following a rainfall or irrigation. When the soil deficit is
replenished with irrigation the possibility always exists that un-
expected rainfall will displace the infiltrated irrigation water so that
it will be lost from the root zone as deep seepage. The severity of
deep seepage loss of irrigation water increases as the water retaining
capacity of the soil decreases. Deep seepage is also more severe if the
soil is relatively wet at the time irrigation water is applied. And for
a given soil, deep seepage of water increases as depth of rooting
decreases. In the final analysis, careful water management of crops in
humid regions can be used to minimize deep seepage loss of water applied
as irrigation. It is neither desirable or practical to completely
eliminate deep seepage in agricultural soils. However, farmers should
attempt to minimize the deep seepage loss of irrigation water. Farmers
can minimize such losses by careful irrigation scheduling based on crop
growth stage and rooting depth, soil water status and water retaining
capacity, evaporative demand of the atmosphere, and rainfall forecasts.

The objectives of the current study were:

1. To test the hypothesis that, for humid regions characterized
by short duration droughts, improved water-use efficiency
can be attained by replenishing a part rather than the full
soil water deficit in the root zone without adverse effects
on crop yield.

2. To determine, under field conditions on deep sands, the
influence of sprinkler irrigation management on the partition
of total water input between evapotranspiration and deep
seepage, and on the water-use efficiency of corn, soybeans,
and peanuts.

3. To develop and validate mathematical models incorporating
total water input amounts and distribution patterns and soil
physical properties for the purpose of describing water
infiltration, redistribution, deep seepage and uptake by
plants.

4. To develop, from the field data and the mathematical models,
efficient water management systems which can be implemented
readily by growers and utilized by water planners and policy-
makers.














CHAPTER II. FIELD EXPERIMENTS, GAINESVILLE

Ten water management field experiments were conducted during the
growing seasons of 1977,'78, and'79. Crops included corn, peanuts and
soybeans. Three soil series were involved: Lake, Arredondo, and Kend-
rick fine sands. Lake fine sand is a member of the hyperthermic, coated
family of Typic Quartzipsamments. In the test sites, this soil has
sandy A and B horizons that extend to a depth of 210 cm or more with a
sandy clay B2t horizon underneath. Arredondo and Kendrick fine sands
are members of the loamy, siliceous, hyperthermic families of Gross-
arenic and Arenic Paleudults, respectively. In the experimental area,
Arredondo fine sand is similar to the Lake fine sand except that a sandy
clay loam B2t horizon begins at depths ranging from 120 to 200 cm. The
Kendrick fine sand profile consists of fine sand material over a fine
sandy loam B2t horizon which begins at depths of 100 to 150 cm.

Irrigation was carried out by three overhead sprinkler systems: (1)
hand-operated fan spray nozzle on garden hose, (2) low pressure "Micro-
jet" sprinklers, and (3) impact sprinklers. Water management treatment
variables included timing of irrigation events and quantities per event.
Irrigation intensities were 2.55 cm per hour or less so that no surface
runoff occurred.

Other soil and crop management practices were approximately equal
to or better than those currently recommended for farmers. The soil
water status was monitored periodically with tensiometer readings of
water suction and neutron and gravimetric measurements of volumetric
water content. Rainfall distribution was nonuniform with time over the
three years of the study. Crop damaging droughts of varying durations
occurred during each year.

Crop response to water management was determined as yield of market-
able grain. Response to irrigation was analyzed using a simple water
balance model (Rao et al. 1976, 1981) which incorporates estimated daily
evapotranspiration (ET) rates (from monthly averages), measured soil
water characteristics (field capacity, permanent wilting percentage, and
water redistribution time), estimated root depth with time, and a water
extraction rate which equals the ET rate until 80% of the available
water has been depleted. At that point, the extraction rate was decreased
linearly with decreasing available water to zero at the wilting point.

Water balance simulations used in the Gainesville and Live Oak
experiments were based on potential ET rates calculated for Jacksonville
(Table 1). These ET rates were calculated using the Penman method from
longterm weather records and from handbook tables of extraterrestial
















radiation. In most of the simulations, a 10% downward adjustment of the
ET rate was made for an incomplete crop canopy (0-25 days) during the
early part of the season, and a 10% upward adjustment was made for later
in the season (after 40 days). The latter adjustment was made in con-
sideration of the non-average weather conditions associated with pro-
longed droughts.


Table 1. Calculated daily potential evapotranspiration rates by months
for three locations.

Potential evapotranspiration
Jacksonville Tampa Miami
------------------------ cm/day ---------------

January 0.112 0.152 0.191
February 0.163 0.206 0.252
March 0.234 0.277 0.323
April 0.343 0.371 0.391
May 0.411 0.432 0.424
June 0.422 0.432 0.422
July 0.429 0.414 0.432
August 0.391 0.399 0.4 17
September 0.312 0.348 0.35(
October 0.226 0.279 0.287
November 0.150 0.191 0.216
December 0.104 0.142 0.180
Annual Total, Jan-Dec (cm)
100.5 111.0 118.4



A. Corn, 1977, Experiment I

Experiment I was designed to test the yield response of three corn
hybrids under irrigation to subsoiling and to multiple sidedressings at a
fixed total nitrogen level.

1. Methods

This experiment was located on Kendrick fine sand soil and a solid
set overhead sprinkler irrigation system delivered water at the rate of
0.51 cm/hour. All plots received irrigation averaging 1.9 cm per appli-
cation when the soil water suction at 15 cm depth reached 500 cm.
Treatments were arranged in a factorial statistical design of three corn
hybrids, two soil conditions, and four nitrogen application schemes.














ThLe 5.5 x 7.6 m plots were arranged in a randomized block with four
replications (Table 2). Prior to planting, subsoiling was performed
with chisel plow 30 cm on center and to a depth of 35 cm. Corn was
planted by hand in 45 cm rows with 30 cm spacing (71,760 plants/ha) on
March 29, 1977 and harvested July 27 to Aug. 7 (soon after maturity).


Values of various parameters
are given in the Appendix Table.
water redistribution are shown in


used in the water balance simulation
The time dependence of root growth and
Figures 1 and 2.


2. Results and Discussion

The rainfall distribution pattern shown in Figure 3 indicates that
the corn growing season of 1977 was extremely deficient in water. Only
12.7 cm of rainfall occurred, and 42.2 cm of water was used as irriga-
tion (Table 3). The simulated and measured seasonal water balance data
are given in Table 4. Actual ET and water drainage (deep seepage)
losses may be higher or lower than these estimates. Nevertheless, the
simulated data provide useful information for evaluating results and
planning further studies.


Table 2. Nitrogen application treatments (sub-subpl Ts)
ment I, Gainesville, 1977.-


for corn Experi-


Number
of applications


1

2


3



4


Date


May 6

April 29
May 13


April
May
May


2/
Nitrogen-2/
kg/ha

224

112
112

56
112
56


April 22
May 6
May 20
June 6


-Main plot treatments: Funk 4810, Pioneer 3369A, and McNair 508 corn
hybrids; subplot treatments: non-subsoiled and subsoiled.
2/
- As NI NO3, additional nitrogen applied at planting as part of mixed
fertilizer (45-39-149-45 kg/ha N-P-K-Mg).









TIME, days from planting
14 20 40 60 80 100 120


20-
SIMULATED
40-
ROOT DEPTH

E 60- CORN
S80-
I--
0 100- T
Lil
120-

140- 1980

160- 1979

180


Estimated depth of corn root zone with time after planting.


Figure 1.









































0 3 6 9 12
DAYS


Figure 2.


Water redistribution time in a Lake fine sand.

















1977

GAINESVILLE


* LI


MAR APR MAY JUN JUL


AUG


8.86

,1


SEP


iIi. *i *iI


OCT NOV


Figure 3. Rainfall distribution at Gainesville, 1977.


1,1 1 .1 11 j I. II Ii I d 1111 11 A


I


ii n iIaIB IIH













Table 3. Schedule for


water application to corn,
Experiment I.


Gainesville, 1977,


Irrigati 1
amount-
cm

1.07
1.52
1.52
1.52
1.78
1.96
2.18
2.54
1.91
0.64
2.16


Date


June 9
11
14
20
27
30
July 6
12
15
22
26


Total


Irrigation
amount
cm

2.03
2.54
2.29
2.11
2.79
2.36
1.78
2.16
1.78
1.78
1.78


42.20


i/Rainfall, 12.7 cm.



Table 4. Corn grain yields and estimated water balances for Experiments
I and II, Gainesville 1977.

Profile
2 water 3 Grain
Treatment- ET Irrig.- depletion-- Drainage yield
- - cm - kg/ha

Experiment I
Irrigated 48.21 42.20 2.13 9.03 8911
Experiment II
No irrigation (1) 20.85 8.02 282 b
Irrigated (2) 47.33 47.59 5.44 18.55 7991 a
Irrigated (3) 47.64 51.12 2.24 18.57 9000 a

-See text for details.
2/Rainfall amounts of 12.92 and 12.85 cm occurred for Experiments I and
II, respectively.
3/Net seasonal loss from the soil profile.


Date


May 1
4
8
12
15
18
21
25
29
June 2
4








-10-


Grain yields (Table 5) for the three hybrids were in the range
expected for corn growing in sandy soils under good water management.
The Pioneer hybrid gave significantly higher grain yields than did the
Funk and McNair hybrids. Subsoiling generally improves root density
distribution in this sandy soil by mechanically disrupting soil zones
which have become compacted due to tillage; however, this improvement
would not be expected to provide increased corn yield during dry years
(1977) when irrigation is used extensively.

Three and four split applications of N gave higher corn yields than
a single application. This result could be attributed to a larger vola-
tilization loss from the single application.


Table 5. Yield of corn grain as affected by hybrid, subsoiling, and
number of nitrogen sidedressings.

Treatment Grain yield-/
kg/ha

Hybrid:
Funk 4810 8671 b
McNair 508 8770 b
Pioneer 3369A 9293 a

Soil condition:
Non-subsoiled 8857 a
Subsoiled 8965 a

Nitrogen sidedressings: 2/

1 8501 b
2 8733 ab
3 9283 a
4 9128 a

1/Yield values followed by the same letter are not different at the 0.05
level.
- See Table 2 for rates at each application.








-11-


B. Corn, 1977, Experiment II

The purpose of this experiment was to determine the response of
corn to irrigation regimes having small quantities of water applied
frequently (small frequent applications) versus medium quantities applied
less frequently (medium infrequent applications).

1. Methods

Replicated (4 each) treatments were: (1) no irrigation, (2) irri-
gation frequently in small amounts, and (3) irrigation less frequently
in medium amounts. Plots 7 x 7 m were located on Lake fine sand. During
irrigation events overhead sprinklers delivered water at the rate of 1.7
cm/hr. Irrigation times and amount are shown in Table 6 and rainfall in
Figure 3. Water was applied when the soil water suction at 15 cm depth
reached approximately 200 cm. DeKalb XL-80 corn hybrid was planted in
45 cm rows at a.population of 95,000 plants/ha on March 24, 1977. Corn
reached maturity on July 27 and was harvested on August 15. Simulated
and measured water data were obtained in the same way as for Experiment
I. See the Appendix Table for input data.

Table 6. Schedule of water application to corn, Gainesville, 1977,
Experiment II.

1/
Irrigation amount- Irrigation amount
Date Treat. 2 Treat. 3 Date Treat. 2 Treat. 3
cm cm cm cm

May 2 -- 1.70 June 12 3.68 --
3 1.70 -- 13 -- 3.60
8 1.30 1.70 15 4.80 1.70
14 1.70 2.55 20 -- 4.38
18 3.12 -- 27 3.82 --
19 -- 3.40 29 -- --
22 3.40 -- 30 3.68 --
23 -- July 6 -- 4.38
25 3.40 -- 10 5.38 --
26 -- 4.25 15 -- 4.24
June 1 3.83 -- 17 4.24 --
3 -- 4.39 24 -- 3.82
5 3.54 --
8 -- 3.40 Total 47.59 51.12

-/Rainfall, 12.85 cm.








-12-


2. Results and Discussion

Seasonal rainfall for Experiment II was approximately the same as
for Experiment I (Figure 3). Irrigation amounts and distribution patterns
for treatment 2 and 3 did not differ as much as anticipated because of
the extremely dry growing season (Table 6). The amounts may be overesti-
mated as much as 10%.

Water balance and grain yield data are given in Table 4 along with
data from Experiment I. Yields from the irrigation treatments in both
experiments are comparable even though the calculated amount of water
applied was larger in Experiment II. The simulation model indicated
excessive irrigation quantities in the latter experiment with a conse-
quent drainage loss of water from the root zone. It is possible that
the actual ET was greater than calculated since the plots were small
enough for a marked oasis effect and the climatic conditions during the
long drought were far from the average conditions assumed in the calcu-
lation of expected ET (Table 1).


C. Peanuts, 1977

The objective of this study was to determine the yield response of
peanuts on a deep, well-drained sandy soil to four Irrigation treatments.

1. Methods

Small field plots 5.5 x 5.5 m located on Lake fine sand, were
planted with 'Florunner' peanuts in 91 cm rows on April 22, 1977. Four
treatments were replicated four times: (1) no irrigation, (2) frequent
irrigation, small amount, (3) infrequent irrigation, medium amount, and
(4) infrequent irrigation, small amount. Water was applied by hand with
a calibrated fan spray nozzle on a garden hose. Repeated passes along
each row provided a rate of application of nearly 42 cm/hr. Irrigation
was scheduled for treatments 2 and 3 when soil water suction at 15 cm
reached 500 cm or more. In treatment 4, plant water stress symptoms were
allowed to develop before irrigation was scheduled. Simulated water
balance data were obtained as before (Appendix Table).

Soil water content and suction were monitored with a neutron meter
and tensiometers, respectively. Data are in a thesis by Nafis (1979).

2. Results and Discussion

Rainfall and irrigation distribution data are shown in Figures 3
and 4. The first three irrigations were also applied initially to the






-13-


1977 GAINESVILLE


APR MAY


PEANUTS


JUN JUL


AUG


Irrigation schedules for three water management
treatments on peanuts, Gainesville, 1977.


2




11LJ11fl.1 -
3




I I _______
4


z
01

I-

r2


1


0


Figure 4.







-14-


no-irrigation treatment in order to get the plants up and established
during a very dry period. Irrigation input, simulated ET, drainage from
the root zone, and pod yields are given in Table 7. Roots were assumed
to increase in depths linearly with time from 4 cm on day 1 to a maximum
of 200 cm by day 90. The simulated ET values may be lower than actual
since there was additional apparent drainage water under irrigation.
Nevertheless, most of the added irrigation was allocated to ET and the
measured pod yields increased linearly with the calculated values of ET
(Figure 5).


Table 7. Water balance and peanut yields, Gainesville, 1977.

Profile
I/ water 2/
Treatment ET Irrig.- depletion- Drainage Yield
- - cm - kg/ha

Non-irrigated 36.95 4.8 2.45 0.53 2261 c
Irrig., frequent 46.91 12.9 1.88 2.89 3817 a
Irrig., infrequent 44.76 10.4 1.96 2.67 3622 a
Irrig., plant stress 41.78 6.7 1.92 1.89 2999 h

- Rainfall, 30.2 cm.
2/
- Net seasonal loss from the soil profile.

In 1975 studies, peanut pod yields were 4500 kg/ha when the amount
of seasonal water depletion was 57 cm, and irrigation of 8 cm did not
cause a yield increase (Varnell, et al. 1976). However, when water
input was decreased by covering the plot during midseason rainfalls the
yields were decreased to 3900 kg/ha with a water depletion of 48 cm.
The latter yield is similar to that predicted from the 1977 results in
Figure 5 if we assume that water depletion is a fair estimate of evapo-
transpiration.

The peanut plant is often considered to be drought tolerant due to
its characteristic deep rooting and reinitiation of blooming and fruit-
ing after drought stress. Nevertheless, these data indicate that peanuts
do respond to water management and that yields can be severely reduced
by an overall seasonal water deficit.

D. Corn, 1978

This study involved two crops corn and peanuts grown in adjacent
plots on the same Lake fine sand site as the 1977 peanut experiment.







-15-


4000






a- 3000
.!


2000


0 36 38 40 42 44 46


EVAPOTRANSPIRATION,


Figure 5.


The relationship of peanut pod yield to estimated
evapotranspiration, Gainesville, 1977.


cm







-16-


The objective was to determine the effect of irrigation strategy (fre-
quency and quantity) on periodic water depletion rates and on irrigation
water-use efficiency.

1. Methods

De Kalb XL-80 field corn was planted in plots 5.54 m x 5.54 m on
March 17, 1978. Plant spacing was 39.48 cm in north-south rows spaced
at 45.72 cm and the plant population was 72,000 plants/ha.

Four treatments were assigned in a four-replicate, randomized block
design: (1) no irrigation, (2) light, frequent irrigation irrigate
with just enough water to wet the soil to 30 cm depth when soil water
suction at 15 cm was between 150 and 500 cm of water, (3) medium, in-
frequent irrigation irrigate as in 2 except to 45 cm depth, and (4)
light, infrequent irrigation same schedule and rate as 2 except irri-
gation frequency was decreased during periods of grain filling.

The irrigation system consisted of "Microjet" sprinklers fastened
to black polyethylene tubing which was placed on the soil surface
between rows. The system delivered 1.7 cm/hr at 25 PSI.

2. Results and Discussion

The seasonal rainfall and irrigation distributions in 1978 are
shown in Fig. 6. The irrigation treatment numbers 1, 2, and 3 fn the
figure correspond to treatments 2, 3 and 4, respectively. Water balance
and yield data are given in Table 8. Yields were lower than obtained in
either of the 1977 experiments (Table 4) even though more total water
input occurred in 1978. The predicted drainage shows that the input
water could not be used as efficiently in meeting ET demand as apparently
was the case on a dif-ferent soil type and under different water input
conditions in 1977.

The simulated data (ET, net profile depletion, and water drainage),
though unverified by direct means in the experiment, reveal interesting
facts. The particular rainfall distribution pattern resulted in deep
seepage losses of water from irrigated treatments. The magnitude of the
losses was influenced by the evapotranspiration model used. Higher
actual ET values would be balanced by less deep seepage loss and/or more
net depletion of the soil profile. However, the latter quantity is simply
the difference between the storage at the beginning and end of the
season and does not indicate the level of depletion which may have
occurred during the season. In the unirrigated treatment, rainEall and
an adequately depleted profile must have occurred concurrently through-
out the season in order to produce the low level of predicted outflow.

















Planti.g,March 17,1978. arvestAu ?.
. '""" I IIIiiii 1 IIll




Irrigation 1 i 111




Rainfall


0 a I. .,.. It. .I


(0o D 100
DAYS AFTER PLANTING


140 N1


Rainfall distribution and irrigation schedules for three water
management treatments on corn, Gainesville, 1978.


Figure 6.







-18-


Nevertheless, without irrigation, the total water available to the corn
crop over the season and during critical periods was less than that
needed to produce grain yields which would offset the costs of produc-
tion.

Table 8. Effect of water management on yields of corn and peanuts and
simulated water balance, Gainesville, 1978.

Profile
I/ 2/ water 3/ 4/
Treatment-/ ET Irrig.-/ depletion-/ Drainage Yield-'
-- -- cm - kg/ha

Corn

Non-Irrigated (1) 35.77 -- 1.37 1.80 2.110 b
frrigaLtd (2) 44.92 25.46 0.44 17.19 7720 a
Irrigated (3) 44.36 24.23 0.47 16.55 7080 a
irrigated (4) 44.04 21.65 0.50 14.31 7070 a

Peanuts

Non-irrigated (1) 44.03 -- 8.72 26.40 3780 a
Irrigated (2) 50.68 15.44 1.75 28.32 4190 a
Irrigated (3) 48.64 13.29 2.43 28.78 4390 a

-/See text for details.
2/
-/Rainfall, 36.2 cm on corn and 61.7 cm on peanuts.
1/Net seasonal loss from the soil profile.
A/Grain for corn and pods for peanuts.

From the standpoint of irrigation water-use efficiency in a humid
climate with sandy soils, these data are not very encouraging. On the
other hand, a non-irrigated crop production system would have been an
economic disaster. The irrigation scheduling strategy was designed to
minimize leaching losses s and at the same time minimize crop damage
from drought. The degree of success actually achieved with the water
management strategy used is not easy to measure, but a brief analysis
of tensiometer and water content data is instructional.

It is evident from hydraulic head values during the dry period from
53 to 75 days (Fig. 7) that irrigation in treatment 2 was producing a
net downward water flow below 45 cm depth. In contrast, the hydraulic
head distribution with depth in the non-irrigated treatment (Fig. 8)
showed a net flow gradient upward, reaching increasing depths with time








-19-


HYDRAULIC
-800 -600


HEAD (cm)
-400


Hydraulic head distribution with soil depth under
irrigated corn during a dry period 53 to 75 days
after planting, Gainesville, 1978.


-200


I







I




PLOT 11-2 /


53 DAYS '

4.---A 61 DAYS
.-. 68 DAYS
a---* 4 73 DAYS
X-X 75 DAYS '


S.dH
, dz
e -


100

I-
120 "'


140


160


160


Figure 7.








-20-


HYDRAULIC


-00


HEAD (cm)

-/.O


Hydraulic head distribution with soil depth under
non-irrigated corn during a dry period 53 to 75
days after planting, Gainesville, 1978.


-600


-200


I
I,
% I


















PLOT III-1

53 DAYS
-- 61 DAYS
C. '


















e-w 6 8 DAYS /

.... 73 DAYS

---- 75 DAYS
,_K75 DAYS


--o 400 -..0


I00

I-




140





180


Figure 8.








-21-


Lo 150 cm. Changes in the neutron measured water content distribution
with depth and time for the non-irrigated treatment are shown in Fig. 9.

Thlie waLer flux condition in the 150-180 cm zone throughout the
season for the same two treatments is shown in Fig. 10. Since flow
occurs from higher to lower hydraulic head values, the net flow direc-
tion was downward for nearly the whole season in treatment 2 (irrigated).
In treatment 1 (non-irrigated) the soil dried out more than in treatment
2 and Lhelre was an 18-day period (112-120 days) when the net flow direc-
Lion was upward. The water flux through the 150-180 cm zone could be
calculated from the Darcy flow equation using the hydraulic head gradi-
ent and the water content dependent hydraulic conductivity.


Table 9. Effect of water management on periodic water depletion rates
of corn and peanuts, Gainesville, 1978.


2/
Water depletion rate on treatments-2/

Day:s1 _1 2 3 4
- - mm/day - - -

Corn

4-11 3.66 2.92 3.64 --
12-26 2.29 2.57 1.70 --
27-40 2.08 1.79 2.52 --
41-68 3.71 4.02 5.88 6.69
69-77 3.65 9.86 7.34 8.93
78-89 6.19 6.66 5.56 6.60
90-102 3.31 9.62 7.77 8.01
103-138 5.62 8.05 8.27 8.27

Peanuts

25-33 5.11 4.39 5.09 --
34-47 5.24 6.92 5.30 --
48-57 1.92 3.21 4.43 --
58-68 4.75 6.02 5.77 --
69-103 11.06 11.20 11.48 --
104-117 4.18 7.43 5.28 --
118-130 1.28 4.73 4.52 --


- Days from planting: corn, March 17, 1978, and peanuts,


May 10, 1978.








-22-


%/ VOLUMETRIC WATER CONTENT


0 5


Figure 9. Soil water content distribution with depth under
non-irrigated corn during a dry period 53 to 75
days after planting, Gainesville, 1978.*


U


I
H"


UJ
0
. LD
















DAYS AFTER


PLANTING


-200


-250


- -300
0

S-350


I -400



-200


-250


-300


I t (0









cvor
a H



mo
' 0r



m rt



a C

0
H' H







(D(








-24-




Further information on the water balance resulting from the irriga-
tion strategies in this experiment is provided by the measured daily
water depletion data of Table 9. Water input data and periodic neutron
measured soil water contents were used to make the calculations. It is
likely that plants were under water stress during these periods when
daily depletion was less than expected ET (Table 1). This condition
existed in all treatments from days 12-40, and in treatment 1 for days
41-77 and 90-102. During the 12-40 day period, water depletion was low
due to incomplete ground cover rather than plant water stress. Periods
of potential drainage loss from the soil profile were indicated by water
depletion rates in excess of the expected ET rates. This was the situa-
tion for all periods other than the potential drought stress periods
indicated above. These depletion data, like the simulated water balance
and tensiometer data do not reveal the true success of the irrigation
strategy, because measured water input and soil water contents are sub-
ject to sampling error.


E. Peanuts, 1978

The objective of this experiment was to determine the effect of
irrigation strategy on periodic water depletion rates and on IrrigatlJon
water-use efficiency. The peanut experimental ploLs were located ;mat-
cent to the above corn plots which were a part ol Lihe overall study.

1. Methods

'Florunner' peanuts were planted in north-south rows on May 10,
1978. The distance between rows was 0.92 meter. The field plot design
was a randomized block with three water management treatments and four
replica-tions. Plot size was 6.15 m x 5.54 m. The three treatments
were (1) control (rainfall only) (2) light, frequent irrigation -
irrigate to 30 cm soil depth when soil water suction at 15 cm was greater
than 200 cm and (3) medium, infrequent irrigation irrigate to 45 cm
soil depth when suction at 15 cm was greater than 200 cm.

As in the c.orn experiment "Microjet" sprinklers were used but the
system was redesigned to apply water over the peanut canopy. The black
polyethylene tubing was fastened to narrow wooden slats mounted about
0.6 m above the ground. Water was applied at 1.36 cm/hr at 25 1'1,.

2. Results and Discussion

Seasonal rainfall and irrigation distributions are shown in Fig.
11. Irrigation treatment numbers in the figure correspond to treatment
numbers 2 and 3 respectively. Yield and water balance data are given
in Tables 8 and 9.

























Planting, May 10,1978. Harvest, Sept.19;78.




2-
SIrrigation 2 I I I II II
2-




8 Rainfall




4l


-10 0


20 40
DAYS AFTER


60 O90
PLANTING


100 120 140


Figure 11.


Rainfall distribution and irrigation schedules for two
treatments on peanuts, Gainesville, 1978.


water management








-26-


Much more total water was available in 1978 than in 1977, and the
yields were slightly higher. Yield increases due to irrigation were
not significant. However, late season leaf spot damage contributed to a
large variability between plots. The unirrigated treatment was wilted
for the last 12-15 days of the season. Apparently the plants had already
produced a near normal pod load before the drought became severe.

Excessive rainfall in the 60-90 day period contributed to a large
simulated drainage from the profile in all treatments. Note that in
contrast to the corn experiment the irrigation schedule interacted with
the rainfall distribution in such a way as to minimize the increased
water outflow over the unirrigated treatment. However, the nonirrigated
soil profile was left in a more depleted state at the end of the season.
The predicted net profile depletion level at the end of the season was
lower than the neutron measured level (4.65 cm vs. 5.98 cm). This is
not a great difference, but when coupled with differences in the opposite
direction for the irrigated treatments (treatment 2, 11.53 cm vs. 7.16
cm; treatment 3, 10.85 cm vs. 7.51 cm) it means that the differences
between the nonirrigated and irrigated treatments in simulated ET values
were not as large as they were in reality. Adjustments on the basis of
the measured profile water contents would be in the direction of Lower-
ing the simulated ET for the nonirrigated treatment and increasing these
values on the irrigated treatments. On the other hand, the discrepancy
could be due to uncertainties in the amounts of irrigation input. An
overestimate of the input would result in a calculated water content
higher than measured.

These and earlier data support the not-so- obvious fact that irriga-
tion water must increase ET over non-irrigated ET in order to avoid an
increase in profile outflow equal to the amount of irrigation water
applied. In a humid climate with sandy soils it will not be possible to
manage the irrigation of crops for an economical production without some
additional contribution to the water drained below the root zone. The
estimated ET values obtained in the corn and peanut experiments in 1.978
appear to be reasonable (Table 8).

Daily water depletion rates (Table 9)"were calculated for periods
of varying length from measured water input and neutron measured changes
in soil profile water content. These depletion rates, when compared
with the appropriate expected ET rates of Table 1, show the periods of
potential drought stress (depletion < expected ET) and of potential
drainage outflow (depletion > expected ET). For an irrigation treatment,
a depletion rate less than expected ET would indicate that irrigation
quantities were not adequate to prevent plant water stress. Treatment
2 in the 48-57 day period is an example. In reality errors in measure-
ment of the soil water content or of the amount and uniformity of water
input could result in inaccurate water depletion estimates.








-27-


RooL density distributions were measured in treatments 1 and 2 on
';,p)Litembeur 31. 1978 (Fig. 12). Sampling was not extensive enough to
Imseasure any treatment differences, but the results are typical of those
found by Robertson et al. (1979, 1980). In this experiment, roots were
observed at 225 cm depth.

F. Soybeans, 1978

A newly installed irrigation field plot area with solid set impact
sprinklers was used to study the response of soybeans to irrigation
scheduling strategies.

1. Methods

The experimental area was located on Arredondo fine sand and con-
sisted of twenty-four 13.7 m x 13.7 m water management plots arranged as
a randomized block in four replications. Impact sprinklers on the
corners of each plot delivered water at the rate of 2.54 cm/hr in a
quarter circle pattern with a radius of 14.3 m. Five of the six-treatments
available were chosen for this study. Water management treatments
(main) were: (1) no irrigation, (2) irrigation, light rate and frequent,
(3) irrigation, medium rate and infrequent, (4) irrigation, light rate
and infrequent, and (5) irrigation, mixed light and medium rates.
Irrigation scheduling for treatments 2 and 3 was based on a tensiometer
reading of approximately 150 cm at a depth of 15 cm. Treatment 4 was
irrigated at a suction of 300 cm at a depth of 30 cm. The light rate of
irrigation in treatment 5 was scheduled in the same way as for treatment
2, but when the suction at 30 cm reached 300 cm a medium rate of irrigation
was scheduled. Soybean response was measured in terms of yield of
mature beans.

Subplot treatments were cultivars of soybeans Bragg and Cobb.
Planting was on June 15, 1978 in 76 cm rows with a population of about
258,000 plants/ha. Field dry beans were harvested November 7, 1978.

2. Results and Discussion

Rainfall distribution can be seen in Fig. 11. No rainfall occurred
between Oct. 1 (day 144, peanuts) and the end of the soybean season
(November 28). Irrigation distribution for the various irrigation treat-
ments are given in Table 10. Soybean yields increased linearly with the
amount of water applied (Figure 13). Apparently quantity was more
important for this season than the strategies of application.

Yields are given in Table 11 along with simulated water balance
results. Simulated ET was nearly the same for all irrigated treatments
indicating that the model was not sensitive enough to detect the expected
real ET decrease for treatment 4 where there was a measured decrease in






-28-


ROOT LENGTH DENSITY (CM ROOT / CM3 SOIL)


1.0


2.0


1.0


15

30

45


60


75

90

105

120

135

150


TREATMENT 2
(IRRIGATED)


TREATMENT 1
(NON-IRRIGATED)


Figure 12.


Peanut root density distribution with depth in
irrigated and non-irrigated treatments, Gainesville,
1978.


2.0


105

120

135

150








-29-


Table 10. Irrigation schedule, soybeans, Gainesville, 1978.


Irrigation Amount on Treatment


3


Date


4
cm -


2


1.27
1. 27
1.27
1.27

1.07
1.07
1.07
1.07

1.07

1.07
0.84
0.84

0.84


0.84

0.69

0.84

0.69
0.84


0.97
18.89


Numbers-/

5


1..27
1.27
1.27
1.27
--

1.07
1.70

1.19
--
1.47

1.27

1.07

1.07
--


0.84

0.69

0.84

1.19

1.27
0.97
19.72


- There were two irrigations
five treatments, but this
season total of 55.1 cm.


(June 16 and 19) of 0.84 cm each on all
was included as rainfall, making it a


1.27





1.07
1.07

1.47


1.47

0.84
1.14



1.27


1.27



1.07

1.27



13.21
E--


.liue 30
Aug. 18
23
25
26
28
30
Sept. 1
3
5
7
8
10
13
16
17
19
20
21
22
23
26
27
29
Oct. 4
6
10
11
12
21
Total


1.27
1.91
1.91

1.27
1.07
1.47

1.47

1.52
E--
1.07
1.27

1.07

1.07


1.27
E--



1.07

0.84

0.97

1.07
0.97

22.56






-30-


4 8 12 16 20


IRRIGATION,


cm


Figure 13.


Soybean yields as influenced by irrigation
amounts, Gainesville, 1978.


3000






2000


I000






0-
0







-31-


Table 11. Effect of water management on yield of soybeans and simulated
water balance, Gainesville, 1978.


Profile
it 2' water 3/
Trea Litmnent -- ET Irrig.- depletion- Drainage Yield
- - -cm- - kg/ha

Non-irrigated (1) 35.42 -- 11.24 30.91 976 c
Irrigated (2) 46.76 18.89 4.10 31.32 2690 a
Irrigated (3) 47.01 22.56 2.39 33.03 2832 a
Irrigated (4) 46.86 13.21 9.88 31.32 2311 b
Irrigated (5) 47.01 19.72 3.15 30.94 2668 a


- See text for details.
2/Rainfall, 55.1 cm.

- Net seasonal loss from the soil profile.


yield. The discrepancy could result from any one or a combination of
the Collowing input parameter deficiencies: an underestimate of ET, an
overestimate of field capacity, and an overestimate of the time needed
Ior water redistribution in the soil profile.

In contrast to other water balance data presented thus far, these
data predict very little influence of irrigation on drainage loss of
water. Nearly all of the irrigation input was used for ET and storage
in the soil profile. This is the desired objective of the irrigation
management in humid regions. However, in this case all of the irriga-
tion was applied during a prolonged drought so that a few small rainfalls
did not cause an overfilling of the soil reservoir. The relatively
large drainage outflow for all treatments was due to high rainfall
during the period June 12 through August 11.

The neutron-measured soil water profile status, the approximate
field capacity profile, and the soybean root density distribution are
given in Fig. 14. On October 27 the non-irrigated plot was essentially
at the permanent wilting point to a depth of about 135 cm. The irrigation
regime for treatment 2 (light, frequent) allowed the lower part of the
profile to become water-depleted as planned. Root densities were a
little larger for soybeans than for peanuts (Fig. 12). The decrease in
root density in the 30-60 cm soil depth zone was evident in both the
soybean and peanut data and was attributed to a compact soil zone created by
tillage.













ROOTS, cm/cm3


SOYBEANS 197B




LIGHT IRRIG
9/26


Figure 14.


Soil water content and soybean root density distribution with depth,
Gainesville, 1978.


60


90


120


150


180


FO CONTENT,







-33-


G. Corn, 1979

The purpose of this study was to determine the effect of drought on
grain yield of corn for different stages of plant growth and for two
Irrigation scheduling practices.

I. Methods

The field experiment was established on Lake fine sand with a solid
set, .impact sprinkler irrigation system delivering 2.54 cm of water per
hour. The plots were 13.7 x 13.7 m in size and arranged in a randomized
block statistical design in four replications. Water management treat-
ments were: (1) no irrigation, (2) irrigation, light rate and frequent,
(3) irrigation, medium rate and infrequent, (4) irrigation, same as
treatment number 2 except beginning at tassel, and (5) irrigation, same
as treatment number 2 except no irrigation during tasseling and silking.
Irrigation was scheduled when readings from tensiometers at 15 cm depths
exceeded 150-250 cm of water suction. Funk G-4507 corn hybrid seed were
planted on March 13, 1979 in 90 cm rows at a population of 71,000 plants/
ha.

2. Results and Discussion

Rainfall distribution and irrigation schedule data are shown in
FJLg. 1! and Table 12. Yields and simulated water balance data are given
In Table 13. Due to previous experiments in the field plot area, yields
were highly variable. Irrigated treatment differences were not signifi-
cant. Nevertheless, the omission of irrigation during tasseling and
silking (treatment 5) resulted in a yield not significantly larger than
the non-irrigated treatment.

The water balance data shows the effect of poor rainfall distribu-
Lton in relation to the water retaining properties of the soil. Had the
11 cm outflow from the non-irrigated treatment been available for ET,
the expected yield would have been near the maximum obtained in the
experiment.

A summary of results from the four Gainesville corn experiments is
presented in Fig. 16 and 17. In 1977, the very low yield for nonirri-
gated corn did not fall in the region of linear response and was omitted.
The reasonably good regression between yield and irrigation (Fig. 16)
indicates a remarkable similarity of drought factors for the three
years, although a much better linear fit of the data would be obtained
if the 1977 data were excluded. On the other hand, the response may be
curvilinear as found by (Skogerboe, 1979). It is apparent that total
water input was more important to grain yield than scheduling strategy
during these years. The regression coefficient of 143 kg/ha-cm (5.79
bu/acre-in) is a 3-year average measure of irrigation water-use efficiency.












8.20
, I .. . .


41


10


30


CORN


50


II I I, I


70


1979


90


1 I I


110


130


TIME, days from planting
Figure 15. Rainfall distribution and irrigation schedule for treatment 2
of corn experiment, Gainesville, 1979.


EO
U
z'
0

(9


I* ... ... ... .... .... .


n








-35-


NoLe that coefficients would be higher if data for each year were taken
.I;e para l l ,y.

Tlhe regressi n of grain yield on simulated ET(Fig. 17) yields a
si lghtly higher R -value and a larger regression coefficient(661 kg/ ha-
cm). Again, the data suggest that seasonal ET was more important than
irrigation scheduling strategy. The regression coefficient is much
larger than the 147 kg/ha-cm value found in Colorado (Skogerboe, 1979).
However, calculated values from corn grain data of Hillel and Guron
(1973) in Israel were 540, 440, and 450, respectively for 1968, '69, and
'70. Our ET simulations may be forcing the range of values within
unrealistically small bounds, especially for the higher levels of irri-
gation where the model allocated a considerable amount of the water
input to drainage outflow.


Table 12.


Irrigation schedule, corn, Gainesville, 1979.-


Irrigation amount


on treatment numbers-


Date


March 22
Apr il 20
May 7
17
20
23
26
28
30
June 6
10
13
18


Total


1- Treatment
42.47 cm.


0.84
0.84
1.07
1.27
1.91
1.91

1.91

1.91
1.91

1.91

15.48


3


0.84
1.65
1.65
1.91
2.54

2.67

2.54
2.54

2.54
2.54

21.42


4
cm -

0.84


1.27
1.91
1.91

1.91

1.91
1.91

1.91

13.65


5


0.84
1.06
1.07
--






1.91




4.88


number 1 also was irrigated with 0.84 cm on 3/22; rainfall,







-36-


Table 13. Effect of water management on yield of corn and simulated water
balance, Gainesville, 1979.


Treatment


Nbn-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)
Irrigated (5)


Profile
2/ water 3/
ET Irrig.- depletion-- Drainage
- - - cm - -


37.81
43.80
43.80
42.68
39.60


0.84
15.84
21.42
13.64
4.87


5.58
2.54
2.64
3.09
5.34


11.08
16.72
22.60
16.52
13.07


1See text for further details.
2Rainfall, 42.47 cm.
Net water loss from the soil profile.


Grain
yield
kg/ha


1434
5120
5187
4248
2690






-37-


10 20 30
IRRIGATION,


Figure 16.


40 50
cm


Yield of corn grain as affected by irrigation
amounts, Gainesville, 1977, 1978, and 1979.


9


8

c7

It?


VO
0




LJ






O
z
0
0


0-
0







-38-


9- GAINESVILLE


8 1977 .
ao 1978
x 1979
M7-


0
-6-
CV

5 0)



Z 4- 4-
< /


z 3-
CC x
(0


x

I l





0 30 35 40 45 50 55
EVAPOTRANSPIRATION (ET), cm

Figure 17. Yield of corn grain as affected by estimated
evapotranspiration, Gainesville, 1977, 1978,
and 1979.








-39-


CHAPTER III. FIELD EXPERIMENTS, LIVE OAK


Three water management field experiments were conducted on corn
during 1977, 1978, and 1979. The experimental site of three hectares
conusisted of predominantly Lakeland fine sand, a member of the thermic,
(onted f.ami ly of Typic Quartzipsamments. A subsurface asphalt layer
(ioil mol!.;tLre barrJer) was previously (1967 and 1970) installed in one
ol I wo-tLreatment main pots (0.37 ha) Ln four replicatLons (Saxena, et
ail., 1973). The barrier to water and nutrient flow was placed in a
cont iinuous strip (0.3 cm thick) at a depth of approximately 65 cm with
ov(erlapphing passes of a special sweep plow. However, the process did
niot resulL in a complete seal at the lines of overlap; consequently,
p)rrchiing of a water table on the barrier occurred only for very short
LIm.'s and only during large and intense rainfalls. Nevertheless, a
uniquely modified soil profile was created--one in which water was
retained in larger amounts and for longer time periods in soil located
above the barrier than for the case of naturally rapid drainage in the
untreated soil. With a life expectancy of more than 25 years, the
barrier system has the potential for increasing the proportion of rain-
fall utilized by plants. Moreover, the field facility provides an
unusual opportunity to study selected strategies for irrigation manage-
meint and their influence upon crop yield and water and nutrient balance
in sandy soils of humid regions.

'our water management subplots (24 x 24 m) were maintained in the
same' local lon over the three seasons. One of the subplots was non-
I rr i'igat ed and the other three were irrigated overhead from impact
pr i ik.l rs mounted on portable aluminum pipe.

Water balance simulations were performed similarly to that described
lor the Cainesville Experiments (Chapter II). Also, the basic irrigation
management strategy was to irrigate frequently at light rates (small
quantLLties of water per event) leaving part of the water-depleted soil
profile unfilled and available to store rainfall.

A. Corn Experiment, 1977

The purpose of this experiment was to determine interrelationships
between moisture barrier, irrigation strategy, plant population, nitrogen
fertilizer level, and grain yield for two corn hybrids. A second purpose
was to measure downward flux of water below the root zone under different
water management treatments.







-40-


1. Methods

The four water management subplots described above were split into
eighteen sub-subplots to which factorial treatments (two corn hybrids,
three plant populations, and three nitrogen rates) were assigned randomly.
Water management treatments were: (1) no irrigation, (2) frequent
irrigation with light rate (small quantities per event) when the soil
water suction at 15 cm depth exceeded 120 cm of water, (3) same as 2
except infrequently with medium rate, and (4) same as 2 except irriga-
tion was applied when soil water suction exceeded 600 cm. Water manage-
ment treatments on the asphalt barrier plots were designated by adding
the letter "A" to the above numbers. Corn hybrids DeKalb XL-80 and Funk
G-4507, were planted March 22, 1977 and harvested July 25. Nitrogen
fertilizer was applied only as a sidedressing, 30% at 16 days after
planting 70% at 45 days. Nitrogen levels were 134, 202, and 269 kg/ha.
Plant populations were 39, 59, and 79 thousand plants/ha in rows 76 cm
apart. Irrigation was by overhead impact sprinklers at 1.9 cm per hour.
Soil water data used in water balance simulatLons are gLven Iu the
Appendix Table. An extensive network of mercury manometer tens iometers
was installed in only sub-subplots planted with Funk G-4507 corn at the
highest plant population and nitrogen levels of all water management
subplots of two replications. Tensiometers were placed in depth inter-
vals varying from 5 to 30 cm with maximum depths of 45 and 300 cm,
respectively for the barrier and non-barrier treatments. Readings were
taken once-daily on Monday through Friday of each week. Evaluation of
these data will be reported elsewhere. However, calculations of down-
ward flux of water at 240 cm depths were obtained by multiplying gradients
of hydraulic head from tensiometer readings with hydraulic conductivities
obtained in an earlier study (Parra, 1971).

2. Results and Discussion

Rainfall and distribution and irrigation scheduling data are shown
in Figs. 18 and 19. Drought conditions were serious in most of May and
for shorter periods in April, June, and July. These water input distri-
butions are reflected in the soil water status as measured by LonsiomeLuru
(Figs. 20-24). Tensiometer data were shown as hydraulic heads aL vuaroius
depths and in two to four-day intervals. Thus, one can obtain a equal I La-
tive view of both water content and vertical water Flux directions.
Moreover, during dry periods the hydraulic head readings revealed Lhe
presence or absence of water absorption by roots.

At a given depth, water content decreases as the hydraulic head
becomes more negative. Water flows vertically in the direction of














11.30


1977

LIVE OAK


APR MAY JUN JUL


AUG SEP OCT


Figure 18. Rainfall distribution, Live Oak, 1977.


I ~


MAR


I li IL I, l 11L, l 11 ill111111lll U I I L, I.


S I I i


. .






-42-


1977 LIVE OAK


2A




3




3A




-4
'4




- 4A


MAR APR MAY


JUN JUL


Figure 19.


Irrigation
treatments
ments with
subsurface


distribution for six irrigaitlon
on corn, Live Oak, 1977. Trent-
the letter "A" included a
asphalt barrier treatment.


2

zO
0

<2


0


I -


I I


I I | | | | | I








APR MAY JUN JUL

1977 LIVE OAK DEPTH
-*- 15 cm
TRT 1 0-0- 45cm
S--- 120cm
E --- 150cm
U-100-









>-300 I-- .

-400-
>--300---- r f k-.--




-500
Figure 20. Hydraulic head distribution with depth and time under non-irrigated corn,
Live Oak, 1977. Soil surface was the datum.









APR MAY JUN JUL
O -PT
1977 LIVE OAK EP TH
--o-- 45cm
TRT 2 ---- 120cm
--0-- 150cm

E -oo -
U







3-200.---- ;

I
>- -300 ----- - f l *- -^-------


-400 -


-500--
Figure 21. Hydraulic head distribution with depth and time under irrigated corn,
treatment 2, Live Oak, 1977. Soil surface was the datum.










APR MAY JUN JUL
DEPTH
1977 LIVE OAK -15cm
----- 45cm
TRT 3 --- 120 cm
--o--- 150 cm







E -100








-400 -
>- -300----- -------------





-500
Figure 22. Hydraulic head distribution with depth and time under irrigated corn,
treatment 3, Live Oak, 1977. Soil surface was the datum,









APR MAY JUN JUL
DEPTH
1977 LIVE OAK --- 15cm
--o- 45cm
TRT 4 --a-- 120cm
-o- 150cm

E -100 -
U




-200 -W '





300--
I

-400 -


-500 -
Figure 23. Hydraulic head distribution with depth and time under irrigated corn,
treatment 4, Live Oak, 1977. Soil surface was the datum.











0




E
u-100



LU
I
200
U
(1
D


>- -300
I

-400


MAY


-500
Figure 24.


Hydraulic head distribution with time at a depth of 45 cm under non-irrigated
and irrigated corn grown on soil modified by a subsurface asphalt barrier,
Live Oak, 1977.


APR


JUL


JUN







-48-


decreasing (more negative) hydraulic head. For example in treatment 1,
non-irrigated (Fig. 20) water flow in April was downward between all
depths shown. In May, as the soil became dry, net water movement became
upward, first in the 15 to 45 cm zone and next in the 45 to 120 cm zone.
Wide temporal swings in hydraulic heads at the two upper tensiometers
resulted from alternate periods of droughts and rainfall. However, in
the 120 and 150 cm zone net water flow was downward throughout the
season. At the 120 cm soil depth soil water content decreased with not
much evidence of water withdrawal by roots until late season. Likely,
plants stunted in "above-ground" growth by the May droughts were also
limited in root growth.

Irrigation (Figs. 21, 22, and 23) attenuated temporal variability
in hydraulic head readings at 15 and 45 cm. Water flow throughout the
season was vertically downward at depths below 45 cm in treatments 2 and
3. In contrast, the less frequently irrigated treatment 4 exhibited
three periods of upward movement in the 45 to 120 cm zone. Root activity
as indicated by tensiometer readings was not easily discernable in the
120 to 150 cm data of the irrigated treatments because the soil profiles
were maintained in a moist state.

The influence of the asphalt barrier on hydraulic head at 45 cm
depth in non-irrigated and irrigated treatments is shown In Vig. 24.
Irrigation essentially eliminated drying of the soll at 45 cm since the
strategy was to restore to field capacity (Appendix Table) only the top
30 cm of the soil profile. Comparing the early and late season hydraul Ic
head values in Fig. 24 with those at the same depth in Flgs. 20 to 23
reveals the higher soil water retention in soil above barrier. Tn
addition, the fluctuation of hydraulic head was less over the barrier.

Simulated (model) water balance data and corn yields are given in
Table 14. Yields will be discussed later. In the non-irrigated treat-
ments all input water and a considerable soil profile depletion was used
for evapotranspiration. On the other hand, irrigation caused a substan-
tial deep seepage or drainage loss in some cases. Assuming that the
simulated ET was correct, it is evident that several rainfall events
occurred when the soil profile was not sufficiently depleted to retain the
amount which fell. The more water-conserving treatments 4 and 4A reduced
drainage loss but also resulted in reduction in yield. One of the more
striking results in Table 14 is the effect of the increased soil water
retention by the moisture barrier in producing larger yields with less
irrigation. The results are unique in demonstrating improved produc-
tivity of drought sandy soils from an alteration of a purely physical.
soil property. Finally, these data call attention to the need, in humid
regions to utilize irrigation strategies which maximize use-efficiency
of water, fertilLzers, and pesticides. It's unlikely that one can







-49-


Table 1.4. Effect of water management on yield of corn and simulated
water balance, Live Oak, 1977.


1/eat
'1r ea tment t


Profile
2/ water 3/
ET Irrig.- depletion- Drainage
- - cm - -


Grain
Yield
kg/ha


Non-barrier


Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)


33.58
42.18
42.85
40.62


--
28.61
32.65
14.09


6.58
1.24
1.25
3.35


0.54
15.22
18.60
4.37


3015
7741
8081
6217


Barrier


Non-Irrigated (1A)
Irrigated (2A)
Irrigated (3A)
Irrigated (4A)


34.89
42.55
43.06
41.40


21.98
23.43
12.39


7.80
1.87
1.87
3.13


0.46
8.85
9.79
1.67


3367
9363
9266
8349


I/Se texL for details.


R./ra.nlal., 27.55 cm.

3/ Net seasonal water


loss from soil profile.


Table 15.


Estimates


of seasonal drainage in relation to water manage-
ment treatments, Live Oak, 1977.


1/Seasonal deep seepage at 240 cm depth
Seasonal deep seepage at 240 cm depth-


Tensiometer


2/
Treatment-/


Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)
Fallow


3/
Model-/
cm

0.54
15.22
18.60
4.37


Rep. I
cm

0.23
5.27
1.38
1.00


Rep. III
cm

2.47
0.73
7.93
2.77
12.33


1/During 121 days.

'/See text for details.
-/From Table 14.








-50-


always win the water balance game without losing crop yield. The
implications for agricultural production, economics, resource-use effi-
ciency, and pollution control have not been fully appreciated or explored.

In an attempt to obtain a measure of the real water balance (non-
barrier treatments) we calculated seasonal downward water flux at 240 cm
depths from hydraulic head readings at 210 and 240 cm. The results
along with simulated values from Table 14 are given in Table 15. Obvi-
ously, there is a need for more replication, and one might question
whether the frequency of readings were sufficient for such calculations.
Other problems include spatial variability in soil and in irrigation
water application. Treatment differences were not clearly delineated
from the tensiometer data, but the range of values is less than that
simulated by the model and possibly indicates that the model has under-
estimated ET and overestimated water outflow by deep seepage.

Yield data (Tables 16 and 17) were organized to show the following
significant two-way interactions: barrier x population, irrigation x
population, variety x population, variety x fertilizer, and irrigation x
fertilizer. Tukey's honestly significant difference (THSD) values were
calculated as an aid in making comparisons among the various average
yields. In Table 16, THSD values for comparison of barrier, water
management, and variety treatment means at the same population were 943,
1289, and 294, respectively. And THSD values for comparison of population
means at the same barrier, water management, and variety treatments wei-r
353, 499, and 353, respectively. In Table 17 THSD values for comparison
of water management and variety treatment means at the same fertil.z-,r
level were 1289, and 294, respectively. And THSD values for comnparilon
of fertilizer means at the same water management and variety treatments
were 499 and 353, respectively. The results of these comparisons will be
stated in general terms only.

The barrier improved yields at all populations, and yield increases
with population were greater on the barrier treatment. All irrigation
treatment yields were higher than the non-irrigated treatment at each
population. In addition, at the highest population, treatment 2 and 3
yields were significantly higher than treatment 4. Considering population
effects at the same irrigation treatment, there was no response For
treatment 1. Response to population in treatment 4 leveled off at
59,000 plants/ha. In contrast, for irrigation treatments 2 and 3, there
was an increase in yield with each increase in plant population. Funk.
G-4507 corn hybrid yields increased with increasing population while
DeKalb XL-80 leveled off at the intermediate population. There was no
difference in corn hybrids at the smallest population level, but highest
yields were obtained from Funk C-4507 at the intermedl.ate and highly plant
populations.







-51-


Table 16.


Corn grain yields as affected by plant populat-ion, moisture
barrier, water management, and corn hybrid, Live Oak, 1977.


Grain yield


r1/
Treatment-


Non-barrier
Barrier

Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)

DoKalb XL-80
Funk C-4507

-
2n1 an1.i/ha.


Tabl I e 17.


39,000J2
kg/ha

5593
6554

3105
7313
7340
6535

6182
5964


59,000
kg/ha

6559
7928

3371
8918
8981
7705

7062
7424


79 000


6640
8278

3098
9426
9702
7609

7102
7816


Corn grain yields as affected by nitrogen fertilizer level,
water management and corn hybrid, Live Oak, 1977.


Grain yield


r1/
Treatment-/


Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)

DeKalb XL-80
Funk G-4507


234?/
kg/ha

3322
8285
8240
7166

6447
7060


--See text for details.
-4ertilizer N, kg/ha.


202
kg/ha

3186
8723
8701
7246

6916
7013


269
kg/ha

3066
8648
9081
7437

6984
7132








-52-


Corn response to irrigation interacted with fertilizer level.
Grain yields increased with increasing fertilizer level only for irriga-
tion treatment 3 and only at the highest nitrogen level. For each level
of nitrogen applied, all irrigation treatments gave higher corn yields
than non-irrigation treatments. However, at the intermediate and high
nitrogen levels, yields were lower in irrigation treatment 4 than in the
other irrigation treatments. In relation to variety and fertilizer
level, Funk G-4507 corn was higher yielding than DeKalb XL-80 only at
the low nitrogen level and did not respond to nitrogen levels. Crain
yield of DeKalb XL-80 increased with increasing nitrogen application only
up to the intermediate level.

B. Corn Experiment, 1978

The overall objective was the same as for the 1977 experiments ex-
cept that corn hybrid was not included as a variable.

1. Methods

The four water management subplots described earlier were split Into
six sub-subplots (3 x 12 m) to accommodate factorial treatments of throw,
nitrogen fertilizer levels (134, 268, and 336 kg of N/ha) and two planil
populations (59 and 90 thousand plants/ha). Corn hybrid DeKalb XL-80 was
planted in 76 cm rows on March 16, 1978 and harvested on July 18. Water
management treatments and irrigation methods were the same as in the 1977
experiment. Soil water was monitored in the highest nitrogen plots in
replicates 1 and 3 using tensiometric and neutron techniques. Input
data used in the water balance simulations are given in the Appendix Table.

2. Results and Discussion

Rainfall and irrigation schedule data are shown in Figs. 25 and 26.
In Fig. 26, the late July through October data are from a soybean experi-
ment not reported here. Note that the drought period for the corn experi-
ment extended from early May through June.

Simulated water balance data and corn yields are given in Table 18.
Grain yields were less than in 1977. Yields for treatments 2A and 3A
(barrier) were more than 1.3 times those for treatments 2 and 3. A I of
the irrigation treatments resulted in a maximum ET of around 42 cm while
the ET of both non-irrigated treatments were nearly the same aL 36 cm.
These values are very close to those obtained in 1977 (Table 14). One
third or more of the irrigation water applied was allocated to drainage.
However, as in 1977, irrigation water was used more efficiently with the
barrier treatment. These results challenge us to develop water manage-
ment systems which will equal the effectiveness of the barr.ier-lrrigat Ion












1978
LIVE OAK


MAR APR


.1,


MAY JUN JUL


AUG


SEP OCT NOV


Figure 25. Rainfall distribution, Live Oak, 1978.


5_


4
-J
.J3

z
<2
0::


I ., .







-54-


1978 LIVE OAK

2
2-



2A
2-


E 3
u 3
2-



S_ 3A
< 2 -



4
2

0 -- 1 -- --- J -- I L -
4A
2 -


MAY JUN JUL AUG SEP OCT

Figure 26. Irrigation distribution for six irrigant lu.
treatments on corn, Live Oak, 1978. Treait-
ments with the letter "A" Included a
subsurface asphalt barrier treatment.









-55-


system. There was more drainage loss from the non-irrigated treatment
In 1978 than in 1977, although in both years the amount represented a
small percentage of the rainfall.

In contrast to several two-way interactions obtained in 1977, only
one was found in 1978--irrigation by population (Table 19). On the non-
irrigated treatment, the corn yield for the high population treatment
was less than that for the lower population. Population had no effect
within the irrigated treatments. This result is opposite to the 1977
results where marked response to population increase was obtained for
IrrigatLon treatment 2 and 3. However, the maximum number of plants/ha
was 79,000 in 1977 versus 90,000 in 1978. Considering the data for both
years, it is evident that for DeKalb XL-80, 90,000 plants/ha was larger
than required for optimum grain yield. At the same population, treatment
I means were smaller than all the irrigated treatment means (THSD =
L295). Yields were affected by both barrier and nitrogen fertilizer
level (no interaction) as shown in Table 20.

Table 18. Effect of water management on yield of corn and simulated
water balance, Live Oak, 1978.

Profile
1/2/ water 3/ Grain
Treatment- ET Irrig.- depletion- Drainage yield
- - -cm - - kg/ha

Non-barrier

Non-irrigated (1) 35.77 -- 8.05 3.42 2475
Irrigated (2) 41.81 16.20 1.06 6.59 6290
Irrigated (3) 41.81 19.02 1.04 9.39 6353
Irrigated (4) 41.81 23.32 1.14 13.79 6821

Barrier

Non-irrigated (1A) 36.53 -- 8.06 2.68 3272
Irrigated (2A) 42.10 19.06 0.88 9.00 8509
Irrigated (3A) 42.10 19.02 0.91 8.98 8624
Irrigated (4A) 42.10 18.56 0.91 8.52 7942

- See text for details.
2/Rainfall, 31.15 cm.
3/
- Net seasonal water loss from soil profile.







-56-


Table 19. Corn grain yields as affected by water management
and plant population, Live Oak, 1978.


Grain yield


STreatment-l/


Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)


Average


59,000o2


3324
7244
7578
7535


6420


90,000 Average
- kg/ha - -


2423
7555
7398
7228


2874
7400
7488
7382


6151


- See text for details.
2/
- Number of plants/ha.


Table 20. Corn grain yields as affected by
and nitrogen fertilizer level,


subsurface moisture bairr-Lr
Live Oak, 1978.


Nitrogen/
added -
kg/ha

134
268
336


Grain yield

No barrier Barrier Average
- - kg/ha - -


5114
5819
5521

5485


Average


6589
7606
7064

7087


5852
6713
6305


.N as NH NO3.

Response to nitrogen was curvilinear. The additional 67 kg N/ha
over the intermediate level was applied at tasseling and resulted in a
reduction in yield in comparison with the yield at the intermediate
level. We have not found an explanation for this puzzling result, although
there is evidence in Table 17 (1977 data) that response to nitrogen was
near maximum at the 269 kg/ha rate.

C. Corn Experiment, 1979

The overall objective was the same as for the 1977 and 1978 experi-
ments. However, plant population was not included as a variable in 1979.








-57-


1. Methods

The four water management treatments described earlier were split
into six sub-subplots (3 x 12 m) to accommodate factorial treatments of
three nitrogen fertilizer levels (84, 168, and 252 kg of N/ha) and two
corn hybrids (Coker 77 and McCurdy 67-14). Planting was in 76 cm rows
to give a population of 59,000 plants/ha. Water management treatments
were the same as in 1977 and '78, and they were assigned to the same
field plot locations. The same overhead irrigation system was utilized.
Soil water conditions were monitored in replicates 1 and 3 with tensio-
ieloers and a neutron moisture meter. Water balance simulations were
made In the same way as in 1977 and '78 (see the Appendix Table). Corn
was planted March 8, 1977 and harvested August 17.

2. Results and Discussion

Rainfall and irrigation scheduling data are shown in Figs. 27 and 28.
1The rainfall data in Fig. 28 after July apply to another experiment not
reported here. As in 1978, drought conditions were prevalent in May and
June of 1979.

Simulated water balance and corn yield data are given in Table 21.
Again we have maximum ET values around 42 cm with irrigation. For the
unirrigated plots, the low rainfall resulted in ET values of only 30-31
cm which were the smallest values for the three years. Yields from non-
irrigated treatments were surprisingly close for the three years. The
strange drainage results for the barrier treatments and treatments 1 and
4 (non-barrier) were a consequence of a single drainage-producing rain-
fall event in early April (Fig. 27).

In spite of a heavy seasonal irrigation for treatment 3, the yield
wa,; not comparable to treatments 2A and 3A on the barrier. Comparable
yields were obtained on treatments 3 and 4A, but note that 64% less
Irrigation was required on the barrier. It is likely that yields were
limited for treatment 3 by leaching losses of plant nutrients as well as
the relatively high deep seepage loss of water.

Irrigated treatment yields in 1979 were equal to or slightly better
than in 1977 with somewhat less irrigation use. There was one signifi-
cant main effect (variety) and two significant two-way interactions:
irrigation x fertilizer and irrigation x barrier. The short season
variety, McCurdy 67-14, produced a significantly lower yield than the
full season variety, Coker 77 (6074 vs 7826 kg/ha). The irrigation x
fertilizer interaction is shown in Fig. 29. Yields are averaged over
the barrier treatments. Corn did not respond to nitrogen applications
when there was a drought condition (treatment 1). Treatment 4, both
with and without the barrier, received the least amount of water of the

















1979
LIVE OAK


II,


MAR APR


MAY JUN JUL


AUG SEP


OCT NOV


Figure 27. Rainfall distribution, Live Oak, 1979.


I .


I A1


L I


.I






-59-


1979


LIVE OAK


2




2A




3




3A



4




4A
44


APR MAY JUN JUL


AUG SEP OCT


Figure 28. Irrigation distribution for six irrigation treat-
ments on corn, Live Oak, 1979. Irrigations after
July are not relevant to current study. Treatments
with the letter "A" included a subsurface asphalt
barrier treatment.


zO
I-
<2
(D
_0o


2

0

2

0






-60-


10,000



8000



6000



4000



2000


0


Irrigation
Schedules


1





I S I


84


168


252


TOTAL N


Figure 29.


(kg/ha)


Influence of nitrogen s idedress Ing rate s iid
water management treatments on yield of cori
grain, Live Oak, 1979.


'


)








-61-


Irrigated treatments, and response was limited to the intermediate level
of N.

Irrigation x barrier interaction is evident in the data of Table
22. The appropriate significant difference values (THSD) for testing
two means are: 1727 kg/ha for two irrigation treatments means at the
same barrier treatment, and 1964 kg/ha for two barrier means at the same
Irrigation treatment. Barrier treatment means were different only for
irrigation treatment 2 where yields were higher with the barrier.
Yields on the no-barrier treatment were larger for irrigation treatments
2, 3, and 4 than for 1. On the barrier treatment, yields were higher
for Irrigation treatments 2 and 3, and the yield for irrigation treat-
ment 4 was larger than for non-irrigated treatment 1.

Yield and water data for the three corn experiments at Live Oak
were subjected to the same regression analysis as the Gainesville corn
data (see Chapter II, C. Corn, 1979; Figures 16 and 17). The Live Oak
results are given in Figs. 30 and 31. Differences in response to the
barrier treatments were of sufficient magnitude to suggest separate
regression lines. For irrigation and ET, the response was steeper with
the subsurface barrier. The best-fitting regression equations were
obtained with irrigation as a variable rather than ET. Thus, we have
further evidence that the simulated ET values may not represent very
well the actual ET. Analysis of the tensiometer and neutron data
(beyond the scope of this report) should provide information useful in
resolving this question.

However, in the Gainesville analysis, the variable ET gave a better
fit than irrigation amounts. The response per unit of ET was larger at
Gainesville and Live Oak than the response per unit of irrigation. And,
in a further comparison of Gainesville and Live Oak data, note that the
best fit was obtained for Live Oak using irrigation amounts and for
(;ulne.ville using ET amounts. The response per unit of irrigation water
was Larger at Live Oak than Gainesville; while the response per unit ET
was larger at Gainesville. Some of the steep response at Gainesville
can be attributed to the depressed yields in 1979 due to soil fertility
problems associated with an earlier study in the experimental site. It
is unlikely that actual ET was reduced in proportion to the yield reduc-
t ion.

Finally, as noted earlier for Gainesville,the climatic factors at
Live Oak during the three growing seasons must have been very similar to
produce the results obtained. Further evaluation of these results and
addition of 1980 data may show that Gainesville and Live Oak data can be
combined for a regression relationship which will be useful for the
North Florida region.








-62-


Table 21. Effect of water management on yield of corn and simulated
water balance, Live Oak, 1979.


TreatmentI-


Profile
water /
depletion-
- cm -


ET Irig.2


Non-barrier


Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)



Non-irrigated (1A)
Irrigated (2A)
Irrigated (3A)
Irrigated (4A)


./See text for details.
2/Rainfall 25.29 cm.
3/Net seasonal water loss


from soil profile.


Table 22. Effect of water management and a subsurface moisture barrier
on yield of corn, Live Oak, 1979.


Grain yield


Treatment


No barrier
Barrier


Average


11/ 2


2955
3442
3198


6779
9874
8327


3 4 Average
- kg/ha - - -


8021
9761
8891


6814
7953
7383


97.93
123.69


-/Water management treatments; see text for details.


Drainage


Grain
yield
kg/ha


31.61
41.50
41.50
40.72


24.01
33.45
16.77


10.11
0.26
-0.17
2.45


3.79
8.05
17.06
4.79


2955
6779
8021
6814


Barrier


30.43
41.65
41.65
41.63


17.71
16.75
12.01


9.48
2.15
3.11
7.84


4.34
4.34
4.34
4.34


3442
9874
9701i
7953






-63-


5 10


15 20 25 30 35


IRRIGATION,


cm


Figure 30.


Influence of irrigation amounts and a subsurface
asphalt barrier on yields of corn grain, Live
Oak, 1977, 1978, and 1979.


10


9




O







-64-


IO0


LIVE OAK

No Barrier
1977
x 1978
a 1979
Barrier
o 1977
1978
o 1979












3 m


* *
* U


0 0


32 34 36 38


40 42 44


Figure 31.


EVAPOTRANSPIRATION (ET), cm
Influence of evapotranspirat lon an.d a sib.s.mrlnc'
asphalt barrier on yields of corn graln, Live
Oak, 1977, 1978, and 1979.


9F


a-



0

-aJ
O


uJ

z
r0


0*'
O
0


5h


2


030







-65-


CHAPTER IV. SUMMARY DISCUSSION

Results presented in Chapters II and III have important implica-
tions to Florida agriculture for scheduling strategies and agricultural
water policy. These findings represent an initial effort to fulfill the
need for water management information in planning and implementing the
efficient use of all resources. Recognition of this need and the ini-
tiation of appropriate research programs is of relatively recent origin
in Florida.

A. Water-Use Efficiency

Water-use efficiency (WUE) by plants may be expressed in several
ways depending on the nature of yields and water-use data available. WUE
is commonly defined simply as a yield per unit of water used per season
in ET, or total depletion (ET + runoff + drainage). Total depletion is
readily measurable since it is also the sum of rainfall, irrigation, and
change in soil water storage. Actual ET is not easily measured, hence
the common use of estimates. Crop yield, Y, may be expressed as total
dry matter or some fraction of it, for example, grain. Crop yield increase
(Y-Yo, where Yo is yield without irrigation) per unit of irrigation (I)
provides another useful measure of water-use efficiency. In this report
total grain yield, yield differences due to irrigation, irrigation
amounts and estimated ET are used to calculate water use efficiencies
based on irrigation [WUE(I) = (Y-Yo)/I] and on ET [WUE(ET) = Y/ET].

Calculated WUE from the corn experiments are given in Table 23 for
Gainesville and Table 24 for Live Oak. Although these data are related
to the regression analysis in Figs. 16, 17, 30 and 31, they permit a
somewhat more detailed evaluation of the individual treatment effects in
each experiment. The slope from the regression equations (irrigation-
based, Figs. 16 and 30) are comparable to individual treatment WUE(I)
values. However, the WUE(ET) values are based on the whole of ET and
not on a threshold value as is the case in the regression analysis.
Consequently, the slope of the regression equation is larger than the
calculated WUE(ET) values. Also, this line of reasoning suggests that
WUE(I) should usually be larger than WUE(ET) for a specific irrigation
treatment. Irrigation in excess of ET needs by the crop or a large
rainfall soon after irrigation will reduce WUE(I).

Large WUE(I) values may be obtained when small amounts of irriga-
tion water are needed only at critical stages in plant growth. For
example, WUE(I) values calculated from the corn data of Robertson, et
al. (1973) in experiments on the asphalt barrier plots at Live Oak in
1971 were 0.28 tons of grain per ha/cm of irrigation on the non-barrier
treatment and 0.49 tons/ha-cm on the barrier. Corn data of Rhoads and
Stanley (1973) gave smaller WUE(I) values, 0.09 and 0.18 for 1970 and
1971. WUE(I) values calculated from corn data of Robertson et al. (1980)







-66-


ranged from 0.16 to 0.27 tons/ha-cm. In Colorado, Skogerboe, et al.
(1979), found WUE(ET) values for corn ranging from 0.11 to 0.15 tons/ha-
cm. The "best treatments" regression coefficient (Y/ET) was 0.18.
Calculated WUE(ET) values from three years of corn grain data of Hillel
and Guron (1973) in Israel ranged from 0.03 to 0.22 tons/ha-cm. Regres-
sion coefficients from their data were 0.54, 0.44, and 0.45 tons/ha-cm
in 1968, '69, and'70, respectively. These coefficients are in the range
of our three-year average values given in Figs. 17 and 31.

Table 23. Water-use efficiency of corn, Gainesville, 1977-79.

2/
Water-use efficiency (WUE)2/
1977 1977
Treatment- Exp. I Exp. II 1978 1979
- tons/ha-cm - -

Irrigation Basis

Non-irrigated (1) -
Irrigated (2) 0.20 0.16 0.22 0.24
Irrigated (3) 0.09 0.21 0.18
Irrigated (4) 0.23 0.21
Irrigated (5) 0.26

ET Basis

Non-irrigated (1) 0.01 0.06 0.04
Irrigated (2) 0.19 0.17 0.17 0.12
Irrigated (3) 0.19 0.16 0.12
Irrigated (4) 0.16 0.10
Irrigated (5) 0.07

-/See Chapter II for details.

- Calculated from yield and water balance data in Tables, 4, 8, and 13.
Water-use efficiency irrigation basis and ET basis abbreviated in text
thusly: WUE(I) and WUE(ET).

In Tables 23 and 24, we find several treatments with WUE(I) values
less than WUE(ET). In all cases, these treatments received rather large
irrigation applications. For treatments 2 and 3, Gainesville 1977, the
irrigation amounts were equal to or greater than ET (Table 4). Thus
either ET was underestimated or irrigation amounts were greater than
needed. On the other hand, irrigation amounts were not measured but
calculated for the spinkler system. For treatment 3 with no barrier
(Live Oak 1977 and 1979), the WUE(I) values were reasonable in magni-
tude, but still less than WUE(ET).












Table 24. Yield and water-use efficiency of corn, Live Oak, 1977-79.


Grain yield-=


3/WUE, irrigation basis
WUE- irrigation basis


Treatment-/


No barrier
kg/ha


Barrier
kg/ha


No barrier


-Barrier No barrier
- -ton/ha-cm- -


1977


Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)



Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)



Non-irrigated (1)
Irrigated (2)
Irrigated (3)
Irrigated (4)


2756
8487
9552
6315


2889
10290
11285
9466


0.20
0.21
0.25


0.34
0.36
0.53


0.08
0.20
0.22
0.16


0.08
0.24
0.26
0.23


1978


2603
5802
6617
6711


3810
8342
8781
8279


0.20
0.21
0.18


0.24
0.26
0.24


0.07
0.14
0.16
0.16


0.10
0.20
0.21
0.20


1979


2899
9045
11683
8764


4553
13680
12422
9783


0.26
0.26
0.35


0.52
0.47
0.44


0.09
0.22
0.28
0.22


0.15
0.33
0.30
0.23


--See Chapter II for details. See also Tables 14, 18, and 21 for irrigation and ET data.
21
-Yields differ slightly from those given earlier since the following treatments only were
selected: 1977, population and fertilizer levels of 79,000 plants/ha and 269 kg N/ha; 1978,
population and fertilizer levels of 59,000 plants/ha and 252 kg N/ha.
-IWater-use efficiency. Irrigation and ET basis designated as WUE(I) and WUE(ET) in text.


3/
WUE-z-


ET basis


Barrier


-







-68-


An interesting comparison of WUE(I) and WUE(ET) values is provided
by treatment 5, Gainesville 1979 (Table 23). A small irrigation amount
produced a large yield increase, but the WUE(ET) value was still very
low. This results from the nature of the response curve; low yields give
low WUE(ET) values and these measures of WUE should be used together
with seasonal regression coefficients (as in Figs. 5 and 13) in evaluat-
ing irrigation treatment effects.

The advantage of the moisture barrier system over the no-barrier
system is clearly evident in the differences in water-use efficiency
values (Table 24). In addition, the fact that water-use efficiency was
substantially increased by irrigation has important implications on
resource management. When there is appreciable drought, the decision
not to irrigate can be a wasteful use of water as well as of other
resources

Peanutsand soybeans gave lower water use efficiencies than corn
(Table 25) as expected for high protein and oil producing crops.

Table 25. Water-use efficiency of peanuts and soybeans, Gainesville.

Water-use efficiency (WUE)1/
Treatment- Peanuts, 1977 Peanuts, 1978 Soybeans, 1978
- - tons/ha-cm - -- - -

Irrigation Basis

Non-irrigated (1) -
Irrigated (2) 0.12 0.03 0.09
Irrigated (3) 0.13 0.05 0.08
Irrigated (4) 0.11 0.10
Irrigated (5) 0.09

ET Basis

Non-irrigated 0.06 0.09 0.03
Irrigated (2) 0.08 0.08 0.06
Irrigated (3) 0.08 0.09 0.06
Irrigated (4) 0.07 0.05
Irrigated (5) 0.06

1ee Chapter II for details.

1-alculated from yields and water balance data in Tables 7, 8, and 11.
Water-use efficiency irrigation basis and ET basis abbreviated in text
thusly: WUE(I) and WUE(ET).








-69-


In most cases, irrigation increased water use efficiency, and treatments
2 and were the best. The 1978 peanut data show very low WUE(I) values,
anld they are less than WUE(ET) values. This probably indicates that too
much Irrigation was used in that year on a deep-rooted crop. Note that
WUE(]ET) values were about the same in both years, except for the non-
IrrigaLed treatments.

The 1977 peanut data in Table 25 and in Figure 5 provide cn inter-
cstLLug contrast in data presentation and evaluation. Calculated WUE(ET)
UquatL ons because there is a threshold ET value, 23 cm. This is not the
case for the irrigation-based regression; the coefficient for soybeans,
Figure 13, is about the same as WUE(I) values in Table 25. The excellent
fit of the peanut and soybean data indicates that no significant between-
treatment variations in other management factors were present. A regres-
sion coefficient of 0.138 tons/ha-cm was calculated from peanut (Florunner)
data of Pallas et al. (1979) in Georgia.

Figure 5 provides another comparison in data representation one
year versus three years of data as in Figs. 17 and 31. As expected,
there was a considerable year to year variation in the simulated ET for
non-irrigated treatments. Since the yields were nearly the same this
coniLrbuted to a much lower R value than would have been obtained with
annuaIl regression equations.

ComparLson of irrigation and ET annual regression equations con-
trasts with comparisons of WUE(ET) and WUE(I) values. Theoretically,
the two regression coefficients could be nearly equal. The required
conditions are: (1) ET is actual seasonal ET, and (2) accurately
measured seasonal irrigation inputs contribute to ET in the same way for
each treatment, i.e. the quantity of irrigation not used to increase ET
(runoff and drainage losses and increase in soil storage) is the same.
Even if actual ET data were available, it is not likely that the second
requirement would be fulfilled in most irrigation experiments. Thus, in
view of the nature of the crop-response, water-use relationship discussed
earlier, it is reasonable to expect that the irrigation-based regression
coefficient will be less than the ET based coefficient. This was the
case for all experiments in the current study.

Regression of peanut yield on ET (Fig. 5) gives a regression
coefficient of 162 kg/ha-cm, a value larger than the regression coeffi-
cient for yield on irrigation:

Y = 124X + 2241 (R =0.99)

where Y is pod yield, kg/ha, and X is seasonal irrigation amount in cm.








-70-


Comparison of these two regression coefficients provides a very reveal-
ing test of the success attained in irrigation management. The ideal
ratio of b /bET (where b is the regression coefficient) would be one.
In this case we have 124/162 = 0.765, the highest ratio obtained in the
studies reported here (Table 26). These surprising results suggest
further evaluation in terms of economics of irrigation as well as of the
variable irrigation effectiveness achieved.

Table 26. Water-use efficiencies from the water management experiments
of this report in terms of yield versus water-use regression
coefficients and their ratios.


Location of Total Regression coefficient
Crop experiment years Y vs. I Y vs. ET
kg/ha-cm kg/ha-cm

Peanuts Gainesville 1 124 162
Soybeans Gainesville 1 84 144
Corn Gainesville 3 143 661
Corn Live Oak 3 160 471
Corn-- Live Oak 3 276 591


Ratio,-
b /bE
1 VI


0.76)
0.583;
0.216
0.340
0.467


- b and bET are regression coefficients of yield on irrigation and evapo-
transpiration, respectively.
2/
- Soil profile modified with a subsurface asphalt barrier.



B. Irrigation Scheduling

Results from the present study show that an irrigation management
strategy of frequent application at rates which only partially fill the
ET and drainage-depleted soil profile will conserve water while meeL.ing
the water needs of crops.

An irrigation scheduling plan is presented here whereby this con-
cept can be put into practice by farmers and other growers in Florida.
Perhaps more importantly, the basic concepts of the plan can be used by
industry in developing new irrigation systems and in designing new and
replacement irrigation installations. Moreover, there is a current
rapid development of commercial irrigation scheduling services based on
a growing sophisticatl ion In communication, measurement of the physical
system, computer simulaLtion, and crop growth modt'l development TooIn
needed by the farmer to utilize this method of schledulUng Irrlpgatiol lor
crop production are a low-cost rain gauge, a radio For weather forc.'1;Ii,
a chart of daily potential ET estimates (Table 1), and a shovel to aid
in making periodic observations of the top 30 to 45 cm of the soil
profile.








-71-


The basic plan is to irrigate the top 30 cm of the soil profile
oaltenii eiougIh to prevent more than short-term ( a few hours) wilting of
plantsL. The amount of water needed can be estimated by examining the
soil proflle for depth of water percolation 12 to 24 hours after a
measured application. A starting test irrigation amount per irrigation
event for sandy soils should be in the range of 1.5 to 3.0 cm (approxi-
mately 3/2 to 1 inch).

Once irrigation has been initiated, subsequent scheduling during
the drought is based upon the estimated daily ET and the amount of
water added in the last irrigation. For example, assume an irrigation
of 1.7 cm applied to a crop in Gainesville on May 15. The estimated ET
value from Table 1 is 0.411 cm/day. Thus, dividing 1.7 by 0.411 we have
four days or May 19 before the second irrigation is scheduled. A rain-
fall of about 0.5 cm would delay irrigation for a day.

There is an upper limit characteristic of the soil type, for the
quantity of water which can be stored in the soil root zone. And, only
a fraction of this amount can be transpired by the plant before tempo-
rary wilt develops. It is this fraction the usable soil water capacity,
which must serve as a maximum in the above calculations. New estimates
ar4' 1Odod throughout the season as root growth extends to increasing
soI depths. To obtain the estimate, observe the growing crop over a
period of rainless days following a rainfall which established a wet
soil profile throughout the root zone. When the plants begin to show
stress (temporary wilting) by mid afternoon or earlier in the day, it is
time for the initial irrigation. Add up the days since rainfall and
multiply by the expected ET for that period to obtain the estimated
usable water storage. To continue the above example, assume an elapse of
7 days without rainfall prior to the May 15 irrigation. Seven days
times 0.411 cm equals 2.88 cm of maximum usable storage for the particu-
lar soil and plant root depth.

Use of the maximum usable storage value in irrigation scheduling
can be seen in a further step of the example above. First of all, the
value is not needed as long as irrigation or rainfall does not completely
restore to capacity the partially depleted soil profile. Assume, a
rainfall of 1.6 cm one day after the May 15 irrigation (1.7 cm). The
maximum usable storage has been restored since rainfall plus irrigation
minus one day of ET equals 2.89 cm. Dividing 2.88 by 0.411 predicts an
irrigation date (May 23) 7 days after the rainfall on the 16th. If the
rainfall had been larger than 1.6 cm, the maximum usable storage value
of 2.88 would still be used in the calculations. On the other hand, in
both cases, the soil root zone would be restored to capacity, and one
could begin anew to determine an initial irrigation by observation.
There is a second point in the illustration just given a justification
for an irrigation amount which leaves some storage for rainfall. The








-72-


rainfall of 1.6 cm did not produce deep seepage loss of water and
nutrients.

In actual practice, the irrigation scheduling plan can be altered
in line with keen and experienced observation of weather conditions,
crop appearance and growth stage, and the soil water status. A probabi-
lity of rain may justify a decision to delay a scheduled irrigation or
to apply a smaller quantity if the irrigation system will permit. On
the other hand, unusually hot days with low relative humidity will cause
the actual ET to be higher than predicted and a shorter irrigation
interval will be necessary. Additional help in using the irrigation
scheduling plan can be obtained with soil water measuring devices such
as the tensiometer.

Some current irrigation installations cannot be used effectively to
apply the small amount per event on a frequent schedule. However, the
basic scheduling principles of the plan can be helpful in getting the
most efficiency from the available installation or in making modifica-
tions to it, and above all in planning of new installations. The above
plan is a simple one, but it is workable and will lay the ground work
for the more advanced irrigation systems and scheduling plans already in
the research and development stages.

C. Water Policy for Agriculture in Florida

The findings from the current study as well as others in the broad
field of soil-water-climate relationships can be integrated with in-
formation on the geology-soil-climate-hydrologic system in Florida to
develop a physically-based philosophy on water-use policy in agricultural
production.

The current study revealed an important characteristic of water use
by plants. Crop yield increased linearly with increasing seasonal ET.
Reduction in actual ET at one stage of crop growth cannot be recovered
at a later stage and yield reduction occurs irreversibly. Consequently,
water use is less efficient, and returns from other production inputs.
(capital, fertilizer, pesticides, fuel, and labor) are reduced. The
conclusion is that agricultural production must be based on a full
utilization of water up to the maximum ET demand of the atmosphere. The
adoption of an agricultural water policy based on this management object-
ive is compatible with Florida's unique soil-climate-hydrologic resource.

The average annual rainfall in Florida ranges from about 132 to 165
cm, while potential ET has been estimated to range around 100 cm.
Consequently, the excess of rainfall over actual ET recharges the water
storage capacity of soils, aquifers, lakes, and streams, and maintains








-73-


a net outflow of water from the State through streams and underground
seepage along the coast. The storage components of the hydrologic cycle
are the source of water for municipal, industrial and agricultural use.

Important differences in water requirement and use among the above
users need to be recognized in water planning and policymaking. Much of
municipal and industrial use is non-consumptive. It is not lost to the
atmosphere, but disposal of the physically and/or chemically altered
water is a necessary phase of the water-use operation. Multiple cycles
of use before discharge reduces the overall quantities needed. Muni-
cipal and industrial activities are commonly so concentrated on land
areas that the water requirement exceeds the rainfall input of those
areas. The resultant imbalance in local hydrology must be eventually
offset by water input from adjacent or remote sources.

On the other hand, agricultural water use is largely consumptive
since most of it is transpired through plants and evaporated from soil.
Water evapotranspired or consumptively used is finite in quantity, the
upper limit being equal to the evaporative demand of the atmosphere. In
Florida, this is less than average rainfall. Consequently, if only the
consumptive use is recognized as an agricultural use, then agriculture
does not cause an imbalance in the local hydrology.

There is a seasonal crop production water need which turns out to
be larger than consumptive-use need during that season. The reason for
this is the uneven rainfall distribution on a soil root zone of limited
water storage capacity. During droughts, the farmer must recall water
from stored sources mentioned earlier in order to meet evaporative
demands and to keep crops growing and producing efficiently. It is
inevitable that not all rainfall and irrigation inputs will be used for
evapotranspiration and storage in the soil; some will contribute to
runoff and deep seepage. The latter quantities are not consumptive
uses, do not represent a loss from the available water resource, and
should not be so designated if assigned as an agricultural use. Never-
theless, the agricultural producer must do his part to minimize surface
runoff and deep seepage during the crop production season, since non-
consumptive uses represent increased production costs in pumped water
not used by the crop and in fertilizers and pesticides lost by leaching.

Crop production enterprises in Florida which utilize seepage
irrigation require total water inputs greatly in excess of consumptive-
use or evapotranspiration needs. Consequently, alternative crop pro-
duction and water management systems may be needed in some of these
agricultural areas in order to meet the economic demands as well as to
avoid undue disturbance of the hydrologic balance. It should be stated
that the majority of the irrigated cropland in Florida is irrigated by
overhead sprinklers. Moreover, on an annual basis, except for perennial
crops, crop production water needs are usually less than potential








-74-


evapotranspiration because the land is fallow during part of the year.
In fact, ET water needs of agriculture are less than for natural vegeta-
tion.

In view of the characteristics of water use in agriculture and the
findings of this study, the objectives of research and education in this
field should be to develop crop production systems which maximize return
from production factors other than water, but which incorporate a water
management scheme which meets actual ET demands while minimizing the
loss of water by runoff and drainage. The implication in terms of water
policy are:

1. The Florida climate, characterized by rainfall in excess of
ET, justifies a full use of ET for crop production, especially for
annuals where the farmer can be considered to have accumulated a stored-
water credit from earlier rainfall in excess of ET. Irrigation to
maintain potential ET is the most efficient use of water.

2. In view of the economic and energy waste incurred by water
deficit during crop growth, there should be no drought-triggered re-
duction in allocation of water to agricultural enterprises which use
water within the bounds of average annual ET. Further justification in
terms of food and fiber needs can be made on the basis that not all of
agricultural cropland is under irrigation. Thus, production from
irrigated farms will be needed to make up for production inefficiencies
of non-irrigated farming.








-75-


LITERATURE CITED


1. Bruce, R. R., J. L. Chesness, T. C. Keisling, J. E. Pallas, Jr.,
1). A. Smittle, J. R. Stansell, and A. W. Thomas. 1980. Irrigation
of crops in the southeastern United States. USDA-SEA ARM-S-9. 54 pages.

2. Hillel, D., and Y. Guron. 1973. Relation between evapotranspira-
tion rate and maize yield. Water Resources Research. 9:743-748.

3. N.r is, A. W. B. 1979. Influence of irrigation strategy on water
use and yields of corn and peanuts. Master of Science Thesis,
University of Florida, Gainesville, Florida. 163 pages.

4. Pallas, J. E., Jr., J. R. Stansell, and T. J. Koske. 1979. Effects
of drought on florunner peanuts. Agronomy Journal. 71:853-858.

5. Parra, J. V. 1971. Effect of an underground asphalt barrier on
hydraulic properties of Lakeland fine sand. Master of Science
Thesis, University of Florida, Gainesville, Florida. 81 pages.

h. Rao, P. S. C., J. M. Davidson, and L. C. Hammond. 1976. Estima-
tion oF nonreactive and reactive solute front locations in soils.
In: Residual Management by Land Disposal. Proc. Hazardous Waste
Residue Symposium. Tucson, AZ. p. 235-242.

7. Rao, P. S. C., J. M. Davidson, and R. E. Jessup. 1981. Simulation
of nitrogen behavior in the root zone of cropped land areas receiv-
ing organic wastes. In: Simulation of Nitrogen Behavior in Soil-
Plant Systems (edited by: M. J. Frissell and H. van Veen). Pudoc,.
Wagoningen, The Netherlands, (in press).

8. Ran, P. S. C., J. M. Davidson, R. E. Jessup, and K. R. Reddy. 1981.
Simulation of nitrogen behavior in croplands receiving organic wastes.
[FAS Monograph and U.S.E.P.A. Ecol. Res. Series, (in press).

9. Rawlins, S. L., and P. A. C. Raats. 1975. Prospects for high-
frequency irrigation. Science. 188:604-609.

10. Rhoads, F. M., and R. L. Stanley, Jr. 1973. Response of three
corn hybrids to low levels of soil moisture tension in the plow
layer. Agronomy Journal. 65:315-318.

11. Robertson, W. K., L. C. Hammond, G. K. Saxena, and H. W. Lundy.
1973. Influence of water management through irrigation and a sub-
surface asphalt layer on seasonal growth and nutrient uptake of
corn. Agronomy Journal. 65:866-868.








-76-


12. Robertson, W. K., L. C. Hammond, J. T. Johnson, and G. M. Prine.
1979. Root distribution of corn, soybeans, peanuts, sorghum, and
tobacco in fine sands. Soil and Crop Sci. Soc. Fla. Proc. 38:
54-59.

13. Robertson, W. K., L. C. Hammond, J. T. Johnson, and K. J. Boote.
1980. Effects of plant-water stress on root distribution of corn,
soybeans, and peanuts in sandy soil. Agronomy Journal. 72:548-550.

14. Saxena, G. K., L. C. Hammond, and W. K. Robertson. 1973. Effects
of subsurface asphalt layers on corn and tomato root systems.
Agronomy Journal. 65:191-194.

15. Skogerboe, G. V., J. W. H. Barrett, B. J. Treat, and D. B. McWhorter.
1979. Potential effects of irrigation practices on crop yields in
Grand V\alley. EPA-600/2-79-149. 193 pages.

16. Varnell, R. J., H. Mwandemere, W. K. Robertson, and K. J. Boote.
1975. Peanut yields affected by soil water, no-till and gypsum.
Soil and Crop Sci. Soc. Fla. Proc. 35:56-59.








-77-


ABSTRACTS OF PUBLISHED PAPERS

1. Rao, P. S. C., J. M. Davidson, and L. C. Hammond. 1976. Estima-
tion of nonreactive and reactive solute front locations in soils.
In: Residual Management by Land Disposal. Proc. Hazardous Waste
Residue Symposium. Tucson, AZ p. 235-242.

A technique, based on the physical principles of water and solute
transport, was used to describe the position of nonreactive and/or
reactive solute fronts in a soil profile. The procedure estimates the
solute front location after infiltration and redistribution of the soil
water to "field capacity," and includes extraction of soil water by
plant roots between irrigation/rainfall events. Linear equilibrium
adsorption-desorption of the reactive solutes was assumed. The approx-
imation procedure was based on the principles that (i) the soil water
residing in all pore sequences participates in the transport processes,
and (ii) the soil water initially present in the profile is completely
displaced ahead of the water entering at the soil surface. An analysis
of published field and laboratory data on infiltration of nonreactive
solutes (Cl and NO3 ) indicated that these assumptions were valid.
Agreement between predicted solute front location using a sophisticated
one-dimensional transient flow models and the above procedure further
support the validity of the assumptions. Field data for chloride
movement in a sandy soil, in the presence of a fully established millet
crop, during a 60-day period were in agreement with the simplified
model. The major drawback of the present technique is in its failure to
describe the attenuation or spreading of a solute pulse as it is leached
through the soil profile.

2. Selim, H. M., L. C. Hammond, and R. S. Mansell. 1977. Soil
water movement and uptake by plants during water infiltration
and redistribution. Soil and Crop Sci. Soc. Fla. Proc. 36:101-107.

A numerical solution of the equation governing soil water movement
and uptake by plants during infiltration and redistribution was used to
investigate the influence of the amount of irrigation and soil non-
uniformity on deep seepage loss and soil water storage in the root zone.
For a uniform Lakeland sandy soil the deep seepage loss increased as the
amount of water applied in an irrigation event increased. However, the
presence of a low permeability lower layer in a two-layered soil profile
was beneficial in increasing soil water storage and in minimizing deep
seepage loss. Such a decrease in deep seepage loss was more pronounced
when large amounts of irrigation were applied and the hydraulic con-
ductivity of the lower layer was extremely small. For small irriga-
tions, the influence of a low permeability lower layer on deep seepage
loss was essentially negligible.








-78-


3. Rhoads, F. M., R. S. Mansell, and L. C. Hammond. 1978. Influence
of water and fertilizer management on yield and water-input
efficiency of corn. Agronomy Journal. 70:305-308.

Damaging plant water stress develops in corn grown on coarse-
textured, low water retaining soil of the southeastern U.S. during 1 to
2-week periods without rainfall. However, in most years rainstorms
cause leaching of soluble fertilizers from the root zone. This study
was conducted to evaluate efficiency of water input in terms of corn
grain yield per unit of water, with two fertilizer systems on a Troup
loamy sand (Grossarenic Paleudult). Water management consisted of (a)
control--natural rainfall only, (b) trickle irrigation scheduled daily
(0.64 cm/day), and (c) trickle irrigation scheduled by tensiometer (1.30
cm/application). Tensiometers were placed in each treatment at six
depths between 15 and 150 cm below the soil surface and readings were
recorded daily. Methods of applying fertilizers were designated (a)
conventional--1/3 of N and all P and K applied broadcast preplant, and
remainder of N applied in two sidedressings; (b) program fertilization--
N-P-K applied broadcast in small increments (5, 5, 10, 20, 20, 20, and
20%) at 2-week intervals after corn emerged. Average grain yields for
the above water management treatments were 2,790, 4,160, and 5,700 kg/ha
respectively. Conventional fertilization had an average grain yield of
3,680 kg/ha and program fertilization 4,760 kg/ha. Water-input eflIl-
ciencies based on grain yields and total water inpul were 57, 42, and 76
kg/ha/cm for no irrigation, daily irrigation, and tensiometer scheduled
irrigation respectively. Highest irrigation water-input efficiency (150
kg/ha/cm) occurred with program fertilization and tensiometer scheduled
irrigation. Irrigtion water-input efficiency was lowest (10 kg/ha/cm)
with corn receiving daily irrigation and conventional fertilization.

4. Robertson, W. K., L. C. Hammond, J. T. Johnson, and G. M. Prine.
1978. Root distribution of corn, soybeans, peanuts, sorghum,
and tobacco in fine sands. Soil and Crop Science Soc. of Fla.
Proc. 38:54-59.

A knowledge of the root distribution of plants contributes to
decisions on how to fertilize, irrigate, select cultivars, and till the
soil more effectively. In this paper, we report root patterns to a
depth of 150 cm for corn (Zea mays L.) on Lakeland fs, a thermic,
coated, Typic Quartzipsamment; soybeans (Glycine max (L.) Merr.) on
Kendrick fs, a loamy, siliceous, hyperthermic, Arenic Paleudult; peanuts
(Arachis hypogaea L.) on Lake fs, a hyperthermic, coated, Typic Quartz-
ipsamment; sorghum (Sorghum bicolor L. Moench) on Arredondo fs, a loamy,
siliceous, hyperthermic, Grossarenic Paleudult; and tobacco (NicotLana
tabacum L.) on Lakeland fs, a thermic, coated, Typic Quartzipsamment.
The line intercept technique was used to determine root density in ').0
cm diameter soil cores. Lakeland fs was located on the Agricultural
Research Center near Live Oak and the remaining soils were on the








-79-


Agricultural Experiment Station farm, Gainesville. All soils were fine
sands; however, depth to clay for Kendrick and Arredondo was shallower
(120 to 150 cm) than for the remaining soils. The latter two soils had
higher Al contents associated with the clay but it was not believed to
be at a high enough level to be toxic to plants.

The crops ranked in order of root length density as follows: soy-
beans > corn > tobacco > peanuts > sorghum; but on a single plant basis
the order was: tobacco > corn > soybeans > peanuts > sorghum. On the
basis of length per unit weight of root the order of crops was: peanuts >
soybeans > corn = tobacco = sorghum. In most instances, root size was
greater and root length density less in the area of the plow sole at 30-
45 cm. The roots of corn, soybeans, and sorghum were finer, tobacco
coarser, and peanuts the same size in the middle between the rows as
compared to under the row.

5. Robertson, W. K., L. C. Hammond, J. T. Johnson, and K. J. Boote.
1980. Effects of plant-water stress on root distribution of corn,
soybeans, and peanuts in sandy soil. Agronomy Journal. 72:548-550.

The efficient recovery by crops of added nutrients and water is
Influenced by plant rooting characteristics. Published data on the
relationship of irrigation water and root distribution of certain crops
grown on sandy soils in humid regions are limited and needed. This
field ol study was a part of an investigation on three soil types to
determine the effect of plant-water stress and irrigation on root dis-
tribution of corn (Zea mays L.), soybeans [Glycine max (L.) Merr.], and
peanuts (Arachis hypogaea L.). The basic irrigation plan was to re-
plenish the water deficit in only the top 30 to 60 cm of the soil profile.
Depth of wetting and the degree of soil water depletion below the irri-
gated soil layer varied with irrigation frequency and amount of water
per application. Although treatments were not the same in the four
experiments reported, the major part of the study involved four water
management treatments: 1) no irrigation; 2) light, infrequent irri-
gation; 3) light, frequent irrigation; and 4) medium, infrequent irri-
gation. Seed yields were obtained at maturity and root length measure-
ments were made at full canopy. Root lengths per unit volume of soil
were measured by the line intercept method. Yields of corn and peanuts
increased with total amount of irrigation water used, but there was no
yield response to irrigation in the soybean experiment. Peanut and
soybean root growth (root length per unit area to a depth of 150 cm) was
not affected by water management. In the corn experiments, irrigation
increased the length of roots in the 150 cm soil profile. The largest
root length value was found in the light, infrequent irrigation treat-
ment. Crops vary in rooting response to plant-water stress and irri-
gation strategy. Limited rooting of corn under stress very likely
decreases the efficiency of water and fertilizer use.








-80-


APPENDIX TABLE


Soil water characteristic data


used in water balance


simulat ions.


1/
Soil type-



Lakeland, f.s.




Lakeland, f.s.
with barrier








Arredondo, f.s.






Lake, f.s.



Kendrick, f.s.


Depth
cm

0-32
32-92
92-152
152-180


0-8
8-16
16-32
32-40
40-44
44-52
52-64

0-32
32-60
60-132
132-148
148-168

0-28
28-180
180-200

0-16
16-32
32-128
128-136
136-148


Field capacity
cm3/cm3

0.100
0.085
0.075
0.065


0.100
0.120
0.120
0.150
0.200
0.220
0.260

0.100
0.085
0.070
0.085
0. 150

0.075
0.065
0.065

0.090
0.080
0.075
0.110
0.210


Lakeland soil located at Live Oak;
Abbreviation f.s. means fine sand.


others located at i nesv I I I .


1 5-b;ir
cm3 /cm3


0.025
0.022
0.022
0.022


0.025
0.025
0.025
0.022
0.022
0.022
0.022

0.0? 5
0.02?
0. 0()
0.020
0. 02()

0.022
0.022
0.022

0.022
0.022
0.020
0.025
0.100




Full Text

PAGE 1

'. IRRIGATION EFFICIENCY AND CONTROLLED ROOT-ZONE WETTING IN DEEP SANDS By L. C. Hammonu, H. S. Mansell, W. K. J. T. Johnson, and H. M. Selim PUBLICATION NO. 52

PAGE 2

t. "J ro IRRIGATION EFFICIENCY AND CONTROLLED ROOT-ZONE WETTING IN DEEP SANDS By L. C. Hammond, H. S. Mansell, W. K. Robertson, J. T. Johnson, and H. M. Se1im PUBLICATION NO. 52 FLORIDA WATER RESOURCES RESEARCH CENTER RESEARCH PROJECT TECHNICAL COMPLETION REPORT O"'RT Pro.iect Number A-034-FLA Annual Allotment Agreement Numbers 14-34-0001-7019 14-34-0001-7020 14-34-0001-8010 14-34-0001-9010 Report Submitted January, 1981 The work upon which this report is based was supported in part by funds provided by the United States Department of the Interior, Office of Water Research and Technology as Authorized under the Water Resources Research Act of 1964 as amended.

PAGE 3

-ii-TABLE OF CONTENTS Title TnhI<.\ of Contents Acknowledgements AhstrAct Chapter I. Introduction Chupter II. Field Experiments, Gainesville A. Corn, 1977 Experiment I B. Corn, 1977 Experiment II C. PeAnuts, 1977 n. Corn, 197R I':. 1'(';lT1Uts, 197H F. SnylH';llls, 197H (: Corn, 1979 Clwpter I IT. Field Experiments, Live Oak A. Corn Experiment, 1977 B. Corn Experiment, 1978 C. Corn Experiment, 1979 Chapter IV. Summary Discussion A. Water-Use Efficiency B. Irrigation Scheduling C. Water Policy for Agriculture in Florida Literature Cited Abstracts of Published Papers Apfwndix Table .' i if iii iv 1 3 4 11 12 14 24 27 33 39 39 52 56 65 65 70 72 75 77 80

PAGE 4

-iiiACKNOWLEDGEMENTS are grateful (or contributions from other professional members o( u1:Is(.)ciated research teams: J. M. Bennett. K. J. Boote, J. A. Cornell, .1. M. Davidson, J. W. Jones, P. S. C. Rao, and A. G. Smajstrala. A host o[ technical staff in the Departments of Agricultural Engineering, Agronomy, and Soil Science and at the Agricultural Research Center, Live Oak. participated in this research activity. We are thankful for the administrative support of W. H. Morgan, J. P. Heaney, and Mary Robinson of the Florida Water Resources Center. In addition, this is to express appreciation to Ron Jessup for the computer simulations, to Jennifer Johnson for the drafting, and to Sheila Whitlock and Barbara Stokes for typing the manuscript.

PAGE 5

-ivABSTRACT Ten field water management experiments were conducted with corn, soybeans and peanuts during 1977, 1978, and 1979 on well-drained sandy soils at Gainesville and Live Oak. Irrigation scheduling treatments varied from one to eight per experiment. A simple computer simulation model provided an estimate of seasonal evapotranspiration (ET), drainage, and change in soil storage. These data and calculated water use efficiencies served as a test of the effectiveness of various irrigation scheduling strategies, and provided information on crop response to water stress. Yields increased linearly with estimated ET. Regression of corn yields on estimated ET gave the following 3-year average results for Gainesville and Live Oak, respectively: Y = 66lX -22,895 (R2 0.90) Y = 47lX -12,964 (R2 = 0.73) where Y is grain yield, kg/ha-cm, and X is ET, cm. These findings were interpreted to mean that the most crop production use of water is obtainl::!d when water Is to meeL the full easonal ET needs imposed by the atmol:lphcre. [n udd.ltion. u I:Itrategy of light, frequent irrigation of only the top JO cm of Bandy soils will produce high yields with minimum deep seepage loss of water and nutrients.

PAGE 6

-1-CHAPTER I. INTRODUCTION Although the Florida climate is characterized by relatively high rainfall, two factors create a need for irrigation of crops for effi cient production: (1) inconsistent and often unfavorable rainfall distribution patterns, and (2) the predominance of sandy soils with low water and plant nutrient storage capacities. The annual rainfall distribution is highly'nonuniform with approximately 80-90% of the total occurring during the summer. Even during years with average rainfall distribution, irrigation is commonly required during spring and late autumn. Reasonably successful irrigation and other management practices have evolved over the years from farmer experience and scientific findings; nevertheless, recent population growth has accelerated the demand on our large but finite'water resource, and have made it necessary that nIl uscrs--agriculture, municipalities, and industry--develop more effi.cient water-use systems. Favorable factors in the Florida climate Buch as long growing season, warm temperatures, and high total rainfall justify the continued development of more efficient irrigation practices through application of present knowledge and the discovery of new knowledge through research. In fact, the need for new information exists for most of the humid southeastern United States where there has been a major increase in irrIgated agriculture during the past five years (Bruce et a1., 1980). In particular, it has been found that crop production practices associated with irrigated agriculture in arid regions are not applicable to humid regions without .considerable modification and adaptation. Humid region agriculture depends mostly upon rainfai1 for crop production, and irrigation is needed during relatively short but numerousdroughts. Consequently, when uneven rainfall distribution patterns are coupled with soils which have characteristically restricted root zones and thus limited water storage capacities, there is created a major problem in the scheduling of irrigation. Timing, application intensity, method of application, and amounts .of water applied affect the fraction of added water used by the plant, the leaching losses of ,pesticides and fertHizers, and. in many cases the aeration condition in the root zone. These problems have not yet received research attention commensurate with the high return potential for solution to them. Of considerable current interest and research attention, is a water management system based on high-frequency irrigation with shallow wetting of the root zone (Rawlings and Raats, 1979). In this system, small quantities of water are applied frequently to meet crop needs in a manner such that a closely following rainfall replenishes water depleted from deeper portions of the root zone without excessive loss to deep seepage.

PAGE 7

-2-In Florida, most well-drained sandy soils temporarily store less than one inch of water per foot.of soil"depth, so that the growing crop can develop water stress within 3 to 7 days (depending on rooting depth) following a rainfall or irrigation. When the soil deficit is replenished with irrigation the possibility always exists that unexpected rainfall will displace the infiltrated irrigation water so that it will be lost from the root zone as deep seepage. The severity of deep seepage loss of irrigation water increases as the water retaining capacity of the soil decreases. Deep seepage is also more severe if the soil is relatively wet at the time irrigation water is applied. And for a given'soil, deep seepage of water increases as depth of rooting decreases. In the final analysis, careful water management of crops in humid regions can be used to minimize deep seepage loss of water applied as irrigation. It is neither desirable or practical to completely eliminate deep seepage in agricultural soils. However, farmers should attempt to minimize the deep seepage loss of irrigation water. Farmers can minimize such losses by careful irrigation scheduling based on crop growth stage and rooting depth, soil water status and water retaining capacity, evaporative demand of the atmosphere, and rainfall The objectives of the current study were: 1. To test the hypothesis that, for humid reg10ns charactcrf.zcd by short duration droughts, improved water-use efficiency can be attained by replenishing a part rather than the f:ull soil water deficit in the root zone without adverse effects on crop yield. 2. To determine, under field conditions on deep sands, the influence of sprinkler irrigation management on the partition of total water input between evapotranspiration and deep seepage, and on the water-use efficiency of corn, soybeans, and peanuts. 3. To develop and validate mathematical models incorporating total water input amounts and distribution patterns and soil physical properties for the purpose of describing water infiltration, redistribution, deep seepage and uptake by plants. 4. To develop, from the field data and the mathematical models, efficient water management systems whIch can be :f.mplementecl readily by growers and utilized by water planners and pol:l.cymakers.

PAGE 8

-3CHAPTER II. FIELD EXPERIMENTS, GAINESVILLE Ten water management field experiments were conducted during the growing seasons of 1977,'78, and'79. Crops included corn, peanuts and soybeans. Three soil series were involved: Lake, Arredondo, and Kendrick fine sands. Lake fine sand is a member of the hyperthermic, coated family of Typic Quartzipsamments. In the test sites, this soil has sandy A and B horizons that extend to a depth of 210 cm or more with a saudy clay B2t horizon underneath. Arredondo and Kendrick fine sands are members of the loamy, siliceous, hyperthermic families of Grossarcn1.c and Arenic Paleudults, respectively. In the experimental area, Arredondo fine sand is similar to the Lake fine sand except that a sandy clay loam B2t horizon begins at depths ranging from 120 to 200 cm. The Kendrick fine sand profile consists of fine sand material over a fine sandy loam B2t horizon which begins at depths of 100 to 150 em. Irrigation was carried out by three overhead sprinkler systems: (1) hand-operated fan spray nozzle on garden hose, (2) low pressure "Micro jet" sprinklers, and (3) impact sprinklers. Water management treatment variables included timing of irrigation events and quantities per event. Irrigation intensities were 2.55 cm per hour or less so that no surface runoff occurred. Other soil and crop management practices were approximately equal to or better than those currently recommended for farmers. The soil water status was monitored periodically with tensiometer readings of waLer suction and neutron and gravimetric measurements of volumetric water content. Rainfall distribution was nonuniform with time over the three years of the study. Crop damaging droughts of varying durations occurred during each year. Crop response to water management was .determined as yield of marketable grain. Response to irrigation was analyzed using a simple water balance model (Rao et al. 1976, 1981) which incorporates estimated daily evapotranspiration (ET) rates (from monthly averages), measured soil water characteristics (field capacity, wilting percentage, and water redistribution time), estimated root depth with time, and a water extraction rate which equals the ET rate until 80% of the available water has been depleted. At that point, the extraction rate was decreased linearly with decreasing available water to zero at the wilting point. Water balance simulations used in the Gainesville and Live Oak experiments were based on potential ET rates calculated for Jacksonville (Table. 1). These ET rates were calculated using the Penman method from longterm weather records and from handbook tables of extraterrestial

PAGE 9

-4-radiation. In most of the simulations, a 10% downward adjustment of the ET rate was made for an incomplete crop canopy (0-25 days) during the early part of the season, and a 10% upward adjustment was made for later in the season (after 40 days). The latter adjustment was made in consideration of the non-average weather conditions associated with prolonged droughts. Table 1. Calculated daily potential evapotranspiration rates by months for three locations. Potential evapotranspiration Jacksonville Tampa Miami ------------------------cm/day ---------------January 0.112 0.152 0.191 February 0.163 0.206 0.252 March 0.234 0.277 0.323 April 0.343 0.371 0.39l May 0.411 0.432 0.1124 June 0.422 0.432 0.1.22 July 0.429 0.414 0.432 August 0.391 0.399 0.111.7 September 0.312 0.348 O.35h October 0.226 0.279 0.287 November 0.150 0.191 0.216 December 0.104 0.142 0.180 ---Annual Total, Jan-Dec (cm) 100.5 111.0 118.4 A. Corn, 1977, Experiment I Experiment I was designed to test the yield response of three corn hybrids under irrigation to subsoiling and to multiple sidedressings at a fixed total nitrogen level. 1. Methods This experiment was located on Kendrick fine sand soil and a solid set overhead sprinkler irrigation system delivered water at the rate o[ 0.51 cm/hour. All plots received irrigation averaging 1. 9 cm per appli. cation when the soil water suction at 15 depth reached 500 em. Treatments were arranged in a factorial statistical design of thrcl' corn hybrids, two soil conditions, and four nitrogen application schcmeH.

PAGE 10

-5Tho 5.5 x 7.6 m plots were arranged in a randomized block with four replications (Table 2). Prior to planting, subsoiling was performed with chIsel plow 30 em on center and to a depth of 35 cm. Corn was planted by hand in 45 em rows.with 30 cm spacing (71,760 plants/ha) on March 29, 1977 and harvested July 27 to Aug. 7 (soon after maturity). Values of various parameters used in the water balance simulation are given in the Appendix Table. The time dependence of root growth and water redistribution are shown in Figures 1 and 2. 2. Results and Discussion The rainfall distribution pattern shown in Figure 3 indicates that the corn growing season of 1977 was extremely deficient in water. Only 12.7 em of rainfall occurred, and 42.2 em of water was used as irrigation (Table 3). The simulated and measured seasonal water balance data are given in Table 4. Actual ET and water drainage (deep seepage) losses may be higher or lower than these estimates. Nevertheless, the simulated data provide useful information for evaluating results and planning further studies. Table 2. Nitrogen application treatments for corn Experiment I, Gainesville, 1977.-Number of applications Date NitrogeJ-/ kg/ha 1 May 6 224 2 April 29 112 May 13 112 3 April 29 56 May 13 112 May 27 56 4 April 22 56 May 6 56 May 20 56 June 6 56 l/Main plot treatments: Funk 4810, Pioneer 3369A, and MCNair 508 corn hybrids; subplot treatments: non-subsoiled and suhsoiled. 2/ -As IDiuN03' additional nitrogen applied at planting as part of mixed fertilizer (45-39-149-45 kg/ha N-P-K-Mg).

PAGE 11

014 20 40 60 E u 80 I w 140 160 ., TIMEj 40 days from planting 60 80 100 SIMULATED ROOT DEPTH CORN 120 1980 Figure 1. Estimated depth of corn root zone with time after planting. I 0'1 I

PAGE 12

II \ w r-Z 8 10 a:: \ w r-
PAGE 13

5 E4 u -13 -1 z <{2 cr 1 o f-f8.86 1977 GAINESVILLE L I I I J I. I III .. & J.L Il.. I I I I MAR APR MAY JUN JUL AUG SEP OCT NOV Figure 3. Rainfall distribution at Gainesville, 1977. I ex> I

PAGE 14

-9-Table 3. Schedule for water application to corn, Gainesville, 1977, Experiment I. Irrigation Date Date amount em em May 1 1.07 June 9 2.03 4 1.52 11 2.54 8 1. 52 14 2.29 12 1.52 20 2.11 15 1. 78 27 2.79 18 1. 96 30 2.36 21 2.18 July 6 1. 78 25 2.54 12 2.16 29 1.91 15 1. 78 June 2 0.64 22 1. 78 4 2.16 26 1. 78 Total 42.20 1/ --Rainfall, 12.7 cm. Tabl<.> L,. Corn grain yields and estimated water balances for Experiments I and II, Gainesville 1977. 11 ET Irrigated 48.21 42.20 No irrigation (1) 20.85 Irrigated (2) 47.33 47.59 Irrigated (3) 47.64 51.12 !/See text for details. Profile water 3/ depletioIr --em --Experiment I 2.13 Experiment II 8.02 5.44 2.24 Drainage 9.03 18.55 18.57 Grain yield kg/ha 8911 282 b 7991 a 9000 a amounts of 12.92 and 12.85 em occurred for Experiments I and II, respectively. l/Net seasonal loss from the soil profile.

PAGE 15

-10Grain yields (Table 5) for the three hybrids were in the range expected for corn growing in sandy soils under good water management. The Pioneer hybrid gave significantly higher grain yields than did the Funk and McNair hybrids. Subsoiling generally improves root density distribution in this sandy soil by mechanically disrupting soil zones which have become compacted due to tillage; however, this improvement would not be expected to provide increased corn yield during dry years (1977) when irrigation is used extensively. Three and four split applications of N gave higher corn yields than a single application. This result could be attributed to a larger volatilization loss from the single application. Table 5. Yield of corn grain as affected by hybrid, subsoiling, and number of nitrogen sidedressings. Treatment Hybrid: Funk 4810 ,McNair 508 Pioneer 3369A Soil condition: Non-subsoiled Subsoiled Nitrogen sidedressings:!/ Grain yieid.!I kg/ha 8671 b 8770 b 9293 a 8857 a 8965 a 1 8501 b 2 8733 ab 3 9283 a 4 9128 a l/Yield values followed by the same letter are not different at the 0.05 leveL 2/ See Table 2 for rates at each application.

PAGE 16

.. -11-B. Corn, 1977, Experiment II The purpose of this experiment was to determine the response of corn to irrigation regimes having small quantities of water applied frequently (small frequent applications) versus medium quantities applied less frequently (medium infrequent applications). 1. Methods Replicated (4 each) treatments were: (1) no irrigation, (2)'irrigation frequently in small amounts, and (3) irrigation less frequently in medium amounts. Plots 7 x 7 m were located on Lake fine sand. During irrigation events overhead sprinklers delivered water at the rate of 1.7 cm/hr. Irrigation times and amount are shown in Table 6 and rainfall in Figure 3. Water was applied when the soil water suction at 15 em depth reached approximately 200 cm. DeKalb XL-80 corn hybrid was planted in 45 cm rows at a,popu1ation of 95,000 p1ants/ha on March 24,1977. Corn reached maturity on July 27 and was harvested on August 15. Simulated and measured water data were obtained in the same way as ,for Experiment 1. See the Appendix Table for input data. Table 6. Schedule of water application to corn, 1977, Experiment II. ]rrigation Irrigation amount Date Treat. 2 Treat. em em May 2 1. 70 3 1. 70 8 1.30 1. 70 14 1. 70 2.55 18 3.12 19 3.40 22 3.40 23 25 3.40 26 4.25 June 1 3.83 3 4.39 5 3.54 8 3.40 1./ Rainfall, 12.85 em 3 Date --June 12 13 15 20 27 29 30 July 6 10 15 17 24 Total Treat,. em 3.68 4.80 3.82 3.68 5.38 4.24 47.59 2 Treat. 3 em 3.60 1. 70 4.38 4.38 4.24 3.82 51.12

PAGE 17

-12-2. Results and Discussion Seasonal rainfall for Experiment II was approximately the same as for Experiment I (Figure 3). Irrigation amounts and distributi.on patterns for treatment 2 and 3 did not differ as much as anticipated because of the extremely dry growing season (Table 6). The amounts may be overestimated as much as 10%. Water balance and grain yield data are given in Table 4 along with data from Experiment I. Yields from the irrigation treatments in both experiments are comparable even though the calculated amount of water applied was larger in Experiment II. The simulation model indicated excessive irrigation quantities in the latter experiment with a consequent drainage loss of water from the root zone. It is possible that the actualET was greater than calculated since the plots were small enough for a marked oasis effect and the climatic conditions during the long drought were far from the average conditions assumed in the calculation of expected ET (Table 1). C. Peanuts, 1977 The objective of this study was to determine the yield rC!:Iponse of peanuts on a deep, well-drained sandy soil to four Jrrlgat:lon trl.'atments. 1. Methods Small field plots 5.5 x 5.5 m located on Lake fine sand, were planted with 'Florunner' peanuts in 91 em rows on April 22, 1977. Four treatments were replicated four times: (1) no irrigation, (2) frequent irrigation, small amount, (3) infrequent irrigation, medium amount, and (4) infrequent irrigation, small amount. Water was applied by hand with a calibrated fan spray nozzle on a garden hose. Repeated passes along each row provided a rate of application of nearly 42 em/hr. Irrigation was scheduled for treatments 2 and 3 when soil water suction at 15 em reached 500 cm or more. In treatment 4, plant water stress symptoms were allowed to develop before irrigation was scheduled. Simulated water balance data were obtained as before (Appendix Table). Soil water content and suction were monitored with a neutron meter and tensiometers, respectively. Data are in a thesis by Nafis (1979). 2. Results and Discussion Rainfall and irrigation distribution data are shown in Figures 3 and 4. The first three irrigations were also applied initially to the

PAGE 18

-13-1977 GAINESVILLE PEANUTS -2 .. 1 I I II I 3 I I I r4 1 ro I I J APR MAY JUN JUL AUG Figure 4. Irrigation schedules for three water management treatments on peanuts, Gainesville, 1977.

PAGE 19

-14-no-irrigation treatment in order to get the plants up and established during a very dry period. Irrigation input, simulated ET, drainage from the root zone, and pod yields are given in Table 7. Roots were assumed to increase in depths linearly with time from 4 cm on day 1 to a maximum of 200 cm by day 90. The simulated ET values may be lower than actual since there was additional apparent drainage water under irrigation. Nevertheless, most of the added irrigation was allocated to ET and the measured pod yields increased linearly with the calculated values of ET (Figure 5). Table 7. Water balance and peanut yields, Gainesville, 1977. Treatment Non-irrigated Irrig., frequent Irrig., infrequent Irrig., plant stress II -Rainfall, 30.2 cm. ET 36.95 46.91 44.76 41. 78 Profile II water 21 Irrig.-depletion-Drainage 4.8 12.9 10.4 6.7 --cm -2.45 1.88 1.96 1.92 0.53 2.89 2.67 1.89 seasonal loss from the soil profile. kg/ha 2261 c 3817 a 3622 u 2999 h In 1975 studies, peanut pod yields were 4500 kg/ha when the amount of seasonal water depletion was 57 cm, and irrigation of 8 cm did not cause a yield increase (Varnell, et al. 1976). However, when water input was decreased by covering the plot during midseason rainfalls the yields were decreased to 3900 kg/ha with a water depletion of 48 cm. The latter yield is similar to that predicted from the 1977 results in Figure 5 if we assume that water depletion is a fair estimate of evapotranspiration. The peanut plant is often considered to be drought tolerant due to its characteristic deep rooting and reinitiation of blooming and fruiting after drought stress. Nevertheless, these data indicate that peanutH do respond to water management and that yields can be severely by an overall seasonal water deficit. D. Corn, 1978 This study involved two crops -corn and peanuts grown in adjacent plots on the same Lake fine sand site as the 1977 peanut exper.Lment.

PAGE 20

-15-4000--------------------------------------. PEANUTS 1977 o 36 38 40 42 44 46 48 EVAPOTRANSPIRATION, em Figure 5. The relationship of peanut pod yield to estimated evapotranspiration, Gainesville, 1977.

PAGE 21

-16-The objective was to determine the effect of irrigation strategy (frequency and quantity) on periodic water depletion rates and on irrigation water-use efficiency. 1. Methods De Kalb XL-80 field corn was planted in plots 5.54 m x 5.54 m on March 17, 1978. Plant spacing was 39.48 cm in north-south rows spaced at 45.72 cm and the plant population was 72,000 plants/ha. Four treatments were assigned in a four-replicate, randomized block design: (1) no irrigation, (2) light, frequent irrigation -irrigate with just enough water to wet the soil to 30 cm depth when soil water suction at 15 cm was between 150 and 500 cm of water, (3) mediunl, infrequent irrigation -irrigate as in 2 except to 45 cm depth, and (4) light, infrequent irrigation same schedule and rate as 2 except irrigation frequency was decreased during periods of grain filling. The irrigation system consisted of "Microjet" sprinklers fastened to black polyethylene tubing which was placed on the soil surface between rows. The system delivered 1.7 cm/hr at 25 PSI. 2. Results and Discussion The seasonal rainfall and irrigation distributions in 1978 are shown in Fig. 6. The irrigation treatment numbers 1, 2, and 3 In the figure correspond to treatments 2, 3 and 4, respectively. Water balance and yield data are given in Table 8. Yields were lower than obtalned in either of the 1977 experiments (Table 4) even though more total water input occurred in 1978. The predicted drainage shows that the input water could not be used as efficiently in meeting ET demand as apparently was the case on a dif-ferent soil type and under different water input conditions in 1977. The simulated data (ET, net profile depletion, and water drainage), though unverified by direct means in the experiment, reveal interesting facts. The particular rainfall distribution pattern resulted in deep seepage losses of water from irrigated treatments. The magnitude of the losses was influenced by the evapotranspiration model used. Higher actual ET values would be balanced by less deep seepage loss and/or more net depletion of the soil profile. However, the latter quantity is simply the difference between the storage at the beginning and end of the season and does not indicate the level of depletion which may have occurred during the season. In the unirrigated treatment, ral.ll[all and an adequately depleted profile must have occurred concurrently through--out the season in order to produce the low level of predicted outfiow.

PAGE 22

'I' "alltill" .arch 17 ,1971. :l Irr i.atioll 1 R I II 'IHB H H BlaB I 5! 2-:. Irri,ltioll 2 Cl' 01 II I! I! I! II g II II II 1111 !! Irri,ltioll 1 01 A BBBBggg UBgBqBBB lainflll ; C 2 01 II. -'U. II -10 o 20 40 DAYS AFlU "nest,AII I,1971. l II1II Figure 6. Rainfall distribution and irrigation schedules for three water management treatments on corn, Gainesville, 1978. I ..... --.J I

PAGE 23

-18-Nevertheless, without irrigation, the total water available to the corn crop over the season and during critical periods was less than that needed to produce grain yields which would offset the costs of production. Table 8. Effect of water management on yields of corn and peanuts and simulated water balance, Gainesville, 1978. Tr ea tmen t1./ Non-1.rrigated (1) Trrigal<.!d (2) Irrigated (3) Irrigated (4) Non-irrigated (1) Irrigated (2) Irrigated (3) ET 35.77 44.92 44.36 44.04 44.03 50.68 48.64 l/See text for details. Irrig }:./ Profile water 3/ depletion------cm 25.46 24.23 21. 65 15.44 13 .29 Corn 1.37 0.44 0.47 0.50 Peanuts 8.72 1. 75 2.43 36.2 em on corn and 61.7 cm on peanuts. l/Net seasonal loss from the soil profile. for corn and pods for peanuts. Drainage 1.80 17.19 16.55 14.31 26.40 28.32 28.78 kg/ha 2.110 b 7720 11 7080 a 7070 11 3780 a 4190 a 4390 a From the standpoint of irrigation water-use efficiency in a humid climate with sandy soils, these data are not very encouraging. On the other hand. a non-irrigated crop production system would have' been nn economic disaster. The irrigation scheduling strategy was designed to minimize leaching 10.saes s and at the same time minimIze crop damage from drought. The degree of success actually achieved with the water management strategy used is not easy to measure, but a brief analysis of tensiometer and water content data is instructional. It is evident from hydraulic head values during the dry period from 53 to 75 days (Fig. 7) that irrigation in treatment 2 was producing a net downward water flow below 45 cm depth. In contrast, the hydraulic head distribution with depth in the non-irrigated treatment (Fig. 8) showed a net flow gradient upward, reaching increasing depths with time

PAGE 24

-19HYDRAULIC HEAD (em) -800 -bOO -400 -200 0 PLOT 1I-2 53 DAYS I (!) Q I ..... 61 DAYS '-. 68 DAYS It----&!I 73 DAYS )t K 75 DAYS f dH 1 fer-I Z I Figure 7. Hydraulic head distribution with soil depth under irrigated corn during a dry period 53 to 75 days after planting, Gainesville, 1978. 40 bO 80 100 120 140 160 ISO '"' E (J '-J :c I-Cl. lIJ 0

PAGE 25

-800 HYDRAULIC -foOD PLOT 111-1 -53 DAYS Aw---... 61 DAYS m---GI 68 DAYS .. ---4 73 DAYS -75 DAYS \ -20HEAD (em) -400 -. \ -200 \ \ \ I I dH=l dz 0 I I I Figure 8. Hydraulic head distribution with soil depth under non-irrigated corn during a dry period 53 to 75 days after planting, Gainesville, 1978. 0 20 40 80 E u '-" 100 :J: a.. 120 w a 140 'bO 180

PAGE 26

I -21-Lo 150 em. ChangeR in the neutron measured water content distribution wIth deplh and time for the non-irrigated treatment are shown in Fig. 9. The waleI' flux condition in the 150-180 em zone throughout the (or the. same two treatments is shown in Fig. 10. Since flow oeeurs from higher. to lower hydraulic head values, the net flow direc tIon WllS downward (or nearly the whole season in treatment 2 (irrigated). In In'ntmcnt 1 (non-irrigated) the soil dried out more than in treatment 2 and t111're was an 18-day period (112-120 days) when the net flow direc-: lion was upward. TIle water flux through the 150-180 cm zone could be calculated from the Darcy flow equation using the hydraulic head gradient and the water content dependent hydraulic conductivity. Table 9. Effect of water management on periodic water depletion rates of corn and peanuts, 1978. Water depletion rate on 1 2 3 4 rom/ day :::-:: Corn 4-11 3.66 2.92 3.64 12-26 2.29 2.57 1. 70 27-40 2.08 1. 79 2.52 41-68 3.71 4.02 5.88 6.69 69-77 3.65 9.86 7.34 8.93 78-89 6.19 6.66 5.56 6.60 90-102 3.31 9.62 7.77 8.01 103-138 5.62 8.05 8.27 8.27 Peanuts 25-13 .11 4.39 5.09 34-47 5.24 6.92 5.30 48-57 1.92 3.21 4.43 ':>8-68 4.75 6.02 5.77 69-103 11.06 11.20 11.48 104-117 4.18 7.43 5.28 118-130 1.28 4.73 4.52 1lDays from planting: corn, March 17, 1978, and peanuts, May 10, 1978.

PAGE 27

.......... E: U '-./ :r: r-Q w 0 ..J 0 If) -22% VOLUMETRIC WATER CONTENT 0 5 10 15 20 0 20 \ 53 DAYS er--. 61 DAYS 40 6-.-Q 68 DAYS -73 DAYS bO \ ):. 75 DAYS I I 80 I 4 I 100 4-I t:.. 120 \ \ \ A 140 160 180 \ \ l-I \ I 200 \ A '-, 220 ---"". -, =---= 7,CiJ-_-. 240 Figure 9. Soil water content distribution with depth umkr non-irrigated corn during a dry period 53 to days after planting, Gainesville, 197H. 25

PAGE 28

-' "=j 1-'-ClQ 'i DAYS AFTER PLANTING ro .... 0 0 0 110 1 0 0 -200 --.... 150cm .... 1llC') ""'''''--, .. ----\.OHl::r -..J Hl III .. 180cm ...... -(Xl ro ::l ... nOQ .. /'1' ro .-ro -250 p. 1-'-C"::l Non-irrigated \ -300 \ 1-'-p. \ 'i 'i 'i III (Treatment 1) 'JIpward 1-'-0 OQ .... \ III 1-'-flow rt n -350 \ 1-'-0 \ I 0 ::r -< \ N ::l ro \ w III I I p. I -400 .. III 0 '-_ ... III /'1' H OQ /'1' H -... ;:J Irrigated ..", 0 -200 .. "" ... 150cm rtp. (Treatment 2) 180cm ro 0"0 Hlrt :x: .. ---------.... ::r -250 ........ n 00 0 'i ::l 1-'-.. /'1' ::r -300 0 III /'1' 1-'1-'-::l a ro ro 00 <: III 1-'-00 .... .... ro ..

PAGE 29

-24-Further information on the water balance resulting from the irrigation strategies in this experiment is provided by the measured daily water depletion data of Table 9. Water input data and periodic neutron measured soil water contents were used to make the calculations. It is likely that plants were under water stress during these periods when daily depletion was less than expected ET (Table 1). This condition existed in all treatments from days 12-40, and in treatment 1 for days 41-77 and 90-102. During the 12-40 day period, water depletion was low due to incomplete ground cover rather than plant water stress. of potential drainage loss from the soil profile were indicated by water depletion rates in excess of the expected ET rates. This was the situation for all periods other than the potential drought stress periods indicated above. These depletion data, like the simulated water balance and tensiometer data do not reveal the true success of the irrigation strategy, because measured water input and soil water contents subject to sampling error. E. Peanuts, 1978 The objective of this experiment was to determine the of irr I.gation on periodic water rates and on Irr lol\l JOII water-u!;e ef f lciency. '111e peanut pI aL::; atl.ln', cent to the above corn which were a part 01 Llw uvcrall Bllldy. 1. Methods 'Florunner' peanuts were planted in north-south rows on May 10, 1978. The distance between rows was 0.92 meter. The field plot design was a randomized block with three water management treatments and four rep1ica-tions. Plot size was 6.15 m x 5.54 m. The three treatments were (1) control (rainfall only) (2) light, frequent irrigation -irrigate to 3Q cm soil depth when soil water suction at 15 cm was greater than 200 cm and (3) medium, infrequent irrigation -irrigate to 45 cm soil depth when suction at 15 cm was greater than 200 cm. As in the c,orn "Microjet" sprinklers were but the system was redesigned to apply water over the peanut canopy. The black polyethylene tubing was fastened to narrow wooden slats abouL 0.6 m above the ground. Water was applied at 1.36 em/In: at 25 PSI. 2. Results and Discussion Seasonal rainfall and irrigation distributions are shown in FJg. 11. Irrigation treatment numbers in the figure correspond to treatment numbers 2 and 3 respectively. Yield and water balance data arc glven in Tables 8 and 9. :-,

PAGE 30

(! Pla"tilll, M., 10,1978. HI r UI t, Se pt.19:78. 1 -:IE 2 u ..., Ini ,It ie" Z l ________ __ __________ ; 01 !! I c ,,2 R RRRRRRA R R R R 01 Irri tiOll 1 8 "" c4 ... :IE :2 lIi.'.1I -10 40 120 DA'S AFUI 140 Figure 11. Rainfall distribution and irrigation schedules for two water management treatments on peanuts, Gainesville, 1978. .It I N VI I

PAGE 31

-26-Much more total water was available in 1978 than in 1977, and the yields were slightly higher. Yield increases due to irrigation were not significant. However, late season leaf spot damage contributed to a large variability between plots. The unirrigated treatment was wilted for the last 12-15 days of the season. Apparently the plants had already produced a near normal pod load before the drought became severe. Excessive rainfall in the 60-90 day period contributed to a large simulated drainage from the profile in all treatments. Note that in contrast to the corn experiment the irrigation schedule interacted with the rainfall distribution in such a way as to minimize the increased water outflow over the unirrigated treatment. However, the nonirrigated soil profile was left in a more depleted state at the end of the season. The predicted net profile depletion level at the end of the season was lower than the neutron measured level (4.65 cm vs. 5.98 cm). This is not a great difference, but when coupled with differences in the oppos.lle direction for the irrigated treatments (treatment 2, 11.53 cm VB. 7.1 cm; treatment 3, 10.85 cm vs. 7.51 cm) it means that the differem:cs between the nonirrigated and irrigated treatments in simulated ET va lues were not as large as they were in reality. Adjustments on tile hasi::; of the measured profile water contents would be in the direction or Lower ing the simulated ET for the nonirrigated treatment and Increu::; Lng, these values on the irrigated treatments. On the other hand, the dl::;crcpuney could be due to uncertainties in the amounts or irrigation l.npul. All overestimate of the input would result in a calculated water content higher than measured. These and earlier data support the not-so-obvious fact that irrigation water must increase ET over non-irrigated ET in order to avoid an increase in profile outflow equal to the amount of irrigation water applied. In a humid climate with sandy soils it will not be possible to manage the irrigation of crops for an economical production without some additional contribution to the water drained below the root zone. The estimated ET values obtained in the corn and peanut experiments in 1978 appear to be reasonable (Table 8). Daily water depletion rates (Table 9)-were calculated for pcrJods of varying length from measured water input and neutron measured chunges in soil profile water content. These depletion rates, when compared with the appropriate expected ET rates of Table 1, show the periods or potential drought stress (depletion < expected ET) and of potent lal drainage outflow (depletion> expected ET). For an irrigatIon treutment, a depletion rate less than expected ET would indicate that irrigation quantities were not adequate to prevent plant water stress. Treatment 2 in the 48-57 day period is an example. In reality errors in measurement of the soil water content or of the amount and uniformity of water input could result in inaccurate water depletion estimates.

PAGE 32

-27Roal density distributions were measured in treatments 1 and 2 on ] 1, 1978 (Fig. 12). Sampling was not extensive enough to IIWUl.iure any treatment differences, but the results are typical of those found by l<.obert!:;on et a!. (1979, 1980). In this experiment, roots \vere obHerved at 225 cm depth. F. Soybeans, 1978 A newly installed irrigation field plot area with solid set impact sprinklers was used to study the response of soybeans to irrigation scheduling strategies. 1. Methods The experimental area was located on Arredondo fine sand and consisted of twenty-four 13.7 m x 13.7 m water management plots arranged as U randomized block in four replications. Impact sprinklers on the corners of each plot delivered water at the rate of 2.54 cm/hr in a quarter circle pattern with a radius of 14.3 m. Five of the six-treatments avallable were chosen for this study. Water management treatments (main) were: (1) no irrigation, (2) irrigation, light rate and frequent, (:3) irrigation, medium rate and infrequent, (4) irrigation, light rate and InfrL'.quent, and (5) irrigation, mixed light and medium rates. Lrrigation scheuuling for treatments 2 and 3 was based on a tensiometer reading of approximately 150 cm at a depth of 15 cm. Treatment 4 was irrigated at a suction of 300 cm at a depth of 30 cm. The light rate of irrigation in treatment 5 was scheduled in the same way as for treatment 2, but when the suction at 30 cm reached 300 cm a medium rate of irrigation was scheduled. Soybean response was measured in terms of yield of mature beans. Subplot treatments were cultivars of soybeans -Bragg and Cobb. Planting was on June 15, 1978 in 76 cm rows with a population of about 258,000 plants/ha. Field dry beans were harvested November 7, 1978. 2. Results and Discussion Rainfall distribut.ion can be seen in Fig. 11. No rainfall occurred between Oct. 1 (day 144, peanuts) and the end of the soybean season (November 28). Irrigation distribution for the various irrigation treatments are given in Table 10. Soybean yields increased linearly with the amount of water applied (Figure 13). Apparently quantity was more important for this season than the strategies of application. Yields are given in Table 11 along with simulated water balance results. Simulated ET was nearly the same for all irrigated treatments indicating that the model was not sensitive enough to detect the expected real ET decrease for treatment 4 where there was a measured decrease in

PAGE 33

-28ROOT LENGTH DENSITY (CM ROOT /CM) SOIL) 0 1.0 2.0 0 1.0 2.0 0 0 15 15 )0 )0 45 45 -:a 60 60 U -::r: 75 75 8 P-! 0 90 90 H !-oj 0 TREATMENT 2 TREATMENT 1 en 105( IRRIGA TED) 105 (NON-IRRIGATED) 120 120 1)5 1)5 150 150 Figure 12. Peanut root density distribution with depth in irrigated and non-irrigated treatments, Gainesville, 1978. ,1>

PAGE 34

-29Table 10. Irrigation schedule, soybeans, Gainesville, 1978. -_ ..... .. __ ._. __ .. _----Irrigation Amount on Treatment NumberJI Date 2 3 4 5 em ---.IUllt' ]0 1. 27 1. 27 1.27 1.27 Aug. 18 1.27 1.91 1.27 23 1.27 1.91 1.27 25 1.27 1. 27 26 1.27 28 1.07 1.07 1.07 30 1.07 1.47 1.07 1.07 Sept. 1 1.07 1. 70 3 1.07 1.47 1.47 5 1.19 7 1.07 1.52 8 1.47 1.47 10 1.07 1.07 13 0.84 1.27 0.84 1.27 16 0.84 1.14 17 1.07 1.07 19 0.84 20 1.07 1.07 21 1.27 22 0.84 23 1.27 0.84 26 0.69 1.27 27 1.07 0.69 29 0.84 Oct. 4 0.84 0.84 6 0.69 1.07 10 0.84 0.97 1.19 11 1.27 12 1.07 1. 27 21 0.97 0.97 0.97 Total 18.89 22.56 13.21 19.72 !/ There were two irrigations (June 16 and 19) of 0.84 em eaeh on all five treatments, but this was included as rainfall, making it a season total of 55.1 em ..

PAGE 35

.. o ....J -30-2000 1000 SOYBEANS 1978 >y=84x+1045 r2 = .98 4 8 12 16 IRRIGATION, em 20 24 Figure 13. Soybean yields as influenced by irrigation amounts, Gainesville, 1978.

PAGE 36

-31-Table 11. Effect of water management on yield of soybeans and simulated water balance, Gainesville, 1978. 1/ Tr ea tment-ET Non-irrigated (1) 35.42 Irrigated (2) 46.76 Irrigated (3) 47.01 Irrigated (4) 46.86 Irrigated (5) 47.01 !/See for details. 2/ -Rainfall, 55.1 cm. Profile 2/ water 3/ Irrig.-depletion-Drainage -----cm-----18.89 22.56 13.21 19.72 11.24 4.10 2.39 9.88 3.15 30.91 31.32 33.03 31.32 30.94 J/ Net seasonal loss from the soil profile. Yield kg/ha 976 c 2690 a 2832 a 2311 b 2668 a yield. The discrepancy could result from anyone or a combination of tIl(' (ollow1.ng l.nput parameter deficiencies: an underestimate of ET, an e of field capacity. and an overestimate of the time needed '"or wuter redislrlbution in the soil profile. In contrast to other water balance data presented thus far, these data predict very little influence of irrigation on drainage loss of water. Nearly all of the irrigation input was used for ET and storage in the soil profile. This is the desired obj eetive of the irrigation management in humid regions. However, in this case all of the irrigation was applied during a prolonged drought so that a few small rainfalls did not cause an overfilling of the soil reservoir. The relatively large drainage outflow for all treatments was due to high rainfall during the period June 12 through August 11. The neutron-measured soil water profile status, the approximate field capacity profile, and the soybean root density distribution are given in Fig. 14. On October 27 the non-irrigated plot was essentially at the permanent wilting pOint to a depth of about 135 em. The irrigation regime for treatment 2 (light, frequent) allowed the lower part of the profile to become water-depleted as planned. Root densities were a little larger for soybeans than for peanuts (Fig. 12). The decrease in root density in the 30-60 em soil depth zone was evident in both the soybean and peanut data and was attributed to a compact soil zone created by tillage.

PAGE 37

E: u I-a.. w 0 CONTENT, e ROOTS, crryem3 0 01 Q05 15 3.0 45 6.0 0 \ 0 30 '0 I 6 60 I SOYBEANS 1978 0 I 0 90 I 10 10/27 120 \ 150 l. -...sLIGHT IRRIG 180 l 10;5 Figure 14. Soil water content and soybean root density distribution with depth, Gainesville, 1978. I w N I

PAGE 38

-33-G. Corn, 1979 The purpose of this study was to determine the effect of drought on grain yield of corn for different stages of plant growth and for two Jrrlgatlon scheduling practices. 1. Methods The field experiment was established on Lake fine sand with a solid Ht'l, lmpact sprinkler irrigation system delivering 2.54 cm of water per hour. The plots were 13.7 x 13.7 m in size and arranged in a randomized block statistical design in four replications. Water management treat ments were: (1) no irrigation, (2) irrigation, light rate and frequent, 0) lrrlgation, medium rate and infrequent, (4) irrigation, same as treatment number 2 except beginning at tassel, and (5) irrigation, same as tre.atment number 2 except no irrigation during tasseling and si1king. Irrigation was scheduled when readings from tensiometers at 15 cm depths exceeded 150-250 cm of water suction. Funk G-4S07 corn hybrid seed were planted on March 13, 1979 in 90 em rows at a population of 71,000 plants/ ha. 2. Results and Discussion Rainfall distrJbution and irrigation schedule data are shown in F 19. II.> and Table 12. Yields and simulated water balance data are given In 'l'ubh' 1:3. Due to previous experiments in the field plot area, yields wen' hi ghly variable. .I rrigated treatment differences were not signifi Nevertheless, the omission of irrigation during tasseling and (treatment 5) resulted in a yield not significantly larger than lhe non-irrigated treatment. T'he water balance data shows the effect of poor rainfall distribu lloLl In Tl!lation to the water retaining properties of the soil. Had, the 11 em outflow from the non-irrigated treatment been available for ET, the expected yield would have been near the maximum obtained in the experiment. A summary of results from the four Gainesville corn experiments is presented in Fig. 16 and 17. In 1977, the very low yield for nonirrigated corn did not fall in the region of linear response and was omitted. The reasonably good regression between yield and irrigation (Fig. 16) indicates a remarkable similarity of drought factors for the three years, although a much better linear fit of the data would be obtained if the 1977 data were excluded. On the other hand, the response may be curvilinear as found by (Skogerboe, 1979). It is apparent that total water input was more important to grain yield than scheduling strategy during these years. The regression coefficient of 143 kg/ha-cm (5.79 bu/acre-in) is a 3-year average measure of irrigation water-use efficiency.

PAGE 39

8.20 EO u I I I I z .. o I I I I I I I I I I I I (92 I I I I I I 0:: 0::. CORN 1979 4 E u .. -12 -1 L1 z
PAGE 40

.. ... -35Lhat coefficilmts would be higher if data for each year were taken 'l'lw regress1211 of gra:l.n yIeld on simulated ET(Fig. 17) yields a 1:1.1 Ir,btly higher R -value and a larger regression coefficient(661 kg/ hacm). Again, the data suggest that seasonal ET was more important than .irrIgation scheduling strategy. The regression coefficient is much larger than the 147 kg/ha-cm value found in Colorado (Skogerboe, 1979). However, calcu1ate.d values from corn grain data of Hillel and Guron (191'3) in Israel were 540, 440, and 450, respectively for 1968, '69, and '70. Our ET simulations may be forcing the range of values within unrealistically small bounds, especially for the higher levels of irri gatIon where the model allocated a considerable amount of the water i.nput to drainage outflow. Table 12. Irrigation schedule, corn, Gainesville, 1979., -._------------------------------------------------------------------, 1/ Irrigation amount on treatment numbersDale 2 3 4 5 ----cm -Mureh 22 0.84 0.84 0.84 0.84 AprU 20 0.84 1.65 1.06 May 7 1.07 1.65 1.07 17 1.27 1. 91 1.27 20 1.91 2.54 1.91 23 1. 91 1.91 26 2.67 28 1.91 1.91 30 2.54 June 6 1.91 2.54 1.91 1.91 10 1.91 1.91 --13 2.54 18 1.91 2.54 1.91 Total 15.48 21.42 13.65 4.88 l/Treatment number 1 also was irrigated with 0.84 cm on 3/22; rainfall, 42.47 cm.

PAGE 41

-3613. Effect of water management on yield of corn and balance, Gainesville, 1979. Treatment'll ET Irrig}.:l NJn-irrigated (1) 37.81 Irrigated (2) 43.80 Irrigated (3) 43.80 Irrigated (4) 42.68 Irrigated (5) 39.60 lSee text for further details. 2Rainfa1l, 42.47 em. 0.84 15.84 21.42 13.64 4.87 3 Net water loss from the soil profile. Profile water 3/ deE1etionDrainage em -5.58 11.08 2.54 16.72 2.64 22.60 3.09 16.52 5.34 13.07 simulated water Grain Xie1d kg/ha 1434 b 5120 a 5187 a 4248 a 2690 ab --.-.. -...

PAGE 42

-37-9 GAINESVILLE 8 b. a:s ..c ev' "7 b.b. 0') .:s:. "-,.., )C, 0 6 "? "" '" "" .. 0 5 .....J W ->-z 4
PAGE 43

ctS .s:::. ........ C) ,., 0 0 ..J W >z a:: (!) z a:: 0 0 -38-9 GAINESVILLE 8 1977 A 1978 x 1979 7 6 5 x 4 3 2 00 30 35 40 45 50 55 EVAPOTRANSPIRATION (ET), em Figure 17. Yield of corn grain as affected by estimated evapotranspiration, Gainesville, 1977, 1978, and 1979.

PAGE 44

-39CHAPTER III. FIELD EXPERUIENTS, LIVE OAK Three water management field experiments were conduc'ted on corn c1ur.lng .1.977, 1978, and 1979. The experimental site of three hectares o[ predominantly Lakeland fine sand, a member of the thermic, eoall'd [am lly of Typic Quartzlpsamments. A subsurface asphalt layer (:1011 1II0Ii.:tllre harr.ler) was prevlous]y (1967 and 1970) i.nstal1ed i.n one or two-tn'atmont maIn plots (0.37 ha) in four rep1ieations (Saxena, et it'! 191"3). The harr j (,T to water and nutrient flow waR placed in a l'ollLlllllom: t;tri.p (0.3 cm thick) at a depth of approximately 65 cm with uvpr.l.app.lllg passes of a special sweep plow. However, the process did not n'Hult 1n a complete seal at the lines of overlap; consequently, l)[ a water table on the barrier occurred only for very short t1.nll's and only durtng large and intense rainfalls. Nevertheless', a uniqudy modified soil profile was created--one in which water was retahwd in larger amounts and for longer time periods in soil located the barrier than for the case of naturally rapid drainage in the untreated soil. With a life expectancy of more than 25 years, the barrier system has the potential for increasing the proportion of rainfall utllized by plants. Moreover, the field facility provides an II1luHlIa1 opportunity to study selected strategies for irrigation manage ment and their influence upon crop yield and water and nutrient balance III Hllndy soils of humid regions. FOlll" watec m:magement subplots (24 x 24 m) were maintained in the Hamt' l OVt'r l he three seasons. One of the subplo ts was non )I'rlgatl'd a III I the OLl1l')" t.hree were irrigated overhead from impact Hpr mounted 011 portable aluminum pipe. WaLl'r balanct' simulations, were performed similarly to that described lor" tilt' en Incsv Ille Experiments (Chapter II). Also, the basic irrigation was to irrigate frequently at light rates (small ljllilnllL:lcs of water per event) leaving part of the water-depleted soil pr.o [lle unf illed and available to store rainfall. A. Corn Experiment, 1977 The purpose of this experiment was to determine interrelationships between moisture barrier, irrigation strategy, plant population, nitrogen fertilizer level, and grain yield for two corn hybrids. A second purpose was to measure downward flux of water below the root zone under different water management treatments.

PAGE 45

-401. Methods The four water management subplots described above were split into eighteen sub-subplots to which factorial treatments (two corn hybrids, three plant populations, and three nitrogen rates) were assigned randomly. Water management treatments were: (1) no irrigation, (2) frequent irrigation with light rate (small quantities per event) when the soil water suction at 15 cm depth exceeded 120 cm of water, (3) same as 2 except infrequently with medium rate, and (4) same as 2 except irrigation was applied when soil water suction exceeded 600 cm. Water manage ment treatments on the asphalt barr.ier plots were designated by adding the letter "A" to the above numbers. Corn hybrids DeKalbXL-80 and Funk G-4507, were planted March 22, 1977 and harvested July 25. Nitrogen fertilizer was applied only as a sidedressing, 30% at 16 days after planting 70% at 45 days. Nitrogen levels were 134, 202, 269 kg/ha. Plant populations were 39, 59, and 79 thousand plants/ha in rows 7() em apart. Irrigation was by overhead impact sprlnklcnl at 1.9 cm jwr hour. SoLl water data used in water balance simuJ at lons g lVl\lt .I.n tile Appendix Table. An extensive network or mercury manometer tcns.i was installed in only sub-subplots pla1,lted with Funk (:-4S07 COTn at. t.h(' highest plant population and nitrogen levels of all water subplots of two replications. Tensiometers were placed in depth inter vals varying from 5 to 30 cm with maximum depths of 45 and 300 Clll, respectively for the barrier and non-barrier treatments. Readings were taken once-daily on Monday through Friday of each week. Evaluation of these data will be reported elsewhere. However, calculations of down ward flux of water at 240 cm depths were obtained by multiplying gradients of hydraulic head from tensiometer readings with hydraulic conductivities obtained in an earlier study (Parra, 1971). 2. Results and Discussion Rainfall and distribution and irrigation scheduling data are shown in Figs. 18 and 19. Drought conditions were serious 1.n most of May and for shorter periods in April, June, and July. These water input bUlions are reflected in the 60i1 water status as measured by lellfJiollloluru (Figs. 20-24). Tenslometcr data were shown as hydraull.c }wads vur Lum. depths and in two to four-day intervals. Thus, onc (:<111 obta1.n a quul I.Lu tive view of both water content and vertical water flux direclions. Moreover, during dry periods the hydraulic head read.l.ugs revealed Lhe presence or absence of water absorption by roots. At a given depth, water content decreases as the hydraulic head becomes more negative. Water flows vertically in the direction or

PAGE 46

5 E u 4 j 3 z 4:2 0::: 1 1977 LIVE OAK I I II I o MAR APR MAY 11.30 II 11-.. .II II ,I I I I I JUN JUL AUG SEP OCT Figure 18. Rainfall distribution, Live Oak, 1977. I ./:'f-' I

PAGE 47

2 2 2A 2 E 0 u 3 2 -3A <.9 0:: -42-1977 LIVE OAK I II 0:: 4 2 4A 2 o MAR APR MAY JUN JUL Figure 19. Irrigation distribution for six irrigiltioll treatments on corn, Live Oak, 1977. Tr0ill ments with the letter "A" included a subsurface asphalt barrier treatment.

PAGE 48

APR MAY JUN JUL 0 1977 LIVE OAK DEPTH ----15cm TRT 1 r 45cm ... -120cm --0-150cm .. o w I u -1 ::J 0::: o -200 ......... .. ... \ \ I 'V >-300 I 1\/ \ 6 I I V 1 I I I I II "-0'&.-0.'0"'-""'j'\ U' .... I -4-00 -500 I 9 I I I '" I I Figure 20. Hydraulic head distribution with depth and time under non-irrigated corn, Live Oak, 1977. Soil surface was the datum, I w I

PAGE 49

APR MAY JUN JUL E -100 u o w I u -.J :J -200 1977 LIVE OAK TRT 2 0:: -"0"1 o ,. p'n I >_ 300 rJ C\-o-_<\ /' 1 I u \ -400 \ \ \ b. -500' ". -Q..,o.,,Q. DEPTH --+15cm --0-45cm ..... 120cm 150cm f...... ,,,, ,ri .... I ... .... .0'0. "Q. '0--{l. --,crD' Figure 21. Hydraulic head distribution with depth and time under irrigated corn, treatment 2. Live Oak, 1977. Soil surface was the datum. I I

PAGE 50

0 E -100 u w I -200 APR MAY 1977 LIVE OAK TRT 3 u -...J ::::> 0::: O I o-<}o-
PAGE 51

, o w I U -.J :J 0:: APR 0 1977 MAY JUN LIVE OAK TRT 4 p-P "O'()---o--d JUL DEPTH --15cm --045cm -_e_-120cm --0--150cm ,. I ... I 'tI............ .... P-I \ I -oJ..O-o-Oo-rl o 300 I \ ,,}, II IT \ I II' -400 -500 I \11 I I Figure 23. Hydraulic head distribution with depth and time under irrigated corn, treatment 4. Live Oak, 1977. Soil surface was the datum. I .j::-(J'\ I

PAGE 52

.. APR MAY JUN JUL 1977 LIVE OAK BARRIER TRT 45cm Depth E u-100 -o w ./IRRIGATED (2A) NON/ I -200 IRRIGATED (1A) u ---' ...J 0::: o I. I I -400 Figure 24. Hydraulic head distribution with time at a depth of 45 cm under non-irrigated and irrigated corn grown on soil modified by a subsurface asphalt barrier, Live Oak, 1977. I ....., I

PAGE 53

-48(more negative) hydraulic head. For example in treatment I, non-irrigated (Fig. 20) water flow in April was downward between all depths bhown.. In May, as the soil became dry, net water movement became upward, first in the 15 to 45 cm zone and next in the 45 to 120 cm zone. Wide temporal swings in hydraulic heads at the two upper tensiometers resulted from alternate periods of droughts and rainfall. However, in the 120 and 150 cm zone net water flow was downward throughout the season. At the 120 cm soil depth soil water content decreased with not much evidence of water withdrawal by roots until late season. Likely, plants stunted in "above-ground" growth by the May droughts were also limited in root growth. Irrigation (Figs. 21, 22, and 23) attenuated temporal variability in hydraulic head readings at 15 and 45 cm. Water flow throughout the season was vertically downward at depths below 45 cm in treatments 2 and 3. In contrast, the less frequently irrigated treatment 4 exhibited three periods of upward movement in the 45 to 120 cm zone. Root activity as indicated by tensiometer readings was not easily discernable in the 120 to 150 cm data of the irrigated treatments because the. soil prof:l.les were maintained in a moist state. The influence of the asphalt barrier on hydraulic head at 4S em depth in non-irrigated and irrigated treatments i1; shown In 24. Irrigation essentially eliminated drying of the ut 45 cm sJuee the otrategy was to restore to field capacity (Appendl.x Table) only the top 30 cm of the soil prof lle. Comparing tlw early and late hydrau]:l.e head values in Fig. 24 with those at the same depth 1n F.lgH. 20 to 2.1 reveals the higher soil water retention in soil above barrier. Tn addition, the fluctuation of hydraulic head was less over the barrier. Simulated (model) water balance data and corn yields are given in Table 14. Yields will be discussed later. In the non-irrigated treatments all water and a considerable soil profile depletion was used for evapotranspiration. On the other hand, irrigation caused a substantial deep seepage or drainage loss in some cases. Assuming that the simulated ET was correct, it is evident that several rainfall events occurred when the 'soil profile was not sufficiently depleted to retaln the amount which fell. The more water-conserving treatments 4 and 4A reduced drainage loss but also resulted in reduction in yield. One of the more striking results in Table 14 is the. effect of the increased soil wnil'r retention by the moisture barrier in producing larger yJolds wjl:lt lesH irrigation. The results are unique in demonstrating Jmproved proc.1uctivity of droughty sandy soils from an alteration of
PAGE 54

-49Tub.! t' VI. Effect of water management on yield of corn and simulated water balance, Live Oak, 1977. Non-in igated (1) Irr igll-ted (2) lrr.igatcd (3) Irrigated (4) Non-irr igated (IA) Irrl gated (2A) Irr j gated (3A) Ir riga ted (4A) Us. '1./ f J] '27 r:r.: Hu. nl a.. ..u em. ET 33.58 42.18 42.85 40.62 34.89 42.55 43.06 41.40 Irrig 28.61 32.65 14.09 21.98 23.43 12.39 Profile water 3/ depletion-ern Non-barrier 6.58 1.24 1. 25 3.35 Barrier 7.80 1.87 1.87 3.13 'J/ Net seasonal water loss from soil profile. Drainage 0.54 15.22 18.60 4.37 0.46 8.85 9.79 1.67 Grain Yield kg/ha 3015 7741 8081 6217 3367 9363 9266 8349 Table 15. Estimates of seasonal drainage in relation to water managementtreatments, Live Oak, 1977. Tr ea tmen t!:./ Non-irrigated (1) Irrigated (2) Irrigated (3) Irrigated (4) Fallow l/Durlng 121 days. 2/ See text for details. }./ From Table 14. Seasonal deep seepage at 240 ern Tensiometer em 0.54 15.22 18.60 4.37 ReE' em 0.23 5.27 1.38 1.00 I ReE' HI ern 2.47 0.73 7.93 2.77 12.33

PAGE 55

-50always win the water balance game without losing crop yield. TIle implications for agricultural production, economics, resource-use effi ciency, and pollution control have not been fully appreciated or explored. In an attempt to obtain a measure of the real water balance (nonbarrier treatments) we calculated seasonal downward water flux at 240 em depths from hydraulic head readings at 210 and 240 cm. The results along with simulated values from Table 14 are given in Table 15. Obv:l. ously, there is a need for more replication, and one might question whether the frequency of readings were sufficient for such calculations. Other problems include spatial variability in soil and in irr:l.gation water application. Treatment differences were not clearly delineated from the tensiometer data, but the range of values is less than. that simulated by the model and possibly indicates that the model has underestimated ET and overestimated water outflow by deep seepage. Yield data (Tables 16 and 17) were organized to show the following significant two-way interactions: barrier x population, irrigatJ.on x population, variety x population, variety x fertilizer, and irr1.gat:lon x fertilizer. Tukey's honestly significant difference (THSD) values were calculated as an aid in making comparisons among the various average yields. In Table 16, THSD values for comparison of barrier, water management, and variety treatment means at the same populat1.on were 94 'J. 1289, and 294, respect1.vely. And THSD values [or comparlson of populaLlon means at the same barrier, water management, and va'riety "'l'rl' 353, 499, and 353, respectively. In 17 'l11SD valucH ror. complldmUl of water management and variety treatment means at the sallie f ertJ.1 J level were 1289, and 294, respectively. ,And THSD values for comparJ.hOoll '. of fertilizer means at the same water management and var lety trl.!atments were 499 and 353, respectively. The results of these comparisons ",Hl be stated in general terms only. The barrier improved yields at all populations, and yield increases with population were greater on the barrier treatmeent. All irrigation treatment yields were higher than the non-irrigated treatment at each population. In addition, at the highest population, treatment 2 and J yields were signficantly higher than treatment 4. Considering poptllatJon effects at the same irrigation treatment, there was no response [or treatment 1. Response to population in treatment 4 leveled off at plants/ha. In contrast, for irrigation treatments 2 and J, there was an increase in yield with each increase in plant population. Funk. G-4507 corn hybrid yields increased with increasing population while DeKalb XL-SO leveled off at the intermediate population. TIlere was no difference in corn hybrids at the smallest populat1.on level, but highest yields were obtained from Funk C-4507 at the intermed 1.<1 le ilnd Jdgll P 1 illl t populations.

PAGE 56

-51Table 16. Corn grain yields as affected by plant popu1ation, moisture barrier, water management, and corn hybrid, Live Oak, 1977. ..... __ ... -.--.--Grain yield 39,OOo?--/ 59,000 792000 ---.. kg/ha kg/ha kg/ha Non-harder 5593 6559 6640 Burrier 6554 7928 8278 Non-irrigated (1) 3105 3371 3098 Irrigated (2) 7313 8918 9426 Irrigated (3) 7340 8981 9702 Irrigated (4) 6535 7705 7609 DeKa1b XL-80 6182 7062 7102 Funk G-4507 5964 7424 7816 If' ------.. ... 4. text for details. Tnh I (\ ] J Corn grain yIelds as affected by nitrogen fertilizer level, wntl'r management and corn hybrid, Live Oak, 1977. -.-.. -. _. ---_.-_._----------------------------Treatment!/ Non-irrigated (1) Irrigated (2) Irrigated (3) Irrigated (4) DcKa1b XL-80 Funk G-4S07 !'ee text for details. N, kg/ha. 1342:./ kg/ha 3322 8285 8240 7166 6447 7060 Grain yield 202 kg/ha 3186 8723 8701 7246 6916 7013 269 kg/ha 3066 8648 9081 7437 6984 7132

PAGE 57

-52Corn response to irrigation interacted with fertilizer level. Crain yiC;;!lds increased with increasing fertilizer levC;;!l only for irrigation treatment 3 and only at the highest nitrogen level. For each level of nitrogen applied, all irrigation treatments gave higher corn yields than non-irrigation treatments. However, at the intermediate and high nitrogen levels, yields were lower in irrigation treatment 4 than in the other irrigation treatments. In relation to variety and fertilizer level, Funk G-4507 corn was higher yielding than DeKalb XL-80 only at the low nitrogen level and did not respond to nitrogen levels. Crain yield of DeKalb XL-80 increased with increasing nitrogen application only up to the intermediate level. B. Corn Experiment, 1978 The overall objective was the same as for the 1977 experiments except that corn hybrid was not included as a variable. 1. Methods The four water management subplots del:icribed carllE!r were spllt .I nto six 1mb-subplots (3 x 12 m) to accomodate factor la1 lrmltlllcntH of thrvl! nitrogen fertilizer levels (134, 268, and 336 kg or N/lli:1) and two 1'11\1\1. populations (59 and 90 thousand plants/ha). Corn hybrid DcKalh XL-SO wny planted in 76 cm rows on March 16, 1978 and harvestcd on July 18. Wuler management treatments and irrigation methods were the same as 1n the 1.977 experiment. Soil water was monitored in the highest nitrogen plots 1n replicates 1 and 3 using tensiometric and neutron techniques. Input data used in the water balance simulations are given in the Appendix Table. 2. Results and Disc4ssion Rainfall and irrigation schedule data are shown in Figs. 25 and 26. In Fig. 26, the late July through October data are from a soybean experiment not reported here. Note that the drought period for the corn experi ment extended from early May through June. Simulated water balance data and corn yields are given in Table 18. Grain yields were less than in 1977. Yields for treatments 2A and 3A (barrier) were more than 1.3 times those for treatments 2 and 3. Allor the irr.igation treatments resulted in a maximum ET of aroUlld 42 em wIdJ (\ the ET of both non-irrigated treatments were nearly snn,,:! .:IL 36 em. These values are very close to those obtained in 1977 ('rable 14). OIW third or more of the irrigation water applied was allocated to draln:lgl" However, as in 1977, irrigation water was used more efficiently with the barrier treatment. These results challenge us to develop water manop-(' ment systems which will equal the effectiveness of the bnrr.ier-Jrrlgntl.on

PAGE 58

6 5 E 4 u j3 z a:: 1 o 1978 LIVE OAK I. I I I .. ,t II L I I .1 1.1 MAR APR MAY JUN JUL AUG SEP OCT NOV Figure 25. Rainfall distribution, Live Oak, 1978. >''i I U1 W I

PAGE 59

-54-1978 LIVE OAK 2 2 E u 3 2 z o 2
PAGE 60

.. -55was more drainage loss from the non-irrigated treatment 1.n PJ78 than in 1977, although in both years the amount represented a 911U111 percentage of the rainfall. In contrast to several two-way interactions obtained in 1977, only one was found in 1978--irrigation by population (Table 19). On the non l.t:rlgated treatment, the corn yield for the high population treatment Willi I.el:lf:l than that for the lower population. Population had n.) effect w:Lthl.11 the :i.rrigated treatments. This result is opposite to the 1977 where marked response to population increase was obtained for lrr.Lgat"1on treatment 2 and 3. However, the maximum number of p1ants/ha 79,000 in 1977 versus 90,000 in 1978. Considering the data for both years, it is evident that for DeKalb XL-80, 90,000 p1ants/ha was larger than required for optimum grain yield. At the same treatment 1 means were smaller than all the irrigated treatment means (THSD = L295). Yields were affected by both barrier and nitrogen fertilizer level (no interaction) as shown in Table 20. Table 18. Effect of water management on yield of corn and simulated water balance, Live Oak, 1978. ------.-.-. ----------------,---------------NOll-Jrr (1) Ixr 19atl!U (2) lrr lr,l.ltcd (3) Irrigated (4) Non-irrigated (lA) Irrigated (2A) Irrigated (3A) Irrigated (4A) ET 35.77 41.81 41.81 41.81 36.53 42.10 42.10 -42.10 Profile water 3/ depletion--em -Non-barrier 16.20 19.02 23.32 19.06 19.02 18.56 Barrier 8.05 1.06 1.04 1.14 8.06 0.88 0.91 0.91 !/See text for details. 2/Rainfall, 31.15 cm. 3/ Net seasonal water loss from soil profile Drainage Grain yield kg/ha 3.42 2475 6.59 6290 9.39 6353 13.79 6821 2.68 9.00 8.98 8.52 3272 8509 8624 7942

PAGE 61

-56-Table 19. Corn grain yields as affected by water management and plant population, Live Oak, 1978. Grain TreatmenJ:/ 59,00Of:-/ 90,000 Average -kg/ha Non-irrigated (1) 3324 2423 2874 Irrigated (2) 7244 7555 7400 Irrigated (3) 7578 7398 7488 Irrigated (4) 7535 7228 7382 Average 6420 6151 -1/ -See text ------.-.-. for details. 2/ Number of p1ants/ha. Table 20. Corn grain yields as affected by !iubsurfacc barr Lt')' and nitrogen fertilizer level, Live Ouk, J978. Nitrogen1 / added -kg/ha 134 268 336 Average l./N NO as NH4 3 No barrier 5114 5819 5521 5485 Barrier AveruJ[ kg/ha -6589 7606 7064 7087 5852 6713 6305 Response to nitrogen was curvilinear. The addit:l.onal 67 kg N/ha over the intermediate level was applied at tasseling and result.ed 1n a reduction in yield in comparison with the yield at the interllled1.ut.e level. We have not found an explanation [or this puzzling result. nlthough there is evidence in Table 17 (1977 data) that response to nitrogen waH near maximum at the 269 kg/ha rate. c. Corn Experiment, 1979 The overall objective was the same as for the 1977 and 1978 experI ments. However, plant population was not included as a variable in 1979.

PAGE 62

-57-1. Methods The four water management treatments described earlier were split into six sub-subplots (3 x 12 m) to accommodate factorial treatments of three ni.trogen fertilizer levels (84, 168, and 252 kg of N/ha) and two corn hybrids (Coker 77 and McCurdy 67-14). Planting was in 76 em rows to give a population of 59,000 plants/ha. Water management treatments were the same as in 1977 and '78, and they were assigned to the same [ll'ld plot: locations. The same overhead. irrigation system was utilized. Soll water conditions were monitored in replicates 1 and 3 with tensio mcLeril llnd
PAGE 63

6 r5 r4 -l -l Lt3 fz <{ Ct 2 1 o .. 1979 LIVE OAK I .. II I LJ LL I I I MAR APR MAY JUN JUL AUG SEP OCT NOV Figure 27. Rainfall distribution, Live Oak, 1979 I II I VI 00 I

PAGE 64

-59-1979 LIVE OAK 2 2 -o L---LLJ.L II ---Ll..Ll.LL1111 L-J.......LJll.----L.'-l-1 2A 2to ----L1L.-J II:...L.-JILI tl.JJ....L...11I ---I-.I....L--,' -..i.---...i E r 3 u 2 I-I 3A
PAGE 65

a ..c --tJ) ........, 0 --1 W >8000 6000 4000 2000 Irrigation Schedules 84 -60168 TOTAL N 3 2 4 1 252 (kg/ha) FIgure 29. Influence of nitrogen rilll'S IIl1d water management treatments Oil ylt'ld of corll grain, Live Oak, 1979.

PAGE 66

.. -61irrigated treatments, and response was limited to the intermediate level of N. Irrigation x barrier interaction is evident in the data of Table 22. The appropriate significant difference values (THSD) for testing two means are: 1727 kg/ha for two irrigation treatments means at the same barrier treatment, and 1964 kg/ha for two barrier means at the same 1 rr( treatment. Barrier treatment means were different only for irrigation treatment 2 where yields were higher with the barrier. Yields on the no-barrier treatment were larger for irrigation treatments 2, 3, and 4 than for 1. On the barrier treatment, yields were higher for .irrigation treatments 2 and 3, and the yield for irrigation treat ment 4 was larger than for non-irrigated treatment 1. Yield and water data for the three corn experiments at Live Oak were Bubj ected to the same regression analysis as the Gainesville corn dutu (see Chapter II, C. Corn, 1979; Figures 16 and 17). The Live Oak results are given in Figs. 30 and 31. Differences in response to the barrier treatments were of sufficient magnitude to suggest separate regresBion lines. For irrigation and ET, the response was steeper with the !:mbsurface barrier. The best-fitting regression equations were obtained with irrigation as a variable rather than ET. Thus, we have further evidence that the simulated ET values may not represent very well the actual ET. Analysis of the tensiometer and neutron data (bpyonJ the scope of this report) should provide information useful in resolving this question. However, in the Gainesville analysis, the variable ET gave a better fit than irrigation amounts. The response per unit of ET was larger at Gainesville and Live Oak than the response per unit of irrigation. And, :Ln u further comparison of Gainesville and Live Oak data, note that the hest rl t was obtained for Live Oak using irrigation amounts and for (:u t.lll'HV 1 Lle using ET amounts. The response per unit of irrigation water wus larger at Live Ouk than Gainesville; while the response per unit ET wus larger at Gainesville. Some of the steep response at Gainesville ean be attributed to the depressed yields in 1979 due to soil fertility .problems associated with an earlier study in the experimental site. It is unlikely that actual ET was reduced in proportion to the yield reduc'tion. Finally, as noted earlier for Gainesville, the climatic factors at Live Oak during the three growing seasons must have been very similar to produce the results obtained. Further evaluation of these results and addition of 1980 data may show that Gainesville and Live Oak data can be combined for a regression relationship which will be useful for the North Florida region

PAGE 67

-62-Table 21. Effect of water management on yield of corn 'and simulated water balance, Live Oak, 1979. Tr ea tmen t1..1 ET Non-irrigated (1) 31.61 Irrigated (2) 41.50 Irrigated (3) 41.50 Irrigated (4) 40.72 Non-irrigated (lA) 30.43 Irrigated (2A) 41.65 Irrigated (3A) 41. 65 Irrigated (4A) 41. 63 Profile / water 3/ Irrig.1 depletion-Drainage --------cm -------24.01 33.45 16.77 17.71 16.75 12.01 Non-barrier 10.11 0.26 -0.17 2.45 Barrier 9.48 2.15 3.11 7.84 3.79 8.05 17.06 4.79 4.34 4.34 4.34 4 .14 Crain yield kg/ha 2955 6779 8021 6814 3'.42 9874 9761 7<)53 \/ for .'.'. --" '2/ -Rainfall 25.29 em. l/Net seasonal water loss from soil profile. Table 22. Effect of water management and a subsurface moisture barrier on yield of corn, Live Oak, 1979. Grain Treatment 11/ 2 3 4 Av eras e ---------kg/ha -----------No barrt"er 2955 6779 8021 6814 97.93 Barrier 3442 9874 9761 7953 123.69 Average 3198 8327 8891 7383 -----_._--1/ Water management treatments; see text for details. 4

PAGE 68

.. -6310 00 LIVE OAK 9 ro .s:::. .......... 0 e ,., 0 7 .. "vjo 0 ":> ..J 6 'Of).. W r:o+ C!J'b cv '1-"" >" -\ z 5 l('b <{ 0;No Barrier a::: ,'0 1977 '" ", x 1978 Z 4 1979 a::: 0 Barrier () o 1977 3 1978 .0 1979 2 00 5 10 15 20 25 30 35 IRRIGA TION, em Figure 30. Influence of irrigation amounts and a subsurface asphalt barrier on yields of corn grain, Live Oak, 1977, 1978, and 1979

PAGE 69

-6410 0 ..t= "-C) ,." 0 .. 0 ..J W r z ":> '0" "" ". (0' 5 / <"" ,,":> ,":r / lJ,+ 4 1><0) ,., 0 3 32 34 36 38 40 42 44 EVAPOTRANSPIRATION CET), em Figure 31. Influence of ;J1lt1 ;1 Hllk:llrlll('(' asphalt barrier on yields of corn gratn, L [VI' Oak, 1977, 1978, and 1979.

PAGE 70

-65CHAPTER IV. SUMMARY DISCUSSION Results presented in Chapters II and III have important implica tions to Florida agriculture for scheduling strategies and agricultural water policy. These findings represent an initial effort to fulfill the need for water management information in planning and implementing the efficient use of all resources. Recognition of this need and the initiation of appropriate research programs is of relatively recent origin in Florida. A. Water-Use Efficiency Water-use efficiency (WUE) by plants may be expressed in several ways depending on the nature of yields and water-use data available. WUE is commonly defined simply as a yield per unit of water used per season in ET, or total depletion (ET + runoff + drainage). Total depletion is readily measurable since it is also the sum of rainfall, irrigation, and change in soil water storage. Actual ET is not easily measured, hence the common use of estimates. Crop yield, Y, may be expressed as total dry matter or some fraction of it, for example, grain. Crop yield increase (Y-Yo, where Yo is yield without irrigation) per unit of irrigation (I) provides another useful measure of water-use efficiency. In this report total grain yield, yield differences due to irrigation, irrigation amounts and estimated ET are used to calculate water use efficiencies based on irrigation [WUE(I) = (Y-Yo)/I] and on ET [WUE(ET) = Y/ET]. Calculated from the corn experiments are given in Table 23 for Gainesville and Table 24 for Live Oak. Although these data are related to the regression analysis in Figs. 16, 17, 30 and 31, they permit a somewhat more detailed evaluation of the individual treatment effects in each experiment. The slope from the regression equations (irrigationbased, Figs. 16 and 30) are comparable to individual treatment WUE(I) values. However, the WUE(ET) values are based on the whole of ET and not on a threshold value as is the case in the regression analysis. Consequently, the slope of the regression equation is larger than the calculated WUE(ET) values. Also, this line of reasoning suggests that WUE(I) should usually be, larger than WUE(ET) for a specific irrigation treatment. Irrigation in excess of ET needs by the crop or a large rainfall soon after irrigation will reduce WUE(I). Large WUE(I) values may be obtained when small amounts of irrigation water are needed only at critical stages in plant growth. For example, WUE(I) values calculated from the corn data of Robertson, et a1. (1973) in experiments on the asphalt barrier plots at Live Oak in 1971 were 0.28 tons of grain per ha/cm of irrigation on the non-barrier treatment and 0.49 tons/ha-cm on the barrier. Corn data of Rhoads and Stanley (1973) gave smaller WUE(I) values, 0.09 and 0.18 for 1970 and 1971. WUE(I) values calculated from corn data of Robertson et a1. (1980)

PAGE 71

-66-ranged from 0.16 to 0.27 tons/ha-em. In Colorado, Skogerboe, et ale (1979), found WUE(ET) values for corn ranging from 0.11 to 0.15 tons/haem. The "best treatments" regression coefficient (Y/ET) was 0.18. Calculated WUE(ET) values from three years of corn grain data of Hillel and Guron (1973) in Israel ranged from 0.03 to 0.22 tons/ha-cm. Regression coefficients from their data were 0.54, 0.44, and 0.45 tons/ha-cm in 1968, '69, and'70, respectively. These coefficients are in the range of our three-year average values given in Figs. 17 and 31. Table 23. Water-use efficiency of corn, Gainesville, 1977-79. 1/ Treatment-Non-irrigated (1) Irrigated (2) Irrigated (3) Irrigated (4) Irrigated (5) Non-irrigated (1) Irrigated (2) Irrigated (3) Irrigated (4) Irrigated (5) Water-use efficiency 1977 1977 Exp. I 0.20 0.19 Exp. II 1978 tons/ha-cm Irrigation Basis 0.16 0.09 ET Basis 0.01 0.17 0.19 0.22 0.21 0.23 0.06 0.17 0.16 0.16 1979 0.24 0.18 0.21 0.26 0.04 0.12 0.12 0.10 0.07 !/See Chapter II for details. 2/ Calculated from yield and water balance data in Tablec, 4, 8, and 13. Water-use efficiency irrigation basis and ET basis abbreviated in thusly: WUE(I) and WUE(ET). In Tables 23 and 24, we find several treatments with WUE(I) values less than WUE(ET). In all cases, these treatments received rather large irrigation applications. For treatments 2 and 3, Gainesville 1977, the irrigation amounts were equal to or greater than ET (Table 4). Thus either ET was or irrigation amounts were greater than needed. On the other hand, irrigation amounts were not measured but calculated for the spinkler system. For treatment 3 with no barrier (Live Oak 1977 and 1979), the WUE(I) values were reasonable in magnitude, but still less than WUE(ET).

PAGE 72

" Table 24. Yield and water-use efficiency of corn, Live Oak, 1977-79. Treatment1 / Non-irrigated (1) Irrigated (2) Irrigated (3) Irrigated (4) Non-irrigated (1) Irrigated (2) Irrigated (3) Irrigated (4) Non-irrigated (1) Irrigated (2) Irrigated (3) Irrigated (4) 21 Grain yie1d-No barrier kg/ha 2756 8487 9552 6315 2603 5802 6617 6711 2899 9045 11683 8764 Barrier kg/ha 2889 10290 11285 9466 3810 8342 8781 8279 4553 13680 12422 9783 wuEl/, irrigation basis WUE3/, ET basis No barrier .. Barrier No barrier Barrier -ton/ha-cm0.20 0.21 0.25 0.20 0.21 0.18 0.26 0.26 0.35 1977 1978 1979 0.34 0.36 0.53 0.24 0.26 0.24 0.52 0.47 0.44 0.08 0.20 0.22 0.16 0.07 0.14 0.16 0.16 0.09 0.22 0.28 0.22 0.08 0.24 0.26 0.23 0.10 0.20 0.21 0.20 0.15 0.33 0.30 0.23 Chapter II for details. See also Tables 14, 18, and 21 for irrigation and ET data. differ slightly from those given earlier since the following treatments only were selected: 1977, population and fertilizer levels of 79,000 p1ants/ha and 269 kg N/ha; 1978, population and fertilizer levels of 59,000 p1ants/ha and 252 kg N/ha. efficiency. Irrigation and ET basis designated as WUE(I) and WUE(ET) in text. I cr....... I

PAGE 73

-68-An interesting comparison of WUE(I) and WUE(ET) values is provided by treatment 5, Gainesville 1979 (Table 23). A small irrigation amount produced a large yield increase, but the WUE(ET) value'was still very low. This results from the nature of the response curve; low yields give low WUE(ET) values and these measures of WUE should be used together with seasonal regression coefficients (as in Figs. 5 and 13) in evaluating irrigation treatment effects. The advantage of the moisture barrier system over the no-barrier system is clearly evident in the differences in water-use efficiency values (Table 24) ,. In addition, the fact that water-use efficiency was substantially increased by irrigation has important implications on resource management. When there is appreciable drought, the decision not to irrigate can be a wasteful use of water as well as of other resources I Peanuts,'and soybeans gave lower water use efficiencies than corn (Table 25) as expected for high protein and oil producing crops. Table 25. Water-use efficiency of peanuts and soybeans, Gainesville. Water-use efficiency (WUE)!/ Treatment1 / Non-irrigated (1) Irrigated (2) Irrigated (3) Irrigated (4) Irrigated (5) Non-irrigated Irrigated (2) Irrigated (3) Irrigated (4) Irrigated (5) Peanuts, 1977 0.12 0.13 0.11 0.06 0.08 0.08 0.07 Chapter II for details. Peanuts, 1978 Soybeans, 1978 -tons/ha-cm -----------Irrigation Basis 0.03 0.05 ET Basis 0.09 0.08 0.09 0.09 0.08 0.10 0.09 0.03 0.06 0.06 0.05 0.06 !6a1culated from yields and water balance data in Tables 7, 8, and 11. Water-use efficiency irrigation basis and ET basis abbreviated in text thusly: WUE(I) and WUE(ET).

PAGE 74

-69-1.1\ mOHt eal:les, irrigation increased water use efficiency, and treatments 'j and '\ were the best.Tbe 1978 peanut data show very low WE(r) values, and LileY are less than WUE(ET) values. This probably indicates that too lIlut:h Lrrlgution was used in that year on a deep-rooted crop. Note that WUE(J':T) values were about the same in both years, except for the non LrrigaLeJ treatments. The 1977 peanut data in Table 25 and in Figure 5 provide c,n interl!HL i.ng contrast in datu presentation and evaluation. Calculated WUE(ET) e(jllutlollS because there Is a threshold ET value, 23 cm. This is not the case for the irrigation-based regression; the coefficient for soybeans, Figure 13, is about the same as WUE(r) values in Table 25. The excellent fit of the peanut and soybean data indicates that no significant betweentreatment variations in other management factors were present. A regression coefficient of 0.138 tons/ha-cm was calculated from peanut (Florunner) data of Pallas et al. (1979) in Georgia. F.igure 5 provides another comparison in data representation -one year Versus three years of data as in Figs. 17 and 31. As expected, thl!rl! was a considerable year to year variation in the simulated ET for lIon-irrigat.ed treatments. Si2ce the yields were nearly the same this contributed to a much lower R value than would have been obtained with U1uluul regre8s1.on eq"uatlol1s. COlllJlurl80n o[ ir.rigation and ET annual regression equations con l1-asts w.Lth comparisons of WUE(ET) and WE(r) values. Theoretically, UK' two regression coefficients could be nearly equal. The required condHiol1s are: (1) ET is actual seasonal ET, and (2) accurately measured seasonal irrigation inputs contri,bute to ET in the same way for each treatment, Le. the quantity of irrigation not used to increase ET (runoff and drainage losses and increase in soil storage) is the same. Even if actual ET data were available, it is not likely that the second requirement would be fulfilled in most irrigation experiments. Thus,in view of the nature of the crop-response, water-use relationship discussed earlier, it is reasonable to expect that the irrigation-based 'regression coefficient will be less than the ET based coefficient. This was the case for all experiments in the current study. Regression of peanut yield on ET (Fig. 5) gives a regression coefficient of 162 kg/ha-cm, a value larger than the regression coefficient for yield on irrigation: Y = 124X + 2241 (R2 =O.99) whe:r;e Y is pod yield, kg/ha, and X is seasonal ir.rigation amount in cm.

PAGE 75

-70Comparison of these two regression coefficients provides a very revealing test of the success attained in irrigation management. The ideal ratio of br/bET (where b is the regression coefficient) would be one. In this case we have 124/162 = 0.765, the highest ratio obtained in the studies reported here (Table 26). These surprising results suggest further evaluation in terms of economics of irrigation as well as of the variable irrigation effectiveness achieved. Table 26. Water-use efficiencies from the water management experiments of this report in terms of yield versus water-use regression coefficients and their ratios. Location of Total Regression coefficient -----+1 Ratlo .. ------Crop eXEeriment years Y vs. I Y VS. E1' b}/hET kg/ha-cm ki!l;;;'=-cm Peanuts Gainesville 1 124 162 0.7 be) Soybeans Gainesville 1 84 144 O.SS] Corn Gainesville 3 143 661 0.2.16 Corn21 Live Oak 3 160 471 o. ]l,O CorrrLive Oak 3 276 591 0.467 11 b I and bE'T' are regression coefficients of yield on irrigation and evapotranspiration, respectively. llSoil profile modified with a subsurface asphalt barrier. B. Irrigation Scheduling Results from the present study show that an irrigation managemenl strategy of frequent application at rates which only partially f i.ll tilt! ET and drainage-depleted soil profile will conserve water while lllt'eLillg the water needs of crops. An irrigation scheduling plan is presented here whereby this concept can be put into practice by farmers and other growers in Florida. Perhaps more importantly, the basic concepts of the plan can be used by industry in developing new irrigation systems and in designing new and replacement irrigation installations. Moreover, there is a current rapid development of commercial irrigation scheduling servi.ccs hased on a gr-owl.ng sopiti8Lic..:uLion In l:ommunicatioll, meU8urelllcnt or the pllyt-.;(clll system, computer simulatIoJl, and crop growth mod,'1 dc!vc.loPlllt'llt.. 'J'noln needed by the farmer to utilize this method of schcJullng Irrlp,;H lOll lor crop production are a low-cost rain gauge, a radio [or weather rOrt'Clllll:lI, a chart of daily potential ET estimates (Table 1), and a shovel to aid in making periodic observations of the top 30 to 45 cm of the so11 profile. ..

PAGE 76

.. -71Tile.' basic plan is to irrigate the top 30 cm of the soil profile ()j l ell (.'JlolJgh to prevent more than short-term ( a few hours) wilting of plaJlLH. The amount of water needed can be estimated by examining the Boll prorlle for depth of water percolation 12 to 24 hours after a application. A starting test irrigation amount per irrigation event for sandy soils should be in the range of 1. 5 to 3.0 cm (approximately ]/2 to 1 inch). Once irrigation has been initiated, subsequent schedulin.g during the drought is based upon the estimated daily ET and the amount of water added in the last irrigation. For example, assume an irrigation or 1.7 em applied to a crop in Gainesville on May 15. The estimated ET value from Table 1 is 0.411 cm/day. Thus, dividing 1. 7 by 0.411 we have rour daYB or May 19 before the second irrigation is scheduled. A rain [aU of about 0.5 cm would delay irrigation for a day. There is an upper limit characteristic of the soil type, for the ql\antlty of water which can be stored in the soil root zone. And, only II fraction of this amount can be transpired by the plant before tempo rary wilt develops. It is this fraction the usable soil water capacity, wit lell must serve as a maximum in the above calculations. New estimates art' IH'NINI throughout the season as root growth extends to increasing HO 1.1 depths. To obtain the estimate, observe the growing crop over a ])(>[" loJ or ra:1.nless days following a rainfall which established a wet 80 1.1. pr.oLlle throughout the root zone. When the plants begin to show stress (temporary Wilting) by mid afternoon or earlier in the day, it is time [01" the initial irrigation. Add up the days since rainfall and multipLy by the expected ET for that period to obtain the estimated usable water storage. To continue the above example, assume an elapse of 7 day!:! w:i thout rainfall prior to the May 15 irrigation. Seven days liuWH 0.411 cm equals 2.88 cm of maximum usable storage for the particu lar Ho11 and plant root depth. UHC of the maximum usable storage value in irrigation scheduling can be seen in a further step of the example above. First of all, the value is not needed as long as irrigation or rainfall does not completely restore to capacity the partially depleted soil profile. Assume, a rainfall of 1.6 cm one day after the May 15 irrigation (1.7 cm). The maximum usable storage has been restored since rainfall plus irrigation minus one day of ET equals 2.89 cm. 2.88 by 0.411 predicts an irrigation date (May 23) 7 days after the rainfall on the 16th. If the rainfall had been larger than 1.6 cm, the maximum usable storage value of 2.88 would still be used in the calculations. On the other hand, in both cases, the soil root zone would be restored to capacity, and one could begin anew to determine an initial irrigation by observation. lberc is a second point in the illustration just given -a justification for
PAGE 77

-72-rainfall of 1.6 cm did not produce deep seepage loss of water and nutrients. In actual practice, the irrigation scheduling plan can be aitered in line with keen and experienced observation of weather conditions, crop appearance and growth stage, and the soil water status. A probability of rain may justify a decision to delay a scheduled irrigation or to apply a smaller quantity if the irrigation system will permit. On the other hand, unusally hot days with low relative humidity will cause the actual ET to be higher than predicted and a shorter irrigation interval will be necessary. Additional help in using the irrigation scheduling plan can be obtained with so:il water measuring devices such as the tensiometer. Some current irrigation installations cannot be used effectively to apply the small amount per event on a frequent schedule. However, the basic scheduling principles of the plan can be helpful in getti.ng the most efficiency from the available installation or in making modlfications to it, and above all in planning of new installations. The above plan is a simple one, but it is workable and will lay the ground work for the more advanced irrigation systems and scheduling plans already 1n the research and development stages. C. Water Policy for Agriculture in Florida The findings from the current study as well as others in the broad field of soil-water-climate relationships can be integrated with i.n formation on the geology-soil-climate-hydrologic system in Florida to develop a physically-based philosphy on water-use policy in agricultural production. The current study revealed an important characteristic of lISC by plants. Crop yield increased linearly with increasing seasonal ET. Reduction in actual ET at one stage of crop growth cannot be recovered at a later stage and yield reduction occurs irreversibly. Consequently, water use is less efficient, and returns from other production inputs. (capital, fertilizer, pesticides, fuel, and labor) are reduced. The conclusion is that agricultural production must be based on a full utilization of water up to the maximum ET.demand of the atmosphere. The adoption of an agricultural water policy based on this management objective is compatible with Florida's unique soil-climate-hydrologic resource. The average annual rainfall in Florida ranges from about 132 to 165 cm, while potential ET has been estimated to range around 100 cm. Consequently, the excess of rainfall over actual ET recharges the water storage capacity of soils, aquifers, lakes, and streams, and maintains ..

PAGE 78

-73-a net outflow of water from the State through streams and underground seepage along the coast. The storage components of the hydrologic cycle are the source of water for municipal, industrial and agricultural use. Important differences in water requirement and use among the above users need to be recognized in water planning and policymaking. Much of municipal and industrial use is non-consumptive. It is not lost to the atmosphere, but disposal of the physically and/or chemically altered water is a necessary phase of the water-use operation. Multiple cycles of use before discharge reduces the overall quantities needed. Muni cipal and industrial activitie's are commonly so concentrated on land areas that the water requirement exceeds the rainfall input of those areas. The resultant imbalance in local hydrology must be eventually offset by water input from adjacent remote sources. On the other hand, agricultural water use is largely consumptive since most of it is transpired through plants and evaporated from soil. Water evapotranspired or consumptively used is finite in quantity, the upper limit being equal to the evaporative demand of the atmosphere. In Florida, this is less than average rainfall. Consequently, if only the consumptive use is recognized as an agricultural use, then agriculture uoes not cause an imbalance in the local hydrology. There is a seasonal crop production water need which turns out to be larger than consumptive-use need during that season. The reason for this is the uneven rainfall distribution on a soil root zone of limited water storage capacity. During droughts, the farmer must recall water from stored sources mentioned earlier in order to meet evaporative demands and to keep crops growing and producing efficiently. It is inevitable that not all rainfall and irrigation inputs will be used for evapotranspiration and storage in the soil; some will contribute to runoff and deep seepage. The latter quantities are not consumptive uses, do not represent a loss from the available water resource, and should not be so designated if assigned as an agricultural use. Nevertheless, the agricultural producer must do his part to minimize surface runoff and deep se.epage during the crop production season, since nonconsumptive uses represent increased production costs in pumped water not used by the crop and in fertilizers and pesticides lost by leaching. Crop production enterprises in Florida which utilize seepage irrigation require total water inputs greatly in excess of consumptiveuse or evapotranspiration needs. Consequently, alternative crop production and water management systems may be needed in some of these agricultural areas in order to meet the economic demands as well as to avoid undue disturbance of the hydrologic balance. It should be stated that the majority of the irrigated cropland in Florida is irrigated by overhead sprinklers. Moreover, on an annual basis, except for perennial crops, crop production water needs are usually less than potential

PAGE 79

-74-evapotranspiration because the land is fallow during part of the year. In fact, ET water needs of agriculture are less than for natural vegetation. In view of the characteristics of water use in agriculture and the findings of this study, the objectives of research and education in this field should be to develop crop production systems which maximize return from production factors other than water, but which incorporate a water management scheme which meets actual ET demands while minimizing the loss of water by runoff and drainage. The implication in terms of water policy are: 1. The Florida climate, characterized by rainfall in excess of ET, justifies a full use of ET for crop production, especially for annuals where the farmer can be considered to have accumulated a storedwater credit from earlier rainfall in excess of ET. Irrigation to maintain potential ET is the most efficient use of water. 2. In view of the economic and energy waste incurred by water deficit during crop growth, there should be no drought-triggered reduction in allocation of water to agricultural enterprises which use water within the bounds of average annual ET. Further justification in terms of food and fiber needs can be made on the basis that not all of agricultural cropland is under irrigation. Thus, production from irrigated farms will be needed to make up for production inefficiencies of non-irrigated farming.

PAGE 80

-75-LITERATURE CITED I. l\rtH'C, R. R., J. L. Chesness, T.C. Keisling, J. E. Pallas, Jr., 2. D. A. Smittle, J. R. Stansell, and A. W. Thomas. 1980. Irrigation of crops in the southeastern United States. USDA-SEA ARM-S-9. 54 pages. Hillel, D., and Y. Guron. tion rate and maize yield. 1973. Relation between evapotranspira Water Resources Research. 9:743-748. 3. N:lrLS, A. B. ]979. Influence of irrigation strategy on water moll' Hnd yields of corn and peanuts. Master of Science Thesis, UnJversity of Florida, Gainesville, Florida. 163 pages. PalJas, J. E., Jr., J. R. Stansell, and T. J. Koske. of drought on florunner peanuts. Agronomy Journal. 1979. Effects 71:853-858 '5. Parra, J. V. 1971. Effect of an underground asphalt barrier on hydraulic properties of Lakeland fine sand. Master of Science Thcsi.s, University of Florida, Gainesville, Florida. 81 pages. 6. 1{;lO, P. S. C., J. M. Davidson, and L. C. Hammond. 1976. Estima I-Ion 0 F nonrenctive and reactive solute front locations in soils. In: Residual Management by Land Disposal. Proc. Hazardous Waste Residue Symposium. Tucson, AZ. p. 235-242. 7. Rao, P. S. C., .T. M. Davidson, and R. E. Jessup. 1981. Simulation_ or nitrogen behavior in the root zone of cropped land areas receiv Ing nrgnnic wast('s. Tn: Simulation of Nitrogen Behavior in SoilPl
PAGE 81

-76-12. Robertson, W. K., 1. C. Hammond, J. T. Johnson, and G. M. Prine. 1979. Root distribution of corn, soybeans, peanuts, sorghum, and tobacco in fine sands. Soil and Crop Sci. Soc. Fla. Proc. 38: 54-59. 13. Robertson, W. K., L. C. Hammond, J. T. Johnson, and K. J. Boote. 1980. Effects of plant-water stress on root distribution of corn, soybeans, and peanuts in sandy soil. Agronomy Journal. 72:548-550. ]4. Saxena, G. K., L. C. Hammond, and W. K. Robertson. 1973. Effects of asphalt layers on corn and tomato root systems. Agronomy 65:191-194. 15. Skogerboe, G. V., J. W. H. Barrett, B. J. Treat, and D. B. McWhorter. 1979. Potential effects of irrigation practices ()n crop yields In Grand Valley. EPA-600/2-79-149. 193 pages. 16. Varnell, R. J., H. Mwandemere, W. K. Robertson, and K. J. Boott'. 1975. Peanut yields affected by soil water, no-till and gypsum. Soil and Crop Sci. Soc. Fla. Pioc. 35:56-59. ..

PAGE 82

-77ABSTRACTS OF PUBLISHED PAPERS 1. Rao, P. S. C., J. M. Davidson, and L. C. 1976. Estimation of nonreactive and reactive solute front locations in soils. In: Residual Management by Land Disposal. Proc. Hazardous Waste Resl.due Symposium. Tucson, AZ p. 235-242. A technique, based on the physical principles of water and solute transport, was used to describe the position of nonreactive and/or reactive solute fronts in a soil profile. The procedure estimates the solute front location after infiltration and redistribution of the soil water to "field capacity," and includes extraction of soil water by plant roots between irrigation/rainfall events. Linear equilibrium adsorption-desorption of the reactive solutes was assumed. The approximation procedure was based on the principles that (i) the soil water residing in all pore sequences participates in the transport processes, and (ii) the soil water initially present in the profile is completely displaced ahead of the water entering at the soil surface. An analysis of published field and laboratory data on infiltration of nonreactive solutes (Cland N03-) indicated that these assumptions were valid. Agreemellt between predicted solute front location using a sophisticated transien.t flow models and the above procedure further !Hlpport the validity of the assumptions. Field data for chloride II\OVllllIlmt 1n a Handy Boil, in the presence of a fully established millet c.rop, during a 60-day were in agreement with the simplified model. The major drawback of the present technique is in its failure to describe the attenuation or spreading of a solute pulse as it is leached through the soil profile. 2. Selim, H. M., L. C. Hammond, and R. S. Mansell. 1977. Soil water movement and uptake by plants during water infiltration and redistribution. Soil and Crop Sci. Soc. Fla. Proc. 36:101-107. A numerical solution of the equation governing soil water movement and uptake by plants during infiltration and redistribution was used to investigate the influence of the amount of irrigation and soil nonuniformity on deep seepage loss and soil water storage in the root zone. For a uniform Lakeland sandy soil the deep seepage loss increased as the amount of water applied in an irrigation event increased. However, the presence of a low permeability lower layer in a two-layered soil profile was beneficial in increasing soil water storage and in minimizing deep seepage loss. Such a decrease in deep seepage loss was more pronounced when large amounts of irrigation were applied and the hydraulic conductivity of the lower layer was extremely small. For small irrigations, the influence of a low permeability lower layer on deep seepage loss was essentially negligible.

PAGE 83

-78-3. Rhoads, F. M., R. S. Mansell, and L. C. Hanunond. 1978. Influ(;!nc(;! of water and fertilizer management on yield and water-input efficiency of corn. Agronomy Journal. 70:305-308. Damaging plant water stress develops in corn grown on coarsetextured, low water retaining soil of the southeastern U.S. during 1 to 2-week periods without rainfall. However, in most years rainstorms cause leaching of soluble fertilizers from the root zone. This study was conducted to evaluate efficiency of water input in terms of corn grain yield per unit of water, with two fertilizer systems on a Troup loamy sand (Grossarenic Paleudult). Water management consisted of (a) control--natural rainfall only,. (b) trickle irrigation scheduled da:U y (0.64 cm/day), and (c) trickle irrigation scheduled by tensiometer (1.30 cm/application). Tensiometers were placed in each treatment at six depths between 15 and 150 cm below the soil surface and readings recorded daily. Methods of applying fertilizers were (a) conventional--l/3 of N and all F and K applied broadcast preplant, and remainder of N applied in two sidedressings; (b) program fertUizat.lonN-P-K applied broadcast in small increments (5, 5, 10, 20, 20, 20, and 20%) at 2-week intervals after corn emerged. Average grain yields for the above water management treatments were 2,790, 4,160, and .5,700 kg/hu respectively. Conventional fertilization had an 8V(;!rag<.! grain y.leld of 3,680 kg/ha and program fertilization 4,760 kg/ha. Water-input l'[[f ciencies based on grain yields and total water input were 57, 42, anci 7b kg/ha/cm for no irrigation, daily irrigation, and tensiometer scheduled irrigation respectively. Highest irrigation water-input efficiency (150 kg/ha/cm) occurred with program fertilization and tensiometer scheduled irrigation. Irrigtion water-input efficiency was lowest (10 kg/ha/cm) with corn receiving daily irrigation and conventional fertilization. 4. Robertson, W. K., L. C. Hammond, J. T. Johnson, and G. M. Prine. 1978. Root of corn, soybeans, peanuts, sorghum, and tobacco in fine sands. Soil and Crop Science Soc. of Fla. Froc. 38:54-59. A knowledge of root distribution of plants contributes to d.ecisions on how to fertilize, irrigate, select cultivarf,;, and t.ill till' soil more effectively. In this paper, we report root patterns to a depth of 150 cm for corn (Zea mays L.) on Lakeland fs, a thermic, coated, Typic Quartzipsanunent; soybeans (Glycine (L.) Merr.) on Kendrick f s, a loamy, siliceous, hyperthermic, Arenic Pul(;!udul t; ts (Arachis hypogaea L.) on Lake fs, a hyperthermic, coated, Typic Quartz ipsamment; sorghum (Sorghum bicolor L. Moench) on Arredondo f s, a loamy. siliceous, hyperthermic, Grossarenic Paleudult; and tobacco tabacum L.) on Lakeland fs, a thermic, coated, Typic Quartzipsamment. The line intercept technique was used to determine root density in ').0 cm diameter soil cores. Lakeland fs was located on the Agricultural Research Center near Live Oak and the remaining soils were on the

PAGE 84

" -79-Agricultural Experiment Station farm, Gainesville. All soils were fine sands; however, depth to clay for Kendrick and Arredondo was shallower (120 to 150 em) than for the remaining soils. The latter two soils had hlgber AI contents associated with the clay but it was not believed to ue at a high enough level to be toxic to plants. Tlw cropt:; ranked in order of root length density as follows: soybe.ans > corn> tobacco> peanuts> sorghum; but on a single plant basis the ordar was: tobacco> corn> soybeans> peanuts> sorghum. On the basis of length per unit weight of root the order of crops was: peanuts> soybeans> corn = tobacco ... sorghum. In most instances, root size was greater and root length density less in the area of the plow sole at 3045 em. The roots of corn, soybeans, and sorghum were finer, tobacco coarser, and peanuts the same size in the middle between the rows as compared to under the row. 5. Robertson, W. K., L. C. Hammond, J. T. Johnson, and K. J. Boote. 1980. Effects of plant-water stress on root distribution of corn, soybeans, and peanuts in sandy soil. Agronomy Journal. 72:548-550. The efficient recovery by crops of added nutrients and water is influenced by plant rooting characteristics. Published data on the relutionship of irrigation water and root distribution of certain crops grown all sundy soils in humid regions are limited and needed. This or study was a part of an investigation on three soil types to untl!rmlne the effect of plant-water stress and irrigation on root distrJbut:ion of corn JZea mays L.), soybeans [Glycine (L.) Merr.], and p(.'anuts (Arachis hypogaea L.). The basic irrigation plan was to replenish the water deficit in only the top 30 to 60 cm of the soil profile. Depth of wetting and the degree of soil water depletion below the irri gated soil layer varied with irrigation frequency and amount of water per application. Although treatments were not the same in the four experiments reported, the major part of the study involved four water management treatments: 1) no irrigation; 2) light, infrequent irrigation; 3) light, frequent irrigation; and 4) medium, infrequent irrigation. Seed yields were obtained at maturity and root length measurements were made at full canopy. Root lengths per unit volume of soil were measured by the line intercept method. Yields of corn. and peanuts increased with total amount of irrigation water used, but there was no yield response to irrigation in the soybean experiment. Peanut and soybean root growth (root length per unit area to a depth of 150 cm) was not affected by water management. In the corn experiments, irrigation increased the length of roots in the 150 em soil profile. The largest root length value was found in the light, infrequent irrigation treatment. Crops vary in rooting response to plant-water stress and irrigation strategy. Limited rooting of corn under stress very likely decreases the efficiency of water and fertilizer use.

PAGE 85

-80-APPENDIX TABLE Soil water characteristic data used in water balance simulatiolls. 11 Soil type-Lakeland, f. s. Lakeland, f.s. with barrier Arredondo, f.s. Lake, f.s. Kend.rick, f. s. Depth cm 0-32 32-92 92-152 152-180 0-8 8-16 16-32 32-40 44-52 52-64 0-32 32-00 hO-I32 132L4H 148-168 0-28 28-180 180-200 0-16 16-32 32-128 128-136 136-148 -------.Field capacity 15-h;lr ----_._--cm3/cm3 cm"lcm 3 0.100 0.025 0.085 0.022 0.075 0.022 0.065 0.022 0.100 0.025 0.120 0.025 0.120 0.025 0.150 0.022 0.200 0.022 0.220 o.on 0.200 o.on 0.100 0.0:)', O.OH,) o.n:):) 0.070 O.O:)() O.OR,) O.O:W (). I 'j () o.():!() 0.075 D.On 0.065 0.022 0.065 0.022 0.090 0.022 0.080 0.022 0.075 0.020 0.110 0.025 0.210 0.100 1Lake1and soil located at Live Oak; others locatt'd ilt (;;IiIH'Svllll'. Abbreviation f.s. means fine sand. .!